ENCYCLOPEDIA OF
Mathematics
James Tanton, Ph.D.
Encyclopedia of Mathematics Copyright © 2005 by James Tanton, Ph.D...
1735 downloads
4776 Views
7MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
ENCYCLOPEDIA OF
Mathematics
James Tanton, Ph.D.
Encyclopedia of Mathematics Copyright © 2005 by James Tanton, Ph.D. All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage or retrieval systems, without permission in writing from the publisher. For information contact: Facts On File, Inc. 132 West 31st Street New York NY 10001 Library of Congress CataloginginPublication Data Tanton, James Stuart, 1966– Encyclopedia of mathematics/James Tanton. p. cm. Includes bibliographical references and index. ISBN 0816051240 1. Mathematics—Encyclopedia. I. Title. QA5.T34 2005 510 ′.3—dc22 2004016785 Facts On File books are available at special discounts when purchased in bulk quantities for businesses, associations, institutions, or sales promotions. Please call our Special Sales Department in New York at (212) 9678800 or (800) 3228755. You can find Facts On File on the World Wide Web at http://www.factsonfile.com Text design by Joan M. Toro Cover design by Cathy Rincon Illustrations by Richard Garratt Printed in the United States of America VB Hermitage 10 9 8 7 6 5 4 3 2 1 This book is printed on acidfree paper.
CONTENTS Acknowledgments v Introduction vi A to Z Entries 1 Feature Essays: “History of Equations and Algebra”
9 “History of Calculus”
57 “History of Functions”
208 “History of Geometry”
226 “History of Probability and Statistics”
414 “History of Trigonometry”
510 Appendixes: Appendix I Chronology 539
Appendix II Bibliography and Web Resources 546 Appendix III Associations 551 Index 552
ACKNOWLEDGMENTS My thanks to James Elkins for reading a substantial portion of the manuscript, for his invaluable comments, and for shaping my ideas in writing a number of specific entries. Thanks also go to Frank K. Darmstadt, executive editor at Facts On File, for his patience and encouragement, and to Jodie Rhodes, literary agent, for encouraging me to pursue this project. I also wish to thank Tucker McElroy and John Tabak for taking the time to offer advice on finding archives of historical photographs, and the staff of The Image Works for their work in finding photographs and granting permission to use them. But most of all, thanks to Lindy and Turner for their love and support, always.
v
INTRODUCTION Mathematics is often presented as a large collection of disparate facts to be absorbed (memorized!) and used only with very specific applications in mind. Yet the development of mathematics has been a journey that has engaged the human mind and spirit for thousands of years, offering joy, play, and creative invention. The Pythagorean theorem, for instance, although likely first developed for practical needs, provided great intellectual interest to Babylonian scholars of 2000 B.C.E., who hunted for extraordinarily large multidigit numbers satisfying the famous relation a2 + b2 = c2. Ancient Chinese scholars took joy in arranging numbers in square grids to create the first “magic squares,” and Renaissance scholars in Europe sought to find a formula for the prime numbers, even though no practical application was in mind. Each of these ideas spurred further questions and further developments in mathematics—the general study of Diophantine equations, semimagic squares and Latin squares, and publickey cryptography, for instance—again, both with and without practical application in mind. Most every concept presented to students today has a historical place and conceptual context that is rich and meaningful. The aim of Facts On File’s Encyclopedia of Mathematics is to unite disparate ideas and provide a sense of meaning and context. Thanks to the encyclopedic format, all readers can quickly find straightforward answers to questions that seem to trouble students and teachers alike: • • • • • • • •
Why is the product of two negative numbers positive? What is π, and why is the value of this number the same for all circles? What is the value of π for a shape different than a circle? Is every number a fraction? Why does the longdivision algorithm work? Why is dividing by a fraction the same as multiplying by its reciprocal? What is the value of ii? What is the fourth dimension?
vi
Introduction vii This text also goes further and presents proofs for many of the results discussed. For instance, the reader can find, under the relevant entries, a proof to the fundamental theorem of algebra, a proof of Descartes’s law of signs, a proof that every number has a unique prime factorization, a proof of Bretschneider’s formula (generalizing Brahmagupta’s famous formula), and a derivation of Heron’s formula. Such material is rarely presented in standard mathematical textbooks. In those instances where the method of proof is beyond the scope of the text, a discussion as to the methods behind the proof is at least offered. (For instance, an argument is presented to show how a formula similar to Stirling’s formula can be obtained, and the discussion of the CayleyHamilton theorem shows that every matrix satisfies at least some polynomial equation.) This encyclopedia aims to be satisfying to those at all levels of interest. Each entry contains crossreferences to other items, providing the opportunity to explore further context and related ideas. The reader is encouraged to browse. As a researcher, author, and educator in mathematics, I have always striven to share with my students the sense of joy and enthusiasm I experience in thinking about and doing mathematics. Collating, organizing, and describing the concepts a highschool student or beginning collegelevel student is likely to encounter in the typical mathematics curriculum, although a daunting pursuit, has proved to be immensely satisfying. I have enjoyed the opportunity to convey through the writing of this text, hopefully successfully, a continued sense of joy and delight in what mathematics can offer. Sadly, mathematics suffers from the ingrained perception that primary and secondary education of the subject should consist almost exclusively of an acquisition of a set of skills that will prove to be useful to students in their later careers. With the push for standardized testing in the public school system, this mindset is only reinforced, and I personally fear that the joy of deep understanding of the subject and the sense of play with the ideas it contains is diminishing. For example, it may seem exciting that we can produce students who can compute 584 × 379 in a flash, but I am saddened with the idea that such a student is not encouraged to consider why we are sure that 379 × 584 will produce the same answer. For those students that may be naturally inclined to pause to consider this, I also worry about the response an educator would give upon receiving such a query. Is every teacher able to provide for a student an example of a system of arithmetic for which it is no longer possible to assume that a×b and b×a are always the same and lead a student through a path of creative discovery in the study of such a system? (As physicists and mathematicians have discovered, such systems do exist.) By exploring fundamental questions that challenge basic assumptions, one discovers deeper understanding of concepts and finds a level of creative play that is far more satisfying than the performance of rote computation. Students encouraged to think this way have learned to be adaptable, not only to understand and apply the principles of a concept to the topic at hand, but also to apply those foundations and habits of mind to new situations that may arise. After all, with the current
viii Introduction advances of technology in our society today, we cannot be sure that the rote skillsets we deem of value today will be relevant to the situations and environments students will face in their future careers. We need to teach our students to be reflective, to be flexible, and to have the confidence to adapt to new contexts and new situations. I hope that this text, in some small ways, offers a sense of the creative aspect to mathematical thinking and does indeed gently encourage the reader to think deeply about concepts, even familiar ones. Encyclopedia of Mathematics contains more than 800 entries arranged in alphabetical order. The aim of the historical notes, culturespecific articles, and the biographical portraits included as entries, apart from providing historical context, is to bring a sense of the joy that mathematics has brought people in the past. The back matter of this text contains a timeline listing major accomplishments throughout the historical development of mathematics, a list of current mathematics organizations of interest to students and teachers, and a bibliography.
A AAA/AAS/ASA/SAS/SSS Many arguments and proofs
the AAS and ASA rules apply. This is sometimes called the “HA congruence criterion” for right triangles.)
presented in the study of GEOMETRY rely on identifying similar triangles. The SECANT theorem, for instance, illustrates this. Fortunately, there are a number of geometric tests useful for determining whether or not two different triangles are similar or congruent. The names for these rules are acronyms, with the letter A standing for the word angle, and the letter S for the word side. We list the rules here with an indication of their proofs making use of the LAW OF SINES and the LAW OF COSINES:
c. The SAS rule: If two triangles have two sides of matching lengths with matching included angle, then the two triangles are congruent. The law of cosines ensures that the third sidelengths of each triangle are the same, and that all remaining angles in the triangles match. By the AAS and ASA rules, the triangles are thus congruent. As an application of this rule, we prove EUCLID’s isosceles triangle theorem:
a. The AAA rule: If the three interior angles of one triangle match the three interior angles of a second triangle, then the two triangles are similar.
The base angles of an isosceles triangle are equal. Suppose ABC is a triangle with sides AB and AC equal in length. Think of this triangle as representing two triangles: one that reads BAC and the other as CAB. These two triangles have two matching sidelengths with matching included angles, and so, by the SAS rule, are congruent. In particular, all corresponding angles are equal. Thus the angle at vertex B of the first triangle has the same measure as the corresponding angle of the second triangle, namely, the angle at vertex C.
The law of sines ensures that pairs of corresponding sides of the triangles have lengths in the same ratio. Also note, as the sum of the interior angles of any triangle is 180°, one need only check that two corresponding pairs of interior angles from the triangles match. b. The AAS and ASA rules: If two interior angles and one sidelength of one triangle match corresponding interior angles and sidelength of a second triangle, then the two triangles are congruent. By the AAA rule the two triangles are similar. Since a pair of corresponding sidelengths match, the two triangles are similar with scale factor one, and are hence congruent. (Note that any two right triangles sharing a common hypotenuse and containing a common acute angle are congruent: all three interior angles match, and
This result appears as Proposition 5 of Book I of Euclid’s famous work THE ELEMENTS. d. The SSS rule: If the three sidelengths of one triangle match the three sidelengths of a second triangle, then the two triangles are congruent. 1
2
abacus
The law of sines ensures that all three interior angles match, and so the SAS rule applies. EUCLIDEAN GEOMETRY takes the SAS rule as an AXIOM, that is, a basic assumption that does not require proof. It is then possible to justify the validity of the remaining rules by making use of this rule solely, and to also justify the law of sines. (The fact that the sine of an angle is the same for all right triangles containing that angle relies on SAS being true.) See also CONGRUENT FIGURES; SIMILAR FIGURES.
abacus Any counting board with beads laid in parallel grooves, or strung on parallel rods. Typically each bead represents a counting unit, and each groove a place value. Such simple devices can be powerful aids in performing arithmetic computations. The fingers on each hand provide the simplest “set of beads” for manual counting, and the sand at one’s feet an obvious place for writing results. It is not surprising then that every known culture from the time of antiquity developed, independently, some form of counting board to assist complex arithmetical computations. Early boards were simple sunbaked clay tablets, coated with a thin layer of fine sand in which symbols and marks were traced. The Greeks used trays made of marble, and the Romans trays of bronze, and both recorded counting units with pebbles or beads. The Romans were the first to provide grooves to represent fixed placevalues, an innovation that proved to be extremely useful. Boards of this type remained the stan
A simple abacus
A Chinese abacus from before 1600. Notice that two beads, each representing five units, are placed in each column above the bar. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
dard tool of European merchants and businessmen up through the Renaissance. The origin of the word abacus can be traced back to the Arabic word abq for “dust” or “fine sand.” The Greeks used the word abax for “sand tray,” and the Romans adopted the word abacus. The form of the abacus we know today was developed in the 11th century in China and, later, in the 14th century in Japan. (There the device was called a soroban.) It has beads strung on wires mounted in a wooden frame, with five beads per wire that can be pushed up or down. Four beads are used to count the units one through four, and the fifth bead, painted a different color or separated by a bar, represents a group of five. This provides the means to represent all digits from zero to nine. Each wire itself represents a different power of ten. The diagram at left depicts the number 35,078. Addition is performed by sliding beads upward (“carrying digits” as needed when values greater than 10 occur on a single wire), and subtraction by sliding beads downward. Multiplication and division can be computed as repeated addition and subtraction. Historians have discovered that the Chinese and Japanese scholars also devised effective techniques for computing square and cube roots with the aid of the abacus. The abacus is still the popular tool of choice in many Asian countries—preferred even over electronic
absolute convergence calculators. It is a useful teaching device to introduce young children to the notion of placevalue and to the operations of basic arithmetic. See also BASE OF A NUMBER SYSTEM; NAPIER’S BONES.
Abel, Niels Henrik (1802–1829) Norwegian Algebra Born on August 5, 1802, Niels Abel might have been one of the great mathematicians of the 19th century had he not died of tuberculosis at age 26. He is remembered, and honored, in mathematics for putting an end to the threecenturylong search for a SOLUTION BY RADICALS of the quintic equation. His theoretical work in the topics of GROUP THEORY and ALGEBRA paved the way for continued significant research in these areas. Abel’s short life was dominated by poverty, chiefly due to the severe economic hardships his homeland of Norway endured after the Napoleonic wars, exacerbated by difficult family circumstances. A schoolteacher, thankfully, recognized Abel’s talent for mathematics as a young student and introduced him to the works of LEONHARD EULER, JOSEPHLOUIS LAGRANGE, and other great mathematicians. He also helped raise money to have Abel attend university and continue his studies. Abel entered the University of Christiania in the city of Christiania (presentday Oslo), Norway, in 1821. During his final year of study, Abel began working on the solution of quintic equations (fifthdegree polynomial equations) by radicals. Although scholars for a long time knew general formulae for solving for QUADRATIC, cubic, and QUARTIC EQUATIONs using nothing more than basic arithmetical operations on the COEFFICIENTs that appear in the equation, no one had yet found a similar formula for solving quintics. In 1822 Abel believed he had produced one. He shared the details of his method with the Danish mathematician Ferdinand Degen in hopes of having the work published by the Royal Society of Copenhagen. Degen had trouble following the logic behind Abel’s approach and asked for a numerical illustration of his method. While trying to produce a numerical example, Abel found an error in his paper that eventually led him to understand the reason why general solutions to fifth and higherdegree equations are impossible. Abel published this phenomenal discovery in 1825 in a selfpublished pamphlet “Mémoire sur les équations algébriques où on démontre l’impossibilité de la résolution de l’équation générale du
3
cinquième degré” (Memoir on the impossibility of algebraic solutions to the general equations of the fifth degree), which he later presented as a series of seven papers in the newly established Journal for Pure and Applied Mathematics (commonly known as Crelle’s Journal for its German founder August Leopold Crelle). At first, reaction to this work was slow, but as the reputation of the journal grew, more and more scholars took note of the paper, and news of Abel’s accomplishment began to spread across Europe. A few years later Abel was honored with a professorship at the University of Berlin. Unfortunately, Abel had contracted tuberculosis by this time, and he died on April 6, 1829, a few days before receiving the letter of notification. In 1830 the Paris Academy awarded Abel, posthumously, the Grand Prix for his outstanding work. Although Abel did not write in terms of the modernday concepts of group theory, mathematicians call groups satisfying the COMMUTATIVE PROPERTY “Abelian groups” in his honor. In 2002, on the bicentenary of his birth, the Norwegian Academy of Science and Letters created a new mathematics prize, the Abel Prize, similar to the Nobel Prize, to be awarded annually. Research in the field of commutative algebra continues today using the approach developed by Abel during his short life. His influence on the development of ABSTRACT ALGEBRA is truly significant.
∞
absolute convergence A
SERIES
∑ an
containing
n=1
positive and negative terms is said to converge absolutely if the corresponding series with all terms ∞
made positive,
∑ an , converges. For example, the series
n=1
1−
1 1 1 1 1 + − + − + L converges absolutely because 2 4 8 16 32
the corresponding series 1 + 1 + 1 + 1 + 1 + 1 + L 2 4 8 16 32 converges. (See CONVERGENT SERIES.) The “absolute convergence test” reads: ∞
If
∑ an
∞
converges, then the original series
n=1
∑ an
n=1
also converges. It can be proved as follows:
4 absolute value Let pn =  an  – an. Then each value pn is either zero or equal to 2  an , depending on whether an is positive or negative. In particular we have that 0 ≤ pn ≤ 2 an . ∞
2 ln 2 = (2 − 1) −
L = 1−
∞
Consequently, 0 ≤ ∑ pn ≤ 2 ∑ an n =1 n =1 the
and so, by
1 ⎛ 2 1⎞ 1 ⎛ 2 1 ⎞ + ⎜ − ⎟ − + ⎜ − ⎟ +L 2 ⎝ 3 3⎠ 4 ⎝ 5 5⎠ 1 1 1 1 + − + − L = ln 2 2 3 4 5
and so 2 = 1.
∞
COMPARISON TEST,
so does
∑ pn
converges. Conse
n=1
quently ∞
∞
n =1
n =1
∑ an = ∑ ( an
)
− pn =
∞
∞
n =1
n =1
∑ an − ∑ pn .
This test does not cover all cases, however. It is still ∞
possible that a series
∑ an
Paradoxes like these show that it is not permissible to rearrange the order of terms of a conditionally convergent series. Mathematicians have shown, however, that rearranging the terms of an absolutely convergent series is valid. See also ABSOLUTE VALUE.
containing positive and
n=1
∞
negative terms might converge even though
∑ an
n=1
does not. For example, the alternating SERIES
1−
1 1 1 1 1 + − + − +L 2 3 4 5 6
HARMONIC
converges,
yet
1 1 1 1 1 + + + + + L does not. A series that 2 3 4 5 6 converges “on the condition that the negative signs 1+
∞
remain present,” that is, one for which
∑ an
n=1 ∞
converges but
∑ an
does not, is called “condition
n=1
ally convergent.” Manipulating conditionally convergent series can lead to all sorts of paradoxes. For example, the following argument “proves” that 1 = 2: Start with the observation that: 1−
1 1 1 1 1 + − + − + L = ln 2 ≈ 0.69 2 3 4 5 6
(This follows from the study of the harmonic series or from MERCATOR’S EXPANSION.) Consequently: ⎛ ⎞ 1 1 1 1 1 1 1 2 ln 2 = 2⎜1 − + − + − + − + L⎟ 2 3 4 5 6 7 8 ⎝ ⎠ = 2 −1+
2 1 2 1 2 1 − + − + − +L 3 2 5 3 7 4
Collecting terms with a common denominator gives:
absolute value (modulus) Loosely speaking, the absolute value of a REAL NUMBER is the “positive version of that number.” Vertical bars are used to denote the absolute value of a number. For example, the absolute value of negative three is  –3 = 3, and the absolute value of four is  4  = 4. The absolute value of a real number a is typically envisioned three ways: 1.  a equals a itself if a is positive or zero, and equals –a if it is negative. (For example,  –3 = –(–3) = 3 and  3 = 3.) 2.  a equals the positive square root of a2. (For example, −3 = (−3)2 = 9 = 3 .) 3.  a is the distance between the points a and 0 on the real number line. (For example,  –3 = 3 =  3 since both –3 and 3 are three units from the origin.) More generally, if a and b are two points on the number line, then the distance between them on the number line is given by  a – b . (For example, the points 4 and –7 are  4–(–7) =  4 + 7 = 11 units apart.) By examining each of the cases with a and b positive or negative, one can check that the absolute value function satisfies the following properties: i.  a + b ≤  a +  b ii.  a – b ≤  a +  b iii.  a · b =  a ·  b Knowing the absolute value of a quantity determines the value of that quantity up to sign. For example, the equation  x+2 = 5 tells us that either x + 2 = 5
acute angle or x + 2 = –5, and so x equals either 3 or –7. Alternatively, one can read the equation as  x – (–2) = 5, interpreting it to mean that x is a point a distance of five units from –2. Five units to the left means x is the point –7; five units to the right means x is 3. The notion of absolute value was not made explicit until the mid1800s. KARL WEIERSTRASS, in 1841, was the first to suggest a notation for it—the two vertical bars we use today. Matters are currently a little confusing, however, for mathematicians today also use this notation for the length of a VECTOR and for the MODULUS of a COMPLEX NUMBER.
abstract algebra Research in pure mathematics is motivated by one fundamental question: what makes mathematics work the way it does? For example, to a mathematician, the question, “What is 263 × 178 (or equivalently, 178 × 263)?” is of little interest. A far more important question would be, “Why should the answers to 263 × 178 and 178 × 263 be the same?” The topic of abstract algebra attempts to identify the key features that make ALGEBRA and ARITHMETIC work the way they do. For example, mathematicians have shown that the operation of ADDITION satisfies five basic principles, and that all other results about the nature of addition follow from these. 1. Closure: The sum of two numbers is again a number. 2. Associativity: For all numbers a, b, and c, we have: (a + b) + c = a + (b + c). 3. Zero element: There is a number, denoted “0,” so that: a + 0 = a = 0 + a for all numbers a. 4. Inverse: For each number a there is another number, denoted “–a,” so that: a + (–a) = 0 = (–a) + a. 5. Commutativity: For all numbers a and b we have: a + b = b + a. Having identified these five properties, mathematicians search for other mathematical systems that may satisfy the same five relations. Any fact that is known about addition will consequently hold true in the new system as well. This is a powerful approach to matters. It avoids having to reprove THEOREMS and facts about a new system if one can recognize it as a familiar one in disguise. For example, MULTIPLICATION essentially satisfies the same five AXIOMS as above, and so for any fact about addition, there is a corresponding fact about multiplication. The set of symmetries of a geometric
5
figure also satisfy these five axioms, and so too all known results about addition immediately transfer to interesting statements about geometry. Any system that satisfies these basic five axioms is called an “Abelian group,” or just a GROUP if the fifth axiom fails. GROUP THEORY is the study of all the results that follow from these basic five axioms without reference to a particular mathematical system. The study of RINGs and FIELDs considers mathematical systems that permit two fundamental operations (typically called addition and multiplication). Allowing for the additional operation of scalar multiplication leads to a study of VECTOR SPACEs. The theory of algebraic structures is highly developed. The study of vector spaces, for example, is so extensive that the topic is regarded as a field of mathematics in its own right and is called LINEAR ALGEBRA.
acceleration See VELOCITY.
actuarial science The statistical study of life expectancy, sickness, retirement, and accident matters is called actuarial science. Experts in the field are called actuaries and are employed by insurance companies and pension funds to calculate risks and relate them to the premiums to be charged. British mathematician and astronomer, Edmund Halley (1656–1742) was the first to properly analyze annuities and is today considered the founder of the field. See also LIFE TABLES.
acute angle An ANGLE between zero and 90° is called an acute angle. An acuteangled triangle is one whose angles are all acute. According to the LAW OF COSINES, a triangle with sidelengths a, b, and c and corresponding angles A, B, C opposite those sides, satisfies: cosC =
a 2 + b2 − c 2 2ab
The angle C is acute only if cosC > 0, that is, only if a2 + b2 > c2. Thus a triangle a, b, c is acute if, and only if, the following three inequalities hold:
6 addition a2 + b2 > c2 b2 + c2 > a2 c2 + a2 > b2 See also
OBTUSE ANGLE; PERIGON;
PYTHAGORAS’S
THEOREM; TRIANGLE.
addition The process of finding the sum of two numbers is called addition. In the elementary ARITHMETIC of whole numbers, addition can be regarded as the process of accumulating sets of objects. For example, if a set of three apples is combined with a set of five apples, then the result is a set of eight apples. We write: 3 + 5 = 8. Two numbers that are added together are called addends. For instance, in the equation 17 + 33 = 50, the numbers 17 and 33 are the addends, and the number 50 is their sum. Addition can also be regarded as the process of increasing one number (an addend) by another (called, in this context, an augend). Thus when 17 is augmented by 33 units, the result is 50. The PLACEVALUE SYSTEM we use today for writing numbers simplifies the process of adding large integers. For instance, adding together 253 and 589 yields 2 + 5 = 7 units of 100, 5 + 8 = 13 units of 10, and 3 + 9 = 12 units of 1. So, in some sense, it is reasonable to write the answer to this addition problem simply as 7  13  12 using a vertical bar to separate units of powers of 10. Since 13 units of 10 is equivalent to one unit of 100 and three units of 10, this is equivalent to 8  3  12. Noting, also, that 12 units of one 12 is equivalent to one unit of 10 and two single units, this can be rewritten as 8  4  2. Thus we have: 253 + 589 = 842. The latter process of modifying the figures into singledigit powers of 10 (that is, in our example, the process of rewriting 7  13  12 as 8  4  2) is called “carrying digits.” Students in schools are usually taught an algorithm that has one carry digits early in the process of completing an addition problem rather than leaving this work as the final step. Either method is valid. (The term “carry a digit” dates back to the time of the ABACUS, where beads on rods represented counts of powers of 10 and the person had to move— “carry”—counters from one rod to another if any count was greater than a single digit.) The process of addition can be extended to NEGATIVE NUMBERS (yielding an operation called SUBTRAC
TION), the addition of FRACTIONs (completed with the aid of computing COMMON DENOMINATORs), REAL NUMBERS, COMPLEX NUMBERS, VECTORs, and MATRIX addition. The number ZERO is an additive IDENTITY ELEMENT in the theory of arithmetic. We have that a + 0 = a = 0 + a for any number a. The sum of two realvalued functions f and g is the function f + g whose value at any input x is the sum of the outputs of f and g at that input value: (f + g)(x) = f(x) + g(x). For example, if f(x) = x2 + 2x and g(x) = 5x + 7, then (f + g)(x) = x2 + 2x + 5x + 7 = x2 + 7x + 7. A function with the property that f(x + y) = f(x) + f(y) for all inputs x and y is called “additive.” For example, f(x) = 2x is additive. The addition formulae in TRIGONOMETRY assert:
sin(x + y) = sin x cos y + cos x sin y cos(x + y) = cos x cos y − sin x sin y tan x + tan y tan(x + y) = 1 − tan x tan y The symbol + used to denote addition is believed to have derived from a popular shorthand for the Latin word et meaning “and” and was widely used by mathematical scholars in the late 15th century. The symbol first appeared in print in Johannes Widman’s 1489 book Behennde unnd hüpsche Rechnung auf fallen Kauffmannschaften (Neat and handy calculations for all tradesmen). See also ASSOCIATIVE ; CASTING OUT NINES ; COMMUTATIVE PROPERTY; DISTRIBUTIVE PROPERTY; MULTIPLICATION; SUMMATION.
affine geometry The study of those properties of geometric figures that remain unchanged by an AFFINE is called affine geometry. For example, since an affine transformation preserves straight lines and RATIOs of distances between POINTs, the notions of PARALLEL lines, MIDPOINTs of LINE segments, and tangency are valid concepts in affine geometry. The notion of a CIRCLE, however, is not. (A circle can be transformed into an ELLIPSE via an affine transformation. The equidistance of points on the circle from the circle center need not be preserved.) Affine geometry was first studied by Swiss mathematician LEONHARD EULER (1707–83). Only postulates TRANSFORMATION
Agnesi, Maria Gaëtana 7 1, 2, and 5 of EUCLID’S affine geometry.
POSTULATES
remain valid in
affine transformation Any map from the
PLANE to itself that transforms straight LINES into straight lines and preserves RATIOs of distances between POINTs (so that the midpoint of a line segment, for instance, remains the midpoint after the transformation) is called an affine transformation. One can prove that any affine transformation must be a LINEAR TRANSFORMATION followed by a translation. Thus an affine transformation T is completely specified by a MATRIX A and a VECTOR b so that T (x) = Ax + b for any vector x representing a point in the plane. An affine transformation T satisfies the relation:
T(sx + ty) = sT(x) + tT(y) for any two vectors x and y and any two real numbers s and t such that s + t = 1. This is sometimes taken as the definition of what it means for a transformation to be affine. Affine transformations generally do not preserve the lengths of line segments nor the measure of ANGLEs between segments. It is possible to transform a CIRCLE into an ELLIPSE, for instance, via an affine transformation. See also AFFINE GEOMETRY.
Agnesi, Maria Gaëtana (1718–1799) Italian Calculus Born on May 16, 1718, to a wealthy family of silk merchants, Maria Agnesi is best remembered for her influential expository text outlining the methods and techniques of the newly invented CALCULUS. Written with such clarity and precision, Istituzioni analitiche (Analytical institutions) garnered her international fame. Agnesi is considered the first major female mathematician of modern times, and she holds the distinction of being the first woman to be awarded a professorship of mathematics on a university faculty. Agnesi demonstrated remarkable academic talents as a young child. By age 13 she had mastered many languages and had published translations of academic essays. Although little consideration was given to educating women at the beginning of the 18th century, Agnesi’s father encouraged her intellectual develop
ment, provided tutors of the highest quality, and provided forums for her to display her talents to Italian society. In preparation for these events, Agnesi had prepared discourses on a wide variety of topics in science and philosophy, which she published as a collection of 190 essays at age 20. After the death of her mother, Agnesi undertook the task of instructing her younger brothers in the subject of mathematics. In 1738 she began preparing a textbook for their use, and found the topic so compelling that she devoted her complete intellectual attention to mathematics. Ten years later, her famous twovolume text Istituzioni analitiche was published. The work was the first comprehensive overview of the subject of calculus. Although designed for young students beginning their studies of the subject, Agnesi’s work was recognized as providing hitherto unnoticed connections between the different approaches of SIR ISAAC NEWTON (1642–1727) and GOTTFRIED WILHELM LEIBNIZ (1646–1716), independent coinventors of the subject. Her piece also provided, for the first time, clear explanations of previously confusing issues in the topic. Her text collated and explained the work of other contributors to the subject from several different countries, a task no doubt facilitated by her skills in translation. Her talents, not just as an expository writer, but also as a great scholar in mathematics, were apparent. Mathematicians at the time recognized her text as a significant contribution to the further development of the topic of calculus. In 1750 Agnesi was appointed the chair of mathematics at the University of Bologna in recognition of her great accomplishment. Curiously, she never officially accepted or rejected the faculty position. It is known that she never visited the city of Bologna, even though her name appears on university records over a span of 45 years. After the death of her father in 1752, Agnesi withdrew from mathematics and devoted her life to charitable work, helping the sick and the poor. In 1783 she was made director of a women’s poorhouse, where she remained the rest of her life. Having given all her money and possessions away, Agnesi died on January 9, 1799, with no money of her own and was buried in a pauper’s grave. The city of Milan, where she had lived all her life, today publicly honors her gravesite. John Colson, Lucasian professor of mathematics at Cambridge, published an English translation of Istituzioni analitiche in 1801. He said that he wanted to
8 Alembert, Jean Le Rond d’ give the youth of Britain the same opportunity to benefit from this remarkable text as the young scholars from Italy had been able to enjoy.
Alembert, Jean Le Rond d’ (1717–1783) French Differential equations, Analysis, Philosophy Born on November 17, 1717, in Paris, France, scholar Jean le Rond d’Alembert is best remembered for his 1743 treatise Traité de dynamique (Treatise on dynamics), in which he attempted to develop a firm mathematical basis for the study of mechanics. D’Alembert pioneered the study of partial DIFFERENTIAL EQUATIONs and their use in physics. He is also noted for his work on vibrating strings. After briefly pursuing theology and medicine at the Jansenist Collège des Quatre Nations, d’Alembert settled on mathematics as his choice of academic study.
Jean Le Rond d’Alembert, an eminent mathematician of the 18th century, pioneered the study of differential equations and their application to mechanics. (Photo courtesy of Topham/ The Image Works)
He graduated from the Collège in 1735 to then pursue interests in fluid mechanics. In 1740 he presented a series of lectures on the topic to members of the Paris Academy of Science, which earned him recognition as a capable mathematician and admittance as a member of the academy. He remained with the institution for his entire career. D’Alembert came to believe that the topic of mechanics should be based on logical principles, not necessarily physical ones, and that its base is fundamentally mathematical. In his 1743 treatise Traité de dynamique, he attempted to refine the work of SIR ISAAC NEWTON (1642–1727) and clarify the underpinnings of the subject. The following year d’Alembert published a second work, Traité de l’équilibre et du mouvement des fluides (Treatise on the equilibrium and movement of fluids), that applied his results to the study of fluid motion and introduced some beginning results on the study of partial differential equations. He developed these results further over the following years. In 1747 d’Alembert submitted a paper “Réflexions sur la cause générale des vents” (Reflections on the general cause of air motion) for consideration for the annual scientific prize offered by the Prussian Academy. He did indeed win. At the same time d’Alembert also began work as a writer and science editor for the famous French Encyclopédie ou dictionnaire raisonné des sciences, des arts, et des métiers (Encyclopedia and dictionary of the rationales of the sciences, arts, and professions), taking responsibility for the writing of the majority of mathematical entries. The first volume of the 28volume work was published in 1751. D’Alembert published several new mathematical results and ideas in this epic work. For instance, in volume 4, under the entry differential, he suggested, for the first time, that the principles of CALCULUS should be based on the notion of a LIMIT. He went so far as to consider defining the derivative of a function as the limit of a RATIO of increments. He also described the new ratio test when discussing CONVERGENT SERIES. D’Alembert’s interests turned toward literature and philosophy, and administrative work, in the latter part of his life. He was elected as perpetual secretary of the Académie Française in 1772. He died 11 years later on October 29, 1783. His work in mathematics paved the way for proper development of the notion of a limit in calculus, as well
algebra 9 as advancement of the field of partial differential equations. In mechanics, he is honored with a principle of motion named after him, a generalization of Sir Isaac Newton’s third law of motion. See also DIFFERENTIAL CALCULUS.
algebra The branch of mathematics concerned with the general properties of numbers, and generalizations arising from those properties, is called algebra. Often symbols are used to represent generic numbers, thereby distinguishing the topic from the study of ARITHMETIC. For instance, the equation 2 × (5 + 7) = 2 × 5 + 2 × 7 is a (true) arithmetical statement about a specific set of numbers, whereas, the equation x ×
History of Equations and Algebra Finding solutions to equations is a pursuit that dates back to the ancient Egyptians and Babylonians and can be traced through the early Greeks’ mathematics. The RHIND PAPYRUS, dating from around 1650 B.C.E., for instance, contains a problem reading: A quantity; its fourth is added to it. It becomes fifteen. What is the quantity? Readers are advised to solve problems like these by a method of “false position,” where one guesses (posits) a solution, likely to be wrong, and adjusts the guess according to the result obtained. In this example, to make the division straightforward, one might guess that the quantity is four. Taking 4 and adding to it its fourth gives, however, only 4 + 1 = 5, onethird of the desired answer of 15. Multiplying the guess by a factor of three gives the solution to the problem, namely 4 × 3, which is 12. Although the method of false position works only for LINEAR EQUATIONs of the form ax = b, it can nonetheless be an effective tool. In fact, several of the problems presented in the Rhind papyrus are quite complicated and are solved relatively swiftly via this technique. Clay tablets dating back to 1700 B.C.E. indicate that Babylonian mathematicians were capable of solving certain QUADRATIC equations by the method of COMPLETING THE SQUARE. They did not, however, have a general method of solution and worked only with a set of specific examples fully worked out. Any other problem that arose was
(y + z) = x × y + x × z is a general statement describing a property satisfied by any three numbers. It is a statement in algebra. Much of elementary algebra consists of methods of manipulating equations to either put them in a more convenient form, or to determine (that is, solve for) permissible values of the variables that appear. For instance, rewriting x2 + 6x + 9 = 25 as (x + 3)2 = 25 allows an easy solution for x: either x + 3 = 5, yielding x = 2, or x + 3 = –5, yielding x = –8. The word algebra comes from the Arabic term aljabr w’almuqa¯bala (meaning “restoration and – – A reduction”) used by the great MUHAMMAD IBN MUS – – ALKHW ARIZM I (ca. 780–850) in his writings on the topic.
matched with a previously solved example, and its solution was found by adjusting the numbers appropriately. Much of the knowledge built up by the old civilizations of Egypt and Babylonia was passed on to the Greeks. They took matters in a different direction and began examining all problems geometrically by interpreting numbers as lengths of line segments and the products of two numbers as areas of rectangular regions. Followers of PYTHAGORAS from the period 540 to 250 B.C.E., for instance, gave geometric proofs of the DISTRIBUTIVE PROPERTY and the DIFFERENCE OF TWO SQUARES formula, for example, in much the same geometric way we use today to explain the method of EXPANDING BRACKETS. The Greeks had considerable trouble solving CUBIC EQUATIONs, however, since their practice of treating problems geometrically led to complicated threedimensional constructions for coping with the product of three quantities. At this point, no symbols were used in algebraic problems, and all questions and solutions were written out in words (and illustrated in diagrams). However, in the third century, DIOPHANTUS OF ALEXANDRIA introduced the idea of abbreviating the statement of an equation by replacing frequently used quantities and operations with symbols as a kind of shorthand. This new focus on symbols had the subtle effect of turning Greek thinking away from geometry. Unfortunately, the idea of actually using the symbols to solve equations was ignored until the 16th century. The Babylonian and Greek schools of thought also influenced the development of mathematics in ancient India. The scholar BRAHMAGUPTA (ca. 598–665) gave solutions to (continues)
10 algebra
History of Equations and Algebra (continued) quadratic equations and outlined general methods for solving systems of equations containing several variables. (He also had a clear understanding of negative numbers and was comfortable working with zero as a valid numerical quantity.) The scholar Bh–askara (ca. 1114–85) used letters to represent unknown quantities and, in working with quadratic equations, suggested that all positive numbers have two square roots and that negative numbers have no (meaningful) roots. A significant step toward the development of modern algebra occurred in Baghdad, Iraq, in the year 825 when the – –A ALKHW –ARIZM–I (ca. Arab mathematician MUHAMMAD IBN MUS 780–850) published his famous piece Hisab aljabr w’almuqa¯bala (Calculation by restoration and reduction). This work represents the first clear and complete exposition on the art of solving linear equations by a new practice of performing the same operation on both sides of an equation. For example, the expression x – 3 = 7 can be “restored” to x = 10 by adding three to both sides of the expression, and the equation 5x = 10 can be “reduced” to x = 2 by dividing both –ı also showed how – sides of the equation by five. AlKhw arizm to solve quadratic equations via similar techniques. His descriptions, however, used no symbols, and like the ancient –ı wrote everything out in words. None– Greeks, alKhwarizm – theless, alKhwarizm–ı’s treatise was enormously influential, and his new approach to solving equations paved the way for modern algebraic thinking. In fact, it is from the word aljabr in the title of his book that our word algebra is derived. AlKhw –arizm–ı’s work was translated into Latin by the Italian mathematician FIBONACCI (ca. 1175–1250), and his efficient methods for solving equations quickly spread across Europe during the 13th century. Fibonacci translated the word shai used by alKhw –arizm–ı for “the thing unknown” into the Latin term res. He also used the Italian word cosa for “thing,” and the art of algebra became known in Europe as “the cossic art.” In 1545 GIROLAMO CARDANO (1501–76) published Ars magna (The great art), which included solutions to the cubic and QUARTIC EQUATIONs, as well as other mathematical discoveries. By the end of the 17th century, mathematicians were comfortable performing the same sort of symbolic manipulations we practice today and were willing to accept negative numbers and irrational quantities as solutions to equations. The French mathematician FRANÇOIS VIÈTE (1540–1603) introduced an efficient system for denoting powers of variables and was the first to use letters as coefficients before variables, as in “ax2 + bx + c,” for instance. (Viète also introduced the signs “+” and “–,” although he never used a sign for equality.) RENÈ DESCARTES (1596–1650) introduced the convention of denoting unknown quantities
by the last letters of the alphabet, x, y, and z, and known quantities by the first, a, b, c. (This convention is now completely ingrained; when we see, for example, an equation of the form ax + b = 0, we assume, without question, that it is for “x” we must solve.) The German mathematician CARL FRIEDRICH GAUSS (1777–1855) proved the FUNDAMENTAL THEOREM OF ALGEBRA in 1797, which states that every POLYNOMIAL equation of degree n has at least one and at most n (possibly complex) roots. His work, however, does not provide actual methods for finding these roots. Renaissance scholars SCIPIONE DEL FERRO (1465–1526) and NICCOLÒ TARTAGLIA (ca. 1500–57) both knew how to solve cubic equations, and in his 1545 treatise Ars magna, Cardano published the solution to the quartic equation discovered by his assistant LUDOVICO FERRARI (1522–65). For the centuries that followed, mathematicians attempted to find a general arithmetic method for solving all quintic (fifthdegree) equations. LEONHARD EULER (1707–83) suspected that the task might be impossible. Between the years 1803 and 1813, Italian mathematician Paolo Ruffini (1765–1822) published a number of algebraic results that strongly suggested the same, and just a few years later Norwegian mathematician NIELS HENRIK ABEL (1802–29) proved that, indeed, there is no general formula that solves all quintic equations in a finite number of arithmetic operations. Of course, some degreefive equations can be solved algebraically. (Equation of the form x5 – a = 0, for instance, have – solutions x = 5√a. ) In 1831 French mathematician ÉVARISTE GALOIS (1811–32) completely classified those equations that can be so solved, developing work that gave rise to a whole new branch of mathematics today called GROUP THEORY. In the 19th century mathematicians began using variables to represent quantities other than real numbers. For example, English mathematician GEORGE BOOLE (1815–64) invented an algebra symbolic logic in which variables represented sets, and Irish scholar SIR WILLIAM ROWAN HAMILTON (1805–65) invented algebraic systems in which variables represented VECTORs or QUATERNIONs. With these new systems, important characteristics of algebra changed. Hamilton, for instance, discovered that multiplication was no longer commutative in his systems: a product a × b might not necessarily give the same result as b × a. This motivated mathematicians to develop abstract AXIOMs to explain the workings of different algebraic systems. Thus the topic of ABSTRACT ALGEBRA was born. One outstanding contributor in this field was German mathematician AMALIE NOETHER (1883–1935), who made important discoveries about the nature of noncommutative algebras. See also ASSOCIATIVE; BABYLONIAN MATHEMATICS; CANCELLATION; COMMUTATIVE PROPERTY; EGYPTIAN MATHEMATICS; FIELD; GREEK MATHEMATICS; INDIAN MATHEMATICS; LINEAR ALGEBRA; RING.
alternating series 11 In modern times the subject of algebra has been widened to include ABSTRACT ALGEBRA, GROUP THEORY, and the study of alternative number systems such as MODULAR ARITHMETIC. BOOLEAN ALGEBRA looks at the algebra of logical inferences, matrix algebra the arithmetic of MATRIX operations, and vector algebra the mechanics of VECTOR operations and VECTOR SPACEs. An algebraic structure is any set equipped with one or more operations (usually BINARY OPERATIONs) satisfying a list of specified rules. For example, any group, RING, FIELD, or vector space is an algebraic structure. In advanced mathematics, a vector space that is also a field is called an “algebra.” See also BRACKETS; COMMUTATIVE PROPERTY; DISTRIBUTIVE PROPERTY; EXPANDING BRACKETS; FUNDAMENTAL THEOREM OF ALGEBRA; HISTORY OF EQUATIONS AND ALGEBRA (essay); ORDER OF OPERATION.
algebraic number A number is called algebraic if it is the root of a POLYNOMIAL with integer coefficients. — For example, (1/2) (5 + √13) is algebraic since it is a solution to the equation x2 – 5x + 3 = 0. All RATIONAL NUMBERS are algebraic (since a fraction a/b is the solution to the equation bx – a = 0), and all square, cube, n— and higher roots of integers are algebraic (since √ a is a solution to xn – a = 0). At first thought it seems that all numbers are algebraic, but this is not the case. In 1844 French mathematician JOSEPH LIOUVILLE made the surprising discovery that the following number, today called “Liouville’s constant,” cannot be a solution to any integer polynomial equation: L=
∞
1
∑ 10n!
= 0.11000100000000000000000100...
braic, with a different– from 0 or 1, and b irrational. (Thus, for example, 2√ 3 is transcendental.) The German mathematician GEORG CANTOR (1845–1918) showed that the set of algebraic numbers is COUNTABLE. As the set of real numbers is uncountable, this means that most numbers are transcendental. The probability that a real number chosen at random is algebraic is zero. Although it was proven in 1929 that eπ is transcendental, no one to this day knows whether or not ππ is algebraic. In analogy with algebraic numbers, a FUNCTION y = f(x) is called “algebraic” if it can be defined by a relation of the form pn(x)yn + pn– 1(x)yn – 1+…+p1(x)y + p0(x) = 0 where the functions pi(x) are polynomials in x. For — example, the function y = √ x is an algebraic function, since it is defined by the equation y2 – x = 0. A transcendental function is a function that is not algebraic. Mathematicians have shown that trigonometric, logarithmic, and exponential functions are transcendental. See also CARDINALITY.
algorithm An algorithm is a specific set of instructions for carrying out a procedure or solving a mathematical problem. Synonyms include “method,” “procedure,” and “technique.” One example of an algorithm is the common method of LONG DIVISION. Another is the EUCLIDEAN ALGORITHM for finding the GREATEST COMMON DIVISOR of two positive integers. The word algorithm is a distortion of “alKhw–arizm–ı,” the name of a Persian mathematician (ca. 820) who wrote an influential text on algebraic methods. See also BASE OF A NUMBER SYSTEM; MUHAMMAD – – – – IBN MUSA ALKHWARIZM ı.
n =1
Numbers that are not algebraic are called “transcendental.” In 1873 French mathematician Charles Hermite (1822–1901) proved that the number e is transcendental, and, nine years later in 1882 German mathematician CARL LOUIS FERDINAND VON LINDEMANN established that π is transcendental. In 1935 Russian mathematician Aleksandr Gelfond (1906–68) proved that any number of the form ab is transcendental if a and b are both alge
alternating series A SERIES whose terms are alternately positive and negative is called an alternating series. For 1 1 1 π is example, the GREGORY SERIES 1 − + − + K = 3 5 7 4 an alternating series, as is the (divergent) series: 1 – 1 + 1 – 1 + 1 – 1 +… Alternating series have the form ∞
∑ (−1)n−1 an = a1 − a2 + a3 − a4 + K ,
n =1
number.
with each ai positive
12 altitude In 1705, GOTTFRIED WILHELM LEIBNIZ noticed that many convergent alternating series, like the Gregory series, have terms ai that decrease and approach zero: i. a1 ≥ a2 ≥ a3 ≥… ii. an → 0 He managed to prove that any alternating series satisfying these two conditions does indeed converge, and today this result is called the “alternating series test.” (One can see that the test is valid if one physically paces smaller and smaller steps back and forth: a1 feet forward, a2 feet backward, a3 feet forward, and so on. This motion begins to “hone in” on a single limiting location.) We see, for example, that the series 1 1 1 1 1 + − + − + K converges. Unfortunately, 4 9 16 25 36 the alternating series test gives us no indication as to what the value of the sum could be. Generally, finding the limit value is a considerable amount of work, if at all possible. The values of many “simple” alternating series are not known today. (One can show, however, that the above series above converges to π2/12. See CONVERGENT SERIES.) See also ZETA FUNCTION. 1−
altitude A line segment indicating the height of a two or threedimensional geometric figure such as a POLYGON, POLYHEDRON, CYLINDER, or CONE is called an altitude of the figure. An altitude meets the base of the figure at a RIGHT ANGLE. Any TRIANGLE has three distinct altitudes. Each is a LINE segment emanating from a vertex of the triangle meeting the opposite edge at a 90° angle. The LAW OF SINES shows that the lengths ha, hb, and hc of the three altitudes of a triangle ABC satisfy: ha = c sin β = b sin γ hb = a sin γ = c sin α hc = b sin α = a sin β where a, b, and c are the sidelengths of the triangle, and α, β, and γ are the angles at vertices A, B, and C, respectively. Here ha is the altitude meeting the side of length a at 90°. Similarly, hb and hc are the altitudes meeting sides of length b and c, respectively. It also follows from this law that the following relation holds:
Proving altitudes are concurrent
ha hb h c =
(abc)3 8r 3
where r is the radius of the circle that contains the points A, B, and C. The three altitudes of a triangle always meet at a common point called the orthocenter of the triangle. Surprisingly, this fundamental fact was not noticed by the geometer EUCLID (ca. 300 B.C.E.). The claim can be proved as follows: Given a triangle ABC, draw three lines, one through each vertex and parallel to the side opposite to that vertex. This creates a larger triangle DEF. By the PARALLEL POSTULATE, alternate angles across parallel lines are equal. This allows us to establish that all the angles in the diagram have the values as shown. Consequently, triangle DAB is similar to triangle ABC and, in fact, is congruent to it, since it shares the common side AB. We have that DA is the same length as BC. In a similar way we can show that AE also has the same length as BC, and so A is the midpoint of side DE of the large triangle. Similarly, B is the midpoint of side DF, and C the midpoint of side EF. The study of EQUIDISTANT points establishes that
angle 13 the perpendicular bisectors of any triangle are CONCURRENT, that is, meet at a point. But the perpendicular bisectors of triangle DEF are precisely the altitudes of triangle ABC. One can also show that the three altitudes of a triangle satisfy: 1 1 1 1 + + = ha hb hc R where R is the radius of the largest circle that sits inside the triangle. See also EULER LINE.
solution to a problem “analysis.” He wrote about “a method of analysis” somewhat vaguely in his geometry text Collection, which left mathematicians centuries later wondering whether there was a secret method hidden behind all of Greek geometry. The great RENÉ DESCARTES (1596–1650) developed a powerful method of using algebra to solve geometric problems. His approach became known as analytic geometry. See also ANALYTIC NUMBER THEORY; CARTESIAN COORDINATES.
analytic number theory The branch of NUMBER THEthat uses the notion of a LIMIT to study the properties of numbers is called analytic number theory. This branch of mathematics typically deals with the “average” behavior of numbers. For example, to answer: ORY
amicable numbers (friendly numbers) Two whole numbers a and b are said to be amicable if the sum of the FACTORs of a, excluding a itself, equals b, and the sum of the factors of b, excluding b itself, equals a. For example, the numbers 220 and 284 are amicable: 284 has factors 1, 2, 4, 71, and 142, and their sum is 220 220 has factors 1, 2, 4, 5, 10, 11, 20, 22, 44, 55, and 110, and their sum is 284 The pair (220, 284) is the smallest amicable pair. For many centuries it was believed that this pair was the only pair of amicable numbers. In 1636, however, PIERRE DE FERMAT discovered a second pair, (17296, 18416), and in 1638, RENÉ DESCARTES discovered the pair (9363584, 9437056). Both these pairs were also known to Arab mathematicians, perhaps at an earlier date. By 1750, LEONHARD EULER had collated 60 more amicable pairs. In 1866, 16yearold Nicolò Paganini found the small pair (1184, 1210) missed by all the scholars of preceding centuries. Today more than 5,000 different amicable pairs are known. The largest pair known has numbers each 4,829 digits long. See also PERFECT NUMBER.
analysis Any topic in mathematics that makes use of the notion of a LIMIT in its study is called analysis. CALcomes under this heading, as does the summation of infinite SERIES, and the study of REAL NUMBERS. Greek mathematician PAPPUS OF ALEXANDRIA (ca. 320 C.E.) called the process of discovering a proof or a CULUS
On average, how many square factors does a number possess? one notes that all numbers have 1 as a factor, onequarter of all numbers have 4 as a factor, oneninth have the factor 9, onesixteenth the factor 16, and so on. Thus, on average, a number possesses 1 1 1 π2 + + + L= ≈ 1.64 square factors. This par4 9 16 6 ticular argument can be made mathematically precise. See also ANALYSIS; ZETA FUNCTION. 1+
angle Given the configuration of two intersecting LINEs,
line segments, or RAYs, the amount of ROTATION about the point of intersection required to bring one line coincident with the other is called the angle between the lines. Simply put, an angle is a measure of “an amount of turning.” In any diagram representing an angle, the lengths of the lines drawn is irrelevant. For example, an angle corresponding to oneeighth of a full turn can be represented by rays of length 2 in., 20 in., or 200 in. The image of a lighthouse with a rotating beam of light helps clarify the concept of an angle: each ray or line segment in a diagram represents the starting or ending position of the light beam after a given amount of turning. For instance, angles corresponding to a quarter of a turn, half a turn, and a full turn appear as follows:
14 angle A unit of measurement called a “gradian” is sometimes used in surveying. In this system, a full turn is considered 400 gradian (and, consequently, a quarter turn is divided into exactly 100 parts). This system is rarely used in mathematics, if at all. Angles are classified according to their measure:
Some simple angles
Mathematicians sometimes find it convenient to deem an angle measured in a counterclockwise sense as positive, and one measured in a clockwise sense as negative. Babylonian scholars of ancient times were aware that the year is composed of 365 days but chose to operate with a convenient calendar composed of 12 months of 30 days. Thus the number 360 came to be associated with the notion of a full cycle. Today, harking back to the Babylonians, angles are measured in units of degrees, in which a full rotation corresponds to 360 degrees (written 360°). Thus a half turn corresponds to 180°, and a quarter turn to 90°. A single degree corresponds to 1/360 of a turn. Each degree is divided into 60 smaller units called minutes, denoted with an apostrophe, and each minute is divided into 60 smaller units called seconds, denoted with two apostrophes. Thus, for instance, 15°46′23′′ represents an angle of 15 degrees, 46 minutes, and 23 seconds. Mathematicians prefer to use a unit of angle measurement independent of the circumstance that we live on the Earth, i.e., one that is natural to mathematics. The chosen unit of measurement is called a radian. Working with the simplest CIRCLE possible, namely, a circle of radius one, mathematicians match the measure of a full turn with the distance around that circle, namely 2π, the circumference of the circle. Thus one full rotation equals 2π radian. A half turn is measured as half of this, namely, π radian, and a quarter turn as 1 π ⋅ 2π = radian. 4 2 To convert between degree and radian measures, one simply notes that 360 degrees corresponds to 2π 2π radian. (Thus one degree equals ≈ 0.017 radian, 360 and one radian equals
360 ≈ 57.3o .) 2π
• An angle of zero degree is called a null angle. • An angle between 0° and 90° is called acute. • An angle of 90° is called a right angle. (It is the angle formed when one makes a perfect right turn.) • An angle between 90° and 180° is called obtuse. • An angle of 180° is called a straight angle. • An angle between 180° and 360° is called a reflex angle. • An angle of 360° is called a PERIGON or a round angle. Two angles are said to be congruent if they have the same measure. If two angles have measures summing to a right angle, then they are said to be complementary, and two angles are supplementary if their measures sum to a straight angle. Special names are also given to angles that appear in a diagram involving a TRANSVERSAL. The angle of elevation of a point P above the ground relative to an observer at position Q on the ground is defined to be the angle between the line connecting Q to P, and the line connecting Q to the point on the ground directly below P. If P lies below ground level, then an angle of depression is computed similarly. The angle between two VECTORs is computed via the DOT PRODUCT. Using TRIGONOMETRY one shows that the angle A between two lines in the plane of slopes m1 and m2, respectively, is given by: tan A =
m1 − m2 1 + m1m2
(This follows by noting that the angle the first line makes with the xaxis is tan–1(m1) and the angle of the second line is tan–1(m2). Thus the angle we seek is A = tan–1(m1) – tan–1(m2). The trigonometric identity tan(α– β) = tan α − tan β now establishes the result.) 1 + tan α tan β Consequently, if m1m2 = –1, the lines are PERPENDICULAR. The angle between two intersecting curves in a plane is defined to be the angle between the TANGENT lines to the curves at the point of intersection. The link between the measure of an angle and the length of arcs of a unit circle to define radian measure can be extended to associate a measure of “angle” with tan(α − β ) =
Apollonius of Perga 15 regions on a unit SPHERE. A SOLID ANGLE of a region is simply the measure of surface AREA of that region projected onto a unit sphere. Units of solid angle are called steradians. The full surface area of the sphere is 4π steradian. See also BABYLONIAN MATHEMATICS; DIHEDRAL; SLOPE; TRIANGLE; TRISECTING AN ANGLE.
angle trisection See TRISECTING AN ANGLE. annulus (plural, annuli) An annulus is the region between two CONCENTRIC circles in a plane. The of the annulus is the difference of the areas of the two circles. If a cyclist rides a perfect circle, the region between the tracks made by the front and rear wheels is an annulus. If the bicycle is r feet long (axel to axel), then, surprisingly, the area of this annulus is πr2 feet squared, irrespective of the size of the circle the cyclist traces. AREA
antidifferentiation (integration) The process of finding a function with a given function as its DERIVATIVE is called antidifferentiation. For example, x2 is an d 2 (x ) = 2x . The MEAN dx VALUE THEOREM shows that two antiderivatives of the same function differ only by a constant. Thus all the antiderivatives of 2x, for example, are functions of the form x2 + C. The antiderivative of a function f(x) is denoted ∫f(x)dx and is called the indefinite integral of f. It is defined up to a constant, and so we write, for example, ∫2x dx = x2 + C. (The constant C is referred to as a “constant of integration.”) The notation is deliberately b suggestive of a definite integral of a function, ∫a f(x)dx, for the area under the curve y = f(x) over the INTERVAL [a,b]. The FUNDAMENTAL THEOREM OF CALCULUS shows that the two notions are intimately connected. See also INTEGRAL CALCULUS. antiderivative of 2x, since
antilogarithm See LOGARITHM. antipodal points (antipodes) Two points on a SPHERE
at the opposite ends of a diameter are said to be
antipodal. For example, the north and south poles are antipodal points on the EARTH, as are any two points EQUIDISTANT from the equator, with longitudes differing by 180°. The famous BorsukUlam theorem, first conjectured by Stanislaw Ulam and then proved by Karol Borsuk in 1933, states: Let f be a continuous function that assigns two numerical values to each and every point on the surface of a sphere. Then there must exist two antipodal points which are assigned precisely the same pair of values. An amusing interpretation reads: At any instant there exist two antipodal points on the Earth’s surface of precisely the same air temperature and air pressure. Although the proof of this theorem is difficult, a onedimensional version of the result follows as an easy consequence of the INTERMEDIATEVALUE THEOREM.
apex (plural, apices) The point at the top of a POLYGON or a POLYHEDRON, such as the vertex of a triangle opposite its BASE or the vertex of a pyramid, is called the apex of the figure. The distance from the base of the figure to its apex is called the height of the figure.
Apollonius of Perga (ca. 262–190 B.C.E.) Greek Geometry Born in Perga, Greek Ionia, now Antalya, Turkey, Apollonius worked during the Golden Age of Greek mathematics and has been referred to throughout history as the Great Geometer. His famous work, The Conics, written in eight volumes, greatly influenced the development of mathematics. (The names ELLIPSE, PARABOLA, and HYPERBOLA for the three CONIC SECTIONS, for instance, are said to have been coined by Apollonius.) Copies of the first four volumes of this work, written in the original Greek, survive today. Arabic translations of the first seven volumes also exist. Little is known of Apollonius’s life other than what can be gleaned from incidental comments made in the prefaces of his books. As a young man it is known that he traveled to Alexandria to study with the followers of EUCLID, who then introduced him to the topic of conics.
16 Apollonius’s circle The first volume of The Conics simply reviews elementary material about the topic and chiefly presents results already known to Euclid. Volumes two and three present original results regarding the ASYMPTOTEs to hyperbolas and the construction of TANGENT lines to conics. While Euclid demonstrated a means, for instance, of constructing a circle passing through any three given points, Apollonius demonstrated techniques for constructing circles tangent to any three lines, or to any three circles, or to any three objects be they a combination of points, lines, or circles. Volumes four, five, six, and seven of his famous work are highly innovative and contain original results exploring issues of curvature, the construction of normal lines, and the construction of companion curves to conics. Apollonius also applied the theory of conics to solve practical problems. He invented, for instance, a highly accurate sundial, called a hemicyclium, with hour lines drawn on the surface of a conic section. Apollonius also played a fundamental role in the development of Greek mathematical astronomy. He proposed a complete mathematical analysis of epicyclic motion (that is, the compound motion of circles rolling along circles) as a means to help explain the observed retrograde motion of the planets across the skies that had confused scholars of his time. Apollonius’s work was extraordinarily influential, and his text on the conics was deemed a standard reference piece for European scholars of the Renaissance. JOHANNES KEPLER, RENÉ DESCARTES, and SIR ISSAC NEWTON each made reference to The Conics in their studies. See also CIRCUMCIRCLE; CYCLOID.
Apollonius’s circle Let A and B be two points of the plane and let k be a constant. Then the set of all points P whose distance from A is k times its distance from B is a CIRCLE. Any circle obtained this way is referred to as one of Apollonius’s circles. Note that when k = 1 the circle is “degenerate,” that is, the set of all points EQUIDISTANT from A and B is a straight line. When k becomes large, the Apollonius’s circle approaches a circle of radius 1. To see that the locus of points described this way is indeed a circle, set A to be the origin (0,0), B to be the point (k + 1, 0) on the xaxis, and P to be a general point with coordinates (x,y). The DISTANCE FOR
MULA
then
gives
an
equation
of
the
form
x 2 + y 2 = k (x − k − 1)2 + y 2 . This is equivalent to 2
2
k k2 x − + y2 = , which is indeed the equa k − 1 k − 1 tion of a circle, one of radius
k . APOLLONIUS k −1
OF
PERGA used purely geometric techniques, however, to establish his claim.
Apollonius’s theorem If a, b, and c are the sidelengths of a triangle and a median of length m divides the third side into two equal lengths c/2 and c/2, then the following relation holds: a 2 + b2 =
c2 + 2m2 2
This result is known as Apollonius’s theorem. It can be proved using two applications of the LAW OF COSINES as follows: Let B be the ANGLE between the sides of length a and c. Then m2 = a2 + (c/2)2–ac cos(B) and b2 = a2 + c2 – 2ac cos(B). Solving for ac cos(B) in the first equation and substituting into the second yields the result. See also MEDIAN OF A TRIANGLE.
apothem (short radius) Any line segment from the center of a regular POLYGON to the midpoint of any of its sides is called an apothem. If the regular polygon has n sides, each one unit in length, then an exercise in TRIGONOMETRY shows that each apothem of the figure has length
r=
1 180 . 2 tan n
An analog of PI (π) for a regular polygon is the of its PERIMETER to twice the length of its apothem. For a regular nsided polygon, this ratio has value n tan(180/n). The SQUEEZE RULE shows that this quantity approaches the value π as n becomes large. See also LONG RADIUS. RATIO
Arabic mathematics 17
applied mathematics The study and use of the mathematical techniques to solve practical problems is called applied mathematics. The field has various branches including STATISTICS, PROBABILITY, mechanics, mathematical physics, and the topics that derive from them, but the distinction from PURE MATHEMATICS might not be sharp. For instance, the general study of VECTORs and VECTOR SPACEs can be viewed as either an abstract study or a practical one if one later has in mind to use this theory to analyze force diagrams in mechanics. Many research universities of today possess two departments of mathematics, one considered pure and the other applied. Students can obtain advanced degrees in either field.
approximation A numerical answer to a problem that is not exact but is sufficient for all practical purposes is called an approximation. For example, noting that 210 is approximately 1,000 allows us to quickly estimate the value of 2100 = (210)10 as 1030. Students are often encouraged to use the fraction 22/7 as an approximate value for π. Mathematicians use the notation “≈” to denote approximately equal to. Thus, for example, π ≈ 22/7. Physicists and engineers often approximate functions by their TAYLOR SERIES with the higherorder x3 x 5 + terms dropped. For example, sin x ≈ x − , at 3! 5! least for small values of x. The theory of INTEGRAL CALCULUS begins by approximating areas under curves as sums of areas of rectangles. See also ERROR; FACTORIAL; NUMERICAL DIFFERENTIATION; NUMERICAL INTEGRATION.
Arabic mathematics Mathematical historians of today are grateful to the Arabic scholars of the past for preserving, translating, and honoring the great Indian, Greek, and Islamic mathematical works of the scholars before them, and for their own significant contributions to the development of mathematics. At the end of the eighth century, with the great Library of Alexandria destroyed, Caliph alMa’mun set up a House of Wisdom in Baghdad, Iraq, which became the next prominent center of learning and research, as well as the repository of important academic texts. Many scholars were employed by the caliph to translate the mathemat
ical works of the past and develop further the ideas they contained. As the Islamic empire grew over the following seven centuries, the culture of intellectual pursuit also spread. Many scholars of 12thcentury Europe, and later, visited the Islamic libraries of Spain to read the texts of the Arabic academics and to learn of the advances that had occurred in the East during the dark ages of the West. A significant amount of mathematical material was transmitted to Europe via these means. One of the first Greek texts to be translated at the House of Wisdom was EUCLID’s famous treatise, THE ELEMENTS. This work made a tremendous impact on the Arab scholars of the period, and many of them, when conducting their own research, formulated theorems and proved results precisely in the style of Euclid. Members of the House of Wisdom also translated the works of ARCHIMEDES OF SYRACUSE, DIOPHANTUS OF ALEXANDRIA, MENELAUS OF ALEXANDRIA, and others, and so they were certainly familiar with all the great Greek advances in the topics of GEOMETRY, NUMBER THEORY, mechanics, and analysis. They also translated – the works of Indian scholars, ARYABHATA and – BHASKARA, for instance, and were familiar with the theory of TRIGONOMETRY, methods in astronomy, and further topics in geometry and number theory. Any Arab scholar who visited the House of Wisdom had, essentially, the entire bulk of human mathematical knowledge available to him in his own language. – A – Arab mathematician MUHAMMAD IBN MUS – – ALKHWARIZM ı (ca. 800) wrote a number of original texts that were enormously influential. His first piece simply described the decimal placevalue system he had learned from Indian sources. Three hundred years later, when translated into Latin, this work became the primary source for Europeans who wanted to learn the new system for writing and manipulating numbers. But – –ı ’s piece Hisab aljabr more important was alKhwarizm w’almuq¯abala (Calculation by restoration and reduction), from which the topic of “algebra” (“aljabr”) – –ı was fortunate to have all sources arose. AlKhwarizm of mathematical knowledge available to him. He began to see that the thendisparate notions of “number” and “geometric magnitude” could be unified as one whole by developing the concept of algebraic objects. This represented a significant departure from Greek thinking, in which mathematics is synonymous with geometry. – –ı’s insight provided a means to study both AlKhwarizm arithmetic and geometry under a single framework, and
18 arc his methods of algebra paved the way for significant developments in mathematics of much broader scope than ever previously envisioned. The mathematician alMahani (ca. 820) developed refined approaches for reducing geometric problems to algebraic ones. He showed, in particular, that the famous problem of DUPLICATING THE CUBE is essentially an algebraic issue. Other scholars brought rigor to the subject by proving that certain popular, but complicated, algebraic methods were valid. These scholars were comfortable manipulating POLYNOMIALs and developed rules for working with EXPONENTs, They solved linear and QUADRATIC equations, as well as various SYSTEMs OF EQUATIONS. Surprisingly, no one of the time thought to ease matters by using symbols to represent quantities: all equations and all manipulations were described fully in words each and every time they were employed. With quadratic equations well understood, the scholar OMAR KHAYYAM (ca. 1048–1131) attempted to develop methods of solving degreethree equations. Although he was unable to develop general algebraic methods for this task, he did find ingenious geometric techniques for solving certain types of cubics with the aid of CONIC SECTIONS. He was aware that such equations could have more than one solution. In number theory, Thabit ibn Qurra (ca. 836–901) found a beautiful method for generating AMICABLE NUMBERS. This technique was later utilized by alFarisi (ca. 1260–1320) to yield the pair 17,296 and 18,416, which today is usually attributed to LEONHARD EULER (1707–83). In his writing, Omar Khayyam referred to earlier Arab texts, now lost, that discuss the equivalent of PASCAL’s TRIANGLE and its connections to the BINOMIAL THEOREM. The mathematician alHaytham (ca. 965–1040) attempted to classify all even PERFECT NUMBERs. Taking advantage of the ease of the Indian system of decimal placevalue representation, Arabic scholars also made great advances in numeric computations. The great 14thcentury scholar JAMSHID ALKASHI developed effective methods for extracting the nth root of a number, and evaluated π to a significant number of decimal places. Scholars at the time also developed effective methods for computing trigonometric tables and techniques for making highly accurate computations for the purposes of astronomy. On a theoretical note, scholars also advanced the general understanding of trigonometry and explored
problems in spherical geometry. They also investigated the philosophical underpinnings of geometry, focusing, in particular, on the role the famous PARALLEL POSTULATE plays in the theory. Omar Khayyam, for instance, attempted to prove the parallel postulate—failing, of course—but did accidentally prove results about figures in nonEuclidean geometries along the way. The mathematician Ibrahim ibn Sinan (908–946) also introduced a method of “integration” for calculating volumes and areas following an approach more general than that developed by Archimedes of Syracuse (ca. 287–212 B.C.E.). He also applied his approach to the study of CONIC SECTIONS and to optics. See also BASE OF A NUMBER SYSTEM; HISTORY OF EQUATIONS AND ALGEBRA (essay); HISTORY OF GEOMETRY (essay); HISTORY OF TRIGONOMETRY (essay).
arc Part of a continuous curve between two given points on the curve is called an arc of the curve. In particular, two points on a CIRCLE determine two arcs. If the circumference of the circle is divided by them into two unequal parts, then the smaller portion is usually called the minor arc of the circle and the larger the major arc.
Archimedean spiral See SPIRAL OF ARCHIMEDES. Archimedes of Syracuse (287–212
B.C.E.) Greek Geometry, Mechanics Born in the Greek colony of Syracuse in Sicily, Archimedes is considered one of the greatest mathematicians of all time. He made considerable contributions to the fields of planar and solid GEOMETRY, hydrostatics, and mechanics. In his works Measurement of a Circle and Quadrature of the Parabola, Archimedes solved difficult problems of mensuration in planar geometry by inventing an early technique of INTEGRAL CALCULUS, which he called the “method of exhaustion.” This allowed him to compute areas and lengths of certain curved figures. Later, in his works On the Sphere and Cylinder and On Conoids and Spheroids, he applied this technique to also compute the volume and surface area of the sphere and other solid objects. In his highly original work On Floating Bodies, Archimedes developed the mathematics of hydrostatics and equilibrium, along with the LAW OF THE LEVER, the notion of specific gravity, and techniques for computing the CENTER OF GRAVITY of a
Archimedes of Syracuse 19
A 1547 woodcut depicting Archimedes’ realization that a tub of water can be used to compute the volumes and densities of solid figures (Photo courtesy of ARPL/Topham/The Image Works)
variety of bodies. In mathematics, Archimedes also developed methods for solving cubic equations, approximating square roots, summing SERIES, and, in The Sand Reckoning, developed a notation for representing extremely large numbers. Except for taking time to study at EUCLID’s school in Alexandria, Archimedes spent his entire life at the place of his birth. He was a trusted friend of the monarch of the region, Hiero, and his son Gelon, and soon developed a reputation as a brilliant scientist who could solve the king’s most troublesome problems. One famous story asserts that the king once ordered a goldsmith to make him a crown, and supplied the smith the exact amount of metal to make it. Upon receiving the newly forged crown, Hiero suspected the smith of ill doing,
substituting some cheaper silver for the gold, even though the crown did have the correct weight. He could not prove his suspicions were correct, however, and so brought the problem to Archimedes. It is said that while bathing and observing the water displaced by his body Archimedes realized, and proved, that the weight of an object suspended in liquid decreases in proportion to the weight of the liquid it displaces. This principle, today known as Archimedes’ principle, provided Archimedes the means to indeed prove that the crown was not of solid gold. (It is also said that Archimedes was so excited upon making this discovery that he ran naked through the streets shouting, “Eureka! Eureka!”) Archimedes is also purported to have said, “Give me a place to stand and I shall move the Earth.” Astonished
20 Archimedes of Syracuse by the claim, King Hiero asked him to prove it. Archimedes had, at this time, discovered the principles of the levers and pulleys, and set about constructing a mechanical device that allowed him, singlehandedly, to launch a ship from the harbor that was too large and heavy for a large group of men to dislodge. Dubbed a master of invention, Archimedes also devised a waterpumping device, now known as the Archimedes screw and still used in many parts of the world today, along with many innovative machines of war that were used in the defense of Sicily during the notinfrequent Roman invasions. (These devices included parabolic mirrors to focus the rays of the sun to burn advancing ships from shore, catapult devices, and springloaded cannons.) But despite the fame he received for his mechanical inventions, Archimedes believed that pure mathematics was the only worthy pursuit. His accomplishments in mathematics were considerable.
Archimedes of Syracuse, regarded as one of the greatest scientists of all time, pioneered work in planar and solid geometry, mechanics, and hydrostatics. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
By bounding a circle between two regular polygons and calculating the ratio the perimeter to diameter of each, Archimedes found one of the earliest estimates for the value of π, bounding it between the values 3 10/71 and 3 1/7. (This latter estimate, usually written as 22/7, is still widely used today.) Archimedes realized that by using polygons with increasingly higher numbers of sides yielded better and better approximations, and that by “exhausting” all the finite possibilities, the true value of π would be obtained. Archimedes also used this method of exhaustion to demonstrate that the length of any segment of a parabola is 4/3 times the area of the triangle with the same base and same height. By comparing the crosssectional areas of parallel slices of a sphere with the slices of a cylinder that encloses the sphere, Archimedes demonstrated that the volume of a sphere is 2/3 that of the cylinder. The volume of the sphere then follows: V = (2/3)(2r × πr2) = (4/3)πr3. (Here r is the radius of the sphere.) Archimedes regarded this his greatest mathematical achievement, and in his honor, the figures of a cylinder and an inscribed sphere were drawn on his tombstone. Archimedes also computed the surface area of a sphere as four times the area of a circle of the same radius of the sphere. He did this again via a method of exhaustion, by imagining the sphere as well approximated by a covering of flat tiny triangles. By drawing lines connecting each vertex of a triangle to the center of the sphere, the volume of the figure is thus divided into a collection of triangular pyramids. Each pyramid has volume onethird its base times it height (essentially the radius of the sphere), and the sum of all the base areas represents the surface area of the sphere. From the formula for the volume of the sphere, the formula for its surface area follows. One cannot overstate the influence Archimedes has had on the development of mathematics, mechanics, and science. His computations of the surface areas and volumes of curved figures provided insights for the development of 17thcentury calculus. His understanding of Euclidean geometry allowed him to formulate several axioms that further refined the logical underpinnings of the subject, and his work on fluids and mechanics founded the field of hydrostatics. Scholars and noblemen of his time recognized both the theoretical and practical importance of his work. Sadly, Archimedes died unnecessarily in the year 212 B.C.E. During the conquest of Syracuse by the Romans, it is
Archytas of Tarentum 21
A 19thcentury engraving of Archimedes’ water screw, a device for pumping water (Photo courtesy of AAAC/Topham/The Image Works)
said that a Roman soldier came across Archimedes concentrating on geometric figures he had drawn in the sand. Not knowing who the scholar was, or what he was doing, the soldier simply killed him.
Archytas of Tarentum (ca. 428–350 B.C.E.) Greek Geometry, Philosophy The Greek scholar Archytas of Tarentum was the first to provide a solution to the classic DUPLICATING THE CUBE problem of antiquity. By reducing the challenge to one of constructing certain
ratios and proportions, Archytas developed a geometric construct that involved rotating semicircles through certain angles in threedimensional space to produce a length essentially equivalent to the construct of the cube root of two. (Creating a segment of this length is the chief stumbling block to solving the problem.) Although his innovative solution is certainly correct, it uses tools beyond what is permissible with straightedge and compass alone. In the development of his solution, Archytas identified a new mean between numbers, which he called the HARMONIC MEAN.
22 arc length Archytas lived in southern Italy during the time of Greek control. The region, then called Magna Graecia, included the town of Tarentum, which was home to members of the Pythagorean sect. Like the Pythagoreans, Archytas believed that mathematics provided the path to understanding all things. However, much to the disgust of the Pythagoreans, Archytas applied his mathematical skills to solve practical problems. He is sometimes referred to as the Founder of Mechanics and is said to have invented several innovative mechanical devices, including a mechanical bird and an innovative child’s rattle. Only fragments of Archytas’s original work survive today, and we learn of his mathematics today chiefly through the writings of later scholars. Many results established by Archytas appear in EUCLID’s famous text THE ELEMENTS, for instance.
arc length To measure the length of a curved path, one could simply lay a length of string along the path, pull it straight, and measure its length. This determines the arc length of the path. In mathematics, if the curve in question is continuous and is given by a formula y = f(x), for a < x < b say, then INTEGRAL CALCULUS can be used to find the arc length of the curve. To establish this, first choose a number of points (x1, y1),…,(xn, yn) along the curve and sum the lengths of the straightline segments between them. Using the DISTANCE FORMULA, this gives an approximate value for the length s of the curve: n
(xi − xi −1 )2 + (yi − yi −1 )2
s≈∑
i =1
Rewriting yields: n
s≈∑
i =1
⎛ y − yi −1 ⎞ 1+ ⎜ i ⎟ ⎝ xi − xi −1 ⎠
2
(xi − xi −1 )
The MEANVALUE THEOREM shows that for each i there is yi − yi −1 a value ci between xi–1 and xi so that x − x = f ′(ci ) , i i −1 and so the length of the curve is well approximated by the formula: n
s ≈ ∑ 1 + ( f ′(ci )) ( xi − xi −1 ) i =1
2
Of course, taking more and more points along the curve gives better and better approximations. In the limit, then, the true length of the curve is given by the formula: n
s = lim n→∞ ∑ 1 + ( f ′(ci )) ( xi − xi −1 ) 2
i =1
This is precisely the formula for the integral of the function
1 + ( f ′(x))
2
over the domain in question.
Thus we have: The arc length of a continuous curve y = f(x) over the interval [a,b] is given by s = ∫
b
a
1 + ( f ′(x)) dx 2
Alternatively, if the continuous curve is given by a set of PARAMETRIC EQUATIONS x = x(t) and y = y(t), for a < t < b say, then choosing a collection of points along the curve, given by t1,…,tn say, making an approximation to the curve’s length, and taking a limit yields the formula:
s = lim n→∞
(x(t i ) − x(t i −1) )2 + (y(t i ) − y(t i −1) )2 2
2
⎛ x(t ) − x(t i −1) ⎞ ⎛ y(t i ) − y(t i −1) ⎞ = lim n→∞ ⎜ i ⎟ +⎜ ⎟ (t i − t i −1) ⎝ t i − t i −1 ⎠ ⎝ t i − t i −1 ⎠ =∫
b
a
2
2
⎛ dx ⎞ ⎛ dy ⎞ ⎜ ⎟ + ⎜ ⎟ dt ⎝ dt ⎠ ⎝ dt ⎠
In a similar way one can show that if the continuous curve is given in POLAR COORDINATES by formulae and x = r(θ)cos(θ) and y = r(θ)sin(θ), for a < θ < b, then the arc length of the curve is given by: s=∫
b
a
2
⎛ dr ⎞ r 2 + ⎜ ⎟ dθ ⎝ dθ ⎠
The presence of squareroot signs in the integrands often makes these integrals very difficult, if not impossible, to solve. In practice, one must use numerical techniques to approximate integrals such as these. See also NUMERICAL INTEGRATION.
area Loosely speaking, the area of a geometric figure is the amount of space it occupies. Such a definition
area 23 appeals to intuitive understanding. In general, however, it is very difficult to explain precisely just what it is we mean by “space” and the “amount” of it occupied. This is a serious issue. (See BanachTarski paradox on following page.) As a starting point, it seems reasonable to say, however, that a 1 × 1 square should have “area” one. We call this a basic unit of area. As four of these basic units fit snugly into a square with sidelength two, without overlap, we say then that a 2 × 2 square has area four. Similarly a 3 × 3 square has area nine, a 4 × 4 square area 16, and so on. A 3 × 6 rectangle holds 18 basic unit squares and so has area 18. In general, a rectangle that is l units long and w units wide, with both l and w whole numbers, has area l × w: area of a rectangle = length × width This is a fundamental formula. To put the notion of area on a sound footing, we use this formula as a defining law: the area of any rectangle is to be the product of its length and its width. Although it is impossible to fit a whole number of unit squares into a rectangle that is 5 3/4 units long and — √ 7 units wide, for example, we declare, nonetheless, that the area of such a rectangle is the product of these two numbers. (This agrees with our intuitive idea that, — with the aid of scissors, about 5 3/4 × √ 7 ≈ 15.213 unit squares will fit in this rectangle.) From this law, the areas of other geometric shapes follow. For example, the following diagram shows that the area of an acute TRIANGLE is half the area of the rectangle that encloses it. This leads to the formula: area of a triangle = 1/2 × base × height This formula also holds for obtuse triangles.
The areas of basic shapes
By rearranging pieces of a PARALLELOGRAM, we see that its area is given by the formula: Area
area of a parallelogram = base × height
24 area In general we can calculate the area of any POLYGON as the sum of the areas of the triangles that subdivide it. For example, the area of a TRAPEZOID is the sum of the areas of two triangles, and the area of a regular POLYGON with n sides is the sum of the areas of n triangles. Curved Figures It is also possible to compute the area of curved figures. For example, slicing a circle into wedgedshape pieces and rearranging these slices, we see that the area of a circle is close to being the area of a rectangle of length half the circumference and of width r, the radius of the circle. If we work with finer and finer wedgedshape pieces, the approximation will better approach that of a true rectangle. We conclude that the area of a circle is indeed that of this ideal rectangle: area of a circle = 1/2 × circumference × r (Compare this with the formula for the area of a regular polygon.) As PI (π) is defined as the ratio of the circumference of a circle to its diameter, circumference π = ––––––––––––– , the area of a circle can thus be 2r written: area = 1/2 × 2πr × r. This leads to the famous formula: area of a circle = πr2 The methods of INTEGRAL CALCULUS allow us to compute areas of other curved shapes. The approach is analogous: approximate the shape as a union of rectangles, sum the areas of the rectangular pieces, and take the LIMIT of the answers obtained as you work with finer and finer approximations.
Establishing the area of a circle
Theoretical Difficulties Starting with the principle that a fundamental shape, in our case a rectangle, is asserted to have “area” given by a certain formula, a general theory of area for other geometric shapes follows. One can apply such an approach to develop a measure theory for measuring the size of other sets of objects, such as the notion of the surface area of threedimensional solids, or a theory of VOLUME. One can also develop a number of exotic applications. Although our definition for the area of a rectangle is motivated by intuition, the formula we developed is, in some sense, arbitrary. Defining the area of a rectangle as given by a different formula could indeed yield a different, but consistent, theory of area. In 1924 STEFAN BANACH and Alfred Tarski stunned the mathematical community by presenting a mathematically sound proof of the following assertion: It is theoretically possible to cut a solid ball into nine pieces, and by reassembling them, without ever stretching or warping the pieces, form TWO solid balls, each exactly the same size and shape as the original. This result is known as the BanachTarski paradox, and its statement—proven as a mathematical fact—is abhorrent to our understanding of how area and volume should behave: the volume of a finite quantity of material should not double after rearranging its pieces! That our intuitive understanding of area should eventually lead to such a perturbing result was considered very disturbing. What mathematicians have come to realize is that “area” is not a welldefined concept: not every shape in a plane can be assigned an area (nor can every solid in threedimensional space be assigned a volume). There exist certain nonmeasurable sets about which speaking of their area is meaningless. The nine pieces used in the BanachTarski paradox turn out to be such nonmeasurable sets, and so speaking of their volume is invalid. (They are extremely jagged shapes, FRACTAL in nature, and impossible to physically cut out.) In particular, interpreting the final construct as “two balls of equal volume” is not allowed. Our simple intuitive understanding of area works well in all practical applications. The material presented in a typical highschool and college curriculum, for example, is sound. However, the BanachTarski paradox points out that extreme care must be taken
argument 25 when exploring the theoretical subtleties of area and volume in greater detail. See also SCALE.
Argand, Jean Robert (1768–1822) Swiss Complex number theory Born on July 8, 1768, in Geneva, Switzerland, amateur mathematician Jean Argand is remembered today for his famous geometrical interpretation for COMPLEX NUMBERS. An ARGAND DIAGRAM uses two perpendicular axes, one representing a real number line, the second a line of purely complex numbers, to represent complex numbers as points in a plane. It is not well known that Argand, in fact, was not the first to consider and publish this geometric approach to complex numbers. The surveyor Casper Wessel (1745–1818) submitted the same idea to the Royal Danish Academy in 1797, but his work went unnoticed by the mathematics community. At the turn of the century, Argand independently began to interpret the complex number i geometrically as a rotation through 90°. He expounded on the convenience and fruitfulness of this idea in a small book, Essai sur une manière de représenter les quantités imaginaires dans les constructions géometriques (Essay on a method for representing imaginary quantities through a geometric construction), which he published privately, at his own expense, in 1806. He never wrote his name in the piece, and so it was impossible to identify the author. By chance, French mathematician Jacques Français came upon the small publication and wrote about the details of the work in an 1813 article, “A Memoir on the Geometric Representation of Imaginary Numbers,” published in the Annales de Mathématiques. He requested that the unknown originator of the ideas come forward and receive credit for the work. Argand made himself known by submitting his own article to the same journal, presenting a slightly modified and improved approach to his methods. Although historians have since discovered that the mathematicians JOHN WALLIS (1616–1703) and CARL FRIEDRICH GAUSS (1777–1855) each considered their own geometric interpretations of complex numbers, Argand is usually credited as the discoverer of this approach. Argand was the first to develop the notion of the MODULUS of a complex number. It should also be noted that Argand also presented an essentially complete proof of the FUNDAMENTAL THEOREM OF ALGEBRA in
his 1806 piece, but has received little credit for this accomplishment. Argand was the first to state, and prove, the theorem in full generality, allowing all numbers involved, including the coefficients of the polynomial, to be complex numbers. Argand died on August 13, 1822, in Paris, France. Although not noted as one of the most outstanding mathematicians of his time, Argand’s work certainly shaped our understanding of complex number theory. The Argand diagram is a construct familiar to all advanced highschool mathematics students.
Argand diagram (complex plane) See
COMPLEX
NUMBERS.
argument In the fourth century
B.C.E., Greek philosopher ARISTOTLE made careful study of the structure of reasoning. He concluded that any argument, i.e., a reasoned line of thought, consists, essentially, of two basic parts: a series of PREMISEs followed by a conclusion. For example:
If today is Tuesday, then I must be in Belgium. I am not in Belgium. Therefore today is not Tuesday. is an argument containing two premises (the first two lines) and a conclusion. An argument is valid if the conclusion is true when the premises are assumed to be true. Any argument has the general form: If [premise 1 AND premise 2 AND premise 3 AND…], then [Conclusion] Using the symbolic logic of FORMAL LOGIC and TRUTH TABLEs, the above example has the general form: p→q ¬q Therefore ¬p The argument can thus be summarized: ((p → q) ∨ (¬q)) → (¬p). One can check with the aid of a truth table that this statement is a tautology, that is, it is a true statement irrespective of the truthvalues of the component statements p and q. (In particular, it is true when both premises have truthvalue T.) Thus the argument presented above is indeed a valid argument.
26 Aristotle An argument that does not lead to a tautology in symbolic logic is invalid. For example, If a bird is a crow, then it is black. This bird is black. Therefore it is a crow. ∨ is an invalid argument: ((p → q) q) → p is not a tautology. (Informally, we can assert that a black bird need not be a crow.) The following table contains the standard forms of argument commonly used, along with some invalid arguments commonly used in error.
Valid Arguments
Invalid Arguments
Direct Reasoning (modus ponens) p→q p therefore q
Fallacy of the Converse
Contrapositive Reasoning (modus tollens) p→q ¬q therefore ¬p
Fallacy of the Inverse p→q ¬p therefore ¬q
Disjunctive Reasoning p∨q p∨q ¬p ¬q therefore q therefore p
Misuse of Disjunctive Reasoning p∨q p∨q p q therefore ¬q therefore ¬p
p→q q therefore p
Transitive Reasoning p→q q→r therefore p → r
In the mid1700s LEONHARD EULER invented an elegant way to determine the validity of syllogisms, that is, arguments whose premises contain the words all, some, or no. For example, All poodles are dogs. All dogs bark. Therefore all poodles bark. is a syllogism, and Euler would depict such an argument as a diagram of three circles, each representing a set mentioned in one of the premises. The validity of the argument is then readily apparent:
Euler diagrams
An argument of the following structure, for example, can be demonstrated as invalid by arranging circles as shown: All As are Bs. Some Bs are Cs. Therefore, some As are Cs. Any diagram used to analyze the validity of an argument is called an Euler diagram. See also DEDUCTIVE/INDUCTIVE REASONING; QUANTIFIER.
Aristotle (384–322 B.C.E.) Greek Logic, Philosophy, Physics, Medicine Born in Stagirus, Macedonia, Aristotle is remembered in mathematics for his systematic study of deductive logic. In laying down the foundations of FORMAL LOGIC, Aristotle identified the fundamental LAWS OF THOUGHT, the laws of reasoning, and the fundamental principles that lie at the heart of any mathematical ARGUMENT. His work in this area so deeply affected the attitudes and approaches of scientific thinking that Western intellectual culture as a whole is often referred to as Aristotelian. At age 17, Aristotle joined PLATO’s Academy in Athens and remained there for 20 years. He worked closely with Plato, and also EUDOXUS, nephew of Plato. The equivalent of a modernday research university, the Academy brought together scholars from all disciplines and provided a culturally rich environment that encouraged learning and promoted the advancement of knowledge. Due to internal politics, however, Aristotle decided to leave the Academy after Plato’s death in 347 B.C.E.
arithmetic 27 After tutoring the heir of King Phillip II, the future Alexander the Great, for a number of years, Aristotle returned to Athens in 335 B.C.E. to found his own school, the Lyceum. He intended the school to be as broadbased as possible, exploring a wide range of subjects but with prominence given to the study of the natural world. While at the Lyceum, Aristotle wrote 22 texts covering an astonishing range of topics: logic, physics, astronomy, meteorology, theology, metaphysics, ethics, rhetoric, poetics, and more. He founded a theory of kinematics, a study of space, time, and motion, and he established principles of physics that remained unchallenged for two millennia. With regard to mathematics, Aristotle is remembered for his writings in logic, a subject he identified as the basis of all scientific thought. He invented the syllogism, a form of argument that comes in three parts: a major premise, a minor premise, and a conclusion. Although a straightforward notion for us today, this work represented a first fundamental step toward understanding the structure of reasoning. He presented the following line of thought as an example of a syllogism: Every Greek is a person. Every person is mortal. Therefore every Greek is mortal. Aristotle recognized that any line of reasoning following this form is logically valid by virtue of its structure, not its content. Thus the argument: Every planet is made of cheese. Every automobile is a planet. Therefore every automobile is made of cheese. for example, is a valid argument, even though the validity of the premises may be in question. Removing content from structure was a sophisticated accomplishment. Aristotle called his field of logic “analytics” and described his work on the subject in his book Prior and Posterior Analytics. He wanted to demonstrate the effectiveness of logical reasoning in understanding science. Aristotle also discussed topics in the philosophy of mathematics. He argued, for instance, that an unknowable such as “infinity” exists only as a potentiality, and never as a completed form. Although, for example, from any finite set of prime numbers one can always construct one more, speaking of the set of prime numbers as a single concept, he argued, is meaningless.
(Today we say that Aristotle accepted the “potentially infinite” but rejected the “actual infinite.”) It is recorded that Aristotle would often walk through the gardens of the Lyceum while lecturing, forcing his pupils to follow. His students became known as the peripatetics, the word peripatetic meaning “given to walking.” Copies of Aristotle’s lecture notes taken by the peripatetics were regarded as valuable scholarly documents in their own right and have been translated, copied, and distributed across the globe throughout the centuries. Political unrest forced Aristotle to leave Athens again in 322 B.C.E. He died soon afterward at the age of 62 of an unidentified stomach complaint. Aristotle’s analysis of critical thinking literally shaped and defined the nature of logical thought we exercise today in any academic pursuit. One cannot exaggerate the profundity of Aristotle’s influence. By identifying valid modes of thought and clarifying the principles of logical reasoning, Aristotle provided the tools necessary for sensible reasoning and astute systematic thinking. These are skills today deemed fundamental to basic goals of all levels of education. See also CARDINALITY; DEDUCTIVE/INDUCTIVE REASONING; PARADOX.
arithmetic The branch of mathematics concerned with computations using numbers is called arithmetic. This can involve a number of specific topics—the study of operations on numbers, such as ADDITION, MULTIPLICATION, SUBTRACTION, DIVISION, and SQUARE ROOTs, needed to solve numerical problems; the methods needed to change numbers from one form to another (such as the conversion of fractions to decimals and vice versa); or the abstract study of the NUMBER SYSTEMS, NUMBER THEORY, and general operations on sets as defined by GROUP THEORY and MODULAR ARITHMETIC, for instance. The word arithmetic comes from the Greek work arithmetiké, constructed from arithmós meaning “number” and techné meaning “science.” In the time of ancient Greece, the term arithmetic referred only to the theoretical work about numbers, with the word logistic used to describe the practical everyday computations used in business. Today the term arithmetic is used in both contexts. (The word logistics is today a predominantly military term.) See also BASE OF A NUMBER SYSTEM; FUNDAMENTAL THEOREM OF ARITHMETIC; ORDER OF OPERATION.
28 arithmetic–geometricmean inequality
arithmetic–geometricmean inequality See MEAN. arithmetic mean See MEAN. arithmetic sequence (arithmetic progression) A
SEof numbers in which each term, except the first, differs from the previous one by a constant amount is called an arithmetic sequence. The constant difference between terms is called the common difference. For example, the sequence 4, 7, 10, 13, … is arithmetic with common difference 3. An arithmetic sequence can be thought of as “linear,” with the common difference being the SLOPE of the linear relationship. An arithmetic sequence with first term a and common difference d has the form: QUENCE
a, a + d, a + 2d, a + 3d, … The nth term an of the sequence is given by an = a + (n – 1)d. (Thus, the 104th term of the arithmetic sequence 4,7,10,13,…, for example, is a104 = 4 + [103 × 3] = 313.) The sum of the terms of an arithmetic sequence is called an arithmetic series: a + (a + d) + (a + 2d) + (a + 3d)+… The value of such a sum is always infinite unless the arithmetic sequence under consideration is the constant zero sequence: 0,0,0,0,… The sum of a finite arithmetic sequence can be readily computed by writing the sum both forward and backward and summing columnwise. Consider, for example, the sum 4 + 7 + 10 + 13 + 16 + 19 + 22 + 25 + 28 + 31. Call the answer to this problem S. Then: 4 + 7 + 10 + 13 + 16 + 19 + 22 + 25 + 28 + 31 = S 31 + 28 + 25 + 22 + 19 + 16 + 13 + 10 + 7 + 4 = S and adding columns yields: 35 + 35 + 35 + 35 + 35 + 35 + 35 + 35 + 35 + 35 = 2S That is, 2S = 10 × 35 = 350, and so S = 175. In general, this method shows that the sum of n equally spaced numbers in arithmetic progression, a + b + … + y + z, is n times the average of the first and last terms of the sum:
S =n×
a+z 2
It is said that CARL FRIEDRICH GAUSS (1777–1855), as a young student, astonished his mathematics instructor by computing the sum of the numbers 1 though 100 in a matter of seconds using this method. (We have 1 + 2 1 + 100 + … + 100 = 100 × = 5,050.) 2 See also GEOMETRIC SEQUENCE; SERIES.
arithmetic series See ARITHMETIC SEQUENCE.
array An ordered arrangement of numbers or symbols is called an array. For example, a VECTOR is a onedimensional array: it is an ordered list of numbers. Each number in the list is called a component of the vector. A MATRIX is a twodimensional array: it is a collection of numbers arranged in a finite grid. (The components of such an array are identified by their row and column positions.) Two arrays are considered the same only if they have the same number of rows, the same number of columns, and all corresponding entries are equal. One can also define three and higherdimensional arrays. In computer science, an array is called an identifier, and the location of an entry is given by a subscript. For example, for a twodimensional array labeled A, the entry in the second row, third column is denoted A23. An ndimensional array makes use of n subscripts. _ Aryabhata (ca. 476–550
C.E.) Indian Trigonometry, Number theory, Astronomy Born in Kusumapura, – now Patna, India, A ryabhata (sometimes referred to as – Aryabhata I to distinguish him from the mathematician of the same name who lived 400 years later) was the first Indian mathematician of note whose name we know and whose writings we can study. In the section Ganita (Calculation) of his astronomical treatise – Aryabhatiya, he made fundamental advances in the theory of TRIGONOMETRY by developing sophisticated techniques for finding and tabulating lengths of halfchords in circles. This is equivalent to tabulating values – of the sine function. Aryabhata also calculated the value of π to four decimal places (π ≈ 62,832/20,000 = 3.1416) and developed rules for extracting square and
asymptote 29 cube roots, for summing ARITHMETIC SERIES, and finding SUMS OF POWERS. – As an astronomical treatise, Aryabhatiya is written as a series of 118 verses summarizing all Hindu mathematics and astronomical practices known at that time. A number of sections are purely mathematical in context and cover the topics of ARITHMETIC, TRIGONOMETRY, and SPHERICAL GEOMETRY, as well as touch on the theories of CONTINUED FRACTIONs, QUADRATIC equa– tions, and SUMMATION. Aryabhata also described methods for finding integer solutions to linear equations of the form by = ax + c using an algorithm essentially equivalent to the EUCLIDEAN ALGORITHM. – Historians do not know how Aryabhata obtained his highly accurate estimate for π. They do know, how– ever, that Aryabhata was aware that it is an IRRATIONAL NUMBER, a fact that mathematicians were not able to prove until 1775, over two millennia later. In – practical applications, however, Aryabhata preferred — to use √10 ≈ 3.1622 as an approximation for π. Scholars at the time did not think of sine as a ratio of sidelengths of a triangle, but rather the physical length of a halfchord of a circle. Of course, circles of different radii give different lengths for corresponding halfchords, but one can adjust figures with the use of proportionality. Working with a circle of radius 3,438, – Aryabhata constructed a table of sines for each angle from 1° to 90°. (He chose the number 3,438 so that the circumference of the circle would be close to 21,600 = 360 × 60, making one unit of length of the circumference matching one minute of an angle.) Thus, in his table, sine of 90° is recorded as 3,438, and the sine of 30°, for example, as 1,719. – With regard to astronomy, Aryabhatiya presents a systematic treatment of the position and motions of the – planets. Aryabhata calculated the circumference of the Earth as 24,835 miles (which is surprisingly accurate) and described the orbits of the planets as ELLIPSEs. European scholars did not arrive at the same conclusion until the Renaissance.
associative A BINARY OPERATION is said to be associative if it is independent of the grouping of the terms to which it is applied. More precisely, an operation * is associative if: a * (b * c) = (a * b) * c
for all values of a, b, and c. For example, in ordinary arithmetic, the operations of addition and multiplication are associative, but subtraction and division are not. For instance, 6 + (3 + 2) and (6 + 3) + 2 are equal in value, but 6 – (3 – 2) and (6 – 3) – 2 are not. (The first equals 6 – 1 = 5, and the second is 3 – 2 = 1.) In VECTOR analysis, the addition of vectors is associative, but the operation of taking CROSS PRODUCT is not. From the basic relation a * (b * c) = (a * b) * c, it follows that all possible groupings of a finite number of fixed terms by parentheses are equivalent. (Use an INDUCTION argument on the number of elements present.) For example, that (a * b) * (c * d) equals (a * (b * c)) * d can be established with two applications of the fundamental relation as follows: (a * b) * (c * d) = ((a * b) * c) * d = (a * (b * c)) * d. As a consequence, if the associative property holds for a given set, parentheses may be omitted when writing products: one can simply write a * b * c * d, for instance, without concern for confusion. These considerations break down, however, if the expression under consideration contains an infinite number of terms. For instance, we have: 0=0+0+0+… = (1 – 1) + (1 – 1) + (1 – 1) + … If it is permissible to regroup terms, then we could write: 0 = 1 + (–1 + 1) + (–1 + 1) + (–1 + 1) + … =1+0+0+0+… =1 This absurdity shows that extreme care must be taken when applying the associative law to infinite sums. See also COMMUTATIVE PROPERTY; DISTRIBUTIVE PROPERTY; RING.
asymptote A straight line toward which the graph of a function approaches, but never reaches, is called an asymptote for the graph. The name comes from the Greek word asymptotos for “not falling together” (a: “not;” sym: “together;” ptotos: “falling”). For example, the function y = 1/x has the lines x = 0 and y = 0 as asymptotes: y becomes infinitely small, but never reaches zero, as x becomes large, and vice versa. The function y = (x + 2)/(x – 3) has the vertical line x = 3
30 automaton as asymptote: values of the function become infinitely large as x approaches the value 3 from the right, and infinitely large and negative as x approaches the value 3 from the left. A function y = f(x) has a horizontal asymptote y = L if limx→∞ f(x) = L or limx→–∞ f(x) = L. For example, the function
y=
3x 2 has horizontal asympx2 + x + 1
tote y = 3, since lim x→∞ =
3x 2 = lim x→∞ x2+ x + 1
3 = 3. 1+ 0 + 0
3 1 1 1+ + 2 x x
An asymptote need not be horizontal or vertical, 3 2 however. For example, the function y = x + x + x + 2 x2 + 1
one place to the left. Thus, for instance, six pennies placed into the 1 ← 2 machine “fire” four times to yield a final distribution that can be read as “1 1 0.” This result is the number six written as a BINARY NUMBER, and this machine converts all numbers to their basetwo representations. A 1 ← 3 machine yields basethree representations, and a 1 ← 10 machine yields the ordinary baseten representations. The process of LONG DIVISION can be explained with the aid of this machine. Variations on this idea can lead to some interesting mathematical studies. Consider, for example, a 2 ← 3 machine. This machine replaces three pennies in one box with two pennies in the box one place to the left. In some sense, this is a “base one and a half machine.” For instance, placing 10 pennies in this machine yields
1 approaches the line y = x + 1 as x x +1 becomes large, thus y = x + 1 is a “slant asymptote” for the curve. 2 2 If a HYPERBOLA is given by the equation x − y = 1, a 2 b2 = x +1+
2
2
2
b y b2 then manipulating yields the equation − = 2 . a x x The right hand side tends to zero as x becomes large, showing that the curve has slant asymptotes given by 2
2
b b y − = 0 , that is, by the lines y = a x and a x b y=− x. a Extending the definition, we could say that the 3 curve y = x + 1 = x 2 + 1 has the parabola y = x2 as x x an asymptote.
automaton (plural, automata) An abstract machine used to analyze or model mathematical problems is called an automaton. One simple example of an automaton is a “numberbase machine,” which consists of a row of boxes extending infinitely to the left. One places in this machine a finite number of pennies in the rightmost box. The machine then redistributes the pennies according to a preset rule. A “1 ← 2” machine, for example, replaces a pair of pennies in one box with a single penny in the box
An automaton in action
axiom of choice 31 the representation “2 1 0 1,” and it is true that 3
2
1
0
3 3 3 3 2 × + 1 × + 0 × + 1 × = 10 . 2 2 2 2 See also BASE OF A NUMBER SYSTEM.
axiom (postulate) A statement whose truth is deemed selfevident or to be accepted without proof is called an axiom. The name comes from the Greek word axioma for “worth” or “quality.” The alternative name “postulate” comes from postulatum, Latin for “a thing demanded.” One of the great achievements of the great Greek geometer EUCLID and his contemporaries of around 300 B.C.E. was to recognize that not every statement in mathematics can be proved: certain terms remain undefined, and basic rules (postulates) about their relationships must simply be accepted as true. One must develop a mathematical theory with a “big bang,” as it were, by simply listing a starting set of assumptions. From there, using the basic laws of reasoning, one then establishes and proves further statements, or THEOREMs, about the system. For example, in a systematic study of EUCLIDEAN GEOMETRY, the terms point, line, and plane are undefined, and one begins a systematic study of the subject by studying a list of basic axioms that tells us how these quantities are meant to interrelate. (One axiom of Euclidean geometry, for instance, asserts that between any two points one can draw a line.) All the results presented in a typical highschool text on geometry, for example, are logical consequences of just five principal assumptions. In SET THEORY, the terms set and element of a set are undefined. However, rules are given that define the equality of two sets, that guarantee the existence of certain sets, and establish the means of constructing new sets from old ones. In NUMBER THEORY, PEANO’S POSTULATES provide a logical foundation to the theory of numbers and arithmetic.
A statement in a mathematical system that appears true, but has not yet been proved, is called a conjecture. See also DEDUCTIVE/INDUCTIVE REASONING; ERNST FRIEDRICH FERDINAND ZERMELO.
axiom of choice First formulated by German mathematician ERNST FRIEDRICH FERDINAND ZERMELO (1871–1953), the axiom of choice is a basic principle of SET THEORY that states that from any given collection C of nonempty sets, it is possible to construct a set S that contains one element from each of the sets in C. The set S is called a “choice set” for C. For example, if C represents the three sets {1,2,3}, {2,4,6,8,…}, and {5}, then S = {1,6,5} is a choice set for C. So too is the set S = {2,5}. The axiom of choice has been considered counterintuitive when interpreted on a practical level: although it is possible to select one element from each of a finite collection of sets in a finite amount of time, it is physically impossible to accomplish the same feat when presented with an infinite collection of sets. The existence of a choice set is not “constructive,” as it were, and use of the axiom is viewed by mathematicians, even today, with suspicion. In 1938 Austrian mathematician KURT GÖDEL proved, however, that no contradiction would ever arise when the axiom of choice is used in conjunction with other standard axioms of set theory. Zermelo formulated the axiom to prove that every ordered set can be wellordered. The axiom of choice also proves (and in fact is equivalent to) the trichotomy law, which states that for any pair of REAL NUMBERS a and b, precisely one of the following holds: i. a > b ii. a < b iii. a = b Although this statement, on one level, appears obvious, its validity is fundamental to the workings of the real numbers and so needs to be properly understood. See also WELLORDERED SET.
B Babbage, Charles (1791–1871) British Computation Born on December 26, 1791, in London, England, Charles Babbage is best remembered for his work on the design and manufacture of a mechanical calculator, the forerunner of a computer. After first constructing a “difference machine,” Babbage devoted the remainder of his life to the construction of a superior “analytic engine” capable of performing all mathematical operations. His work toward this goal laid the foundations of computer design used today. Partly due to lack of funding, however, the machine was never completed. Babbage entered Trinity College, Cambridge, in 1810. While a student, he and a fellow undergraduate coauthored Differential and Integral Calculus, an influential memoir on the history of calculus. After transferring to Peterhouse College, Babbage received his bachelor’s degree in mathematics in 1814 to then begin a career in mathematical research. Babbage published a number of influential papers on the topic of functional equations and was honored with election to the ROYAL SOCIETY in 1816. In 1827 he became the Lucasian Professor of Mathematics at Cambridge. Much of Babbage’s theoretical work relied on consulting tables of logarithms and trigonometric functions. Aware of the inaccuracy of human calculation, Babbage became interested in the problem of using a mechanical device to perform complex computations. In 1819 he began work on a small “difference engine,” which he completed three years later. He announced his invention to the scientific community in an 1822 paper, “Note on the Application of
Charles Babbage, an eminent mathematician of the 19th century, is best known for his design and manufacture of a mechanical computer. (Photo courtesy of the Science Museum, London/ TophamHIP/The Image Works)
32
Babylonian mathematics 33 he published a book, On the Economy of Machinery and Manufactures, offering a theoretical discussion on the topic. This could be considered the first published work in the field of OPERATIONS RESEARCH. Unfortunately, due to financial and technological difficulties, the machine was never completed. (The metalwork technology of the mid1800s was not capable of the levels of precision Babbage’s machine demanded.) The device in its unfinished state is preserved today in the Science Museum of London. Although he never realized his dream of building an operational, mechanical computer, his design concepts have since been proved correct. It is not an exaggeration to say that the modern computers constructed on Babbage’s theoretical design have revolutionized almost all aspects of 20thcentury life. Babbage died in London, England, on October 18, 1871.
Babylonian mathematics The Babylonians of 2000 Charles Babbage completed work on his “difference engine,” the world’s first sophisticated mechanical computer, in 1822. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
Machinery to the Computation of Astronomical and Mathematical Tables.” Although the machine was capable of performing relatively simple, but highly accurate, computations (using the method of FINITE DIFFERENCES to compute values of POLYNOMIAL functions), his invention was well received and was understood to be a first step toward a new era in computational capabilities. Babbage was awarded a gold medal from the Astronomical Society and was given a grant from the Chancellor of the Exchequer to construct a larger, more complex, difference engine. In 1801 JosephMarie Jacquard invented a loom capable of weaving complex patterns by making use of a set of instructions set out on cards punched with holes. Two decades later Babbage decided to follow the same idea and design a steampowered engine that would accept instructions and data from punched cards. With the assistance of Lord Byron’s daughter, LADY AUGUSTA ADA LOVELACE, Babbage took to work on creating a sophisticated calculating device. In 1832
B.C.E. lived in Mesopotamia, the fertile plain between the Euphrates and Tigris Rivers in what is now Iraq. We are fortunate that the peoples of this region kept extensive records of their society—and their mathematics—on hardy sunbaked clay tablets. A large number of these tablets survive today. The Babylonians used a simple stylus to make marks in the clay and developed a form of writing based on cuneiform (wedge shaped) symbols. The mathematical activity of the Babylonians seems to have been motivated, at first, by the practical everyday needs of running their society. Many problems described in early tablets are concerned with calculating the number of workers needed for building irrigation canals and the total expense of wages, for instance. But many problems described in later texts have no apparent practical application and clearly indicate an interest in pursuing mathematics for its own sake. The Babylonians used only two symbols to represent numbers: the symbol to represent a unit and the
symbol to represent a group of ten. A simple additive system was used to represent the numbers 1 through 59. For example, the cluster represents “32.” A base60 PLACEVALUE SYSTEM was then used to represent numbers greater than 59. For instance, the number 40,992, which equals 11 × 602 + 23 × 60 + 12, was written: between clusters of symbols.
. Spaces were inserted
34 Babylonian mathematics Historians are not clear as to why the Babylonians chose to work with a SEXAGESIMAL system. A popular theory suggests that this number system is based on the observation that there are 365 days in the year. When rounded to the more convenient (highly divisible) value of 360, we have a multiple of 60. Vestiges of this number system remain with us today. For example, we use the number 360 for the number of degrees in a circle, and we count 60 seconds in a minute and 60 minutes per hour. There were two points of possible confusion with the Babylonian numeral system. With no symbol for zero, it is not clear whether the numeral represents 61 (as one unit of 60 plus a single unit), 3601 (as one unit of 602 plus a single unit), or even 216,060, for instance. Also, the Babylonians were comfortable with
A seventhcentury cuneiform tablet from northern Iraq records observations of the planet Venus. (Photo courtesy of the British Museum/TophamHIP/The Image Works)
fractions and used negative powers of 60 to represent them (just as we use negative powers of 10 to write fractions in decimal notation). But with no notation for the equivalent of a decimal point, the symbol could also be interpreted to mean 1 + (1/60), or (1/60) + (1/602), or even 60 + (1/604), for instance. As the Babylonians never developed a method for resolving such ambiguity, we assume then that it was never considered a problem for scholars of the time. (Historians suggest that the context of the text always made the interpretation of the numeral apparent.) The Babylonians compiled extensive tables of powers of numbers and their reciprocals, which they used in ingenious ways to perform arithmetic computations. (For instance, a tablet dated from 2000 B.C.E. lists all the squares of the numbers from one to 59, and all the cubes of the numbers from one to 32.) To compute the product of two numbers a and b, Babylonian scholars first computed their sum and their difference, read the squares of those numbers from a table, and divided their difference by four. (In modern notation, this corresponds to the computation: ab = (1/4) [(a + b)2 – (a – b)2].) To divide a number a by b, scholars computed the product of a and the reciprocal 1/b (recorded in a table): ab = a × (1/b). The same table of reciprocals also provided the means to solve LINEAR EQUATIONs: bx = a. (Multiply a by the reciprocal of b.) Problems in geometry and the computation of area often lead to the need to solve QUADRATIC equations. For instance, a problem from one tablet asks for the width of a rectangle whose area is 60 and whose length is seven units longer than the width. In modern notation, this amounts to solving the equation x(x + 7) = x2 + 7x = 60. The scribe who wrote the tablet then proffers a solution that is equivalent to the famous quadratic formula: x = √(7/2)2 + 60 – (7/2) = 5. (Square roots were computed by examining a table of squares.) Problems about volume lead to cubic equations, and the Babylonians were adept at solving special equations of the form: ax3 + bx2 = c. (They solved these by setting n = (ax)/b, from which the equation can be rewritten as n3 + n2 = ca2/b3. By examining a table of values for n3 + n2, the solution can be deduced.) It is clear that Babylonian scholars knew of PYTHAGORAS’S THEOREM, although they wrote no general proof of the result. For example, a tablet now housed in the British museum, provides the following problem and solution:
Barrow, Isaac 35 If the width of a rectangle is four units and the length of its diagonal is five units, what is its breadth? Four times four is 16, and five times five is 25. Subtract 16 from 25 and there remains nine. What times what equals nine? Three times three is nine. The breadth is three. The Babylonians used Pythagoras’s theorem to compute the diagonal length of a square, and they found an approximation to the square root of two accurate to five decimal places. (It is believed that they used a method analogous to HERON’S METHOD to do this.) Babylonian scholars were also interested in approximating the areas and volumes of various common shapes by using techniques that often invoked Pythagoras’s theorem. Most remarkable is a tablet that lists 15 large PYTHAGOREAN TRIPLES. As there is no apparent practical need to list these triples, this strongly suggests that the Babylonians did indeed enjoy mathematics for its own sake. See also BASE VALUE OF A NUMBER SYSTEM.
Banach, Stefan (1892–1945) Polish Analysis, Topology Born on March 30, 1892, in Kraków, now in Poland, Stefan Banach is noted for his foundational work in ANALYSIS and for generalizing the notion of a VECTOR SPACE to a general theory of a space of functions. This fundamental work allows mathematicians today to develop a theory of FOURIER SERIES, in some sense, in very abstract settings. Banach is also remembered for his work leading to the famous BanachTarski paradox that arises in the study of AREA and volume. Banach began his scholarly career with a university degree in engineering from Lvov Technical University. His academic plans were interrupted, however, with the advent of World War I. During this time Banach was forced to work building roads, although he did manage to find time to also teach at local Kraków schools during this period. Soon after the war Banach joined a mathematics discussion group in Lvov and soon impressed mathematical scholars with his abilities to solve mathematical problems. Within a week of joining the group, Banach had drafted a coauthored research paper on the topic of measure theory, a theory that generalizes the concept of area. Banach continued to
produce important results in this field at an extremely rapid rate thereafter. Banach was offered a lectureship at Lvov Technical University in 1920 and quickly set to work on a doctoral thesis. Despite having no previous official university qualifications in mathematics, Banach was awarded a doctorate in 1922 by the Jan Kazimierz University in Lvov. Banach’s contributions to mathematics were significant. His generalized work on Fourier series founded a branch of mathematics now called functional analysis. It has connections to the fields of measure theory, integration, and SET THEORY. He and his colleague Alfred Tarski presented their famous paradoxical result in 1926 in their paper “Sur la décomposition des ensembles de points en partiens respectivement congruent” (On the decomposition of figures into congruent parts). His 1932 paper, “Théorie des opérations linéaires” (Theory of linear operators), which develops the notion of a normed VECTOR SPACE (that is, a vector space with a notion of length attached to its vectors), is deemed his most influential work. As well as conducting research in mathematics, Banach also wrote arithmetic, algebra, and geometry texts for highschool students. In 1939 Banach was elected as president of the Polish Mathematical Society. Banach was allowed to maintain his university position during the Soviet occupation later that year, but conditions changed with the 1941 Nazi invasion. Many Polish academics were murdered, but Banach survived, although he was forced to work in a German infectious diseases laboratory, given the task to feed and maintain lice colonies. He remained there until June 1944, but he became seriously ill by the time Soviet troops reclaimed Lvov. Banach died of lung cancer on August 31, 1945. Banach’s name remains attached to the type of vector space he invented, and research in this field of functional analysis continues today. The theory has profound applications to theoretical physics, most notably to quantum mechanics.
bar chart (bar graph) See STATISTICS: DESCRIPTIVE.
Barrow, Isaac (1630–1677) British Calculus, Theology Born in London, England, (his exact birth date is
36 base of a logarithm EUCLID’s THE ELEMENTS in 1655. He became professor of geometry at Gresham College, London, in 1662, and was elected as one of the first 150 fellows of the newly established ROYAL SOCIETY in 1663. He returned to Cambridge that same year to take the position of Lucasian Chair of Mathematics, at Trinity College, and worked hard to improve the standards of mathematics education and interest in mathematical research at Cambridge. With this aim in mind, Barrow gave a series of lectures on the topics of optics, geometry, NUMBER THEORY, and the nature of time and space. His discussions on geometry proved to be highly innovative and fundamentally important for the new perspective they offered. Newton advised Barrow to publish the notes. In 1669 Barrow resigned from the Lucasian Chair to allow Newton to take over, and he did no further mathematical work. He died in London, England, on May 4, 1677, of a malignant fever. Barrow’s influence on modernday mathematics is oblique. His effect on the development of the subject lies chiefly with the inspiration he provided for others. Isaac Barrow, a mathematician of the 17th century, is noted chiefly for the inspiration he provided others in the development of the theory of calculus. He may have been the first scholar to recognize and understand the significance of the fundamental theorem of calculus. (Photo courtesy of ARPL/Topham/The Image Works)
not known) Isaac Barrow is remembered in mathematics for his collection of lecture notes Lectiones geometricae (Geometrical lectures), published in 1670, in which he describes a method for finding tangents to curves similar to that used today in DIFFERENTIAL CALCULUS. Barrow may have also been the first to realize that the problem of finding tangents to curves is the inverse problem to finding areas under curves. (This is THE FUNDAMENTAL THEOREM OF CALCULUS.) The lectures on which his notes were based were extremely influential. They provided SIR ISAAC NEWTON, who attended the lectures and had many private discussions with Barrow, a starting point for his development of CALCULUS. Barrow graduated from Trinity College, Cambridge, with a master’s degree in 1652, but was dissatisfied with the level of mathematics instruction he had received. After leaving the college, Barrow taught himself GEOMETRY and published a simplified edition of
base of a logarithm See LOGARITHM. base of an exponential See
EXPONENT; EXPONEN
TIAL FUNCTION.
base of a number system (radix, scale of a number system) The number of different symbols used, perhaps in combination, to represent all numbers is called the base of the number system being used. For example, today we use the ten symbols 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9 to denote all numbers, making use of the position of these digits in a given combination to denote large values. Thus we use a base10 number system (also called a decimal representation system). We also use a placevalue system to give meaning to the repeated use of symbols. When we write 8,407, for instance, we mean eight quantities of 1,000 (103), four quantities of 100 (102), and seven single units (100). The placement of each DIGIT is thus important: the number of places from the right in which a digit lies determines the power of 10 being considered. Thus the numbers 8, 80, 800, and 8,000, for instance, all represent different quantities. (The system of ROMAN NUMERALS, for example, is not a placevalue system.)
base of a number system 37 For any positive whole number b, one can create a placevalue notational system of that base as follows: Write a given number n as a sum of powers of b: n = akbk + ak–1 bk–1 + … + a2b2 + a1b + a0 with each number ai satisfying 0 ≤ ai < b. Then the base b representation of n is the kdigit quantity ak ak–1…a2a1a0. Such a representation uses only the symbols 0, 1, 2,…, b – 1. For example, to write the number 18 in base four— using the symbols 0, 1, 2, and 3—observe that 18 = 1 × 42 + 0 × 4 + 2 × 1, yielding the base4 representation: 102. In the reverse direction, if 5,142 is the base6 representation of a number n, then n is the number 5 × 63 + 1 × 62 + 4 × 6 + 2 × 1 = 1,142. One may also make use of negative powers of the base quantity b. For example, using a decimal point to separate positive and negative powers of ten, the number 312.407, for instance, represents the fractional 1 1 quantity: 3 × 102 + 1 × 10 + 2 × 1 + 4 × 10 + 0 × 102 1 + 7 × 103 . In base 4, the number 33.22 is the quantity 1 1 2 2 3 × 4 + 3 × 1 + 2 × 4 + 2 × 42 = 15 + 4 + 16 , which is 15.625 in base 10. The following table gives the names of the placevalue number systems that use different base values b. The Babylonians of ancient times used a sexagesimal system, and the Mayas of the first millennium used a system close to being purely vigesimal.
base b 2 3 4 5 6 7 8 9 10 11 12 16 20 60
The representation of numbers can be wellrepresented with the aid of a simple AUTOMATON called a numberbase machine. Beginning with a row of boxes extending infinitely to the left, one places in the rightmost box a finite number of pennies. The automaton then redistributes the pennies according to a preset rule. A “1 ← 2” machine, for example, replaces a pair of pennies in one box with a single penny in the box one place to the left. Thus, for instance, six pennies placed into the 1 ← 2 machine “fire” four times to yield a final distribution that can be read as “1 1 0.” This result is the number six written as a BINARY NUMBER and this machine converts all numbers to their basetwo representations. (The diagram in the entry for automaton illustrates this.) A 1 ← 3 machine yields basethree representations, and a 1 ← 10 machine yields the ordinary baseten representations. Long Division The process of long division in ARITHMETIC can be explained with the aid of a numberbase machine. As an example, let us use the 1 ← 10 machine to divide the number 276 by 12. Noting that 276 pennies placed in the 1 ← 10 machine yields a diagram with two pennies in the 100s position, seven pennies in the 10s position, and six pennies in the units position, and that 12 pennies appears as one penny in a box with two pennies in the box to its right, to divide 276 by 12, one must simply look for “groups of 12” within the diagram of 276 pennies and keep count of the number of groups one finds.
number system Binary Ternary Quaternary Quinary Senary Septenary Octal Nonary Decimal Undenary Duodecimal Hexadecimal Vigesimal Sexagesimal Long division base ten
38 base of a polygon/polyhedron powers of five, and “unfire” a group of five pennies, one can continue the long division process to compute, in base 5, that 1432 ÷ 13 = 110.1111… See also BABYLONIAN MATHEMATICS; BINARY NUMBERS; DIGIT; MAYAN MATHEMATICS; NESTED MULTIPLICATION; ZERO.
base of a polygon/polyhedron The base of a trian
Long division base five
Notice that we find two groups of 12 at the 10s position (that is, two groups of 120), and three groups of 12 at the units position. Thus: 276 ÷ 12 = 23. The standard algorithm taught to school children is nothing more than a recording system for this process of finding groups of twelve. Notice too that one does not need to know the type of machine, that is, the base of the number system in which one is working in order to compute a longdivision problem. If we simply write the base number as x, and work with a 1 ← x machine, then the same computation provides a method for dividing polynomials. In our example, we see that: (2x2 + 7x + 6) ÷ (x + 2) = 2x + 3 Thus the division of polynomials can be regarded as a computation of long division. There is a technical difficulty with this: a polynomial may have negative coefficients, and each negative coefficient would correspond to a negative number of pennies in a cell. If one is willing to accept such quantities, then the numberbase machine model continues to work. (Note, in this extended model, that one can insert into any cell an equal number of positive and negative pennies without changing the system. Indeed, it might be necessary to do this in order to find the desired groups of pennies.) The process of long division might produce non zero remainders. For example, in base 5, dividing 1432 by 13 yields the answer 110 with a remainder of 2 units. (In base 10, this reads: 242 ÷ 8 = 30 with a remainder of 2.) If one is willing to work with negative
gle, or of any POLYGON, is the lowest side of the figure, usually drawn as a horizontal edge parallel to the bottom of the page. Of course other edges may be considered the base if one reorients the figure. The base of a POLYHEDRON, such as a cube or a pyramid, is the lowest FACE of the figure. It is the face on which the figure would stand if it were placed on a tabletop. The highest point of a geometric figure opposite the base is called the APEX of the figure, and the distance from the base to the apex is called the height of the figure.
basis See LINEARLY DEPENDENT AND INDEPENDENT.
Bayes, Rev. Thomas (1702–1761) English Probability, Theology Born in London, England, (the exact date of his birth is not known), theologian and mathematician Reverend Thomas Bayes is best remembered for his influential article “An Essay Towards Solving a Problem in the Doctrine of Chances,” published posthumously in 1763, that outlines fundamental principles of PROBABILITY theory. Bayes developed innovative techniques and approaches in the theory of statistical inference, many of which were deemed controversial at the time. His essay sparked much further research in the field and was profoundly influential. The work also contains the famous theorem that today bears his name. An ordained minister who served the community of Tunbridge Wells, Kent, England, Bayes also pursued mathematics as an outside interest. As far as historians can determine, he published only two works during his lifetime. One was a theological essay in 1731 entitled “Divine Benevolence, or an Attempt to Prove that the Principal End of the Divine Providence and Government is the Happiness of His Creatures.” The other was a mathematical piece that he published anonymously in 1736, “Introduction to the Doctrine of Fluxions, and a
Bernoulli family 39 Defense of the Mathematicians Against the Objections of the Author of The Analyst,” defending the logical foundations of SIR ISAAC NEWTON’s newly invented CALCULUS. Despite the apparent lack of published mathematical work, Bayes was nonetheless elected a fellow of the prestigious academic ROYAL SOCIETY in 1742. Bayes retired from the ministry in 1752 but remained in Tunbridge Wells until his death on April 17, 1761. His friend, Richard Price, discovered the nowfamous paper on probability theory among his belongings and submitted it for publication. A second paper, “A Letter on Asymptotic Series from Bayes to John Canton,” one on asymptotic series, was also published after Bayes’s death. The theoretical approach of inferential statistics Bayes proposed remains an active area of research today. See also BAYES’S THEOREM; STATISTICS: INFERENTIAL.
Here let A be the event “a red ball is chosen” and B1 and B2 the events “a ball is selected from bag 1 / bag 2,” respectively. Then P(B1) = 1/2 = P(B2), P(AB1) = 5/7, and P(AB2) = 7/11. Thus the probability we seek, P(B1A), is given by:
P(B1  A) =
P(A  B1)P(B1) P(A  B1)P(B1) + P(A  B2 )P(B2 )
5 1 × 7 2 = 5 1 7 1 × + × 7 2 11 2 =
55 104
bearing The ANGLE between the course of a ship and Bayes’s theorem In his 1763 paper, published posthumously, REV. THOMAS BAYES established a fundamental result, now called Bayes’s theorem, that expresses the CONDITIONAL PROBABILITY P(AB) of an event A occurring given that event B has already occurred in terms of the reverse conditional probability P(BA). Precisely: P(A  B) = P(B  A)
P(B) P(A)
This formula is easily proved by noting that P(A ∩ B) P(A ∩ B) P(A  B) = and P(B  A) = . P(B) P(A) More generally, suppose B1, B2,…, Bn is a mutually exclusive and exhaustive set of events, that is, a set of nonoverlapping events covering the whole SAMPLE SPACE. Suppose also that we have been told that another event A has occurred. Then the probability that event Bi also occurred is given by: P(Bi  A) =
P(A  Bi )P(Bi ) P(A  B1)P(B1) + L + P(A  Bn )P(Bn )
To illustrate: suppose that bag 1 contains five red balls and two white balls, and bag 2 contains seven red balls and four white balls. If a bag is selected at random and a ball chosen from it is found to be red, what is the probability that it came from bag 1?
the direction of north is called the ship’s bearing. The angle is measured in degrees in a clockwise direction from north and is usually expressed as a threedigit number. For example, a ship heading directly east has a bearing of 090 degrees, and one heading southwest has a bearing of 225 degrees. The word “bearing” is also used for the measure of angle from north at which an object is sighted. For example, a crewman on board a ship sighting a lighthouse directly west will say that the lighthouse has bearing 270 degrees.
Bernoulli family No family in the history of mathematics has produced as many noted mathematicians as the Bernoulli family from Basel, Switzerland. The family record begins with two brothers, Jacob Bernoulli and Johann (Jean) Bernoulli, respectively, the fifth and 10th children of Nicolaus Bernoulli (1623–1708). Jacob (December 27, 1654–August 16, 1705) is noted for his work on CALCULUS and PROBABILITY theory, being one of the first mathematicians to properly understand the utility and power of the newly published work of the great WILHELM GOTTFRIED LEIBNIZ (1646–1716). Jacob applied the calculus to the study of curves, in particular to the logarithmic spiral and the BRACHISTOCHRONE, and was the first to use POLAR COORDINATES in 1691. He also wrote the first text concentrating on probability theory Ars conjectandi (The art of conjecture), which was published
40 Bernoulli numbers posthumously in 1713. The BERNOULLI NUMBERS also appear, for the first time, in this text. Jacob occupied the chair of mathematics at Basel University from 1687 until his death. Johann (July 17, 1667–January 1, 1748) is also known for his work on calculus. Being recognized as an expert in the field, Johann was hired by the French nobleman MARQUIS DE GUILLAUME FRANÇOIS ANTOINE L’HÔPITAL (1661–1704) to explain the new theory to him, first through formal tutoring sessions in Paris, and then through correspondence when Johann later returned to Basel. L’Hôpital published the contents of the letters in a 1696 textbook Analyse des infiniment petits (Analysis with infinitely small quantities), but gave little acknowledgment to Johann. The famous rule that now bears his name, L’HÔPITAL’S RULE, is due to Johann. Johann succeeded his brother in the chair at Basel University. The two brothers, Jacob and Johann, worked on similar problems, and each maintained an almost constant exchange of ideas with Leibniz. The relationship between the two siblings, however, was not amicable, and they often publicly criticized each other’s work. Nicolaus (I) Bernoulli (October 21, 1687–November 29, 1759), nephew to Jacob and Johann, also achieved some fame in mathematics. He worked on problems in GEOMETRY, DIFFERENTIAL EQUATIONs, infinite SERIES, and probability. He held the chair of mathematics at Padua University, once filled by GALILEO GALILEI (1564–1642). Johann Bernoulli had three sons, all of whom themselves became prominent mathematicians: Nicolaus (II) Bernoulli (February 6, 1695–July 31, 1726) wrote on curves, differential equations, and probability theory. He died—by drowning while swimming—only eight months after accepting a prestigious appointment at the St. Petersberg Academy. Daniel Bernoulli (February 8, 1700–March 17, 1782), the most famous of the three sons, is noted for his 1738 text Hydrodynamica (Hydrodynamics), which laid the foundations for the modern discipline of hydrodynamics. (Daniel’s father, Johann, jealous of his son’s success, published his own text on hydrodynamics in 1739 but placed on it the publishing date of 1732 and accused his son of plagiarism.) Daniel also worked on the mathematics of vibrating strings, the kinetic theory of gases, probability theory, and
partial differential equations. He was awarded the Grand Prize from the Paris Academy no fewer than 10 times. Johann (II) Bernoulli (May 28, 1710–July 17, 1790) studied the mathematics of heat flow and light. He was awarded the Grand Prize from the Paris Academy four times and succeeded his father in the chair at Basel University in 1743. Johann (II) Bernoulli had three sons, two of whom, Johann (III) Bernoulli (November 4, 1744–July 13, 1807) and Jacob (II) Bernoulli (October 17, 1759– August 15, 1789), worked in mathematics. Johann (III) studied astronomy and probability, and wrote on recurring decimals and the theory of equations. He was a professor of mathematics at Berlin University at the young age of 19. Jacob (II) Bernoulli wrote works on the mathematics of elasticity, hydrostatics, and ballistics. He was professor of mathematics at the St. Petersburg Academy, but, like his uncle, drowned at the age of 29 while swimming in the Neva River. Members of the Bernoulli family had a profound effect on the early development of probability theory, calculus, and the field of continuum mechanics. Many concepts (such as the Bernoulli numbers, a probability distribution, a particular differential equation) are named in their honor.
Bernoulli numbers See SUMS OF POWERS.
Bertrand’s paradox French mathematician JosephLouis François Bertrand (1822–1900) posed the following challenge:
Answering Bertrand’s paradox
Bh–askara II 41 Imagine an equilateral TRIANGLE drawn inside a CIRCLE. Find the PROBABILITY that a CHORD chosen at random is longer than the sidelength of the triangle. There are two possible answers: 1. Once a chord is drawn we can always rotate the picture of the circle so that one end of the selected chord is placed at the top of the circle. It is clear then that the length of the chord will be greater than the sidelength of the triangle if the other endpoint lies in the middle third of the perimeter of the circle. The chances of this happening are 1/3, providing the answer to the problem. 2. Rotating the picture of the circle and the selected chord, we can also assume that the chord chosen is horizontal. If the chord crosses the solid line shown, then it will be longer than the sidelength of the triangle. One observes that this solid line is half the length of the diameter. Thus the chances of a chord being longer than the sidelength of the triangle are 1/2. Surprisingly, both lines of reasoning are mathematically correct. Therein lies a PARADOX: the answer cannot simultaneously be 1/3 and 1/2. The problem here lies in defining what we mean by “select a chord at random.” There are many different ways to do this: one could spin a bottle in the center of the circle to select points on the perimeter to connect with a chord, or one could roll a broom across a circle drawn on the floor, or perhaps even drop a wire from a height above the circle and see where it lands. Each approach to “randomness” could (and in fact does) lead to its own separate answer. This paradox shows that extreme care must be taken to pose meaningful problems in probability theory. It is very difficult to give a precise definition to “randomness.”
Bh–askara II (Bhaskaracharya) (1114–1185) Indian
Algebra, Arithmetic Born in Vijayapura, India, Bh–askara (often referred to as Bh–askara II to distinguish him from the seventhcentury mathematician of the same name) is considered India’s most eminent mathematician of the 12th century. He revised and continued the studies of the great BRAHMAGUPTA, making corrections
and filling in gaps in his work, and reached a level of mastery of ARITHMETIC and ALGEBRA that was not matched by a European scholar for several centuries to come. Bh–askara wrote two influential mathematical treatises: Lilavati (The beautiful), on the topic of arithmetic, and the Bijaganita (Seed arithmetic) on algebra. Bh–askara was head of the astronomical observatory in Ujjain, the nation’s most prominent mathematical research center of the time. Although much of Indian mathematics was motivated by problems and challenges in astronomy, Bh–askara’s writings show a keen interest in developing mathematics for its own sake. For example, the text Lilavati, consisting of 13 chapters, begins with careful discussions on arithmetic and geometry before moving on to the topics of SEQUENCEs and SERIES, fractions, INTEREST, plane and solid geometry, sundials, PERMUTATIONs and COMBINATIONs, and DIOPHANTINE EQUATIONs (as they are called today). For instance, Bh–askara shows that the equation 195x = 221y + 65 (which he expressed solely in words) has infinitely many positive integer solutions, beginning with x = 6, y = 5, and x = 23, y = 20, and then x = 40, and y = 35. (The xvalues increase in steps of 17, and the yvalues in steps of 15.) In his piece Bijaganita, Bh–askara develops the arithmetic of NEGATIVE NUMBERS, solves quadratic equations of one, or possibly more, unknowns, and develops methods of extracting SQUARE and CUBE ROOTs of quantities. He continues the discussions of Brahmagupta on the nature and properties of the number ZERO and the use of negative numbers in arithmetic. (He denoted the negative of a number by placing a dot above the numeral.) Bh–askara correctly points out that a quantity divided by zero does not produce zero (as Brahmagupta claimed) and suggested instead that a/0 should be deemed infinite in value. Bh–askara solves complicated equations with several unknowns and develops formulae that led him to the brink of discovering the famous QUADRATIC formula. Bh–askara also wrote a number of important texts in mathematical astronomy and made significant strides in the development of TRIGONOMETRY, taking the subject beyond the level of just a tool of calculation for astronomers. Bh–askara discovered, for example, the famous addition formulae for sine: sin(A + B) = sin A cos B + cos A sin B sin(A – B) = sin A cos B – cos A sin B
42 bias However, as was the tradition at the time, Bh–askara did not explain how he derived his results. It is conjectured that Indian astronomers and mathematicians felt it necessary to conceal their methods regarding proofs and derivations as “trade secrets” of the art. Bh–askara’s accomplishments were revered for many centuries. In 1817, H. J. Colebrook provided English translations of both Lilavati and Bijaganita in his text Algebra with Arithmetic and Mensuration.
bias A systematic error in a statistical study is called a bias. If the sample in the study is large, errors produced by chance tend to cancel each other out, but those from a bias do not. For example, a survey on the shopping habits of the general population conducted at a shopping mall is likely to be biased toward people who shop primarily at malls, omitting results from people who shop from home through catalogs and online services. This is similar to a loaded die, which is biased to produce a particular outcome with greater than onesixth probability. Surprisingly, American pennies are biased. If you delicately balance 30 pennies on edge and bump the surface on which they stand, most will fall over heads up. If, on the other hand, you spin 30 pennies and let them all naturally come to rest, then most will land tails. See also POPULATION AND SAMPLE.
biconditional In
FORMAL LOGIC,
a statement of the form “p if, and only if, q” is called a biconditional statement. For example, “A triangle is equilateral if, and only if, it is equiangular” is a biconditional statement. A biconditional statement is often abbreviated as p iff q and is written in symbols as p ↔ q. It is equivalent to the compound statement “p implies q, and q implies p” composed of two CONDITIONAL statements. The truthvalues of p and q must match for the biconditional statement as a whole to be true. It therefore has the following TRUTH TABLE:
p
q
p↔q
T T F F
T F T F
T F F T
The two statements p and q are said to be logically equivalent if the biconditional statement p ↔ q is true. See also ARGUMENT.
bijection See FUNCTION. bimodal See STATISTICS: DESCRIPTIVE. binary numbers (base2 numbers) Any whole number can be written as a sum of distinct numbers from the list of powers of 2: 1, 2, 4, 8, 16, 32, 64, … (Simply subtract the largest power of 2 less from the given number and repeat the process for the remainder obtained.) For instance, we have: 89 = 64 + 25 = 64 + 16 + 9 = 64 + 16 + 8 + 1 No power of 2 will appear twice, as two copies of the same power of 2 sum to the next power in the list. Moreover, the sum of powers of 2 produced for a given number is unique. Using the symbol 1 to denote that a particular power of 2 is used and 0 to denote that it is not, one can then encode any given number as a sequence of 0s and 1s according to the powers of 2 that appear in its presentation. For instance, for the number 89, the number 64 is used, but 32 is not. The number 16 appears, as does 8, but not 4 or 2. Finally, the number 1 is also used. We write: 89 = 10110012 (It is customary to work with the large power of 2 to the left.) As other examples, we see that the code 100010112 corresponds to the number 128 + 64 + 32 + 16 + 8 + – 4+ 2 + 1 = 139, and the code 101112 to the number 16 + 8 + 4 + 2 + 1 = 23. Numbers represented according to this method are called binary numbers. These representations correspond precisely to the representations made by choosing 2 as the BASE OF A NUMBER SYSTEM. If one introduces a decimal point into the system and interprets positions to the right of the point as negative powers of 2, then fractional quantities can also be represented in binary notation. For instance, 0.1012 represents the quantity 2–1 + 2–3 = 1 + 1 = 5 , and 2 8 8 1 1 1 0.010101…2 the quantity + + + …, which, 4 16 64
binomial distribution 43 according to the GEOMETRIC SERIES formula, is the fraction 1 . 3 Any DYADIC fraction has a finite binary decimal expansion. As the study of those fractions shows, the binary code of a dyadic can be cleverly interpreted as instructions for folding a strip of paper to produce a crease mark at the location of that dyadic fraction. The process of RUSSIAN MULTIPLICATION also uses binary numbers in an ingenious manner. Binary numbers are used in computers because the two digits 0 and 1 can be represented by two alternative states of a component (for example, “on” or “off,” or the presence or absence of a magnetized region). It is appropriate to mention that the powers of 2 solve the famous “five stone problem”: A woman possesses five stones and a simple twoarm balance. She claims that, with a combination of her stones, she can match the weight of any rock you hand her and thereby determine its weight. She does this under the proviso that your rock weighs an integral number of pounds and no more than 31 pounds. What are the weights of her five stones? As every number from 1 through 31 can be represented as a sum of the numbers 1, 2, 4, 8, and 16, the woman has stones of weights corresponding to these first five powers of 2.
binary operation A rule that assigns to each pair of elements of a set another element of that same set is called a binary operation. For example, the addition of two numbers is a binary operation on the set of real numbers, as is the product of the two numbers and the sum of the two numbers squared. “Union” is a binary operation on sets, as is “intersection,” and CROSS PRODUCT is a binary operation on the set of vectors in threespace. However, the operation of DOT PRODUCT is not a binary operation on the set of vectors; the results of this operation are numbers, not other vectors. If the set under consideration is denoted S, then a binary operation on S can be thought of as a FUNCTION f from the set of pairs of elements of the set, denoted S × S, to the set S: f : S × S → S. See also OPERATION; UNARY OPERATION.
binomial Any algebraic expression consisting of two terms, such as 2x + y or a + 1, is called a binomial. See also MONOMIAL; POLYNOMIAL; TRINOMIAL.
binomial coefficient See BINOMIAL THEOREM.
binomial distribution The distribution that arises when considering the question What is the probability of obtaining precisely k successes in n runs of an experiment? is called the binomial distribution. Here we assume the experiment has only two possible outcomes—“success or failure,” or “heads or tails,” for example—and that the probability of either occurring does not change as the experiment is repeated. The binomial distribution itself is a table of values providing the answers to this question for various values of k, from k = 0 (no successes) to k = n (all successes). To illustrate: the chance of tossing a “head” on a fair coin is 50 percent. Suppose we choose to toss the coin 10 times. Observe that the probability of attaining any specific sequence of outcomes (three heads, followed by two tails, then one head and four tails, for 1 1 total of 10 times 1 ⎛ 1 ⎞ … × =⎜ ⎟ example) is 2 × 2 × 2 ⎝ 2⎠
10
=
1 1024 . In
particular, the probability of seeing no heads (all tails) is also 1/1024, as is the chance of seeing 10 heads in a row. There are 10 places for a single head to appear among 10 tosses, thus the chances of seeing precisely ⎛ 1⎞ one head out of 10 tosses is 10 × ⎜⎝ 2 ⎟⎠
10
1 percent. According to the theory of
=
10 1024 , about
COMBINATIONs,
⎛10⎞ 10! there are ⎜⎝ 2 ⎟⎠ = 2! 8! = 45 ways for two heads to appear among 10 places, and so the probability of seeing precisely two heads among the 10 tosses is ⎛10⎞ ⎛ 1 ⎞ ⎜ ⎟ × ⎜⎝ 2 ⎟⎠ ⎝ 2⎠
10
=
45 1024 , about 4.4 percent.
Continuing this way, we obtain the binomial distribution for tossing a fair coin 10 times:
44 binomial theorem
Number of Heads
Probability 10 10 1 1 ≈ 0.1% × = 1024 0 2
0
10
1
10 1
2
10 1 45 ≈ 4.4% × = 1024 2 2
1 × 2
=
10 ≈ 1.0% 1024
10
10
3
10 3
4
10 1 210 ≈ 20.5% × = 1024 4 2
1 × 2
=
120 ≈ 11.7% 1024
n
n
n
n
(p + q)n = 0 p0qn + 1 p1qn−1+ 2 p2qn−2 + L + n pnq0 This explains the name of the distribution. The binomial distribution has mean value (EXPECTED n n µ = k pkqn− k , which equals np. (To see this, ∑ VALUE) k k= 0
n
differentiate the formula
n
(p + q)n = ∑ k pkqn−k k= 0
with
respect to p.) The standard deviation is σ = √npq . (See STATISTICS: DESCRIPTIVE.)
10
10
5
10 5
6
10 1 210 ≈ 20.5% × = 1024 6 2
7
10 7
8
10 1 45 ≈ 4.4% × = 1024 8 2
9
10 9
1 × 2
=
252 ≈ 24.6% 1024
10
1 × 2
10 =
120 ≈ 11.7% 1024
10
1 × 2
10 =
10 ≈ 1.0% 1024
10
10 1 1 ≈ 0.1% × = 1024 10 2
10
If the coin is biased—say the chances of tossing a head are now only 1/3—then a different binomial distribution would be obtained. For example, the probability of attaining precisely eight heads among 10 tosses is 8
2
10 1 2 now only 8 3 3 ≈ 0.003 , about 0.3 percent. In general, if p denotes the probability of success, and q = 1 – p is the probability of failure, then the binomial distribution is given by the formula n k n− k p q , the probability of attaining precisely k k successes in n runs of the experiment. This quantity is the kth term of the binomial expansion formula from the BINOMIAL THEOREM:
The Poisson Distribution It is difficult to calculate the binomial distribution if n is very large. Mathematicians have shown that the binomial distribution can be well approximated by the NORMAL DISTRIBUTION for large values of n, provided the value p is neither extremely small nor close to one. For these troublesome values of p, SIMÉONDENIS POISSON µk –µ e —with µ = np, for showed in 1837 that the values k! k = 0, 1, 2, …—provide a sequence of values close to the values one would expect from the binomial distribution. The distribution provided by these approximate values is called a Poisson distribution. The Geometric Distribution Alternatively one can ask: what is the probability that the first success in a series of experiments occurs on the nth trial? If p is the probability of success and q = 1 – p is the probability of failure, then one obtains a first success on the nth experiment by first obtaining n – 1 failures and then a success. The probability of this occurring is: P(n) = pq n–1. The distribution given by this sequence of probability values (for n = 1, 2, 3, …) is called the geometric distribution. It has mean µ = 1/p – and standard deviation σ = √ q/p. The geometric distribution is a special example of Pascal’s distribution, which seeks the probability Pk(n) of the kth success occurring on the nth trial. See also HISTORY OF PROBABILITY AND STATISTICS (essay).
binomial theorem (binomial expansion) The identities (x + a)2 = x2 + 2xa + a2 and (x + a)3 = x3 + 3x2a +
bisection method 45 3xa2 + a3 are used in elementary ALGEBRA. These are both special cases of the general binomial theorem that asserts, for any positive integer n, we have: (x + a)n =
n
⎛ n⎞
⎛ n⎞
⎛ n⎞
∑ ⎜⎝ k⎟⎠ x n−kak = x n + ⎜⎝ 1⎟⎠ x n−1a + ⎜⎝ 2⎟⎠ x n−2 a2
k= 0
⎛ n ⎞ n−1 n +L+ ⎜ ⎟ xa + a ⎝ n − 1⎠ ⎛ n⎞ n! Here each number ⎜ ⎟ = is a COMBINATORIAL ⎝ k⎠ k!(n − k)! COEFFICIENT, also called a binomial coefficient. The binomial theorem is proved by examining the process of EXPANDING BRACKETS, thinking of the quantity (x + a)n as a product of n factors: (x + a)(x + a)… (x + a). To expand the brackets, one must select an entry from each set of parentheses (“x” or “a”), multiply together all the selected elements, and add together all possible results. For example, there is one way to obtain the term xn: select x from every set of parentheses. There are n ways to create a term of the form x n–1a: select a from just one set of parentheses, and x from the ⎛ n⎞ remaining sets. In general there are ⎜ ⎟ ways to select k ⎝ k⎠ as and n – k xs. Thus, in the expansion, there will be ⎛ n⎞ ⎜ ⎟ terms of the form x n–ka k. ⎝ k⎠ The combinatorial coefficients are the entries of PASCAL’S TRIANGLE. The binomial theorem applied to (1 + 1)n explains why the elements of each row of Pascal’s triangle sum to a power of two: ⎛ n⎞ ⎛ n⎞ ⎛ n⎞ 2n = (1 + 1)n = ⎜ ⎟ + ⎜ ⎟ + L + ⎜ ⎟ ⎝ 0⎠ ⎝ 1⎠ ⎝ n⎠ Applying the theorem to (1 – 1)n explains why the alternating sum of the entries is zero:
112 = (10 + 1)2 = 100 + 2 × 10 + 1 113 = (10 + 1)3 = 1,000 + 3 × 100 + 3 × 10 + 1 114 = (10 + 1)4 = 10,000 + 4 × 1,000 + 6 × 100 + 4 × 10 + 1 (The correspondence would remain valid if we did not carry digits when computing higher powers of 11.) The binomial theorem can be used to approximate high powers of decimals. For example, to estimate (2.01)10 we observe: 2.0110 = (2 + 0.01)10 = 210 + 10 × 29 × 0.01 + 28 + 45 × 28 × 0.012 + … ≈ 1024 + 10 × 512 × 0.01 + 45 × 256 × 0.00001 = 1024 + 51.2 + 1.152 ≈ 1076 In 1665 SIR ISAAC NEWTON, coinventor of CALCULUS, discovered that it is possible to expand quantities of the form (x + a)r where r is not equal to a whole number. This leads to the generalized binomial theorem: If r is an arbitrary real number, and x < a, then: (x + a)r = x r + rx r −1a + +
r(r − 1) r − 2 2 x a 2!
r(r − 1)(r − 2) r −3 3 x a +L 3!
The formula is established by computing the TAYLOR SERIES of f(x) = (x + a)r at x = 0. In 1826 Norwegian mathematician NIELS ABEL proved that the series converges for the range indicated. Notice that if r is a positive integer, then the theorem reduces to the ordinary binomial theorem. (In particular, from the n + 1’th place onward, all terms in the infinite sum are zero.) ⎛ n⎞ The combinatorial coefficients ⎜ ⎟ arising in the ⎝ k⎠
⎛ n⎞ ⎛ n⎞ ⎛ n⎞ ⎛ n⎞ 0 = (1 − 1)n = ⎜ ⎟ − ⎜ ⎟ + ⎜ ⎟ − L ± ⎜ ⎟ ⎝ 0⎠ ⎝ 1⎠ ⎝ 2⎠ ⎝ n⎠
binomial theorem are sometimes called binomial coefficients. The generalized combinatorial coefficients appear in expansions of quantities of the type (x + y + z)n and (x + y + z + w)n, for example. See also COMBINATION.
Applying the theorem to (10 + 1)n explains why the first few rows of Pascal’s triangle resemble the powers of 11:
bisection method (dichotomous line search, binary line search) Often one is required to find a solution to an
46 bisector equation of the form f(x) = 0 even if there are no clear algebraic means for doing so. (For instance, there are no general techniques helpful for solving √x + √x + 2 + √x + 3 – 5 = 0. The bisection method provides the means to find, at least, approximate solutions to such equations. The method is based on the fact that if two function values f(a) and f(b) of a CONTINUOUS FUNCTION have opposite signs, then, according to the INTERMEDIATE VALUE THEOREM, a ROOT of the equation f(x) = 0 lies between a and b. The method proceeds as follows: 1. Find two values a and b (a < b) such that f(a) and f(b) have opposite signs. 2. Set m = (a + b)/2, the midpoint of the interval, and compute f(m). 3. If f(m) = 0, we have found a zero. Otherwise, if f(a) and f(m) have opposite signs, then the zero of f lies between a and m; repeat steps 1 and 2 using these new values. If, on the other hand, f(m) and f(b) have opposite signs, then the zero of f lies between m and b; repeat steps 1 and 2 using these new values. In either case, a new interval containing the zero has been constructed that is half the length of the original interval. 4. Repeated application of this procedure homes in on a zero for the function. To solve the equation f(x) = √x + √x + 2 + √x + 3 – 5 = 0, for example, notice that f(1) = –0.268 < 0 and f(2) = 0.650 > 0. A zero for the function thus lies between 1 and 2. Set m = 1.5. Since f(1.5) = 0.217 > 0 we deduce that, in fact, the zero lies between 1 and 1.5. Now set m = 1.25 to see that the zero lies between 1.25 and 1.5. One can find the location of a zero to any desired degree of accuracy using this method. For example, repeating this procedure for the example above six more times shows that the location of the zero lies in the interval [1.269,1.273]. This shows that to three significant figures the value of the zero is 1.27. The bisection method will fail to locate a root if the graph of the function touches the xaxis at that location without crossing it. Alternative methods, such as NEWTON’S METHOD, can be employed to locate such roots.
bisector Any line, plane, or curve that divides an angle, a line segment, or a geometric object into two
equal parts is called a bisector. For example, the equator is a curve that bisects the surface of the EARTH. A straight line that divides an angle in half is called an angle bisector, and any line through the MIDPOINT of a line segment is a segment bisector. If a segment bisector makes a right angle to the segment, then it is called a perpendicular bisector.
Bolyai, János (1802–1860) Hungarian Geometry Born on December 15, 1802, in Kolozsvár, Hungary, now Cluj, Romania, János Bolyai is remembered for his 1823 discovery of NONEUCLIDEAN GEOMETRY, an account of which he published in 1832. His work was independent of the work of NIKOLAI IVANOVICH LOBACHEVSKY (1792–1856), who published an account of HYPERBOLIC GEOMETRY in 1829. Bolyai was taught mathematics by his father Farkas Bolyai, himself an accomplished mathematician, and had mastered CALCULUS and mechanics by the time he was 13. At age 16 he entered the Royal Engineering College in Vienna and joined the army engineering corps upon graduation four years later. Like many a scholar throughout the centuries, Farkas Bolyai had worked, unsuccessfully, on the challenge of establishing the PARALLEL POSTULATE as a logical consequence of the remaining four of EUCLID’S POSTULATES. He advised his son to avoid working on this problem. Fortunately, János Bolyai did not take heed and took to serious work on the issue while serving as an army officer. During the years 1820 and 1823 Bolyai prepared a lengthy treatise outlining the details of a new and consistent theory of geometry for which the parallel postulate does not hold, thereby settling once and for all the problem that had troubled scholars since the time of EUCLID: the parallel postulate cannot be proved a consequence of the remaining postulates of Euclid. In Bolyai’s system of hyperbolic geometry it is always the case that, for any point P in the plane, there are an infinite number of distinct lines through that point all PARALLEL to any given line not through P. (In ordinary geometry, where the parallel postulate holds, there is only one, and only one, line through a given point P parallel to a given direction. This is PLAYFAIR’S AXIOM.) In his new geometry, angles in triangles sum to less than 180°, and the ratio of the circumference of a circle to its diameter is greater than π.
Bombelli, Rafael 47 Just before publishing his work, Bolyai learned that the great CARL FRIEDRICH GAUSS (1777–1855) had already anticipated much of this theory, even though he had not published any material on the matter. Bolyai decided to delay the release of his work. In 1832 he printed the details of his new theory only as a 24page appendix to an essay his father was preparing. Later, in 1848, Bolyai discovered that Lobachevsky had published a similar piece of work in 1829. Bolyai never published the full version of his original treatise. His short 24page piece was practically forgotten until Richard Blatzer discussed the work of both Bolyai and Lobachevsky in his 1867 text Elemente der Mathematik (Elements of mathematics). At that point, Bolyai’s piece was recognized as the first clear account of the mathematics of a new type of geometry. He is today regarded as having independently founded the topic. Bolyai died on January 27, 1860, in Marosvásárhely, Hungary (now TirguMures, Romania). In 1945 the University of Cluj honored Bolyai by including his name in its title. It is today known as the BabesBolyai University of Cluj.
Bolzano, Bernard Placidus (1781–1848) Czech Analysis, Philosophy, Theology Born on October 5, 1781, in Prague, Bernard Bolzano is remembered as the first mathematician to offer a rigorous description of what is meant by a CONTINUOUS FUNCTION. The related theorem, the INTERMEDIATEVALUE THEOREM, is sometimes named in his honor. Bolzano studied philosophy and mathematics at the University of Prague and earned a doctoral degree in mathematics in 1804. He also completed three years of theological study at the same time and was ordained a Roman Catholic priest two days after receiving his doctorate. Choosing to pursue a career in teaching, Bolzano accepted a position as chair of philosophy and religion at the university later that year. In 1810 Bolzano began work on understanding the foundations of mathematics and, in particular, the logical foundations of the newly discovered CALCULUS. He found the notion of an INFINITESIMAL troublesome and attempted to provide a new basis for the subject free from this concept. In his 1817 paper “Rein Analytischer Beweis” (Pure analytical proof), Bolzano explored the concept of a LIMIT—anticipating the foundational approach of AUGUSTINLOUIS CAUCHY offered four
years later—and proved the famous intermediate value theorem. Bolzano was also the first to provide an example of a function that is continuous at every point but differentiable at no point. Bolzano also anticipated much of GEORG CANTOR’S work on the infinite. In his 1850 article “Paradoxien des Unendlichen” (Paradoxes of the infinite), published by a student two years after his death, Bolzano examined the nature of infinite sets and the paradoxes that arise from them. This piece contains the first use of the word “set” in a mathematical context. Bolzano died on December 18, 1848, in Prague, Bohemia (now the Czech Republic). His work paved the way for providing rigorous underpinnings to the subject of calculus. In particular, Bolzano identified for the first time in “Rein Analytischer Beweis” the “completeness property” of the real numbers.
Bombelli, Rafael (1526–1572) Italian Algebra Born in Bologna, Italy, in 1526 (the day and month of his birth date are not known), scholar Rafael Bombelli is remembered for his highly influential 1572 book L’Algebra (Algebra). In this work, Bombelli published rules for the solution to the QUADRATIC, CUBIC, and QUARTIC EQUATIONS, and was one of the first mathematicians to accept COMPLEX NUMBERS as solutions to equations. Bombelli began his career as an engineer specializing in hydraulics and worked on a number of projects to turn salt marshes into usable land. Having read the great Ars magna (The great art) by GIROLAMO CARDANO (1501–76), Bombelli decided to write an algebra text that would make the methods developed there accessible to a general audience and be of interest and use to surveyors and engineers. He intended to write a fivevolume piece but only managed to publish three volumes before his death in 1572. Bombelli noted that Cardano’s method of solving cubic equations often leads to solutions that, at first glance, appear unenlightening. For instance, examination of the equation x3 = 15x + 4 leads to the solution: x = 3 2 + −121 + 3 2 − −121 Although scholars at the time rejected such quantities (because of the appearance of the square root of a negative quantity) Bombelli argued that such results
48 Boole, George should not be ignored and, moreover, that they do lead to real solutions. After developing the algebra of complex numbers, Bombelli could show that the answer presented above, for instance, is just the number x = 4 in disguise. Bombelli went on to show that, in fact, every equation of the form x3 = ax + b, with a and b positive, has a real solution, thereby justifying the method of complex numbers. In addition to developing complex arithmetic, Bombelli developed the basic algebra of NEGATIVE NUMBERS. In particular, he provided a geometric argument to help explain why a negative number times itself must be positive—a notion that still causes many people difficulty today. Bombelli died in 1572. (The exact date of his death is not known.) GOTTFRIED WILHELM LEIBNIZ (1646–1716), codiscoverer of CALCULUS, taught himself mathematics from Bombelli’s L’Algebra and described the scholar as “an outstanding master of the analytic art.”
Boole, George (1815–1864) British Logic Born on November 2, 1815, in Lincolnshire, England, algebraist George Boole is remembered for his highly innovative work in the field of logic. In his pioneering piece, An Investigation of the Laws of Thought, on Which are Founded the Mathematical Theories of Logic and Probability, published in 1854, Boole established the effectiveness of symbolic manipulation as a means to represent and perform operations of reasoning. Boole is considered the founder of the field of symbolic logic. Boole received no formal education in mathematics. As a young man he read the works of JOSEPHLOUIS LAGRANGE and PIERRESIMON LAPLACE, and by age 20 began publishing original results. His early work in the field of DIFFERENTIAL EQUATIONs garnered him national attention as a capable scholar. In 1845 Boole was honored with a gold medal from the ROYAL SOCIETY of London, England’s most prestigious academic society. Four years later, in 1849, he was appointed chair of the mathematics department at Queens College, Cork, Ireland, despite having no university degree. Boole stayed at this college for the rest of his life, devoting himself to teaching and research. Boole began work in mathematical logic before moving to Ireland. At the time, logic was considered to be a topic of interest only to philosophers, but in 1847, in his pamphlet The Mathematical Analysis of
George Boole, an eminent mathematician of the 19th century, established the field of mathematical logic. (Photo courtesy of Topham/The Image Works)
Logic, Boole successfully argued that the topic has merit in the art of mathematical reasoning. By using symbols to represent statements, Boole developed an “algebra of logic” whose rules and valid manipulations matched the processes of reasoning. Thus mathematical arguments and lines of thought could themselves be reduced to simple algebraic manipulations. For instance, if the symbol x is used to represent “all butterflies,” then 1 – x represents all that is not a butterfly. If y represents the color blue, then xy is the set of all objects that are both butterflies and blue, that is, all blue butterflies. The expression (1 – x)(1 – y) represents all the nonblue nonbutterflies. Algebraically the quantity (1 – x)(1 – y) equals 1 – x – y + xy. One could argue that the INCLUSIONEXCLUSION PRINCIPLE is at play here, saying that the set of nonblue nonbutterflies is the set of all objects that are neither butterflies, nor blue, with an adjustment
bound 49 made to account for the fact that the set of blue butterflies has been excluded twice. These adjustments, however, are awkward. Boole invented a new system of “algebra” that avoids such modifications. The axioms it obeys differ from those of ordinary arithmetic. The algebra Boole invented proved to be of fundamental importance. It gave 20thcentury engineers the means to instruct machines to follow commands and has since been used extensively in all computer design and electrical network theory. In a real sense, Boole was the world’s first computer scientist, despite the fact that computers were not invented for another century to come. Boole died unexpectedly in 1864 at the age of 49 from pneumonia. (The exact date of his death is not known.)
Boolean algebra In the mid1800s GEORGE BOOLE developed a system of algebraic manipulations suitable for the study of FORMAL LOGIC and SET THEORY, now called Boolean algebra. He assumed that one is given a set of elements, which we will denote x, y, z, …, on which one can perform two operations, today called Boolean sum, x + y, and Boolean product, x · y. These operations must satisfy the following rules:
Other rules for Boolean algebra follow from the four axioms presented above. For example, one can show that two ASSOCIATIVE laws hold: x + (y + z) = (x + y) + z and x · (y · z) = (x · y) · z. See also DE MORGAN’S LAWS.
Borromean rings The term refers to a set of three rings linked together as a set, but with the property that if any single ring is cut, all three rings separate. The design of three such rings appeared on the coat of arms of the noble Italian family, BorromeoArese. (Cardinal Carlo Borromeo was canonized in 1610, and Cardinal Federico Borromeo founded the Ambrosian Art Gallery in Milan, Italy.) The curious property of the design attracted the attention of mathematicians. It is an amusing exercise to arrange four rings such that, as a set, they are inextricably linked together, yet cutting any single ring would set all four free. Surprisingly this feat can be accomplished with any number of rings. See also KNOT THEORY.
bound A function is bounded if it takes values no 1. The operations are COMMUTATIVE, that is, for all elements x and y we have x + y = y + x and x · y = y · x. 2. There exist two special elements, denoted “0” and “1,” which, for all elements x, satisfy x + 0 = x and x · 1 = x. 3. For each element x there is an inverse element “–x” which satisfies x + (–x) = 1 and x · (–x) = 0. 4. The following DISTRIBUTIVE laws hold for all elements in the set: x · (y + z) = (x · y) + (x · z) and x + (y · z) = (x + y) · (x + z). One can see that the Boolean operations “ + ” and “ · ” are very different from the addition and multiplication of ordinary arithmetic and so cannot be interpreted as such. However, thinking of Boolean addition as the “union of two sets” and Boolean product as “the intersection of two sets,” with 0 being the empty set and 1 the universal set, we see that the all four axioms hold, making SET THEORY a Boolean algebra. Similarly, the FORMAL LOGIC of propositional calculus is a Boolean algebra if one interprets addition as the DISJUNCTION of two statements (“or”) and product as their CONJUNCTION (“and”).
higher than some number M and no lower than some second value L. For example, the function f(x) = sin x is bounded between the values –1 and 1. We call M an upper bound for the function and L a lower bound.
Borromean rings
50 Bourbaki, Nicolas A function is bounded above if the function possesses an upper bound (but not necessarily a lower bound), and bounded below if it possesses a lower bound (but not necessarily an upper bound). For example, f(x) = x2 is bounded below by the value L = 0, since all output values for this function are greater than or equal to zero. The EXTREMEVALUE THEOREM ensures that every CONTINUOUS FUNCTION defined on a closed INTERVAL is bounded. A set of numbers is bounded above if every number in the set is less than or equal to some value M, bounded below if every number in the set is greater than or equal to some value L, and bounded if it is both bounded above and bounded below. For example, the set S = {0.6, 0.66, 0.666, …} is bounded below by 0.6 and bounded above by 1. The smallest possible upper bound for a set is called the least upper bound, and the largest lower bound, the greatest lower bound. The set S has 2/3 as its least upper bound and 0.6 as its greatest lower bound. The REAL NUMBERS have the property that any subset S of them that is bounded above possesses a least upper bound, and, similarly, any subset that is bounded below possesses a greatest lowest bound. This is a key property that shows that no numbers are “missing” from the real number line. (See DEDEKIND CUT.) This is not true of the set of rational numbers, for instance. The set of all rationals whose square is less than 2, for example, is bounded above, by 3/2 for example, but possesses no least upper bound in the set of rationals: the square root of two is “missing” from the set of rationals. A sequence is bounded if, as a set of numbers, it is bounded. A geometric figure in the plane is bounded if it can be enclosed in a rectangle of finite area. For example, a CIRCLE is bounded but a HALFPLANE is not.
Bourbaki, Nicolas Taking the name of a junior Napoleonic officer, a group of French mathematicians of the 1930s adopted the pseudonym of Nicolas Bourbaki to publish a series of books, all under the title Éléments de mathématiques (Elements of mathematics), that attempt to present a complete, definitive, and utterly rigorous account of all modern mathematical knowledge. This project continues today. Contributors to the work remain anonymous and change over the years. To date, over 40 volumes of work have been produced.
The material presented through Bourbaki is austere and abstract. The goal of the founding work was to develop all of mathematics on the axioms of SET THEORY and to maintain the axiomatic approach as new concepts are introduced. The work, devoid of narrative and motivational context, is difficult to read and not suitable for use as textbooks. During the 1950s and 1960s, however, there were often no graduatelevel texts in the developing new fields, and the volumes of Bourbaki were the only sources of reference. It is unlikely that today a graduate student in mathematics would consult the work of Bourbaki.
brachistochrone See CYCLOID.
brackets Any pair of symbols, such as parentheses ( ) or braces { }, that are used in an arithmetic or an algebraic expression to indicate that the quantity between them is to be evaluated first, or treated as a single unit in the evaluation of the whole, are called brackets. For example, in the expression (2 + 3) × 4, the parentheses indicate that we are required to first calculate 2 + 3 = 5 and then multiply this result by 4. In complicated expressions, more than one type of bracket may be used in the same equation. For instance, the expression 3{2 + 8[2(x + 3) – 5(x – 2)]} is a little easier to read than 3(2 + 8(2(x + 3) – 5(x – 2))). Before the advent of the printing press in the 15th century, the VINCULUM was used to indicate the order of operations. Italian algebraist RAFAEL BOMBELLI (1526–1572) was one of the first scholars to use parentheses in a printed algebraic equation, but it was not until the early 1700s, thanks chiefly to the influence of LEONHARD EULER, GOTTFRIED WILHELM LEIBNIZ, and members of the BERNOULLI FAMILY, that their use in mathematics became standard. Angle brackets are typically only used to list the components of a VECTOR or a finite SEQUENCE. Matters are a little confusing, for in the theory of quantum mechanics, angle brackets are used to indicate the DOT PRODUCT of two vectors (and not the vectors themselves). The left angle bracket “” a “ket.” In SET THEORY, braces { } are used to list the elements of a set. Sometimes the elements of a sequence are listed inside a set of braces.
Brahmagupta’s formula 51 If x is a real number, then the bracket symbols x, x, and {x} are used to denote the floor, ceiling, and fractional part values, respectively, of x. Square brackets [ ] and parentheses ( ) are placed at the end points of an INTERVAL on the real number line to indicate whether or not the end points of that interval are to be included. See also EXPANDING BRACKETS; FLOOR/CEILING/ FRACTIONAL PART FUNCTIONS; ORDER OF OPERATION.
Brahmagupta (ca. 598–665) Indian Arithmetic, Geometry, Astronomy Born in Ujjain, India, scholar Brahmagupta is recognized as one of the important mathematicians of the seventh century. His famous 628 text Brahmasphutasiddhanta (The opening of the universe) on the topic of astronomy includes such notable mathematical results as his famous formula for the AREA of a cyclic QUADRILATERAL, the integer solution to certain algebraic equations, and methods of solution to simultaneous equations. This work is also historically significant as the first documented systematic use of ZERO and negative quantities as valid numbers in ARITHMETIC. Brahmagupta was head of the astronomical observatory at Ujjain, the foremost mathematical center of ancient India, and took an avid interest in the development of astronomical observation and calculation. The first 10 of the 25 chapters of Brahmasphutasiddhanta pertain solely to astronomy, discussing the longitude of the planets, lunar and solar eclipses, and the timing of planet alignments. Although rich in mathematical computation and technique, it is the remainder of the work that offers an insight into Brahmagupta’s farreaching understanding of mathematics on an abstract level. Brahmagupta goes on to describe the decimal PLACEVALUE SYSTEM used in India at his time for representing numerals and the methods for doing arithmetic in this system. (For instance, he outlines a method of “long multiplication” essentially equivalent to the approach we use today.) Brahmagupta permits zero as a valid number in all of his computations, and in fact gives it the explicit status of a number by defining it as the result of subtracting a quantity from itself. (Until then, zero acted as nothing more than a placeholder to distinguish 203 from 23, for instance.) He also explains the arithmetical properties of zero—that adding zero to a number leaves that number unchanged and multiplying any number by zero
produces zero, for instance. Brahmagupta also detailed the arithmetic of negative numbers (which he called “debt”) and suggested, for the first time, that they may indeed be valid solutions to certain problems. Brahmagupta next explores problems in ALGEBRA. He develops some basic algebraic notations and then presents a series of methods for solving a variety of linear and quadratic equations. For instance, he devised an ingenious technique for finding integer solutions to equations of the form ax2 + c = y2. (For example, Brahmagupta correctly asserted that x = 226,153,980 and y = 1,766,319,049 are the smallest positive integer solutions to 61x2 + 1 = y2.) Brahmagupta also presents the famous SUMS OF POWERS formulae: 1+ 2 +L+ n =
n(n + 1) 2
12 + 22 + L + n 2 =
n(n + 1)(2n + 1) 6
13 + 23 + L + n3 =
n 2 (n + 1)2 4
as well as algorithms for computing square roots. Unfortunately, as was the practice of writing at the time, Brahmagupta never gave any word of explanation as to how his solutions or formulae were found. No proofs were ever offered. In the final sections of Brahmasphutasiddhanta, Brahmagupta presents his famous formula for the area of a cyclic quadrilateral solely in terms of the lengths of its sides. Curiously, Brahmagupta does not state that the formula is true only for quadrilaterals inscribed in a CIRCLE. In a second work, Khandakhadyaka, written in 665, Brahmagupta discusses further topics in astronomy. Of particular interest to mathematicians, Brahmagupta presents here an ingenious method for computing values of sines. Brahmagupta’s methods and discoveries were extremely influential. Virtually every text that discusses Indian astronomy describes or uses some aspect of his work. See also BRAHMAGUPTA’S FORMULA.
Brahmagupta’s formula Seventhcentury Indian mathematician and astronomer BRAHMAGUPTA derived a formula for the AREA of a QUADRILATERAL inscribed in a
52 braid CIRCLE solely in terms of the lengths of its four sides. His formula reads:
area = √(s – a)(s – b)(s – c)(s – d) where a, b, c, and d are the four sidelengths and a+b+c+d s= is the figure’s semiperimeter. 2 If p and q are the lengths of the figure’s two diagonals then PTOLEMY’S THEOREM asserts that pq = ac + bd. Brahmagupta’s formula follows from BRETSCHNEIDER’S FORMULA for the area of a quadrilateral: area =
1 4(pq)2 − (b2 + d 2 − a2 − c 2 )2 4
by substituting in this value for pq. If one of the sides of the quadrilateral has length zero, that is, the figure is a TRIANGLE, then Brahmagupta’s formula reduces to HERON’S FORMULA. See also CYCLIC POLYGON.
braid A number of strings plaited together is called a braid. The theory of braids examines the number of (essentially distinct) ways a fixed number of strings, held initially in parallel, can be braided. One can combine two braids on a fixed number of strings by repeat
ing the pattern of the second braid at the end of the first braid. If, after completing this maneuver, the act of physically shaking the system of strings settles the strands to the unbraided state, then we say that the two braids are “inverse braids.” For example, the two braids shown in the diagram are inverse braids. If a braid consists of n strings, then the symbol σi is used to record the act of switching of the ith string over the (i + 1)th string (for 1 ≤ i ≤ n – 1) and σi–1 for the act of switching of the same two strings but in the opposite sense. A general braid is then described as a string of these symbols (called a “word”). For instance, the two braids shown in the diagram below, at left, are represented by the words σ1σ2–1σ1 and σ1–1σ2σ1–1, respectively. A braid with no crossings (that is, in which no strings cross) is denoted “1,” and the process of combining braids corresponds precisely to the process of concatenating words. Two braids are inverse braids, if, after performing the suggested symbolic manipulations, their resulting concatenated word is 1. For instance, in our example, we have: σ1σ2–1σ1σ1–1σ2σ1–1 = σ1σ2–1σ2σ1–1 = σ1σ1–1 = 1. It is possible that two different words can represent the same physical braid. (For instance, on three strings, the braids σ1σ2σ1 and σ2σ1σ2 are physically equivalent.) Each set of braids on a fixed number of strings forms a GROUP called a braid group. Austrian mathematician Emil Artin (1898–1962) was the first to study these groups and solve the problem of determining precisely when two different words represent the same braid.
Bretschneider’s
formula German mathematician Carl Anton Bretschneider (1808–78) wrote down a formula for the AREA of a QUADRILATERAL solely in terms of the lengths of its four sides and the value of its four internal angles. If, reading clockwise around the figure, the sidelengths of the quadrilateral are a, b, c, and d, and the angles between the edges are A, B, C, and D (with the angle between edges d and a being A), then Bretschneider established that the area K of the quadrilateral is given by: K = √(s – a)(s – b)(s – c)(s – d) – abcd cos2(θ) a+b+c+d is the semiperimeter of the figure 2 and θ is the average of any two opposite angles in the Here s = A braid and its inverse
Briggs, Henry 53 B+ D A +C or θ = . (Since the interior 2 2 angles of a quadrilateral sum to 360 degrees, each quantity yields the same value for the cosine.)
16K2 = 16(s – a)(s – b)(s – c)(s – d) – 8abcd (cos B cos D – sin B sin D + 1)
figure: θ =
Many texts in mathematics state Bretschneider’s result without proof. Although algebraically detailed, the derivation of the result is relatively straightforward. One begins by noting that area K is the sum of the areas of triangles ABC and ADC. We have: K=
1 1 ab sin B + cd sin D 2 2
Multiplying by four and squaring yields: 16K2 = 4a2b2 sin2B + 4c2d2 sin2D + 8abcd sin B sin D Call this equation (i). Applying the LAW OF COSINES to each of the two triangles yields the relationship: a2 + b2 – 2ab cos B = c2 + d2 – 2cd cos D which can be rewritten:
The following identities from trigonometry: cos B cos D – sin B sin D = cos(B + D) and x cos(x) + 1 = 2 cos2 2 now give: B + D 16K 2 = 16(s − a)(s − b)(s − c)(s − d) − 16abcd cos 2 2 which directly yields the famous result. If, further, the opposite angles of the quadrilateral sum to 180 degrees (in which case the quadrilateral is a CYCLIC POLYGON), then Bretschneider’s formula reduces to BRAHMAGUPTA’S FORMULA for the area of a cyclic quadrilateral:
a2 + b2 – c2 – d2 = 2ab cos B – 2cd cos D
K = √(s – a)(s – b)(s – c)(s – d)
Call this equation (ii). Also note that, with the aid of this second equation:
If one of the sides has length zero, say d = 0, then the quadrilateral is a triangle and we have HERON’S FORMULA for the area of a triangle:
4(s – c)(s – d) = (a + b – c + d)(a + b + c – d) = a2 + b2 – c2 – d2 + 2ab + 2cd = 2ab cos B – 2cd cos D + 2ab + 2cd = 2ab(1 + cos B) + 2cd(1 – cos D) Similarly: 4(s – a)(s – b) = (–a + b + c + d)(a – b + c + d) = 2ab(1–cos B) + 2cd(1 + cos D) Multiplying these two equations together gives: 16(s – a)(s – b)(s – c)(s – d) = 4a2b2(1– cos2 B) + 4c2d2(1– cos2 D) + 4abcd((1 + cos B)(1 + cos D) + (1–cos B) (1–cos D)) = 4a2b2 sin2 B + 4c2d2 sin2 D + 8abcd(1 + cos B cos D) and substituting back into equation (i) produces:
K = √s(s – a)(s – b)(s – c) This is valid, since every triangle can be inscribed in a circle and so is indeed a cyclic polygon.
Briggs, Henry (1561–1630) British Logarithms Born in February 1561 (the exact birth date is not known) in Yorkshire, England, scholar Henry Briggs is remembered for the development of base10 logarithms, revising the approach first taken by the inventor of logarithms, JOHN NAPIER (1550–1617). Today such common logarithms are sometimes called Briggsian logarithms. In 1617, after consulting with Napier, Briggs published logarithmic values of the first 1,000 numbers and, in 1624, in his famous text The Arithmetic of Logarithms, the logarithmic values of another 30,000 numbers, all correct to 14 decimal places.
54 Brouncker, Lord William Briggs graduated from St. John’s College of Cambridge University in 1581 with a master’s degree and was appointed a lectureship at the same institution 11 years later in 1592 to practice both medicine and mathematics. Four years later Briggs became the first professor of geometry at Gresham College, London, when he also developed an avid interest in astronomy. Around this time Napier had just developed his theory of logarithms as a mathematical device specifically aimed at assisting astronomers with difficult arithmetical computations. Briggs read Napier’s text on the subject in 1614 and, with keen excitement, arranged to visit the Scottish scholar in the summer of 1615. The two men agreed that the theory of logarithms would indeed be greatly simplified under a base10 system, and two years later Briggs published his first tables of logarithmic values in Logarithmorum chilias prima (Logarithms of numbers 1 to 1,000). Later, in his famous 1624 piece, Briggs published logarithmic values for the numbers 1 through 20,000 and from 90,000 to 100,000. (The gap of 70,000 numbers was filled three years later by the two Dutch scholars, Adriaan Vlacq and Ezechiel de Decker.) Briggs was appointed chair of geometry at Oxford University in 1619. He remained at Oxford pursuing interests in astronomy and classical geometry until his death on January 26, 1630. In his inaugural lecture at Gresham College, ISAAC BARROW (1630–77) expressed gratitude on behalf of all mathematicians for the outstanding work Briggs had accomplished through the study of logarithms.
With the restoration of the monarchy in 1660 and the election of King Charles II to the throne, Brouncker was appointed chancellor to Queen Anne and keeper of the Great Seal. Brouncker was also appointed president of the newly created ROYAL SOCIETY of London in 1662. In mathematics, Brouncker studied infinite SERIES and made a number of remarkable discoveries. He devised a series method for computing LOGARITHMs and a surprising continuedfraction expression for π: 4 = 1+ π
12 32 2+ 52 2+ 2 +L
Brouncker used this formula to correctly calculate the value of π to the 10th decimal place. The great English mathematician JOHN WALLIS later published these results on Brouncker’s behalf. Brouncker also studied DIOPHANTINE EQUATIONs and found a general method for solving equations of the form nx2 + 1 = y2. LEONHARD EULER later called this equation “Pell’s equation,” after English mathematician John Pell (1611–84), not realizing that it was Brouncker, not Pell, who had studied it so intensively. Pell’s name, unfortunately, remains attached to this equation to this day. Brouncker died on April 5, 1684, in London, England. He never received fame as a great mathematician for he tended to focus on solving problems posed by others rather than forging original pathways in mathematics research.
Brouncker, Lord William (ca. 1620–1684) British Calculus William Brouncker is best remembered for his work in the early development of CALCULUS and also as one of the first mathematicians in Britain to use CONTINUED FRACTIONs. Very little is known of Brouncker’s early life, including the exact year of his birth, for example, and his nationality. Records do show, however, that he entered Oxford University at the age of 16 to study mathematics, languages, and medicine. He received a doctorate of medicine in 1647, pursuing mathematics and its applications to music, mechanics, and experimental physics as an outside interest throughout his life. Brouncker held Royalist views and took an active part in the political turmoil of the time in England.
brute force The method of establishing the validity of a claim by individually checking each and every instance of the claim is called brute force. Mathematicians much prefer devising general arguments and principles to demonstrate the validity of a claim, rather than resort to this method. Some claims, however, seem amenable only to the technique of brute force. For instance, the following numbernaming puzzle can be solved only via brute force: Think Count obtain letters
of a number between one and 100. the number of letters in its name to a second number. Count the number of in the name of the second number to
Buffon needle problem 55 obtain a third number. Continue this way until the chain of numbers obtained this way ends on a number that repeats. This repeating number, in every case, is the number four. As every number between one and 100 can be written in 12 or fewer letters, the second number obtained will lie between three and 12. One checks, by brute force, that each of these numbers eventually leads to the number four. At present, the bruteforce method is the only known technique guaranteed to yield an optimal solution to the famous TRAVELINGSALESMAN PROBLEM. This problem is of significant practical importance. Unfortunately, even with the fastest computers of today, the bruteforce approach cannot be carried out in any feasible amount of time. The famous FOURCOLOR THEOREM was solved in the 1970s by reducing the problem to a finite, but extraordinarily large, number of individual cases that were checked on a computer by brute force. See also GOLDBACH’S CONJECTURE.
Solving the Buffon needle problem
In the diagram we see that if the needle lands within an angle as indicated by either sector labeled θ, then it will not fall across the upper crack, but it will do so if instead it lands in the sector of angle π –2θ. Thus the probability that the needle will fall across the crack, given that it lands a distance x units below a crack, is π − 2θ 2 Px = = 1 − sin −1x . Summing, that is, integrating, π π over all possible values of x gives us the total probability P we seek: P=
1
Buffon needle problem (BuffonLaplace problem)
= 1−
In 1733 French naturalist Georges Buffon proposed the following problem, now known as the Buffon needle problem:
=
A needle one inch long is tossed at random onto a floor made of boards one inch wide. What is the probability that the needle lands crossing one of the cracks? One can answer the puzzle as follows: Suppose that the needle lands with lowest end a distance x from a crack. Note that 0 ≤ x ≤ 1.
1
2
2
1
∫0 Px dx = ∫0 1 − π sin −1 x dx = 1 − π ∫0 sin −1 xdx 1 2 2 π x sin −1 x + 1 − x 2 = 1 − − 1 0 π π2
2 π
In principle, this problem provides an experimental method for computing PI: simply toss a needle onto the floor a large number of times, say 10,000, and count the proportion that land across a crack. This proportion 2 should be very close to the value π . In practice, however, this turns out to be a very tedious approach. See also MONTE CARLO METHOD.
C calculus (infinitesimal calculus) The branch of math
An amusing activity in mathematics searches for fractions that remain unchanged by the action of “anomalous cancellation.” For example, deleting the digit 6 from each of the numerator and denominator of the fraction 65/26 produces the number 5/2. Curiously, 65/26 does equal five halves. Other fractions with this 64 4 98 8 95 5 property include: — = – , — = – , and — = – . Of 16 1 49 4 19 1 course deleting digits this way is an invalid mathematical operation, and the equalities obtained here are purely coincidental. In ALGEBRA the word cancellation is used in two settings. Akin to the process of cancellation for fractions, one may simplify a rational expression of the form xy/xz by canceling the factor x from both the numerator and denominator. This gives xy/xz as equal
ematics that deals with the notion of continuous growth and change is called calculus. It is based on the concept of INFINITESIMALs, exceedingly small quantities, and on the concept of a LIMIT, quantities that can be approached more and more closely but never reached. By treating continuous changes as if they consisted of infinitely small increments, DIFFERENTIAL CALCULUS can be used, for example, to find the VELOCITY of a moving object at any particular instant. INTEGRAL CALCULUS represents the reverse process, finding the aggregate endresult if the continuous change is already known. For example, by integrating the instantaneous velocity of an object over a given time period, one finds the total distance the object moved during this time. The word calculus comes from the Latin word calx for “pebble,” which in turn is derived from the Greek word chalis for “limestone.” Small beads or stones arranged in a counting board or on an ABACUS were often used to aid mathematical calculations, and the word calculus came to refer to all mathematical activity. Today, however, the word is used almost exclusively to denote the study of continuous change. See also HISTORY OF CALCULUS (essay).
cancellation In
to y/z. Thus, for instance, the expression
x 2 + 3x + 2 x2 + x
x+2 (x + 1)(x + 2) simplifies to . x +1 x(x + 1) The process of removing two equal quantities from an equation via subtraction is also called cancellation. For example, in the equation x + 2y = 5 + 2y, the term 2y can be cancelled to leave x = 5. Both actions fall under the umbrella of a general cancellation law. In an abstract setting, a BINARY OPERATION “*” is said to satisfy such a law if whenever a*x = b*x holds, a = b follows. For the operation of addition, this reads: =
ARITHMETIC,
the process of dividing the numerator and the denominator of a FRACTION by a common factor to produce a simpler fraction is called cancellation. For example, the fraction 18/12, which is (6 × 3)/(6 × 2), can be simplified to 3/2 by canceling “6” from the fraction.
a + x = b + x implies a = b 56
cancellation 57
History of Calculus The study of calculus begins with the study of motion, a topic that has fascinated and befuddled scholars since the time of antiquity. The first recorded work of note in this direction dates back to the Greek scholars PYTHAGORAS (ca. 569–475 B.C.E) and ZENO OF ELEA (ca. 500 B.C.E.), and their followers, who put forward the notion of an INFINITESIMAL as one possible means for explaining the nature of physical change. Motion could thus possibly be understood as the aggregate effect of a collection of infinitely small changes. Zeno, however, was very much aware of fundamental difficulties with this approach and its assumption that space and time are consequently each continuous and thus infinitely divisible. Through a series of ingenious logical arguments, Zeno reasoned that this cannot be the case. At the same time, Zeno presented convincing reasoning to show that the reverse position, that space is composed of fundamental indivisible units, also cannot hold. The contradictory issues proposed by Zeno were not properly resolved for well over two millennia. The concept of the infinitesimal also arose in the ancient Greek study of area and volume. Scholars of the schools of PLATO (428–348 B.C.E.) and of EUDOXUS OF CNIDUS (ca. 370 B.C.E.) developed a “method of exhaustion,” which attempted to compute the area or volume of a curved figure by confining it between two known quantities, both of which can be made to resemble the desired object with any prescribed degree of accuracy. For example, one can sandwich a circle between two ngons, one inscribed and one circumscribed. As one can readily compute the area of a regular ngon, the formula for the area of a circle follows by taking larger and larger values of n. (See AREA.) The figure of a regular ngon as n grows differs from that of a true circle only by an infinitesimal amount. ARCHIMEDES OF SYRACUSE (287–212 B.C.E.) applied this method to compute the area of a section of a PARABOLA, and 600 years later, PAPPUS OF ALEXANDRIA (ca. 300–350 C.E.) computed the volume of a SOLID OF REVOLUTION via this technique. Although successful in computing the areas and volumes of a select collection of geometric objects, scholars had no general techniques that allowed for the development of a general theory of area and volume. Each individual calculation for a single specific example was hailed as a great achievement in its own right. The resurgence of scientific investigation in the mid1600s led European scholars to push the method of exhaustion beyond the point where Archimedes and Pappus had left it. JOHANNES KEPLER (1571–1630) extended the use of infinitesimals to solve OPTIMIZATION problems. (He also developed new mathematical methods for computing the volume of wine barrels.) Others worked on the problem of finding tangents to curves, an important practical problem
in the grinding of lenses, and the problem of finding areas of irregular figures. In 1635, Italian mathematician BONAVENTURA CAVALIERI wrote the first textbook on what we would call integration methods. He described a general “method of indivisibles” useful for computing volumes. The principle today is called CAVALIERI’S PRINCIPLE. French mathematician Gilles Personne de Roberval (1602–75) was the first to link the study of motion to geometry. He realized that the tangent line to a geometric curve could be interpreted as the instantaneous direction of motion of a point traveling along that curve. Philosopher and mathematician RENÉ DESCARTES (1596–1650) developed general techniques for finding the formula for the tangent line to a curve at a given point. This technique was later picked up by PIERRE DE FERMAT (1601–65), who used the study of tangents to solve maxima and minima problems in much the same way we solve such problems today. As a separate area of study, Fermat also developed techniques of integral calculus to find areas between curves and lengths of arcs of curves, which were later developed further by BLAISE PASCAL (1623–62) and English mathematicians JOHN WALLIS (1616–1703) and ISAAC BARROW (1630–77). At the same time scholars, including Wallis, began studying SERIES and INFINITE PRODUCTs. Scottish mathematician JAMES GREGORY (1638–75) developed techniques for expressing trigonometric functions as infinite sums, thereby discovering TAYLOR SERIES 40 years before BROOK TAYLOR (1685–1731) independently developed the same results. By the mid1600s, certainly, all the pieces of calculus were in place. Yet scholars at the time did not realize that all the varied problems being studied belonged to one unified whole, namely, that the techniques used to solve tangent problems could be used to solve area problems, and vice versa. A fundamental breakthrough came in the 1670s when, independently, GOTTFRIED WILHELM LEIBNIZ (1646–1716) of Germany and SIR ISAAC NEWTON (1642–1727) of England discovered an inverse relationship between the “tangent problem” and the “area problem.” The discovery of the FUNDAMENTAL THEOREM OF CALCULUS brought together the disparate topics being studied, provided a beautiful and natural perspective on the subject as a whole, and allowed scholars to make significant advances in solving geometric and physical problems with spectacular success. Despite the content of knowledge that had been established up until that time, it is the discovery of the fundamental theorem of calculus that represents the discovery of calculus. Newton approached calculus through a concept of “flowing entities.” He called any quantity being studied a “fluent” and its rate of change a FLUXION. Records show that (continues)
58 cancellation
History of Calculus (continued) he had developed these ideas as early as 1665, but he did not publish an account of his theory until 1704. Unfortunately, his writing style and choice of notation also made his version of calculus accessible only to a select audience. Leibniz, on the other hand, made explicit use of an infinitesimal in his development of the theory. He called the infinitesimal change of a quantity x a DIFFERENTIAL, denoted dx. Leibniz invented a beautiful notational system for the subject that made reading and working with his account of the theory immediately accessible to a wide audience. (Many of the symbols we use today in differential and integral calculus are due to Leibniz.) Leibniz formulated his approach in the mid1670s and published his account of the subject in 1684. Although it is now known that Newton and Leibniz had made their discoveries independently, matters at the time were not clear, and a bitter dispute arose over the priority for the discovery of calculus. In 1712 the ROYAL SOCIETY of England formed a special committee to adjudicate the issue. Applying the techniques to problems of the real world became the main theme of 18thcentury mathematics. Newton’s famous 1687 text Principia paved the way with its analysis of the laws of motion and the mechanics of the solar system. The Swiss brothers Jakob Bernoulli (1654–1705) and Johann Bernoulli (1667–1748) of the famous BERNOULLI FAMILY, champions of Leibniz in the famous dispute, studied the newly invented calculus and were the first to give public lectures on the topic. Johann Bernoulli was hired to teach differential calculus to the French nobleman GUILLAUME FRANÇOIS DE L’ HOPITAL (1661–1704) via written correspondence. In 1696 L’Hopital then published the content of Johann’s letters with his own name as author. Italian mathematician MARIE GAETANA AGNESI (1718–99) wrote the first comprehensive textbook dealing with both differential and integral calculus in 1755. The Swiss mathematician LEONHARD EULER (1707–83) and French mathematicians JOSEPHLOUIS LAGRANGE (1736–1813) and PIERRESIMON LAPLACE (1749–1827) were prominent in developing the theory of DIFFERENTIAL EQUATIONs. Euler also wrote extensively on the subject of calculus, showing how the theory can be applied to a vast range of pure and applied mathematical problems. Yet despite the evident success of calculus, some 18thcentury scholars questioned the validity and the soundness of the subject. The sharpest critic of Newton’s and Leibniz’s work was the Anglican Bishop of Coyne, George Berkeley (1685–1753). In his scathing essay, “The Analyst,” Berkeley demonstrated, convincingly, that both Newton’s notion of a fluxion
and Leibniz’s concept of an infinitesimal are illdefined, and that the foundations of the subject are consequently insecure. (Some historians suggest that Berkeley’s vehement criticisms were motivated by a personal disdain for the apparent atheism of the type of mathematician who argues that science is certain and that theology is based on speculation.) Mathematicians consequently began looking for ways to put calculus on a sound footing. Significant progress was not made until the 19th century, when French mathematician AUGUSTINE LOUIS CAUCHY (1789–1857) suggested that the notion of an infinitesimal should be replaced by that of a LIMIT. German mathematician KARL WEIERSTRASS (1815–97) developed this idea further and was the first to give absolutely clear and precise definitions to all concepts used in calculus, devoid of any mystery or reliance on geometric intuition. The work of German mathematician RICHARD DEDEKIND (1831–1916) highlighted the role properties of the real number system play in ensuring the validity of the INTERMEDIATEVALUE THEOREM and EXTREMEVALUE THEOREM and all the essential results that follow from them. Initially, calculus was deemed a theory pertaining only to continuous change and CONTINUOUS FUNCTIONs. German mathematician BERNHARD RIEMANN (1826–66) was the first to consider, and give careful discussion on, the integration of discontinuous functions. His definition of an integral is the one typically presented in textbooks today. At the end of the 19th century, French mathematician HENRI LÉON LEBESGUE (1875–1941) literally turned Riemann’s approach around and developed a concept of integration that can be applied to a much wider class of functions and class of settings. In constructing a Riemann integral, one begins by subdividing the range of inputs, the xaxis, into small intervals and adding areas of rectangles above these intervals of heights given by the function. This is akin to counting the value of a pocketful of coins by taking one coin out at a time, and adding the outcomes as one goes along. Lebesgue suggested, on the other hand, subdividing the range of outputs, the yaxis, into small intervals and measuring the size of the sets on the xaxis for which the function gives the desired output on the yaxis. This is akin to counting coins by first collecting all the pennies and determining their number, all the nickels and ascertaining the size of that collection, and so forth. In order to do this, Lebesgue had to develop a general “measure theory” for determining the size of complicated sets. His new theory proved to be fundamentally important, and it now has profound applications to a wide range of mathematical topics. It proved to be especially important to the sound development of PROBABILITY theory. See also CALCULUS; DIFFERENTIAL CALCULUS; GRAPH OF A FUNCTION; INTEGRAL CALCULUS; VOLUME.
Cardano, Girolamo 59 and for multiplication, assuming that x is not zero: a × x = b × x implies a = b The cancellation law holds for any mathematical system that satisfies the definition of being a GROUP. With the guaranteed existence of inverse elements, we have a*x = b*x yields a*x*x – 1 = b*x*x – 1, which is the statement a = b. It holds in MODULAR ARITHMETIC in the following context: ax ≡ bx (mod N) implies a ≡ b (mod N), only if x and N are COPRIME This follows since the statement ax ≡ bx (mod N) holds only if x(a – b) is a multiple of N. If x and N share no prime factors, then it must be the case that the term a – b contains all the prime factors of N and so is a multiple of N. Consequently, a ≡ b(mod N). The fact that 4 × 2 is congruent to 9 × 2 modulo 10, without 4 and 9 being congruent modulo 10, shows that the cancellation law need not hold if x and N share a common factor. See also ASSOCIATIVE; COMMUTATIVE PROPERTY; DISTRIBUTIVE PROPERTY.
Cantor, Georg (1845–1918) German Set theory Born on March 3, 1845, in St. Petersburg, but raised in Wiesbaden and in Frankfurt, Germany, mathematician Georg Cantor is remembered for his profound work on the theory of sets and CARDINALITY. From the years 1874 to 1895, Cantor developed a clear and comprehensive account of the nature of infinite sets. With his famous DIAGONAL ARGUMENT, for instance, he showed that the set of rational numbers is DENUMERABLE and that the set of real numbers is not, thereby establishing for the first time that there is more than one type of infinite set. Cantor’s work was controversial and was viewed with suspicion. Its importance was not properly understood at his time. Cantor completed a dissertation on NUMBER THEORY in 1867 at the University of Berlin. After working as a school teacher for a short while, Cantor completed a habilitation degree in 1869 to then accept an appointment at the University of Halle in 1869. He worked on the theory of trigonometric SERIES, but his studies soon required a clear understanding of the IRRATIONAL NUMBERs. This need turned Cantor to the study of general sets and numbers.
In 1873 Cantor proved that the set of all rationals and the set of ALGEBRAIC NUMBERs are both COUNTABLE, but that the set of real numbers is not. Twenty years earlier, French mathematician JOSEPH LIOUVILLE (1809–82) established the existence of TRANSCENDENTAL NUMBERs (by exhibiting specific examples of such numbers), but Cantor had managed to show, in one fell swoop, that in fact almost all numbers are transcendental. Having embarked on a study of the infinite, Cantor pushed forward and began to study the nature of space and dimension. In 1874 he asked whether the points of a unit square could be put into a onetoone correspondence with the points of the unit interval [0,1]. Three years later Cantor was surprised by his own discovery that this is indeed possible. His 1877 paper detailing the result was met with suspicion and was initially refused publication. Cantor’s friend and colleague, the notable JULIUS WILHELM RICHARD DEDEKIND (1831–1916), intervened and urged that the work be printed. Cantor continued work on transfinite sets for a further 18 years. He formulated the famous CONTINUUM HYPOTHESIS and was frustrated that he could not prove it. Cantor suffered from bouts of depression and mental illness throughout his life. During periods of discomfort, he turned away from mathematics and wrote pieces on philosophy and literature. (He is noted for writing essays arguing that Francis Bacon was the true author of Shakespeare’s plays.) Cantor died of a heart attack on January 6, 1918, while in a mental institution in Halle, Germany. Even though Cantor’s work shook the very foundations of established mathematics of his time, his ideas have now been accepted into mainstream thought. Beginning aspects of his work in SET THEORY are taught in elementary schools.
capital See INTEREST.
Cardano, Girolamo (Jerome Cardan) (1501–1576) Italian Algebra Born on September 24, 1501, in Pavia, Italy, scholar Girolamo Cardano is remembered as the first to publish solutions to both the general CUBIC EQUATION and to the QUARTIC EQUATION in his 1545 treatise Ars magna (The great art). Even though these results were due to SCIPIONE DEL FERRO
60 Cardano’s formula
Cardano’s formula (CardanoTartaglia formula) See CUBIC EQUATION.
cardinality In common usage, the cardinal numbers are the counting numbers 1, 2, 3, … These numbers represent the sizes of FINITE sets of objects. (Unlike the ORDINAL NUMBERS, however, the cardinal numbers do not take into account the order in which elements appear in a given set.) In the late 1800s German mathematician GEORG CANTOR (1845–1918) extended the notion of cardinality to include meaningful examination of the size of infinite sets. He defined two sets to be of the same cardinality if their members can be matched precisely in a onetoone correspondence. That is, each element of the first set can be matched with one element of the second set, and vice versa. For example, the set of people {Jane, Lashana, Kabeer} is of the same cardinality as the set of dogs {Rover, Fido, Spot}, since one can draw leashes between owners and dogs so that each owner is assigned just one dog, and each dog is leashed to one owner. Each of these sets is said to have cardinality 3. (Both sets are of the same cardinality as the set {1, 2, 3}.) Two sets of the same cardinality are said to be equipotent (equipollent, equinumerable, or, simply, equivalent). The set of all counting numbers {1, 2, 3, …} is equipotent with the set of all integers {…, –2, –1, 0, 1, 2, …}. This can be seen by arranging each set of numbers in a list and matching elements according to their positions in the list:
↔
↔
↔
↔
↔
1 2 3 4 5 6 … ↔
(1465–1526), NICCOLÒ TARTAGLIA (1499–1557), and to his assistant LUDOVICO FERRARI (1522–65), Cardano unified general methods. He was an outstanding mathematician of the time in the fields of ALGEBRA, TRIGONOMETRY, and PROBABILITY. At the age of 19, Cardano entered Pavia University to study medicine but quickly transferred to the University of Padua to complete his degree. He excelled at his studies and earned a reputation as a top debater. He graduated with a doctorate in medicine in 1526. Cardano set up a small medical practice in the village of Sacco, but it was not at all successful. He obtained a post as a lecturer in mathematics at the Piatti Foundation in Milan, where he pursued interests in mathematics while continuing to treat a small clientele of patients. In 1537 Cardano published two mathematics texts on the topics of arithmetic and mensuration, marking the start of a prolific literary career. He wrote on such diverse topics as theology, philosophy, and medicine in addition to mathematics. In 1539 Cardano learned that an Italian mathematician by the name of Tartaglia knew how to solve cubic equations, a topic of interest to Cardano since he and Ferrari, his assistant, had discovered methods for solving quartics, if the method for cubics was clear. Tartaglia revealed his methods to Cardano under the strict promise that the details be kept secret. (At the time, Renaissance scholars, such as Tartaglia, were often supported by rich patrons and had to prove their worth in public challenges and debates. Keeping methods secret was thus of key importance.) When Cardano later learned that another scholar by the name of del Ferro had discovered methods identical to those of Tartaglia decades earlier, Cardano no longer felt obliged to keep the solution secret and published full details in his famous 1545 piece Ars magna (The great art). Tartaglia was outraged by this act, and a bitter dispute between the two men ensued. Although Cardano properly credits Ferrari and Tartaglia as the first scholars to solve the cubic equation, it should be noted that Cardano properly identified general approaches that unified previous methods. Cardano also recognized that solutions would often involve COMPLEX NUMBERS and was the first scholar to make steps toward understanding these quantities. He died on September 21, 1576, in Rome. See also RAFAEL BOMBELLI.
0 –1 1 –2 2 –3 … This procedure shows that any two sets whose elements can be listed are equipotent. Such a set is said to be DENUMERABLE, and Cantor denoted the cardinality of any denumerable set ℵ0, pronounced “aleph null.” Cantor’s first DIAGONAL ARGUMENT shows that the set of all counting numbers, the set of integers, and the set of all rational numbers are each denumerable sets and so each have cardinality ℵ0. Mathematicians have shown that the set of all ALGEBRAIC NUMBERs is also denumerable. Not all infinite sets, however, are denumerable, as Cantor’s second diagonal argument shows. For
cardinality 61 instance, it is not possible to place the set of all real numbers in a list, and so, in some welldefined sense, the set of real numbers is “more infinite” than the set of counting numbers or the rationals. Cantor denoted the cardinality of the set of real numbers c, for “continuum.” We have:
= .1010101… .) And conversely, given a real number x between zero and one, create a subset B of counting numbers as dictated by the placement of 1s in its binary expansion. (Thus, for instance, the real number x = .010000… corresponds to the subset {2}.)
ℵ0 < c
(There is one technical difficulty with this correspondence. The numbers 0.01000… and 0.001111…, for instance, represent the same real number, yet correspond to different subsets of the counting numbers. Mathematicians have shown that this difficulty can be obviated.) We have:
In general, the cardinality of one set A is said to be less than the cardinality of another set B if it is possible to match each element of A with a unique element of B, but not vice versa. (Thus A is equipotent with a proper subset of B but not equipotent with B itself.) For example, the cardinality of {Jane, Lashana, Kabeer}, which we denote as 3, is less than the cardinality of the set {Rover, Fido, Spot, Tess, Rue, Tucker, Jet}, which we denote as 7. Although we can match the owners with distinct dogs, we cannot match the dogs with distinct owners. An infinite set has the property that it is equipotent with a proper subset of itself. (This is usually taken as the definition of what it means for a set to be infinite.) For example, since the set of integers is equipotent with the subset of counting numbers, the set of integers is indeed an infinite set. The graph of the tangent function π and x = – π shows that each y = tan x between x = – – 2 2 point on the yaxis is matched with a unique point on π π the xaxis in the interval − 2 , 2 , and vice versa. Thus the set of all points on the entire number line (the yaxis) has the same cardinality as the set of all points just in a finite interval. The set of all real numbers is indeed an infinite set. The set of all subsets of a set A is called the “power set” of A, denoted P(A). For example, the power set of {a,b,c,} is the set of eight elements: {Ø,{a}, {b},{c},{a,b},{a,c},{b,c},A}. (In general, the power set of a set with n elements has 2n elements.) In some sense, the power set of the set of all counting numbers A = {1,2,3,…} matches precisely with the set of all real numbers between zero and one. This can be seen as follows: Given a subset B of counting numbers, let x be the real number written as a decimal in base two whose kth digit is 1 if k belongs to B, and 0 otherwise. (Thus, for instance, the subset of odd counting numbers yields the real number x
P(N) = R where N denotes the set of natural numbers and R the set of real numbers. The cardinality of the counting numbers is ℵ0, and the above argument shows that the cardinality of its power set is c. This suggests, as for finite sets, that the power set of a set is always of “larger” cardinality than the original set. Cantor used the following argument to prove that this is indeed the case: Let A be a set and consider its power set P(A). Since every element a of A gives rise to the element {a} of P(A), we can certainly match the elements A with distinct elements of P(A). We now show, however, that it is not possible to reverse the process and match the elements of P(A) with distinct elements of A. Suppose, to the contrary, we have specified a way to associate with each subset of A a distinct element of A. For example, the subset {a,b,c} might be matched with the element a, and the subset {a,c,e,g,…} with b. Notice that the first subset contains the element a with which it is matched, but the second subset does not contain the element b with which it is matched. Let U be the set of all elements of A that are used in the above correspondence, but are not elements of the subsets to which they are assigned. (For instance, b above is an element of U, but a is not.) The subset U must be assigned some element of A, call it u. Now ask: Is u a member of U? The set U cannot contain u by the very definition of U. But in that case u satisfies the definition of being in U. So by not being in U, u must
62 cardioid be in U, and by being in U, u cannot be in U. This absurdity shows that there cannot be a meaningful correspondence that assigns distinct elements of A to subsets of A after all. Given an infinite set A, Cantor had thus shown that the sets P(A), P(P(A)), P(P(P(A))), … form a neverending chain of increasingly larger infinite sets. Thus, in a very definite sense, there are infinitely many different types of infinity. At the other end of the spectrum, the study of denumerable sets shows that every infinite set contains a denumerable subset. Thus of all the infinite sets, denumerable sets are the “smallest” type of infinite sets. The CONTINUUM HYPOTHESIS asks whether or not there is an infinite set with cardinality that lies somewhere between that of N and P(N) = R. One might suppose that P(R), the power set of the set of all points on the real number line, is R2, the set of all points in the plane, or, equivalently, that the power set of the set of all points in the unit interval [0,1] is the set of all points inside the unit square [0,1] × [0,1]. Surprisingly, this is not the case: there are just as many points in the unit square as there are in a unit interval. This is seen as follows: Associate to each point (x,y) with 0 ≤ x ≤ 1 and 0 ≤ y ≤ 1, each written as an infinite decimal expansion, x = 0.x1x2x3… and y = 0.y1y2y3…, the real number r = 0.x1y1x2y2x3y3… in the interval [0,1], and, conversely, match each real number r = 0.r1r2r3r4r5r6… with the point (0.r1r3r5…, 0.r2r4r6…) in the unit square.
(x2 + y2 – ax)2 = a2(x2 + y2). The PARAMETRIC EQUATIONS of the curve are x = acosθ(1 + cosθ) and y = asinθ(1 + cosθ). The curve has area oneandahalf times the area of either generating circle, and perimeter eight times the radius. The cardioid was first studied extensively by Italian mathematician Johann Castillon, who also coined its name in 1741. See also CYCLOID.
Cartesian coordinates (orthogonal coordinates, rectangular coordinates) One of the biggest breakthroughs in the development of mathematics occurred when geometry and algebra were united through the invention of the Cartesian coordinate system. Credited to 17thcentury French mathematician and philosopher RENÉ DESCARTES (whose name Latinized reads Cartesius), Cartesian coordinates provide a means of representing each point in the plane via a pair of numbers. One begins by selecting a fixed point O in the plane, called the origin, and drawing through it two perpendicular number lines, called axes, one horizontal and one vertical, and both with the point O at the zero position on the line. It has become the convention to set the positive side of the horizontal number line to the right of O, and the positive side of the vertical number line above O, and to call the horizontal axis the xaxis, and the vertical one the yaxis. The Cartesian coordinates of a point P in the plane is a pair of numbers (x,y) which then describes the location of that point as follows:
(Again there is a technical difficulty caused by those real numbers that have two different decimal representations. For instance, onehalf can be written both as 0.5000… and 0.4999… Mathematicians have shown that this difficulty can be obviated.) It turns out that P(R) corresponds to the set of all possible realvalued functions y = f(x). See also INFINITY; PEANO’S CURVE.
The xcoordinate, or “abscissa,” is the horizontal distance of the point from O along the horizontal axis. (A positive distance represents a point to the right of the vertical axis; a negative distance one to the left.) The ycoordinate, or “ordinate,” is the vertical distance of the point from O along the vertical axis. (A positive distance represents a point located above the horizontal axis, and a negative distance one located below.)
cardioid The heartshaped curve traced by a point on the circumference of one circle as it rolls around another circle of equal size is called a cardioid. In POLAR COORDINATES, the cardioid is given by an equation of the form r = a(1– cosθ) where a is the common radii of the circles, and in CARTESIAN COORDINATES by
For example, if the bottom left corner of this page is the origin of a Cartesian coordinate system, with xand yaxes marked in units of inches, then the point with coordinates (4, 1) lies four inches to the right of the left edge of the page, and one inch above the bottom of the page.
casting out nines 63 Extending this idea to threedimensions, points in space can be specified by a triple of numbers (x,y,z) representing the distances along three mutually perpendicular number lines. The coordinate axes are usually called the x, y, and zaxes. They intersect at a point O, called the origin, which is zero on all three number lines. The axes could be oriented to either form a lefthanded or a righthanded system. Coordinate Geometry The advent of a coordinate system allowed mathematicians, for the first time, to bring the power of algebra to the study of geometry. For example, straight lines are represented as sets of points (x,y) that satisfy equations of the form y = mx + b. Multiplying the SLOPE m of one line with the slope of another quickly ascertains whether or not those two lines are perpendicular, for example. (The product of the slopes of two perpendicular lines is –1.) French mathematician NICOLE ORESME (1323–82) was the first to describe a way of graphing the relationship between an independent variable and a dependent one, and thus the first to make steps toward uniting geometry and algebra. The explicit construction of what we would call a coordinate system first appeared with the work of French lawyer and amateur mathematician PIERRE DE FERMAT (1601–65). Starting with some horizontal reference line to represent an independent variable x, Fermat would graphically depict the relationship of a second variable y to it as a line segment, held at a fixed angle to the reference line, whose length would vary according to the variable y as it slides along the xaxis. Fermat did not think in terms, however, of identifying a second axis, nor did he require the line segment representing y to be perpendicular to the xaxis. In his famous 1637 text La géométrie (Geometry), René Descartes independently described similar methods for representing algebraic relationships graphically. Because the work of Fermat was not published until after his death, the discovery of coordinate geometry was attributed to Descartes. Because Fermat and Descartes interpreted the unknown variable y in an algebraic relationship as a physical length, both scholars only ever considered positive coordinates. English mathematician JOHN WALLIS (1616–1703) was the first to introduce the possibility of negative coordinates. The idea of setting a fixed second
axis, the yaxis, perpendicular to the xaxis was not popular until the mid 1700s. It was an idea that seemed to evolve gradually. SIR ISAAC NEWTON (1642–1727) is considered the originator of POLAR COORDINATES. See also COORDINATES; GRAPH OF A FUNCTION.
Cartesian product (cross product, external direct product, product set, set direct product) Given two sets A and B, their Cartesian product, denoted A × B, is the set of all ordered pairs (a,b), where a ∈ A and b ∈ B. For example, if A = {1,2,3} and B = {α,β}, then: A × B = { (1,α), (2,α), (3,α), (1,β), (2,β), (3,β) } This is different from the set B × A. If sets A and B are both finite, with n and m elements, respectively, then A × B is a finite set with nm elements. German mathematician GEORG CANTOR (1845–1918) showed that if A and B are both infinite COUNTABLE sets, then their Cartesian product A × B is again countable. The Cartesian product of three sets A, B, and C, denoted A × B × C, is defined as the set of all ordered triples (a,b,c), with a ∈ A, b ∈ B, and c ∈ C. The Cartesian product of any finite collection of sets is defined similarly. Any SEQUENCE can be thought of as an element of the Cartesian product of a countable number of sets. If two sets A and B have a particular structure (they might both be GROUPs or VECTOR SPACEs, for instance), then it is usually possible to give the Cartesian product A × B the same structure. For example, if A and B are groups with group operations * and •, respectively, then A × B has the structure of a group with group operation given by: (a1, b1) · (a2, b2) = (a1 * a2, b1 • b2) The Klein fourgroup is the Cartesian product of the twoelement group Z2 = {0,1} with itself. (The group operation for Z2 is addition in mod 2 MODULAR ARITHMETIC.) See also SET THEORY.
casting out nines The
DIVISIBILITY RULES show that the remainder of any number, when divided by 9, is the sum of its digits. For example, 59,432,641 leaves a
64 Catalan, Eugène Charles remainder of 5 + 9 + 4 + 3 + 2 + 6 + 4 + 1 = 34 when divided by 9, which corresponds to a remainder of 3 + 4 = 7. Any sets of digits that sum to 9, such as the 5 and the 4 in the first and third positions of the number above, can be ignored when performing this calculation, for they will not contribute to the remainder. The method of “casting out nines” is the process of deleting groups of digits summing to 9. The sum of the digits that survive is the remainder that number yields upon division by 9. For example, 59,432,641 → 932,641 → 934 → 34 shows, again, that this number leaves a remainder of 3 + 4 = 7 when divided by 9. This method is often used to check arithmetical work. For example, we can quickly determine that 563 × 128 cannot equal 72,364. Upon division by 9, 563 leaves a remainder of 5, 128 a remainder of 2, and so their product leaves a remainder of 5 × 2 = 10, which is 1. Yet 72,364 has a remainder of 4. Long lists of additions, subtractions, and multiplications can be quickly checked this way. Of course, errors may still be present if, by chance, remainders happen to match. For example, casting out nines will not detect that 632 × 723 = 459,636 is incorrect.
Catalan, Eugène Charles (1814–1894) Belgian Number theory Born on May 30, 1814, mathematician Eugène Catalan is best remembered for his work in NUMBER THEORY and for the famous series of numbers that bears his name. In 1844 Catalan conjectured that 8 and 9 are the only two consecutive integers that are both nontrivial powers (8 = 23 and 9 = 32). Establishing this claim, today known as the CATALAN CONJECTURE, stymied mathematicians for over a century. It was only recently resolved. Catalan’s career in academia was turbulent. After entering the École Polytechnique in 1833 he was expelled the following year for engaging in radical political activity. He was later given permission to return to complete his degree and in 1838 was offered a post as lecturer at the institution, which he accepted. His political conduct, however, hampered his ability to advance beyond this entrylevel position. Catalan worked in the field of CONTINUED FRACTIONs and achieved some fame for publishing a simplified solution to LEONHARD EULER’s “polygon division problem.” This challenge asks for the number of ways to divide a regular polygon into triangles
using nonintersecting diagonals. While not being the first to solve the problem (in fact the problem was first stated and solved by 18thcentury Hungarian mathematician J. A. Segner and then studied by Euler), Catalan used an approach that was particularly elegant. His 1838 paper on the topic, “Note sur une équation aux differences finie” (Note on a finite difference equation), was very influential because of the method it detailed. The sequence of numbers that arise in the study of the problem are today called the CATALAN NUMBERS. They remain his standing legacy. Catalan died on February 14, 1894, in Liège, Belgium.
Catalan conjecture In his 1844 letter to Crelle’s Journal, EUGÈNE CHARLES CATALAN conjectured that the integers 8 and 9 are the only two consecutive integers that are both powers (8 = 23 and 9 = 32). He was not able to prove his claim, and establishing the truth or falsehood of the conjecture became a longstanding open problem. In April 2002, amateur mathematician Preda Mihailescu announced to the mathematics community that he had completed a proof demonstrating Catalan’s assertion to be true. Beforehand, Mihailescu had proved a series of related results, all while working at a Swiss fingerprinting company and exploring mathematics as an outside interest. Mathematicians are currently reviewing his final step of the work.
Catalan numbers In 1838 EUGÈNE CHARLES CATALAN studied the problem of finding the number of different ways of arranging n pairs of parentheses. For example, there is one way to arrange one set: ( ), two ways to arrange two pairs: ( ) ( ) and (( )), and five ways to arrange three pairs: ((( ))), (( )( )), (( )) ( ), ( ) (( )), and ( ) ( ) ( ). Is there a general formula for the number of ways to arrange n pairs of parentheses? This puzzle is today known as “Catalan’s problem.” As Catalan showed, the solution is given by the formula:
Cn =
2 ⋅ 6 ⋅ 10 ⋅ L ⋅ (4n − 2) (n + 1)!
yielding the sequence of numbers C1 = 1, C2 = 2, C3 = 5, C4 = 14, C5 = 42, …, now called the Catalan numbers. It is convenient to set C0 = 1. Some algebraic manipulation
catenary 65 shows that the Catalan numbers can also be expressed in terms of BINOMIAL COEFFICIENTs: Cn =
2n ⋅ 1 ⋅ 3 ⋅ 5 ⋅ L ⋅ (2n − 1) (n + 1)!
=
2 ⋅ 4 ⋅ 6 ⋅ L ⋅ (2n) 1 ⋅ 3 ⋅ 5 ⋅ L ⋅ (2n − 1) ⋅ (n + 1)! n!
=
1 (2n)! ⋅ n + 1 n! n!
=
1 ⎛ 2n⎞ n + 1 ⎜⎝ n ⎟⎠
Thus the Catalan numbers can be found in PASCAL’s TRIANGLE as the middle entry of every alternate row divided by one more than the row number, regarding the apex of the triangle as row zero. One can show that the Catalan numbers satisfy the relationship: C0 = 1 Cn = C0Cn–1 + C1Cn–2 + … + Cn–1C0 The Catalan numbers appear as the solution to a surprising number of different mathematical problems. We list here just a few examples. 1. Euler’s Polygon Division Problem: How many ways are there to divide an (n + 2)sided polygon into n triangles using nonintersecting diagonals of the polygon? A threesided polygon, that is, a triangle, is already appropriately subdivided. There is one solution to the problem, namely, do nothing. A square can be subdivided into two triangles two different ways. One can check that a pentagon can be so subdivided five different ways. In general the solution to this puzzle is the nth Catalan number. 2. Laddered Exponents: How many ways can one interpret a laddered exponent? For example, 32 has only one interpretation: it means 4 3 × 3 = 9. The expression 23 , however, can be inter4 4 preted two ways: (23) = 4096 or 2(3 ) = 2417851639229258349412352. In general, a laddered exponent with (n + 1) terms can be interpreted
Cn different ways. (This problem is equivalent to Catalan’s original parentheses puzzle.) 3. Handshakes across a Table: In how many different ways can n pairs of people sitting at a circular table shake hands simultaneously? No pair of handshakes may cross. Two people sitting at a table can shake hands only one way. Four people can accomplish the feat in only two ways. (Diagonal handshakes cross.) In general, n pairs of people can shake hands Cn different ways, for one can interpret two hands shaking as a pair of parentheses. 4. Stair Climbing: Starting at the base of a flight of stairs, in how many ways can one take n steps up and n steps down, in any order? (You will necessarily return to the base of the steps on completion of the walk.) There is one way to take two steps: one step up followed by one step down, and two ways to take four steps: two up, two down, or one up, one down, repeated twice. In general there are Cn ways to accomplish this task. (Thinking of a left parenthesis as an “up step” and a right parenthesis as a “down step,” we can see that this puzzle too is equivalent to Catalan’s original problem.) 5. Summation Problem: Select n numbers from the set {1,2,3,…,2n} so that their sum is a multiple of n + 1. Can this be done? If so, in how many different ways? Consider the case n = 3, for example. There are five ways to select three numbers from the set {1, 2, 3, 4, 5, 6} with sum divisible by four: 1 + 2 + 5 = 8; 1 + 3 + 4 = 8; 1 + 5 + 6 = 12; 2 + 4 + 6 = 12; 3 + 4 + 5 = 12. In general, these puzzles can always be solved, and there are Cn ways to do them. See also CATALAN CONJECTURE.
catenary The shape of the curve formed by a uniform flexible cable hanging freely between two points, such as an electric cable between two telegraph poles, is called a catenary. GALILEO GALILEI (1564–1642) thought this curve to be a PARABOLA, but German scholar Joachim Jungius (1587–1657) later proved that this could not be the case. Jacques Bernoulli (1654–1705) of the famous
66 Cauchy, AugustinLouis BERNOULLI FAMILY was the first to write down the formula for the catenary. Up to constants, it is given by: y=
ex + e−x = cosh(x) 2
This is the hyperbolic cosine function from the set of HYPERBOLIC FUNCTIONs. Engineers, in designing suspension bridges such as the Golden Gate Bridge in San Francisco, make extensive use of this function in their work. The name catenary comes from the Latin word catena for “chain.”
Cauchy, AugustinLouis (1789–1857) French Analysis, Calculus, Number theory Born on August 21, 1789, French scholar AugustinLouis Cauchy is remembered as one of the most important mathematicians of
AugustinLouis Cauchy, an eminent mathematician of the 19th century, was the first to use the notion of a “limit,” as it is now known, to develop a sound model of continuity and convergence in the theory of calculus. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
his time. With 789 mathematical papers and seven influential textbooks to his credit, Cauchy made significant contributions to the study of NUMBER THEORY, ANALYSIS, GROUP THEORY, DIFFERENTIAL EQUATIONs, and PROBABILITY. In his famous 1821 text Cours d’analyse (A course of analysis), Cauchy was the first to provide an exact, rigorous meaning of the terms derivative and integral as used in CALCULUS through the development of the notion of a LIMIT. Cauchy also properly defined the terms continuity and convergence. His insistence on the absolute need for rigor and clarity in all of mathematics had a lasting effect and set the standards of rigor required today of all mathematical research. After graduating from the L’École Polytechnique in 1807 with a degree in mathematics, Cauchy pursued a career in engineering. Noted as a promising practitioner in the field, Cauchy was assigned to the Ourcq Canal project and by age 21 managed to receive a highranking commission in Cherbourg as a military engineer. Despite his busy work life, Cauchy continued to pursue interests in mathematics. In 1811 he proved a result on the geometry of polyhedra, which he submitted for publication. He received considerable praise for this accomplishment and decided to change careers and pursue research in mathematics full time. A year later he returned to Paris and began looking for a faculty position at an academic institution. Cauchy was finally awarded an assistant professorship at the L’École Polytechnique in 1815. All the while, Cauchy continued to produce and publish mathematical results. The same year as his appointment, Cauchy won the Grand Prize of the L’Académie Royale des Sciences for his outstanding mathematical discoveries on the theory of waves. This recognition garnered him some notice in the scientific community, but real fame came to Cauchy when, another year later, he solved an outstanding problem posed by PIERRE DE FERMAT (1601–1665) on the properties of FIGURATE NUMBERS. Cauchy had now proved himself an expert in a surprisingly large number of disparate fields. Cauchy published an incredible number of papers during this early period of his life, at a rate of as many as two a week. He was so prolific that the editors of the French journal Comptes Rendu imposed a quota on him. In response, Cauchy persuaded a family member, who worked in the publishing field, to create a new journal that contained nothing but papers by him!
Cavalieri’s principle 67 The main thrust of much of Cauchy’s work was to make mathematics rigorous and precise. He insisted on providing clarity, precision, and rigor in all the courses he taught and in his published work. His famous 1821 text was in fact a course book for students developed with the intention of “doing calculus the correct way.” It is fair to say that Cauchy influenced the entire course of mathematical research by pointing out, and demanding, the need for absolute clarity and care in the development of new (and even previously established) ideas. It is said that the great mathematician PIERRESIMON LAPLACE (1749–1827), after attending a lecture given by Cauchy on the importance of the convergence of an infinite series, quickly ran home to check the convergence of all the infinite series he had used in his already published popular text Mécanique céleste on celestial mechanics. In 1826 Cauchy published seminal works in the field of number theory, and in 1829 he defined, for the first time, the notion of a complex function of a complex variable. Cauchy left Paris and the brewing politics of the royal regime in 1830. Upon his return a year later, he refused to swear an oath of allegiance to the new regime and consequently lost his academic position. It was not until the overthrow of Louis Philippe in 1848 that Cauchy regained his university position. Even though Cauchy’s publication rate slowed considerably during this trying time, he did accomplish important work on the theory of differential equations and applications to mathematical physics during this period. Cauchy left a standing mark on the development of calculus with his work on refining the logical basis of the subject and greatly influenced the study of complex functions. A number of fundamental concepts in the field of analysis are named in his honor, including a pair of equations known as the CauchyRiemann equations that determine whether or not a complex function is differentiable. Cauchy died in Sceaux, near Paris, France, on May 22, 1857. His collected works, Oeuvres completes d’Augustin Cauchy (The complete works of Augustin Cauchy), collated under the auspices of the Académie des Sciences, were published throughout the years 1882 to 1970 in a total of 27 volumes.
Cavalieri, Bonaventura Francesco (1598–1647) Italian Geometry Born in Milan, Italy, in 1598 (his exact birth date is not known), mathematician, and disciple of
GALILEO, Bonaventura Cavalieri is best remembered for his 1635 work Geometria indivisibilibus continuorum (A new geometry of continuous indivisibles) in which he introduced his famous “method of indivisibles” for determining the areas and volumes of curved figures. This work is considered a forerunner to the entire theory of INTEGRAL CALCULUS. While still a boy, Cavalieri joined the religious order Jesuati in Milan. In 1616, he transferred to the monastery in Pisa, where he met Galileo and developed an interest in mathematics. Even though Cavalieri taught theology for many years and became a deacon in the order, he actively pursued employment as a mathematician. In 1629 he received a position as a chair of mathematics at Bologna. By this time Cavalieri had developed his method of indivisibles. Based on ARCHIMEDES’ method of exhaustion and JOHANNES KEPLER’s theory of the infinitely small, Cavalieri’s technique provided a means to rapidly compute the area and volumes of certain geometric figures previously deemed too difficult for analysis. Cavalieri’s famous 1635 work describing these methods, however, was not well received and was widely attacked for its lack of rigor. In response, Cavalieri published a revised piece, Exercitationes geometricae sex, which successfully settled all concerns. This second piece was acknowledged as a masterpiece and deemed a necessary text of study for all 17thcentury scholars. Cavalieri also studied and wrote extensively on the topics of LOGARITHMs, CONIC SECTIONs, TRIGONOMETRY, optics, and astronomy. He developed a general rule for computing the focal length of lenses, and described the principles and design of a reflecting telescope. Cavalieri died on November 30, 1647, in Bologna, Italy. His name appears in all highschool geometry textbooks of today for the principle he devised. See also CAVALIERI’S PRINCIPLE.
Cavalieri’s principle Italian mathematician BONAVENCAVALIERI (1598–1647) identified a general principle today known as Cavalieri’s principle:
TURA
Solids of equal height have equal volumes if crosssections made by planes parallel to the bases at the same distances from these bases have equal areas.
68 Cayley, Arthur It is based on the idea that the volume of a deck of cards, for example, does not change even if the deck is skewed. A close examination of VOLUME explains why Cavalieri’s principle is true.
umes had been produced. The remaining six volumes were edited by A. R. Forsyth. See also CAYLEYHAMILTON THEOREM.
CayleyHamilton theorem English mathematician Cayley, Arthur (1821–1895) British Matrix theory, Geometry, Abstract algebra, Analysis Born on August 16, 1821, in Richmond, England, Arthur Cayley is remembered as a prolific writer, having produced 967 papers in all, covering nearly every aspect of modern mathematics. His most significant work, Memoir on the Theory of Matrices (1858) established the new field of matrices and MATRIX algebra. Cayley studied abstract groups and was the first to study geometry in ndimensional space with n a number greater than three. Cayley demonstrated a great aptitude for mathematics as a child. A schoolteacher recognized his talent and encouraged Cayley’s father to allow him to pursue studies in mathematics rather than leave school and enter the family retail business. Cayley attended Trinity College, Cambridge, and graduated in 1842. Unable to find an academic position in mathematics, Cayley pursued a law degree and practiced law for 14 years. During this time, however, Cayley actively studied mathematics and published over 250 research papers. In 1863 he was finally appointed a professorship in mathematics at Cambridge. In 1854, while working as a lawyer, Cayley wrote “On the Theory of Groups Depending on the Symbolic Equation θn = 1” and other significant papers that defined, for the first time, the notion of an abstract GROUP. At that time, the only known groups were PERMUTATION groups, but Cayley realized that the mathematical principles behind these structures also applied to matrices, number systems, and geometric transformations. This work allowed Cayley to begin analyzing the geometry of higherdimensional space. This, in turn, coupled with his newly developed matrix algebra, provided the foundation for the theory of quantum mechanics, as developed by Werner Heisenberg in 1925. Cayley was elected president of the British Association for the Advancement of Science in 1883. From the years 1889 to 1895, Cayley’s entire mathematical output was collated into one 13volume work, The Collected Mathematical Papers of Arthur Cayley. This project was supervised by Cayley himself until he died on January 26, 1895, at which point only seven vol
ARTHUR CAYLEY (1821–95) and Irish mathematician SIR WILLIAM ROWAN HAMILTON (1805–65) noted that any square MATRIX satisfies some polynomial equation. To see this, first note that the set of all square n × n matrices with real entries forms a VECTOR SPACE over the real numbers. For example, the set of 2 × 2 square matrices is a fourdimensional vector space with basis elements: 1 0 0 1 0 0 0 0 , , , and 0 0 0 0 1 0 0 1 a b and any 2 × 2 matrix A = is indeed a linear c d combination of these four linearly independent matrices: a b 1 0 0 1 0 0 0 0 A= = a + b + c + d 0 0 0 0 1 0 0 1 c d In general, the set of all n × n square matrices is an n2dimensional vector space. In particular then, for any n × n matrix A, the n2 + 1 matrices I, A, A2, A3, 2 …, An must be linearly dependent, that is, there is a linear combination of these elements that yields the zero matrix: 2
c0I + c1A + c2A2 + … + cn2An = 0 for some numbers c0, …, cn2 . This shows: Any n × n square matrix satisfies a polynomial equation of degree at most n2. Cayley and Hamilton went further and proved that any square matrix satisfies its own “characteristic polynomial”: Let A be an n × n square matrix and set x to be a variable. Subtract x from each diagonal entry of A and compute the DETERMINANT of the resulting matrix. This yields a polynomial in x
centrallimit theorem 69 of degree n. Then the matrix A satisfies this particular polynomial equation. 2 1 For example, consider the 2 × 2 matrix A = . 3 4 Then the “characteristic polynomial” of this matrix is: 2 − x 1 2 det = (2 − x)(4 − x) − 3 ⋅ 1 = x − 6x + 5 4 − x 3 One now checks that 2
2 1 2 1 1 0 A2 − 6A + 5I = −6 +5 3 4 3 4 0 1 7 6 12 6 5 0 = − + 18 19 18 24 0 5 0 0 = 0 0 is indeed the zero matrix. See also IDENTITY MATRIX; AND INDEPENDENT.
center of gravity of this balance point and the third mass. This procedure can, of course, be extended to find the center of gravity of any finite collection of masses. (A location computed this way is technically the center of mass of the system. If the force of gravity is assumed to be uniform, then the center of mass coincides with the center of gravity of the system.) This principle can be extended to locate the center of gravity of arbitrary figures in the plane (viewed as flat, uniformly dense objects held parallel to the ground). If the figure is composed of a finite collection of rectangles glued together, one locates the center of each rectangle, the mass of each rectangle, and then regards the system as a collection of individual masses at different locations. Applying Archimedes’ law of the lever as above locates the figure’s center of gravity. If a figure can only be approximated as a union of rectangles, one can find the approximate location of the center of gravity via this principle, and then improve the approximation by taking the LIMIT result of using finer and finer rectangles in the approximations. This approach will yield an INTEGRAL formula for the location of the center of gravity. See also CEVA’S THEOREM; SOLID OF REVOLUTION.
LINEARLY DEPENDENT
centrallimit theorem In the early 1700s scientists ceiling function See FLOOR/CEILING/FRACTIONAL PART FUNCTIONS.
center of gravity (balance point) The location at which the weight of the object held in space can be considered to act is called the object’s center of gravity. For example, a uniform rod balances at its midpoint, and this is considered its center of gravity. A flat rectangular plate made of uniform material held parallel to the ground balances at its center. This point is the figure’s center of gravity. Archimedes’ LAW OF THE LEVER finds the balance point P of a system of two masses m1 and m2 held in space. The twomass system can then be regarded as a single mass m1 + m2 located at P. The center of gravity of a system of three masses in the space can be found by finding the balance point of just two masses, using ARCHIMEDES’ law of the lever, and then applying the law a second time to find the
from a wide range of fields began to notice the recurring appearance of the NORMAL DISTRIBUTION in their studies and experiments. Any measurement that represents an average value of a sample, or an aggregate value of a series of results, tends to follow this classical bellshaped distribution. The work of MARQUIS DE PIERRESIMON LAPLACE in 1818 and Aleksandr Mikhailovich Lyapunov in 1901, and others, led to the establishment of the centrallimit theorem: If an experiment involves the repeated computation of the average value of N measurements (a different set of N measurements each time), then the set of average values obtained very closely follows a normal distribution—even if the original experiments do not. The larger the value of N, the better the approximation to a normal curve. One can go further and say that if the original experiments have mean µ and standard deviation σ, then the collection of average values also has mean µ, but
70 Ceva’s theorem
standard deviation
σ
. For example, a factory may
N produce light bulbs packaged in large shipping cartons, 100 per carton. Even though the lifespan of individual light bulbs may vary wildly following no recognizable distribution of values, the central limit theorem asserts that the average lifespan of the bulbs per carton is given by a normal distribution. Since the height of an individual is the aggregate effect of the growth rate of a large number of individual cells, the distribution of heights of men and woman is essentially normal, as is the distribution of heights of most anything that grows—cats, maple trees, or carrots, for example. Another version of the centrallimit theorem is useful when trying to ascertain what proportion p percent of the entire population possesses a certain property (such as “Has blood type AB” or “Will vote Republican next November”). As it is impossible to examine every individual on the globe, or poll every individual in the nation, one can examine a sample of individuals and compute the percentage in this sample with the desired property. The centrallimit theorem also asserts: If many different samples of N individuals are examined, then the distribution of the percentages of those samples possessing a particular property very closely follows a normal distribution (and the larger the value of N, the better is the approximation to a normal curve). This distribution has mean p, the true percentage of the population with this property, and standard deviation σ =
p(100 − p) . N
Both versions of the centrallimit theorem allow statisticians to make inferences and predictions based on statistical data. See also GEORGE PÓLYA; STATISTICS: INFERENTIAL.
Ceva’s theorem Let P, Q, and R be, respectively, points on sides BC, CA, and AB of a triangle ABC. (One is permitted to extend one or more sides of the triangle.) Then the lines connecting P to A, Q to B, and R to C are CONCURRENT if, and only if: BP CQ AR ⋅ ⋅ =1 PC QA RB
Here BP, for instance, represents the distance between points B and P, and the ratio BP is considered positive PC if the direction from B to P is the same as the direction from P to C, and negative if they are in opposite directions. This result is due to Italian mathematician Giovanni Ceva (1647–1734). It is equivalent to the statement that the operation of finding the CENTER OF GRAVITY of point masses is ASSOCIATIVE. For instance, given three masses at locations A, B, and C in a plane, one could locate the center of mass of the entire system by first computing the center of mass of just two points and then compute the center of mass of that result with the third point. Ceva proved that the same final result is produced no matter which two points are chosen initially. Ceva’s theorem can be proved mathematically by making repeated use of MENELAUS’S THEOREM.
chain rule (function of a function rule) If y = f(u) is a function of a quantity u, which in turn is a function of another quantity x, u = g(x) say, then y itself can be thought of as a function of x as a COMPOSITION of functions: y = f(g(x)). The chain rule states that the rate of change of y with respect to the quantity x is given by the formula: dy dy du = ⋅ dx du dx This can also be written: (f(g(x)))′ = f′(g(x))·g′ (x). For example, to differentiate y = (x2 + 2)100, we can write y = u100, where u = x2 + 2 and so: dy dy du d 100 d 2 = ⋅ = (u ) ⋅ (x + 2) = 100u 99⋅ 2x dx du dx du dx = 200x(x 2 + 1)99 The chain rule can be proved by making use of the formal definition of a derivative as a LIMIT: d f (g(x + h)) − f (g(x)) f (g(x)) = lim h→0 dx h = lim h→0
f (g(x + h)) − f (g(x)) g(x + h) − g(x) ⋅ g(x + h) − g(x) h
= f ′(g(x) ⋅ g ′(x)
Chebyshev, Pafnuty Lvovich 71 Intuitively, the concept is easy to grasp if we think of derivatives as rates of change. For example, if y changes a times as fast as u, and u changes b times as fast as x, then we expect y to change ab times faster than x. The chain rule extends to functions of more than one variable. For example, if z = f(x, y) is a function of two variables with each of x and y a function of t, then one can show that the total derivative of z with respect to t is given by: df ∂f dx ∂f dy = ⋅ + ⋅ dt ∂x dt ∂y dt See also PARTIAL DERIVATIVE.
change of variable See INTEGRATION BY
SUBSTITUTION.
chaos A situation in which a DYNAMICAL SYSTEM can appear to be random and unpredictable is called chaos. More precisely, mathematicians define a dynamical system to be chaotic if the set of all equilibrium points for the system form a FRACTAL. The term chaos was introduced by American mathematician James Yorke and Chinese mathematician TienYien Li in their 1975 seminal paper on iterations of functions on the real number line.
characteristic polynomial See CAYLEYHAMILTON THEOREM.
Chebyshev, Pafnuty Lvovich (Tchebyshev) (1821– 1894) Russian Number theory, Analysis, Statistics Born on May 16, 1821, in Okatova, Russia, Pafnuty Chebyshev is remembered for his significant contributions to NUMBER THEORY, ANALYSIS, PROBABILITY theory, and the development of inferential statistics. In 1850 he proved a conjecture posed by French mathematician JosephLouis François Bertrand (1822–1900) stating that for any value n > 3, there is at least one PRIME between n and 2n – 2. Chebyshev is also noted for founding an influential school of mathematics in St. Petersburg. Chebyshev entered Moscow University in 1837 and graduated four years later with an undergraduate degree in mathematics. Driven by an unabashed
desire to achieve international recognition, Chebyshev immersed himself in mathematical work. He earned a master’s degree in 1846 at the same institution, all the while publishing results on integration theory and methods, the convergence of TAYLOR SERIES, and the development of analysis. Chebyshev also examined the principles of probability theory and developed new insights that prove the main results of the theory in an elementary, but rigorous, way. In particular, Chebyshev was able to offer an elegant proof of SIMÉONDENIS POISSON’s weak LAW OF LARGE NUMBERS. In 1849 Chebyshev wrote a thesis on the theory of MODULAR ARITHMETIC, earning him a doctorate in mathematics from Moscow University, as well as a prize from the Russian Academy of Science in recognition of its originality and its significance. In his study of prime numbers, Chebyshev not only established Bertrand’s conjecture, but also made significant steps toward proving the famous PRIMENUMBER THEOREM. Chebyshev was elected as a full professor in mathematics at the University of St. Petersburg in 1850, and, by this time, had indeed achieved international fame. He traveled extensively throughout Europe and collaborated with many scholars on research projects on topics as diverse as mechanics, physics, mechanical inventions, and the construction of calculating machines, as well as continued work in mathematics. He was awarded many honors throughout his life, including membership to the Berlin Academy of Sciences in 1871, the Bologna Academy in 1873, the ROYAL SOCIETY of London in 1877, the Italian Royal Academy in 1880, and the Swedish Academy of Sciences in 1893. Every Russian university elected him to an honorary faculty position, and Chebyshev was even awarded honorary membership to the St. Petersburg Artillery Academy, as well as to the French Légion d’Honneur. He died on December 8, 1894, in St. Petersburg, Russia. A number of results and concepts are today named in Chebyshev’s honor. For example, in analysis, the “Chebyshev polynomials” provide a basis for the VECTOR SPACE of CONTINUOUS FUNCTIONs and have important applications to approximation theory. In statistics and probability theory, CHEBYSHEV’S THEOREM provides a “weak law of large numbers.” See also STATISTICS: INFERENTIAL.
72 Chebyshev’s theorem
Chebyshev’s theorem (Chebyshev’s inequality) This result, due to the Russian mathematician PAFNUTY LVOVICH CHEBYSHEV (1821–94), can be thought of as an extension of the 689599.7 rule for the NORMAL DISTRIBUTION to one applicable to all distributions. It states that if an arbitrary DISTRIBUTION has mean μ and standard deviation σ, then the probability that a measurement taken at random will have value differing from μ by more than k standard deviations is at most 1/k2. This shows that if the value σ is small, then all DATA values taken in an experiment are likely to be tightly clustered around the value μ. Manufacturers make use of this result. For example, suppose a company produces pipes with mean diameter 9.57 mm, with a standard deviation of 0.02 mm. If manufacturer standards will not tolerate a pipe more than four standard deviations away from the mean (0.08 mm), then Chebyshev’s theorem implies that on average about 1/16, that is 6.25 percent, of the pipes produced per day will be unusable. The LAW OF LARGE NUMBERS follows as a consequence of Chebyshev’s theorem. See also STATISTICS: DESCRIPTIVE.
chicken See PRISONER’S DILEMMA. Ch’in Chiushao (Qin Jiushao) (1202–1261) China Algebra Born in Szechwan (now Sichuan), China, mathematician and calendarmaker Ch’in Chiushao is remembered for his 1243 text Shushu jiuzhang (Mathematical treatise in nine sections), which contains, among many methods, an effective technique of iterated multiplication for evaluating polynomial equations of arbitrary degree. (In modern notation, this technique is equivalent to replacing a polynomial such as 4x3 + 7x2 – 50x + 9, for instance, with its equivalent form as a series of nested parentheses: ((4x + 7)x – 50)x + 9. In this example, only three multiplications are needed to evaluate the nested form of the polynomial compared with the six implied by the first form of the expression. In practice, this technique saves a considerable amount of time.) This approach was discovered 500 years later in the West independently by Italian mathematician Paolo Ruffini (1765–1822) and English scholar William George Horner (1786–1837). Ch’in Chiushao also extended this method to find solutions to polynomial equations.
His text is also noted for its development of MODUIn particular, Ch’in Chiushao proved the following famous result, today known as the Chinese remainder theorem: LAR ARITHMETIC.
If a set of integers mi are pairwise COPRIME, then any set of equations of the form x ≡ ai (mod mi) has a unique solution modulo the product of all the mi. For example, this result establishes that there is essentially only one integer x that leaves remainders of 1, 11, and 6, respectively, when divided by 5, 13, and 16 (namely, 726, plus or minus any multiple of 5 × 13 × 16 = 1040). After serving in the army for 14 years, Ch’in Chiushao entered government service in 1233 to eventually become provincial governor of Qiongzhou. His 1243 piece Shushu jiuzhang was his only mathematical work.
Chinese mathematics Unfortunately, very little is known about early Chinese mathematics. Before the invention of paper around 1000 C.E., the Chinese wrote on bark or bamboo, materials that were far more perishable than clay tablets or papyrus. To make matters worse, just after the imperial unification of China of around 215 B.C.E., Emperor Shi Huangti of the Ch’ih dynasty ordered that all books from earlier periods be burned, along with the burying alive of any scholars who protested. Only documents deemed “useful,” such as official records and texts on medicine, divination, and agriculture were exempt. Consequently very little survived beyond this period, although some scholars did try to reconstruct lost materials from memory. The art of mathematics was defined by ancient Chinese scholars as suan chu, the art of calculation. Often the mathematics studied was extremely practical in nature, covering a wide range of applications, including engineering, flood control, and architecture, as well astronomy and divination. Practitioners of the art were capable scientists. Records show, for example, that the Chinese had invented seismographs to measure earthquakes by the year 1000 C.E., and used compasses made with magnetic needles a century later. Evidence of mathematical activity in China can be dated back to the 14th century B.C.E.. Tortoise shells and cattle bones inscribed with tally marks indicate that the
Chinese mathematics 73 people of the ancient Shang dynasty had developed a base10 notational system utilizing place value. This establishes that the Chinese were one of the first civilizations to invent a DECIMAL REPRESENTATION system essentially equivalent to the one we use today. Like other civilizations of the time, however, the Chinese had not developed a notation for zero and so wrote, for example, the numbers 43 and 403 the same way, namely as “ ,” relying on context to distinguish the two. The most important early Chinese mathematical text is Jiuzhang suanshu (Nine chapters on the mathematical art) dating from the period of the Han dynasty (206 B.C.E. to 220 C.E.). The author of the work is unknown, but it is believed to be a summary of all mathematical knowledge possessed in China up to the third century C.E., and may well have been the result of several authors contributing to the same work. The text is a presentation of 246 problems replete with solutions and general recipes for solving problems of a similar type. The work is generally very practical in nature, with three chapters devoted to issues of land surveying and engineering, and three to problems in taxation and bureaucratic administration. But the text does describe sophisticated mathematical techniques of an abstract nature, and it offers many problems of a recreational flavor. The document thus also clearly demonstrates that scholars of the time were also interested in the study of mathematics for its own sake. Jiuzhang suanshu is clearly not written for beginners in the art of mathematics: many basic arithmetic processes are assumed known. Another text of the same period, Chou pei suan ching (Arithmetic classic of the gnomon and the circular paths of heaven), describes basic mathematical principles such as working with fractions (and establishing common denominators), methods of extracting square roots, along with basic principles and elements of geometry, and surprisingly, what appears to be a proof of what we today call PYTHAGORAS’S THEOREM. At the very least, the Chinese of this period knew the theorem for right triangles with sides of length 3, 4, and 5 as the diagram of hsuanthu—four copies of a 3–4–5 right triangle arranged in a square—appears in the text. Although this diagram in itself does not constitute a “proof” of Pythagoras’s theorem, the idea embodied in the diagram can nonetheless easily be expanded upon to establish a general proof of the result. For this reason, it is believed that the Chinese had independently established the same
famous result. This is certainly verified in the text of Jiuzhang suanshu, as many problems posed in the piece rely on the reader making use of the theorem. Many scholars from the second to the 15th centuries wrote commentaries on the work Jiuzhang suanshu and extended many of the results presented there. Perhaps the most famous of these were the commentaries of Liu Hui who, in 263 C.E., offered written proofs of the formulae for the volumes of a square pyramid and a tetrahedron presented in Jiuzhang suanshu, as well as developed a more precise value for π than presented in the text. Liu Hui later went on to write Haidao suanjing (Sea island mathematical manual), in which he solved problems related to the surveying and mapping of inaccessible objects using a refined method of “double differences” arising from pairs of similar triangles. This extended the work of proportions presented in Jiuzhang suanshu. Liu Hui also claimed that the material of Jiuzhang suanshu dates back to 1100 B.C.E., but added that the actual text was not written until 100 B.C.E. Historians today differ about how seriously to take his claim. The text Jiuzhang suanshu also contains recipes for extracting square and cube roots, and methods for solving systems of linear equations using techniques very similar to the methods of LINEAR ALGEBRA we use today. Liu Hui’s commentary gives justification for many of the rules presented. Although not formal proofs based on axioms, it is fair to describe Liu Hui’s justifications as valid informal proofs. It seems that mathematicians for the centuries that followed remained satisfied with simple informal arguments and justifications, and no formal rigor was deemed necessary. Chinese scholars also made significant contributions to the study of COMBINATORICS, NUMBER THEORY, and ALGEBRA. Mathematician CH’IN CHIUSHAO (ca. 1200 C.E.) developed inventive methods for evaluating polynomial expressions and solving polynomial equations. He also established the famous “Chinese remainder theorem” of number theory. The work of LI YE of the same period also establishes an algebra of polynomials. Although Chinese mathematicians of the time were familiar with negative numbers, they ignored negative solutions to equations, deeming them absurd. The famous text Suyuan yuchien (The precious mirror of the four elements) written by the scholar CHU SHIHCHIEH (ca. 1300 C.E.) contains a diagram of what has in the West become known as PASCAL’S TRIANGLE.
74 chisquared test This text was written 300 years before French mathematician BLAISE PASCAL was born. (Some historians believe that this work in fact dates back 200 years earlier to the writings of mathematician Jia Xian.) Scholars of this time routinely used the triangle to approximate nth roots of numbers using the equivalent of the BINOMIAL THEOREM of today. They preferred their procedural methods of extracting square roots to solve QUADRATIC equations, rather than make use of the general quadratic formula. Soon after JOHN NAPIER (1550–1617) of the West published an account of his new calculating aid, the NAPIER’S BONES, the Chinese developed an analogous system of graded bamboo rods that could be used to quickly compute long multiplications and divisions. It is not known if the Chinese invented this system independently, or whether the idea was perhaps brought to them by 17thcentury Jesuit missionaries. Along with the ABACUS developed in China 500 years earlier, the calculating rods allowed for improved arithmetic computations, especially useful for the precise computations needed in astronomy. Early scholar ZU CHONGZHI (ca. 500 C.E.) computed the volume of a sphere by a principle identical to that of BONAVENTURA CAVALIERI (1598–1647). See also MAGIC SQUARE.
chisquared test The chisquared test is a statistical test (see STATISTICS: INFERENTIAL) used to determine whether or not two characteristics of a population are independent or associated in some way. For example, imagine a social study looking for a possible correlation between the type of milk people prefer on their cereal and the number of body piercings they possess. Five hundred people were surveyed and the results obtained are displayed in a CONTINGENCY TABLE.
No One or two More than two piercings piercings piercings FatFree Milk
47
33
22
102
2% Milk
40
80
44
164
113
37
84
234
200
150
150
500
Whole Milk
Observe, in this study, that 102/500 = 0.204 of the participants are fatfree milk users. If milk choice
bears no relationship to body piercings, we would expect then about 0.204 of the 200 folk with no piercings to use fatfree milk. We observed a value of 47 (the observed frequency) but expect a value of 0.204 × 200 = 40.8 (the expected frequency). Similarly, the expected value for fatfree milk users with more than two piercings is 0.204 × 150 = 30.6 and for whole milk users with one or two piercings: (234/500) × 150 = 70.2. In this way we compute all expected frequencies, here shown in parentheses:
No One or two More than two piercings piercings piercings FatFree Milk 2% Milk Whole Milk
47 (40.8)
33 (30.6)
22 (30.6)
102
40 (65.6)
80 (49.2)
44 (49.2)
164
113 (52.9)
37 (70.2)
84 (70.2)
234
200
150
150
500
Denoting the observed frequencies by the letter o and the expected frequencies by e, we compute the chisquared statistic, χ2, as: χ2 = ∑
(o − e)2 e
where the sum is over all entries in the table. (In the 1800s it was customary to convert all differences to a positive value by use of the squaring function rather than the ABSOLUTE VALUE function. This way, techniques of calculus could be readily applied—it is straightforward to differentiate the square function, for example.) A large value for χ2 indicates that there is considerable discrepancy between observed and expected values, suggesting that the two features of the population are not independent, i.e., that there is a CORRELATION. A small χ2 value suggests that there is no correlation. Our particular example yields the value: χ2 =
(47 − 40.8)2 (33 − 30.6)2 (22 − 30.6)2 + + 40.8 30.6 30.6 +
(40 − 65.6)2 (80 − 49.2)2 (44 − 49.2)2 + + 65.6 49.2 49.2
+
(113 − 52.9)2 (37 − 70.2)2 (84 − 70.2)2 + + 52.9 70.2 70.2
= 120.1
circle 75 The chisquared distribution is a DISTRIBUTION representing the values one would expect χ2 to adopt given the assumption that the two features being studied are independent. There is one distribution for each table of given dimensions. Statistics texts usually present lists of values for this statistic. It turns out that the χ2 value in this example is extraordinarily high, suggesting that this study indicates a strong correlation between milk choice and body piercings. The chisquared test does not give any information about the nature of the correlation detected, only that it seems to exist. Further examination of the data by alternative methods may provide details of the association. See also KARL PEARSON.
equals
n(n + 1)(n + 2) ), and also provides general 6
techniques for summing arbitrary series. He also provides methods for solving equations via a process of successive approximations. Many historians claim that Chu ShihChieh’s impressive work represents the peak of ancient Chinese mathematics, noting that relatively little progress was made for a long time after the publication of this piece. Four years before the release of Suyuan yuchien, Chu ShihChieh wrote a mathematical text intended to help beginners in the subject. Extremely little is known of his personal life.
circle The set of all points in a plane a fixed distance chord If A and B are two points on a continuous curve, then a straightline segment connecting A to B is called a chord to the curve. This is not to be confused with the ARC of the curve connecting A to B. The CIRCLE THEOREMS show that any chord of a circle is bisected by a radius that is perpendicular to it. BERTRAND’S PARADOX shows that the act of selecting chords of a circle at random can lead to philosophical difficulties. See also CIRCLE.
Chu ShihChieh (Zhu Shijie) (ca. 1270–1330) Chinese Algebra Often regarded as one of China’s greatest mathematicians, Chu ShihChieh is remembered for his influential 1303 text Suyuan yuchien (Precious mirror of the four elements). It contains a diagram of PASCAL’S TRIANGLE, as it has become known in the West, representing one of the earliest appearances of the figure in the history of mathematics. In the work, Chu ShihChieh uses the triangle to describe a general method for extracting roots to equations. He also presents a system of notation for polynomials in four unknowns, which he calls the four elements—namely, the celestial, the earthly, the material, and the human— and provides effective techniques for manipulating them to solve problems. His section on FINITE DIFFERENCES gives formulae for the sums of the first n terms of each diagonal of Pascal’s triangle (for instance, that 1 + 2 + 3 + … + n equals
n(n + 1) n(n + 1) , and that 1 + 3 + 6 + 10 + … + 2 2
r from a given point O forms a closed curve in the plane called a circle. The length r is called the radius of the circle, and the point O its center. If the center point has CARTESIAN COORDINATES 0 = (a,b), then the DISTANCE FORMULA shows that any point (x,y) on the circle satisfies the equation: (x – a)2 + (y – b)2 = r2 If the point (x,y) on this circle makes an angle θ with a horizontal line through the center of the circle, then we have: x = a + rcosθ y = b + rsinθ These are the PARAMETRIC EQUATIONS of a circle of radius r and center (a,b). The DIAMETER of a circle is the maximal distance between two points on the circle. It equals twice the radius of the circle. A circle is a figure of CONSTANT WIDTH. The length of the curve closed to form a circle is called the circumference of the circle. Scholars since the time of antiquity have observed that the ratio of the circumference of a circle to its diameter is the same for all circles. This constant value is called PI, denoted π. We have: π = 3.14159265… That all circles yield the same value for π is not immediately obvious. This is a property of EUCLIDEAN GEOMETRY of the plane, and a careful study of SCALE explains why it must be true. (The value of π varies
76 circle theorems from 2 to 3.141592… for different circles drawn on the surface of a SPHERE, for instance, since the diameter of a circle must be measured as the length of a curved line on the surface.) If C is the circumference of a planar circle and D = 2r is its diameter, then, by definition, π = C/D. This yields a formula for the circumference of a circle: C = 2πr A study of AREA also shows that the area A of a circle is given by: A = πr2 It is not immediate that the value π should also appear in this formula. A study of EQUIDISTANCE shows that it is always possible to draw a circle through any three given points in a plane (as long as the points do not lie in a straight line), or, equivalently, it is always possible to draw a CIRCUMCIRCLE for any given TRIANGLE. APOLLONIUS OF PERGA (ca. 262–190 B.C.E.) developed general methods for constructing a circle TANGENT to any three objects in the plane, be they points, lines, or other circles. Any line connecting two points on a circle is called a CHORD of the circle. It divides the circle into two regions, each called a segment. A chord of maximal length passes through the center of a circle and is also called a diameter of the circle. (Thus the word diameter is used interchangeably for such a line segment and for the numerical value of the length of this line segment.) A radius of a circle is any line segment connecting the center of the circle to a point on the circle. Two different radii determine a wedgeshaped region within the circle called a sector. If the angle between the two radii is θ, given in RADIAN MEASURE, then the area of this θ 1 segment is ⋅ π r 2 = θ r 2 . The length of the ARC of 2π 2 θ ⋅ 2π r = rθ . the circle between these two radii is 2π Any two points P = (a1,b1) and Q = (a2,b2) in the plane determine a circle with the line segment connecting P to Q as diameter. The equation of this circle is given by:
outside. (This seemingly obvious assertion is not true for circles drawn on a TORUS, for example.) Two intersecting circles divide the plane into four regions; three intersecting circles can be arranged to divide the plane into eight regions; and four mutually intersecting circles can divide the plane into 14 regions. In general, the maximal number of regions into which n intersecting circles divide the plane is given by the formula: n2 – n + 2. The region formed at the intersection of two intersecting circles of the same radius is called a lens. There are a number of CIRCLE THEOREMS describing the geometric properties of circles. A circle is a CONIC SECTION. It can be regarded as an ELLIPSE for which the two foci coincide. If one permits the use of COMPLEX NUMBERS, then any two circles in the plane can be said to intersect. For example, the two circles each of radius one centered about the points (0,0) and (4,0), respectively, given by the equations x2 + y2 = 1 and (x – 4)2 + y2 = 1 intersect – – at the points (2, i√ 3) and (2,–i√ 3). The threedimensional analog of a circle is a SPHERE: the locus of all points equidistant from a fixed point O in threedimensional space. In onedimension, the analog of a circle is any pair of points on a number line. (Two points on a number line are equidistant from their MIDPOINT.) The midpoint theorem asserts that all midpoints of line segments connecting a fixed point P in the plane to points on a circle C form a circle of half the radius of C. See also APOLLONIUS’S CIRCLE; BRAHMAGUPTA’S FORMULA; CYCLIC POLYGON; FAREY SEQUENCE; NINEPOINT CIRCLE; UNIT CIRCLE; VENN DIAGRAM.
circle theorems A CIRCLE is defined as the set of points in a plane that lie a fixed distance r, called the radius, from some fixed point O, called the center. This simple definition has a number of significant geometric consequences: 1. Tangent Theorems
(x – a1)(x – a2) + (y – b1)(y – b2) = 0
The point of contact of a TANGENT line with a circle is the point on that line closest to the center point O. As a consequence of PYTHAGORAS’S THEOREM, the line connecting the point of contact to O is at an angle 90° to the tangent line. This proves:
The JORDAN CURVE THEOREM establishes that a circle divides the plane into two regions: an inside and an
The tangent to a circle is PERPENDICULAR to the radius at the point of contact.
circle theorems 77
Tangent theorems
This is illustrated in the diagram above, left. The above diagram to the right shows that two tangents through a common point P outside the circle produce line segments PA and PB of equal length. This follows from the fact that the two triangles produced are both right triangles of the same height with a shared hypotenuse, and are hence congruent. Thus: If PA and PB are tangents to a circle at points A and B, respectively, then PA and PB have the same length. 2. InscribedAngle Theorems In the diagram below, left, angles α (the peripheral angle) and β (the central angle) are subtended by the same ARC. Thus we have: For angles subtended by the same arc, the central angle is always twice that of the peripheral angle.
Proving the centralangle/peripheralangle theorem
isosceles triangles. Following the lefthand side of the next diagram, and noting that the interior angles of a triangle sum to 180°, we thus have x + y = α and (180 – 2x) + (180 – 2y) + β = 360, from which it follows that β = 2α. A modification of this argument shows that the result is still true even if the peripheral angle is located as shown in the righthand side of the diagram, or if the arc under consideration is more than half the PERIMETER of the circle. The next three results follow (see diagram below, right): i. All angles inscribed in a circle subtended by the same arc are equal, ii. All angles inscribed by a diameter are right angles. (This is known as the theorem of Thales.)
This is proved by drawing a radius from the center O to the point at which angle α lies to create two
Central and peripheral angles
Consequences of the centralangle/peripheralangle theorem
78 circumcircle PERGA, who showed how to construct a circle tangent to any three points, lines, or circles in the plane.) The center of the circumcircle of a triangle is called its circumcenter. Clearly, it is a point EQUIDISTANT from the three vertices, and so it must lie on each line of points equidistant from any two vertices. The circumcenter of a triangle can thus be found by drawing the perpendicular bisectors of the sides of the triangle and locating where these three lines meet. By the LAW OF SINES, the radius of the circumcircle a of a triangle is given by r = , where a is 2sin(A) a sidelength of the triangle and A is the angle opposite that chosen side. If the remaining two sides of the triangle have lengths b and c, then the area of the triangle 1 can be written: area = bc sin(A) . By HERON’S 2 FORMULA, this area can also be computed via a+b+c . Equating area = s(s − a)(s − b)(s − c) where s = 2 these two equations, solving for sin(A), and substituting into the formula above produces a formula for the radius of the circumcircle of a triangle solely in terms of its side lengths: NIUS OF
Chord theorems
iii. Opposite angles in a cyclic sum to 180°.
QUADRILATERAL
3. TwoChord Theorem In the diagram above, we have: ab = cd. This is proved by connecting the endpoints of the chords to create two triangles. The inscribedangle theorems show that these two triangles are similar by the AAA rule. Consequently, a/c = d/b. 4. RadiusChord Theorem In the righthand side of the diagram, we see that: The radius of a circle bisects a chord (that is, we have a = b) if, and only if, the radius is perpendicular to the chord. This is proved by drawing two radii to the endpoints of the chord to produce a large isosceles triangle. If the central radius is at 90° to the chord, this produces two congruent right triangles, and so a = b. Conversely, if we are told a = b, then the LAW OF COSINES, applied to each base angle of the isosceles triangle, shows that the central radius intercepts the chord at an angle of 90°. See also AAA/AAS/ASA/SAS/SSS; CONGRUENT FIGURES; CYCLIC POLYGON; PTOLEMY’S THEOREM; SECANT; SIMILAR FIGURES.
circumcircle A circle that passes through all three vertices of the triangle is called a circumcircle for the triangle. That three distinct points in the plane determine a unique circle was first proved by EUCLID in his treatise, THE ELEMENTS, Book III. Euclid also presented a general method for actually constructing the circumcircle of a triangle. (This result was later generalized by APOLLO
r=
abc 4 s(s − a)(s − b)(s − c)
The circumcircle of a regular POLYGON is that circle that passes through all the vertices of the polygon. The radius of the circumcircle of a square, for example, – of sidelength x is r = x/√ 2. See also CONCURRENT; LONG RADIUS; TRIANGLE.
circumscribe/inscribe If A and B are two geometric figures, with A inside B, drawn so that the two figures have points in common but do not have edges that cross, then we say that figure A is inscribed in B, or, alternatively, that figure B is circumscribed about A. For example, a polygon lying inside a circle with all its vertices on that circle is said to be inscribed in the circle, and a circle inside a polygon touching each side of the polygon is inscribed in the polygon. A circle that passes through all three vertices of a triangle circumscribes that triangle, and the smallest square that surrounds a circle circumscribes that circle. See also CIRCUMCIRCLE; INCIRCLE.
clock math See MODULAR ARITHMETIC.
collinear 79
closure property A
on a set S is said to be closed if the combination of two elements in that set yields another member of that set. For example, the set of positive whole numbers is closed under addition, since the sum of any two positive integers is itself a positive integer. This set is also closed under multiplication (the product of two positive whole numbers is a positive integer), but not subtraction: if n and m are positive whole numbers, then n – m could be zero or negative and thus no longer in the set of positive whole numbers. (For instance, 3 – 7 = –4.) For a more unusual example, consider the set S of all whole numbers that can be expressed as the sum of two SQUARE NUMBERS. We have S = {0, 1, 2, 4, 5, 8, 9, 10, 13, 16, 17, 20, 25,…}. (For instance, 0 = 02 + 02, 5 = 12 + 22, 20 = 22 + 42, and 25 = 02 + 52 = 32 + 42.) Surprisingly, this set is closed under multiplication. For example, both 5 and 8 belong to S, and so does 5 × 8 = 40. (We have 40 = 22 + 62.) Also, 10 and 13 belong to S, and so too does 130. (We have 130 = 32 + 112.) This general observation follows from the algebraic identity that if N = a2 + b2 and M = c2 + d2; then N × M = (ac + bd)2 + (ad – bc)2. This set S is also closed under exponentiation: if N and M are each a sum of two squares, then so is NM. For example, 513 = (12,625)2 + (31,250)2. See also SQUARE. BINARY OPERATION
coefficient A numerical or constant multiplier of the variables in a term of an algebraic expression is called the coefficient of that term. For example, consider the equation 5x3 – 2x + 7 = 0, where x is the variable, the coefficient of x3 is 5, the coefficient of x is –2, and the coefficient of x2 is zero. In the equation 3 cos y – 4xy2 = 7, the coefficients of cos y and xy2 are 3 and –4, respectively. Sometimes the value of a coefficient is not known and a symbol is used in its stead. For instance, in the expression ax2 + bx + c with x the variable, the numbers a and b are coefficients (and c is a constant term). Although the values of a, b, and c are not specified, it is understood that their values do not change even as the value of x varies. In a more general context, the term coefficient is used for any number that serves as a measure of some property or characteristic of a set of data or a physical property. For instance, a CORRELATION COEFFICIENT in
gives a measure of the extent to which two data sets are interdependent, while the heat coefficient in physics gives a measure as to how well a material conducts heat. See also BINOMIAL COEFFICIENT; COMBINATORIAL COEFFICIENT; CONSTANT; LEADING COEFFICIENT; POLYNOMIAL. STATISTICS
Collatz’s conjecture (“3n + 1” mapping problem) Consider the following process: Select a positive integer. If it is odd, triple it and add one; otherwise, divide the number by two. Now perform the same operation again on the result. Repeat this process indefinitely to produce a sequence of numbers. The number 7, for example, yields the sequence: 7, 22, 11, 34, 17, 52, 26, 13, 40, 20, 10, 5, 16, 8, 4, 2, 1, 4, 2, 1, 4, 2, 1,… Notice that this sequence finally falls into a 421 cycle. In 1937 German mathematician Lothar Collatz conjectured that, no matter the starting integer selected, all sequences lead to the same 421 cycle. Collatz was unable to prove this claim, but he was also unable to find an example of a starting number that does not behave this way. To this day, no one knows whether or not Collatz’s conjecture is true. All integers up to 2.702 × 1016 have been checked. As a first step toward understanding this problem, mathematicians have proved that 421 is the only cycle of reasonable size that could possibly appear; it has been established that any other cycle that might appear would be at least 275,000 numbers long.
collinear Any number of points are said to be collinear if they all lie on the same straight line. Two points are always collinear. Three points in a plane A = (a1,a2), B = (b1,b2), and C = (c1,c2) are collinear only if the lines connecting points A and B and the points connecting A and C have the same SLOPE. This means that the following relationship must hold: b2 − a2 c2 − a2 = b1 − a1 c1 − a1
80 combination Three points A, B, and C in threedimensional space are collinear if the triangle they form has zero area. Equivalently, the three points are collinear if the → → angle between the VECTORS AB and AC is zero, and → → consequently the CROSS PRODUCT AB × AC equals the zero vector. The collinearity of points in a plane is a topic of interest to geometers. In the mid1700s, LEONHARD EULER discovered that several interesting points constructed from triangles are collinear, yielding his famous EULER LINE. In 1893 British mathematician James Sylvester (1814–97) posed the question of whether it is possible to arrange three or more points in a plane, not all on a line, so that any line connecting two of the points from the collection passes through a third point as well. Forty years later Tibor Gallai (1912–92) proved that there is no such arrangement. Two or more distinct PLANEs are said to be collinear if they intersect in a common straight line. In this case, the vectors normal to each plane all lie in a plane perpendicular to the common line. Thus one can determine whether or not a collection of planes is collinear by noting whether or not the cross products of pairs of normal vectors are all parallel. See also GRADIENT; NORMAL TO A PLANE.
number of ways to arrange n distinct objects in a row. There are, of course, n! different ways to do this. (See FACTORIAL.) Alternatively, we can imagine selecting which k objects are to be arranged in the n first k positions along the row (there are k ways to do this), ordering those k items (there are k! different ways to do this), and then arranging the remaining n – k objects for the latter part of the row (there are (n – k)! different ways to accomplish this). This n yields k k!(n − k)! different ways to arrange n objects in a row. Since this quantity must equal n!, we have n n! the formula k = k!(n − k)! for the combinatorial coefficient. It is appropriate to define 0! as equal to one. In this way, the formula just established holds even for k = n. (There is just one way to select n objects from n n! a collection of n items, and so n = n!0! should equal n n! one.) It then follows that 0 = 0! n! = 1 . (There is just one way to select no objects.) Mathematicians set
combination (selection, unordered arrangement) Any set of items selected from a given set of items without regard to their order is called a combination. Repetition of choices is not permitted. For example, there are six distinct combinations of two letters selected from the sequence A,B,C,D, namely: AB, AC, AD, BC, BD, and CD. (The selection BA, for example, is deemed the same as AB, and the choice AA is not permitted.) The number of combinations of k items selected n from a set of n distinct objects is denoted k . The 4 number 2 , for instance, equals six. The quantity n is called a combinatorial coefficient and is read as k “n choose k.” Given their appearance in the BINOMIAL THEOREM, these numbers are also called binomial coefficients. n One develops a formula for k by counting the
n to be zero if k is negative or greater than n. k The combinatorial coefficients appear as the entries of PASCAL’S TRIANGLE. They also satisfy a number of identities. We list just four, which we shall phrase in terms of the process of selecting k students to be in a committee from a class of n students. n n 1. = k n − k (Selecting k students to be in a committee is the same as selecting n – k students not to be in the committee.) n n − 1 n − 1 2. = + k k − 1 k (Any committee formed either includes, or excludes, a particular student John, say. If John is to be on the committee, then one must select k – 1 more students
combination 81 from the remaining n – 1 students. If John is not to be on the committee, then one must select k students from the pool of n – 1 students that excludes John.) n
3.
n
n
n
n
n
∑ k = 0 + 1 + 2 + L + n = 2n
k= 0
n n n (There are 0 + 1 + L + n possible committees of any size. But this number can also be computed by deciding, student by student, whether or not to put that student in the committee. As there are two possibilities for each student, in or out, there are 2n possible committees. These counts must be the same.) n
4.
n
n
n
n
n
∑ k k = 1 + 2 2 + 3 3 + L + n n = n2n−1
k= 0
(Suppose, in the committee, one student is to be selected as chair. In a committee of size k there are k n n possible choices for chair. Thus ∑ k k counts the k= 0
total number of committees possible, of any size, with one student selected as chair. But this quantity can also be computed by selecting some student to be chair first—there are n choices for this—and then deciding, student by student, among the remaining n – 1 students whether that student should be on the committee. This yields n2n–1 possibilities.) Property 1 explains why Pascal’s triangle is symmetric. Property 2 shows that each entry in Pascal’s triangle is the sum of the two entries above it, and property 3 shows that the sum of all the entries in any row of Pascal’s triangle is a power of two. n In 1778 LEONHARD EULER used the notation k for the combinatorial coefficients, which, three years n later, he modified to . In the 19th century, mathek maticians started following Euler’s original notation, dropping the VINCULUM for the purposes of easing typesetting. Many textbooks today use the notation n nCk, or C k, or even C(n,k), for the combinatorial n coefficient k .
Generalized Coefficients
n The generalized combinatorial coefficient k k K k , 1 2 r where k1,k2,…,kr are nonnegative integers summing to n, is defined to be the number of ways one can select, from n items, k1 objects to go into one container, k2 objects to go into a second container, and so forth, up to kr objects to go into an rth container. (Notice that n is the ordinary combinatorial coefficient.) k n − k Mimicking the argument presented above, note that one can arrange n items in a row by first selecting which k1 items are to go into the first part of the row and ordering them, which k2 items are to go in the next portion of the row and ordering them, and so on. n This shows that n! = k1! k2 !K kr ! , yielding k k k K 1 2 r the formula: n n! = k ! k !K k ! k1 k2 K kr 1 2 r Generalized combinatorial coefficients show, for 7 7! example, that there are 1 1 3 2 = 1!1!3! 2! = 420 ways to rearrange the letters CHEESES: Of the seven slots for letters, one must choose which slot is assigned for the letter C, which one for the letter H, which three for the letter E, and which two for letter S. The generalized combinatorial coefficients also appear in generalizations to the BINOMIAL THEOREM. For example, we have the trinomial theorem: n k k k (x + y + z)n = ∑ x iy 2z 3 k1 k2 k3 where the sum is taken over all triples k1,k2,k3 that sum to n. The proof is analogous to that of the ordinary binomial theorem.
MultiChoosing 4 The quantity 2 , read as “n multichoose k,” counts the number of ways to select k objects from a collection
82 combinatorial coefficient of n items, where order is not important, but repetition is allowed. For example, there are 10 ways to multichoose two objects from the set A, B, C, and D, namely: AB, AC, AD, BC, BD, CD, and AA, BB, CC, 4 and DD. Thus = 10 . One can show that a multi 2 choose coefficient equals an ordinary combinatorial coefficient: n n + k − 1 k = k
combinatorial coefficient See COMBINATION. combinatorics (combinatorial analysis) The branch of mathematics concerned with the theory and practices of counting elements of sets and the construction of specified arrangements of objects, along with the study of COMBINATIONs and PERMUTATIONs, is called combinatorics. GRAPH THEORY is also regarded as an aspect of combinatorics. The technique of “double counting,” that is, counting the same set of objects in two different ways, is a common practice in combinatorics used to yield interesting results. For example, counting the dots in an n × n square array along diagonals as opposed to across the rows gives the surprising formula: 1 + 2 + 3 + … + (n – 1) + n + (n – 1) + … + 3 + 2 + 1 = n2 Counting the number of subsets of a set of n elements, either by summing the number of subsets containing, in turn, 0, 1, 2, up to n elements, or by noting that each subset is decided by making n choices between two options—whether or not each element in turn is to be in the subset—yields the formula: n n n n n + + + ... + = 2 × 2 × L × 2 = 2 0 1 2 n EULER’S THEOREM can be considered a result in combinatorial geometry. See also DISCRETE; FIGURATE NUMBERS.
commensurable Two quantities having a common measure, meaning that they can be measured in terms of
whole numbers of a common unit, are said to be commensurable. For example, the quantities one month and one week are commensurable because they can both be measured in terms of a whole number of days. In GEOMETRY, two line segments are said to be commensurable if there is another segment whose measure goes evenly, without remainder, into the measures of each segment. For instance, segments of lengths 20 and 12 in. are commensurable for they can each be evenly divided into lengths of 1 (or 2 or 4) in. In general, two segments of lengths a and b units are commensurable if the ratio a/b – is a RATIONAL NUMBER. As √ 2 is irrational, segments of – length 1 and √ 2 (respectively, the sidelength and the diagonal of a unit square) are incommensurable. A study of the EUCLIDEAN ALGORITHM shows that if given two commensurable line segments of lengths a and b, say, then repeatedly subtracting the shorter length from the longer to produce a new pair of lengths eventually produces two line segments equal in length. This final shared measure is the largest length that divides evenly into the two original segments. (If a and b are wholenumber measurements, then the length of the final measure is the GREATEST COMMON DIVISOR of a and b.) If, on the other hand, one can demonstrate that the process of repeatedly erasing the shorter line segment from the longer will continue indefinitely without ever producing two line segments equal in length, then the original two segments cannot be commensurable. Around 425 B.C.E. Greek mathematician THEODORUS OF CYRENE used precisely this observa– tion to prove the irrationality of √ 2. In NUMBER THEORY, two real numbers a and b are said to be commensurable if their ratio is rational. For – – instance, the numbers √48 and √ 3/2 are commensurable. No one to this day knows whether or not π and e are commensurable. The numbers log5(3) and log5(7) are incommensurable. (If log 5 7 = p for some whole log 5 3 q numbers p and q, then 7q = 3p, which is absurd since every power of 7 is 1 more than a multiple of 3.)
common denominator Two or more fractions are said to have a common denominator if the denominator of each fraction is the same. For example, the 5 3 fractions 12 and 12 have a common denominator of 12. It is a straightforward matter to add and subtract
commutative property 83 fractions with a common denominator. For instance, 5 3 8 5 3 2 12 + 12 = 12 and 12 – 12 = 12 . It is possible to rewrite the terms of an arbitrary collection of fractions so that they all share a common 1 2 5 denominator. For instance, the rewriting 3 and 5 as 15 6 and 15, respectively, shows that the two fractions have a common denominator 15. In fact, any COMMON MULTIPLE of 3 and 5 serves as a common denominator 1 2 1 10 2 12 of 3 and 5 . For instance, we have 3 = 30 and 5 = 30 , 1 15 2 18 and 3 = 45 and 5 = 45 . The LEAST COMMON MULTIPLE of the denominators of a collection of fractions is called the least common denominator of the fractions. For example, the least 1 2 common denominator of 3 and 5 is 15, and the least 3 3 5 common denominator of 8 , 4 , and 6 is 24. One adds and subtracts arbitrary fractions by rewriting those fractions in terms of a common denominator.
common factor (common divisor) A number that divides two or more integers exactly is called a common factor of those integers. For example, the numbers 20, 30 and 50 have 2 as a common factor, as well as 1, 5, and 10 as common factors. It is always the case that the largest common factor a set of integers possesses is a multiple of any other common factor. In our example, 10 is a multiple of each of 1, 2, and 5. The value 1 is always a common factor of any set of integers. The FUNDAMENTAL THEOREM OF ARITHMETIC shows that any number can be uniquely expressed as a product of PRIME factors. Any common factor of two or more integers is a product of primes common to all those integers, and the largest common factor is the product of all the primes in common, with repetition permissible. (This explains why the largest common factor is a multiple of any other common factor.) If the integers have no primes in common, then their largest common factor is one. See also GREATEST COMMON DIVISOR; RELATIVELY PRIME.
common multiple A number that is a multiple of two or more other numbers is called a common multiple of
those numbers. For example, 60 is a common multiple of 5, 6, and 10. The lowest number that is a common multiple of a given set of numbers is called their LEAST COMMON MULTIPLE. In our example, 30 is the least common multiple of 5, 6, and 10. Every common multiple is a multiple of the least common multiple. The FUNDAMENTAL THEOREM OF ARITHMETIC shows that any number can be uniquely expressed as a product of PRIME factors. Any common multiple of two or more integers is the product of, at the very least, all the primes that appear in the factorizations of the given integers, with the necessary repetitions, with perhaps additional factors. With no additional factors present, one obtains the least common multiple.
commutative property A
BINARY OPERATION is said to be commutative if it is independent of the order of the terms to which it is applied. More precisely, an operation * is commutative if:
a*b = b*a for all values of a and b. For example, in ordinary arithmetic, the operations of addition and multiplication are commutative, but subtraction and division are not. For instance, 2 + 3 and 3 + 2 are equal in value, but 2 – 3 and 3 – 2 are not. If an operation is both commutative and ASSOCIATIVE, then all products of the same set of elements are equal. For example, the quantity a*(b*c) equals (a*c)*b and b*(c*a). In this case, one is permitted to simply write a*b*c, with terms in any order, without concern for confusion. In SET THEORY, the union and intersection of two sets are commutative operations. In VECTOR analysis, the addition and DOT PRODUCT of two vectors are commutative operations, but the CROSS PRODUCT operation is not. The multiplication of one MATRIX with another is not, in general, commutative. Geometric operations generally are not commutative. For example, a reflection followed by a rotation does not usually produce the same result as performing the rotation first and then applying the reflection. One could also say that the operations of putting on one’s shoes and one’s socks are not commutative. A GROUP is called commutative, or Abelian, if the operation of the group is commutative. See also NIELS HENRIK ABEL; DISTRIBUTIVE PROPERTY; RING.
84 comparison test
comparison test See CONVERGENT SERIES. completing the square A QUADRATIC quantity of the form x2 + 2bx can be regarded, geometrically, as the formula for the area of an incomplete square. Adding the term b2 completes the picture of an (x + b) × (x + b) square. We have:
This formula is known as the quadratic formula. To see why it is correct, divide the given equation through by a and add a term to complete the square of resultant b portion x 2 + x . We have: a x2 +
x2 + 2bx + b2 = (x + b)2
2
x2 +
This process of completing the square provides a useful technique for solving quadratic equations. For example, consider the equation x2 + 6x + 5 = 21. Completing the square of the portion x2 + 6x requires the addition of the constant term 9. We can achieve this by adding 4 to both sides of the equation. We obtain:
The solutions of a quadratic equation ax2 + bx + c = 0, with a ≠ 0, are given by: x=
Completing the square
−b ± b2 − 4ac 2a
b b c b x+ + = a a 2a 2a
2
2
b c b2 x + + = 2 2a a 4a 2
b b2 c b2 − 4ac x + = 2 − = 2a a 4a 4a2
x2 + 6x + 5 + 4 = 21 + 4 x2 + 6x + 9 = 25 (x + 3)2 = 25 from which it follows that x + 3 equals either 5 or –5, that is, that x equals 2 or –8. The process of completing the square generates a general formula for solving all quadratic equations. We have:
b c x+ =0 a a
x+
b b2 − 4ac =± 2a 2a
x=
−b ± b2 − 4ac 2a
For example, to solve x2 + 6x + 5 = 21, subtract 21 from both sides of the equation to obtain x2 + 6x – 16 = 0. By the quadratic formula: x=
−6 ± 36 − 4 ⋅ (−16) −6 ± 100 −6 ± 10 = = = 2 or − 8 2 2 2
The quadratic formula shows that the two roots r1 and r2 of a quadratic equation ax2 + bx + c = 0 (or the single double root if the DISCRIMINANT b2 –4ac c b equals zero) satisfy r1 + r2 = − and r1r2 = . It also a a shows that every quadratic equation can be solved if one is willing to permit COMPLEX NUMBERS as solutions. (One may be required to take the square root of a negative quantity.) There do exist analogous formulae for solving CUBIC EQUATIONs ax3 + bx2 + cx + d = 0 and QUARTIC EQUATIONs ax4 + bx3 + cx2 + dx + e = 0 in terms of the coefficients that appear in the equations. Algebraist NIELS HENRIK ABEL (1802–29) showed that there can be no analogous formulae for solving fifth and higherdegree equations. See also FACTORIZATION; FUNDAMENTAL THEOREM OF ARITHMETIC; HISTORY OF EQUATIONS AND ALGEBRA (essay); SOLUTION BY RADICALS.
complex numbers 85
complex numbers There is no real number x with 2
the property that x = –1. By introducing an “imaginary” number i as a solution to this equation we obtain a whole host of new numbers of the form a + ib with a and b real numbers. These new numbers form the system of complex numbers. It is customary to use the variable z to denote an arbitrary complex number: z = a + ib. If b = 0, then z is a real number. Thus the set of complex numbers includes the set of real numbers. If a = 0 so that z is of the form z = ib, then z is said to be purely imaginary. In general, if z = a + ib, then a is called the real part of z and b the imaginary part of z. We write: Re(z) = a and Im(z) = b. The number i is usually regarded as the square – root of negative one: i = √ –1. (One must be careful as there are, in fact, two square roots of this quantity, namely i and –i.) The roots of other negative quanti– – – – – ties follow: √ –9 = √ –1 . √ 9 = ±3i and √ –30 = ± i √30, for instance. The set of all complex numbers is denoted C. Arithmetic can be performed on the complex numbers by following the usual rules of algebra and replacing i2 by –1 whenever it appears. For example, we have: (2 + 3i) + (4 – i) = 6 + 2i (2 + 3i) – (4 – i) = –2 + 4i (2 + 3i)(4 – i) = 8 + 12i – 2i – 3i2 = 8 + 10i + 3 = 11 + 10i The QUOTIENT of two complex numbers can be computed by the process of RATIONALIZING THE DENOMINATOR: 2 + 3i 2 + 3i 4 + i (2 + 3i)(4 + i) 5 + 14i 14 = ⋅ = = = 1+ i 4−i 4 − i 4 + i (4 − i)(4 + i) 5 4 − i2 One can show that with these arithmetic properties, the set of complex numbers constitutes a mathematical FIELD. In the early 18th century, French mathematician ABRAHAM DE MOIVRE noticed a striking similarity between complex multiplication and the ADDITION formulae of the sine and cosine functions from TRIGONOMETRY. Given that: (a + ib) · (c + id) = (ac – bd) + i(ad + bc) and:
cos(x + y) = cos(x)cos(y) – sin(x)sin(y) sin(x + y) = sin(x)cos(y) + cos(x)sin(y) we obtain the compact formula: (cos(x) + isin(x))(cos(y) + isin(y)) = cos(x + y) + isin(x + y) This observation formed the basis for the famous formula that now bears his name: (cos(x) + isin(x))n = cos(nx) + isin(nx) A few years later LEONHARD EULER (1707–83) took matters one step further and used the techniques of calculus to establish his extraordinary formula: eix = cos(x) + isin(x) from which DE MOIVRE’S FORMULA follows easily. (Use (eix)n = e i (nx).) Moreover, this result shows that de Moivre’s formula also holds for noninteger values of n. That the cosine and sine functions appear as the real and imaginary parts of a simple EXPONENTIAL FUNCTION shows that all of trigonometry can be greatly simplified by rephrasing matters in terms of complex numbers. Although some might argue that complex numbers do not exist in the real world, the mathematics of the complex number system has proved to be very powerful and has offered deep insights into the workings of the physical world. Engineers and physicists phrase a great deal of their work in terms of complex number theory. (Engineers prefer to use the symbol j instead of i.) It is a surprise to learn that the introduction of a single new number i as a solution to the equation x2 + 1 = 0 provides all that is needed to completely solve any POLYNOMIAL equation anxn + an–1xn–1 + … + a1x + a0 = 0. The FUNDAMENTAL THEOREM OF ALGEBRA asserts that a polynomial equation of degree n has precisely n roots (counted with multiplicity) in the complex number system. It is possible to raise a real number to a complex power to obtain a real result. For example, by EULER’S FORMULA, we have: eiπ = cos(π) + isin(π) = –1 + i · 0 = –1
86 complex numbers
Multiplying by –1 and multiplying by i
π 2
π π = cos( ) + i sin( ) = i , we see that 2 2 it is also possible to raise a complex number to a complex power to obtain a real result: Also, noting that e
i
i
π iπ − ii = e 2 = e 2 π i( + 2kπ )
(Technically, since i can also be expressed as e 2 for any whole number k, there are infinitely different (real) values for the quantity ii.) The Geometry of Complex Numbers Multiplying the entries of the real NUMBER LINE by –1 has the effect of rotating the line about the point 0 through an angle of 180°. It is natural to ask: multiplication by which number creates a 90° rotation about the point zero? Call the desired number x. Multiplication by x twice, that is, multiplication by x × x = x2, would have the effect of performing two 90° rotations, namely, a
rotation by 180°. Thus multiplication by x2 has the same effect as multiplication by –1, and the desired number x must therefore satisfy the equation x2 = –1. This shows that x = i and that it is natural to interpret purely imaginary numbers of the form ib as members of a vertical number line. This model provides a natural correspondence between complex numbers and points in the plane. The horizontal axis is called the real axis, the vertical axis the imaginary axis, and an arbitrary complex number z = a + ib appears as the point with coordinates (a,b) on the plane. This representation of complex numbers as points on a plane is called an Argand diagram in honor of JEAN ROBERT ARGAND (1768–1822) who, along with surveyor Casper Wessel (1745–1818), first conceived of depicting complex numbers in this way. The plane of all complex numbers is also called the complex plane. The angle θ that a complex number z = a + ib makes with the positive real axis is called the argument of the complex number, and the distance r of the complex number from the origin is called its modulus (or, simply, absolute value), denoted z. PYTHAGORAS’S THEOREM and the DISTANCE FORMULA show that: z = √a2 + b2 Using trigonometry we see that the values a and b can be expressed in terms of r and θ in a manner akin to POLAR COORDINATES. We have: a = r cos θ b = r sin θ With the aid of Euler’s formula, this shows that any complex number z can be expressed in polar form: z = r cos θ + ir sin θ = reiθ From this it follows, for instance, that the product of two complex numbers z1 = r1eiθ1 and z2 = r2eiθ2 has modulus r1r2 and argument θ1 + θ2: z1z2 = r1eiθ1r2eiθ2 = r1r2ei(θ1 + θ2) It is convenient to define the conjugate of a complex number z = a + ib to be the number –z = a – ib. We have:
Polar coordinates of a complex number
z · –z = a2 + b2 = z2
composition 87 and, in polar form, if z = reiθ, then –z = re –iθ. Taking the conjugate of a complex number has the geometric effect of reflecting that number across the real axis. History of Complex Numbers The first European to make serious use of the square root of negative quantities was GIROLAMO CARDANO (1501–76) of Italy in the development of his solutions to CUBIC EQUATIONs. He noted that quantities that arose in his work, such as an expression of the 3 3 form 2 + −121 + 2 − −121 for instance, could be manipulated algebraically to yield a real solution to
equations. (We have 3 2 + −121 + 3 2 − −121 = 4.) Nonetheless, he deemed such a manipulation only as a convenient artifice with no significant practical meaning. French philosopher RENÉ DESCARTES (1596–1650) agreed and coined the term imaginary for roots of negative quantities. During the 18th century, mathematicians continued to work with imaginary roots, despite general skepticism as to their meaning. Euler introduced the symbol i – for √ –1, and Argand and Wessel introduced their geometric model for complex numbers, which was later popularized by CARL FRIEDRICH GAUSS (1777–1855). His proof of the fundamental theorem of algebra convinced mathematicians of the importance and validity of the complex number system. Irish mathematician SIR WILLIAM ROWAN HAMILTON (1805–65) is credited as taking the final step to demystify the meaning of the complexnumber system. He extended the notion of the complex numbers as arising from 90° rotations by showing that any rotation in threedimensional space can naturally and easily be represented in terms of complex numbers. He also noted that the complex numbers are nothing more than ordered pairs of numbers together with a means for adding and multiplying them. (We have (a,b) + (c,d) = (a + c,b + d) and (a,b)·(c,d) = (ac – bd,ad + bc).) In Hamilton’s work, the number i became nothing more than the point (0,1). See also NEGATIVE NUMBERS; STEREOGRAPHIC PROJECTION.
composite Used in any context where it is possible to speak of the multiplication of two quantities, the term composite means “having proper factors.” For exam
ple, the number 12, which equals 3 × 4, is a COMPOSITE and y = x2 + 2x – 3 = (x–1)(x + 3) is a composite polynomial (not to be confused with the COMPOSITION of two polynomials). A quantity that is not composite is called irreducible, or, in the context of number theory, PRIME. NUMBER,
composite number A whole number with more than two positive factors is called a composite number. For example, the number 12 has six positive factors, and so is composite, but 7, with only two positive factors, is not composite. The number 1, with only one positive factor, also is not composite. Numbers larger than one that are not composite are called PRIME. The sequence 8, 9, 10 is the smallest set of three consecutive composite numbers, and 24, 25, 26, 27, 28 is the smallest set of five consecutive composites. It is always possible to find arbitrarily long strings of composite numbers. For example, making use of the FACTORIAL function we see that the string (n + 1)! + 2, (n + 1)! + 3,…,(n + 1)! + (n + 1) represents n consecutive integers, all of which are composite. (This shows, for example, that there are arbitrarily large gaps in the list of prime numbers.) See also FACTOR.
composition (function of a function) If the outputs of one function f are valid inputs for a second function g, then the composition of g with f, denoted g°f, is the function that takes an input x for f and returns the output of feeding f(x) into g: (g°f )(x) = g(f(x)) For example, if feeding 3 into f returns 5, and feeding 5 into g returns 2, then (g ° f )(3) = 2. If, alternatively, 1 f(x) = x2 + 1 and g(x) = 2 + x , then (g°f )(x) = g(f(x)) = g(x2 + 1) = 2 + 21 x +1 Typically g ° f is not the same as f ° g. In our last example, for instance,
88 compound interest (f° g)(x) = f ( g(x)) ⎛ 1⎞ = f ⎜2 + ⎟ x⎠ ⎝ 2
⎛ 1⎞ = ⎜2 + ⎟ + 1 x⎠ ⎝ which is a different function. As another example, if M is the function that assigns to each person of the world his or her biological mother, and F is the analogous biological father function, then (M° F)(John) represents John’s paternal grandmother, whereas (F° M)(John) is John’s maternal grandfather. The notation g° f is a little confusing, for it needs to be read backwards. The function f is called the “core function” and needs to be applied first, with the “external function” g applied second. The composition of three functions f, g, and h is written h° f° g (here h is the external function), and the repeated composition of a function f with itself is written f(n). Thus, for example, f (4) denotes the composition f° f° f° f. A set of repeated compositions is called a DYNAMICAL SYSTEM. Mathematicians have shown that the composition of two CONTINUOUS FUNCTIONs is itself continuous. Precisely, if f is continuous at x = a, and g is continuous at x = f(a), then g° f is continuous x = a. The composition of two differentiable functions is differentiable. The CHAIN RULE shows that the DERIVATIVE of g° f is given by (g° f)′ (x) = g′(f(x))·f ′(x).
compound interest See INTEREST.
compound statement See TRUTH TABLE.
computer An electronic device for automatically performing either arithmetic operations on DATA or sequences of manipulations on sets of symbols (as required for ALGEBRA and SET THEORY, for instance), all according to a precise set of predetermined instructions, is called a computer. The most widely used and versatile computer used today is the digital computer in which data are represented as sequences of discrete electronic pulses. As each pulse could either be “on” or “off,” it is natural to think of
sequences of 0s and 1s in working in computer theory and, consequently, to work with the system of BINARY NUMBERS to represent data. A digital computer has a number of separate parts: 1. An input device, such as a keyboard, for entering a set of instructions (program) and data. 2. A central processing unit (CPU) that codes information into binary form and carries out the instructions. (This unit consists of a series of electronic circuit boards on which are embedded a large number of “logic gates,” akin to the CONJUNCTION and DISJUNCTION configurations.) 3. Memory units, such as disks and magnetic tape. 4. An output device for displaying results, such as a monitor or a printer. The study of computer science typically lends itself to the theoretical capabilities of computing machines defined in terms of their programs, not the physical properties of actual computers. The HALTING PROBLEM and the question of being NP COMPLETE, for instance, are issues of concern to scientists in this field. See also ABACUS; CHARLES BABBAGE; DIGIT.
concave/convex A curve or surface that curves inward, like the circumference of a circle viewed from the interior, or the hollow of a bowl, for example, is called concave. A curve or surface that curves outward, such as the boundary of a circle viewed from outside the circle, or the surface of a sphere, is called convex. A geometric shape in the plane or a threedimensional solid is called convex if the boundary of the shape is a convex curve or surface. For example, triangles, squares, and any regular POLYGON are convex figures. Cubes and spheres are convex solids. Any shape that is not convex is called concave. A deltoid QUADRILATERAL, for example is a concave polygon. A polygon is convex if each of its interior angles has value less than 180°. Equivalently, a polygon is convex if the figure lies entirely on one side of any line that contains a side of the polygon. A POLYHEDRON is convex if it lies entirely on one side of any plane that contains one of its faces. A convex figure can also be characterized by the property that, for any two points inside the figure, the line segment connecting them also lies completely within the figure.
conditional 89
concave up/concave down The graph of a function y = f(x) may be described as concave up over an interval if, over that interval, the slope of the tangent line to the curve increases as one moves from left to right. Assuming the function is twice differentiable, this means that the DERIVATIVE f ′(x) is a strictly increasing function and, consequently, the double derivative satisfies f ′′(x)>0. (See INCREASING/DECREASING.) For example, the double derivative of f(x) = x2 is always positive, f ′′(x) = 2>0, and the parabola y = x2 is concave up. A concaveup graph also has the property that any CHORD joining two points on the graph lies entirely above the graph. The graph of a function y = f(x) is concave down over an interval if, over that interval, the slope of the tangent line to the curve decreases as one moves from left to right. Assuming the function is twice differentiable, this means that f′(x) is a strictly decreasing function and, consequently, f ′′(x) gives a vector the same length as a = and perpendicular to it. In fourdimensional space one can always find a fourth vector perpendicular to each of any given three vectors. See also ORTHOGONAL; TRIPLE VECTOR PRODUCT; VECTOR EQUATION OF A PLANE.
cube 111
cryptography The practice of altering the form of a message by codes and ciphers to conceal its meaning to those who intercept it, but not to those who receive it, is called cryptography. If letters of the alphabet and punctuation marks are replaced by numbers, then mathematics can be used to create effective codes. In 1977 three mathematicians, Ron Rivest, Adi Shamir, and Leonard Adleman, developed a publickey cryptography method in which the method of encoding a message can be public to all without compromising the security of the message. The RSA encryption method, as it is known today, is based on the mathematics of the MODULAR ARITHMETIC and relies on the fact that it is extraordinarily difficult to find the two factors that produce a given large product. It is the primary encryption method used today by financial institutions to transmit sensitive information across the globe. The RSA encryption method is based on the following result from modular arithmetic: Suppose p and q are distinct PRIME numbers. If n is a number with neither p nor q a factor, then n(p–1)(q–1) ≡ 1(mod pq). Moreover, we have nm(p–1)(q–1)+1 ≡ n(mod pq) for any two natural numbers n and m, even if n is a multiple of p or q. (We prove this result at the end of this entry.) One proceeds as follows: 1. Encode your message as a string M of numbers. 2. Choose two large prime numbers p and q so that their product N = pq is larger than M. Let k = (p – 1)(q – 1) and choose a number e with no common factor to k. The numbers N and e can be made public. 3. Raise the number M to the eth power, modulo N. This gives the encoded message M′:
If one works with very large prime numbers, say, 10,000digit primes, it is virtually impossible to factor the public number N and find k and d. Thus the RSA system is extremely secure. (This use of large primes also explains the current excitement over the discovery of larger and larger prime numbers.) On the other hand, multiplying and raising large numbers to powers is easy for computers to do, and so the RSA method is also very easy to implement. Proof of Result Suppose first that n is not a multiple of p and consider the numbers 1,2,…,p – 1. Multiply each by n. If, for two numbers x and y in the list, we have nx ≡ ny(mod p), then n(x – y) is a multiple of p. This can only happen if x and y are the same number. Thus, up to multiples of p, the products n·1,n·2,…,n·(p – 1) are distinct and so must represent a rearrangement of the original list, modulo p. Consequently, n·1·n·2·…·n(p – 1) ≡ 1·2·…·(p – 1) (mod p), and the factor np–1 in the left must be congruent to 1 modulo p: np–1 ≡ 1(mod p). Similarly, we have nq–1 ≡ 1(mod q) if n is not a multiple of q. It follows from these two results that if n is not a multiple of p or q, then: n(p–1)(q–1) = (np–1)q–1 ≡ 1(mod p) n(p–1)(q–1) = (nq–1)p–1 ≡ 1(mod q) which shows that n(p–1)(q–1) – 1 is divisible by both p and q, and hence by pq. This proves the first claim made. The second follows by noting that nm(p–1) (q–1)+1 = m(q–1) n·(np–1) ≡ n (mod p) is a true statement even if n is a multiple of p, and nm(p–1)(q–1)+1 ≡ n (mod q) is also true for all values of n. Any quantity nm(p–1)(q–1)+1 – n is thus always divisible by both p and q.
M′ ≡ Me (mod N) 4. Since the GREATEST COMMON DIVISOR of k and e is one, the EUCLIDEAN ALGORITHM shows that we can find numbers d and m so that 1 = de – mk. In particular there is a number d such that ed ≡ 1(mod k). Keep the number d secret. 5. To decode the message, raise M′ to the dth power. By the result stated above, this does indeed return the original message M: M′d = Med = Mkm+1 = Mm(p–1)(q–1)+1 ≡ M (mod N)
cube (hexahedron) The third PLATONIC SOLID, the cube, is the solid figure bounded by six identical square faces that meet at right angles. It has eight vertices and 12 edges. Because the VOLUME of a cube of sidelength a is given by a3, any number raised to the third power is sometimes called a cube. The “cubic numbers” are the cubes of the counting numbers: 0, 1, 8, 27, 64, 125,… It is possible to subdivide a large cube into 27 smaller cubes with six planar cuts. This number of cuts cannot be improved upon even if one is permitted to
112 cube root/nth root “stack” pieces of the cube during the slicing process so as to cut through several pieces at once. (To see this, consider the innermost cube among the 27 little cubes. It has six planar faces, each of which need to be cut. As no single planar cut will slice two of those faces, a minimum of six cuts is indeed required to form the 27 little cubes.) It is possible, however, to dice a large cube into 4 × 4 × 4 = 64 smaller cubes in fewer than nine slices if “stacking” is permitted (six will suffice). Nine slices suffice to dice a cube into 5 × 5 × 5 = 125 smaller cubes. In general, the minimum number of slices required to subdivide a large cube into n × n × n = n3 smaller cubes is given by the formula: 3 log2n making use of the CEILING FUNCTION. The fourdimensional analog of a cube is a HYPERCUBE. See also FLOOR/CEILING/FRACTIONAL PART FUNCTIONS; DUPLICATING THE CUBE; PARALLELEPIPED; PRISM.
cube root/nth root The cube root of a number a is a
3– value x such that x3 = a. We write √ a for the cube root of a. Every real number a has exactly one real cube root. For instance, the cube root of 27 is 3 (since 3 × 3 × 3 equals 27), and the cube root of –8 is –2 (since –2 × –2 × –2 equals –8). If one works within the realm of complex numbers, then the FUNDAMENTAL THEOREM OF ALGEBRA shows that every number has exactly three complex cube roots. For instance, the three cube roots
3 3 3 3 3 3 +i , − −i , and 3. A study of 2 2 2 2 the nth roots of unity shows that the three complex cube roots of a number lie on the vertices of an equilateral triangle in the complex plane. In general, the nth root of a number a is a value x such that xn = a. Again, the fundamental theorem of algebra shows that every number has exactly n complex nth roots. If a is a real number, then a has a real nth root if a is positive. (For example, a fourth root of 16 is 2.) If a is a negative real number, then a real nth root exists only if n is odd. (For example, –243 has a fifth root, namely –3, but no real fourth root.) The real n– nth root of a number a, if it exists, is denoted √ a. See also CUBIC EQUATION; ROOT; SQUARE ROOT. of 27 are −
cubic equation Any degreethree
equation of the form ax + bx + cx + d = 0 with a ≠ 0 is called a cubic equation. During the Renaissance, scholars sought for a general arithmetic formula in terms of the coefficients a, b, c, and d that would solve all cubic equations (one akin to the famous QUADRATIC formula for solving degreetwo equations). At the time, however, scholars were not comfortable working with NEGATIVE NUMBERS, or with ZERO as a number, and wrote equations in a form that avoided their appearance. (For instance, the cubic x3 – 2x + 5 = 0 was cast as x3 + 5 = 2x.) Mathematicians consequently thought that there were eight different types of cubic equations to solve. Italian scholar SCIPIONE DEL FERRO (1465–1526) was the first to make progress in solving certain cubics, but never published his results. Later, the scholar NICCOLÒ TARTAGLIA (ca. 1499–1557) succeeded in solving some additional classes of cubics. GIROLAMO CARDANO (1501–76) published Tartaglia’s work (without Tartaglia’s consent) in his epic 1545 piece Ars magna (The great art) and developed a general approach that solves all cubic equations. The formula he devised is today called “Cardano’s formula” or the “CardanoTartaglia formula.” We describe it here using modern notation: By dividing the cubic equation through by the leading coefficient a, we can assume that we are working with a cubic of the form: 3
POLYNOMIAL
2
x3 + Bx2 + Cx + D = 0 b , C=— c , and D = — d . Substituting for numbers B = — a a a B x=y–— simplifies the equation further to one without 3 a square term: y3 + py + q = 0 B2 2B3 BC (Here p = C – — and q = — – — + D .) This form 27 3 3 of the cubic is called the reduced cubic, and any solution y to this equation corresponds to a solution b x=y–— of the original equation. 3a Assume that the equation has a solution that can be written as the sum of two quantities: y = u + v. Substituting these variables yields:
cubic equation (u3 + v3) + (3uv + p)(u + v) + q = 0
2
This equation will be satisfied if we can choose u and v so that u3 + v3 = –q and 3uv + p = 0. This yields a pair of equations for u3 and v3: u3 + v 3 = − q p u3v3 = − 3
3
Solving for v3 in the first equation and substituting the result into the second shows that u3 must satisfy the quadratic equation:
v2 = w ×
3
q p q − − + 2 2 3 2
v3 = w 2 × 3 −
(u ) 3
p + q u3 − = 0 3
( )
and using the quadratic formula, this gives u3, and consequently v3 = –u3 – q, to be the two numbers: p − q ± q2 + 4 3 u3 and v3 = 2
3 2
=−
q p q ± + 2 2 3
3
We must now take the cube root of these quantities. Note first that any number M has three cube roots: one 3– 3– real, denoted √M, and two imaginary, w × √M, and −1 + i 3 3– w2 × √M, where w = . Set: 2 2
u1 = 3 −
q p q + + 2 2 3
3
2
u2 = w × 3 −
q p q + + 2 2 3 2
u3 = w 2 × 3 −
2
v1 = 3 −
q p q − + 2 2 3
3
3
is called the discriminant of the cubic, and it determines the nature of the solutions: If ∆ > 0, then the equation has one real root and two complex roots. If ∆ = 0, then the equation has three real roots, at least two of which are equal. If ∆ < 0, then the equation has three distinct real roots. In the third case, one is required to take combinations of cube roots of complex numbers to yield, surprisingly, purely real answers. For example, Cardano’s method applied to the equation x3 = 15x + 4 yields as one solution the quantity: x = 3 2 + −121 + 3 2 − −121
3
q p q + + 2 2 3
and
3
One can now check that the three quantities u1 + v1, u2 + v2, and u3 + v3 represent the three solutions to the reduced cubic y3 + py + q = 0. They constitute Cardano’s formula. The quantity under the square root sign: q p ∆ = + 2 3
3
3
q p q − + 2 2 3
2
2
113
3
It is not immediate that this number is x = 4. This confusing phenomenon of using complex quantities to produce real results was first explored by Italian mathematician RAFAEL BOMBELLI (1526–72). French mathematician FRANÇOISE VIÈTE (1540–1603) used trigonometric formulae as an alternative approach to identifying the three distinct real roots that appear in this puzzling scenario. Another Method French mathematician Viète also developed the following simpler approach to solving cubic equations. This method was published posthumously in 1615.
114 curl Take the reduced form of the cubic y3 + py + q = 0 and rewrite it as:
cusp See TANGENT.
y3 + 3ry = 2t
cyclic polygon A
p q r t = – — . Substitute y = — – z to where r = — and 3 2 z obtain the simpler equation: (z3)2 + 2t(z3) – r3 = 0 which is a quadratic in z3. All one need do now is solve for z3 and extract its three cube roots. The solution to the cubic equation is needed in the solution to the QUARTIC EQUATION. See also FUNDAMENTAL THEOREM OF ALGEBRA; HISTORY OF EQUATIONS AND ALGEBRA (essay).
curl See DIV. curve A set of points that form a line, either straight or continuously bending, is called a curve. The GRAPH OF A FUNCTION plotted in CARTESIAN COORDINATES, for example, is a curve. A curve can also be considered as the path of a moving particle (and, consequently, PARAMETRIC EQUATIONS can be used to describe it). As the graph of a function, a curve is called algebraic if it is given by a formula y = f(x) with f an algebraic function, and transcendental if f is transcendental. (See ALGEBRAIC NUMBER.) For example, the PARABOLA y = x2 is algebraic, as is any CONIC SECTION, but the sine curve y = sin x is transcendental. A curve that lies in a plane is called a planar curve. A curve in threedimensional space that does not remain in a plane, such as a HELIX, is called skew or twisted. Any curve that lies in threedimensional space is called a space curve (whether or not it is twisted). A curve is called closed if one can traverse the curve and return to the same point an indefinite number of times. For example, a CIRCLE is closed. In some settings it is appropriate to allow for a point at infinity, in which case a straight line would also be considered a closed curve. (One can head to infinity in one direction and return from infinity from the other direction and repeat this journey an indefinite number of times.) A curve with ENDPOINTs is called open. See also ARC; LOCUS.
POLYGON is called cyclic if all its vertices (corners) lie on a circle. As every triangle can be inscribed in a circle, all triangles are cyclic. So too are all squares, rectangles, and all regular polygons. However, not every quadrilateral or highersided polygon is cyclic BRAHMAGUPTA (598–665) gave a formula for the area of a cyclic quadrilateral. The CIRCLE THEOREMS show that opposite angles of any cyclic quadrilateral are supplementary. The converse is also true: any quadrilateral with opposite angles summing to 180° is cyclic. See also BRAHMAGUPTA’S FORMULA; CIRCUMCIRCLE; PTOLEMY’S THEOREM.
cycloid The shape traced out by a point on the circumference of a circle rolling along a straight line is called a cycloid. (In particular, it is the curve traced out by a piece of gum stuck to the rim of a bicycle wheel.) This curve has many remarkable geometric properties and was studied extensively by mathematicians of the 16th and 17th centuries, and later. GALILEO GALILEI (1564–1642) was the first to study the curve and gave it its name. In 1644 Galileo’s disciple Evangelista Torricelli (who invented the barometer) proved that the area under one arch of the cycloid equals three times the area of the rolling circle. In 1658 English architect Christopher Wren showed that the length of one arch of the cycloid is four times the diameter of the circle. In 1696 Johann Bernoulli of the famous BERNOULLI FAMILY posed and solved the nowfamous brachistochrone problem: Imagine a small ball starting at a point A and rolling down along a curve to a lower point B to the right of A. The ball is propelled only by the force of gravity. What shape curve connecting A to B allows the ball to travel between them the fastest? Surprisingly, a straight line does not give the shortest time, but an upsidedown cycloid does. In 1658 Dutch scientist Christiaan Huygens considered the cycloid in his work on pendulums. He discovered that a simple pendulum in which the bob is
cylindrical coordinates 115 forced to follow a cycloidshaped path always has the same period irrespective of the length of the pendulum. This is called the tautochrone property of the cycloid. (It is also the case that if a ball starts at rest at any point of an inverted cycloid and travels along the curve under the force of gravity, then the time it takes to reach the lowest point of the curve is independent of the starting location of the ball.) Related curves can be considered by following the path traced by a point on the circumference of a circle as that circle rolls along another circle. If the circle rolls on the inside of a fixed circle, then the curve traced is called a hypocycloid. If a circle rolls on the outside of a fixed circle, then the curve traced is called an epicycloid. In both cases the fixed circle is called the deferent, and the moving circle is the epicycle. Some special names are given to the curves created in particular situations. For example, when the two circles have the same radius, the epicycloid produced is heartshaped and is called a CARDIOID. When the rolling outer circle has diameter onefourth that of the fixed circle, the fourpointed curve produced is called an astroid. An epicycloid with five cusps is called a ranunculoid. The epicycloid was known to APOLLONIUS OF PERGA of the third century B.C.E., who used it in his descriptions of planetary motion.
cylinder In threedimensional space, a cylinder is the surface formed by an infinite collection of parallel straight lines, each passing through one point of a fixed closed curve drawn in a plane. The closed curve is called the directrix of the cylinder, and the lines drawn are called the generators of the cylinder. Often the term cylinder is used for the solid figure of finite volume confined between two parallel planes. In this setting, the cylinder has three faces: the two parallel planar regions, each called a base of the figure, and the lateral surface given by the straight lines that generate the cylinder. The base of a cylinder need not be a circle. For example, a CUBE satisfies the definition of being a cylinder. If the lateral surface is at right angles to the base, then the cylinder is called a right cylinder. All other cylinders are called oblique. The height of a cylinder is the perpendicular distance between the two bases.
All horizontal crosssections of a cylinder are the same size and shape as the base of the cylinder. CAVALIERI’S PRINCIPLE then shows that the volume V of a cylinder is given by V = Ah, where h is the height of the cylinder and A is the area of its base. ARCHIMEDES OF SYRACUSE (ca. 287–212 B.C.E.) showed that the volume of a SPHERE is twothirds that of the volume of the cylinder that contains it. The formula for the volume of a sphere readily follows. See also CONE.
cylindrical coordinates (cylindrical polar coordinates) In threedimensional space, the location of a point P can be described by three coordinates—r, θ, and z—called the cylindrical coordinates of P, where (r, θ) are the POLAR COORDINATES of the projection of P onto the xyplane, and z is the height of P above the xyplane. Cylindrical coordinates are useful for describing surfaces with circular symmetry about the zaxis. For example, the equation of a cylinder of radius 5 with a central axis, the zaxis can be described by the simple equation r = 5. (As the angle θ varies between zero and 360°, and the height z varies through all values, points on an infinitely long cylinder are described.) The surface defined by the equation θ = c, for some constant c (allowing r and z to vary), is a vertical HALFPLANE with one side along the zaxis, and the surface z = c is a horizontal plane. A point P with cylindrical coordinates (r, θ, z) has corresponding CARTESIAN COORDINATES (x, y, z) given by: x = r cos θ y = r sin θ z=z These formulae follow the standard conversion formulae for polar coordinates. It is usual to present the angle θ in RADIAN MEASURE. In this case, a triple integral of the form ∫∫v∫ f(x,y,z)dx dy dz over a volume V described in Cartesian coordinates converts to the corresponding integral ∫∫v∫ f(r cosθ, r sinθ,z) r dr dθ dz in cylindrical coordinates. The appearance of the term r in the integrand follows for the same reason that r appears in the conversion of a DOUBLE INTEGRAL from planar Cartesian coordinates to polar coordinates. See also ANGLE; SPHERICAL COORDINATES.
D data (singular, datum) Information of a numerical
day, namely Wednesday; and April 1, 30 days later, fell on a weekday two days after Wednesday, namely Friday. In this way we can determine the weekday of any first day of the month, and from there, the weekday of any particular day of that month. It is convenient to label Sunday as “day 0,” Monday as “day 1,” up to Saturday as “day 6.” As the weekdays cycle in units of 7, it is also appropriate to ignore all multiples of 7 and work only with the remainders of numbers upon division by 7. (That is, we shall work in a base7 system of MODULAR ARITHMETIC.) For example, dates 16 and 30 days into a year fall on the same weekday as the day 2 days into the year: all numbers involved here are 2 more than a multiple of 7. We shall call the numbers 16, 30, and 2 “equivalent” and write 16 ≡ 2 and 30 ≡ 2, for instance. January 1, 2000, fell on day 6. As we have noted, February 1 falls 31 ≡ 3 days later and so lands on day 6 + 3 ≡ 2, Tuesday. March 1 falls another 29 ≡ 1 days later, and so lands on day 6 + 3 + 1 ≡ 3, Wednesday. In general, the following table shows the amount by which a particular date must be adjusted depending on the month in which it lies:
nature is called data. For example, records of the daily numbers of visitors to a tourist attraction, or the measurement of growth rates of a yeast culture under different temperature conditions, would be examples of data. Direct counts from observational studies or measurements from experiments like these are called primary or raw data. Numerical information describing the raw data (such as the average value, largest value, range of values, and so on) is sometimes called secondary data. The science of developing methods for collecting, organizing, and summarizing data is called descriptive statistics. See also STATISTICS; STATISTICS: DESCRIPTIVE.
daysoftheweek formula It is a challenging exercise to derive a mathematical formula that determines the day of the week on which a particular calendar date falls. As a first step, knowing the day on which January 1 of a year falls, it is reasonably straightforward to determine on which day any other date of that year falls. For example, New Year’s Day in the year 2000 was a Saturday. As the days of the week cycle in periods of seven, it follows that January 8, January 15, and January 22 of that year were also Saturdays. Since there are 31 days in January, and because 31 is three more than a multiple of 7, it follows that February 1, 2000, fell on the weekday three days later than Saturday, namely, Tuesday. As there were 29 days in February that year, and 29 is one more than a multiple of 7, March 1 fell on the weekday that directly follows Tues
116
Month Add
Jan. 0
Feb. 3
March 3 (4)
April 6 (0)
May 1 (2)
June 4 (5)
Month Add
July 6 (0)
Aug. 2 (3)
Sept. 5 (6)
Oct. 0 (1)
Nov. 3 (4)
Dec. 5 (6)
daysoftheweek formula 117 The numbers in parentheses pertain to leap years. Thus, for example, July 1, 2000, fell on day 6 + 0 = 6, Saturday; and December 1, 2000, fell on day 6 + 6 ≡ 5, Friday. Now it is a matter of determining on which weekday the first day of any given year fell. Assume, for the sake of the mathematical argument, that the Gregorian calendar has been in use for two millennia, and that January 1 in the year 0 was day d. Consider the day the weekday of New Year’s Day N years later. As each ordinary year contains 365 (≡1) days, the day on which January 1 falls advances one weekday each year. For each leap year, it advances an additional day. We need to determine the number of leap years over a period of N years. In general, a leap year occurs every 4 years, yielding N 4 possible occurrences of a leap year, including year zero. (Here we are making use of the CEILING FUNCTION.) However, no leap year occurs on a year value N that is a multiple of 100—and this occurs 100 times—except if N is a multiple of 1,000, which occurs N 1000 times. (The year 1900, for instance, was not a leap year, but the year 2000 was.) The total number of leap years L that occur in a period of N years from year zero is thus given by: N N N L= − + 4 100 1000
That is, working with remainders upon division by 7, d + 5 + 6 = 6, yielding d = 2. Thus, in our theory, January 1 in the year zero was a Tuesday. We now have the following ALGORITHM for computing the weekday of any given date. Assume we wish to compute the weekday of the Dth day, of month M, in year N. 1. Consider the year number N and compute its remainder upon division by 7. N N N 2. Compute L = − + and its 4 100 1000 remainder upon division by 7. 3. Sum the answers of the previous two steps and add 2. Compute the remainder of this number, if necessary, when divided by 7. This is the weekday number of January 1 of year N. 4. To this weekday number add the day D, subtract 1, and add the appropriate month number from the table above. Look at the remainder upon division by 7, if necessary. This final result is the weekday number of the desired day. For example, for the date of March 15, 2091, N = 2091 ≡ 5, L = 523 – 21 + 3 = 505 ≡ 1, yielding January 1 of that year to be day 2 + 5 + 1 ≡ 1, a Monday. To this we add 14 days (the number of days later is 1 less than the date D) with a month adjustment of value 3 (this is not a leap year). Thus March 15, 2091, will fall on day 1 + 14 + 3 ≡ 4, a Thursday. (Warning: as the Gregorian calendar was not used before October 15, 1582, this algorithm cannot be applied to dates earlier than this. )
For instance, February 29 appeared, in theory, 2000 2000 2000 4 − 100 + 1000 = 500 − 20 + 2 = 482 times before the date of January 1, 2000. Thus, the weekday of January 1, year N, is given by:
Simplifying the Procedure This method can be simplified to some extent. Write the year number as mcyy, with m for millennia, c for century, and yy as the twodigit year number. More precisely, we mean:
d + N + L N = 1000m + 100c + yy (January 1 year zero, was day d. There is an advance for each of the N years, and an advance of an additional day for each of the L leap years.) Knowing that New Year’s day, 2000, was day 6, we deduce then that the appropriate value of d is given by: d + 2000 + 482 = 6
with 0 ≤ c ≤ 9 and 0 ≤ yy ≤ 99. For example, the year 3261 will be written: N = 1000 × 3 + 100 × 2 + 61 Notice that
118 decibel N = 1000m + 100c + yy ≡ 6m + 2c + yy N yy yy 4 = 250m + 25c + 4 ≡ 5m + 4c + 4 N yy yy 100 = 10m + c + 100 ≡ 3m + c + 100 N c yy 1000 = m + 10 + 1000 yielding: N N N N+L= N+ − + 4 100 1000 yy c yy yy = 2m − 2c + yy + + + − 100 . 4 10 1000 c yy yy One can check that the quantity + − 10 1000 100 equals 1 if yy = 00 and c ≠ 0, and zero in every other case. Rewriting steps 1 and 2 from the above algorithm, we have: January 1 of year mcyy falls on day number
same is true of 9/5 (September 5) and 7/11 (July 11), and their inverses 5/9 and 11/7. (One can remember this with the mnemonic, “I work from 9 to 5 at the 7to11 store.”) So, for example, if you know that December 20, 2002, was a Friday, and you wish to determine the day on which April 10 of that year fell, quickly compute that December 12 was a Thursday, yielding April 4 also as a Thursday, making April 10 a Wednesday.
decibel Denoted dB, a decibel is a measure of the intensity of sound, with zero decibels representing the lowest intensity at which a sound can be heard. Decibels follow a base10 LOGARITHMIC SCALE. This means that each increase of 1 dB represents a 10fold increase in the intensity of the sound. For example, a note played at 1 dB is 10 times as strong as the softest sound, and one played at 2 dB is 100 times as strong as the softest sound. Normal human speech is at a level of about 60 dB, and a whisper is around 20 dB. The threshold of pain for the human ear is about 90 dB. Rock concerts have been known to reach levels of 120 dB at a distance of 50 m from the sound system. The decibel level of a sound is computed by the formula:
yy 2 + 2m − 2c + yy + + ε 4
P 10 log10 P0
where ε is 1 if the year is a turn of the century, but not the turn of a millennium, and zero otherwise.
where P is the intensity of the note being played, and P0 is the lowest intensity at which that note can be heard. An interval of 10 dB is called a bel. Sound intensity was originally measured in terms of bel, the name being chosen in honor of the American inventor of the telephone, Alexander Graham Bell. Today, however, the unit of a decibel is considered more useful and is the one most commonly used.
Thus, for example, consider New Year’s Day of the year 3261. The millennium number here is m = 3, the century number is c = 2, and we have yy = 61. Also, 61 4 = 15.25 = 16 , and, since this is not a turn of the century, ε = 0. Thus New Year’s Day of this year will fall on day 2 + (2 × 3) – (2 × 2) + 61 + 16 + 0 ≡ 4, a Thursday. We can now compute the weekday of any other day that year if we wish. An Alternative Method Mathematician Chris McManus has observed that whatever day of the week “March 0” (the last day of February) falls, so do 4/4 (April 4), 6/6 (June 6), 8/8 (August 8), 10/10 (October 10), and 12/12 (December 12). The
decimal representation See
BASE OF A NUMBER
SYSTEM.
decomposition (factorization) The result of expressing a given object or quantity in terms of simpler components is called a decomposition. For example, the FUNDAMENTAL THEOREM OF ARITHMETIC shows that any natural number decomposes as a product of prime
Dedekind cut 119 numbers. Also, any rational function decomposes into a sum of PARTIAL FRACTIONS. The process of GAUSSIAN ELIMINATION shows that any square MATRIX A decomposes into the product of a lower triangular matrix L and an upper triangular matrix U. An example of such an LU factorization is: 2 1 1 0 2 1 = 6 7 3 1 0 4 Any VECTOR decomposes into a sum of basis vectors. For instance: = 3 +2< 0, 1, 0> +< 0, 0, 1> And in GEOMETRY, as any polygon can be divided into triangles, one could say that all polygons “decompose” into a union of triangles. Studying the simpler pieces in the decomposition of an object can lead to general results about the object. For instance, knowing that the interior angles of a triangle sum to 180° allows us to immediately deduce that the interior angles of any quadrilateral (the union of two triangles) sum to 360°, and that the interior angles of any pentagon (the union of three triangles) sum to 540°. See also FACTORIZATION.
Dedekind, Julius Wilhem Richard (1831–1916) German Analysis Born on October 6, 1831, in Braunschweig, now a part of Germany, Richard Dedekind is remembered for his elegant construction of the REAL NUMBER system, which is based on an idea today known as a DEDEKIND CUT. This work represented an important step in formalizing mathematics. In particular, it offered the means to finally put CALCULUS on a sound mathematical footing. Dedekind studied NUMBER THEORY and calculus at the University of Göttingen. He earned a doctoral degree in 1852 under the supervision of CARL FRIEDRICH GAUSS (he was Gauss’s final pupil), and two years later obtained a habilitation degree granting him the right to be a member of the university faculty. In 1858 Dedekind accepted a position at the Polytechnikum in Zürich. Dedekind realized that the foundations of calculus, in particular, the properties of the realnumber system on which calculus rests, were not
properly understood. When faced with the challenge of teaching calculus to students at the Polytechnikum for the first time, Dedekind decided not to sidestep the issue, but rather develop an approach that would properly justify the principles of the subject to himself and to his students. This is when the idea of a Dedekind cut came to him. Dedekind published the details of this construction several years later in his famous 1872 paper “Continuity and Irrational Numbers.” This paper was extremely well received and was admired not only for the brilliant ideas it contained, but also for the manner in which those ideas were detailed. Dedekind exhibited a talent for explaining mathematical concepts with exceptional clarity. In 1862 Dedekind returned to his hometown to accept a position at the Brunswick Polytechnikum. He remained there for the rest of his life. He never married and lived his life with one of his sisters, who also remained unmarried. Dedekind received many honors for his outstanding work, including election to the Berlin Academy in 1880, the Academy of Rome and the Académie des Sciences, Paris, in 1900, as well as honorary doctorates from the Universities of Zurich, Brunswick, and Oslo. Dedekind died in Brunswick, Germany, on February 12, 1916. Dedekind made a lasting impact on the modern understanding of the realnumber system. Most every collegelevel course on the topic of the real numbers will discuss in detail the issues Dedekind explored.
Dedekind cut During the 1800s it became clear to mathematicians that in order to prove that CALCULUS is mathematically sound one needs to properly define what is meant by a real number and, moreover, show that the real number system is “complete,” in the sense that no points are “missing” from it. This is particularly important for establishing the EXTREMEVALUE THEOREM, the INTERMEDIATEVALUE THEOREM, and the MEANVALUE THEOREM. All the key theorems in calculus rely on these three results. Although the RATIONAL NUMBERS Q are relatively easy to define, the system of rationals is certainly not complete: the square root of 2, for example, is not a fraction and so is “missing” from the set of rationals. The task of defining exactly what is meant by an irrational number perplexed scholars for a very long time,
120 deductive/inductive reasoning and the definition of a REAL NUMBER was subject to much debate. In 1872 JULIUS DEDEKIND had the very simple and elegant idea to simply define the irrationals to be the gaps in the rational number line. He noted that each “gap,” like the square root of 2 for example, divides the line of rationals into two pieces—a left piece and a right piece. One can focus one’s attention on just the left piece (for those points that are not in it constitute the right piece) and this left piece L satisfies the following three properties: 1. It is not empty, nor is it the whole set of points. 2. If a is a number in L, and b = k , and so x1 = x0 – ka,y1 = y0 – kb, and z1 = z0 – kc. Since N = (x1,y1,z1) lies in the plane, this point also satisfies the equation of the plane. Algebraic manipulation then ax0 + by0 + cz0 + d shows that k = . By the distance a 2 + b2 + c 2 formula, the distance between P and N is:
d = √(x2 – x1)2 + (y2 – y1)2 + (z2 – z1)2
 ax0 + by0 + cz0 + d 
The set of all points (x,y,z) in space a fixed distance r from a given point C = (a,b,c) form a SPHERE with radius r and center C. The distance formula gives the equation of such a sphere as (x – a)2 + (y – b)2 + (z – c)2 = r2. The distance formula generalizes to points in ndimensional space as the square root of the sum of the differences of the n coordinates squared. This works even for onedimensional space: the distance between two points x1 and x2 on the number line is d = √(x2 – x1)2 . This is precisely the ABSOLUTE VALUE x2 – x1. The LENGTH of a VECTOR v = < a,b,c > is given by the distance formula: If we place the vector at location O = (0,0,0) so that its tip lies at A = (a,b,c), then its length is v = √a2 + b2 + c2 . Distance of a Point from a Plane in ThreeDimensional Space The distance of a point P from a plane is defined to be the distance between P and the point N in the plane closest to P. Suppose that the point P has coordinates P = (x0,y0,z0) and the VECTOR EQUATION OF A PLANE is ax + by + cz + d = 0 where n = < a,b,c > is the normal to the plane. Then N is the point (x1,y1,z1) in the plane with vector NP 90° to the plane. This means that the vector NP is parallel to n, and so NP = kn for some constant k. This
(x0 − x1)2 + (y0 − y1)2 + (z0 − z1)2 = k2 a2 + k2b2 + k2c 2 =  k  a 2 + b2 + c 2 =
 ax0 + by0 + cz0  a 2 + b2 + c 2
This establishes: The distance of a point P = (x0,y0,z0) from the plane ax + by + cz + d = 0 is given by the formula:
a 2 + b2 + c 2
Distance of a Point from a Line in TwoDimensional Space The distance of a point P from a line is defined to be the distance between P and the point N in the line closest to P. The EQUATION OF A LINE is a formula of the form ax + by + c = 0. An argument analogous to the one presented above establishes: The distance of a point P = (x0,y0) from the line ax + by + c = 0 is given by the formula:  ax0 + by0 + c  a 2 + b2 See also COMPLEX NUMBERS.
distribution Any table or diagram illustrating the frequency (number) of measurements or counts from an experiment or study that fall within certain preset categories is called a distribution. (See STATISTICS: DESCRIPTIVE.) For example, the heights of 1,000 8yearold children participating in a medical study can be
distributive property 143 recorded via a histogram. The categories considered are conveniently chosen intervals of height ranges, such as 36.1–40.0 in., 40.1–44.0 in., and so on, for example. If the DATA gathered is numerical and can adopt a continuous array of values, including fractional values (for example, height can adopt fractional values—48 3/4 in., or 52.837 in. are possible measurements), then one can choose narrower and narrower interval ranges for categories. In the LIMIT, the histogram becomes then the graph of a smooth curve representing the distribution of measurements over a continuous spectrum of values. In some sense, the total area under the curve represents the total number of measurements observed, and the area above an interval [a,b] represents the number of measurements that have value greater than a and less than b. To make this more precise, it is appropriate to scale the distribution so that the total area under the curve is one (that is, one draws histograms with vertical bars of heights representing the percentage of measurements recorded within that category—”relative frequencies”—with the total area under such a histogram representing 100 percent). The ideal curve obtained in the limit is called a “probability density function.” The area under the curve above an interval [a,b] here represents the PROBABILITY that a measurement taken at random falls within the range [a,b]. A numerical quantity that can adopt a continuous array of values (such as height, weight, or temperature) is usually called a continuous random variable. One ascertains the distribution (probability density function) of a random variable by conducting experiments or studies—for example by recording the heights of 1,000 8yearold children—or from mathematical reasoning, making use of the CENTRALLIMIT THEOREM, the NORMAL DISTRIBUTION, or perhaps the BINOMIAL DISTRIBUTION, for example. Often the distribution of a random, variable is unknown, and “hypothesis testing”
is used to check the validity of an assumption that a SAMPLE follows a particular distribution. This is part of inferential statistics. If a random variable has probability density function given by a formula f(x), then the area under the curve to the left of a value x gives a new function F(x) called the cumulative distribution function. The quantity F(x) represents the probability that a measurement taken at random has value less than or equal to x. See also STATISTICS: INFERENTIAL.
distributive property Given a mathematical system with two operations, such as addition and multiplication, or union and intersection, we say that one operation distributes over the second if applying the first operation to a set of elements combined via the second produces the same result as applying the first operation to the individual members of the combination, and then combining them via the second. For example, in ordinary arithmetic, multiplication distributes over addition. We have, for instance: 3 × (2 + 5 + 4) = 3 × 2 + 3 × 5 + 3 × 4 that is, tripling a sum of numbers produces the same result as tripling each individual number and then summing. (In arithmetic, the distributive property corresponds to the operation of EXPANDING BRACKETS.) Notice that addition is not distributive over multiplication, however. For instance: 4 + (6 × 7)≠(4 + 6) × (4 + 7). (The first quantity equals 4 + 42 = 46, whereas the second is 10 × 11 = 110.) In arithmetic, the distributive property is usually expressed as a multiplication applied to the sum of just two terms: a × (b + c) = a × b + a × c That it applies to a sum of three or more terms follows from applying this basic law more than once. For instance: a × (b + c + d) = a × ((b + c) + d) = a × (b + c) + a × d =a×b+a×c+a×d
A distribution as a limit
Multiplication also distributes over addition “from the right.” We have:
144 div (a + b) × c = c × (a + b) =c×a+c×b =a×c+b×c
point P, is the angular velocity of an element of fluid located at P, that is, it is a measure of the amount of turning it undergoes. See also GRADIENT.
In SET THEORY, “intersection” distributes over “union” both from the left and from the right:
divergent This term simply means “does not conA∩(B∪C) = (A∩B)∪(A∩C) (A∪B)∩C = (A∩C)∪(B∩C) As a mnemonic device, it is helpful to think of the phrase “distributes over” as synonymous with “sprinkles over.” We have: multiplication “sprinkles over” additions and intersection “sprinkles over” union. In any RING, the distributive property is the single axiom that combines the two defining operations. See also ASSOCIATIVE; COMMUTATIVE PROPERTY.
div A
VECTOR FIELD assigns to every point (x,y,z) in space a vector F = . The divergence of F, denoted div F, is the quantity:
div F =
∂f1 ∂f2 ∂f3 + + ∂x ∂y ∂z
divisibility rules A number is said to be divisible by
given as a sum of PARTIAL DERIVATIVEs. Physicists have shown that this quantity represents the amount of flux leaving an element of volume in space. For example, if F represents the velocity field of a turbulent fluid, and ρ is the density of the fluid, then ρdivF, calculated at a point P, is the rate at which mass is lost from an (infinitely small) box drawn around P. (At any instant, fluid is flowing into and out of this box.) The divergence operator is often written as though it is a DOT PRODUCT of two vectors: divF = · F where ∇ = PRODUCT
∂ ∂ ∂ , , ∂x ∂y ∂z
is the del operator. The
verge.” For example, an infinite SEQUENCE is said to diverge if it has no LIMIT, and an infinite SERIES diverges if the sequence of partial sums diverges. A divergent sequence is said to be “properly divergent” if it tends to infinity. For example, the sequence 1,2,3, … is properly divergent (but the sequences 1, –1,1, –1,1, … and 1, –2,3, –4,5, … are not). One also describes a series as properly divergent if the corresponding sequence of partial sums has this property. For example, the series 1 + 2 + 4 + 8 + 16 +… is properly divergent. An INFINITE PRODUCT is divergent if it has value zero or does not converge. An IMPROPER INTEGRAL diverges if the limit defining it does not exist. See also CONVERGENT SEQUENCE.
CROSS
of with F is called the curl of F:
∂f ∂f ∂f ∂f ∂ f ∂f curl F = ∇× F = 3 − 2 i + 1 − 3 j + 2 − 1 k ∂ y ∂ z z x x ∂ ∂ ∂ ∂y If F is the velocity field of a turbulent liquid, then physicists have shown that 1/2 curl F, calculated at a
n if, working solely within the integers, the number leaves a remainder of zero when divided by n. For example, 37 leaves a remainder of 1 when divided by 3, and so is not divisible by 3. On the other hand, 39, leaving a remainder of zero, is divisible by 3. There are a number of rules to quickly test the divisibility of numbers by small integers. We present divisibility rules for the first 12 integers. Divisibility by 1 All numbers are divisible by 1. Divisibility by 2 As all multiples of 10 are divisible by 2, it suffices to check whether or not the final digit of a number is divisible by 2. For example, 576 = 57 × 10 + 6. That 6 is a multiple of 2 ensures that 576 is too. We have the rule: A number is divisible by 2 only if its final digit is 0, 2, 4, 6 or 8. Divisibility by 3 That 10, 100, 1000,… all leave a remainder of 1 when divided by 3 allows us to quickly determine the remain
divisibility rules 145 der of any number divided by 3. For example, 3,212 equals 3 × 1000 + 2 × 100 + 1 × 10 + 2 × 1 and so leaves a remainder of 3 × 1 + 2 × 1 + 1 × 1 + 2 × 1 = 8 when divided by 3. This is the sum of its digits. Of course, a remainder of 8 is equivalent to a remainder of 2. We have: The remainder of any number divided by 3 is the sum of its digits. Thus a number is divisible by 3 only if the sum of its digits is a multiple of 3. One can make repeated use of this rule to check for divisibility. For example, 55,837 leaves a remainder of 5 + 5 + 8 + 3 + 7 = 28 when divided by 3. This corresponds to a remainder of 2 + 8 = 10, which, in turn, is a remainder of 1 + 0 = 1. This rule shows that scrambling the digits of any multiple of 3 produces a new number that is still a multiple of 3. Divisibility by 4 As any multiple of 100 is divisible by 4, it suffices to check whether or not the final two digits of a number represent a twodigit multiple of 4. For example, 18,736 equals 187 × 100 + 36. As 36 is a multiple of 4 (it can be divided by 2 twice), we are sure then that 18,736 is a multiple of 4. We have: A number is divisible by 4 if its final two digits represent a twodigit number that can be divided by 2 twice. Divisibility by 5 Any number N can be written in the form N = 10a + b where b is the final digit in N. (For example, 739 = 73 × 10 + 9.) As 10 is divisible by 5, we need only check whether or not the final digit b is divisible by 5. This gives: A number is divisible by 5 only if its final digit is 0 or 5. Divisibility by 6 For a number to be divisible by 6 it must both be even and a multiple of 3. This gives: A number is divisible by 6 only if it is an even number whose digits sum to a multiple of 3.
Divisibility by 7 Every number N is of the form N = 10a + b, that is, a multiple of 10 plus a single digit. As 7a and 7b are clearly divisible by 7, we have that 10a + b leaves the same remainder as 10a + b – 7a – 7b = 3(a – 2b) does when divided by 7. Thus it suffices to check whether the quantity 3(a – 2b) is a multiple of 7. This can only occur if a – 2b is a multiple of 7. Noting that a is the original number N with its final digit removed and b is the final digit of N, we have the rule: To test whether or not a number is divisible by 7, remove the last digit and subtract twice that digit from the number remaining. Then the original number is divisible by 7 only if the result of this operation is divisible by 7. For example, to test whether or not 68,978 is divisible by 7, remove the 8 and subtract twice this, 16, from 6,897, the number remaining. This gives 6,897 – 16 = 6,881. We can test whether or not 6,881 is a multiple of 7 the same way: 6,881 → 688 – 2 = 686, and once more: 686 → 68 – 12 = 56. That the final result, 56, is divisible by 7 assures us that 68,978 is a multiple of 7. Divisibility by 8 As any multiple of 1,000 is divisible by 8, it suffices to check whether or not the final three digits of a number represent a threedigit multiple of 8. For example, 648,728 equals 648 × 1000 + 728. As 728 can be divided by 2 three times, and hence is a multiple of 8, we have that 648,728 is divisible by 8. A number is divisible by 8 if its final three digits represent a threedigit number that can be divided by 2 three times. Divisibility by 9 Given that 10, 100, 1000,… all leave a remainder of one when divided by 9, the divisibility rule for 9 is identical to that of 3. The remainder of any number divided by 9 is the sum of its digits. Thus a number is divisible by 9 only if the sum of its digits is a multiple of 9. Again, one may make repeated use of this rule. For example, 76,937 leaves a remainder of 7 + 6 + 9 + 3 + 7 = 32 when divided by 9, which, in turn, corresponds to a remainder of 3 + 2 = 5. This rule is often used to
146 division check arithmetical work via the method of
CASTING
OUT NINES.
Divisibility by 10 Any number N can be written in the form N = 10a + b, where b is the final digit of N. Thus a number is divisible by 10 only if its final digit is a multiple of 10. We have: A number is divisible by 10 only if its final digit is a zero. Divisibility by 11 The numbers 100, 1000,… alternately leave remainders of 1 and –1 when divided by 11. (For example, 100 is 1 more than a multiple of 11, but 1,000 is 1 less.) Thus the remainder of a number when divided by 11 is obtained as the alternate sum of its digits. For example, 69,782, which equals 6 × 10,000 + 9 × 1,000 + 7 × 100 + 8 × 10 + 2 × 1, leaves a remainder 6 × 1 + 9 × (–1) + 7 × 1 + 8 × (–1) + 2 × 1 = 6 – 9 + 7 – 8 + 2 = –2 when divided by 11. (This is equivalent to a remainder of 9.) We have: The remainder of any number when divided by 11 is the alternate sum of its digits. Thus a number is divisible by 11 only if the alternate sum of its digits is a multiple of 11. Divisibility by 12 A number is divisible by 12 only if it is divisible by both 3 and 4. Thus we have: A number is divisible by 12 only if its final two digits represent a twodigit multiple of 4, and the sum of all the digits of the number is a multiple of 3. The divisibility rule for 7 can be extended to other numbers as well. For example, N = 10a + b is divisible by 17 only if 10a + b – 51b = 10(a – 5b) is. This, in turn, shows that N is divisible by 17 precisely when quantity a – 5b, obtained by deleting and subtracting 5 times the final digit, is divisible by 17. Notice here that 51 is the first multiple of 17 that is 1 more than a multiple of 10. In the same way we can use that fact that 111 is the first multiple of 37 that is 1 more than a multiple of 10 to obtain a similar divisibility rule for 37, for exam
ple. Divisibility rules for all PRIME numbers, except 2 and 5, can be created this way.
division The process of finding the QUOTIENT of two numbers is called division. In elementary arithmetic, the process of division can be viewed as repeated SUBTRACTION. For instance, 60 divided by 12 equals 5 because 12 can be subtracted from this number five times before reaching zero: 60 – 12 – 12 – 12 – 12 – 12 = 0. We write: 60 ÷ 12 = 5. Division can also be described as the process of finding how many subsets or magnitudes are contained within a set or given quantity. For instance, 5 ÷ 1/2 = 10 because 10 lengths of onehalf are contained in a length of 5 units. If a number a is divided by a number b to produce a quotient q, a ÷ b = q, then a is called the dividend and b the divisor. The quotient can also be expressed as a FRACTION, a/b, or a RATIO, a:b. In general, the quotient q of two numbers a and b satisfies the equation q × b = a. Thus division may also be thought of as the inverse operation to MULTIPLICATION. Thus, since 5 × 12 = 60, for instance, 5 is indeed the quotient of 60 and 12. This reasoning also shows that, since 0 × b = 0 for any nonzero number b, we have 0 ÷ b = 0. Unfortunately, one cannot give meaning to the quantity 0 ÷ 0. (Given that 53 × 0 = 0, we may be forced to conclude that 0 ÷ 0 = 53. At the same time, since 117 × 0 = 0, we also have that 0 ÷ 0 = 117. We have inconsistency.) It is also not possible to give meaning to the term a ÷ 0 for any nonzero value a. (If a ÷ 0 = q, then q × 0 = a, yielding a CONTRADICTION.) The LONG DIVISION algorithm provides a means to divide large integers. The process of division can be extended to NEGATIVE NUMBERS, FRACTIONs, REAL NUMBERS, and COMPLEX NUMBERS. In all settings, the number 1 acts as an identity element—provided it operates as a divisor: a ÷ 1 = a for all numbers a. The symbol ÷ is called the “obelus” and first appeared in print in Johann Heinrich Rahn’s 1659 text Teutsche algebra. See also DIVISIBILITY RULES; DIVISOR; DIVISOR OF ZERO; EUCLIDEAN ALGORITHM; FACTOR; FACTORIZATION; FACTOR THEOREM; RATIONAL FUNCTION; REMAINDER THEOREM.
divisor Another name for FACTOR.
dot product 147
divisor of zero Two quantities, neither of which are zero, yet multiply together to give zero as their product are called divisors of zero. In ordinary arithmetic, divisors of zero never arise: if a × b = 0, then at least one of a or b must be zero. In MODULAR ARITHMETIC, however, divisors of zero can exist. For example, 3 × 2 equals zero in arithmetic modulo 6. Two nonzero matrices may multiply to give the zero MATRIX, and the product of two nonzero functions could be the zero function. The presence of zero divisors in a mathematical system often complicates the arithmetic one can perform within that system. For example, for arithmetic modulo 10, 2 × 4 equals 2 × 9, but dividing by 2, a divisor of zero in this system, leads to the erroneous conclusion that 4 and 9 are equal in this system. In general, one can never perform division when a divisor of zero is involved.
dot product (inner product, scalar product) Denoted a · b, the dot product of two VECTORs a and b is the sum of the products of respective components of the vectors. Precisely, if a = and b = , then a · b is the number: a · b = a1b1 + a2b2+…+ anbn The result of the dot product operation is always a real number (scalar). For example, the dot product of the two vectors a = < 1,4> and b = < 2,3> is a · b = 1 · 2 + 4 · 3 = 14. The dot product arises in many situations. For example, a local delicatessen receives the following lunch order form:
Triangle of vectors
cost of the order is then the dot product of these two vectors: 2 × 2.80 + 0 × 3.15 + 1 × 1.95 + 3 × 2.50 + 0 × 3.60 = $15.05 Geometrically, the dot product gives a means of computing the angle between two vectors. For example, two twodimensional vectors a = < a1,a2 > and b = < b1,b2 >, with angle θ between them, form a triangle in the plane with sidelengths given by the DISTANCE 2 2 2 2 FORMULA: a = √a1 + a2 , b = √b1 + b2 , and a – b = √(a1 – b1)2 + (a2 – b2)2 . By the LAW OF COSINES we have: a – b2 = a2 + b2 – 2ab cos(θ) from which it follows that: a · b = a1b1 + a2b2 = ab cos(θ)
How Many? 2 1 3
Cost Ham Turkey Egg Salad Tuna Roast Beef
@$2.80 @$3.15 @$1.95 @$2.50 @$3.60 Total Cost:
——— ——— ——— ——— ——— ———
One could interpret the order as a fivedimensional vector to be matched with a fivedimensional “cost vector” . The total
Thus the angle between two vectors a and b can be computed via the formula: cos(θ ) =
a⋅b  a  b 
This formula also holds true for three and higherdimensional vectors. For example, consider the unit vector i = in threedimensional space pointing in the direction of the xaxis, and j = the corresponding vector pointing in the direction of the yaxis. As the angle
148 double integral between these two vectors is 90°, their dot product i · j will be zero (i · j = 1.0 + 0.1 + 0.0 = 0). In general: Two vectors a and b are at right angles if, and only if, their dot product a · b is zero.
result holds true for other shaped regions R as well, as long as they are not too complicated. For example, the volume under the graph z = xy above the rectangle R = [1,2] × [2,3] is:
=∫
ORTHOGONAL;
TRIPLE
VECTOR
PRODUCT;
29
1
CROSS PRODUCT; NORMAL TO A PLANE; VECTOR
EQUATION OF A PLANE.
y =3
R
a·b=b·a a · (b + c) = a · b + a · c See also
2 1
2 3
∫∫ xydA = ∫1 ∫2 xydydx = ∫1 2 xy 2 y =2 dx
The dot product has the following commutative and distributive properties:
2
x − 2xdx = ∫
2
1
5 15 x dx = 4 8
Notice that the integration is performed from the inside out. A triple integral ∫∫∫ f (x, y, z) dV of a function of V
double integral The volume under a graph z = f(x,y) of two variables (which is drawn as a surface sitting in threedimensional space) above a region R in the xyplane is computed via a double integral, denoted:
∫∫ f (x, y)dA R
One approximates this volume by subdividing the region R into small rectangular pieces, drawing a rectangular cuboid above each rectangle with height reaching the surface, and summing the volumes of each of these cuboids. As one takes finer and finer approximations, this process produces better and better approximations to the true volume under the graph. The limit of this process is the double integral:
three variables f(x,y,z), computed over a volume V in space, can often be computed as a triple iterated integral, integrating each variable in turn. Again, the order of the integration, typically, does not matter. See also GEORGE GREEN.
double point A location on a curve where the curve either crosses itself, or is tangential to itself, is called a double point. In the first case, the point of intersection is called a node, and the curve has two distinct tangents at that point. In the second case, the point of contact is called a tacnode or an osculation. The two tangents to the curve coincide at this point. See also ISOLATED POINT; SINGULAR POINT; TANGENT. double root See ROOT.
n
∫∫ f (x, y) dA = lim ∑ f (xk , yk)dAk k=1
R
where dAk denotes the area of the kth rectangular region used to approximate R. GOTTFRIED WILHELM LEIBNIZ (1646–1716) showed that if the region R is itself a rectangle, say, given by a ≤ x ≤ b and c ≤ y ≤ d, then the double integral can be computed as either of the two iterated integrals: b d
∫∫ f (x, y)dA = ∫a ∫c R
d b f (x, y) dy dx = ∫ ∫ f (x, y) dy dx c a
dummy variable A variable appearing in a mathematical expression is a dummy variable if it is assigned no specific meaning and if the letter being used for it could equally well be replaced by another letter. An index of SUMMATION, for instance, is a dummy variable: 4
the sum
of denoting 13 + 23 + 33 + 43, for exam
k=1
4
ple, could equally well be represented as
∑ r3
or
r =1 4
∑ n3 ,
n =1
(In an iterated integral, one integrates one variable at a time, regarding the second variable as a constant.) This
∑ k3
say. The variable used for the integrand of a
is a dummy variable. The two 1 1 expressions ∫0 x2 dx and ∫0 t2 dt, for instance, represent
DEFINITE INTEGRAL
Dürer, Albrecht 149 the same definite integral, a number, and so x and t are dummy variables. (The variable x, however, is not a dummy variable in the INDEFINITE INTEGRAL ∫x2 dx. This expression is a function of the specific variable x.)
3– integers. A similar argument shows that √ 2 does not equal the square root of a rational quantity either.)
Dürer, Albrecht (1471–1528) German Geometry Born duplicating the cube (Delian altar problem, doubling the cube) One of the problems of antiquity (like and TRISECTING AN ANGLE) of considerable interest to the classical Greek scholars is the task of constructing a cube whose volume is twice that of a given cube. Legend has it that this problem, known as duplicating the cube, arose during the Greek plague of 428 B.C.E. It is said that the oracle of Delos instructed the people of Athens to double the size of the cubic altar to Apollo as an attempt to appease the god. They were unable to accomplish this feat. APOLLONIUS OF PERGA (ca. 260–190 B.C.E.) solved the problem with the use of CONIC SECTIONS, but scholars later decided to add the restriction that only the primitive tools of a straightedge (that is, a ruler with no markings) and a compass be used in its solution. The difficulty of the problem increased significantly. If we assume that the sidelength of the original cube is a units long, then one is required to construct a 3— new length b so that b3 = 2a3. Consequently, b = √ 2a, and so the problem essentially reduces to the challenge 3— of constructing a length √ 2 units long using only a straightedge and compass. The theory of CONSTRUCTIBLE numbers shows that, in this setting, any quantity of rational length can be constructed, and that if two lengths l1 and l2 can be produced, then so too can their sum, difference, product and quotient, along with the square root of each 3— quantity. It seems unlikely that a length of √ 2, being neither rational, nor the square root of a rational number, could be produced. Indeed, in 1837, French mathematician Pierre Laurent Wantzel (1814–48) proved that 3— the number √ 2 is not constructible and, consequently, that the problem of duplicating the cube is unsolvable. 3— (To see that √ 2 is not rational, assume to the contrary p 3— that it can be written as a ratio of two integers: √ 2 = – q. Then 2q3 = p3. If the number p has m factors of 2, then the quantity p3 has 3m factors of 2. Consequently, so too must 2q3. But this is impossible, as the number of factors of 2 in 2q3 must be 1 more than a multiple of 3. 3— This absurdity shows that √ 2 cannot be a ratio of two SQUARING THE CIRCLE
on May 21, 1471, in Nürnberg, Germany, artist Albrecht Dürer is remembered in mathematics for his significant accomplishments in the development of descriptive GEOMETRY and its applications to the theory of art. In four famous texts, Dürer explained the theory of proportions and described rulerandcompass techniques for the construction of regular polygons. He explored the art of placing figures in a manner that is pleasing to the eye, thereby beginning a developing theory of PERSPECTIVE, and began a study of the shadows cast by threedimensional objects. Dürer is noted as the first scholar to publish a mathematics book in German, and also as the first Western scholar to give an example of a MAGIC SQUARE. Dürer studied painting and woodcut design as a young man. He apprenticed with the leading producer of altarpieces of his time, Michael Wolgemut, until the age of 20 and learned to appreciate the role mathematics could play in the design of artistic works. After reading the works of EUCLID (ca. 300–260 B.C.E.), as well as a number of famous texts on the theory of architecture, Dürer traveled to Italy, the site of the Renaissance revival of mathematics, to study the mathematics of shape, motion, and perspective. Around 1508 Dürer began collating and processing all the material he had studied with the aim of producing one definitive text on the mathematics of the visual arts. This work was never completed, but he did later publish his four volumes on the theory of proportions Underweysung der Messung mit Zirckel und Richtscheyt in Linien, Ebnen, und gantzen Corporen (Treatise on mensuration with the compass and ruler in lines, planes, and whole bodies) in 1525. Dürer is noted for his inclusion of the following array of numbers in the background of his 1514 engraving Melancholia:
16 5 9 4
3 10 6 15
2 11 7 14
13 8 12 1
150 dyadic Each row and column, as well as the two main diagonals, sum to 34, and so this array represents a fourthorder magic square. It is the first example of a magic square ever recorded in Western Europe. (It has the added feature of including the year the engraving was completed in the two middle cells of the bottom row.) Dürer died in Nürnberg, Germany, on April 6, 1528. His theory of proportions allowed artists who succeeded him to easily perform the transformations on figures needed to translate them across a canvas and maintain the correct sense of perspective. (Prior to Dürer, artists accomplished this feat purely by intuition or by trial and error.) In doing so, Dürer had provided a wellthoughtout theory of geometric perspective that was also valued by mathematicians.
dyadic Any quantity related to the concept of base 2 is sometimes referred to as dyadic. For example, the dyadic rationals are those fractions whose denominators are 3 3 173 173 – –— –— powers of 2. Thus – 4 = 22 and 1024 = 210 , for 1 instance, are dyadic rationals, but – is not. 3 Folding a strip of paper 1 ft long in half produces a crease at the position of the dyadic rational 1/2. If one continues to fold the left or right end of the strip to previously constructed crease marks, then crease marks appear at all the dyadic rationals (and only the dyadic rationals). This paperfolding activity is intimately connected to the construction of BINARY NUMBERS. For instance, to create a crease mark along a strip of paper at the position of the dyadic rational 13/16, write the numerator 13 in binary: 13 = 11012 (This is equivalent to writing the fraction 13/16 as a binary “decimal:” 13/16 = .1101 in base 2.) Now read the binary expansion backwards, interpreting the digit 1 as the instruction “lift right and fold” and the digit 0 as “lift left and fold.” In this example we have: 1 : Lift the right end of the strip and fold to produce a crease at position 1/2. 0 : Lift the left end and fold to the previous crease. This produces a new crease at position 1/4.
1 : Lift the right end and fold to the previous crease. This produces a new crease at position 5/8. 1 : Lift the right end and fold to the previous crease. This produces a new crease at position 13/16, as desired. In general: The binary representation of the numerator of any dyadic rational represents instructions for the construction of that dyadic rational along a strip of paper. In some limiting sense, this procedure also works for fractions that are not dyadic. For example, the number 1/3 written as decimal in base 2 is .010101… If one were to read this as a set of instructions to “fold right and fold left indefinitely,” then the sequence of creases produced do indeed converge to the position 1/3.
dynamical system Any process in which each successive state is a function of the preceding state is called a dynamical system. For instance, the feedback from a microphone as part of a public announcement system is a dynamical system: the amplifier transmits minute erroneous sounds, which the microphone hears and amplifies, which it then hears and amplifies, and so on. In mathematics, if f is a mapping from a space X to itself, then the ITERATION of f defines a dynamical system—the successive states of the system are the iterates of f arising from a given starting point x: x,f(x),f(f(x)),f(f(f(x))),… 1 For example, the mapping f(x) = – 2 x of the real number 1 line to itself, starting with x = 1, gives the iterates 1, – , 2 1 1 – , – ,…. If X is a circle in the plane and f is the function 4 8 that rotates that circle 10° clockwise, then the iterates of any point on the circle constitute 36 evenly spaced points on that circle. The “orbit” of a point x is the sequence of iterates it produces. If the system reaches an equilibrium, that is, tends toward a stable state, or if it cycles between a number of states, then the equilibrium point (or sets of equilibrium points) are called “attractors” of the sys
dynamical system 151 tem. For example, the number zero is an attractor for 1 the system given by f(x) = – 2 x: no matter the starting value, all iterates converge to the value zero. The iterates of very simple functions f can exhibit extremely surprising behavior. Take, for instance, the iterates of the function f(x) = 1 – cx2, with initial value x = 0. (Here c is a constant. Each value of c determines its own dynamical system.) Set c = 0.1. The first five iterates of the system xn+1 = 1 – 0.1xn2 are: 0, 1.000, 0.900, 0.919, 0.916, 0.916 The system seems to stabilize to the value 0.916. One checks that the same type of behavior occurs if we repeat this exercise for c set to any value between 0 and 0.75. There is a marked change in behavior for c = 0.75, however—the system no longer converges to a single value but rather oscillates between two values: 0.60 and 0.72. We say that the value c = 0.75 is a bifurcation point and that the system has undergone “period doubling.” At the value c = 1.25, the system bifurcates again to yield systems that oscillate between four separate
values. For higher values of c, the system continues to bifurcate, until finally a socalled CHAOS is reached, where the results jump around in a seemingly haphazard manner. This phenomenon is typical of many dynamical systems: the behavior they exhibit is highly dependent on the value of some parameter c. (Such dynamical systems are said to be “sensitive” to the parameter set.) Researchers have shown that many natural processes that appear chaotic, such as the turbulent flow of gases and the rapid eye movements of humans, can be successfully modeled as dynamical systems, usually with very simple underlying functions defining them. Meteorologists model weather as a dynamical system, which helps them make forecasts. However, extreme sensitivity to parameters can easily lead to erroneous predictions: one small change in the value of just one parameter may produce very different outcomes. The socalled butterfly effect, for instance, claims that the minute changes in air pressures caused by a butterfly flapping its wings might be all that is needed to tip a meteorological dynamical system into chaos. Iteration of functions with COMPLEX NUMBERS leads to a study of FRACTALs.
E e (Euler’s number) Swiss mathematician LEONHARD
The fourth definition arises from the study of TAYLOR One proves that all four definitions are equivalent as follows: First consider the curve y = 1/x. It has the remarkable property that rectangles touching the curve and just under it have the same area if the endpoints of the rectangles are in the same ratio r. For example, in the first diagram on the opposite page, the rectangles above the intervals [a,ra] – 1 . By taking narrower and and [b,rb] each have area r––– r narrower rectangles, all in the same ratio r, it then follows that the area under the curve above any two intervals of the form [a,ra] and [b,rb] are equal. Following definition 2, let L(x) denote the area under this from position 1 to position x. (If x is less than 1, deem the area negative.) Notice that L(1) = 0. Also, set e to be the location on the xaxis where the area under the curve is 1: L(e) = 1. Notice that the area under the curve from 1 to position ab, L(ab), is the sum of the areas under the curve above the intervals [1,a] and [a,ab]. The first area is L(a) and the second, by the property above, equals L(b). We thus have:
SERIES.
EULER (1707–83) introduced a number, today denoted e, that plays a fundamental role in studies of compound INTEREST, TRIGONOMETRY, LOGARITHMS, and CALCULUS, and that unites these disparate fields. (EULER’S FORMULA, for instance, illustrates this.) The number e has approximate value 2.718281828459045 … and can be defined in any of the following different ways: 1. The number e is the limit value of the expression 1 1 + raised to the nth power, as n increases n indefinitely: 1 e = lim n→∞ 1 + n
n
2. If L(a) denotes the area under the curve y = 1/x above the interval [1,a], then e is the location on the xaxis for which L(e) = 1. 3. If f(x) is a function that equals its own DERIVATIVE, d that is, –– f(x) = f(x), then f(x) is an EXPONENTIAL dx FUNCTION with base value e: f(x) = ex. 1 1 1 4. e is the value of the infinite sum 1 + – + – + – 1! 2! 3! + ….
L(ab) = L(a) + L(b) This shows that L is a function that converts multiplication into addition, which is enough to prove that it is the LOGARITHMIC FUNCTION base e. We have L(x) = loge(x). This function is called the natural logarithm function and is usually written ln(x).
Definition 1 is linked to the problem of computing compound interest. As we show below, definition 2 defines the natural logarithm, and definition 3 arises from studies of natural growth and decay, and consequently the consideration of EXPONENTIAL FUNCTIONs. 152
e
153
By the FUNDAMENTAL THEOREM OF CALCULUS, the derivative of an area function is the original function and so we have: d 1 –– ln(x) = – dx x Now consider the corresponding exponential function y = ex. By taking logarithms, we obtain ln(y) = x. dy dy 1 –– x –– Differentiating yields – y dx = 1, and so dx = y = e . This establishes definition 3 stating that f(x) = ex is the function that equals its own derivative. Now that we know the derivative of y = ex, we can compute its Taylor series. We obtain: ex = 1 + x +
x 2 x3 + +L 2! 3!
Setting x = 1 establishes definition 4. It remains now to establish definition 1. From the graph of the curve y = 1/x, it is clear that the region 1 between x = 1 and x = 1 + – n is sandwiched between a 1 1 1 = rectangle of area n × 1 n + 1 and a rectangle of 1+ n 1 1 – area = – n × 1 = n . Thus: 1 1 1 ≤ ln1 + ≤ n +1 n n Multiplying through by n yields: n n 1 ≤ ln 1 + ≤ 1 n +1 n
n As n becomes large, the quantity n–– + 1 approaches n
1 the value 1. It must be the case, then, that 1 + n approaches a value for which its logarithm is 1. n 1 Consequently lim n→∞ ln 1 + equals 1. This gives: n
The curve y = 1/x
With regard to the issue of compound interest, it n
1 lim n→∞ 1 + = e n
n
r is necessary to compute the limit lim n→∞ 1 + n .
154 Earth We have: r
(n/ r ) 1 r = er lim n→∞ 1 + = lim n→∞ 1 + n n / r) ( n
This represents the value of an investment of $1 after 1 year accruing continuously compounded interest at an interest rate of r percent per annum. One can use the fourth definition to prove that e is an IRRATIONAL NUMBER. One begins by assuming, to p the contrary, that e is a fraction of the form – q! and multiplies the formula presented in definition 4 by q!. The fractional parts that remain cannot add to a whole number. In 1873 French mathematician Charles Hermite (1822–1901) proved that e is a TRANSCENDENTAL NUMBER.
Earth Our planet, the third from the Sun, is often assumed to have the shape of a perfect SPHERE, but detailed measurements show it to be the shape of an oblate spheroid with equatorial radius 3,963 miles (6,378 km) and polar radius 3,950 miles (6,357 km). Its mean orbital distance from the Sun is 9.296 × 107 miles (1.496 × 108 km), one “astronomical unit,” and its mass is 1.317 × 1025 lb (5.976 × 1024 kg). The Earth takes 23 hours, 56 minutes, and 4.1 seconds to complete a sidereal day, that is, one full rotation about its axis as measured relative to the fixed stars. The solar day, the time it takes for a point P on the surface of the Earth initially facing the Sun to return to that position is, by definition, precisely 24 hours. The difference in time measurements is explained by the fact that the Earth advances in its orbit as it rotates: the Earth must turn slightly more than 360° to bring P back to face the Sun as the planet moves forward. The first known attempt to calculate the circumference of the Earth was made by the Greek scholar ERATOSTHENES OF CYRENE (ca. 275–195 B.C.E.). Eratosthenes observed that at the summer solstice, noon on June 21 of every year, the Sun shone directly to the bottom of a well in the city of Syene (present day Aswan). This meant that the Sun was directly
overhead at this time. He also noted that at the same time in Alexandria, a city approximately 500 miles due north of Syene, objects cast shadows, meaning that the Sun was not directly overhead at this location. These observations provided Eratosthenes the means to compute the radius of the Earth. He assumed that the Sun was sufficiently far away from our planet that the rays of light that reach us from it can be regarded as essentially parallel. Eratosthenes measured the angle cast by shadows at Alexandria (by using an object of known height and measuring the length of the shadows cast) and found the angle to be 1/50 of a full turn, about 7.2°. The circumference of the Earth follows: if 1/50th of a full turn corresponds to a distance of 500 miles, then a full turn must correspond to a distance of 50 × 500 = 25,000 miles. Geographical Coordinates The lines of latitude and longitude form a system of COORDINATES on the Earth. The “parallels of latitude” are circles parallel to the equator, labeled according to a measurement of angle. Stated precisely, if O is the center of the Earth, P is a point on a circle of latitude, P′ is the point on the equator directly south of P, and α is the angle POP′, then the circle of latitude containing P is labeled α. The latitude of any point can thus be 0° (on the equator) up to 90° north (the North Pole) or 90° south (the South Pole). New York City, for example, has a latitude of 40°45′06′′ N. Conveniently, the North Star lies directly above the North Pole. By measuring the angle of elevation of the North Star (90° at the North Pole, 0° at the equator), one can quickly determine the latitude of any location in the Northern Hemisphere. Using sextants to measure elevation, sailors of the past relied on the North Star to help determine their locations. The meridians, or lines of longitude, run from the North Pole to the South Pole, perpendicular to the circles of latitude. Each meridian is a semicircle. The equator is divided into 360°, with the meridian passing through Greenwich, England (called the prime meridian), deemed angle zero, and the angle of any other meridian ranges from 180° east to 180° west. Stated precisely, the longitude of a point P on the sphere is given by the angle P′OG′, where P′ is the point on the equator directly south or
Egyptian mathematics 155 duces a new fraction with a smaller numerator. Thus repeated application of this procedure must eventually produce a fraction with unit numerator itself. As an example, for the fraction 5/17 we have:
north of P, and G′ is the point on the equator south of Greenwich. New York City, for example, has longitude 73°59′39′′ W. The longitude of a point P on the Earth’s surface can be measured by identifying the time difference between high noon at P and high noon at Greenwich. As the day is divided into 24 1hour periods, each delay of 1 hour corresponds to 1/24 of a full turn about the Earth’s circumference. Thus, at 360/24 = 15° longitude west, for instance, the Sun reaches its highest point in the sky 1 hour later than it does at Greenwich; at 30° longitude west, it occurs 2 hours later, and so forth. After about 1735, sailors were able to carry accurate chronometers to keep track of Greenwich time. By measuring the time of high noon at any given location, sailors could accurately determine the longitude of that location. See also OBLATE/PROLATE.
Such representations need not be unique. For example, 3 equals both – 1 + – 1 + – 1 and – 1 + – 1. – 10 5 15 30 4 20 See also EGYPTIAN MATHEMATICS.
eccentricity For any
Egyptian mathematics Our knowledge of ancient
CONIC SECTION,
the ratio of the distance of a point on the curve from a fixed point, the focus, to its distance from a fixed line, the directrix, is constant. The value of this ratio, denoted e, is called the eccentricity of the curve, and it gives a measure of the curve’s shape. For a PARABOLA, e equals one. If e lies between zero and one, then the curve is an ELLIPSE. If e is greater than one, then the curve is a HYPERBOLA. The eccentricity of a CIRCLE is defined to be zero. (In this context, the eccentricity e is not to be confused with LEONHARD EULER’s number e.)
Egyptian fractions Any fraction with unit numerator, 1 – 1 1 – such as – 2 , 45 , and 598 , is called an Egyptian fraction. The Egyptians of 4,000 years ago expressed all fractional quantities as sums of distinct Egyptian fractions. 2 1 – 1 3 1 – 1 – – – For example, – 9 was written 6 + 18 , and 10 as 5 + 15 1 – + 30 . In 1202, FIBONACCI began his own investigation of Egyptian fractions and was the first to prove that every fraction can indeed be expressed as a finite sum of distinct Egyptian fractions. (It is not clear whether the ancient Egyptians ever questioned this.) He showed that subtracting a quantity of the form 1/n, with n as small as possible, from a given fraction always pro
5 –– 1 = – 3 (– 1 is too large a quantity to subtract.) – 17 4 68 3 and 3 – – 1 = –— 1 (– 1 is too large a quantity to subtract.) – 68 23 1564 22 giving: 5 =– 1 + – 1 + –— 1 – 17 4 23 1564
Egyptian mathematics from around 2000 B.C.E. comes chiefly from the RHIND PAPYRUS. There we learn, for example, that the Egyptians followed a very natural system for denoting numerals: 1 was a vertical stroke , 2 was two of them , 3 was , and 4 was , and separate symbols were used for 5, 6, 7, 8, and 9, and for 10, 20,…, 100, 200,…, 1000, and so on. All other numbers were represented as groups of these symbols, usually arranged in order from largest to smallest. Like the ROMAN NUMERAL system, the Egyptian system did not use a PLACEVALUE SYSTEM (the symbol for 5, for example, denoted “5” no matter where it appeared in the number). It is very difficult to do pencilandpaper calculations without placevalue notation, but the Egyptians always used a calculating board, much like an ABACUS, to perform arithmetic calculations, and needed only to record the results. They were therefore not hindered by their cumbersome numerical system. The ancient Egyptians were adept at multiplication, using a method of successive doubling to calculate products. This method is today called EGYPTIAN MULTIPLICATION. Division problems lead to FRACTIONs. It did not occur to the ancient Egyptians to express fractions with numerators and denominators. In the Rhind papyrus, the mathematician Ahmes simply placed a dot over a number to indicate its reciprocal, except in the case of 1 ,– 1 ,– 2 , and – 1 , each of which had its the fractions – 2 3 3 4
156 Egyptian multiplication own symbol. Thus the Egyptians only dealt with frac1 (with the exception of twothirds). tions of the form – n Fractions with unit numerators are known today as EGYPTIAN FRACTIONS. All other fractional quantities were expressed as sums of distinct Egyptian fractions. 2 , which equals – 1 + – 1 , was written For example, – 5 3 15 • • • • • 4 – 3+ 15 , and 13 as 4+ 18+ 468 . The Egyptian’s ability to compute such expressions is impressive. The Rhind papyrus provides reference lists of such expressions, and the first 23 problems in the document are exercises in working with such fractional representations. The ancient Egyptians were adept at solving LINEAR EQUATIONs. They used a method called false position to attain solutions. This involves guessing an answer, observing the outcome from the guess, and adjusting the guess accordingly. As an example, problem 24 of the Rhind papyrus asks: Find the quantity so that when 1/7 of itself is added to it, the total is 19. To demonstrate the solution, the author suggests a guess of 7. That plus its oneseventh is 8, by far too small, but multiplying the outcome by 19/8 produces the answer of 19 that we need. Thus 7 × (19/8) must be the quantity we desire. The majority of problems in the Rhind papyrus are practical in nature, dealing with issues of area (of rectangles, trapezoids, triangles, circles), volume (of cylinders, for example), slopes and altitudes of pyramids (which were built 1,000 years before the text was written), and number theoretic problems about sharing goods under certain constraints. Some problems, however, indicate a delight in mathematical thinking for its own sake. For example, problem 79 asks: If there are seven houses, each house with seven cats, seven mice for each cat, seven ears of grain for each mouse, and each ear of grain would produce seven measures of grain if planted, how many items are there altogether? This problem appears in FIBONACCI’s Liber abaci, written 600 years before the Rhind papyrus was discovered. A version of this problem also appears as a familiar nurseryrhyme and riddle, “As I Was Going to St. Ives.”
Egyptian multiplication The RHIND
indicates that the ancient Egyptians of around 2000 B.C.E. used a process of “successive doubling” to multiply numbers. They computed 19 × 35, for example, by repeatedly doubling 35:
1 2 4 8 16
PAPYRUS
35 70 140 280 560
Since 19 = 16 + 2 + 1, summing 560 + 70 + 35 = 665 gives the product. This method shows that knowledge of the twotimes table is all that is needed to compute multiplications. RUSSIAN MULTIPLICATION follows an approach similar to this method. See also EGYPTIAN MATHEMATICS; ELIZABETHAN MULTIPLICATION; FINGER MULTIPLICATION; MULTIPLICATION; NAPIER’S BONES; RUSSIAN MULTIPLICATION.
eigenvalue (evalue, latent root) See EIGENVECTOR.
eigenvector (evector, latent vector, characteristic vector, proper vector) For a square n × n MATRIX A, we say a nonzero VECTOR x is an eigenvector for A if there is a number λ such that Ax = λx. The number λ is called the eigenvalue associated with that eigenvector. If x is an eigenvector of A, then we have that (A – λI)x = 0, where I is the IDENTITY MATRIX. This shows that the matrix A – λI is not invertible, and so must have zero determinant: det(A – λI) = 0. This is a polynomial equation in λ of degree n, called the “characteristic polynomial” of A. As there can only be at most n solutions to such an equation, we have that an n × n matrix A has at most n distinct eigenvalues. Mathematicians have proved that associated with each possible eigenvalue there is at least one corresponding eigenvector. Moreover, it has been established that eigenvectors associated with distinct eigenvalues are linearly independent. The study of eigenvectors and eigenvalues greatly simplifies matrix manipulations. Suppose, for example, a square 3 × 3 matrix A has three distinct eigenvalues λ1, λ2, and λ3. Set D to be the diagonal matrix
Einstein, Albert 157 λ1 0 D = 0 λ2 0 0
0 0 λ 3
and set B to be the matrix whose columns are the corresponding eigenvectors of A. Then it is possible to show that: A = BDB–1 This observation allows one to compute high powers of A with very little work. For instance, the quantity A100, for instance, is just a product of three matrices: A100 = (BDB–1)100 = BD100B–1
Einstein, Albert (1879–1955) German Relativity, Quantum mechanics Born on March 14, 1879, in Ulm, Germany, Albert Einstein is recognized as an outstanding mathematical physicist whose work on the special and general theories of relativity completely revolutionized how scientists think about space, matter, and time. Although he regarded himself as a physicist, Einstein’s work inspired many significant developments in modern mathematics, including the development of TENSOR analysis as the appropriate means to describe the curvature of space. The classical school environment did not suit Einstein well. In 1895 he failed the entrance exam for a Swiss technical school, where he hoped to study electrical engineering. After attending a second school at Aarau, he did eventually manage to enter the Zurich
noting that D100 is simply:
D100
λ 100 1 = 0 0
0 100 λ2
0
0 0 100 λ3
This work also allows mathematicians to define the square root of a matrix: λ1 A = B 0 0
0 λ2 0
0 0 B−1 λ 3
(provided the square roots of the eigenvalues are defined), or to define new quantities, such as the logarithm of a matrix: ln(λ 1) 0 0 ln(A) = B 0 ln(λ 2 ) 0 B−1 0 ln(λ 3 ) 0 Such actions have proved useful in the study of theoretical physics and engineering. The prefix eigen is German for “characteristic” or “own.” The eigenvalues and eigenvectors of a matrix completely characterize the matrix. See also CAYLEYHAMILTON THEOREM; INVERSE MATRIX; LINEARLY DEPENDENT AND INDEPENDENT.
Albert Einstein, an eminent mathematical physicist of the 20th century, revolutionized our understanding of space, matter, and time through his theories of special and general relativity. (Photo courtesy of Topham/The Image Works)
158 Einstein, Albert school to graduate there in 1900 with a degree in teaching. Unable to find a university position, Einstein accepted a job at the Swiss Patent Office in Bern in 1902 and remained there for 7 years. During his time at the patent office, Einstein studied theoretical physics in the evenings, without the benefit of close contact with the scientific literature or colleagues, and managed to produce and publish, all in the year 1905, five truly outstanding papers: • “Über einen die Erzeugung und Verwandlung des Lichtes betreffenden heuristischen Gesichtspunkt” (On a heuristic concerning the production and transformation of light), published in Annalen der Physik, March 1905. • “Die von der molekularkinetischen Theorie der Wärme gefurdete Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen” (On the movement of small particles suspended in a stationary liquid demanded by the molecular kinetic theory of heat), published in Annalen der Physik, May 1905. • “Zur Elektrodynamik bewegter Körper” (On the electrodynamics of moving bodies), published in Annalen der Physik, June 1905. • “Ist die Trägheit eines Körpes von seinem Energieinhalt abhängig?” (Does the inertia of a body depend upon its energycontent?), published in Annalen der Physik, September 1905. • “Eine neue Bestimmung der Moleküldimensionen” (A new determination of molecular dimension), written in April 1905, published in Annalen der Physik, April 1906. The first paper was concerned with the puzzling photoelectric effect observed by scientists of the time. Heinrich Hertz (1857–94) noticed that the number of electrons released from a section of metal bombarded with a beam of light was determined not by the intensity of the beam, but rather by its wavelength. Max Planck (1858–1947) also observed that electromagnetic energy was emitted from radiating objects according to discrete quantities, again in direct proportion to the wavelength of the radiation. Einstein proposed that light and radiation itself travel in discrete bundles, which he called quanta, and described the mathematics that would consequently explain these phenomena. In his second paper, Einstein developed mathematical equations that correctly described the motion of atoms and molecules under “Brownian motion.” In his third paper, Einstein proposed his theory of special relativity. He noted that because light is able to travel
through a vacuum, there is no natural frame of reference for measuring its speed. (The speed of sound, for instance, is measured with respect to the medium of air through which it passes.) Also, since it is impossible to determine whether one is stationary in space or moving through space at a uniform velocity, it follows then that all observers must observe light traveling at the same speed. From this, Einstein developed a series of “thought experiments” that clearly establish that observers traveling at different speeds must hence record different values when measuring quantities such as length and time. (For instance, imagine a light beam bouncing back and forth between two fixed mirrors set at a distance so that the time taken to bounce between the two mirrors is 1 sec. Suppose that a second observer moves past the “clock” with uniform speed. According to this observer, the clock moves past her at uniform speed. A straightforward calculation shows that, since the speed of light is unchanged for this observer, she would see the light beam taking longer than 1 sec to complete a cycle between the two mirrors.) The fourth paper developed the special theory of relativity further, culminating with his famous equation E = mc2, showing that energy and mass are equivalent (with a factor of the speed of light squared incorporated). Einstein submitted his final paper as a doctoral thesis to the University of Zurich to receive a Ph.D. By 1909 Einstein had been recognized as a leading scientific thinker. He resigned from the patent office and was appointed a full professor at the KarlFerdinand University in Prague in 1911. That same year, based on his theory of relativity, Einstein made a prediction about how light rays from distant stars would bend around the Sun, hoping that some day astronomers might be able to observe this effect and verify that his theory of relativity is correct. He also began working on incorporating the role of acceleration (nonuniform motion) into his special theory to develop a general theory of relativity. After a number of false starts, Einstein finally published a coherent general theory in 1915. Four years later, during a solar eclipse, British scientists were able to observe the bending of light rays just as Einstein had predicted. The popular press covered the story, and Einstein immediately received world attention for his achievement. In 1921 Einstein received the Nobel Prize not for his relativity theory, but, surprisingly, for his work on the photoelectric effect. He also received the Copley
Elizabethan multiplication 159 Medal from the ROYAL SOCIETY of London in 1925, and the Gold Medal from the Royal Astronomical Society in 1926. With the rise of antiSemitism in Europe, Einstein accepted a position at the Institute of Advanced Study at Princeton, N.J., in 1933. He stayed there until his death on April 18, 1955. One cannot exaggerate the effect that Einstein’s work has had on modern physics. One of his principal goals was to unite the discrete description of particles and matter with the continuous description of electromagnetic radiation and develop a single unified theory of the two. The result is quantum mechanics. Intense work continues today to incorporate other physical forces, such as gravity, into a grand unified theory.
elementary matrix See GAUSSIAN ELIMINATION.
The Elements No doubt the most influential mathematics text of all time, The Elements, written by EUCLID (ca. 300–260 B.C.E.), provided the model for all of mathematical thinking for the two millennia that followed it. Mathematicians agree that this work defines what the pursuit of PURE MATHEMATICS is all about. More than 2,000 editions of The Elements have been printed since the production of its first typeset version in 1482. Written in 13 volumes (called “books”) The Elements represents a compilation of all the mathematics that was known at the time. Organized in a strict logical structure, Euclid begins the work with a set of basic definitions, “common notions,” and axioms (EUCLID’S POSTULATES), and deduces from them, by the process of pure logical reasoning, some 465 propositions (THEOREMs) on the topics of plane geometry, number theory (typically presented in terms of geometry), and solid geometry. The work is revered for its clarity, precision, and rigor. The work is extremely terse in its presentation. There is no discussion or motivation, and results are simply stated and proved, often referring to a figure accompanying the statement. Each proof ends with a restatement of the proposition studied along with the words, “which was to be proved.” The Latin translation of this phrase is quod erat demonstrandum, and many mathematicians today still like to end a formal proof with the abbreviation Q.E.D.
Although it is generally believed that no result presented in The Elements was first proved by Euclid, the organization of the material and the logical development presented is original. Euclid’s choice of beginning postulates shows remarkable insight and a deep wisdom of the subject. His recognition of the need to formulate the controversial PARALLEL POSTULATE, for instance, shows a level of genius beyond all of those who tried to prove his choice irrelevant during the two millennia that followed. (It was not until the 19th century, with the development of NONEUCLIDEAN GEOMETRY, did mathematicians realize that the parallel postulate was an essential assumption in the development of standard planar geometry.) The first six books of The Elements deal with the topic of plane geometry. In particular, Books I and II establish basic properties of triangles, parallel lines, parallelograms, rectangles, and squares, and Books III and IV examine properties of circles. In Book V, Euclid examines properties of magnitudes and ratios, and applies these results back to plane geometry in Book VI. Euclid presents a proof of PYTHAGORAS’S THEOREM in Book I. Books VII to X deal with NUMBER THEORY. The famous EUCLIDEAN ALGORITHM appears in book VII, and EUCLID’S PROOF OF THE INFINITUDE OF PRIMES in book IX. Book X deals with the theory of irrational numbers, and Euclid actually proves the existence of these numbers in this work. The final three volumes of The Elements explore threedimensional geometry. The work culminates with a discussion of the properties of each PLATONIC SOLID and proof that there are precisely five such polyhedra.
elimination method for simultaneous linear equations Another name for GAUSSIAN ELIMINATION.
Elizabethan multiplication Also known as the galley method and the lattice method, this multiplication technique was taught to students of mathematics in Elizabethan England. To multiply 253 and 27, for example, draw a 2 × 3 grid of squares. Write the first number along the top, and the second number down the right side. Divide each cell of the grid diagonally. Multiply the digits of the top row, in turn, with each of the digits of the right column, writing the products in
160 ellipse Using the notation PF1 and PF2 for the lengths of the line segments connecting P to F1 and F2, respectively, the defining condition of an ellipse can be written: PF1 + PF2 = d
Elizabethan multiplication
the appropriate square cells of the grid as twodigit entries. (Thus compute 3 × 2 as 06, for example.) Add the entries in each diagonal starting with the bottom right diagonal. Write down the units figure and carry any tens figures that appear to the next diagonal. The answer, 6,831, now appears along the left column and bottom row. This procedure works for multidigit multiplications of any size. Its success relies on the DISTRIBUTIVE PROPERTY of arithmetic and the process of EXPANDING BRACKETS. In our example, 253 × 27 = (2 × 102 + 5 × 10 + 3) × (2 × 10 + 7) = (2 × 2) × 103 + (2 × 7) × 102 + (3 × 2) × 10 + (5 × 2) × 102 + (5 × 7) × 10 +3×7
where d denotes the constant sum. The equation of an ellipse can be found by introducing a coordinate system in which the foci are located at positions F1 = (–c,0) and F2 = (c, 0), for some positive number c. It is convenient to write d = 2a, for some a > 0. If P = (x,y) is an arbitrary point on the ellipse, then the defining condition states: √(x + c)2 + y2 + √(x – c)2 + y2 = 2a Moving the second radical to the righthand side, squaring, and simplifying yields the equation: (x − c)2 + y 2 = a −
Squaring and simplifying again yields: y2 x2 – =1 2 + –– 2 a a – c2 Noting that a is greater than c, we can set the positive quantity a2 – c2 as equal to b2, for some positive number b. Thus the equation of the ellipse is: y2 x2 – – + 2 =1 a2 b
Each diagonal corresponds to a powersof10 place, with entries placed in an upper portion of a square cell corresponding to carried figures to the next powersof10 position. Try computing a multiplication problem both the Elizabethan way and the usual way, sidebyside, to see that the two methods do not differ. See also EGYPTIAN MULTIPLICATION; FINGER MULTIPLICATION; MULTIPLICATION; NAPIER’S BONES; RUSSIAN MULTIPLICATION.
ellipse As one of the
CONIC SECTIONS, an ellipse is the plane curve consisting of all points P whose distances from two given points F1 and F2 in the plane have a constant sum. The two fixed points F1 and F2 are called the foci of the ellipse. An ellipse also arises as the curve produced by the intersection of a plane with a single nappe of a right circular CONE.
c x a
Ellipse
ellipsoid 161 Conversely, one can show that any equation of this form, with a > b, does indeed yield an ellipse with foci at positions (± √a2 – b2 , 0), and whose points P have distances from the foci a constant sum 2a. If, on the other hand, b > a, the equation is again an ellipse, but this time with foci along the yaxis at (0,± √b2 – c2 ). The common sum of distances is 2b. (If a equals b, the curve is a CIRCLE.) The equation shows that an ellipse crosses the xaxis at x = ±a and the yaxis at y = ±b. The numbers a and b are called the semimajor axis and the semiminor axis, respectively. Ellipses have the following reflection property: any ray of light emanating from one focus is reflected off the side of the ellipse directly toward the other focus. This can be proved by solving an OPTIMIZATION problem. Any room with walls curved in the shape of an ellipse has the property that any whisper uttered at one focus can be heard by anyone located at the second focus: not only do sound waves bounce off the curved wall directed from one focus to the other, but they also travel the same distance and so arrive synchronized at the second focus. The Mormon Tabernacle in Salt Lake City, Utah, and the Whispering Gallery in the U.S. Capitol building in Washington, D.C., are built to have this property. Elliptical mirrors are used for the treatment of kidney stones. By positioning a mirror so that the kidney stone lies at one focus, medical practitioners can place a highintensity sound wave generator at the second focus. Waves from the generator pass harmlessly through the patient’s body to then concentrate at the stone and destroy it. An ellipse can be drawn using a pencil, a string and two thumbtacks. Tacking each of the two ends of the string at fixed locations (the foci), one pulls the string taut with the tip of the pencil, and then slowly moves the pencil around, all the while keeping the string taut. The curve traced is an ellipse, with constant sum of the distances from the foci being the length of the string. In the process of deriving the equation of an ellipse, we presented the equation (x − c)2 + y 2 = a −
Set
e=
c = a
c x a
a 2 − b2 b2 = 1 − 2 . This is called the a a
ECCENTRICITY of the ellipse and has value between zero and 1. The above equation can be rewritten:
(x − c)2 + y 2 a x − − e
=e
The numerator of the quantity on the left side is the distance of a given point P from a focus, and the denominator is the distance of the point P from the vertical line x = –a/e, called a directrix of the ellipse. This formulation provides an alternative characterization of the ellipse: An ellipse is the set of all points P such that the ratio of its distance from a fixed point (the focus) to its distance from a fixed line (the directrix) equals a constant e with value 0 < e < 1. The ECCENTRICITY of a circle is defined to be e = 0. If e = 1, this characterization gives a PARABOLA. For e > 1, we have a HYPERBOLA. See also APOLLONIUS’S CIRCLE; PROJECTION.
ellipsoid Any geometrical surface or solid sitting in threedimensional space possessing the property that any plane that slices it produces a crosssection that is either an ELLIPSE or a circle is called an ellipsoid. Such a figure has three axes of symmetry. An ellipsoid, centered about the origin (0,0,0) has equation: y2 z2 x2 + – – + –2 = 1 2 a2 c b The points (±a,0,0), (0,±b,0) and (0,0,+c) are the locations where the ellipsoid crosses the x, y, and zaxes, respectively. One can create an ellipsoid by rotating an ellipse about one of its axes. This produces a figure with two of the three quantities a, b, c equal in value. An ellipsoid produced in this way is called a spheroid, but not every ellipsoid is a spheroid. If all three quantities a, b, and c have the same value r, the ellipsoid is a SPHERE of radius r. Mathematicians have shown that the volume of an ellipsoid is given by (4/3)πabc. (Compare this with the equation for the volume of a sphere.) Since the time of LEONHARD EULER (1707–83), mathematicians have attempted to find a simple formula for the surface area
162 empty set of a general ellipsoid. This has proved to be a very difficult problem, and no closedform formula exists. (The surface area of an ellipse can only be expressed in terms of a difficult “elliptic integral.”)
empty set (null set) Any set that contains no elements is called an empty set. For example, the set of all real numbers greater than three and less than two is empty, as is the set of all people with gills. A set A is said to be a subset of a set B, written A ⊃ B, if all elements of A belong to B. Consequently any empty set is a subset of any other set. In particular, if A and B are both empty, then A ⊃ B and B ⊃ A, and the two empty sets are equal. This shows that there is only one empty set. It is usually denoted as Ø, but it can also be written as { }. The set with the empty set as its one member is written {Ø}, and the set with the set containing the empty set as its lone member is written {{Ø}}. In this way we construct a chain of sets: Ø, {Ø}, {{Ø}}, {{{Ø}}},… which naturally corresponds to the sequence of counting numbers 0, 1, 2, 3, … In this context one could argue that all of mathematics arises from the empty set. It is an interesting exercise then to give a numerical interpretation to a twomember set of the form {Ø,{Ø}}, for instance. A set that is not empty is called nonempty. See also SET THEORY.
endpoint See INTERVAL.
epicycle See CYCLOID.
used to denote the equivalence of two quantities, and so we write 2+3 = 5 and {a,b,c} = {c,a,b}. Two algebraic expressions are said to be equal if one can be transformed into the other by the standard rules of algebra. For instance, (x + 1)2 + 3 = x2 + 2x + 4. Two functions are said to be equal if they have the same domains and produce the same output value for each input. For example, the functions f(x) = 9x and g(x) = x
2 + log 3 x log 3 x
, defined on positive values of x, are
equal. The symbol = (a pair of parallel line segments to denote equality) was introduced in 1557 by Welsh mathematician ROBERT RECORDE (ca. 1510–58) “because noe 2 thynges can be more equalle.” See also EQUATION.
equating coefficients Two polynomials f(x) = anxn + an–1xn–1 + … + a1x + a0 and g(x) = bnxn + bn–1xn–1 + … + b1x + b0 are identical as functions, that is, give the same output values for each input value of x, only if the coefficients of the polynomials match: an = bn, an–1 = bn–1,…,a0 = b0. (The general study of POLYNOMIALs establishes this.) The process of matching coefficients if two polynomials are known to be the same is called “equating coefficients.” For example, if x2 equals a polynomial of the form A + B(x – 1) + C(x – 1)(x – 2), then, after EXPANDING BRACKETS, we have x2 = Cx2 + (B – 3C)x + (A – B + 2C). Equating coefficients yields: C = 1, B – 3C = 0 (and so B = 3), and A – B + 2C = 0 (and so A = 1). Thus x2 = 1 + 3(x – 1) + (x – 1)(x – 2). (This technique is often used in the method of PARTIAL FRACTIONS.) As another example, if α and β are the roots of a quadratic equation of the form x2 – mx + n, then: x2 – mx + n = (x – α) (x – β) = x2 – (α + β)x + αβ. We conclude then that m is the sum of the roots, and n their product.
epsilondelta definition See LIMIT. equating real and imaginary parts Two equality Two quantities are said to be equal if, in some meaningful sense, they are equivalent. For example, the quantities 2 + 3 and 5 have the same value and so are equal. The two sets {a,b,c} and {c,a,b} are equal since they contain the same elements. The symbol = is
COMPLEX
NUMBERS a + ib and c + id are equal only if a = c and b = d. Using this fact is called “equating real and imaginary parts.” For example, if (x + iy)(2 + 3i) = 4 + 5i, then we must have 2x – 3y = 4 and 3x + 2y = 5. LEONHARD EULER (1707–83) made clever use of this technique to find formulae for PYTHAGOREAN
equation of a line 163 Also, using his famous formula eiθ = cos(θ) + isin(θ), today called EULER’S FORMULA, many trigonometric identities can be established quickly by equating real and imaginary parts. As another application, TRIPLES.
∞
consider the series
∑ cos(nθ ) . Since cos(nθ) is the real
n =0
part of einθ, the series in question is the real part of ∞
the GEOMETRIC SERIES ∞
∑ (e iθ )
n
=
n −0
∞
n =0
n
. Evaluating, gives
1 1 = 1 − e iθ (1 − cos(θ)) − i sin(θ) =
Thus we have
∑ (e iθ )
sin(θ) 1 +i 2 2 − 2 cos(θ) 1
∑ cos(nθ ) = 2 .
n =0
equation A mathematical statement that asserts that one expression or quantity is equal to another is called an equation. The two expressions or quantities involved are connected by an equals sign, “=.” For instance, the statement (a + b)2 = a2 + 2ab + b2 is an equation, as are the statements 2x + 3 = 11 and 10 = 2 × 5. An equation that is true for all possible values of the variables involved is called an IDENTITY. For instance, y2 – 1 = (y – 1)(y + 1) is an identity: this equation is true no matter which value is chosen for y. An equation that is true only for certain values of the variables is called a conditional equation. For instance, the equation 2x + 3 = 11 is conditional, since it is true only if x is four. The equation 2c + d = 6 is also a conditional, since it holds only for certain values of c and d. The numbers that make a conditional equation true are called the solutions or roots of the equation. For example, the solution to 2x + 3 = 11 is x = 4. An equation may possess more than one solution, and the set of all possible solutions to a conditional equation is called its solution set. For example, the equation y2 = 9 has solution set {3, – 3}, and the solution set of the equation 2c + d = 6 is the set of all pairs of numbers of the form (c, 6 – 2c). The basic principle in solving an equation is to add, subtract, multiply, or divide both sides of the equation by the same number until the desired variable is isolated on one side of the equation. For example, the equation 2x + 7 = 5x + 1 can be solved by subtract
ing 1 from both sides to give 2x + 6 = 5x, then subtracting 2x from both sides to obtain 6 = 3x, and, finally, dividing both sides by 3 to obtain the solution x = 2. This approach works well for LINEAR EQUATIONs of one variable. For QUADRATIC equations, and POLYNOMIAL equations of high degree, one may also be required to take square and higher roots in the process of solving the equation. Not all equations, however, can be solved algebraically, in which case one can seek only a GRAPHICAL SOLUTION. It should be noted that performing the same arbitrary operation on both sides of an equation need not necessarily preserve the validity of the equation. For example, although the statement 9 = 9 is certainly valid, taking a square root on both sides of this trivial equation could be said to yield the invalid result –3 = 3. Although 2(x – 1) = 3(x – 1) is true for the value x = 1, dividing through by the quantity x – 1 yields the 12 3 – = – invalid conclusion 2 = 3. And finally, since 16 4, selecting the numerator of each side of the equation yields the absurdity 12 = 3. Care must be taken to ensure that the operations being used in solving an equation do indeed preserve equality. Even if the application of the same operation on both sides on an equation is deemed valid, such an act may nonetheless yield a new equation not necessarily exactly equivalent to the first. For instance, starting with a = b, squaring both sides yields the equation a2 = b2, which now means a = b or a = –b. Mathematicians use the symbol “⇒” to denote that one equation leads to a second, but that the second need not imply the first. For example, it is appropriate to write: a = b ⇒ a2 = b2. (But a2 = b2 ⇒ / a = b.) The symbol “⇔” is used to indicate that two equations are equivalent, that is, that the first implies the second, and that the second implies the first. For example, we have: 4x = 12 ⇔ x = 3. See also CUBIC EQUATION; HISTORY OF EQUATIONS AND ALGEBRA (essay).
equation of a line A straight
in twodimensional space has the property that the ratio of the difference in ycoordinates of any two points on the line (rise) to the difference of their xcoordinates (run) is always the same. That is, the SLOPE of a straight line is constant and can be computed from any two given points on the line. Precisely, if (a1,b1) and (a2,b2) are LINE
164 equation of a plane two points on the line, and (x,y) are the coordinates of an arbitrary point on the line, then we have: y – b1 b2 – b1 –— x – a1 = ––— a2 – a1 This provides the “twopoint form” equation of the line. For example, the equation of the line passing through y–3 5–3 2 –— – – the points (2,3) and (–1,5) is –— x – 2 = –1 – 2 = 3 . b2 – b1 The quantity ––— a2 – a1 is the slope m of the line. Thus one can rewrite the twopoint form of the equation y – b1 as –— x – a1 . Rearranging yields:
Some authors prefer to divide the general form of the equation of a line through by the constant r and change the names of the remaining labels so that the equation reads: x x — a + — b =1 This is called the intercept form of the equation of a line. In threedimensional space a line is specified by a point (a,b,c) on the line and a VECTOR by v = , representing the direction of the line. Thus the coordinates (x,y,z) of any other point on the line are given by: x = a + tv1 y = b + tv2 z = c + tv3
y – b1 = m(x – a1) This is called the “pointslope form” equation of the line. For example, the equation of a line of slope 4 that passes through the point (5,7) is simply y – 7 = 4(x – 5). Working with this form of equation is useful if the slope of the line is already specified. Rearranging the pointslope equation yields y = mx + (ma1 + b1). Denoting the constant ma1 + b1 simply as b yields the equation:
for some value of the real number t. These are the PARAMETRIC EQUATIONS of the line. (The parametric equations of a line in twodimensional space are analogous.) If the vector v is computed via the difference of coordinates of the point (a,b,c) and a second point (a1,b1,c1) on the line—i.e., v1 = a1 – a,, v2 = b1 – b, and v3 = c1 – c—then solving for t in the parametric equations yields: y–b z–c x–a –— –— a1 – a = –— b1 – b = c1 – c
y = mx + b Noting that if x = 0, then we have y = b. This shows that the constant b is the yintercept of the line. For this reason, the above equation is called the “slopeintercept form” equation of the line. Thus, for example, the equation of a line with slope –1 crossing the yaxis at position 3 is y = –x + 3. One disadvantage of the slopeintercept form is that it does not allow one to write down the equation of a vertical line, that is, one that does not intercept the yaxis at all. Returning to the twopoint form and crossmultiplying yields the equation (a2 – a1)(y – b1) = (x – a1)(b2 – b1). EXPANDING BRACKETS and rearranging terms again yields the general equation: cx + dy = r for some constants c, d, and r. This is called the general form of the equation of a line. For instance, the equation of the vertical line three units to the right of the yaxis is obtained by selecting c = 1, d = 0, and r = 3, to yield the equation x = 3.
These are the “twopoint form” equations of a line in threedimensional space. See also DIRECTION COSINES; LINEAR EQUATION; SIMULTANEOUS LINEAR EQUATIONS; SKEW LINES; VECTOR EQUATION OF A PLANE.
equation of a plane See
VECTOR EQUATION OF A
PLANE.
equiangular A
is said to be equiangular if all of its interior angles are equal. For example, a rectangle is equiangular (each interior angle equals 90°), as is an EQUILATERAL triangle (each interior angle equals 60°). A polygon is called regular if it is both equiangular and equilateral. A point (x,y) in the Cartesian plane is said to be a lattice point if both x and y are integers, and a polygon drawn in the plane is said to be a lattice polygon if its POLYGON
equidistant 165 vertices lie at lattice points. Mathematicians have proved that it is impossible to draw an equiangular lattice polygon with n sides if n is a number different from 4 or 8. (Any foursided equiangular lattice polygon is a rectangle, and any eightsided equiangular lattice polygon has eight interior angles, each equal to 135°.) The square and the octagon are the only two regular lattice polygons. See also CARTESIAN COORDINATES.
equidecomposable Two geometric figures are said to be equidecomposable if it is possible to dissect one figure into a finite number of pieces that can be rearranged, without overlap, to form the second figure. For example, an equilateral triangle of sidelength 1 is equidecomposable with a square of the same area. In the picture, a is of length
3− 4 3 and b is of 4
3 −1 . (The challenge to convert an equilateral 4 triangle into a square by dissection was a puzzle first posed by English puzzlist Henry Ernest Dudeney in 1907. The challenge is also known as Haberdasher’s puzzle.) Scottish mathematician William Wallace (1768–1843) proved that any two polygons of the same area are equidecomposable. Mathematicians have since proved that the result remains valid even for figures with curved boundaries. In particular, Hungarian mathematician Miklov Laczovich demonstrated in 1988 that almost 1050 pieces are needed to convert a circle into a square. Surprisingly, the corresponding result in three dimensions does not hold, even for simple polyhedra. 4
length
German mathematician Max Dehn (1878–1952) proved, for instance, that a cube and a regular tetrahedron of the same volume are not equidecomposable.
equidistant Two points P and Q are said to be equidistant from a third point O if they are the same distance from O. We write: PO = QO. Given a single point O in a plane, the set of all points equidistant from O is a CIRCLE with O as its center. Given two points A and B in a plane, the set of all points equidistant from A and B is the perpendicular bisector of the line segment AB, that is, a straight line perpendicular to AB and passing through the midpoint of AB. (To see this, let M be the midpoint of the line segment AB, and let P be any point in the perpendicular bisector to AB. Suppose that PM = x and AM = y = MB. Then, by PYTHAGORAS’S THEOREM, we have PA = √x2 + y2 = PB, and so P is equidistant from A and B. One can also use Pythagoras’s theorem to check that any point not on this line is not equidistant from those two points.) Given three points A, B, and C in a plane, not in a straight line, there is just one point P equidistant from all three. (To see this, draw the perpendicular bisectors of AB and BC, and let P be the unique point at which they intersect. Then P is equidistant from A and B, and P is also equidistant from B and C. Consequently, P is the same distance from all three points.) Noting that the points A, B, and C can be viewed as the vertices of a TRIANGLE, this proves: The three perpendicular bisectors of the sides of any triangle meet at a common point P. (This observation is used to prove that the three ALTIof any triangle are also CONCURRENT.) Taking matters further, suppose the common distance of P from each of the three points A, B, and C is r. It then follows that a circle of radius r centered about P passes through each of these points. This proves: TUDEs
For any triangle ABC there exists a single circle that passes through each of its vertices A, B, and C.
Equidecomposable figures
This circle is called the CIRCUMCIRCLE of the triangle, and the point P, the common point of intersection of the three perpendicular bisectors of the triangle, is called the circumcenter of the triangle.
166 equilateral In threedimensional space, the set of all points equidistant from a single point O is a sphere with O as its center. Given two points A and B, the set of all points equidistant from them is a plane that passes through the midpoint of the line segment AB and is perpendicular to it. Given three points A, B, and C, not all in a straight line, the set of all points P equidistant from all three is a straight line perpendicular to the plane formed by the three points and passing through the circumcenter of the triangle ABC. There need not be a point equidistant from four given points in threedimensional space. The distance of a point from a line is the length of a line segment from the point meeting that line at right angles. Again, using Pythagoras’s theorem and similar triangles, one can show that the set of all points equidistant from two intersecting lines in a plane is a pair of perpendicular lines that each pass through the point of intersection of the two lines, and each bisects an angle formed by the lines. Furthermore, arguing as above, one can prove: In any triangle, the three lines that bisect the interior angles of the triangle meet at a common point. (Consider first the point of intersection of just two angle bisectors. This point must, in fact, be equidistant from all three sides of the triangle, and so lies on the third angle bisector.) See also EULER LINE; INCIRCLE.
equilateral A POLYGON is said to be equilateral if all of its sides have the same length. For example, a square is equilateral, as is a triangle with each interior angle equal to 60°. A polygon is called “regular” if it is both equilateral and EQUIANGULAR. A point (x,y) in the Cartesian plane is said to be a “lattice point” if both x and y are integers, and a polygon drawn in the plane is said to be a “lattice polygon” if its vertices lie at lattice points. Mathematicians have proved that it is impossible to draw an equilateral lattice polygon with an odd number of sides, although equilateral lattice polygons with any even number of sides do exist. The square and the octagon are the only two regular lattice polygons. See also CARTESIAN COORDINATES.
equivalence relation See PAIRWISE DISJOINT.
Eratosthenes of Cyrene (ca. 275–195 B.C.E.) Greek Geometry, Number theory, Astronomy, Geographer Born in Cyrene, in North Africa, (the exact birth date is not known), Eratosthenes is remembered as the first person to calculate the circumference of the Earth. (See EARTH.) Using the known distance between two particular cities, the lengths of shadows cast by the noonday sun at those cities, and simple geometric reasoning, Eratosthenes determined the circumference of the Earth to be 250,000 “stadia.” Unfortunately, the exact length of a “stade” is not known today, and so it is not possible to be certain of the accuracy of this result. If we take, as many historians suggest, that the likely length of this unit is 515.6 ft (157.2 m), then Eratosthenes’ calculation is extraordinarily accurate. Eratosthenes traveled to Athens in his youth and spent many years studying there. Around 240 B.C.E. he was appointed librarian of the greatest library of the ancient world, the Library of Alexandria. Early in his scholarly career, Eratosthenes wrote the expository piece Platonicus as an attempt to explain the mathematics on which PLATO based his philosophy. Although this work is lost today, scholars of later times referred to it frequently and described it as an invaluable source detailing the mathematics of geometry and arithmetic, as well as the mathematics of music. In this work, Eratosthenes also described the problem of DUPLICATING THE CUBE and provided a solution to it making use of a mechanical device he invented. Eratosthenes also worked on the theory of PRIME numbers and discovered a famous “sieve” technique for finding primes. This method is still used today and is named in his honor. Along with measuring the circumference of the Earth, Eratosthenes also devised ingenious techniques for determining the distance of the Earth from the Sun (which he measured as 804 million stadia), the distance between the Earth and the Moon (780,000 stadia), and the tilt of the Earth’s axis with respect to the plane in which the Earth circles the Sun (which he measured as 11/83 of 180°, that is, 23° 51′ 15′′). Eratosthenes also accurately mapped a significant portion of the Nile River and correctly identified the occurrence of heavy rains near its source as the reason for its erratic flooding near its mouth. He compiled an astronomical cata
error 167 log listing over 675 specific stars and devised an effective calendar system that included leap years. As an extraordinarily wellrounded scholar, Eratosthenes wrote literary works on the topics of theater and ethics, and also wrote poetry. His famous poem “Hermes” was inspired by his studies of astronomy. It is believed that Eratosthenes stayed in Alexandria for the entire latter part of his life. The exact date of his death is not known.
Erdös, Paul (1913–1996) Hungarian Discrete mathematics, Number theory Born on March 25, 1913, in Budapest, Hungary, prolific mathematician Paul Erdös (pronounced “airdish”) is remembered as one of the greatest problemsolvers and problem posers of all time. With an uncanny ability to create problems that led to productive new areas of mathematics research, Erdös is credited as founder of the field of “discrete mathematics,” the mathematics of computer science. With no permanent home, Erdös traveled the globe multiple times throughout his life, collaborating and writing papers with scholars from all countries. His colleagues invented the term Erdös number to describe their close connections to him, assigning an Erdös number of 1 to all those who had coauthored a paper with Erdös, the number 2 to those who had worked with someone who had worked with Erdös, and so on. According to his obituary in the New York Times, 458 mathematicians can claim an Erdös number of 1, and over 4,500 scholars an Erdös number of 2. Erdös entered the University of Budapest in 1930 at the age of 17, and within just a few years, he began making significant contributions to the field of NUMBER THEORY. At age 20 he discovered a new and elementary proof of conjecture of Joseph Bertrand (1822–1900), stating that at least one PRIME lies between any number n and its double 2n. (Russian mathematician PAFNUTY LIVOVICH CHEBYSHEV established the validity of this claim, by complicated means, in 1850.) Later in life, Erdös also found an elementary proof of the famous PRIME NUMBER THEOREM. In 1934, at the young age of 21, Erdös was awarded a Ph.D. in mathematics from the University of Budapest. Because of his Jewish heritage, Erdös was forced to leave Hungary, and he accepted the offer of a postdoctoral fellowship in Manchester, England. As the
situation in Europe worsened, Erdös decided to move to the United States in 1938. Erdös never accepted a permanent academic position. He preferred to devote his entire waking hours to the pursuit of mathematics and traveled from one mathematics conference or seminar to another, building up a growing circle of collaborators. In the latter part of his life, Erdös owned nothing more than a suitcase of clothes and traveled from university to university, and from the home of one mathematician to another. He developed a reputation as a brilliant mathematician and an appalling houseguest. Sleeping only 3 to 5 hours a day, Erdös would often wake his mathematical hosts at all hours of the night, eager to get cracking on more mathematical research. By the end of his life, Erdös had worked on over 1,500 mathematical papers. He died in Warsaw, Poland, on September 20, 1996. Erdös won many prizes during his life, including the 1951 Cole Prize from the American Mathematical Society for his 1949 paper “On a New Method in Elementary Number Theory which Leads to an Elementary Proof of the Prime Number Theorem,” and the 1983 Wolf Prize of $50,000 from the Wolf Foundation. He was also awarded a salary from the Hungarian Academy of Sciences. With no need of money, Erdös often gave it away, either to students in need, or as prizes for solving problems he had posed, of which there were many. Mathematicians today are still publishing papers inspired by those challenges.
error The difference between the approximate value of a quantity and the true numerical value of that quantity is called the error of the approximation. There is some confusion in the literature, however, as to how to interpret this definition. If x is an approximation of the value X, then some texts work with the difference X – x when speaking of the error, whereas other texts use the difference x – X. (Thus, the error in using 3.6 as an approximation for 3.59, say, could be deemed as either 0.01 or –0.01.) For this reason, many authors prefer to work with the “absolute error,” X – x, and avoid the issue of sign altogether. The term error is also used for the uncertainty in a measurement. For example, one can typically read temperature only to the nearest degree Fahrenheit. Thus a temperature recording of 75°F should be written, or at least interpreted as, (75 ± 0.5)°F to indicate that there
168 Euclid is a possible error of as much as half a degree. When a measurement is written in decimal notation, it is generally understood that the absolute error does not exceed a half unit in the last digit. Thus, for example, a recording of 2.3 indicates that the error does not exceed ±0.05, whereas recording 2.30 indicates that the error does not exceed ±0.005. In this context, the final digit recorded is usually understood to be reliable. If a number representing a measurement does not have a decimal part, then a dot is sometimes used to indicate up to which point the digits are reliable. For . example, in recording a measurement as 2300, we are being told that the 3 is the final reliable digit and that the error in this measurement could be as much as ±50 . (half the 100s place). A recorded measurement of 2300, on the other hand, indicates that the first zero is reliable and that the error in this measurement is at most ±5 (half the 10s place). The digits up to, and including, the reliable digit are called the significant figures of the measurement. . Thus the measurement 2300 has two significant figures, . for example, whereas the recorded measurement 2300 has three significant figures. If a result is expressed in SCIENTIFIC NOTATION, p × 10n, it is generally understood that all the digits of p are significant. For example, in writing a value 0.0170 as 1.70 × 10–2, we are indicating that the final 0 is the result of a measurement, and so this digit is reliable. The quantity 0.0170 thus has three significant digits (and the error of this measurement is at most ±0.00005). Similarly a recorded measurement of 0.00030300, for example, has five significant figures. (The initial three zeros of the decimal expansion serve only to place the decimal point correctly. The remaining five digits represent the result of recording a measurement.) When calculating with approximate values, it is important to make sure that the result does not imply an unrealistic level of precision. For example, if the dimensions of the room are measured as 14.3 ft by 10.5 ft, multiplying length by width gives the area of the room as 150.15 ft2. The answer presented this way suggests a level of accuracy up to the nearest 1/100, which is unreasonable given that the initial measurements are made to the nearest 1/10. Generally, the result of a calculation should be presented as no more accurate than the least accurate initial measurement. . For example, in adding measurements .2300 and 1068, the result should be recorded as 3370 (the number
3368 is rounded to the nearest 10). In multiplying 14.3 and 10.5, each with. three significant figures, the result should be written 150 ft2 (again three significant figures). See also PERCENTAGE ERROR; PRECISION; RELATIVE ERROR; ROUNDOFF ERROR.
Euclid (ca.300–260 B.C.E.) Greek Geometry The geometer Euclid is remembered as author of the most famous text in the whole of mathematics, THE ELEMENTS. In 13 books, the work covers all that was known in mathematics at his time, from elementary geometry and number theory, to sophisticated theories of proportion, irrationals, and solid geometry. But Euclid is revered today primarily for his unique approach in organizing the material he presented. Starting with a small set of definitions, “common notions,” and AXIOMs (basic statements whose truth seems to be selfevident), Euclid derived by pure logical reasoning some 465 propositions (THEOREMs) in mathematics. This established standards of rigor and powers of deduction that became the model of all further work in mathematics for the two millennia that followed. It can be said that The Elements defines what PURE MATHEMATICS is about. Close to nothing is known of Euclid’s life. It is believed that he lived and taught in Alexandria, a Greek city near the mouth of the Nile in what is now Egypt, and may have been chief librarian of the great library at the Alexandria Academy. Many ancient historical texts describing the work of Euclid confuse the mathematician Euclid of Alexandria with philosopher Euclid of Megara, who lived about 100 years earlier. Moreover, Euclid was a very common name at the time, and there were many prominent scholars from a variety fields throughout this period. Because of the subsequent confusion and the lack of specific information about the mathematician Euclid, some historians have put forward the theory that Euclid was not, in fact, a historical figure, but the name adopted by a team of mathematicians at the library of Alexandria who published a complete work under the single name Euclid. (Compare this with the fictitious NICOLAS BOURBAKI of the 20th century). This is not the popular view, however. The Elements was deemed a standard text of study for Greek and Roman scholars for 1,000 years. It was translated into Arabic around 800 C.E. and studied extensively by Arab scholars. With the revival of scientific interest during the Renaissance, Euclid’s work
Euclidean algorithm 169 became the model of logical thinking in Europe. More than 2,000 different editions of the text have appeared since the first typeset version produced in the year 1482, and many great scientists of the West, including SIR ISAAC NEWTON (1642–1627) for instance, described their mastering the work of Euclid as a significant part of their development of scientific thinking. Study of The Elements was an integral part of the standard U.S. highschool mathematics curriculum up until the 1950s. Other works attributed to Euclid of Alexandria that have survived today include Data, on the properties of figures; On Divisions, studying the geometric theory of dividing the areas of figures into certain proportions; and Optics, the first Greek work on the theory of PERSPECTIVE. It is also known that Euclid produced at least five other texts in geometry, including a fourbook treatise on CONICS, as well as a work on music and another discussing general scientific principles. See also EUCLID’S POSTULATES.
a:b→c:d→…→u:v→z:z
EUCLID describes a systematic procedure for finding the GREATEST COMMON DIVISOR of any two positive integers. The method is as follows:
Then it is not too difficult to show that z is indeed the greatest common factor of a and b. Firstly, if a and b are both multiples of any number n, then so is their difference. This means that all the pairs of numbers produced by this procedure remain multiples of n. In particular, z is a multiple of n. (In the example above, both 42 and 60 are multiples of 3, for example. All the numbers produced via the procedure remain multiples of 3. They are also multiples of 2 and of 6.) This establishes that z is at least as large as any common factor of a and b. Secondly, working backward through the procedure, we see that the penultimate pair u : v is obtained from z : z via addition. (Look at the example above.) Thus both u and v are multiples of z. Working all the way back, we have in fact that all the numbers appearing in the list must be multiples of z, including both a and b. Thus z is a common factor. These two conclusions show that z is indeed the greatest common factor we seek. As a bonus, the above two paragraphs also show that the greatest common factor of two numbers is a multiple of any other common factor.
1. Write down the pair of numbers. 2. Subtract the smaller number from the larger. 3. Rewrite the pair of numbers but replace the larger number with the answer from step two. 4. Repeat steps two and three until you have two identical numbers. This repeated value is the greatest common factor of the two original numbers.
Linear Combinations A surprising consequence of the Euclidean algorithm is that it also gives a constructive method for writing the greatest common factor of two positive integers a and b as a linear combination of the original numbers. Keeping track of the subtractions performed in the example above, we have:
As an example, we calculate the greatest common factor of 42 and 60:
42 : 60 → 42 : 18 = (42) : (60–42) → 24 : 18 = (42) – (60 – 42) : (60 – 42) = (2 × 42 – 60) : (60 – 42) → 6 : 18 = (2 × 42 – 60) – (60 – 42) : (60 – 42) =(3 × 42 – 2 × 60) : (60 – 42) → 6 : 12 = (3 × 42 – 2 × 60) : (60 – 42) – (3 × 42 – 2 × 60) = (3 × 42 – 2 × 60) : (3 × 60 – 4 × 42) → 6: 6 = (3 × 42 – 2 × 60) : (3 × 60 – 4 × 42) – (3 × 42 – 2 × 60) = (3 × 42 – 2 × 60) : (5 × 60 – 7 × 42)
Euclidean algorithm In his third book of THE ELEMENTS,
42:60 → 42:18 → 24:18 → 6:18 → 6:12 → 6:6 Their greatest common factor is 6. Each step of the procedure produces a pair of numbers with smaller difference. Eventually, a pair with difference zero will result. The Euclidean algorithm is therefore sure to stop after a finite number of calculations.
Why the Algorithm Works Suppose two numbers a and b eventually produce the value z via this procedure:
Thus we can write 6 = 3 × 42 – 2 × 60 (and also 6 = 5 × 60 – 7 × 42.) In general, this shows that it is always
170 Euclidean geometry possible to write the greatest common factor of two numbers a and b in the form:
ntuples of COMPLEX NUMBERS under an analogous set of operations is called a complex Euclidean space.
xa + yb for some integers x and y. This fact is useful for solving the famous JUGFILLING PROBLEM, for example. See also FUNDAMENTAL THEOREM OF ARITHMETIC.
Euclidean geometry The
GEOMETRY based on the definitions and AXIOMs set out in Euclid’s famous work THE ELEMENTS is called Euclidean geometry. The salient feature of this geometry is that the fifth postulate, the PARALLEL POSTULATE, holds. It follows from this that through any point in the plane there is precisely one line through that point parallel to any given direction, that all angles in a triangle sum to precisely 180°, and that the ratio of the circumference of any circle to its diameter is always the same value π. Twodimensional Euclidean geometry is called plane geometry, and the threedimensional Euclidean geometry is called solid geometry. In 1899 German mathematician DAVID HILBERT (1862–1943) proved that the theory of Euclidean geometry is free from CONTRADICTION. See also EUCLID; EUCLID’S POSTULATES; HISTORY OF GEOMETRY (essay); NONEUCLIDEAN GEOMETRY.
Euclidean space (Cartesian space, nspace) The VECof all nTUPLES (x1,x2,…,xn) of real numbers x1, x2,…,xn with the operations of addition and scalar multiplication given by:
TOR SPACE
(x1, x2,…, xn) + (y1,y2,…,yn) = (x1, + y1, x2 + y2,…, xn + yn) k(x1, x2,…,xn) = (kx1, kx2,…,kxn) and equipped with the notion of distance between points x = (x1, x2,…,xn) and y = (y1, y2,…,yn) as given by the DISTANCE FORMULA: d(x,y) = √(x1 – y1)2 + (x2 + y2)2+…+(xn – yn)2 is called a Euclidean space. Elements of a twodimensional Euclidean space can be identified with points in a plane relative to a set of CARTESIAN COORDINATE axes. The vector space of all
Euclid’s postulates EUCLID of Alexandria (ca. 300–260 B.C.E.) began his famous 13volume piece THE ELEMENTS with 23 definitions (of the ilk, “a point is that which has no part” and “a line is that which has no breadth”) followed by 10 AXIOMs divided into two types: five common notions and five postulates. His common notions were: 1. Things that are equal to the same thing are equal to one another. 2. If equal things are added to equals, then the wholes are equal. 3. If equal things are subtracted from equals, then the remainders are equal. 4. Things that coincide with one another are equal to one another. 5. The whole is greater than the part. Euclid’s postulates were: 1. A straight line can be drawn to join any two points. 2. Any straight line segment can be extended to a straight line of any length. 3. Given any straight line segment, it is possible to draw a circle with center one endpoint and with the straight line segment as the radius. 4. All right angles are equal to one another. 5. If two straight lines emanating from the endpoints of a given line segment have interior angles on one given side of the line segment summing to less than two right angles, then the two lines, if extended, meet to form a triangle on that side of the line segment. (His fourth postulate is a statement about the homogeneity of space, that it is possible to translate figures to different locations without changing their basic structure.) It is worth noting that Euclid deliberately avoided any direct mention of the notion of infinity. His wording of the second postulate, for instance, avoids the need to state that straight lines can be extended indefinitely, and his fifth postulate, also known as the PARALLEL POSTULATE, avoids direct mention of parallel lines, that is, lines that never meet when extended indefinitely. From these basic assumptions Euclid deduced, by pure logical reasoning, 465 statements of truth (THEO
Euclid’s proof of the infinitude of primes 171 REMs)
about geometric figures. The systematic approach he followed and the rigor of reasoning he introduced was hailed as a great intellectual achievement. His model of mathematical exploration became the standard for all mathematical research for the next 2,000 years. Euclid’s fifth postulate was always regarded with suspicion. It was never viewed as simple and as selfevident as his remaining four postulates, and Euclid himself did his utmost to avoid using it in his work. (Euclid did not invoke the fifth postulate until his 29th proposition.) Over the centuries scholars came to believe that the fifth postulate could be logically deduced from the remaining four postulates and therefore did not need to be listed as an axiom. Many people proposed proofs for it, including the fifthcentury Greek philosopher Proclus, who is noted for his historical account of Greek geometry. Unfortunately his proof was flawed, as were the proofs proposed by Arab scholars of the eighth and ninth centuries, and by Western scholars of the Renaissance. In 1733 Italian teacher and scholar GIROLAMO SACCHERI (1667–1733) believed that because Euclid’s axioms model the real world, which he thought to be consistent, they cannot lead to a CONTRADICTION. If the first four postulates do indeed imply that the fifth postulate is also true, then assuming the four postulates together with the negation of the fifth postulate should lead to a logical inconsistency. Unfortunately, in following this tact, Saccheri never came across a contradiction. In 1795 Scottish mathematician and physicist John Playfair (1748–1819) proposed an alternative formulation of the famous fifth postulate (today known as PLAYFAIR’S AXIOM). This version of the axiom is considerably easier to handle, and its negation is easier to envision. In an attempt to follow Saccheri’s approach, Russian mathematician NICOLAI IVANOVICH LOBACHEVSKY (1792–1856) and Hungarian mathematician JÁNOS BOLYAI (1802–1860), independently came to the same surprising conclusion: the first four of Euclid’s postulates together with the negation of Playfair’s version of the fifth postulate will not lead to a contradiction. This established, once and for all, that the fifth postulate is an INDEPENDENT AXIOM and cannot be deduced from the remaining four postulates. More important, by exploring the geometries that result in assuming that the fifth postulate does not hold, scholars were led to the discovery of NONEUCLIDEAN GEOMETRY. In the late 1800s the German mathematician DAVID HILBERT (1862–1943) noted that, despite its rigor,
Euclid’s work contained many hidden assumptions. He also realized, despite Euclid’s attempts to describe them, that the notions of “point,” “line,” and “plane” cannot be properly defined and must remain as undefined terms in any theory of geometry. In his 1899 work Grundlagen der Geometrie (Foundations of geometry) Hilbert refined and expanded Euclid’s postulates into a list of 28 basic assumptions that define all that is needed in a complete account of Euclid’s geometry. His axioms are today referred to as Hilbert’s axioms. See also EUCLIDEAN GEOMETRY; HYPERBOLIC GEOMETRY; SPHERICAL GEOMETRY.
Euclid’s proof of the infinitude of primes Around the third century B.C.E., EUCLID proved that there is no such thing as a largest PRIME number, meaning that the list of primes goes on forever. He presented his proof as Proposition IX.20 in his book THE ELEMENTS, and he was the first to recognize and prove this fact about prime numbers. Euclid’s proof relies on the observation that any number N is either prime, or factors into primes. His argument proceeds as follows: Suppose to the contrary that there is a largest prime number p. Then the finite list 2, 3, 5, 7, …, p contains all the prime numbers. But consider the quantity: N = 2 × 3 × 5 × 7 ×…× p + 1 It is not divisible by any of prime numbers in our list (it leaves a remainder of one each time), and so it has no prime factor. It must be the case then that N is prime. Thus we have created a new prime number larger than the largest prime p. This absurdity shows that our assumption that there are only finitely many primes must be false. Euclid’s argument is a classic example of a PROOF BY His argument also provides an ALGORITHM for generating new primes from any finite list of primes. For example, from the list of primes 2, 3, 7, Euclid’s argument yields N = 2 · 3 · 7 + 1 = 43 as a new prime, and from the list 2,3,7,43, we have N = 2 · 3 · 7 · 43 + 1 = 1807 = 13 × 139, yielding 13 as a new prime. CONTRADICTION.
172 Eudoxus of Cnidus Euclid’s argument can be developed further to obtain other interesting facts about prime numbers. For example, we can prove: There are infinitely many primes that leave a remainder of 3 when divided by 4. (That is, there are infinitely many primes of the form 4n + 3.) Again, suppose to the contrary, that the list of such primes is finite: 3, 7, 11,…,p, and this time consider the quantity. N = 4 × (3 × 7 × 11 ×…× p) –1. If this number is prime, then we have found a new prime of the required form. If it is not, then it factors into primes: N = p1 × p2 ×…× pk. Notice that since N is not divisible by 3, none of its prime factors is equal to 3. It also cannot be the case either that all of the prime factors pi leave a remainder of 1 when divided by 4 (for then N would also leave a remainder of 1). Thus at least one of these prime factors is a prime of the form 4n + 3 not already in our list of such primes. It must be the case then that the list of such primes goes on forever. In a similar way (though it is a little more complicated) one can also prove: There are infinitely many primes of the form 6n+5. (Use N = 2 × 3 × 5 ×…× p – 1.) PETER GUSTAV LEJEUNE DIRICHLET (1805–59) generalized these results to prove that if a and b are any two RELATIVELY PRIME whole numbers, then there are infinitely many primes of the form an + b. See also FUNDAMENTAL THEOREM OF ARITHMETIC.
Eudoxus of Cnidus (ca. 408–355 B.C.E.) Greek Geometry, Number theory, Astronomy Born in Cnidus, in Asia Minor (now Turkey), Eudoxus is remembered as one of the greatest mathematicians of antiquity. All of his original work is lost, but it is known from later writers that he was responsible for the material presented in Book V of EUCLID’s famous treatise THE ELEMENTS. In his theory of proportions, Eudoxus developed a coherent theory of REAL NUMBERS using absolute rigor and precision. The full importance of this sophisticated work came to light some two millennia later, when scholars of the 19th century attempted to resolve some
fundamental difficulties with the theory of CALCULUS. They discovered that Eudoxus had already anticipated these fundamental problems and had made significant steps toward resolving them. Eudoxus is also remembered as the first to develop a “method of exhaustion” for computing the AREA of curved figures. As a young man Eudoxus traveled to Tarentum, now in Italy, to study number theory, geometry, and astronomy with ARCHYTAS OF TARENTUM, a follower of PYTHAGORAS. Both men worked to solve the famous problem of DUPLICATING THE CUBE and, in fact, Eudoxus came up with his own geometric solution to the challenge using special curved lines as an aid. (Although the problem calls for the use of nothing more than a compass and a straight edge, this partial solution was nonetheless a significant achievement.) Eudoxus studied the theory of proportions. This blend of GEOMETRY and NUMBER THEORY calls two lengths a and b COMMENSURABLE if they are each a wholenumber multiple of some smaller length t: a = mt and b = nt. In this approach, two ratios a : b and c : d are said to be equal if they are the same multiples of some fundamental lengths t and s: a = mt, b = nt and c = ms, d = ns. For a long time it was believed that all lengths were commensurable and hence all ratios could be compared. Consequently, the Pythagorean discovery of two – incommensurable lengths, namely 1 and √ 2, the side length and the diagonal of a unit square, caused a crisis in the mathematical community. As HIPPASUS OF METAPONTUM (ca. 470 B.C.E.) discovered, there is no – small value t such that 1 = mt and √ 2 = nt. (This is – equivalent to the statement that the number √ 2 cannot be written as a fraction n/m.) Eudoxus came to resolve the crisis of comparing ratios even if they are not commensurable by avoiding all use of a common length t. He defined ratios a : b and c : d to be equal if, for every possible pairs of integers n and m: i. ma < nb if mc < nd ii. ma = nb if mc = nd iii. ma > nb if mc > nd With this formulation, Eudoxus was able to compare lines of any length, either rational or irrational, and obviate all philosophical difficulties associated with incommensurable quantities. Mathematician JULIUS WILHELM RICHARD DEDEKIND (1831–1916) based his
Euler, Leonhard 173 theory of DEDEKIND CUTs on this approach developed by Eudoxus. In geometry, Eudoxus was the first to establish that the volume of a cone is onethird the volume of the cylinder that surrounds it, and also that the volume of a pyramid is one third the volume of a prism of the same base and height. ARCHIMEDES OF SYRACUSE (ca. 287–212 B.C.E.) made use of these results in his famous treatise On the Sphere and Cylinder, citing Eudoxus as the person who first proved them. Eudoxus maintained an active interest in astronomy throughout his life. He built an observatory in the city of Cnidus and made careful note of the motion of the planets and stars across the night skies. Outside of mathematics, Eudoxus is best known for his ingenious theory of planetary motion based on a system of 27 nested spheres. Using advanced techniques in threedimensional geometry, Eudoxus was able to use this model to explain the puzzling retrograde motion of the heavenly bodies.
Euler, Leonhard (1707–1783) Swiss Analysis, Geometry, Number theory, Graph theory, Mechanics, Physics Born on April 15, 1707, in Basel, Switzerland, genius Leonhard Euler was, beyond comparison, the most prolific mathematician of all time. With over 850 books and papers to his name, Euler made fundamental contributions to virtually every branch of mathematics of his day. He formalized the notion of a FUNCTION (and introduced the notation f(x) for it), and thereby changed the focus of mathematics from a study of fixed curves and lines to a more powerful study of transformation and change. (He was the first, for instance, to regard the special functions from TRIGONOMETRY as functions.) Euler published works on ANALYSIS, DIFFERENTIAL CALCULUS, INTEGRAL CALCULUS, DIFFERENTIAL EQUATIONs, NUMBER THEORY, GEOMETRY, LOGIC, COMBINATORICS, approximations for π, planetary motion and astronomy, navigation, cartography, mechanics, and more. He introduced and made popular many of – the standard symbols we use today (such as i for √–1, π for PI, E for his famous number, and Σ for SUMMATION). It is simply not possible in a short piece to give proper justice to the phenomenal quantity of contributions Euler made to the study of mathematics. Euler obtained a master’s degree in philosophy from the University of Basel in 1723, following the
Leonhard Euler, the most prolific of famous mathematicians, made significant contributions to almost every field of pure and applied mathematics studied at his time while also establishing new courses of research. Much of the mathematical notation used today was either introduced or made standard by Euler. (Photo courtesy of Topham/The Image Works)
path his father set for him to study theology. Euler’s interests, however, lay with mathematics, and Euler remained at the university another three years to pursue a course of study in the subject. In 1727 he submitted an entry for the Grand Prize of the Paris Academy of Sciences on the best arrangement of masts on a ship. Euler won second prize, which garnered the attention of the scientific community as an outstanding young graduate. Euler accepted a position at the St. Petersburg Academy of Science that year and was promoted to a full professor of the academy just three years later. In 1736 Euler published Mechanica (Mechanics), a landmark piece that introduced rigorous mathematical techniques as a means for studying the subject. He won the Grand Prize of the Paris Academy in 1738, and again in 1740, and by this time was a highly regarded scholar. At the invitation of Frederick the Great, Euler
174 Eulerian path/circuit joined the Berlin Academy of Science in 1741. He remained there for 25 years, assuming the leadership of the academy in 1759. During his time at Berlin, Euler wrote over 350 articles and several influential books on a wide variety of topics in both pure and applied mathematics. In 1753 he invented the paddle wheel and the screw propeller as means of propelling ships without wind. In 1761 he developed a method of using observational data about the planet Venus to determine the distance of the Earth from the Sun, and also to make measurements of longitude on the Earth’s surface. In 1766 Euler returned to St. Petersburg, but soon afterward fell ill and lost his eyesight. Despite being blind, Euler continued to produce significant pieces of work. (He published approximately 500 books and papers while blind.) Euler also worked to popularize the scientific method and wrote, between the years 1768 and 1772, his famous threevolume piece Letters to a German Princess on the topic of popular science. Euler remained at St. Petersburg until his death in 1783. Euler’s name is attached to at least one fundamental concept in nearly every branch of mathematics. For instance, EULER’S THEOREM is a key result in GRAPH THEORY, linking the number of vertices and edges of a graph to the number of regions it produces. The number e plays a fundamental role in the theory of differential calculus, differential equations, PROBABILITY theory, STATISTICS, the theory of SUBFACTORIALs and derangements, the study of COMPOUND INTEREST, and the study of COMPLEX NUMBERS through EULER’S FORMULA, for instance. In number theory, EULER’S CONSTANT plays a key role in the study of the HARMONIC SERIES, for instance. Euler found a formula for the “Euler totient function” that provides, for a number n, the count of numbers less than n that are RELATIVELY PRIME to n and showed its importance in the theory of MODULAR ARITHMETIC. Euler’s name is intimately associated with the study of the ZETA FUNCTION, with the gamma function in the examination of the FACTORIAL function, with the study of LATIN SQUARES, and with the construction of even PERFECT NUMBERS. (No one to this day knows whether or not examples of odd perfect numbers exist.) Less well known is Euler’s polynomial, n2 – n + 41, which produces a PRIME output for every integer input from –39 through to 40. The 40 × 117 × 240 rectangular block, called “Euler’s brick,” has the property that
any diagonal drawn on the face of this solid also has integer length. Euler showed that there are infinitely many such blocks with integer sidelengths and integer face diagonals. (No one to this day knows whether or not there exists an Euler brick with internal space diagonals also of integer length.) Euler died in St. Petersburg, Russia, on September 18, 1783. It is not an exaggeration to say that Euler offered profound insights on practically every branch of mathematics and mathematical physics studied at his time and, moreover, paved the way for many new branches of mathematics research. In 1915 the Euler Committee of the Swiss Academy of Science began collating and publishing his complete works. Divided into four series—mathematics, mechanics and astronomy, optics and sound, and letters and notebooks—76 volumes of work have been released thus far (covering approximately 25,000 pages of written material), and the committee projects another eight volumes of material still to be released.
Eulerian path/circuit See GRAPH THEORY.
Euler line For any triangle the following are true: 1. The three MEDIANS OF A TRIANGLE meet at a point G, the centroid of the triangle. 2. The three ALTITUDEs of a triangle meet at a point H, the orthocenter of the triangle. 3. The perpendicular bisectors of each side of a triangle meet at a point O, the circumcenter of the triangle.
The Euler line
Euler’s formula 175 These statements are proved through a study of the median of a triangle, the altitude of a triangle, and the consideration of EQUIDISTANT points, respectively. In the mid1700s LEONHARD EULER (1707–83) made the astounding discovery that furthermore, for any triangle, the three points G, H, and O are COLLINEAR, that is, lie on a straight line. This line is called the Euler line of the triangle. Euler proved this observation as follows: If, by chance, the points O and G coincide, then each median of the triangle is also an altitude. This means that the triangle is symmetric about each median, and so must be equilateral. Consequently, the point H occurs at the same location as O and G, and the three points, trivially, lie on a straight line. If, as is more likely the case, O and G do not coincide, then draw a line through them and consider a point J on this line that is situated so that the length of the segment GJ is twice that of OG. Let M be the midpoint of the base of the triangle. From a study of the medians of a triangle, we know that length of the segment AG to that of GM in the diagram above is in ratio of 2 to 1. Consequently, the two shaded triangles are similar, and, in particular, angles AJG and GOM match. By the converse of the PARALLEL POSTULATE, lines OM and AJ are parallel. Since OM makes an angle of 90° to the base of the triangle, so too must line AJ, making this line an altitude to the triangle. Nothing in this argument thus far has relied on vertex A being the object of focus. The same reasoning shows that the altitude from vertex B also passes through the point J, as does the altitude from vertex C. This shows that the point J is in fact the orthocenter H of the triangle. Consequently, O, G, and H do indeed all lie on the same straight line.
Understanding Euler’s constant
1 – 1 1 – 1+ – 2 + 3 + …+ n = ln n + γ + error 1 where the “error” is the term – n minus all the “excess areas” above the curve from position n onward. Notice that these excess areas all fit within the rectangle of 1 1 height – n , so this error is no bigger than – n . In particular, it is negligibly small if n is large. See also HARMONIC SERIES.
Euler’s formula In 1748 LEONHARD EULER noted that the TAYLOR SERIES for the functions ex, sin x, and cos x are intimately connected. Since x x 2 x3 x 4 x 5 x6 + + + + + + 1! 2! 3! 4! 5! 6! x x3 x 5 sin x = − + −L 1! 3! 5! x2 x4 cos x = 1 − + −L 2! 4! ex = 1 +
Euler’s constant In drawing rectangles of width 1 that just cover the curve y = 1/x, one sees that the “excess area” above the curve fits inside the first rectangle of height 1, and so sums to a finite value no larger than 1. The amount of excess area, denoted γ, is called Euler’s constant. To eight decimal places, it has value 0.57721566. No one knows whether γ is a rational or irrational number. As the area under the curve from x = 1 to x = n is n 1 1 1 – – – 1dx = ln n, we have that 1 + 2 + 3 + … + n–1 is approximately equal to ln(n). More precisely, the sum of the areas of the first n rectangles is given by:
∫
setting x = iθ, where i is the square root of –1 and θ is a real number (usually thought of as an angle), yields:
e iθ
2 3 4 5 iθ) (iθ) iθ) iθ) iθ) ( ( ( ( + + + + +L = 1+
1! 2! 3! 4! 5! θ 2 θ 2 3 θ3 4 θ 4 5 θ 5 = 1+ i + i +i +i +i +L 1! 2! 3! 4! 5!
176 Euler’s theorem θ θ2 θ3 θ 4 θ5 − −i + −i +L 1! 2! 3! 4! 5! ⎛ ⎞ ⎛ θ θ3 θ 5 ⎞ θ2 θ4 = ⎜1 − + − L⎟ + i ⎜ − + − L⎟ 2! 4! ⎝ ⎠ ⎝ 1! 3! 5! ⎠ = 1+ i
= cos(θ) + i sin(θ)
The formula eiθ = cosθ + isinθ is today known as Euler’s formula. It has some interesting consequences: Setting θ = π, we obtain: eiπ = cosπ + isinπ = –1 + i · 0 = –1. Mathematicians often deem this as one of the most beautiful facts of mathematics: it is a remarkably simple equation that connects the mysterious, and pervasive, numbers e, π, i, and –1. π i π e 2 = i , which shows that Setting θ = – yields 2 i i = (e
i
π i 2
)
=e
i2
π 2
=e
−
π 2
. Thus raising a complex number to a complex power can yield a real answer as a result. Euler’s formula provides a very simple means for deriving (and memorizing) certain identities from TRIGONOMETRY. For example, since eiA · eiB = ei(A + B) we have:
and the complex number is thus: z = x + iy = r cosθ + ir sinθ = reiθ This is called the polar form of the complex number. If one multiples two complex numbers, z = reiθ and w = seiτ, we see that z · w = rsei(θ+τ), that is: The product of two complex numbers is a new complex number whose distance from the origin is the product of the distances from the origin of the two original numbers, and whose angle with the xaxis is the sum of the two angles made by the two original numbers. Euler’s formula makes the derivation of this fact swift and easy. See also COMPLEX NUMBERS; DE MOIVRE’S FORMULA; E HYPERBOLIC FUNCTIONS.
Euler’s theorem (Euler’s formula, EulerDescartes formula) A GRAPH is a collection of dots, called vertices, connected in pairs by line segments, called edges, subsequently dividing the plane into a finite number of regions. In 1752 LEONHARD EULER showed that if a graph drawn on the plane has v vertices, e edges, and divides the plane into a total of r regions (this includes the large “outer region”), then:
(cosA + isinA) · (cosB + isinB) = cos(A + B) + isin(A + B)
v–e+r=1+c
Expanding the brackets on the left and collecting terms that contain i and those that do not quickly yields:
where c is the number of “connected components” of the graph, that is, the number of distinct pieces of which it is composed. For example, the graph shown is composed of two “distinct pieces” (c = 2) and has nine vertices, 13 edges, and divides the plane into seven distinct regions, and indeed v – e + r equals 3, one more than c. The formula is easily proved via an INDUCTION argument on the number of edges: if a graph has no edges, then it consists solely of v disconnected points. Thus it has c = v components and divides the plane into just one region. The formula v – e + r = 1 + c holds true. One checks that adding an edge either divides a region into two (thereby increasing the value of r by one), creates an extra region if that edge is a loop (again increasing the value of r by 1), or connects two disconnected components of the graph (thereby
cos(A + B) = cosA · cosB – sinA · sinB sin(A + B) = sinA · cosB + cosA · sinB Similarly, the equations (eiA)2 = ei(2A), (eiA)3 = ei(3A), and so forth yield doubleangle and tripleangle formulae, for example. Euler’s formula is also used to represent complex numbers. For example, if z is a point in the complex plane a distance r from the origin, making an angle θ with the xaxis, then its x and ycoordinates can be written: x = r cosθ y = r sinθ
even and odd numbers decreasing the value of c by 1). In all cases the formula v – e + r = 1 + c remains balanced. The formula is usually applied to a graph that is connected, that is, has only one component (c = 1). In this case the formula reads:
177
contains a hole (say, a cube with a square hole drilled through it), one has: v – e + f = 0. (There is one technical difficulty here: one needs to be sure that each region or face under consideration is not itself an ANNULUS. One may need to draw in extra edges to break regions into suitable form.)
v–e+r=2 and this version of the equation is usually called Euler’s theorem. The vertices, edges, and faces of a POLYHEDRON can be thought of as a connected graph. For example, a cube with its top face removed and pushed flat onto a plane yields a graph with eight vertices and 12 edges dividing the plane into six regions. (The large “outer” region corresponds to the top face of the cube that was removed.) We still have: v – e + r = 2. In general, for any polyhedron with v vertices, e edges, and f faces we have: v–e+f=2 This was first observed by RENÉ DESCARTES in 1635. Euler had no knowledge of Descartes’s work when he developed the formula in the more general setting of graph theory. For this reason, this famous formula is also called the EulerDescartes formula. This result holds true only for graphs that lie on the plane (or polyhedra that can be pushed flat onto a plane). One can show that for connected graphs drawn on a TORUS, for example, the formula must be adjusted to read: v – e + r = 0. For example, if a polyhedron
even and odd functions A function y = f(x) is said to be even if, for each x, the function takes the same value at both x and –x, that is, f(–x) = f(x) for all values of x. The graph of an even function is consequently symmetrical about the vertical axis. The functions x2, 3x2 – 5 cos(x), and ——— 1 + x4 , for example, are even functions. A function y = f(x) is said to be odd if, for each x, the function takes opposite values at x and –x, that is, f(–x) = –f(x) for all values of x. The graph of an odd function is consequently symmetric with respect to a 180° rotation about the origin. The functions x, x3, x + 5x3 − 17 x 93 sin(x), and , for example, are odd 3 + 2x 4 functions. Any function g(x) can be expressed as the sum of an even and an odd function. Let: feven (x) =
g(x) + g(− x) 2
fodd (x) =
g(x) − g(− x) 2
Then feven(x) is even, fodd(x) is odd, and g(x) = feven(x) + fodd(x). The FOURIER SERIES of any even function contains only cosine terms, and the Fourier series of any odd function only sine terms. The absolute value of any odd function f(x) is an even function, that is, if y = f(x), then y is even.
A graph with two components
even and odd numbers Working solely in the realm of the whole numbers, a number is said to be even if it is divisible by 2, and odd if it leaves a remainder of 1 when divided by 2. For example 18 is divisible by 2 and so is even, and 23 leaves a remainder of 1 and so is odd. There is a physical interpretation to the evenness or oddness of a number: An even number of pebbles, say, represents a pile that can be split into two equal
178 event parts—18 pebbles split into two piles of 9, for example. An odd pile of pebbles leaves an extra pebble when one attempts to accomplish this feat: 23 pebbles splits into two piles of 11, with one left over. This interpretation shows that the number zero is even: an empty pile of pebbles can be split into two equal piles of nothing. Combining together several piles of pebbles, each of which can be evenly split, produces a large pile that still can be so separated. This shows that the sum of any collection of even numbers is even. Combining two odd piles of pebbles produces a result that is even—the two “errant” pebbles combine to produce a result that can be evenly split. The sum of three odd numbers leaves a single errant pebble, however, and so is odd. In summary we have: The sum of any number of even numbers is even. The sum of an even number of odd numbers is even. The sum of an odd number of odd numbers is odd. These simple ideas are quite powerful. For example, we can quickly ascertain that: The sum of the numbers 1 through 100 will be even. (This sum contains 50 even numbers and 50 odd numbers.) It is impossible to make change for a dollar using a combination of 13 pennies, nickels and quarters. (Thirteen coins of odd denomination will sum to an odd amount. It is impossible then to reach a sum of 100.) If 15 arbitrary sheets are torn from a textbook, then the sum of the missing page numbers is necessarily odd. (Each sheet contains an odd page number on one side and an even page number on the other. The sum of 15 odd numbers and 15 even numbers is necessarily odd.) Seventeen cups are placed upsidedown on a tabletop. Turning four cups over at a time, it is impossible to reach a state in which every cup is upright.
(To be left upright, each cup must be turned over an odd number of times in the process of the game. Thus an odd number of inversions must occur in all, being a sum of 17 odd numbers. But an odd total will never occur when turning an even number of cups over at a time.) See also PARITY.
event Any subset of the SAMPLE SPACE of all possible outcomes of an experiment is called an event. For example, the act of casting a die has sample space {1, 2, 3, 4, 5, 6}—all six possible scores—and the event “the score is even” is the subset {2, 4, 6}. An event could be a single outcome (“rolling a two,” for example, is the subset {2}), the whole sample space (“rolling a number less then 10”), or the empty set (“rolling a multiple of seven”). The probability of an event E occurring is the ratio of the number of outcomes in that event to the total number of outcomes possible. This ratio is denoted 3 1 – P(E). For example, in casting a die, P(even) = – 6 = 2, 2 – 1 0 – P({5,6}) = – 6 = 3 , and P(multiple of 7) = 6 = 0. SET THEORY is useful for analyzing the chances of combinations of events occurring. If A and B represent two events for an experiment, then: The intersection A ∩ B represents the event “both A and B occur.” The union A ∪ B represents the event “either A or B occurs.” The complement A′ represents the event “A does not occur.” To illustrate: if, in casting a die, A is the event {2, 4, 6} and B the event {5,6}, then A ∩ B = {6}, A ∪ B = {2, 4, 5, 6}, and A′ = {1, 3, 5}. The following probability laws hold for two events A and B: i. P(A ∩ B) = P(A) + P(B) – P(A ∪ B). ii. When A and B are disjoint events, that is have no outcomes in common, then P(A ∪ B) = P(A) + P(B). iii. P(A′) = 1 – P(A). iv. When A and B are INDEPENDENT EVENTS, P(A ∩ B) = P(A) × P(B). The study of PROBABILITY explains these rules. See also CONDITIONAL ; MUTUALLY EXCLUSIVE EVENTS; ODDS.
expected value 179 cated than simply understanding the simple process at hand. See also DISTRIBUTIVE PROPERTY ; NESTED MULTIPLICATION.
Computing the product of 23 and 42
expanding brackets The geometric figure of a rectangle explains the process of expanding brackets. Take, for example, a 23by42 rectangle. Its AREA is given by the product 23 × 42. This product can easily be computed by thinking of 23 as 20 + 3 and 42 as 40 + 2. This corresponds to subdividing the rectangle into four pieces: We thus have (20 + 3) × (40 + 2) = 20 × 40 + 20 × 2 + 3 × 40 + 3 × 2, which equals 800 + 40 + 120 + 6, or, 966, which is indeed 23 × 42. Note that each of the four terms in the sum is the product of one number in the first set of parentheses (20 or 3), and one number in the second set (40 and 2), with all possible pairs of numbers appearing. This principle holds in general. For example, the quantity (x + y)(a + b + c) equals the sum of six products: xa + xb + xc + ya + yb + yc (this corresponds to subdividing a rectangle into six pieces), and (r + s + t + u + v)(k + l + m + n + o + p + q) is the sum of 35 individual products. This principle extends to any number of sets of parentheses. For example: (2 + 3) × (4 + 5) × (6 + 7) = 2 × 4 × 6 + 2 × 4 × 7 + 2 ×5×6+2×5×7+3×4 ×6+3×4×7+3×5×6 +3×5×7 (Again select one term from each set of parentheses, making sure to include all possible combinations.) This corresponds to subdividing a cube into eight pieces. It also holds for products containing single terms along with sets of parentheses. For example, we have: (a + b) × x × (c + d) = a × x × c + a × x × d + b × x × c + b × x × d. Many schools teach mnemonic devices for correctly expanding brackets. These can be more compli
expected value (expectation, mean) The expected value of a game of chance involving monetary bets is the average or MEAN profit (or loss) per game you would expect if the game were played a large number of times. The expected value illustrates the extent to which a game is set to, or against, your favor. To demonstrate: imagine you have the opportunity to play the following dice game: You toss a single die and look at the score cast. If a 1 comes up you win $10, and if a 2 appears you win $5. If any other number is cast, you pay a fee of $3 for playing the game. Is this a game worth playing? With 600 plays of this game, one would expect close to onesixth of those rolls (around 100 of them) to yield a 1, and hence a gain of $10, another sixth of the rolls (that is, about 100 rolls) to yield a 2 and a gain of $5, and twothirds of the rolls (around 400 of them) to result in a loss of $3 (that is, a –$3 profit). The average profit over 600 rolls would thus be: 100 × 10 + 100 × 5 + 400 × (−3) 1 1 4 = × 10 + × 5 + × (−3) 600 6 6 6 = 0.50 that is, a gain of 50 cents per game. This positive expected value shows that the game is indeed worth playing. Note, however, that one might still lose money while playing the game. What has been demonstrated here is that, for the long run, the game is set to your favor. Note the appearance of the fractions 1/6, 1/6, and 4/6 in our computation of expected value. These are the probabilities of each identified outcome actually occurring. Such probabilities always appear when computing expected value. In general, if an experiment yields numerical values x1, x2,…,xn, with pi being the probability that outcome xi occurs, then the expected value of the experiment is given by: x1p1 + x2p2 +…+ xn pn
180 exponent For example, in tossing a pair of dice, the expected sum is:
m times
n times
bm × bn = (b × b × L × b) × (b × b × L × b) m + n times
= b × b × L × b × b = bm + n 2×
1 2 3 4 5 6 +3× +4× +5× +6× +7 × 36 36 36 36 36 36 +8×
5 4 3 2 1 +9× + 10 × + 11 × + 12 × =7 36 36 36 36 36
(The possible sums are the numbers 2 through 12, with the probability of casting 2 being 1/36, a 3 being 2/36, and so on.) Expected value is usually denoted by the letter μ from the GREEK ALPHABET. If an experiment has an infinite number of possible outcomes, then the expected value is given by an infinite sum (SERIES). The BINOMIAL DISTRIBUTION is an example of this. If the random phenomenon can produce a continuous array of values (for example, the height of an individual can be any value, including fractional ones), then the expected value is given by an INTEGRAL: ∞
μ = ∫ – ∞ x p(x) dx Here p(x) is the probability density function of the random variable under consideration for the given DISTRIBUTION. The notion of expected value was first developed by Dutch scientist Christiaan Huygens (1629–95) in his treatise On Reasoning in a Dice Game. See also HISTORY OF PROBABILITY AND STATISTICS (essay).
exponent (index) For a real number b and a positive whole number m, the shorthand bm is used for the repeated multiplication of b with itself m times: bm = b × b ×…× b (m times). Thus, for example, 25 = 2 × 2 × 2 × 2 × 2 = 32, (–1)3 = (–1) × (–1) × (–1) = –1, 2
⎛ 1⎞ 1 1 1 1 m ⎜ ⎟ = × = , and 10 = 10. In the expression b , 3 3 9 ⎝ 3⎠ m is called the exponent, or the index, and b is called the base of the exponent. We also call bm a power of b. The product of two expressions bm and bn with the same base b is itself a repeated multiplication of the number b. Precisely:
This establishes the multiplication rule for exponents: To multiply two expressions with the same base, retain the common base and add together the exponents: bm × bn = b m + n. The power rule for exponents, (bm)n = bmn, follows. (One must add m with itself n times.) The multiplication rule is considered fundamental and allows us to define exponential quantities bm for values of m other than whole numbers. We follow the principle that the multiplication rule is to always hold. Consider, for example, the expression 20. This quantity has no meaning when interpreted as “the multiplication of two with itself zero times.” However, one can assign a meaningful value to this expression by multiplying it with another power of two. For example, according to the multiplication rule, it must be the case that 20 × 25 = 20+5 = 25. This says that 20 × 32 = 32, which tells us that 20 must equal one. The multiplication rule thus leads to the rule: The zero exponent for any nonzero base equals 1: b0 = 1. To make sense of the quantity 2–1, again invoke the multiplication rule. We have, for example, 2–1 × 23 = 2–1+3 = 22. This reads: 2–1 × 8 = 4. It must be the case then that 2–1 = 1/2. Similar calculations show that 2–2 must equal 1/4, and that 2–3 must equal 1/8. In general, 1 . This works for any nonzero base b. 2– m = –– 2m A negative exponent indicates that a reciprocal 1. must be taken: b–n = – bn 56 We can make use of this observation to compute –4 , 5 for example. Rewriting, we have
56 1 = 56 × 4 = 56 × 5−4 4 5 5
= 56− 4 = 52 = 25 . To divide two expressions with the same base, retain the common base and subtract bm = bn – m the exponents: – bn
exponential function 181 The multiplication rule also allows us to make1 sense of – fractional exponents. Consider the quantity 12 2 . It must 1 1 – – – be the case that 2 2 × 2 2 = 21 = 2. Thus 2 2 is a value that, when multiplied1 by itself, equals 2. Consequently 1 – – – 2 2 = √ 2. Similarly, 2 3 is a value that, when multiplied 1 3— – by itself three times, equals 2, and so 2 3 = √ 2, the cube 1 – root of 2. In general, 2 n equals the nth root of 2. This works for any nonzero base. A fractional exponent indicates that a root is 1 n— – to be taken: b n = √ b. We use the power rule to make sense of other types of2 – fractional exponents. For example, the quantity 27 3 1
can be computed as
27 3
×2
= 9 . In general, we have: p q b
=
1 ×p q b
=
( )2 = (3 27 ) = 32 1
2
= 27 3
( b) q
p
Finally, to compute a quantity raised to an irrational power, one approximates the exponent by a fraction, computes the corresponding exponential expression, and takes the LIMIT as one uses better and better – approximations. For example, writing √ 2 = 1.414…, 14 1414 – , 141 – –––– we see that any of the fractions 1, 10 100 , 1000 ,… can – be used to approximate √ 2– with better and better degrees of accuracy. We define 2√2 to be the limit of the values: 21 = 2, 1414
14 2 10
=
( 2)
14
10
≈ 2.639 ,
141 2 100
=
( 2) 100
141
≈ 2.657 ,
2 1000 ≈ 2.665 , …. The multiplication and power rules are valid even for irrational exponents. For example, we have: 2
2
2
= 2
2× 2
2
square of an unspecified quantity a tetragon number, meaning a “fourcorner number.” DIOPHANTUS OF ALEXANDRIA (ca. 200–284 C.E.) used the Greek word dynamis, meaning “power,” for the square of an unknown, and called a third power a “cube,” a fourth power a “powerpower,” and fifth and sixth powers “powercube” and “cubecube,” respectively. It took many centuries for scholars to begin using symbols to denote unknown quantities. German mathematician Michael Stifel (ca. 1487–1567) was the first to develop a notational system for powers of an unspecified quantity x. He denoted the fourth power of x simply as xxxx. Other scholars developed alternative notational systems. Scholars eventually settled on the notational system French mathematician and philosopher RENÉ DESCARTES (1596–1650) introduced in 1637, the one we use today. Although Descartes considered only positive integral exponents, later that century the English mathematician SIR ISAAC NEWTON (1642–1727), inspired by the work of JOHN WALLIS (1616–1703), showed that the same notational system can be extended to include negative, fractional, and irrational exponents. LEONHARD EULER (1707–83) later allowed for the possibility of complex exponents. See also COMPLEX NUMBERS; EXPONENTIAL FUNCTION; LOGARITHM.
exponential function Any function or quantity that varies as the power of another quantity is called exponential. Precisely, if b is a positive number different from one, then the function f(x) = bx is called the exponential function with base b. The function is defined for all real numbers x. (This would not be the case if 1 – b were negative: the value b 2 , for example, would x 1 not make sense.) The graphs of y = 2x and y = 2
= 2 =2
The Greek mathematician ARCHIMEDES OF SYRA(287–212 B.C.E.) was one of the first scholars to use a special word for the power of a number. He called the quantity 10,000, 104, a myriad, and he used the phrase “myriad of myriads” for 10,000 squared, 104 × 104 = 108. The ancient Greeks, for whom mathematics was synonymous with geometry, called the CUSE
Exponential functions
182 exponential series illustrate the essential differences between the cases b > 1 and 0 < b < 1. Many types of growth and decay occur at a rate that involves exponential variation. For example, a colony of bacteria might reproduce at a rate that doubles the size of the colony every 24 hours. If initially 500 organisms are present, then after 1 day the culture grows to a size of 1,000 organisms, after 2 days to a size of 2,000 organisms, and by the end of day three there are 4,000 organisms. In general, the population size by the end of day N is given by 500 × 2N. Any formula of the form ABx with b > 1 and A constant is said to represent exponential growth. This formula is a simple example of a POPULATION MODEL. The analogous formula with 0 < b < 1 represents exponential decay. The decay of radioactive material is a typical example of this. For example, if 50 percent of the radioactivity produced by a nuclear explosion disappears after 5 days, then after 10 days only 25 percent of radioactivity remains, and 12.5 percent remains after 15 days. The percentage of radioactivity present 5N 1 after N days is given by the formula . The level of 2 radioactivity decreases but will never reach zero. The DERIVATIVE of an exponential function f(x) = bx can be computed via IMPLICIT DIFFERENTIATION after first taking a LOGARITHM. Precisely, if y = bx, then 1 dy lny = x · lnb. Differentiation yields – · –– = lnb, and so y dx dy x –– = y · lnb = b · lnb. That is: dx
( )
d x b = b x ⋅ ln b dx This formula for the derivative is greatly simplified if one works with base b = e, where e is defined to be the number such that ln e = 1. We have:
exponential series The TAYLOR
of the func2 3 x x x4 tion f(x) = ex, given by e x = 1 + x + + + +K 2! 3! 4! is called the exponential series. The RATIO TEST from the study of CONVERGENT SERIES shows that this series converges for all values of x. LEONHARD EULER (1707–83) made use of this series to prove his famous formula eix = cosx + isinx, today called EULER’S FORMULA.
expression Any meaningful combination of symbols that represent numbers, operations on numbers, or other mathematical entities is called an expression. For example, 2 + x and ab+ c are expressions. One could argue that x + y = 2 is an expression, although mathematicians may prefer to call it an equation. Similarly, 3– √ 8 could be called an expression even though it is equivalent to a single number. In FORMAL LOGIC, compound statements are sometimes called expressions. ∨ For example, ¬(p (q ∨ r)) is an expression. The word express is often used in the sense “to transform.” For example, the product (x – y)(x + y) can be expressed equivalently as the DIFFERENCE OF TWO SQUARES: x2 – y2.
exterior angle An ANGLE contained between one side of a POLYGON and the extension of an adjacent side is called an exterior angle of the polygon. Since two edges of a polygon meet at a vertex, there are two exterior angles at each vertex. One can easily see, however, that these two angles are equal in value. The sum of exterior angles in a convex polygon is 360°, since these angles correspond to one full turn. This result is also true for concave polygons if one deems angles that turn to the left as positive and ones that turn to the right as negative. The EXTERIORANGLE THEOREM, first proved by the geometer EUCLID (ca. 300–260 B.C.E.), states that
The derivative of ex is ex. Consequently, the graph of the exponential function f(x) = ex has the property that the slope of the graph at any point is the same as the value of the function at that point. The function f(x) = ex is sometimes called the exponential function. Because the derivative of this function is particularly simple, it is not surprising that the number e is ubiquitous throughout studies in CALCULUS.
SERIES
Exterior angles
extraction 183 the exterior angle at one vertex of a triangle is greater in value than that of an interior angle at either of the remaining two vertices. See also CONCAVE/CONVEX.
exteriorangle theorem In his famous work THE ELEMENTS, the geometer EUCLID (ca. 300–260 B.C.E.) established the following result, which he called the exteriorangle theorem: For a given triangle ABC with interior angles x and y and exterior angle z as shown, we have z > x and z > y. The result is proved as follows: Let M be the midpoint of side BC, and extend a line from point A through M to a new point D so that AM and DM are the same length. Consider triangles AMB and DMC. They share two sides of the same lengths and a common angle at M. By the SAS principle for similarity, the two triangles are congruent figures. Consequently, angle MCD matches angle MBA, which is y. Since z is clearly larger than angle MCD, we have that z > y. An analogous argument based on a line drawn through the midpoint of side AC establishes that z is also greater than x.
Equal alternate interior angles
than the interior angle x. Similarly, the lines cannot meet to the left to form a triangle by the same reason. (Work with the angle 180 – x.) It must be the case then that the lines are parallel. This result is the CONVERSE of Euclid’s famous, and controversial, PARALLEL POSTULATE. It is important to note that Euclid proved the exteriorangle theorem and its consequence without assuming that the parallel postulate holds. If one is willing to assume that the three angles in a triangle always sum to 180° (a statement equivalent to the parallel postulate), then the proof of the exterior angle theorem is TRIVIAL.
extraction The process of finding the
This theorem has one very important consequence. If two lines cut by a transversal produce equal alternate interior angles, then the two lines are parallel. In the diagram above, if the two angles labeled x are indeed equal, then the lines cannot meet to the right to form a triangle: the exterior angle x cannot be greater
ROOT of a number or the solution to an algebraic equation is sometimes called extraction. For example, extracting the square root of 3 is the process of finding its square root. (One might use HERON’S METHOD, for example, – to compute √ 3 = 1.7320508…) The term digit extraction is often used to describe any method that allows one to compute a specific digit of a number without computing earlier digits. For example, in 1995 mathematician Simon Plouffe discovered the following remarkable formulae for π:
π=
Understanding the exterior angle theorem
∞
4 2 1 1 1 ∑ 8n + 1 − 8n + 4 − 8n + 5 − 8n + 6 16 n =0
n
It has led mathematicians to a technique that computes the Nth digit of π in base 16 without having to calculate the preceding N – 1 digits. (In base 16, the number π appears as 3.243F6A8885A308D … where
184 extrapolation A represents 10, B represents 11, and so forth, with F representing 15. Each decimal place is a power of a sixteenth.) Unfortunately no analogous technique is currently known to compute the ordinary base10 digits of π with ease.
extrapolation The process of estimating the value of a function outside a known range of values is called extrapolation. For example, if the temperature of a cup of tea was initially 200°F and then was measured to be 100°F and 50°F 10 and 20 minutes later, respectively, then one might guess that its temperature after 30 minutes would be 25°F. Methods of extrapolation are normally far less reliable than INTERPOLATION, the process of estimating function values between known values. Scientists generally prefer to avoid making predictions based on extrapolation. Meteorologists, however, must use extrapolation techniques to make weather predictions. Longrange forecasts are generally considered unreliable. See also POPULATION MODELS.
extremevalue theorem This theorem asserts that a f(x), defined on a closed INTERVAL [a,b], reaches a maximum value and a minimum value somewhere within that interval. That is, there is a point c in the interval [a,b] such that f(x), ≤ f(c), for all x in [a,b], and there is another point d in [a,b] such that f(x) ≥ f(d) for all x in [a,b]. For example, the extremevalue theorem tells us that, on the interval [1,5] say, the function f(x) = 3x · cos(x2 + sinx) does indeed adopt a largest value. It does not tell us, however, where that maximum value occurs. A point at which a function has a maximum or minimum value is called an extremum. The theorem is intuitively clear if we think of a continuous function on a closed interval as one whose graph consists of a single continuous piece with no CONTINUOUS
FUNCTION
gaps, jumps, or holes: in moving a pencil from the left end point (a,f(a)) to the right end point (b,f(b)), one would not doubt that there must be a high point on the curve where f(x) reaches its maximum value, and a low point where it attains its minimum value. A rigorous proof of the theorem, however, relies on the notion of the completeness of the real numbers (meaning that no points are missing from the real line). This is a subtle property, one that was not properly understood until the late 1800s with the invention of a DEDEKIND CUT. For example, the function f(x) = 2x – x2 = x(2 – x) has no maximum value on the interval [0, 2] if the value 1 is excluded from our considerations. Although this seems an artificial example, one still needs to be sure that this type of problem can never occur. It is vital that the function under consideration be continuous and that the interval under study be closed for the theorem to be true. For example, the (discontinuous) function 2x − x 2 if x ≠ 1 f (x) = if x = 1 0 does not reach a maximum value in the closed interval 1, [0,2]; nor does the (continuous) function f(x) = – x defined on the open interval (0,2), since the function is arbitrarily large for values x close to zero. The INTERMEDIATEVALUE THEOREM is a companion result to the extremevalue theorem. It asserts that not only does a continuous function on a closed interval actually attain maximum and minimum values, but it also takes on every value between those two extreme values. ROLLE’S THEOREM and the MEANVALUE THEOREM follow from the extremevalue theorem if we further assume that the function under consideration is differentiable. See also DIFFERENTIAL CALCULUS; MAXIMUM/ MINIMUM.
F face A flat surface on the outside of a solid figure,
The factors of a given number have a geometric interpretation. For example, one can arrange 12 pebbles into six different rectangular arrays: a 1 by 12 rectangle, a 2 by 6 rectangle, a 3 by 4 rectangle, a 4 by 3 rectangle, a 6 by 2 rectangle, and finally a 12 by 1 rectangle. The dimensions of these rectangles are precisely the factors of 12. This interpretation shows that the factors of a number come in pairs—unless, one of the rectangles formed is a perfect square (in which case, one factor is “paired with itself”). This shows:
typically a POLYHEDRON, is called a face of the figure. For example, a cube has six identical faces, and a cylinder has two faces. (The lateral surface of a cylinder is not flat.) In the mid1700s, Swiss mathematician LEONHARD EULER established that if all the outside surfaces of a convex solid are flat, then the number of faces f the figure possesses is given by the formula: f=2–v+e Here v is the number of vertices and e is the number of edges the figure has. The angle between two edges of a polyhedron meeting at a common vertex is sometimes called a face angle. In GRAPH THEORY, any region of plane bounded by edges of a planar graph is sometimes called a face of the graph. EULER’S FORMULA v – e + f = 2 also holds for connected planar graphs if one is willing to regard the large unbounded region outside the graph as a face. See also DIHEDRAL.
Square numbers have an odd number of factors. All other numbers have an even number of factors. For example, 36, which equals 6 × 6, has an odd number of factors: 1 and 36, 2 and 18, 3 and 12, 4 and 9, and 6. This simple observation solves the famous prison warden puzzle: A prison warden and 100 inmates, residing in cells numbered 1 through 100, agree to perform the following experiment over 100 days. In the process of the experiment some cell doors will be left unlocked and the prisoners agree not to escape. On the first day, the prison warden will turn the key of each cell door and leave all the doors unlocked. On the second day, the warden will turn the key of every second door. This will lock doors numbered 2, 4, …, 100 and leave the oddnumbered doors open.
factor The term factor is used in two mathematical settings: NUMBER THEORY and ALGEBRA. In number theory, if a, b, and n are whole numbers and if a times b equals n, then a and b are called factors of n. For example, 3 and 4 are both factors of 12 (since 12 = 3 × 4), as are the numbers 1, 2, 6, and 12 (2 × 6 = 12 and 1 × 12 = 12). Any number that divides the given number evenly is a factor. For this reason, factors are sometimes called divisors. 185
186 factorial On the third day, starting with door number 3, he will turn the key of every third door. This will leave a mixture of doors locked and unlocked. On the fourth day he will turn the key of every fourth door, on day five the key of every fifth door, and so on, all the way until day 100, when he will turn the key of every 100th door, namely, just the final door. At this time, any prisoner who finds his door open will be allowed to go free. Which doors will be left unlocked as a result of this experiment? Observe that the warden will turn the key to door number n on each day d that is a factor of n. For example, the key of door number 39 will be turned on days 1, 3, 13, and 39 only, and the key of door number 25 on days 1, 5, and 25. For a door to be left unlocked at the end of the experiment its key must be turned an odd number of times. As only the square numbers have an odd number of factors we have that prisoners 1, 4, 9, 16, 25, 36, 49, 64, 81, and 100 are set to go free. Any number n always has 1 and itself as factors. These are called “improper factors.” Any other factor of n, if there is one, is called a “proper factor.” For example, 1 and 12 are improper factors of 12, and 2, 3, 4, and 6 are proper factors. A number, different from 1, that possesses proper factors is called “composite,” or a COMPOSITE NUMBER, and one that does not is called PRIME. For example, the number 25 has a proper factor, namely 5, and so is composite; whereas 7 has only two factors, both of which are improper, and so is prime. Any proper factor of a composite number, if not itself prime, can be written as a product of two smaller factors. Repeated application of this process shows that any number different from 1, if not already prime, can be written as a product of prime numbers. For example: 180 = 30 × 6 = (6 × 5) × (2 × 3) = 2 × 3 × 5 × 2 × 3. The FUNDAMENTAL THEOREM OF ARITHMETIC asserts that every integer greater than 1 factors into primes in only one way (up to the order of the primes). A factortree is a diagram that describes this factoring process pictorially. Two students drawing different factor trees for the same number will arrive at the same list of primes at the bases of their trees.
In algebra, a factor is a POLYNOMIAL that divides another given polynomial exactly. For example, 2x + 1 and x – 1 are both factors of 2x2 – x – 1, since 2x2 – x – 1 = (2x + 1)(x – 1). It is generally agreed that the factors of a polynomial should themselves be nonconstant polynomials with coefficients that are real numbers. A polynomial that cannot be factored is called irreducible. For example, x2 + 1 cannot be factored over the real numbers and so is irreducible. If one permits COMPLEX NUMBERS as coefficients, then the situation changes: we can write: x2 + 1 = (x + i)(x – i). The FUNDAMENTAL THEOREM OF ALGEBRA asserts that every polynomial factors in the realm of complex numbers. See also COMMON FACTOR; DIVISIBILITY RULES; FACTOR THEOREM; GREATEST COMMON DIVISOR; LONG DIVISION; PERFECT NUMBER; PRIME.
factorial If n is a natural number, then the symbol n! (read as “n factorial”) denotes the product of all positive integers from 1 to n: n! = 1 · 2 · 3·…·n For example, 6! = 1·2·3·4·5·6 = 720. The order in which the terms of the product are multiplied is immaterial. (We have that 6! also equals 6·5·4·3·2·1.) Factorials naturally arise when counting the numbers of ways a collection of objects can be arranged. Given n objects, there are n choices for which object we wish to regard as “first.” Once this selection is made, n – 1 choices remain for which object to select as second, and after the first and second are chosen, there remain n – 2 choices for third. This process continues until two objects remain, yielding two choices for which to regard secondtolast, and just one object to select as last. By the MULTIPLICATION PRINCIPLE, there are thus n · (n – 1) · (n – 2) · … · 2 · 1 = n! different PERMUTATIONs (or arrangements, or orders) of n objects. If one wishes to arrange just r objects chosen from a collection of n different things, there are n choices for which object to regard as first, n – 1 choices for which to regard as second, and so on, to n – r + 1 choices for which object to deem rth, yielding a total of n · (n – 1)· … ·(n – r + 1) different arrangements of r objects selected from n. This formula can be more compactly written as: n! ––— (n – r)!
fair division 187 When r equals n (arrange all n objects) this formula n! reads – 0! . To coincide with our previously computed answer of n!, it is natural to define 0! as equal to 1. The exclamationpoint notation for factorial was first used by Christian Kramp in 1808, in his paper “Élémens d’arithmétique universelle,” though other notations for n! popular at the time, and later, included n_, n′, Π(n) and Γ(n+1). LEONHARD EULER (1707–83) attempted to generalize the factorial function to noninteger values. At the age of 22 he discovered the following LIMIT quantity that helps achieve this: n x n! n→∞ x(x + 1)(x + 2)K (x + n)
Γ(x) = lim
(He called this expression the gamma function to honor ADRIENMARIE LEGENDRE’s use of the symbol Γ for factorial.) This expression has the property, as you may check, that Γ(x + 1) = x · Γ(x) for all positive n ⋅ n! n = lim = 1. real values x. Also Γ(1) = lim n→∞ (n + 1)! n→∞ n + 1 Consequently, Γ(m) = (m – 1)! for all positive integers m. Thus, for example, Γ(7) = 6! = 720. It is now possible to – 1 also define quantities such as ! and (√ 2)! using this 2 expression. INTEGRATION BY PARTS shows that the IMPROPER ∞ INTEGRAL ∫ 0 e –t t m– 1 dt also has value (m – 1)!. Mathematicians have shown that Euler’s gamma function and the corresponding improper integral agree for all positive real values x: +∞
Γ(x) = ∫ 0 e–tt x – 1 dt – 1 This integral has the unexpected value √ π when x = – . 2 1 Thus we may conclude, for example, that − ! = 2 1 1 3 1 1 π Γ = π , and ! = Γ = ⋅ Γ = . 2 2 2 2 2 2 See also BINOMIAL STIRLING’S FORMULA.
THEOREM;
PERMUTATION;
factorization The process or the result of writing a number or a
POLYNOMIAL
as a product of terms is
called factorization. For example, the FUNDAMENTAL THEOREM OF ARITHMETIC asserts that every whole number can be written as a product of PRIME numbers and, up to the order of the terms, this factorization is unique. (For instance, 132 = 2 × 2 × 3 × 11.) Thus every whole number has a unique “prime factorization.” The FUNDAMENTAL THEOREM OF ALGEBRA asserts that, in the realm of COMPLEX NUMBERS, every polynomial factors completely into linear terms. (For instance, 2x3 – x2 – 13x – 6 = (x – 3)(x + 2)(2x + 1) and x2 – 4x + 5 = (x – 2 + i)(x – 2 – i).) If one wishes to remain in the realm of the REAL NUMBERS, then every polynomial with real coefficients is guaranteed to factor into a product of linear terms and irreducible QUADRATIC terms. (For instance, x4 – 1 = (x2 + 1)(x – 1)(x + 1).) See also DECOMPOSITION; FACTOR THEOREM.
factor theorem The
REMAINDER THEOREM shows that if a POLYNOMIAL p(x) is divided by a term of the form x – a for some constant a, then the remainder term is the constant p(a):
p(x) = (x – a)Q(x) + p(a) Thus if the value of the polynomial is zero at x = a, that is, p(a) = 0, then the polynomial factors as p(x) = (x – a)Q(x). This leads to the following factor theorem: A linear term x – a is a factor of a polynomial p(x) if, and only if, p(a) = 0. For example, for p(x) = 2x3 – 4x2 – 10x + 12, we have p(1) = 0, p(–2) = 0, and p(3) = 0. Consequently, x – 1, x + 2, and x – 3 are each factors of the polynomial. (In this example we have p(x) = 2(x – 1)(x + 2)(x – 3).) Since 2 and –2 are clearly each zeros of x6 –64, this polynomial must be divisible by (x – 2)(x + 2) = x2 – 4. See also FUNDAMENTAL THEOREM OF ALGEBRA.
fair division (cake cutting) A classic puzzle asks for a fair way to divide a piece of cake between two greedy brothers. The “you cut, I choose” scheme asks one brother to slice the cake into what he believes to be two equal parts and has the second brother choose one of the two pieces. The first is then guaranteed to receive, in his measure, precisely 50 percent of the cake and the second brother, if he has a different estimation
188 Farey sequence of half, has the advantage of receiving more then 50 percent of the slice according to his own measure. One can avoid this perceived advantage with a different sharing scheme. In the following “knife holding” scheme, each brother is guaranteed to feel that he has the advantage:
Farey sequence (Farey series) For a given positive
Each brother holds a knife vertically across the cake at the location he believes cuts the cake precisely in half. (If the brothers have different estimation of what constitutes “half,” then the knives will be at different locations.) Cutting the cake anywhere between these two positions guarantees each brother a piece, in his estimation, greater than a 50 percent.
0 – 1 1 1 2 1 3 2 3 4 1 – , ,–,–,–,–,–,–,–,–,– 1 5 4 3 5 2 5 3 4 5 1
This scheme generalizes to sharing among any number of players. For instance, it is possible to share a cake among three players in such a manner that each player honestly believes he is receiving more than onethird of the cake. Each player holds a knife vertically across the cake at the location he believes cuts off exactly onethird of the cake from the left. The cake is then cut between the two leftmost positions, giving that piece to the player holding the leftmost knife. This player has received, in his estimation, more than onethird of the cake, and each of the remaining two players believes that more than twothirds of cake remains. The second two brothers then perform the cake division scheme described above for two players. Any cakesharing scheme among n players that guarantees each player, in his estimation, at least 1/n of the cake is called a fairdivision scheme. The scheme described above among three players, however, is not “envy free.” Although each player believes that he received more than his fair share of the cake, it is not assured that he also believes that he received the largest piece ever cut. (Every fair division scheme between two players is envy free.) Complicated envyfree fairdivision schemes do exist for sharing cake among any number of players. There are also fairdivision methods for dividing collections of indivisible objects (such as the furniture in an estate) among two or more people using cash payments to even up the final division.
whole number n, the sequence of all proper fractions, written in reduced form, with denominators no larger than n and arranged in order of magnitude is called the nth Farey sequence. For example, the fifth Farey sequence is:
These sequences have a number of arithmetic properties. a c – For instance, if – b and d are consecutive terms of a Farey sequence, then the numbers ad and bc, arising from taking their cross product, are always consecutive 3 2 – integers. (For example, the consecutive pair – 5 and 3 a – c – e yield consecutive integers 9 and 10.) Also, if – b, d, f are three successive terms of a Farey sequence, then the c a+e ––– middle term – d is the mediant fraction b + f . (For 3 3 3 +3 – 6 – 2 – –—— instance, the term between – 5 and 4 is 5 + 4 = 9 = 3 .) This observation allows one to quickly build up from one Farey sequence to the next: simply compute the mediants between all terms present and retain those whose denominators are not too large. The Farey sequence was first studied by C. Haros in 1802, but interest in the topic did not stir until British geologist John Farey (1766–1826) published his 1816 piece, “On a Curious Property of Vulgar Fractions” in Philosophical Magazine. (Farey was unaware of Haros’s work.) In 1938 American mathematician Lester R. Ford presented a remarkable geometric interpretation of the Farey sequence: a Above each reduced fraction – b on the number 1 – 2 touching the line, draw a circle of radius b number line at that point. Despite expectation, these circles never overlap, although they do often touch. Moreover, two circles at positions a c – – b and d touch precisely when ad and bc are consecutive integers, and furthermore the largest circle that fits in the space between them above the number line is the circle at the a+c mediant –—— b+d.
Fermat’s last theorem 189 (These claims can be proved by making use of the DISTANCE FORMULA to establish that the distance between the centers of two touching circles equals the sum of the radii of the two circles.)
Fermat, Pierre de (1601–1665) French Number theory, Calculus, Probability theory Born on August 17, 1601, in BeaumontdeLomagne, France, Pierre de Fermat is remembered as a leading mathematician in the first half of the 17th century, recognized for his founding work in the theory of numbers. Fermat is also responsible for some pioneering work in CALCULUS and the theory of tangents to curves, PROBABILITY theory, and analytic GEOMETRY. His 1679 piece Isagoge ad locos planos et solidos (On the plane and solid locus), published posthumously, foreshadowed the work of RENÉ DESCARTES (1596–1650) on the application of algebra to geometry, allowing him to define algebraically important curves such as the HYPERBOLA and the PARABOLA. In optics, he is acknowledged as the first scholar to formulate the “fundamental property of least time,” stating that light always follows the shortest paths. Perhaps most notably, Fermat is remembered for the enigmatic comment he scribed in the margin of one of his reading books claiming to have solved a novel problem in number theory. Search for a solution to this problem (if not the one Fermat had in mind) spurred three centuries of important and spirited research in mathematics. FERMAT’S LAST THEOREM was finally resolved in 1994. Fermat received a bachelor’s degree in civil law from the University of Orléans in 1631 and began work as a lawyer for the local parliament of Toulouse that same year. He followed this career path throughout his entire life—accepting a position as a criminal court judge in 1638 and, finally, the high position of king’s counselor in 1648. Fermat’s work in mathematics was an outside interest. Fermat first developed a passion for reading and “restoring” classic Greek texts. This meant completing the mathematics of any passages that were missing from the records that survived from ancient times. His work on the text Plane loci by APOLLONIUS OF PERGA (ca. 262–190 B.C.E.) garnered the attention of the mathematics community at the time, not only for the restoration work itself, but also for the new geometric methods Fermat had devised for computing tangents to curves and solving maxima/minima problems. Fermat
developed a correspondence with French monk MARIN MERSENNE (1588–1648), who served the role of dispersing mathematical information to the notable scholars of the time. Despite the attention Fermat received, he did not seek fame by publishing any of his work. (He published one small piece in his life, which he did anonymously.) Fermat shared his discoveries and results with Mersenne and other scholars, but not his methods for obtaining them. This both inspired and frustrated mathematicians at the time. In 1654 notable scholar BLAISE PASCAL (1623–62) wrote to Fermat with some mathematical questions about gambling and games of chance. The correspondence that ensued led to the joint development of a new mathematical theory of probability. Fermat is today considered one of the founders of the field. However, Fermat had developed a great interest in the theory of numbers, in particular, the properties of whole numbers. This topic was of little interest to mathematicians at the time—perhaps because of its lack of apparent immediate application—but Fermat attempted to spark interest in the subject by posing challenging questions to his contemporaries. He asked scholars to prove, for instance, that the equation x2 + 2 = y3 has only one positive integer solution. His colleagues, however, regarded questions such as these as too specific to be of serious concern and often dismissed then. Fermat, on the other hand, realized that understanding the solutions to such specific questions provides a gateway to great insight on the very general and mysterious properties of whole numbers. It was not until Fermat’s son Samuel published Fermat’s annotated copy of the Arithmetica by the classic scholar DIOPHANTUS OF ALEXANDRIA (ca. 200–284 C.E.)—the text containing the famous marginal note—that interest in number theory was revived and Fermat’s brilliant work on the topic was fully recognized. Fermat died in Castres, France on January 12, 1665. The claim posed in the note scrawled in the margin of Arithmetica is called Fermat’s last theorem. It inspired over three centuries of intense mathematical research in the field of NUMBER THEORY. See also MAXIMUM/MINIMUM.
Fermat’s last theorem Since ancient times, scholars have been aware of many, in fact infinitely many, different integer solutions to the equation x2 + y2 = z2.
190 Ferrari, Ludovico (Solutions to this equation are called PYTHAGOREAN Around 1637 PIERRE DE FERMAT conjectured that no positive integer solutions exist, however, for the equations
TRIPLES.)
xn + yn = zn with n greater than two. In his copy of the translated works of DIOPHANTUS OF ALEXANDRIA, next to a problem about Pythagorean triples, Fermat wrote his now famous note: On the other hand, it is impossible to separate a cube into two cubes, or a fourth power into two fourth powers, or generally any power except a square into two powers with the same exponent. I have discovered a truly admirable proof of this, but the margin is too narrow to contain it. For over 350 years mathematicians tried to reproduce Fermat’s alleged proof. The claim itself became known as Fermat’s last theorem, and it was one of the greatest unsolved problems of all time. Although the problem lends itself to no obvious practical applications, attempts to solve it helped motivate the development of a great deal of important mathematics. It is generally believed that Fermat did not have a proof of the theorem. In his correspondences with colleagues he mentions only the cases n equals 3 and 4 and provides no details of proof even for those special cases. Fermat, again as a marginal note in his copy of Diophantus’s work, does provide a detailed proof of another challenge posed by Diophantus, one about triangles of rational side length. Although not explicitly mentioned, the proof of the n equals 4 case follows readily from mathematical argument he provides. It is thought that Fermat was aware of this. With the case n = 4 taken care of, it is not difficult to see that one need only study the cases where n is an odd prime. For example, if it is known that x7 + y7 = z7 has no positive integer solutions, then x42 + y42 = z42 can have no positive integer solutions either. (Rewrite the latter equation as (x6)7 + (y6)7 = (z6)7.) In the mid1700s, LEONHARD EULER proved that the equation with n = 3 has no positive integer solutions. The extensive work of MARIESOPHIE GERMAIN (1776–1831) during the turn of the century allowed mathematicians to later show that the theorem holds
for all values of n less than 100. During the 19th and 20th centuries mathematicians developed the fields of algebraic geometry and arithmetic on curves. In 1983, Gerd Faltings proved the socalled Mordell conjecture, an important result with the following immediate consequence: any equation of the form xn + yn = zn with n > 3 has, at most, a finite number of positive integer solutions. This led mathematicians a significant step closer to proving Fermat’s last theorem for all values of n: is it possible to show that that finite number is zero in every case? Finally, in 1995, almost 360 years since Fermat’s claim, the English mathematician ANDREW WILES, with the assistance of Richard Taylor, presented a completed proof of Fermat’s last theorem. It is, not surprisingly, very long and highly advanced, relying heavily on new mathematics of the century. Needless to say, the proof is certainly beyond Fermat’s abilities. Although Wiles’s proof is deservedly regarded as a high point of 20thcentury mathematics, mathematicians still search for a simplified argument.
Ferrari, Ludovico (1522–1565) Italian Algebra Born on February 2, 1522, in Bologna, Italian scholar Ludovico Ferrari is remembered as the first person to solve the QUARTIC EQUATION. He worked as an assistant to GIROLAMO CARDANO (1501–76), who published Ferrari’s solution in his famous 1545 work Ars magna (The great art). Assigned to be a servant at the Cardano household at age 14, Ferrari soon impressed his master with his agile mind and with his ability to read and write. Cardano decided to train Ferrari in the art of mathematics. In exchange, Ferrari helped Cardano prepare his manuscripts. Four years after his arrival, and with the blessing of Cardano, Ferrari accepted a post at the Piatti Foundation in Milan as public lecturer in geometry. Ferrari, however, continued to work closely with Cardano. Ferrari discovered his solution to the quartic equation in 1540, but it relied on the methods of solving the CUBIC EQUATION that had been developed by NICCOLÒ TARTAGLIA (ca. 1499–1557) and revealed to Cardano in secrecy. (Mathematicians at the time were supported by patrons and protected their methods as trade secrets: they were often required to prove their worth by solving challenges no other scholar could solve.) Unable to publish the result without breaking a promise, Ferrari and Cardano felt stymied. However, a few years later, Fer
Fibonacci 191 rari learned that another scholar SCIPIONE DEL FERRO (1465–1526) had also developed methods of solving certain types of cubic equations. Although essentially identical to the work of Tartaglia, Ferrari and Cardano decided to publish the solution to the quartic, attributing the work on the cubic needed to del Ferro, with whom no promise of secrecy had been made. Tartaglia was outraged, and a bitter dispute that lasted for many years ensued between Tartaglia and Ferrari. On August 10, 1548, as was common at the time, Tartaglia challenged Ferrari to an open contest and public debate as an attempt to demonstrate that he was in fact the expert on cubic equations. But it was clear from the contest that Ferrari had a more complete understanding of both cubic and quartic equations. Tartaglia left before the contest was over, and victory was given to Ferrari. He immediately garnered national fame and was given many offers of employment, including a request from the emperor himself to act as royal tutor. Ferrari, however, accepted no position offered at the time, left mathematics, and accepted a lucrative position as tax assessor to the governor of Milan. Ferrari died in Bolgna, Italy, in October 1565 (the exact date is not known) and is remembered in mathematics solely for his work on the quartic equation.
Ferro, Scipione del (Ferreo, dal Ferro) (1465–1526) Italian Algebra Born on February 6, 1465, in Bologna, Italy, Scipione del Ferro is remembered as the first mathematician to solve the CUBIC EQUATION. Unfortunately none of his writings survive today, and we learn of his work chiefly through the manuscripts of GIROLAMO CARDANO (1501–76) and LUDOVICO FERRARI (1522–65). Del Ferro was appointed lecturer in arithmetic and geometry at the University of Bologna in 1496, a position he retained for all his life. Little is known of his academic work. Letters to other scholars at the time suggest that del Ferro studied methods for rationalizing rational expressions, rulerandcompass constructions in geometry, and methods for solving cubic equations. Mathematicians of del Ferro’s time were familiar with the general solution to a QUADRATIC equation of the form ax2 + bx + c = 0. (It should be mentioned, however, that 16thcentury scholars did not use zero as a number in an expression, nor permitted the use of negative numbers. Thus the equation x2 – 3x + 2 = 0,
for instance, was written x2 + 2 = 3x.) Mathematicians also knew that, with the appropriate use of substitution, any cubic equation could be reduced to one of two forms: x3 + ax = b or x3 = ax + b. (Here, again, a and b are positive.) Del Ferro was the first mathematician to solve equations of the first type. Some historians suggest that he was able to solve equations of the second type as well. Del Ferro recorded all his results in a personal notebook, which he bequeathed to his soninlaw Hannibal Nave, also a mathematician. Nave later shared the contents of the notebook with Cardano and Ferrari. After seeing the method of solving the cubic fully explained, Cardano and Ferrari realized that del Ferro had in fact solved the famous cubic equation some 30 years before NICCOLÒ TARTAGLIA (ca. 1499–1557), their contemporary, had claimed to do the same. In 1545 Cardano published Ars magna (The great art), outlining Ferrari’s solution to the QUARTIC EQUATION making use of del Ferro’s methods for the cubic. Del Ferro died in Bologna, Italy, some time between October 29 and November 16, 1526. He is remembered in mathematics solely for his work on cubic equations.
Fibonacci (Leonardo Fibonacci, Leonardo of Pisa) (ca. 1170–1250) Italian Arithmetic, Number theory Born in Pisa, Italy (the exact birth date is not known), mathematician Leonardo of Pisa, better known by his nickname Fibonacci, is best remembered for his help in introducing the HINDUARABIC NUMERAL system to the merchants and scholars of Europe. He strongly advocated the system in his famous 1202 text Liber abaci (The book of counting), a treatise on the techniques and practices of arithmetic and algebra, which proved to be extremely influential. This work also contained a large collection of arithmetical problems, including one that leads to the famous sequence of numbers that bears his name. Considered the most important mathematician of the middle ages, Fibonacci also wrote extensively on the topics of EUCLIDEAN GEOMETRY and DIOPHANTINE EQUATIONs. He is recognized as the first scholar in the West to make advances in the field of NUMBER THEORY since the time of DIOPHANTUS OF ALEXANDRIA. Although born in northern Italy, Fibonacci was raised and educated in northern Africa, where his father, a merchant and a government representative,
192 Fibonacci numbers held a diplomatic post. As he grew older, Fibonacci also traveled to Greece, Egypt, Syria, France, and Sicily, and took special note of the arithmetic systems used by local merchants of those areas. He became convinced that the number system used by the Arabs with its roots from India—the actual way they wrote numbers and the way they manipulated them to perform calculations—was far superior to any other arithmetic system he had encountered, including the clumsy system of ROMAN NUMERALS in use at the time in Europe. Upon his return to Pisa around the year 1200, Fibonacci began writing his famous piece. Its aim was to simply explain the HinduArabic numerals, the role of a PLACEVALUE SYSTEM, and illustrate its superior approach. The text begins simply: These are the nine figures of the Indians: 9, 8, 7, 6, 5, 4, 3, 2, 1. With these nine figures, and with this sign 0, … any number may be written, as will be shown. Divided into four sections, the work outlines the methods of addition, multiplication, subtraction, and division. It also discusses fractions (including a discussion on EGYPTIAN FRACTIONS), as well as some geometry and algebra. (Some parts of the text are written from right to left, indicating, perhaps, the extent to which Fibonacci was influenced by Arabic scholars.) Although complete acceptance of the HinduArabic system in the West did not occur until about 300 years later, Fibonacci’s work in this area is recognized as the first significant step in this direction. Fibonacci also wrote extensively in the fields of number theory, trigonometry, and geometry. It is said that the advances Fibonacci presented in his 1225 piece Liber quadratorum (The book of square numbers) were of such interest and value that they sparked renewed interest in theoretical mathematics and revived Western mathematics from its slumber during the Middle Ages. During his life Fibonacci was recognized as a great scholar. Word of his abilities reached the Emperor Frederick II, seated in Palermo, who invited him to compete against other mathematicians of the day in a mathematical tournament. Fibonacci correctly solved all three challenges put before him, garnering him further attention and fame. In 1240 Fibonacci was
awarded a salary from the city of Pisa in recognition of his services to the community. All of Fibonacci’s texts, and their reproductions, were written by hand. Copies of Liber abaci still survive today. Fibonacci’s name is derived from the shortening of the Latin filius Bonacci, meaning the son of Bonaccio, his father’s family name. During his life, Fibonacci was also known as Leonardo of Pisa or, in Latin, Leonardo Pisano. Sometimes, Fibonacci also identified himself as Leonardo Bigollo, following the Tuscan word bigollo for “a traveler.” By introducing the HinduArabic numeral to Europe, his influence on Western mathematics was profound. He died in the city of Pisa, Italy, likely in the year 1250. (The exact date of death is not known.) See also FIBONACCI NUMBERS.
Fibonacci numbers Any one of the numbers that appears in the sequence 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, …, where each number, after the second, is the sum of the two preceding numbers, is called a Fibonacci number. If Fn denotes the nth Fibonacci number, then we have Fn = Fn–1 + Fn–2 with F1 = F2 = 1. These numbers arise from a famous rabbitbreeding problem described in FIBONACCI’s text Liber abaci: How many rabbits would be produced in the nth month if, starting from a single pair, any pair of rabbits of one month produces one pair of rabbits for each month after the next? (The initial pair of rabbits, for example, do not produce another pair of rabbits until month 3. This same pair produces a new pair for each month thereafter.) In any month, the totality of rabbits present consists of all pairs of the previous month together with all the new offspring. The number of offspring equals the population size of two months previous. Thus the solution to the problem is the sequence described above. Any problem whose nth case solution is the sum of the two previous case solutions produces the Fibonacci
Fields medals 193 sequence. For example, there are Fn ways to climb n – 1 steps, one or two steps at a time (consider beginning the climb with either a single step or a double step). There are also Fn ways to tile a 1 × (n – 1) row of squares with 1 × 1 tiles and 1 × 2 dominoes, and there are Fn sequences of 0s and 1s ndigits long beginning and ending with 1 and containing no two consecutive 0s. Regarding a 1 as “tails” and 0 and “heads,” and ignoring the initial and final 1s, this latter example can be used to show that the PROBABILITY of not getting two heads in a row when tossing a coin n times is Fn+2 . One can also use it to show that there are Fn+2 –– 2n subsets of {1,2,…,n} lacking two consecutive numbers as members. Perhaps the most surprising appearances of Fibonacci numbers occur in nature. The seeds in a sunflower’s head, for example, appear to form two systems of spirals—often with 55 spirals arcing in a clockwise tilt, and 34 spirals with a counterclockwise tilt. (Large species of sunflowers have 89 and 144 spirals, again consecutive Fibonacci numbers.) This appears to be typical of all natural objects containing spiral floret, petal, or seed patterns: pineapples, pinecones, and even the spacing of branches around the trunk of a tree. The botanical name for leaf arrangement is phyllotaxis. It is useful to ask whether it is possible to find a value x so that the sequence 1,x,x2,x3,… satisfies the same recursive relationship as the Fibonacci numbers, namely that every term after the second equals the sum of the two preceding terms. This condition therefore requires x to be a number satisfying the equation 1 + x = x2. By the QUADRATIC formula there are two solutions: φ=
1+ 5 2
τ=
1− 5 2
This pineapple has five diagonal rows of hexagonal scales in one direction and eight rows in the other direction. (Note that the sixth unmarked row of scales in the left picture is not a new row; it is a continuation of the bottommost row of scales.) n
Fn =
φn − τ n 5
=
⎛1+ 5 ⎞ ⎛1− 5 ⎞ ⎜ 2 ⎟ −⎜ 2 ⎟ ⎝ ⎠ ⎝ ⎠
n
5
(It is surprising that this formula yields an integer for every value of n). One can use this result to show that Fn the ratio –– of Fibonacci numbers approaches the Fn –1 value ϕ as n becomes large. This happens to be the GOLDEN RATIO. The properties of the Fibonacci numbers are so numerous that there is a mathematical periodical, The Fibonacci Quarterly, devoted entirely to their continued study. See also PASCAL’S TRIANGLE; POLYOMINO.
field See RING. Fields medals These are prizes awarded to young
It follows that any combination of the form aϕ n + bτ n satisfies the same recursive relation as the Fibonacci sequence. Choosing the constants a and b appropriately, so that the first two terms of the sequence produced are both 1, yields the following formula, called Binet’s formula, for the nth Fibonacci number:
researchers for outstanding achievement in mathematics. The awards are regarded as equivalent in stature to Nobel Prizes (which do not exist for mathematics). “International medals for outstanding discoveries in mathematics” were first proposed at the 1924 International Congress of Mathematicians meeting in Toronto.
194 figurate numbers Canadian mathematician John Charles Fields (1863–1932) later donated funds to support this idea, and the awards were created and named in his honor. It was agreed that two gold medals would be awarded every four years—at each quadrennial meeting of the International Congress of Mathematicians. The first awards were given in 1936 and, following a wartime hiatus, were resumed in 1950. In 1966, due to the significant expansion of mathematical research, it was agreed that up to four medals could be awarded at any given meeting. The award itself consists of a cash prize and a medal made of gold. A picture of ARCHIMEDES OF SYRACUSE, along with the quotation, “Transire suum pectus mundoque potiri” (Rise above oneself and take hold of the world), appears on one side of the medal. On the reverse side is the inscription, “Congregati ex toto orbe mathematici ob scripta insignia tribuere” (the mathematicians of the world assembled here pay tribute for your outstanding work). Following Fields’s wish, the awards are presented in recognition of existing work completed by a mathematician, as well as potential for future achievement. For this reason, the awards are usually given to mathematicians under the age of 40. A board of trustees set up by the University of Toronto administers the awards, and a committee of mathematicians appointed by the International Congress of Mathematicians presents the medals to recipients. Laurent Lafforgue of the Institut des Hautes Études Scientifiques, BuressreYvette, France, and Vladimir Voevodsky of the Institute of Advanced Study, Princeton, New Jersey, were the 2002 recipients of the award. Lafforgue made significant contributions to the socalled Langlands program, a series of farreaching conjectures proposed by Robert Langlands in 1967 that, if true, would unite disparate branches of mathematics. Voevodsky was awarded the prize for his work in algebraic geometry, a field that unites number theory and geometry.
Triangular and square numbers
Other geometric shapes are possible, leading to other sequences of figurate numbers. Figurate numbers were of special importance to the Pythagoreans of sixth century B.C.E. Believing that everything in the universe could be explained by the “harmony of number,” they imparted special importance, even personality, to the figurate numbers. For example, 10, being the sum of the first four counting numbers 1 + 2 + 3 + 4, in their belief united the four elements—earth, water, fire, and air—and so was to be held in the greatest of reverence. (They named this number tetraktys, “the holy four.”) Many arithmetic properties of sums can be readily explained by the figurate numbers. For example, the nth triangular number, Tn, is given by the sum: 1 + 2 + … + n. As two triangular configurations placed together produce an n × (n + 1) array of dots, 2Tn, = n × (n + 1), we have: Tn = 1 + 2 + L + n =
figurate numbers Arranging dots to create geometric figures leads to a class of numbers called figurate numbers. For example, the triangular numbers are those numbers arising from triangular arrangements of dots, and the square numbers those from square arrays.
Properties of figurate numbers
n(n + 1) 2
finger multiplication 195 The nth square number is given by the formula: Sn = n × n = n2. Looking at the diagonals of the square reveals the sum: 1 + 2 +…+ (n – 1) + n + (n – 1) +…+ 2 + 1 = n2 Also hidden in a square is the sum of the first n odd numbers. We thus have:
dral numbers, and the like. The nth cube number is given by the formula n3 . It is a threedimensional cubical array of dots, with each layer being a square array of dots. (Thus n layers of n2 dots gives a total of n × n2 = n3 elements.) The nth tetrahedral number is produced by stacking together the first n triangular numbers: T1 + T2 +…+ Tn. This gives the sequence: 1, 4 = 1 + 3, 10 = 1 + 3 + 6, 20 = 1 + 3 + 6 + 10,…
The sum of the first n odd numbers is n2. Adding one to each of these summands gives the sum of the first n even numbers. Thus we have: The sum of the first n even numbers is n2 + n. There is a nice interplay between the triangular and square numbers. For example, the sum of any two consecutive triangular numbers is always a square number: 1 + 3 equals 4, 3 + 6 equals 9, and 6 + 10 equals 16, for instance. In general: Tn–1 + Tn = Sn The center diagram in the figure at the bottom of page 194 explains why this is the case. Similarly, one can arrange eight copies of the one triangle to form a square with its center removed to prove that 8Tn is always one less than a square number. In the same way, one can also establish that the following combination of three consecutive triangular numbers is always square: Tn–1 + 6Tn + Tn+1. The numbers 1 = T1 = S1, 36 = T8 = S6 and 1225 = T49 = S35 are both square and triangular, as are 41,616 and 1,413,721. Mathematicians have proved, using CONTINUED FRACTIONs, that there are infinitely numbers with this property. The alternate triangular numbers 1, 6, 15, 28, … are sometimes called the Bohlen numbers. The nth Bohlen number x is divisible by n and is the unique x multiple of n satisfying: 1 + 2 +…+ – n = x. In 1796 KARL FRIEDRICH GAUSS proved that every natural number is the sum of at most three triangular numbers, and JOSEPHLOUIS LAGRANGE in 1770 proved that every natural number is the sum of no more than four square numbers. One can extend the scope of figurative numbers to three dimensions to produce the cube numbers, tetrahe
One can prove that the nth tetrahedral number is given 1 by the formula – 6 n(n + 1)(n +2). Both the triangular numbers and the tetrahedral numbers appear as diagonals in PASCAL’S TRIANGLE. The successive stacking of tetrahedral numbers produces hypertetrahedral numbers: 1, 5, 15, 35, … These correspond to fourdimensional geometric arrangements of dots. They also appear as a diagonal in Pascal’s triangle. See also NESTED MULTIPLICATION; PERFECT NUMBER; SQUARE; TRIANGLE.
finger multiplication Having memorized the 2, 3, 4, and 5times tables, there is a popular finger method for computing all values of the 6 through 10times tables. It is based on the following rule for encoding numbers: A closed fist represents 5 and any finger raised on that hand adds 1 to that value. Thus a hand with one finger raised, for example, represents 6. A hand with three fingers raised represents 8. To multiply two numbers between 5 and 10, one then follows these steps: 1. Encode the two numbers, one on each hand. 2. Count 10 for each finger raised. 3. Count the number of unraised fingers on each hand and multiply together the two counts. 4. Add the results of steps two and three. This is the desired product. For example, “6 times 8” is represented as one raised finger on the left hand, three on the right hand. There are four raised fingers in all, yielding the number 40 for step 2. The left hand has four lowered fingers, and the right has two fingers lowered. We compute 4 × 2 = 8. Thus the desired product is 40 + 8 = 48. Similarly, “8
196 finite times 8” is computed as 60 + 2 × 2 = 64, “9 times 7” as 60 + 1 × 3 = 63, and “6 times 7” as 30 + 3 × 4 = 42. Notice that one is never required to multiply two numbers greater than five. That the method works is explained by the algebraic identity:
the differences. For example, for the sequence 1,2,4,8,15,26,42,64,… we obtain the difference table: 1 2 4 8 15 26 42 64 ... 1 2 4 7 11 16 22 ...
(5 + a)(5 + b) = 10(a + b) + (5 – a)(5 – b) The identity (N + a)(N + b) = 2N(a + b) + (N – a)(N – b) shows that we can extend this method to use of a different number of digits on each hand. For example, with N = 10, using fingers and toes, one can readily compute 17 × 18 as “seven raised fingers” and “eight raised toes.” Counting each raised digit as 20 (2N) we have: 17 × 18 = 20 × 15 + 3 × 2 = 306. See also EGYPTIAN MULTIPLICATION; ELIZABETHAN MULTIPLICATION; MULTIPLICATION; NAPIER’S BONES; RUSSIAN MULTIPLICATION.
finite Intuitively, a set is said to be finite if one can recite all the elements of the set in a bounded amount of time. For instance, the set {knife, fork, spoon} is finite, for it takes only a second or two to recite the elements of this set. On the other hand, the set of natural numbers {1, 2, 3, …} is not finite, for one can never recite each and every element of this set. Despite our intuitive understanding of the concept, it is difficult to give a precise and direct mathematical definition of a finite set. The easiest approach is to simply define a finite set to be one that is not INFINITE, since the notion of an infinite set can be made clear. Alternatively, since there is a welldefined procedure for mechanically writing down the string of natural numbers 1, 2, 3, …, one can define a finite set to be any set S whose elements can be put in onetoone correspondence with a bounded initial segment of the string of natural numbers. For instance, matching “knife” with 1, “fork” with 2, and “spoon” with 3, the set {knife, fork, spoon} is finite because its elements can be matched precisely with the string of natural numbers {1, 2, 3}.
finite
differences To analyze the terms of a
SEQUENCE,
it can be useful to create a table of successive differences (in the sense of “right minus left”) between the terms of the sequence, and subsequent differences of
1 2 3 4 5
6 ...
1 1 1 1 1 ... 0 0 0 0 ... 0 0 0 ... From the pattern that is now apparent, it is clear that the next number in the original sequence will be 64 + 29 = 93. The entries in the first row under the original sequence are said to be the “first finite differences”; the second row under the sequence depicts the “second finite differences,” and so forth. All the terms that appear in a table of finite differences are completely determined by the values that appear in the leading diagonal. For instance, if the values a, b, c, … shown below are known, then the remainder of the table must appear as follows: a a + b a + 2b + c a + 3b +3c + d a + 4b + 6c + 4d + e ... b b + c b + 2c + d b + 3c + 3d + e ... c c + d c + 2d + e ... d d + e ... e ... The coefficients that appear in the top row match the entries in each row of PASCAL’S TRIANGLE, which are given by the BINOMIAL COEFFICIENTs. This suggests that it would be enlightening to examine the finite difference n n n tables of the sequences , , ,L . We obtain: 0 1 2 n : 0
1 1
1
0 0 0 0
1 0
1 0 ...
0 ...
0 0 ...
...
Fisher, Sir Ronald Aylmer 197 n : 1
first and secondderivative tests See 0
1
2
1
1
3
4
1
0 0
...
0
0 0
1 1
1
Fisher, Sir Ronald Aylmer (1890–1962) British
0 ...
3
6
2 1
...
3 ... 1 ...
0 0 ... Each such sequence produces a table with a straightforward leading diagonal: one that is zero in all places except for the appearance of a single 1. Using this to our advantage, recall that the leading diagonal of the sequence 1, 2, 4, 8, 15, 26, 42, 64, … is 1,1,1,1,0,0,0,…, which is the sum of the leading diagonals for the four n n n n sequences given by , , , and . Thus the nth 0 1 2 3 term in our original sequence must equal the sum of the nth terms of each of the four sequences, namely: n n n n n(n − 1) n(n − 1)(n − 2) + + + + = 1+ n + 2 3 0 1 2 3 =
n3 + 5n + 6 6
Thus we now have a formula for the sequence 1, 2, 4, 8, 15, 26, 42, 64,… In general, one can use this technique to find a formula for any sequence whose difference table eventually contains a row of constant finite differences. Not all sequences, however, have this property. For example, the difference table for the sequence of FIBONACCI NUMBERS cycles indefinitely: 1 1 2 3 0 1 1
5 8 13 21 34 ... 2 3 5
0 0 1 1
2 3
MINIMUM.
1 ...
0 0 ... n : 2
MAXIMUM/
8
13 ... 5
...
One must employ alternative techniques to compute formulae for such sequences.
Statistics, Genetics Born on February 17, 1890, in London, England, Sir Ronald Fisher is considered the most important statistician of the early 20th century. His landmark 1925 text Statistical Methods for Research Workers established methods of designing experiments and analyzing results that have been used extensively by scientists ever since. Fisher was also an able geneticist and made significant contributions to the fields of selection and genetic dominance. After obtaining a degree in astronomy from Cambridge in 1912, Fisher developed an interest in the theory of errors in astronomical observation. This work led him to a general interest in statistical problems and the analysis of ERROR in all disciplines, including those arising in biology. In 1919 Fisher accepted a position at the Rothamsted Agricultural Experiment Station as a biologist. There he developed his key ideas in the theory of genetics while also founding the theory of experimental design described in his 1925 piece. Fisher was professor of genetics at University College, London, from 1933 to 1943, and then professor of genetics, University of Cambridge, until 1957. Upon his retirement, Fisher moved to Australia to become a research fellow at the Division of Mathematics and Statistics, CSIRO, Adelaide. His method of multivariate analysis allowed scientists, for the first time, to properly analyze problems involving more than one variable, and his notion of “likelihood” provided the means to draw general conclusions on the basis of relative probabilities of different events. Fisher also contributed to the science of HYPOTHESIS TESTING by identifying and analyzing new key DISTRIBUTIONS. His work, without doubt, transformed statistics from a general science into a practical and powerful scientific tool. He is considered the founder of modern statistics. Fisher was elected a fellow of the ROYAL SOCIETY in 1929, and was awarded the Royal Medal of the Society in 1938, the Darwin Medal of the Society in 1948, and the Copley Medal of the Society in 1955. He was knighted in 1952 in recognition of his influential work in statistics and for his development of a
198 fixed point statistical theory of natural selection. Fisher died in Adelaide, Australia, on July 29, 1962. See also HISTORY OF PROBABILITY AND STATISTICS (essay); KARL PEARSON.
fixed point Any point that is mapped to itself by a given TRANSFORMATION is called a fixed point. For example, the points x = 0 and x = 1 are fixed points for the function f(x) = x2. Any point on the line of REFLECTION for a reflection in a plane is a fixed point for that reflection. Any continuous map f that maps points in the unit interval [0,1] to points in the same interval must possess a fixed point. (By the INTERMEDIATEVALUE THEOREM, the graphs of y = f(x) and y = x for 0 ≤ x ≤ 1 must intersect.) This is a special case of a more general result proven by Luitzen Egbertus Jan Brouwer in 1915 stating that, for all values n, any continuous map f:[0, 1]n → [0,1]n must possess at least one fixed point. (For n = 2, [0,1]2 = [0,1] × [0,1] is the unit square in the plane, and for n = 3, [0,1]3 = [0,1] × [0,1] × [0,1] is a unit cube in threedimensional space.) This theorem has the following amusing consequences:
The ceiling function, also known as the leastinteger function, takes a real number x and returns the least integer not smaller than x. This quantity is denoted: x. For example, 7.2 = 8, 7.998 = 8 and 7 = 7. Also, –6.34 = –6. The fractional part of a real number x, denoted {x}, is given by: {x} = x – x. For example, {7.2} = 0.2, {7.998} = 0.998 and {7} = 0. Also, {–6.34} = 0.66. The names “floor” and “ceiling,” as well as the notation for these functions, were introduced by Kenneth Iverson in his 1962 computer science text A Programming Language. These functions often arise in N applications of counting. For example, there are 4 multiples of 4 less than, or equal, to N, and, for any two real numbers x and y with x < y, the closed interval [x, y] contains y – x + 1 integers. See also DAYSOFTHEWEEK FORMULA.
floor function See
FLOOR/CEILING/FRACTIONAL PART
FUNCTIONS.
fluxion In his version of
Consider two square sheets of paper, one lying directly on top of the other. Initially each point of the upper sheet lies directly above its corresponding point on the lower sheet. Now crumple the top sheet and rest the crumpled ball anywhere on the lower sheet. By the Brouwer fixedpoint theorem there is still at least one point of the crumpled sheet lying directly above its corresponding point on the bottom sheet. As a thought experiment, imagine the molecules of the liquid in a cup of coffee as the points in a threedimensional cube. After the coffee is stirred, the Brouwer fixed point theorem assures that at least one molecule will return to its original location.
CALCULUS, SIR ISAAC NEWthought of variable x as a flowing quantity, or a fluent, and called the rate of change of x the “fluxion” . of x. He used the notation x for the fluxion of x, x¨ for the fluxion of the fluxion of x, and so forth. Thus if x = f(t), where x is the distance and t the time for a . moving object, then x is the instantaneous VELOCITY of the object, and x¨ its instantaneous ACCELERATION. Today the term fluxion is considered obsolete, and we use the word DERIVATIVE in its stead. The raiseddot notation for derivative, however, is still used by physicists for denoting derivatives with respect to time. See also CALCULUS; HISTORY OF CALCULUS (essay). TON
focal chord Any
See also ITERATION.
floor/ceiling/fractional part functions The floor function, also known as the greatestinteger function, takes a real number x and returns the greatest integer not exceeding x. This quantity is denoted: x. For example, 7.2 = 7, 7.9998 = 7 and 7 = 7. Also, –6.34 = –7.
of a conic curve—a or a HYPERBOLA—that passes through a FOCUS of the conic is called a focal chord. See also CONIC SECTIONS; FOCAL RADIUS. PARABOLA,
an
CHORD
ELLIPSE,
focal radius Any line from the FOCUS of a conic curve—a PARABOLA, an ELLIPSE, or a HYPERBOLA—to a point on the conic is called a focal radius. See also CONIC SECTIONS; FOCAL CHORD.
formula 199
focus (plural, foci) Each
has associated with it one or two special points each called a focus of the conic. See also ELLIPSE; HYPERBOLA; PARABOLA. CONIC SECTION
formal logic (symbolic logic) In mathematics, the systematic study of reasoning is called formal logic. It analyzes the structure of ARGUMENTs, as well as the methods and validity of mathematical deduction and proof. The principles of logic can be attributed to ARISTOTLE (384–322 B.C.E.), who wrote the first systematic treatise on the subject. He sought to identify modes of inference that are valid by virtue of their structure, not their content. For example, “Green and blue are colors; therefore green is a color” and “Cows and pigs are reptiles; therefore cows are reptiles” have the same structure (“A and B, therefore A”), and any argument made via this structure is logically valid. (In particular, the second example is logically sound.) This mode of thought allowed EUCLID (ca. 300–260 B.C.E.) to formalize geometry, using deductive proofs to infer geometric truths from a small collection of AXIOMs (selfevident truths). No significant advance was made in the study of logic for the millennium that followed. This period was mostly a time of consolidation and transmission of the material from antiquity. The Renaissance, however, brought renewed interest in the topic. Mathematical scholars of the time, Pierre Hérigone and Johann Rahn in particular, developed means for representing logical arguments with abbreviations and symbols, rather than words and sentences. GOTTFRIED WILHELM LEIBNIZ (1646–1716) came to regard logic as “universal mathematics.” He advocated the development of a “universal language” or a “universal calculus” to quantify the entire process of mathematical reasoning. He hoped to devise new mechanical symbolism that would reduce errors in thinking to the equivalent of arithmetical errors. (He later abandoned work on this project, assessing it too daunting a task for a single man.) In the mid1800s GEORGE BOOLE succeeded in creating a purely symbolic approach to propositional logic, that part which deals with inferences involving simple declarative sentences (statements) joined by the connectives: not, and, or, if … then…, iff
(These are called the NEGATION, CONJUNCTION, DISJUNCand the BICONDITIONAL, respectively.) He successfully applied it to mathematics, thereby making a significant step to achieving Leibniz’s goal. In 1879 the German mathematician and philosopher Gottlob Frege constructed a symbolic system for predicate logic. This generalizes propositional logic by including QUANTIFIERs: statements using words such as some, all, and, no. (For example, “All men are mortal” as opposed to “This man is mortal.”) At the turn of the century DAVID HILBERT sought to devise a complete, consistent formulation of all of mathematics using a small collection of symbols with welldefined meanings. English mathematician and philosopher BERTRAND RUSSELL, in collaboration with his colleague ALFRED NORTH WHITEHEAD, took up Hilbert’s challenge. In 1925 they published a monumental work. Beginning with an impressively minimal set of premises (“selfevident” logical principles), they attempted to establish the logical foundations of all of mathematics. This was an impressive accomplishment. (After hundreds of pages of symbolic manipulations, they established the validity of “1 + 1 = 2,” for example.) Although they did not completely reach their goal, Russell and Whitehead’s work has been important for the development of logic and mathematics. Six years after the publication of their efforts, however, KURT GÖDEL stunned the mathematical community by proving Hilbert’s (and Leibniz’s) goal to be futile. He demonstrated once and for all that any formal system of logic sufficiently sophisticated to incorporate basic principles of arithmetic cannot attain all the statements it hopes to prove. His results are today called GÖDEL’S INCOMPLETENESS THEOREMS. The vision to reduce all truths of reason to incontestable arithmetic was thereby shattered. Understanding the philosophical foundations of mathematics is still an area of intense scholarly research. See also ARGUMENT; DEDUCTIVE/INDUCTIVE REASONING; LAWS OF THOUGHT. TION, CONDITIONAL,
formula Any identity, general rule, or general expression in mathematics that can be applied to different values of one or more quantities is called a formula. For example, the formula for the area A of a circle is: A = πr2
200 foundations of mathematics where r represents the radius of the circle. The QUADRATIC formula for the roots of a quadratic equation of the form ax2 + bx + c = 0 is:
x=
−b ± b2 − 4ac 2a
foundations of mathematics The branch of mathematics concerned with the justification of mathematical rules, AXIOMs, and modes of inference is called foundations of mathematics. The paradigm for critical mathematical analysis came from the work of the great geometer EUCLID (ca. 300–260 B.C.E.) who, in his work THE ELEMENTS, demonstrated that all geometry known at his time can be logically deduced from a small set of selfevident truths (axioms). LEONHARD EULER (1707–83) produced fundamental results in disparate branches of mathematics and often saw connections between those branches. He too searched for small collections of concepts that were fundamental and, hopefully, common to all fields. In the late 1800s and at the turn of the century with the discovery of RUSSELL’S PARADOX in SET THEORY, mathematicians were led to apparent paradoxes and inconsistencies within the seemingly very basic notions of “set” and “number.” This led to the fervent study of the fundamental principles of elementary mathematics and even to the study of the process of mathematical thinking itself (FORMAL LOGIC). In the 1930s Austrian mathematician KURT GÖDEL (1906–78) stunned the mathematical community by proving, essentially, that any formal system of mathematics that incorporates the principles of arithmetic will contain statements that can neither be proved nor disproved, and, in addition, such a system will necessarily be incapable of establishing that it is free from CONTRADICTIONs. Despite these disturbing conclusions, the study of the founding principles of mathematics is still an active area of research today. See also GEORG CANTOR; GÖDEL’S INCOMPLETENESS THEOREMS; BERTRAND ARTHUR WILLIAM RUSSELL; ALFRED NORTH WHITEHEAD; ERNST FRIEDRICH FERDINAND ZERMELO. fourcolor theorem For centuries, cartographers have known that four colors suffice to color any geographical map (that is, any division of the plane into regions). It is
required that regions sharing a common length of boundary be painted different colors (but two regions meeting at a point, such as the states Arizona and Colorado on a map of the United States, may be painted the same tint). Cartographers had also observed that the same is true for any map drawn on a SPHERE (the globe). The question of whether this observation could be proved true mathematically was first posed by English scholar Francis Guthrie in 1852. Mathematicians AUGUSTUS DE MORGAN (1806–71) and ARTHUR CAYLEY (1821–95) worked to solve the problem and, in 1872, Cayley’s student Alfred Bray Kempe (1849–1922) produced the first attempt at a proof of the fourcolor conjecture. Unfortunately, 11 years later English scholar Percy Heawood (1861–1955) found that Kempe had made an error in his work. In 1890 Headwood later proved that five colors will always suffice to color a planar map, but the proof that just four will actually suffice eluded him. Heawood also looked beyond just planar and spherical maps and made a general conjecture that if a surface contains g holes (such as TORUS with g = 1 hole, or a sphere with g = 0 holes), then any map drawn on that surface can be colored with 7 + 48g + 1 2 colors, and that there do exist examples of maps on these surfaces that do require precisely this many colors. (The brackets indicate to round down to the nearest integer.) In 1968 two mathematicians, Gerhard Ringel and J. W. T. Youngs, proved Heawood to be correct for all surfaces with two or more holes and for the torus. Unfortunately, their work did not apply to the case of a sphere and of a KLEIN BOTTLE. It was not until the next decade when, in 1976, mathematicians Kenneth Appel and Wolfgang Haken finally established that four colors do indeed suffice to color any map on a sphere (and hence the plane, since, by placing a small hole in the center of one of the regions to be painted, a punctured sphere can be stretched and flattened onto the plane, and, vice versa, a planar region can be stretched and molded into a punctured sphere). Appel and Haken’s proof was deemed controversial at the time, since it used some 1,200 hours of computer time to check nearly 2,000 complicated spa
Fourier series 201 tial arrangements. (It was argued that since no single human being could possibly verify that the computer had completed this work correctly, then the outcome produced is not a valid proof.) Most mathematicians today accept the validity of Appel and Haken’s work and consider the famous fourcolor problem solved. (The search for a simple and elegant proof easily checked by human hand, however, continues.) It is interesting to note that Heawood’s conjecture actually fails for the single case that remains: it is known that six colors suffice to color any map drawn on a Klein bottle (g = 1 ), not seven, as predicted by the formula. See also FLOOR/CEILING/FRACTIONAL PART FUNCTIONS.
Fourier, Jean Baptiste Joseph (1768–1830) French Analysis, Engineering Born on March 21, 1768, French scholar Joseph Fourier is remembered in mathematics for his fundamental contributions to the theory of heat conduction and his study of trigonometric series. His groundbreaking advances in these topics appear in his famous 1822 treatise Théorie analytique de la chaleur (Analytic theory of heat). Today, FOURIER SERIES play a fundamental role in the study of physics and engineering, as well as in the development of theoretical mathematics. Placed in a military school as an orphan at age 10, Fourier exhibited a strong interest in mathematics at this early age. He taught at the same military school for four years before entering the newly established teachertraining school L’École Normale (later renamed L’École Polytechnique) in 1794. There Fourier received instruction in mathematics from JOSEPHLOUIS LAGRANGE (1736–1813) and excelled in all his studies. In 1797, when Lagrange stepped down as chair of analysis and mechanics, Fourier was appointed the new department chair. This position was shortlived, however, for the following year Fourier was assigned to Napoleon’s army in the invasion of Egypt. He was charged to oversee a variety of archeological and scientific investigations, and it was during this period of his life, while stationed in the Egyptian desert, that Fourier developed a fascination for the mathematics of heat transfer. Taking a highly original approach, Fourier analyzed conduction by representing complicated oscillat
ing quantities as sums of simpler components. This led him to his theory of trigonometric series. In 1807 Fourier detailed the results of his work in the paper “On the Propagation of Heat in Solid Bodies,” but the mathematicians of the day, including Lagrange, were not convinced of the validity of his approach. Fourier rewrote the paper in 1811, and even received a prize for its content, but still received criticism from the mathematics community for its lack of rigor and lack of generality. For another 10 years Fourier worked to establish the mathematical foundations of his work and detail the reasoning behind his methods. Finally, in 1822, after the publication of his famous treatise, all criticisms of his mathematics were settled, and Fourier’s approach was accepted as valid and fundamentally important. Fourier was elected to the powerful position of secretary to the Académie of Sciences that same year, and in 1827 he was honored with membership to the ROYAL SOCIETY of London. It has been conjectured that Fourier, ironically, may have contracted the exotic illness myxedema while in Egypt, which, among several symptoms, induces an extreme sensitivity to cold. It is said that Fourier kept his living quarters almost unbearably hot and always bundled himself in many layers of clothing. Fourier died of a heart attack on May 16, 1830. Fourier series have found numerous applications to physics and engineering. Practically every branch of science that contends with the transfer of energy through waves (such as acoustics, seismic studies, wireless communications, spectroscopy) utilizes Fourier’s methods. In a more abstract setting, Fourier series are now seen as just one of infinitely many possible approaches to decomposing elements of a VECTOR SPACE of functions in terms of basis elements. This shift in perspective has led to new insights into the study of quantum mechanics.
Fourier series When a musician plays two strings on a violin simultaneously, just one sound wave reaches our ears—the combined effect of the two notes. Nonetheless, the human brain is able to decode the information it receives to “hear” two distinct notes being played. That is, the human brain is able to recognize complicated sound waves as sums of simpler basic sound waves.
202 Fourier series In the 18th century, French mathematician and physicist JEAN LE ROND D’ALEMBERT (1717–83), and the Swiss mathematician LEONHARD EULER (1707–83) worked to describe complicated vibrations of strings as sums of simpler functions. The Swiss mathematician Daniel Bernoulli (1700–82) of the famous BERNOULLI FAMILY introduced the use of trigonometric functions in this study, an approach that was later fully developed by French mathematician and physicist JEAN BAPTISTE JOSEPH FOURIER (1768–1830), although his work was motivated by the study of heat conduction. Fourier showed that many functions could be represented as infinite sums of sine and cosine functions. The result of writing a function as a sum of trigonometric functions is today called a Fourier series. As the trigonometric functions cycle in value every 2π in RADIAN MEASURE, it is assumed in these studies that the functions under consideration are themselves periodic with period 2π. Assume f(x) is such a function. Then a Fourier series for f is an expression of the form: a f (x) = 0 + a1 cos(x) + a2 cos(2x) + L 2 + b1 sin(x) + b2 sin(2x) + L One finds the values of the constants a0,a1,a2,…b1,b2,… π by integrating. For example, since ∫ –π cos(kx) dx = 0 = π ∫ –π sin(kx) dx, we have: π
∫−π
f (x) dx = ∫
π
−π
a0 dx + 0 +0 +L + 0 + 0 + L 2
π –π
1 yielding: a0 = – π ∫ f(x)dx. Multiplying through by sin(x) and integrating gives:
A wave as a sum of a cosine curve and a sine curve
π
π
∫−π f (x)sin(x)dx = ∫−π
π a0 sin(x) dx + ∫ a1 cos(x)sin(x) dx −π 2 π
+ ∫ a2 cos(2x)sin(x) dx + L −π
π
+ L + ∫ b1 sin(x)sin(x) dx −π
π
+ ∫ b2 sin(2x)sin(x) dx + L −π
= 0 + 0 + 0 + L + b1π + 0 + Lb1 1 π f(x)sin(x)dx. showing that b1 = – π ∫ –π One can show that the functions {1,cos(x),cos(2x), …,sin(x),sin(2x),…} are ORTHOGONAL in the sense that the integral of the product of any two different functions from this set is zero. This observation allows us to compute all the values a0,a1,a2,…,b1,b2,… by this method of multiplying through by a trigonometric function and integrating. We have, in general: a0 =
1 π f (x) dx π ∫−π
an =
1 π f (x)cos(nx) dx π ∫−π
bn =
1 π f (x)sin(nx) dx π ∫−π
Mathematicians have shown that if f(x) and its DERIVATIVE f′(x) are both CONTINUOUS FUNCTIONs, then the expansion f (x) =
a0 + a1 cos(x) + a2 cos(2x) + L 2 + b1 sin(x) + b2 sin(2x) + L
is valid. They have also shown that, if interpreted appropriately, the expansion remains valid even if f or
fractal 203 f′ are not continuous for a finite number of locations in the interval [–π,π]. This shows, for example, that functions that zigzag like a sawtooth, or jump up and down in value like a staircase, for example, can still be well approximated by a sum of trigonometric functions. For example, take f(x) to be the Vshaped function f(x) = x on the interval [–π,π], with this section of graph repeated over the entire number line to produce the picture of a sawtooth. One checks: a0 =
1 π 2 π  x  dx = ∫ xdx = π π ∫−π π 0
0 for n even 2 π an = ∫ x cos(nx) dx = 4 π 0 − πn 2 for n odd bn = 0 giving: x =
π 4 cos(x) cos(3x) cos(5x) cos(7 x) − + + + + L 2 2 2 2 π 1 3 5 7
at least on the interval [–π,π]. Note, as a curiosity, that if we place x = 0 into this formula we obtain the remarkable identity: 1 1 1 π2 – – – – 8 = 1 + 32 + 52 + 72 +… See also ZETA FUNCTION.
fractal If we
the picture of a geometric object by a factor k, then its size changes accordingly: any line of length a becomes a line of length ka, any planar region of area A becomes a planar region of area k2A, and any solid of volume V is replaced by a solid of volume k3V. An object can thus be described as ddimensional if its “size” scales according to the rule: SCALE
new size = kd × old size At the turn of the 20th century, mathematicians discovered geometric objects that are of fractional dimension. These objects are called fractals. One such object is Sierpinski’s triangle, devised by Polish mathematician Vaclav Sierpinski (1882–1969). Beginnning with an equilateral triangle, one constructs it by successively removing cen
tral triangles ad infinitum. The final result is an object possessing “selfsimilarity,” meaning that the entire figure is composed of three copies of itself, in this case each at onehalf scale. If the dimension of the object is d and the size of the entire object is S, then according to the scaling rule above, the size of each scaled piece d
1 is × S . As the entire figure is composed of three 2 d
1 of these smaller figures, we have S = 3 × × S . This 2 ln3 tells us that 2d = 3, yielding d = –– ln2 ≈ 1.58. Thus the Sierpinski triangle is a geometric construct that lies somewhere between being a length and an area. In 1904 Swedish mathematician Nils Fabian Helge von Koch (1870–1924) described a fractal curve constructed in a similar manner. Beginning with a line segment, one draws on its middle third two sides of an equilateral triangle of matching size and repeats this construction ad infinitum on all line the segments that appear. The result is called the Koch curve. It too is selfsimilar: the entire figure is composed of four copies of itself, each at onethird scale. The object has fractal ln3 dimension d = –– ln2 ≈ 1.26. The Cantor set, invented by German mathematician GEORG CANTOR (1845–1918), is constructed from a single line segment, by removing its middle third and the middle thirds of all the line segments that subsequently appear. The result is a geometric construct, resembling nothing more than a set of points, but again with the same selfsimilarity property: the entire construct is composed of two copies of itself, each at onethird scale. The Cantor set is a fractal of dimension ln2 d = –– ln3 ≈ 0.63. It is not large enough to be considered onedimensional, but it is certainly “more” than a disconnected set of isolated points. Fractals also arise in the theory of CHAOS and the study of DYNAMICAL SYSTEMs. French mathematician Gaston Maurice Julia (1892–1978) considered the iterations of functions f that take COMPLEX NUMBERS as inputs and give complex numbers as outputs. If z is a complex number and the set of points f(z), f (f(z)), f (f (f(z))), … are all plotted on a graph, then two possibilities may occur: either the sequence is unbounded, or the points jump about in a bounded region. The set
204 fractal dimension
Fractals
of all values z that lead to a bounded sequence is called the “Julia set” for f. It is a region in the complex plane that is almost always a fractal. Even simple functions such as f(z) = z2 + c, with different values of the constant c, can yield surprisingly varied and beautiful structures. Polishborn French mathematician Benoit Mandelbrot considered, alternatively, all the complex values c for which the repeated application of the function f(z) = z2 + c, beginning with the point z = 0, yields a bounded sequence. Plotting all such points c describes a subset of the complex plane today called the Mandelbrot set. It has an extremely complicated structure. In particular, its boundary is a fractal. The Sierpinski triangle also arises from an iterative procedure known as the “chaos game:” Pick a point at random inside an equilateral triangle (diagram (1) above) and then draw the point halfway between it and one of the vertices of the triangle picked at random. (This point will lie somewhere in the shaded region of diagram (2) above.) Now draw the point halfway between this new point and another vertex picked at random. (This point will lie somewhere in the shaded region of diagram (3) above.) Imagine we continue this process indefinitely. Sierpinski’s triangle represents all possible “final” destinations of this point as this game is played an infinite number of times. Many objects in nature seem to possess the selfsimilarity properties of fractals. For example, the boundary shape of a cloud looks just as irregular under magnification as it does when looked at directly. Scientists have found it possible to assign fractal dimensions to various objects in nature. The study of fractals has
since found applications to the study of crystal formation, fluid mechanics, urban growth, linguistics, economics, and many other diverse areas. Much of the work in fractal geometry was pioneered by Benoit Mandelbrot, who also coined the term fractal.
fractal dimension See FRACTAL. fraction Any number written as a
QUOTIENT, that is, as one number a divided by another b, is a fraction. We write a/b and call the dividend a the numerator of the fraction and the divisor b the denominator of the fraction. It is assumed that b is not zero. It is appropriate to regard fractions simply as answers to division problems. For example, if six pies are shared equally among three boys, then each boy receives two pies. We write: 6/3 = 2. Similarly, if one pie is shared among two boys, then the amount of pie each boy receives is written 1/2. We, appropriately, call this quantity “half.” Clearly if a pies are shared with just one boy, then that boy receives all a pies. This yields the observation:
a Unit Denominator Rule: – 1 =a Every fractional property can be explained with this simple piesharing model. For instance, we have: Cancellation Law: For any nonzero value x and a xa – a – fraction – b we have xb = b . For instance, if a pies are shared among b boys, then doubling the number of pies and doubling the number
fraction 205 of boys does not alter the amount of pie each boy 2a a – receives: – 2b = b . We have, in effect, “cancelled” a common factor from the numerator and denominator of the fraction. This principle also shows that it is possible to express the same fraction in an infinite number 3 – 6 – 9 30 – of equivalent forms. For instance, – 5 , 10 , 15 , 50 , and 120 – 200 all represent the same fraction. In continuing the model of sharing pies, we see that if a1 pies are shared among b boys, then another a2 pies are shared among the same boys, then in effect a1 + a2 pies were shared among those boys. This yields the rule: a1 a2 a1 + a2 Addition of Like Fractions: – + – = ––– b b b thereby providing a method for adding fractions sharing the same denominator. To add fractions with different denominators, convert each fraction to forms that share a COMMON DENOMINATOR and then add. For 8×2 2 3 2 3 – – – ––– instance, to sum – 5 and 8 write 5 + 8 = 8 × 5 + 5×3 16 31 – 15 – – ––– 5 × 8 = 40 + 40 to obtain the answer 40 . This yields the rule: a – c da + bc –– Addition of Unlike Fractions: – b + d = bd Subtraction is performed in a similar manner. If, in sharing pies, one wished to double the amount of pie each boy receives, one could simply double the number of pies available. This suggests the rule: a xa – Product Rule: x × – b = b This shows, along with the cancellation law, multiplying a fraction by its denominator produces its numerator as a result: a ba – a – Denominator Product Rule: b × – b = b = 1 =a The product rule also shows how to multiply for 2 4 – fractions. For instance, to compute – 3 × 7 observe 2 ×4 that 2 × 4 = 3 by the product rule. Multiplying 3 7 7
the numerator and denominator each by three gives 2 3× × 4 2×4 3 , which is ––– 3 × 7 . This process is summarized: 3×7 a – c ac – Multiplication of Fractions: – b × d = bd The following division rule for fractions is proved in a similar manner: Division of Fractions:
a b = a × d = ad b c bc c d
(Multiply the numerator and denominator each by b a to obtain b × c . Now multiply the numerator and d ad denominator each by d to yield – bc .) Types of Fractions In writing a generic fraction a/b, it is often assumed that a and b are each a WHOLE NUMBER. If this is indeed the case, then the fraction is called a “common” fraction (or sometimes a “simple” or a “vulgar” fraction). Each common fraction represents a RATIONAL NUMBER. Although there are infinitely many rational numbers, surprisingly, they occupy absolutely no space on the NUMBER LINE. A common fraction with positive numerator and denominator is called “proper” if its numerator is less than its denominator, and “improper” otherwise. (Thus a proper fraction represents a quantity less than one, and an improper fraction a quantity greater than or equal to one.) A “mixed number” is a number consisting of an integer and a proper fraction. For example, 31/2 is a mixed number. A zero fraction is a fraction with numerator equal to zero. If no pies are shared among b boys, then each boy receives zero pie. Thus every zero fraction is equal to zero: 0/b = 0. An “undefined” fraction is a fraction with denominator equal to zero. Such a fraction is invalid, for it cannot have any meaningful value. (If, for instance, 2/0 had value x, then multiplying through by the denominator yields the absurdity 2 = 0 × x = 0. Also, dividing
206 fractional part function zero by zero could, allegedly, have any desired value. For example, one could argue that 0/0 = 17 by noting that 0 × 17 = 0 is a true statement. There is no single appropriate value for this quantity.) A “unit” fraction, also known as an EGYPTIAN FRACTION, is a proper fraction with numerator equal to one. A “complex” fraction is one in which the numerator or denominator, or both, is a fraction. For example, 3/5 – 4/7 is a complex fraction. The divisionoffractions rule provides the means to simplify complex fractions. See also CANCELLATION; CONTINUED FRACTION; DECIMAL REPRESENTATION; PARTIAL FRACTIONS; PERCENTAGE; RATIO; RATIONALIZING THE DENOMINATOR; REDUCED FORM.
fractional
part
function See
FLOOR/CEILING/
FRACTIONAL PART FUNCTIONS.
frequency In STATISTICS, the absolute frequency of an observed value is the number of times that value appears in a data set. For example, in the sample 4, 6, 3, 4, 4, 1, 7, 3, 7, 9, 8, 4, 4, 5, 2, the absolute frequency of the observation 4 is five, and that of 8 is one. The relative frequency of an observed value is the proportion of times it appears. This is computed by dividing the absolute frequency of the observation by the total number of entries in the data set. For example, the relative frequency of 4 in the above data set is 5/15 = 1/3, and that of 8 is 1/15. One can compute the relative frequency of entries in an infinite data set by making use of a LIMIT. For example, mathematicians have proved that 5.8% of the powers of two 1, 2, 4, 8, 16, 32, 64, 128,… begin with a 7. (The first power to do so is 246.) By this they mean that if one were to examine the first N powers of two, approximately 5.8 percent of them begin with a 7. This approximation is made more exact by taking larger and larger values of N. In physics the term frequency is defined as the number of cycles that occur per unit time in a system that oscillates (such as a pendulum, a wave, a vibrating string, or an alternating current). The symbol used for frequency is usually f, although the Greek letter ν is often employed for the frequency of light or other electromagnetic radiation. A unit of frequency is called a hertz (Hz).
frequency distribution See STATISTICS: DESCRIPTIVE. frequency polygon See STATISTICS: DESCRIPTIVE. frieze pattern (band ornament) A design on an infinite strip that consists of repeated copies of a single motif is called a frieze pattern. (In classical architecture, a frieze is a horizontal structure, usually imprinted with decoration, resting along the top of some columns. The modern equivalent is a horizontal strip of wallpaper used to decorate the top portion of a wall just below the ceiling.) Mathematicians are interested in the SYMMETRY properties of frieze patterns. Each frieze pattern, by definition, is symmetrical under a TRANSLATION T in the direction of the strip. A frieze pattern might also be symmetrical about a horizontal REFLECTION (H), a vertical reflection (V), a ROTATION of 180° about a point in the design (R), a GLIDE REFLECTION (G), or some collection of these five basic transformations. The first frieze pattern shown below possesses all five symmetries. Not all combinations of T, H, V, R, and G, represent the symmetries of a frieze pattern. For instance, a frieze pattern cannot possess the symmetries T and H alone, for the composition of these two symmetries produces a glide reflection, and so the pattern must possess symmetry G as well. Similarly, any frieze pattern that possesses symmetries H and V must also possess symmetry R, since the combined effect of a vertical and a horizontal reflection is a rotation. Reasoning this way, one can prove that there are only seven combinations of symmetry types a frieze pattern could possess. These are:
T alone T and V T and R T and G
A frieze pattern
T, H, and G T, V, R, and G T, H, V, R, and G
function 207 For this reason, mathematicians say that there are only seven possible frieze patterns. A wallpaper pattern is the twodimensional analogue of a frieze pattern. Mathematicians have proved that there are exactly 17 different wallpaper patterns.
frustum If a geometric solid is cut by two parallel planes, then the portion of the solid between the two planes is called a frustum. A frustum is also produced by cutting the solid by one plane parallel to the BASE of the solid. For example, a conical frustum, is the shape produced by slicing the top off a cone (with the cut made parallel to the base). A pyramidal frustum is made by slicing off the top of a pyramid, again with the cut made parallel to the base. The altitude h of a frustum is the distance between the two parallel planes. The volume of a conical or pyramidal frustum is given by the formula:
(
1 h A1 + A2 + A1A2 3
)
where A1 and A2 are the areas of the upper and lower faces. See also CONE; SOLID OF REVOLUTION; VOLUME.
Ftest See STATISTICS: INFERENTIAL.
function (mapping) Any procedure or a rule that assigns to each member of one set X one, and only one, element of another set Y is called a function. For example, the relation “is the mother of” is a function from the set X of all the people of the world to the set Y of all women: each person is “assigned” one, and only one, biological mother. The rule that squares numbers takes members of the set X = {1,2,3,4} to corresponding elements of the set Y = {1,4,9,16}. If f is a symbol used to denote a function from a set X to a set Y, then we write f: X → Y and call X the domain of f and Y the codomain of f. We also say that f “maps” elements of X to elements of Y. If a specific element x of the set X is mapped to the element y in Y, then we write y = f(x). (Mathematicians sometimes call x an “input” and the value y the corresponding “output.”) For instance, if f is the squaring function described
above, then we have f(2) = 4 and f(3) = 9. If “M” denotes the “is the mother of” function, and Lawrence’s mother is Trenyce, then we can write M(Lawrence) = Trenyce. Sometimes mathematicians will use arrows with tabs to emphasize the notion of a mapping and write, for example, M: Lawrence → Trenyce. Notice that not all elements of Y need be the result of a mapping. (Not all women are mothers, for example.) Also, it is possible for two or more elements of X to be mapped to the same element of Y. (Two people could have the same mother, for instance.) The set of all possible outputs f(x) for a given function f is called the range (or the IMAGE) of the function. For example, the range of the “is the mother of” function M, is the set of all women who have given birth. In mathematics one typically studies functions between sets of numbers, usually described by formulae. For example, the squaring function is described by the equation f(x) = x2, or just y = x2. The function y = 3x + 2, for example, describes the rule “assign to any number x the number y two more than three times x.” In this context, x is called the independent variable and y the dependent variable. (Its value is dependent on the value of x.) A GRAPH OF A FUNCTION is a pictorial representation of a function that operates on numbers. Equations, such as y = x2 and y = 3x + 2, in which the dependent variable is expressed in terms of the independent variable, are said to be in “explicit form.” In contrast, expressions for which both the independent and dependent variables appear on one side of the equality sign are said to be in “implicit form.” For instance, the equations 2x – y = 5 and x2 + y2 = 9 are in implicit form. Implicit equations might, or might not, define y as a function of x. For example, the first equation presented above can be rewritten y = 2x – 5, yielding the explicit form of a function. The second relation, however, does not represent a function: there is no unique value of y associated to each value of x. (For example, with x = 0, y could be either 3 or –3.) Any set of ordered pairs (x,y) satisfying some given constraint is called a relation. For example, the set of all ordered pairs (x,y) satisfying the equation x2 + y2 = 9 is a relation. Graphically, this relation represents a circle of radius three. A relation is a function if it is never the case that two distinct ordered pairs have the same xcoordinate value (that is, no single xvalue has two different yvalues associated to it). This condition is sometimes called the “vertical line test.” For
208 function
History of Functions
sented functions as sums of sine and cosine functions, but commented that such representations may be valid only for a certain range of values. This later led German mathematician PETER GUSTAV LEJEUNE DIRICHLET (1805–1859) to propose a more precise definition:
Since the time of antiquity, scholars were interested in identifying rules or relationships between quantities. For example, the ancient Egyptians were aware that the circumference of a circle is related to its diameter via a fixed ratio that we now call PI, and Chinese scholars, and later the Pythagoreans, knew that the three sides of a right triangle satisfy the simple relationship given by PYTHAGORAS’S THEOREM. Although these results were not expressed in terms of formulae and symbols (the evolution of algebraic symbolism took many centuries), scholars were aware that the value of one quantity could depend on the value of other quantities under consideration. Although not explicit, the notion of a “function” was in mind. In the mid1300s French mathematician NICOLE ORESME discovered that a uniformly varying quantity (such as the position of an object moving with uniform velocity, for instance) could be represented pictorially as a “graph,” and that the area under the graph represents the total change of the quantity. Oresme was the first to describe a way of graphing the relationship between an independent variable and a dependent one and, moreover, demonstrate the usefulness of the task. In 1694 German mathematician GOTTFRIED WILHELM LEIBNIZ, codiscoverer of CALCULUS, coined the term function (Latin: functio) to mean the SLOPE of the curve, a definition that has very little in common with our current use of the word. The great Swiss mathematician LEONHARD EULER (1707–83) recognized the need to make the notion of a relationship between quantities explicit, and he defined the term function to mean a variable quantity that is dependent upon another quantity. Euler introduced the notation f (x) for “a function of x,” and promoted the idea of a function as a formula. He based all his work in calculus and ANALYSIS on this idea, which paved the way for mathematicians to view trigonometric quantities and logarithms as functions. This notion of function subsequently unified many branches of mathematics and physics. In 1822 French physicist and mathematician JEAN BAPTISTE JOSEPH FOURIER presented work on heat flow. He repre
This idea is based on the fact that the graph of a function is nothing more than a collection of points (x,y) with no two yvalues assigned to the same xvalue. Cantor’s definition is very general and can be applied not only to numbers but to sets of other things as well. Mathematicians consequently came to think of functions as “mappings” that assign to elements of one set X, called the domain of the function, elements of another set Y, called the codomain. (Each element x of X is assigned just one element of Y.) One can thus depict a function as a diagram of arrows in which an arrow is drawn from each member of the domain to its corresponding member of the codomain. The function is then the complete collection of all these correspondences. Advanced texts in mathematics today typically present all three definitions of a function—as a formula, as a set of ordered pairs, and as a mapping—and mathematicians will typically work with all three approaches. See also GRAPH OF A FUNCTION.
example, the points (0,–3) and (0,3) on the circle of radius three have the same firstcoordinate value, and the vertical line at x = 0 intercepts the graph of the circle at two locations. The circle is a relation, but it is not a function. A function f:X → Y is said to be “onto” (or “surjective”) if every element of the codomain Y is the
image of some element of X. (That is, every element y in Y is of the form y = f(x) for some x in X.) For example, the squaring function thought of as a map from the set {1,2,3,4} to the set {1,4,9,16} is onto. It is not onto, however, when thought of as a map from the set of all real numbers to the set of all real numbers: no negative number results from the squaring operation.
A function is a correspondence that assigns a unique value of a dependent variable to every permitted value of an independent variable. This, on an elementary level, is the definition generally accepted today. In the late 19th century, German mathematician GEORG CANTOR (1845–1918) attempted to base all of mathematics on the fundamental concept of a SET. Because the terms variable and relationship are difficult to specify, Cantor proposed an alternative definition of a function: A function is a set of ordered pairs in which every first element is different.
fundamental theorem of algebra 209 A function f:X → Y is said to be “onetoone” (or “injective”) if no two different elements of X yield the same output. For example, the “is the mother of” function is not onetoone: two different people could have the same mother. The squaring function, thought of as a map from the set {1,2,3,4} to the set{1,4,9,16}, is onetoone. It is not onetoone, however, when thought of as a number from the set of all real numbers to the set of all real numbers: the numbers 2 and –2, for example, yield the same output. A function that is both onetoone and onto is called a “bijection” (or sometimes a PERMUTATION). A bijection f:X → Y has the property that each element of Y “comes from” one, and only one, element of X. Thus it is possible to define the inverse function, denoted f–1: Y → X, which associates to each element y of Y the element of X from whence it came. Algebraically, if the dependent variable y is given as an explicit formula in terms of x, then the inverse function determines the independent variable x as a formula in terms of y. For example, the inverse of the function y = 3x + 2 is given by x = (y – 2)/3. (That is, the inverse operation of “tripling a number and adding two” is to “subtract two and then divide by three.”) The roles of the variables x and y have switched, and thus the graph of the inverse of a function can be found by switching the xand yaxes on the graph of the original function. In the theory of CARDINALITY, bijections play a key role in determining whether or not two sets X and Y have the same “size.” See also ALGEBRAIC NUMBER; HISTORY OF FUNCTIONS (essay).
fundamental theorem of algebra (d’Alembert’s theorem) The following important theorem in mathematics is deemed fundamental to the theory of algebra: Every polynomial p(z) = anzn + an–1zn–1 +…+ a1z + a0 with coefficients ai either real or complex numbers, an ≠ 0, has at least one root. That is, there is at least one complex number α such that p(α) = 0. By the FACTOR THEOREM we must then have p(z) = (z – α)q(z) for some polynomial q(z) of degree n – 1. Applying the fundamental theorem of algebra to the polynomial q(z), and again to each polynomial of degree greater than one that appears, shows that the
polynomial p(z) factors completely into n (not necessarily distinct) linear factors. We have as a consequence: Every polynomial p(z) = anzn + an–1z n–1 +…+ a1z + a0 with coefficients ai either real or complex numbers, an ≠ 0, factors completely as p(z) = an(z – α1)(z – α2)…(z – αn) for some complex numbers α1, α2, …, αn. Consequently, in the FIELD of complex numbers, every polynomial of degree n has precisely n roots (when counted with multiplicity). For instance, the polynomial z4 – 2z3 + 2z2 – 2z + 1 factors as (z – 1)(z – 1) (z – i)(z + i) with the root 1 appearing twice. Mathematicians call a field algebraically closed if every degreen polynomial with coefficients from that field has precisely n roots in that field. The set of complex numbers is thus algebraically closed. (The field of real numbers, however, is not. The polynomial p(x) = x2 + 1, for instance, does not factor within the reals.) The fundamental theorem of algebra was first conjectured by Dutch mathematician Albert Girard in 1629 in his investigation of imaginary roots. CARL FRIEDRICH GAUSS (1777–1855) was the first to prove the result in his 1799 doctoral thesis. He later reproved the result several times throughout his life using a variety of different mathematical approaches, and he gave it the name the “fundamental theorem of algebra.” In France, the result is known as d’Alembert’s theorem to honor the work of JEAN LE ROND D’ALEMBERT (1717–83) and his many (unsuccessful) attempts to prove it. To prove the theorem, it suffices to consider a complex polynomial with leading coefficient equal to one: p(z) = zn + an –1z n– 1 +…+ a1z + a0. (Divide through by an if necessary.) Notice that if a0 = 0, then the polynomial has one root, namely z = 0, and there is nothing more to establish. Suppose then that a0 is a complex number different from zero. Using EULER’S FORMULA, regard the variable z as a complex number of the form z = Reiθ = R(cosθ + isinθ), where R is a nonzero real number and θ is an angle. Notice that as θ varies from zero to 360°, z = R(cosθ + isinθ) traces a circle of radius R and zn = Rneinθ = Rn(cos(nθ) + isin(nθ)) wraps around a circle of radius Rn n times. Notice, too, that if R is large, then p(z) = zn + an−1 a a + L + n1−1 + 0n an –1z n– 1 +…+ a1z + a0 = zn 1 + z z z is well approximated as zn(1 + 0 +…+ 0 + 0) = zn. Thus
210 fundamental theorem of arithmetic as θ varies from zero to 360°, again for R large, p(z) closely traces a circle of large radius Rn (again winding around n times). If R is sufficiently large, this circle is sure to enclose the point (0,0) in the complex plane. If R shrinks to the value zero, then the trace of p(z) as θ varies is a circle of radius 0 about the point p(0) = a0. That is, the trace of p(z) is the point a0 in the complex plane. Between these two extremes, there must be some intermediate value of R for which the trace of p(z) passes through the origin (0,0). That is, there is α value on the circle of this radius R for which p(α) = 0. This proves the theorem.
(To see why this is true, suppose that a is not already a multiple of p. Since p is prime, this means that the only factor p and a can have in common is 1. By the EUCLIDEAN ALGORITHM we can thus find numbers x and y so that 1 = px + by. Multiplying through by b gives: b = pbx + aby. The first term in this sum is a multiple of p, and so is the second since ab is. This shows that b must be a multiple of p, if a is not.) Suppose, for example, we found the following two prime factorizations of the same number:
fundamental theorem of arithmetic (unique factor
The quantity on the left is certainly a multiple of 7, which means the quantity on the right is too. By the property described above, this means that one of the factors: 19, 23, 37, or 61 is a multiple of seven. Since each of these factors is prime, this is impossible. In general, this line of reasoning shows that the primes appearing in two factorizations of a number must be the same. (It also shows, for example, that no power of 7 could ever equal a power of 13, and that no power of 6 is divisible by 14.) EUCLID, of around 300 B.C.E., was aware that the prime factorizations of numbers are unique. Note that, in these considerations, it is vital that 1 not be regarded as prime—otherwise every number would have infinitely many different representations as a product of prime factors. (For example, we could write: 6 = 2 × 3 = 1 × 2 × 3 = 1 × 1 × 2 × 3, and so on.) Writing numbers in terms of their prime factorizations helps one quickly identify common factors and common multiples. For example, if a = p1n1 p2n2…pknk and b = p1m1 p2m2…pkmk, with the numbers ni and mi possibly zero (this ensures that each number is expressed via the same list of primes), then the GREATEST COMMON DIVISOR of a and b is the number:
ization theorem) metic asserts:
This fundamental result from arith
Every integer greater than one can be expressed as a product of prime factors in one and only one way, up to the order of the factors. (If the number is already prime, then it is a product with one term in it.) For example, the number 100 can be written as 2 × 2 × 5 × 5. The fundamental theorem of arithmetic asserts that the number 100 cannot be written as a product of a different set of primes. Many elementary school children are familiar with the process of factoring with the aid of a factor tree. It is often taken as selfevident that the prime numbers one obtains as factors will always be the same, no matter the choices one makes along the way to construct the tree. However, a proof of this is required. It is straightforward to see that any number n has, at the very least, some prime factorization: If n is prime, then n is a product of primes with one term in it. If n is not prime, the n can be written as a product of two factors: n = a × b. If both a and b are prime, there is nothing more to do. Otherwise, a and b can themselves be factored. Continue this way. This process stops when all factors considered are prime numbers. That the prime factorization is unique follows from the following property of prime numbers: If a product a × b equals a multiple of a prime number p, then one of a or b must itself be a multiple of p.
7 × 13 × 13 × 29 × 29 × 29 × 41 = 19 × 19 × 23 × 23 × 37 × 61
gcd(a,b) = p1α1 p2α 2…pkαk with each αi the smaller of ni and mi, and the COMMON MULTIPLE of a and b is:
LEAST
lcm(a,b) = p1β1 p2β2…pkβk with each βi the larger of ni and mi. This proves the relationship:
fundamental theorem of calculus 211
lcm(a, b) =
a×b gcd(a, b)
The fundamental theorem of arithmetic is used to prove the irrationality of roots of integers. Precisely, one can show that the square root of a positive integer a is a rational number if, and only if, a is a perfect square (that is, a = b2 for some whole number b). This is established as follows: – – Suppose that √ a is a fraction. Then √ a = n/m for some whole numbers n and m: Squaring and simplifying gives am2 = n2. The prime factorization of n gives a prime factorization of n2 with the property that every prime that appears does so an even number of times. As this is the prime factorization of am2, and the primes that appear in m2 already do so an even number of times, it follows that the primes that appear in the factorization of a do so an even number of times as well. Selecting one prime from each pair of primes that appears produces a whole number b with the property – b2 = a. Thus the assumption that √ a is rational leads us to conclude that a is a perfect square. Consequently, since the number 2, for instance, is not a – perfect square, it cannot be the case that √ 2 is rational. – Thus √ 2 must be an irrational number. (A similar argun– ment establishes that √ a is rational if, and only if, a is a perfect nth power.) See also COMMON FACTOR; COMMON MULTIPLE; FACTOR.
fundamental theorem of calculus The theory of develops methods for calculating two important quantities associated with curves, namely, the slopes of tangent lines to curves and the areas of regions bounded by curves. Although isolated problems dealing with these issues have been studied since the time of antiquity (see HISTORY OF CALCULUS essay), no unified approach or technique for solving them was developed for a very long time. A breakthrough came in the 1670s when GOTTFRIED WILHELM LEIBNIZ in Germany and SIR ISAAC NEWTON in England independently discovered a fundamental inverse relationship between the tangent problem (differentiation) and the area problem (integration). Their result, today known CALCULUS
as the fundamental theorem of calculus, binds together the two parts of the subject and is no doubt the most important single fact in the whole of mathematics. Leibniz and Newton each recognized the importance of the result, developed the ideas that follow from it, and applied the consequent results to solve problems in science and geometry with what can only be described as spectacular success. Their discovery of the fundamental theorem is, in essence, the discovery of calculus. The theorem has two parts. Let y = f(x) be a continuous curve. If A(x) represents the area under the curve from position a to position x, then the RATE OF dA CHANGE of area, –– , is given by: dx dA A(x + h) − A(x) = lim h→0 dx h Noting that the area under the curve between positions x and x + h can be approximated as a rectangle of width h and height f(x), we have A(x + h) – A(x) ≈ dA f (x) ⋅ h f(x)·h. Thus ≈ lim h→0 = f (x) . The details of dx h the argument can be made rigorous to prove: The rate of change of the area under a curve y = f(x) with respect to x is f(x). Loosely speaking, “taking the DERIVATIVE of an INTEreturns the original function.” In particular, the function A(x) is one of the antiderivatives of f(x). The MEANVALUE THEOREM, however, shows that if F (x) is any other antiderivative of f(x), then it must differ from A(x) only by a constant: A(x) = F(x) + c. Noting that A(a) = 0, we obtain: A(x) = F(x) – F(a). In particular, the area under the curve y = f(x) from x = a to x = b is A(b) = F(b) – F(a) . As this area is usually denoted b ∫ a f(x)dx, this establishes the second part of the fundamental theorem of calculus: GRAL
If f(x) is a continuous function, and F(x) is any b antiderivative of f(x), then ∫ a f(x)dx = F(b) – F(a). This second observation is the key result: it transforms the very difficult problem of evaluating areas (via limits of sums) into the much easier problem of finding antiderivatives. For example, to compute the area under the parabola y = x2 from x = 3 to x = 12, simply
212 fundamental theorem of isometries 1 3 is an antiderivative of the x 3 function in question, F′(x) = x2. The desired area is 1 thus (12)3 − 31 (3)3 = 567 units squared. The entire 3 thrust of INTEGRAL CALCULUS is consequently transformed to the study of DIFFERENTIAL CALCULUS employed in reverse. See also ANTIDIFFERENTIATION. note that
F(x) =
vertices A′, B′ and C′ of another congruent triangle. (First reflect along the perpendicular bisector of AA′ to take the point A to the location A′. Next, reflect along a line through A′ to take the image of B to B′. One more reflection may be needed to then send the image of C to C′.) This observation proves the fundamental theorem of isometries: Every isometry of the plane is the composition of at most three reflections.
fundamental theorem of isometries Let ABC be a triangle in the plane. Then the location of any point P in the plane is completely determined by the three numbers that represent its distances from the vertices of the triangle. Any GEOMETRIC TRANSFORMATION that preserves distances is called an isometry. If an isometry takes points A, B, and C to locations A′, B′, and C′, respectively, then it also takes the point P to the unique point P′ with matching distances from A′, B′, and C′. Thus an isometry is completely determined by its effect on its vertices of any triangle in the plane. Three REFLECTIONs, at most, are ever needed to map three vertices A,B, and C of one triangle to three
fuzzy logic In 1965, Iranian electrical engineer Lofti Zadeh proposed a system of logic in which statements can be assigned degrees of truth. For example, whether or not Betty is tall is not simply true or false, but more a matter of degree. Fuzzyset theory assigns degrees of membership to elements of fuzzy sets. These degrees range from 1, when the element is in the set, to zero when it is out of the set. Betty’s membership in the set of tall people is a matter of degree. See also LAWS OF THOUGHT.
G of Pisa. He remained at this post for 3 years before accepting the prestigious position as professor of mathematics at the University of Padua. During his 18 years at Padua, Galileo developed his theories of motion. Using inclined planes to slow the rate of descent, Galileo observed that all objects fall at the same rate of acceleration. From this, using mathematics verified by experiment, he deduced his famous law that the distance traveled by a falling object is proportional to the square of the time of its fall. From this it follows that objects tossed in the air move in parabolic arcs. In 1609 Galileo received word that Dutch scholars had invented a spyglass capable of magnifying images of distant objects. Intrigued, Galileo set to constructing his own version of the spyglass and eventually produced a piece (which he called a “perspicillum”) with a magnification power of eight or nine—far superior to the capabilities of the Flemish telescope. Galileo turned his piece to the heavens and commenced a series of remarkable astronomical discoveries. In his 1610 book Sidereus nuncius (Starry messenger), Galileo described the topography of the Moon, established that the Milky Way is composed of stars, and announced his discovery of four moons orbiting Jupiter. For savvy political reasons, he named these moons the “Medicean stars,” in honor of Cosimo de Medici, the grand duke of Tuscany. This work garnered Galileo considerable fame and, just one month after the book was released, Galileo was appointed chief mathematician at Pisa and “mathematician and philosopher” to the grand duke of Tuscany.
Galilei, Galileo (1564–1642) Italian Mechanics, Astronomy Born on February 15, 1564, in Pisa (now in Italy), Galileo Galilei is remembered for his 1638 Discorsi e dimostrazione matematiche intorno a due nuove scienze (Dialogues on two new sciences) presenting a new approach to the study of kinematics through a combination of experiment and mathematical theory. He formulated and verified the law of acceleration of falling bodies s = 1 qt 2 , established that projectiles 2 follow parabolic paths, and was the first to notice that the period of a pendulum is independent of the weight of the bob. Although not the inventor of the telescope, he was the first to develop a workable design of the device that allowed him to make (outstanding) astronomical observations. In particular, he accrued significant empirical evidence that supported the Copernican theory that the planets travel around the Sun. His belief in this theory brought him in conflict with the Roman Catholic Church and led to his consequent trial and house arrest. Galileo was sent by his parents to the University of Pisa to study medicine at the age of 17, but this course of study never appealed to him. He left the University without obtaining a degree and began teaching mathematics and mechanics in Sienna. In 1586 he wrote La balancitta (The little balance), in which he described the methods of ARCHIMEDES OF SYRACUSE (287–212 B.C.E.) for finding the relative densities of substances with a balance. This piece garnered him significant attention in the scientific community, and in 1589 he was appointed chair of mathematics at the University 213
214 Galilei, Galileo Later that year Galileo turned his telescope toward Saturn, observed its rings, and then turned toward Venus. He discovered that Venus shows phases like those of the Moon, which could only be possible if this planet were to orbit the Sun, not the Earth. Galileo had gathered convincing evidence to support the Copernican theory of the solar system. Although the religious authority at the time viewed this model as antithetical to the Holy Scripture (in which the Earth is perceived as lying at the center of the universe), church leaders were tolerant of scholars who used alternative theories in the guise of mathematical tools for calculating the orbits of the planets. Galileo, unfortunately, went further and later proclaimed in a private communication that he was convinced that the theory
Galileo managed to smuggle his controversial manuscript Dialogo out of Italy, where its sale was banned, for publication in Holland. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
Galileo Galilei, an eminent scientist of the 17th century, was the first to combine theory with experiment to understand dynamics. He established that acceleration due to gravity is independent of mass and that projectiles follow parabolic arcs. Galileo also developed the telescope and made significant, and controversial, astronomical observations. (Photo courtesy of Topham/The Image Works)
was a statement of physical reality. Unfortunately, the letter was intercepted by the church, and for the 15 years that followed, the Catholic Church reviewed Galileo’s work carefully and demanded that he follow church teaching at all times. In 1632, Galileo attempted to publish the text Dialogo (Dialogue) on the theory of planetary motion, cleverly written so as to never actually make the claim that the Copernican system is the physical reality), but the church, feeling that the boundary had been crossed, immediately banned its sale, and tried Galileo for heresy. He was found guilty and was placed under house arrest for the remainder of his life. Galileo did, however, manage to
Galton, Sir Francis 215 publish his famous 1638 piece by having it smuggled out of Italy and printed in Holland. Galileo died on January 8, 1642. He is remembered for having a profound impact on the very nature of scientific investigation, linking together mathematics, speculative philosophy, and physical experiment in the study of the natural world. On October 31, 1992, on the 350th anniversary of Galileo’s death, Pope John Paul II publicly acknowledged that the Catholic Church was in error to convict Galileo of heresy for his theory on the motion of the planets.
ing this time Galois fell in love with the daughter of the resident physician of the prison, StephanieFelice du Motel, but, as letters show, the feelings were not reciprocated. For reasons that are unclear today, Galois fought a duel in her honor soon upon his release. He was wounded in that duel and died the following day, May 31, 1832. Galois’s brother later copied the prison manuscript and sent it to prominent mathematicians of the time. If it were not for his brother’s initiative, the bulk of Galois’s work would have been lost to us today. French mathematician JOSEPH LIOUVILLE (1809–82) published Galois’s seminal piece in 1846.
Galois, Évariste (1811–1832) French Abstract algebra Born on October 25, 1811, in Bourg La Reine, France, scholar Évariste Galois is remembered for his famous results that led not only to the conclusion that there can be no general formula that solves all fifthdegree POLYNOMIAL equations (even though there do exist such formulae for solving QUADRATIC, CUBIC, and QUARTIC EQUATIONs), but more importantly to the classification of which specific equations can be so solved. His seminal 1829 work on the solutions of equations founded the field of GROUP THEORY. Galois enrolled in his first highschool mathematics course at the age of 17 and just a year later published an original paper, “Démonstration d’un théorème sur les fractions continues périodiques” (Proof of a theorem on periodic continued fractions), on the study of CONTINUED FRACTIONs. His teachers at the time, unfortunately, did not understand his work and did not regard him as a gifted student. (Admittedly, Galois had difficulty articulating his sophisticated ideas). Galois hoped to attend the prestigious École Polytechnique, the leading university of Paris at the time, but failed the entrance exam twice. He enrolled, instead, at the École Normale, but was later expelled for objecting to the university policy prohibiting students from joining the Paris rebellion against King Charles X. He joined the Artillery of the National Guard, a radical wing of the military. In 1831, Galois was imprisoned for wearing its uniform when support for the guard was made illegal. Previously Galois had read the work of NIELS HENRIK ABEL (1802–29) on the study of algebraic solutions to equations. The SOLUTION BY RADICALS problem was on his mind and, while incarcerated, Galois wrote a manuscript detailing his thoughts on the problem. Dur
Galton, Sir Francis (1822–1911) British Statistics Born on February 16, 1822, in Sparkbrook, England, Francis Galton is remembered in mathematics for his pioneering work in applying statistical techniques to the analysis of biological problems. Galton’s insights and contributions to the nature of statistical analysis paved the way for the development, and consequent widespread use, of statistics throughout the biological and social sciences in the 20th century. Galton completed a basic bachelor’s degree at Cambridge, taking only enough mathematics courses to meet distribution requirements. (He never received serious training in mathematics.) After inheriting a considerable amount of money from his father, Galton journeyed through southwestern Africa and garnered considerable fame as an intrepid explorer. He developed an interest in the study of human hereditary and began considering the issue of selective breeding as a means to improve the human race. (Such pursuits were deemed acceptable at his time.) Galton also developed a simple mathematical model of ancestral hereditary based on the idea that each parent contributes 1/4 of a genetic trait to a child, each grandparent of 1/16 the trait to the child, and so forth, so that the sum of contributions from all ancestors is unity: 2⋅
1 1 1 1 1 1 + 4⋅ + 8⋅ +L = + + +L = 1 4 16 64 2 4 8
Galton received a knighthood for this work in 1909. The University of London established the Francis Galton Laboratory for National Eugenics in his honor in the late 1800s. The eminent statistician KARL PEARSON
216 gambler’s ruin (1857–1936), much of whose work was greatly influenced by Galton’s studies, was once director of the laboratory. Galton died in Surrey, England, on January 17, 1911. See also HISTORY OF PROBABILITY AND STATISTICS (essay).
gambler’s ruin A class of problems in
in 100 chance that a gambler will win $100, and a one in 1,000 chance that he will win $1000. Casinos are well aware that gamblers are reluctant to stop playing with small profits—especially if a player has had a string of losses and is down to his or her last dollar. The gambler’s ruin shows that in all likelihood, gamblers will end up broke before receiving profits they are content with. See also HARMONIC FUNCTION; RANDOM WALK.
PROBABILITY
theory, all known as problems of gambler’s ruin, contains questions of the following ilk: A gambler with $5 repeatedly bets $1 on a coin toss. Each time he plays he has a 50 percent chance of winning a dollar and a 50 percent chance of losing a dollar. The gambler will play until he either has $10 or has gone broke. What is the probability that he will succeed in doubling his money? Problems like these are typically solved as follows: let P(k) be the probability of reaching the $10 goal when the gambler currently has k dollars in hand. We wish to compute P(5). Note three things: 1. P(0) = 0. (With zero dollars, the gambler has lost and will not win.) 2. P(10) = 1. (With $10 in hand, the gambler has won.) 1 1 P(k − 1) + P(k + 1) P(k − 1) + P(k + 1) = . 2 2 2 (With k dollars in hand, there is a 50 percent chance the gambler will lose a dollar and will have to try to win with k–1 dollars in hand, and a 50 percent chance that he will win a dollar and will continue play with k + 1 dollars in hand.)
3. P(k) =
Thus we have eleven values, P(0),P(1),…,P(10), with the first value equal to zero, the last equal to 1, and all intermediate values equal to the average of the two values around it. One can check that there is only one sequence of numbers satisfying these conditions, 1 2 10 1 namely: 0, –– , –– ,…, –– = 1. This shows that P(1) = –– , 10 10 10 10 2 5 1 P(2) = –– and, in particular, that P(5) = –– = –– . 5 10 2 Notice that with one dollar in hand there is only a 1/10 chance that the gambler will win $10 before going broke. One can similarly show that there is only a one
game theory Game theory is the branch of mathematics that attempts to analyze situations involving parties with conflicting interests for which the outcome of the situation depends on the choices made by those parties. Situations of conflict arise in nearly every realworld problem that involves decision making, and game theory has consequently found profound applications to the study of business competition, economics, politics, military operations, property division, and even the study of personal relationships. Although it is difficult to provide a complete analysis of all the types of games that arise, complete solutions exist for simple “matrix games” (which we describe below) involving a small number of players. These simple games can be used to model morecomplicated multiplayer situations. Game theory was first studied in 1921 by French mathematician Félix Edouard Émile Borel (1871–1956), but the importance of the topic was not properly acknowledged until the 1944 release of the monumental work Theory of Games and Economic Behavior by JOHN VON NEUMANN (1903–57) and Oskar Morgenstern (1902–77). These two scholars completely analyzed situations of conflict satisfying the following basic conditions: i. There are a finite number of players. ii. Each player can select one of a finite number of possible actions. The choice of actions available to each player need not be the same. iii. All players are aware of the actions, and the consequences, others may choose to take. iv. Each player is a rational thinker and will select the action that best suits his or her interests. v. At the play of the game, no participant knows what actions will be taken by the other players. vi. The outcome of the game can be modeled as a set of payments (positive, zero, or negative) to each of the players.
game theory 217 If the sum of payments to all players is zero, then the game is called a zerosum game. This means that any player’s win must be balanced by a loss for some other players. The simplest type of game is a twoplayer zerosum game. The choice of actions and the possible results of the game can be summarized in a “payoff matrix” A, in which the rows represent the possible actions one player, player R, can take; the columns the possible actions of the second player, player C; and the (i,j)th entry of the matrix Aij represents the amount player C must pay to player R if R selects row i and C selects column j in the play of the game. (If Aij is negative, then player R pays C a certain amount.) For example, the following table represents a twoperson zerosum game for which each player can take one of three possible actions:
1 2 3
1 5 1 –3
2 –1 0 –2
3 1 4 2
As a first attempt to analyze this game, consider the actions that player R can take. She hopes that the game will result in the largest positive number possible (in this case 5), for this is the amount player C must pay R if this is the result of the game. Thus player R is tempted to select action 1 (especially since R is aware that player C’s optimal outcome of –3 lies in column 1). Of course, player C is aware that player R is likely to think this way, and would be loath to select column 1 as her action for fear of having 5 be the outcome of the game. Player R is aware that player C will think this way, and will suspect then that player C will choose column 2, knowing the she will likely select row 1, and so R, to foil this plan, is tempted to choose an action different from 1 after all. And so on. As one sees, one can quickly enter a neverending cycle of second guessing. Instead of aiming to maximize her profit from the game, another approach player R could adopt is to minimize her losses. For example, if player R takes action 1, she could potentially lose 1 point (if C happens to choose column 2). If R chooses row 2, the worst outcome would be no gain or loss, and if she chooses row 3, she could potentially be down three
points. Thus the maximal minimum outcome for player R occurs with the choice of row 2. This line of reasoning is called the “maximin strategy.” Similarly, the choice of column 2 results in the minimal maximum outcome in player C’s favor. Thus following a “minimax strategy,” player C would choose column 2, and the outcome of the game is consequently 0. The entry 0 in the above payoff matrix is called a saddle point. It is a minimum in its row and a maximum in its column. If a payoff matrix contains a saddle point, then the optimal strategy for each player is to take actions corresponding to the saddle point. The value of the saddle point, for a game that possesses such a point, is called the value of the game. Not every payoff matrix for a twoperson game possesses a saddle point. For instance, the following game corresponding to two possible actions for each player has no saddle point, and following the minimax or the maximin strategies is not optimal for either player.
1 2
1 1 –3
2 –2 4
(The maximin strategy for player R suggests that she select row 1, and the minimax strategy for player C that she take column 1 yielding the result 1 for the game. Player C, however, can anticipate this and is tempted then to change choice to column 2 to obtain the preferable result of –2. Player R, of course, is aware that player C will likely do this, and so will change her choice to row 2 to obtain the outcome 4, and so forth. The two players again are trapped in an endless cycle of second guessing.) The appropriate strategy in such a game lacking a saddle point is a mixed strategy, in which each player decides to select an action by random choice, appropriately weighted so as to maximize her expected profit from the game. For instance, suppose player R decides to select row 1 with PROBABILITY p and row 2 with probability 1 – p. Under this strategy, if player C chooses column 1, then the expected value of player R’s profit is E1 = 1 × p + (–3) × (1 – p) = 4p – 3. If, on the other hand, player C were to select column 2, then the expected profit for player R is E2 = (–1) × p + 4 × (1 – p) = 4p – 5p. These expected outcomes are equal if we choose p = 7/9. Thus, by selecting row 1 with this
218 Gardner, Martin probability, player R can be guaranteed an expected 7 7 1 profit of 4 − 3 = 4 − 5 = , regardless of what 9 9 9 player C does. Similarly, following a mixed strategy, player C should select column 1 with probability 5/9 so as to minimize her expected loss. In 1928 von Neumann proved the famous “minimax theorem”: If each player adopts her best mixed strategy in a zerosum twoperson game, then one player’s expected gain will equal the other’s expected loss. The shared expected outcome of such a game is called the “value” of the game. Games that are not zerosum games are called variablesum games. The PRISONER’S DILEMMA and CHICKEN are examples of such partialconflict games. In them, one searches for dominant strategies, hoping to encounter an equilibrium for the game. In recent decades the theory of games has successfully been extended to nperson variablesum games and to games with a continuous range of possible actions and strategies. Game theory is now a standard course offered in university economics departments. See also FAIR DIVISION.
Gardner,
Martin (1914– ) American Recreational
mathematics Born on October 21, 1914, in Tulsa, Oklahoma, freelance writer Martin Gardner is regarded today as solely responsible for cultivating and nurturing interest in 20thcentury recreational mathematics. With numerous articles and more than 65 books to his credit, Gardner has achieved worldwide fame as a writer who can make the complex issues of science and mathematics accessible and meaningful to the general audience. Although not a mathematician, Gardner is often credited with having done more to promote and prompt the general pursuit of mathematics than any scholar in the field. Gardner graduated from the University of Chicago with a bachelor’s degree in philosophy in 1936. After serving in the U.S. Navy for four years, Gardner returned to Chicago and began a career as a freelance writer, at first editing and writing short works of fiction. In 1958 Gardner accepted the position of mathematical
games columnist for Scientific American, despite never having taken a mathematics course in college. His first piece explaining the mathematics of a Soma cube, a cube dissection puzzle named after a fictitious addictive drug, illustrated his natural ability to discuss and clarify complex issues with ease. This initial piece garnered him national attention, and Gardner remained a columnist for the publication for 25 years. His articles were extremely influential and have since been collected and republished (multiple times) as books. Wellversed in the practices of illusion and magic, Gardner has also published works on magic, the mathematics of magic, as well as texts that discuss, and debunk, the claims and practices of pseudoscience. He has written two novels, The Flight of Peter Fromm (1994) and Visitors from Oz (1998), and some works on the topics of philosophy and literature. His bestselling work is The Annotated Alice—Alice’s Adventures in Wonderland and through the Looking Glass (1965). Gardner currently lives in Hendersonville, North Carolina, and continues his work as a freelance writer. Despite a lack of formal training in mathematics, Gardner has made some original mathematical discoveries, of a recreational flavor, that have been published in scholarly periodicals.
Gauss, Carl Friedrich (1777–1855) German Number theory, Geometry, Algebra, Analysis, Statistics, Physics, Astronomy Born on April 30, 1777, in Brunswick, Duchy of Brunswick (now Germany), Carl Friedrich Gauss is recognized today as the greatest pure mathematician and physicist of his time. His contributions to both fields were enormous. At the age of 18, he invented the LEAST SQUARES METHOD and made the new discovery that a 17sided regular polygon can be constructed with straightedge and compass alone, signaling that he had accomplished great advances in the theory of CONSTRUCTIBLE numbers. In 1801 he published his masterpiece Disquisitiones arithmeticae (Arithmetical investigations) in which he proved the FUNDAMENTAL THEOREM OF ARITHMETIC, the FUNDAMENTAL THEOREM OF ALGEBRA, and introduced the theory of MODULAR ARITHMETIC. In later work, he proposed general solutions to the problem of determining planetary motion, developed theories of statistics that led to the discovery of the Gaussian DISTRIBUTION, and studied potential theory and electricity and magnetism
Gaussian elimination 219 in physics. He also worked out, but did not publish, the principles of HYPERBOLIC GEOMETRY independently of JÁNOS BOLYAI (1802–60) and NIKOLAI IVANOVICH LOBACHEVSKY (1792–1856), and developed the theory of COMPLEX NUMBERS and complex functions. Gauss demonstrated a talent for mathematics at an early age. (It is said that he astonished his elementary school teachers by summing the integers from 1 to 100 in an instant by observing that the sum amounts to 50 pairs of numbers each summing to 101.) Before leaving high school, Gauss had independently discovered the PRIME NUMBER THEOREM, the BINOMIAL THEOREM, important results in modular arithmetic, and the general arithmetic–geometricmean inequality. He entered the University of Göttingen in 1795 but, for reasons that are not clear to historians, left before completing a degree to return to his hometown of Brunswick. He submitted a doctoral dissertation to the University of Helmstedt on the topic of the fundamental theorem of algebra, which was accepted in absentia. Gauss published his masterpiece Disquisitiones arithmeticae soon afterward. In 1801 Gauss created a sensation in the community of astronomers when he correctly predicted the orbital positions of the asteroid Ceres, discovered January 1, 1801, by Italian astronomer Giuseppe Piazzi. Piazzi had the opportunity to observe an extremely small portion of its orbit before it disappeared behind the sun. Other astronomers published predictions about where it would reappear several months later, as did Gauss, offering a prediction that differed greatly from common opinion. When the comet was observed again December 7, 1801, it was almost exactly where Gauss had predicted. Although Gauss did not reveal his method of prediction at the time, it is known that he used his method of least squares to make the computation. In 1807 Gauss left Brunswick to head the Göttingen observatory, and in 1809 he published his general text on the mathematics of astronomy, Theoria motus corporum coelestium (Theory of the motion of heavenly bodies). While at the observatory, Gauss continued work on mathematics. He began developing a theory of surface curvature and GEODESICs, and studied the convergence of SERIES, integration techniques, STATISTICS, potential theory, and more. In 1818 he was asked to conduct a geodesic survey of the state of Hanover, and 14 years later, he assisted in a project to map the magnetic field
of the Earth. His mathematical work on the theory of differential geometry allowed Gauss to prove a number of properties that the Earth’s field must possess, which allowed him to correctly predict the value of the field at different locations, as well as the location of the magnetic South Pole. (Moreover, he proved mathematically that there can only be two magnetic poles.) Gauss remained at Göttingen for the latter part of his career. He was awarded many honors throughout his life, including election as a foreign member of the prestigious ROYAL SOCIETY of London in 1804 and receiving the Copley Medal from the society in 1838. He also won the Copenhagen University Prize in 1822. Gauss published over 300 significant pieces of work, mostly written in Latin, and kept a large number of unpublished notebooks and correspondences that proved to be as mathematically rich as much of his published work. Gauss died in Göttingen on February 23, 1855. It is impossible to exaggerate the influence Gauss has had on almost every branch of mathematics and mathematical physics. He was a master at solving difficult problems that lay at the heart of complex mathematical ideas, and the very completeness and thoroughness of his work paved the way for significant advances in mathematics. A number of fundamental concepts in number theory, differential geometry, and statistics (such as Gaussian reciprocity, Gaussian curvature, and the Gaussian distribution) are today named in his honor.
Gaussian elimination (pivoting) Named in honor of CARL FRIEDRICH GAUSS (1777–1855), the process of Gaussian elimination provides the means to find the solution (if one exists) for a system of n SIMULTANEOUS LINEAR EQUATIONS in n unknowns by multiplying selected equations with carefully chosen constants and subtracting equations to eliminate variables. The method is best explained with an example. Consider the following system of three linear equations in three unknowns: y + 3z = 0 2x + 4y – 2z = 18 x + 5y + 3z = 14 Interchange the first two equations so that the variable x appears in the first row:
220 Gaussian elimination 2x + 4y – 2z = 18 y + 3z = 0 x + 5y + 3z = 14 Divide the first equation through by 2 so that the coefficient of the leading variable x is 1. x + 2y – z = 9 y + 3z = 0 x + 5y + 3z = 14 Eliminate the appearance of the variable x from the third equation by subtracting the first equation from it. That is, replace the third equation with the equivalent statement: (x + 5y + 3z) – (x + 2y – z) = 14 – 9, that is, 3y + 4z = 5. We have: x + 2y – z = 9 y + 3z = 0 3y + 4z = 5 The second equation contains y as the leading variable with a coefficient of 1. Eliminate y from the third equation by subtracting from it three copies of the second equation, that is, replace the third equation with the equivalent statement, (3y + 4z) – 3(y + 3z) = 5 – 3 · 0, that is, –5z = 5. We now have: x + 2y – z = 9 y + 3z = 0 –5z = 5 Divide the third equation through by –5 so that the leading variable in it is z with a coefficient of 1: x + 2y – z = 9 y + 3z = 0 z = –1 The solution to the system of equations is now easy to compute. By a process of back substitution, we see that z = –1, from which it follows from the second equation that y = –3z = 3, and from the first equation that x = 9 – 2y + z = 9 – 6 – 1 = 2. One checks that this is indeed the solution to the original set of equations.
1. Interchanges two equations 2. Multiplies (or divides) an equation through by a nonzero quantity 3. Adds or subtracts a multiple of one equation from another The process of Gaussian elimination uses elementary row operations to transform a system of linear equations into an equivalent system in “echelon form,” that is, one in which each equation leads, in turn, with one of the variables with coefficient 1. Via the process of back substitution, it is then straightforward to determine the solution to the system of equations. The process illustrated above works for any number of equations with the same number of unknowns. It is possible that during the process of Gaussian elimination, a system of equations might yield an absurd statement (such as 0 = 9, for instance), in which case one would conclude that the system has no solutions, or possibly a vacuous statement (such as, 0 = 0), in which case one would conclude that the system of equations has infinitely many solutions. Note that it is possible to take the process of Gaussian elimination further and reduce a system of equations to a system in which each variable appears just once on each line. For instance, in our example, we obtained: x + 2y – z = 9 y + 3z = 0 z = –1 Subtracting two copies of the second equation from the first yields: x+
– 7z = 9 y + 3z = 0 z = –1
and now adding seven copies of the third equation to the first, and subtracting three copies of the third equation from the second, yields: x y
Some Terminology An elementary row operation is any maneuver on a set of simultaneous linear equations that:
=2 =3 z = –1
The solution to the system is now apparent.
Gaussian elimination 221 Consequences for Matrix Theory Any system of linear equations can be represented via a coefficient MATRIX and a column of constant values. For instance, for our example: y + 3z = 0 2x + 4y – 2z = 18 x + 5y + 3z = 14
1 / 2 0 0 0 1 0 0 0 1 and the act of subtracting the first equation from the first is accomplished by multiplication with the matrix: 1 0 0 0 1 0 −1 0 1
the coefficient matrix, call it A, is given by: 0 1 3 A = 2 4 −2 1 5 3 the column of constant values, call it b, is: 0 b = 18 14 Any elementary row operation performed on the original set of equations corresponds to an operation on the rows of the coefficient matrix A and the column matrix b. For instance, in the example above, our first operation was to interchange the first and second rows. This can be accomplished by multiplying A and b each by the PERMUTATION matrix 0 1 0 1 0 0 0 0 1
In this way one can see that every elementary row operation corresponds to multiplication by an elementary matrix. This observation has an important consequence. The inverse of a square matrix A is simply the product of the elementary matrices that reduce A to the identity matrix. Our example explains this. We used elementary row operations to reduce the system of equations to the equivalent system: x
=2 y =3 z = –1
That is, we found a collection of eight elementary matrices E1, E2, …, E8 such that application of these eight matrices reduced the matrix of coefficients A to the IDENTITY MATRIX I. E8E7E6E5E4E3E2E1A = I
We have: 0 1 0 0 1 3 2 4 −2 1 0 0 2 4 −2 = 0 1 3 0 0 1 1 5 3 1 5 3 0 1 0 0 18 1 0 0 18 = 0 0 0 1 14 14 Similarly, the elementary row operation of dividing the first row through by 2 is accomplished by multiplication with the matrix
If we let B be the matrix E8E7E6E5E4E3E2E1, then we have BA = I, which means that B = A–1, the INVERSE MATRIX to A. (As the DETERMINANT of the identity matrix I is 1, the equation E8E7E6E5E4E3E2E1A = I shows that the determinant of A cannot be zero. Thus, as the study of determinants shows, the matrix A does indeed have an inverse.) Notice that the matrix B is the same elementary row operations applied to the matrix I. This result provides a constructive method for computing the inverse to a matrix A. To compute the inverse of a matrix A, write the matrix A and the matrix I side by side. Perform
222 general form of an equation elementary row operations on the two matrices simultaneously. Reduce A to an identity matrix. The second matrix produced is the inverse matrix to A. As an example, we compute the inverse matrix to the matrix of coefficients above:
⎛ x⎞ x = ⎜ y⎟ ⎜⎜ ⎟⎟ ⎝ z⎠ then the system can be compactly written: Ax = b
0 1 3 1 0 0 2 4 −2 0 1 0 1 5 3 0 0 1
2 4 −2 0 1 0 0 1 3 1 0 0 1 5 3 0 0 1
1 2 −1 0 0.5 0 0 1 3 1 0 0 1 5 3 0 0 1
M
Multiplying through by the inverse matrix A–1 yields:
x=
a −1b
⎛ 2.2 1.2 −1.4⎞ ⎛ 0 ⎞ ⎛ 2⎞ = ⎜ −0.8 −0.3 0.6 ⎟ ⎜ 18⎟ = ⎜ 3⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜⎝ 0.6 0.1 −0.2⎟⎠ ⎜⎝ 14⎟⎠ ⎜⎝ 1⎟⎠
again making the solution to the system apparent. Of course, the work of computing the inverse matrix is equivalent to the original process of Gaussian elimination. This approach, however, has the advantage that it can be readily applied to a different set of constant values b without repeating the elimination process.
general form of an equation A formula that
1 0 0 2.2 1.2 −1.4 0 1 0 −0.8 −0.3 0.6 0 0 1 0.6 0.1 −0.2 Thus
A −1
⎛ 2.2 1.2 −1.4⎞ ⎜ = −0.8 −0.3 0.6⎟ ⎟ ⎜⎜ 0.1 −0.2⎟⎠ ⎝ 0.6
Solving Systems of Linear Equations via Inverse Matrices In our example: y + 3z = 0 2x + 4y – 2z = 18 x + 5y + 3z = 14 if x denotes the column vector of the variables:
describes the general relationship between variables without specifying the constants involved is called the general form of the equation. For example, the general form of a QUADRATIC equation in variable x is ax2 + bx + c = 0. (The equation 2x2 – 3x + 4 = 0, for instance, is a specific quadratic equation.) The formula for the AREA A of a CIRCLE, A = πr2, where r is the radius of the circle, is also an equation in general form.
general linear group (full linear group) The set of all invertible n × n square matrices with real or complex entries forms a GROUP under the operation of MATRIX multiplication. This group is called the nthorder general linear group and is denoted GLn. It is straightforward to see that the four group axioms do indeed hold: Closure. If A and B have inverses, then so does their product AB: (AB)–1 = B–1A–1. Identity. The identity matrix I is invertible and so belongs to this set.
geometric sequence 223 Associativity. This rule holds for all matrices, and hence for invertible matrices in particular. Inverses. The inverse of a matrix A is itself invertible, and so belongs to the set: (A–1)–1 = A. A study of DETERMINANTs and CRAMER’S RULE shows that a matrix is invertible if, and only if, its determinant is not zero. Thus the nthorder general linear group may be defined as the set of all n × n square matrices with nonzero determinants. The subset of all square matrices with determinant equal to 1 forms a subgroup of GLn called the special linear group, denoted SLn.
geodesic The shortest curve connecting two points on a surface, and lying wholly on that surface, is called a geodesic. For example, PYTHAGORAS’S THEOREM shows that a straight line gives the shortest path between two points on a plane, and thus straight lines are the geodesics of a plane. On a SPHERE, geodesics are sections of great circles. (Planes fly along geodesic arcs across the Earth’s surface.) One envisions a geodesic as the path a band of stretched elastic would adopt if held at the two points in question and forced to remain on the surface under study. On a plane there is always only one geodesic between any two given points. On a sphere there are infinitely many geodesics that connect the two poles of the sphere, for example. (If the two points chosen are not antipodal, however, then the geodesic connecting them is unique.) A geodesic dome is a domelike structure made of straightline structural elements held in tension. The straight line segments approximate geodesics of the dome.
geometric distribution See BINOMIAL DISTRIBUTION.
geometric mean See MEAN. geometric sequence (geometric progression) A
SEof numbers in which each term, except the first, is a fixed multiple of the previous one is called a geometric sequence. The constant ratio of terms is
called the common ratio. For example, the sequence 1,3,9,27,… is geometric, with common ratio 3. The 1 1 1 –– –– sequence 1, –– 2 , 4 , 8 ,... is also geometric, with common 1 ratio –– 2. A geometric sequence with first term a and common ratio r has the form: a, ar, ar2, ar3,… The nth term an of the sequence is given by an = arn–1. (It is common, however, to start the count by calling the first term of the sequence the “zeroth term,” so that a0 = ar0 = a and the nth term of the sequence is given by the formula: an = arn.) A geometric sequence can also be described as exponential. The common ratio r is the base of the EXPONENTIAL FUNCTION f(n) = arn. If the value r lies between –1 and 1, then the terms arn of geometric sequence approach the value 0 as n becomes large. If r = 1, then the geometric sequence is the constant sequence a,a,a,… For all other values of r, the sequence diverges. The sum of the terms of a geometric sequence is called a geometric series: a + ar + ar2 + ar3 +… In the study of CONVERGENT SERIES, the ratio test shows that this series sums to a finite value if –1 < r < 1. The value S of the sum can be computed as follows: If S = a + ar + ar2 + ar3 +…, then rS = ar + ar2 + ar3 + ar4 1 +… Subtracting yields: S – rS = a, and so S = –– 1–r 1 In particular we have 1 + r + r2 + r3 + … = –– 1 – r. 1 If r is a fraction of the form –– N (with N ≥ 2), then this formula can be rewritten: 1 1 1 1 1 1 = + + +L = 1 N −1 N N 2 N3 N 1− N This particular expression can also be justified with a physical demonstration, which we illustrate here for the case N = 3:
QUENCE
John takes a piece of paper and tears it into thirds. He hands one piece to Andrea, another
224 geometric series to Beatrice, and tears the third piece into thirds again. He hands one piece to Andrea, a second to Beatrice, and tears the third remaining piece into thirds again. He repeats this process indefinitely. Eventually, John will give all the paper away—half to Andrea and half to Beatrice. But the quantity of paper Andrea receives and the quantity Beatrice receives can also be computed as a third, plus a third of a third, plus a third of a third of a third, and so forth. Thus it 1 1 1 1 –– –– + ... equals –– . must be the case that –– 3 + 9 + 27 2 A repeating decimal can be thought of as a sum of a geometric sequence. For example, the decimal 0.111… 1 1 1 –– –– + ... , which, equals the series –– 10 + 100 + 1000 1 according to the formula above, is –– 9 . (Consequently the 1 repeating decimal 0.999… equals 9 times this, 9 × –– 9, which is 1: 0.999… = 1.) The sum S of just the first n terms of a geometric series, a + ar + ar2 +…+ arn–1 can be computed from the formula S – rS = a – arn. Provided that r ≠ 1, this gives: a + ar + ar 2 + L + ar n−1 = a
1− rn 1− r
(If –1 < r < 1, then rn → 0 as n grows. This again shows 1–0 a –– that a + ar + ar2 + ar3 + … = a · –– 1 – r = 1 – r .) For 1 1 1 –– + … + –– equals example, the sum 1+ –– + 2 4 512 10
1 1− 2 1023 1⋅ = . A similar calculation solves a 512 1 1− 2 famous “chessboard puzzle”: Legend has it that the game of chess was invented for an Indian maharaja who became so delighted with the game that he wanted to reward the inventor with whatever he desired. The inventor asked for nothing more than one grain of rice on the first square of the chessboard, two grains on the second square, four on the third, and so forth, each square containing double the number of grains than the previous square. Given that there are 64 squares on
a chessboard, how many grains of rice did the inventor in fact request? According to the formula, the inventor asked for 1 + 2 1 − 264 + 4 + 8 +…+ 263 grains of rice. This equals 1 ⋅ 1− 2 = 264 – 1 = 18,446,744,073,709,551,615 grains, which is the equivalent of about 25 billion cubic miles of rice, an inconceivable quantity. The inventor fooled the maharaja into making a promise he could not possibly honor. See also ARITHMETIC SEQUENCE; CONVERGENT SEQUENCE; SERIES.
geometric series See GEOMETRIC SEQUENCE. geometric transformation A specified procedure that shifts points in the plane to different positions (and thereby changing the location, and possibly the shapes, of geometric figures) is called a geometric transformation. More precisely, a geometric transformation is a FUNCTION that associates with each point of the plane some other point in the plane. (One may require the function to be onetoone and onto.) For example, the function that shifts each point of the plane one unit to the right is a geometric transformation (called a translation). This transformation preserves the shapes of all geometric figures. Any geometric transformation that preserves distances between points in the plane (and hence the shape and size of geometric figures) is called an isometry or a rigid motion. One that multiplies all distances between points by a constant factor (called the dilation factor) is called a similitude, and a transformation that takes straight lines to straight lines is called a LINEAR TRANSFORMATION. All isometries are linear transformations, for example. These ideas also extend to transformations in threedimensional space. While he never made explicit use of the concept in his writings, the idea of a geometric transformation came from the work of the Greek geometer EUCLID (ca. 300–260 B.C.E.). We list here some classical examples of geometric transformations. Reflection in a Line Given a line l in the plane, a reflection in this line takes a point P on one side of l to the corresponding point P′
geometry 225 on the opposite side of l such that the segment connecting P to P′ is PERPENDICULAR to l and bisected by it. The points on l itself are left unmoved. Any reflection is an isometry that transforms geometric figures to their mirror images. The line used in performing the reflection is called the line of reflection. In a CARTESIAN COORDINATE system, a reflection about the xaxis takes a point with coordinates (x,y) to the point (x,–y), and a refection about the yaxis changes the sign of the xcoordinate: (x,y) becomes (–x,y). The analog of a reflection in a line in threedimensional space is a reflection in a plane.
Translation A geometric transformation that moves all points in the plane a fixed distance in a fixed direction is called a translation. No points are left unmoved by a translation. A translation is an isometry, and all geometric figures are transformed to new figures with the same size, shape, and orientation as the originals. If l1 and l2 are two PARALLEL lines in the plane, d units apart, a reflection in the first line, followed by a reflection in the second, has the same effect as translating all points in the plane a distance of 2d units in a fixed direction perpendicular to the two lines. Thus every translation is equivalent to the COMPOSITION of two reflections. In a Cartesian coordinate system, a translation takes a point with coordinates (x,y) to the point (x + a, y + b) for some fixed values a and b. Rotation A rotation about a point O through an angle θ is the geometric transformation that maps a point P in the plane to the point P′ such that P and P′ are the same distance from O, and the angle POP′ has measure θ. (A counterclockwise turn is applied if θ is positive; a clockwise turn of θ is negative.) Only the location of the point O remains unchanged under a rotation, unless the angle θ is a multiple of 360°, in which case all points are fixed. A rotation is equivalent to two reflections about θ lines that intersect at O making an angle of –– 2 between them. Every rotation is an isometry. The analog of a rotation about a point in threedimensional space is a rotation about a line.
Glide Reflection A reflection in a line followed by a translation in a direction parallel to that line is called a glide reflection. Reflection in a Point A reflection in a point O in the plane is the isometry that takes a point P in the plane to the corresponding point P′ such that O lies at the MIDPOINT of the line segment connecting P and P′. A reflection in a point is equivalent to a rotation of 180° about that point. Dilation A dilation with center O and dilation factor k > 1 is the geometric transformation that leaves O fixed, and moves any point P further away from O, by a factor k, along the ray from O through P. Thus a dilation stretches figures uniformly outward from O. It is possible, for example, to convert a square into a rectangle via a dilation. (A dilation with dilation factor k between O and 1 “shrinks” all points closer to O.) A dilation is not an isometry. Circular Inversion Also called an “inversion in a circle” or a “reflection in a circle,” a circular inversion in a circle, with center O and radius r, takes a point P in the plane a distance d from O, and maps it to the point P′ a distance r2/d from O along the same ray from O through P. Thus points inside the circle are taken outside, and vice versa. Points on the circle itself are left unmoved by the transformation. The image of the center O under a circle inversion is undefined. A circular inversion is not an isometry but proves to be a useful mapping in the study of GEOMETRY. It has the property that circles and straight lines in the plane are converted to new circles and new straight lines. See also FRIEZE PATTERN; FUNDAMENTAL THEOREM OF ISOMETRIES; SYMMETRY; TRANSFORMATION OF COORDINATES.
geometry The branch of mathematics concerned with the properties of space and of figures, lines, curves, and points drawn in space is called geometry. Plane geometry examines objects drawn in a plane (lines, circles, polygons, and the like), solid geometry deals with figures in threedimensional space (polyhedra, lines, planes, and surfaces), and SPHERICAL GEOMETRY studies
226 Germain, MarieSophie
History of Geometry The study of GEOMETRY is an ancient one. Records show that Egyptian and Babylonian scholars of around 1900 B.C.E. had developed sound principles of measurement and spatial reasoning in their architecture and in their surveying of land. Both cultures were aware of PYTHAGORAS’s THEOREM and had developed tables of PYTHAGOREAN TRIPLES. (The Egyptians used knotted ropes to construct “345 triangles” to create RIGHT ANGLES.) Ancient Indian texts on altar construction and temple building demonstrate sophisticated geometry knowledge, and the famous volume JIUZHANG SUANSHU (Nine chapters on the mathematical art) from ancient China also includes work on the Pythagorean theorem. In ancient Greece, mathematical scholars came to realize that many properties of shapes and figures could be deduced logically from other properties. In his epic work THE ELEMENTS the Greek geometer EUCLID (ca. 300–260 B.C.E.) collated a large volume of knowledge on the subject and showed that each and every result could be logically deduced from a very small set of basic assumptions (selfevident truths) about how geometry should work. Euclid’s work was rigorous and systematic, and the notion of a logical PROOF was born. EUCLID’S POSTULATES and the process of logical reasoning became the model of all further geometric investigation for the two millennia that followed. His method of compiling and organizing all mathematical knowledge known at his time was a significant intellectual achievement. Euclid’s rigorous approach was, and still is, modeled in other branches of mathematics. Scholars in SET THEORY, the FOUNDATIONS OF MATHEMATICS, and CALCULUS, for instance, all seek to follow the same process of formal rea
the properties of lines and shapes drawn on the surface of a SPHERE. The word geometry comes from the Greek words ge meaning “earth” and metria meaning “measure.” As the origin of the word suggests, the study of geometry evolved from very practical concerns with regard to the accurate measurement of tracts of land, navigation, and architecture. The Greek mathematician EUCLID (ca. 300–260 B.C.E.) formalized the study of geometry to one of pure logical reasoning and deduction. In his famous work, THE ELEMENTS, Euclid collated a considerable volume of Greek knowledge on the subject and showed that all the results known at the time could be deduced from a very small collection of selfevident truths or AXIOMs, which he stated explicitly. Any result that can be
soning as the correct approach to achieve proper understanding of these topics. The next greatest breakthrough in the advancement of geometry occurred in the 17th century with the discovery of CARTESIAN COORDINATES as a means to represent points as pairs of real numbers and lines and curves as algebraic equations. This approach, described by French mathematician and philosopher RENÉ DESCARTES (1596–1650) in his famous 1637 work La géométrie (Geometry), united the thendisparate fields of algebra and geometry. Unfortunately, Descartes’s interests lay only in advancing methods of geometric construction, not in developing a full algebraic model of geometry. This latter task was pursued by French mathematician PIERRE DE FERMAT (1601–65), who had also outlined the principles of coordinate geometry in an unpublished manuscript that he had circulated among mathematicians before the release of La géométrie. Fermat later published the work in 1679 under the title Isagoge ad locus planos et solidos (On the plane and solid locus). The application of algebra to the discipline provided scholars a powerful new tool for solving geometric problems, and also provided them with a large number of different types of curves for study. Fermat’s work in geometry inspired work on the theory of DIFFERENTIAL CALCULUS and, later, led to the study of “differential geometry” (the application of calculus to the study of shapes and surfaces). This was developed by the German mathematician and physicist CARL FRIEDRICH GAUSS (1777–1855). Neither Descartes nor Fermat permitted negative values for distances. Consequently, neither scholar worked with a full set of coordinate axes as we use them today. The
deduced from these axioms is today described as Euclidean. Euclid’s rigorous approach to geometric investigation remained the standard model of study for two millennia. See also EUCLIDEAN GEOMETRY; EUCLIDEAN SPACE; EUCLID’S POSTULATES; HISTORY OF GEOMETRY (essay); NONEUCLIDEAN GEOMETRY; PARALLEL POSTULATE; PROJECTIVE GEOMETRY.
Germain, MarieSophie (1776–1831) French Number theory, Mathematical physics Born on April 1, 1776, in Paris, France, scholar MarieSophie Germain is remembered in mathematics for her significant contributions to the topic of NUMBER THEORY. Most notably,
Germain, MarieSophie 227
notions of negative distance and negative area were first put forward by SIR ISAAC NEWTON (1642–1727) and GOTTFRIED WILHELM LEIBNIZ (1646–1716), the coinventors of CALCULUS. The 19th century saw other major advances in geometry. It had long been noted that Euclid’s fifth postulate, the socalled PARALLEL POSTULATE, is not necessary for a great deal of geometry. Many Arab scholars of the first millennium attempted, without success, to show that the fifth postulate could be logically deduced from the remaining four (thereby rendering it unnecessary), as did European scholars of the Renaissance. In 1795 Scottish mathematician John Playfair showed that the fifth postulate is equivalent to the statement that, through any point, one can draw one, and only one, line through that point parallel to a given line. (This is today called PLAYFAIR’S AXIOM.) Although not eliminating the need for the fifth postulate, Playfair showed that it could be understood in a more tractable form. In 1829 Russian mathematician NIKOLAI IVANOVICH LOBACHEVSKY (1792–1856) took a bold step and considered a geometric world in which the fifth postulate is false. He assumed that through a given point more than one line could be drawn parallel to a given line. In doing this, Lobachevsky discovered a new, consistent mathematical system free from contradiction, one as logically valid as the geometry of Euclid. (This geometry is today called HYPERBOLIC GEOMETRY.) The philosophical impact of Lobachevsky’s work was enormous: he had shown that mathematics need not be based on a single set of physical truths, and that other equally valid mathematical systems do exist based on alternative, carefully chosen axioms. Lobachevsky had also shown that Euclid’s fifth postulate cannot be established as a consequence of the remaining four axioms: he had pre
sented a valid example of a system in which the first four of Euclid’s postulates hold, but the fifth does not. Surprisingly some of Lobachevsky’s ideas were anticipated well before the 19th century. The great Persian mathematician and poet OMAR KHAYYÁM (ca. 1048–1122) established a number of results that we recognize today as nonEuclidean. These results were later translated into Latin, and extended upon, by Italian priest GIROLAMO SACCHERI (1667–1733). Unfortunately, neither scholar discovered the validity of NONEUCLIDEAN GEOMETRY, as each was focused instead on trying to establish Euclid’s fifth postulate as a consequence of the remaining four. The German mathematician BERNHARD RIEMANN (1826–66) discovered an alternative form of nonEuclidean geometry in which Euclid’s fifth postulate fails in a different way. In a system of SPHERICAL GEOMETRY it is never possible to draw a line through a given point parallel to a given line. Riemann’s contributions to the advancement of geometry were significant. In his famous 1854 lecture “Über die Hypothesen welche der Geometrie zu Grunde liegen” (On the hypotheses that lie at the foundation of geometry), Riemann put forward the view that geometry can be the study of any kind of space of any number of dimensions, and later developed the mathematics needed to properly describe the shape of space. ALBERT EINSTEIN (1879–1955) later used this work to develop his theory of relativity. See also AFFINE GEOMETRY; ARABIC MATHEMATICS; BABYLONIAN MATHEMATICS; JÁNOS BOLYAI; DIMENSION; EGYPTIAN MATHEMATICS; THE ELEMENTS ; GEOMETRIC; TRANSFORMATION; GREEK MATHEMATICS; HISTORY OF EQUATIONS AND ALGEBRA (essay); INDIAN MATHEMATICS; PERSPECTIVE; POSTULATE; PROJECTIVE GEOMETRY; THEOREM; TOPOLOGY.
her work paved the way for other scholars in the field to establish the validity of FERMAT’S LAST THEOREM for all values n < 100, the only substantial step made toward solving the problem during the 19th century. Germain also worked to develop a mathematical theory of elasticity. She is noted as one of the very few women scholars in mathematics before the 20th century. At the age of 13, after reading an account of the life and death of ARCHIMEDES OF SYRACUSE, Germain said she felt impelled to become a mathematician. She pursued her studies by first teaching herself Latin and Greek and then reading the works of SIR ISAAC NEWTON, LEONHARD EULER, and EUCLID. As a woman, Germain was not permitted to enter a university, but she did manage to continue her study of mathematics
by reading course notes borrowed from students who attended the École Polytechnique in Paris. Mathematician JOSEPHLOUIS LAGRANGE (1736–1813) presented a course on ANALYSIS at the university, and Germain submitted to him a thesis on the topic under the pseudonym Louis LeBlanc, a male name. Lagrange was greatly impressed by the contents of the paper and sought out its author. He was not at all perturbed to discover that “M. LeBlanc” was a woman. In fact, he was so impressed with her work that he offered her personal counsel in the study of advanced mathematics. Again under her male pseudonym, Germain later established a correspondence with the German mathematician CARL FRIEDRICH GAUSS (1777–1855), with whom she developed and shared much of her work in
228 glide reflection number theory. Gauss, too, eventually discovered her true identity, but this did not diminish his respect for her as a fine scholar. In 1808 the Institut de France set a competition to find a mathematical theory for the theory of elasticity. Although the necessary mathematics to solve the challenge was not available to scholars at the time, Germain devoted a decade of work toward solving the problem and managed to make significant steps in developing a beginning theory on the subject. Unfortunately, much of her work was ignored by the Institut. Germain never married nor obtained a professional position; she was supported financially by her father throughout her life. Germain died of breast cancer in Paris, France, on June 27, 1831. Her death certificate listed her occupation merely as “property holder.”
glide reflection A reflection in a line followed by a translation in a direction parallel to that line is called a glide reflection. See also GEOMETRIC TRANSFORMATION.
gnomon The Lshaped figure that remains when a PARALLELOGRAM (usually a square) is removed from the corner of a larger similar parallelogram is called a gnomon. For example, in a square 2 × 2 array of dots, the three dots surrounding any corner dot form the shape of a gnomon. (Notice, incidentally, that 4 = 1 + 3.) Adding a gnomon of five dots to the 2 × 2 array produces a square 3 × 3 array (and we have that 9 = 1 + 3 + 5), and adding to this a gnomon of seven dots yields a 4 × 4 array (yielding 16 = 1 + 3 + 5 + 7). This process shows that, in general, the nth square number equals the sum of the first n odd numbers. The pointer on a sundial is also called a gnomon. It, and the shadow it casts, together form the shape of an L. See also FIGURATE NUMBERS.
Gödel, Kurt (1906–1978) AustrianAmerican Logic Born on April 28, 1906, in Brünn, AustriaHungary (now Brno, Czech Republic), Kurt Gödel is today considered the most important logician of the 20th century. His famous pair of incompleteness theorems stunned the mathematical community, dashing the hopes of all those who had been fervently searching for
a set of fundamental axioms from which all mathematics could be logically deduced. Gödel exhibited a talent for academic work at an early age and had completed the equivalent of a university curriculum before leaving secondary school. In 1923 he entered the University of Vienna to pursue a degree in physics, but changed to mathematics in 1926 when he was introduced to the field of FORMAL LOGIC. It soon became clear to the faculty of the department that Gödel would make considerable contributions to the field. In the summer of 1929 Gödel completed a doctoral dissertation on the topic and was awarded a Ph.D. the following year. His thesis outlined the details of his first famous discovery, but the revolutionary impact of his work was not fully understood by the larger mathematical community until he published the result a year later. Gödel had proved that any mathematical system sufficiently sophisticated to incorporate the principles of arithmetic will always contain statements that can neither be proved nor disproved. He later also showed that no mathematical system could be proved to be consistent, that is, free from CONTRADICTION, by making use of just the axioms of the system. It had been a 200yearlong dream, since the time of GOTTFRIED WILHELM LEIBNIZ in fact, to place the whole of mathematics on a firm axiomatic base. Gödel had proved, in essence, that such a dream can never be realized. After joining the faculty of the University of Vienna for three years, Gödel accepted, in 1933, a position at the Institute for Advanced Study in Princeton, New Jersey. After suffering a nervous breakdown, he returned to Vienna after only a year, but later returned to the Institute in 1942 to escape wartorn Europe. Gödel became an American citizen that same year. Many awards were bestowed upon him throughout his life. In 1950 he was one of two recipients of the first Einstein Award, and he was awarded honorary doctorates from Yale University in New Haven, Connecticut, and Harvard University in Cambridge, Massachusetts, in the two years that followed. He was elected to the National Academy of Sciences in 1955. It is said that Gödel held fixed opinions about many matters of life, and felt himself to always be right— especially in the disciplines of mathematics and medicine. Gödel was of a nervous disposition, and after enduring severe bleeding from a duodenal ulcer, he decided to maintain an extremely strict diet of his own
Goldbach’s conjecture 229 devising. He developed a general distrust of doctors and, toward the end of his life, became convinced that he was being poisoned. Gödel eventually stopped eating altogether and died of starvation on January 14, 1978. Gödel’s work in mathematical logic strikes at the very core of the subject. His name and his results are familiar to any undergraduate student having taken a first course in formal logic. See also GÖDEL’S INCOMPLETENESS THEOREMS.
Gödel’s incompleteness theorems In 1900, DAVID posed 23 problems for future mathematicians. The second of his 23 problems challenged the mathematical community to find a logical base for the discipline of mathematics, that is, to develop a formal system of symbols, with welldefined meaning and welldefined rules of manipulation, on which all of mathematics could be based. Mathematicians, such as BERTRAND ARTHUR WILLIAM RUSSEL and his colleague ALFRED NORTH WHITEHEAD, took on the challenge and made some progress in this direction. But in 1931 KURT GÖDEL stunned the mathematical community by proving once and for all that such an aim could never be achieved. Gödel’s first incompleteness theorem states that any mathematical system sufficiently sophisticated to incorporate the principles of arithmetic will necessarily contain statements (theorems) that can neither be proved nor disproved. Informally, this means that it will never be possible to program a computer to answer all mathematical questions. His second incompleteness theorem (which is actually a consequence of the first) asserts that no system of mathematical logic incorporating the principles of arithmetic will ever be capable of establishing its own consistency. That is, the proof that a system of mathematics is free from CONTRADICTION will require axioms and principles not contained in the system. Gödel proved his theorems by assigning numbers to the symbols used in a mathematical system (commas, digits, left and right parentheses, etc.), and then devising an ingenious method of assigning numbers to all mathematical statements (combinations of symbols) in the system. It was then possible to show that the mathematical system contains a statement of the form: “This sentence cannot be proved.” Of course, it is impossible to prove or disprove such a statement. HILBERT
Gödel’s work dashed centuries of hope of finding a small set of basic axioms on which to base all of mathematics. See also CONSISTENT; FORMAL LOGIC; TRUTH TABLE.
Goldbach, Christian (1690–1764) Prussian Number theory Born on March 18, 1690 in Königsberg, Prussia (now Kaliningrad, Russia), Christian Goldbach is best remembered in mathematics for the famous unsolved conjecture that bears his name. Goldbach studied in St. Petersburg and became a professor of mathematics and historian at the university in 1725. He traveled extensively throughout Europe and established personal contact with a number of prominent mathematicians of the time. Although Goldbach studied infinite sums, the theory of equations, and the theory of curves, his most important work was in the field of NUMBER THEORY, much of which was conducted through correspondence with the Swiss mathematician LEONHARD EULER (1707–83). One particular letter later garnered international attention. In it, Goldbach mentioned that every even number seems to be the sum of two PRIME numbers. For instance, he noted that 4 equals 2 + 2, 6 equals 3 + 3, 10 equals 3 + 7, and 1,000 equals 113 + 887. Unable to find an instance where this was not the case, Goldbach asked Euler whether or not this was indeed a property of all even numbers. Euler attempted to resolve the issue but found that he was unable to establish the claim, nor find a COUNTEREXAMPLE to it. Later, English mathematician Edward Waring published the problem in his popular 1770 text Meditationes algebraicae (Meditations on algebra), properly attributing the problem to Goldbach. The challenge garnered considerable notoriety. Today called GOLDBACH’S CONJECTURE, the problem remains one of the most famous unsolved challenges in mathematics. Goldbach died in Moscow, Russia, on November 20, 1764. Apart from the conjecture that bears his name, Goldbach is also remembered for his correspondence with PIERRE DE FERMAT (1601–1655) and for being one of the few mathematicians at the time to understand the implications of Fermat’s approach to number theory.
Goldbach’s conjecture In his 1742 letter to LEONHARD
EULER, Prussian mathematician CHRISTIAN
230 golden ratio GOLDBACH (1690–1764) conjectured that every even number greater than 2 is the sum of two PRIME numbers. For example, 4 = 2 + 2, 6 = 3 + 3,…, 20 = 3 + 17,…, 50 = 13 + 37,…, 100 = 3 + 97, and so forth. No one to this day has been able to establish whether this claim holds true for each and every even number. The problem has since become known as Goldbach’s conjecture, and it remains one of the most famous unsolved mathematical problems of today. Between March 20, 2000, and March 20, 2002, a $1million prize was offered to anyone who could solve Goldbach’s conjecture. The prize went unclaimed. Goldbach’s conjecture is equivalent to the challenge of proving that every integer greater than 5 is the sum of three primes. (If the number is odd, subtract 3 and write the resultant number as a sum of two primes. If it is even, subtract 2.) The source of the difficulty with this problem is that primes are defined in terms of multiplication, while the problem involves addition. It is often very difficult to establish connections between these two separate operations on the integers. In 1931, Russian mathematician L. Schnirelmann (1905–38) proved that every positive integer is the sum of at most 300,000 primes. Although this result seems ludicrous in comparison to the original problem, it is a significant first step to solving the conjecture: it shows at least that it is possible to put a bound on the number of primes representing the integers. In 1937 another Russian mathematician, Ivan Vinogradoff (1891–1983), proved that every sufficiently large number can be written as a sum of four primes. What constitutes “sufficiently large” is not known (Vinogradoff’s work only shows that there cannot be infinitely many integers that require more than four prime summands)—but at least a small bound has been placed on “most” integers. In 1973 JingRun Chen proved that every sufficiently large even number is the sum of a prime and a number that is either prime or has two prime factors. Work on solving the Goldbach conjecture continues. As of the year 2003, it has been confirmed by computer that the conjecture holds true for all numbers up to 6 × 1016. See also BRUTE FORCE.
golden ratio (divine proportion, extreme and mean ratio, golden mean, golden section) A line segment
connecting two points A and B is said to be divided by a third point P in the golden ratio if the ratio of the whole length AB to the length AP is the same as the RATIO of the length AP to the length PB: AB AP –– = –– AP PB The value of this ratio is denoted ϕ (the Greek letter phi) and can be computed as follows: Set the length PB to be one unit. Then, since AP –– = ϕ, the length AP is ϕ units, and the length PB of the entire segment AB is ϕ + 1. We have: ϕ+1 ϕ –––– = –– ϕ 1 yielding the QUADRATIC equation ϕ2 = ϕ + 1. This has two solutions. Selecting the solution that is larger than 1 yields: – 1 + √5 ϕ = –––– = 1.618033988… 2 (The second solution to this quadratic equation is – 1 + √5 –––– = 1 – ϕ = –0.618033988….) 2 The golden ratio was studied and made famous by the Pythagoreans, the followers of the Greek mathematician PYTHAGORAS (ca. 500 B.C.E.). They discovered it in their study of the PENTAGRAM, the figure that appears when one draws in the diagonals of a regular pentagon. The sides of a pentagram divide each other in the golden ratio. The following method shows how to construct a point P that divides a given line segment AB into the golden ratio: Given a line segment AB, draw a perpendicular line through A. Find the MIDPOINT M of AB and locate on the perpendicular line a point C whose distance from A is the same as the length AM. (Use a circle with center A and radius AM as an aid.) Now draw a circle with center C and radius BC to locate a second point D on the perpendicular line such that length BD equals length BC. Let P be the point on the line segment AB such that length AP equals length AD. (Use a circle with center A
golden rectangle 231 and radius AD to find this point.) Then point P divides the segment AB in the golden ratio. (To see why this works, suppose that the line segment AB is of unit length, and the length of AP is x. Then triangle ABC is a right triangle with hypotenuse 1 –– + x. PYTHAGORAS’S THEOREM now shows that x = 2 AB 1 ϕ – 1 = –– , giving –– = ϕ.) ϕ AP Noting that the construction of midpoints, circles, and perpendicular lines can all be accomplished with a straight edge (that is, a ruler with no markings) and a compass alone, the above procedure shows that the golden ratio is a CONSTRUCTIBLE number. Dividing the relation ϕ2 = ϕ + 1 through by ϕ yields 1 ϕ = 1 + –– . Substituting this formula into itself multiple ϕ times establishes: ϕ = 1+
1 = 1+ ϕ
1 1 1+ ϕ
1
= 1+
= ...
1
1+
1+
1 ϕ
\
Repeating this process indefinitely shows that the golden ratio has the following simple CONTINUED FRACTION expansion: 1
ϕ = 1+
1
1+
1
1+ 1+
1 1 + ...
If one terminates this expansion after a finite number of steps, then ratios of the FIBONACCI NUMBERS appear: 1=
3 F 1 1 F2 1 2 F 1 = ,1 + = = 3 ,1 + = = 4 ,1 + 1 1 F1 1 1 F2 1 2 F2 1+ 1+ 1+1 1
5 F5 = , and so forth. (That this pattern persists can 3 F4 be proved by INDUCTION.) We have: =
lim n→∞
Fn+1 =ϕ Fn
An induction argument also proves that:
ϕn = Fnϕ + Fn–1 – Substituting the formula ϕ = √ 1+ϕ into itself multiple times gives an expression for ϕ as a sequence of nested radicals akin to VIÈTE’S FORMULA for π. We have: ϕ = 1 + 1 + 1 + 1 + ... The golden ratio also occurs in
TRIGONOMETRY
⎛π⎞ ϕ ⎛ 2π ⎞ ⎛π⎞ 1 1 cos⎜ ⎟ = , cos⎜ ⎟ = , and sin⎜ ⎟ = , ⎝ 10 ⎠ 2ϕ ⎝ 5⎠ 2 ⎝ 5 ⎠ 2ϕ
as for
instance. The number ϕ also appears in a number of unexpected places in nature and throughout mankind’s artistic pursuits. Since the golden ratio is well approximated by the fraction 16/10, some scholars suggest that the ancient Egyptians of 3000 B.C.E. used the golden ratio repeatedly in the construction of their tombs. The “golden chamber” of the tomb of Rameses IV measures 16 ells by 16 ells by 10 ells, that is, approximately the ratio ϕ : ϕ : 1; other tombs are found in the approximate ratio ϕ2 : ϕ : 1; and Egyptian furniture found in those tombs often had overall shape based on the ratio ϕ : 1 : 1. German artist ALBRECHT DÜRER (1471–1528) wrote a fourvolume text, Treatise on Human Proportions, detailing occurrences of the ratio ϕ in the human body. (He claimed, for instance, that ratio of the length of the human face to its width is approximately ϕ, and also that the elbow divides the human arm, shoulder to fingertip, in the golden ratio.) Artists of that time came to view the golden ratio as a “divine PROPORTION” and used it in all forms of artistic work. The GOLDEN RECTANGLE was deemed the rectangular shape most pleasing to the eye.
golden rectangle Any rectangle whose sides are in –
1 + √5 the ratio 1 to ϕ, where ϕ = –––– is the GOLDEN 2 RATIO, is called a golden rectangle. Such a rectangle has the property that excising the largest square possible from one end of the figure leaves another rectangle in the same proportion. (The remaining rectangle has proportions ϕ – 1 to 1. Since the golden ratio satisfies the ϕ–1 1 –––– equation ϕ2 = ϕ + 1, we have: –– ϕ = 1 .) By this method, new golden rectangles can be constructed from
232
googol/googolplex
googol/googolplex In 1938 Milton Sirotta, nineyearold nephew of American mathematician Edward Kasner, coined the term googol for the number 1 followed by 100 zeros (10100). At the same time he coined the term googolplex for the number 1 followed by a 100 googol zeros (10googol = 10(10 )). Physicists believe that there are only about 1080 particles in the entire universe, considerably less than a googol.
Gosset, William Sealy (1876–1937) British StatisThe golden rectangle
a given golden rectangle ad infinitum. Drawing arcs between the nonadjacent corners of the squares produces a spiral with the remarkable property that it cuts any radius from the center of the spiral at the same ANGLE (approximately 73°). Throughout history the golden rectangle has been considered to have a particularly pleasing shape. The pillars of the Greek Parthenon of the 5th century B.C.E. are spaced to produce golden rectangles in its facade (although no one can be certain that the designers intended this to be the case). Many European scholars of the 1500s based much of their work on the golden rectangle, and German engraver ALBRECHT DÜRER (1471–1528) used the golden ratio in his analysis of PROPORTIONS found in the human body. Many rectangles used today in magazine advertisements, for instance, are surprisingly close to golden rectangles. The golden rectangle also appears in the fifth PLATONIC SOLID, the icosahedron. If all the edges of the figure are assumed to be one unit long, then the distance between any given edge and the (parallel) edge opposite to it on the other side of the figure is ϕ. Thus any two opposite edges of an icosahedron form two sides of a golden rectangle. A “golden triangle” is an isosceles triangle whose sides to its base are in the ratio ϕ to 1. Such a triangle has internal angles 72°, 36°, and 72°. If one of the base angles is bisected, then the figure is divided into two smaller triangles, one of which is a new golden triangle. See also SPIRAL OF ARCHIMEDES.
tics Born on June 13, 1876, in Canterbury, England, William Gosset is remembered for his important work in statistics, most notably his invention of the ttest, details of which were published in 1908 under the title STUDENT’S TTEST. Gosset studied chemistry and mathematics at New College, Oxford, before accepting a position as a chemist at the Guinness brewery in Dublin in 1899. His interests in statistics were motivated by the practical problems of measuring and maintaining production quality. Rather than test each and every sample of product from the brewery, Gosset looked for mathematical techniques that would allow him to deduce reliable information based on just a small number of samples. His famous ttest is the result of his efforts. In an effort to protect trade secrets, the brewery forbade its employees to publish, but Gosset printed his method nonetheless under the pseudonym of “Student.” Gosset’s work was deemed of great value to the brewery. In 1922 he was given an assistant in statistics, and he was allowed to build up a small department of statistics within the company. The Student ttest is today considered a fundamental technique in the repertoire of tools used by industry and science when dealing with concerns of quality control and general statistical inference. In 1935 Gosset left Ireland to manage a new Guinness brewery in London. He died two years later on October 16, 1937. His statistical analysis of variance had a profound effect on the practices of 20thcentury industry. See also HISTORY OF PROBABILITY AND STATISTICS (essay).
grad See GRADIENT.
graph 233
gradient The
of a line is sometimes called its gradient or grade. For example, the gradient of the line connecting the two points A = (a1,a2) and B = (b1,b2), denoted mAB, is given by: SLOPE
mAB =
"rise" b2 − a2 = "run" b1 − a1
Three points A, B, and C in the plane are COLLINEAR if the gradients mAB and mAC are equal. In the setting of multivariable calculus—the study of CALCULUS applied to FUNCTIONS of more than one variable—the gradient of a function f(x,y,z) of three VARIABLES, also called the “grad of f,” is defined to be the VECTOR of its PARTIAL DERIVATIVEs. It is denoted grad(f) or f and is given by: ∇f =
∂x ∂y ∂z
Such a quantity proves to be useful in computing the of a function. In physics, f is also used to describe the spatial variation in the magnitude of a force, such as a gravitational force or a magnetic force. The study of directional derivatives shows that the quantity f calculated at a point represents the direction from that point in which the rate of change of the force f is a maximum. See also DIV.
DIRECTIONAL DERIVATIVE
graph (network) Any diagram of points and line segments connecting pairs of points is called a graph. The points are usually called vertices or nodes, and the line segments are called edges. More than one edge is allowed to connect the same pair of vertices to yield a set of multiple edges. One can also permit an edge connecting a vertex to itself via a loop. Edges can intersect, but the places where they cross are not considered vertices. For example, the following picture is a graph. It has just six vertices and comes in two disconnected pieces. A graph that comes in just one piece is called connected. This means that it is always possible to travel from any one vertex to any other by traversing a sequence of edges. Graphs can be used to codify information. For example, a graph might represent the network of possi
ble flight routes between cities, the flow of information between departments in a large organization, or even the set of acquaintances among people attending a party. (Each vertex represents a person in the room, and an edge is drawn between two vertices if the corresponding two people know each other.) The general study of graphs can translate into interesting facts about travel possibilities, streamlining data flow, acquaintanceships, and the like. The degree (or valence) of a vertex is the number of edges that meet at that vertex. Loops are counted twice. A graph is called complete if every vertex is connected to each and every other vertex by a single edge. For example, the complete graph on four vertices looks like a square with the two diagonals drawn in. Each vertex has degree three. A graph is called planar if it can be drawn on a plane without two edges crossing. The complete graph on four vertices is planar if one draws one of the diagonals “outside the square.” The THREEUTILITIES PROBLEM is an example of a graph that is not planar. A cycle in a graph is a sequence of edges that starts and ends at the same vertex and does not travel over the same edge twice. Finding an “Euler circuit,” i.e., is, a cycle that traverses each and every edge in a graph precisely once, is an old problem. (See GRAPH THEORY.) A connected graph containing no loops or cycles is called a tree. These graphs look like a series of forking branches, and hence the name tree. Any tree diagram, such as a PROBABILITY, tree is an example of a graph that is a tree. It follows from EULER’S THEOREM that any connected graph with n vertices and n–1 edges must be a tree. See also CRITICAL PATH; EULERIAN PATH/CIRCUIT; HAMILTONIAN PATH/CIRCUIT; TOURNAMENT.
A typical graph
234 graphical solution
graphical solution Any solution to a pair of simultaneous equations found by plotting the graph of each equation and visually inspecting the location of a point of intersection is called a graphical solution. For example, in plotting the graphs of y = 2x + 3 and y = 9 – 4x, one sees that the lines intersect at the point (1,5) and so x = 1, y = 5 is a solution to the pair of equations. A graphical approach is often the only feasible method of solving a complicated pair of equations. Graphing calculators are equipped with a “trace” button that allows one to quickly find the coordinates of points of intersection up to some degree of accuracy. A single equation f(x) = 0 can be solved graphically by locating the xintercepts of the function. For example, to find the square root of 2, one can plot a graph of the function f(x) = x2 –2 and attempt to identify the location of the positive xintercept. This graphical method makes it clear that a pair of SIMULTANEOUS LINEAR EQUATIONS: ax + by = p cx + dy = q has either one solution (the lines intersect at a unique point), no solutions (the lines are parallel), or infinitely many solutions (the lines coincide). dy A DIFFERENTIAL EQUATION of the form –– dx = f(x,y) can be solved graphically. We seek a curve y = g(x) in dy the plane whose slope –– dx at any point (x, y) is given by f(x,y). Thus if, for a large selection of points (x,y) across the plane, we draw short line segments of slope f(x,y), the shapes of curves following these slopes may be visually apparent. If we are also told the value of the function y = g(x) at one particular point (x, y), then following the slope of the line segments from that point onward describes a particular solution to the equation. See also BISECTION METHOD; NEWTON’S METHOD.
graph of a function A drawing or a visual representation that shows the relationship between two or more variables is called a graph. It is usual to draw the graph of a FUNCTION of a single variable y = f(x) on a CARTESIAN COORDINATE system with an xaxis and a yaxis at right angles. The graph of the function is then the set of all points (x,y) that satisfy the equation y = f(x) drawn as a curve in the plane. For example, the set of all
points (x, y) that satisfy y = x2 forms a PARABOLA, while the graph of the function y = 3x is a straight line through the origin with slope 3. A general approach to graphing a function is to make a table of (x, y) pairs that satisfy the equation under consideration and then to locate these points on a coordinate system. If sufficiently many points are drawn, then a smooth curve connecting the dots is likely to be a good representation of the function. Graphing calculators employ this technique when displaying the graph of a function. As one develops familiarity with basic equations, graphing simple formulae becomes a matter of routine. For example, one can establish that a LINEAR EQUATION of the form y = mx + b yields a straightline graph of slope m crossing the yaxis at position b, and that an equation of the form x2 + y2 = r2 represents a CIRCLE of radius r. To determine the graphs of morecomplicated functions, scholars—for many decades—could only resort to the tedious task of making tables and plotting individual points until a general picture emerged. The advent of CALCULUS, however, at the turn of the 18th century brought with it the power to quickly identify and examine the basic shape and structure of complicated graphs. By examining the first DERIVATIVE of a function, for instance, one can determine where the graph increases and decreases, as well as the location of any local maxima and minima. The second derivative provides information about the shape of the curve— whether it is concave up or concave down. This information, together with knowledge of the x and yintercepts of the curve and any ASYMPTOTEs it might possess, is enough to draw a reasonably accurate picture of the graph without having to plot individual points. Application of these newly discovered techniques from calculus was literally an eyeopening experience for scholars of the time. French mathematician NICOLE ORESME (ca. 1323–82) was the first to draw the graph of a function and to find an interesting interpretation for the area of the region under it. French lawyer and amateur mathematician PIERRE DE FERMAT (1601–65) developed the idea further, defining a general procedure for associating curves to formulae. Given a relationship between two variables A and B, say, Fermat drew a horizontal reference line for the independent variable A and imagined a second line sliding along this reference line at a fixed angle whose length B varied according to the
graph theory 235 relationship defined. He did not require that the line representing B be at right angles to the reference line. French mathematician and philosopher RENÉ DESCARTES (1596–1650) independently developed the same technique, but he had the further insight to introduce algebraic symbolism to describe relationships via formulae and to label graphs appropriately. Descartes too, however, envisioned functions as sliding lengths along a single reference line. It was not until decades later that mathematicians began drawing an explicit second coordinate axis, the yaxis, at a fixed 90° angle to the horizontal reference xaxis. Quantities other than functions can also be graphed. For example, an inequality of the form y < x + 3 defines the HALFPLANE below the line y = x + 3. The graph of a function of two variables z = f(x, y) is a surface sitting in threedimensional space. For example, the graph of the function z = x2 + y2 is a PARABOLOID. A logarithmic graph is one in which both axes of the coordinate system are marked in LOGARITHMIC SCALE. Such plots are particularly useful for examining equations of the form y = axn. A polarcoordinate graph is a plot of an equation of the form r = f(θ) given in POLAR COORDINATES. Polarcoordinate graph paper assists in plotting such graphs. Graphs of other types of numerical relationships are studied in descriptive statistics. See also CONCAVE UP/CONCAVE DOWN; COORDINATES; GRAPHICAL SOLUTION; HISTORY OF FUNCTIONS (essay); INCREASING/DECREASING; MAXIMUM/MINIMUM; STATISTICS: DESCRIPTIVE.
graph theory The mathematical study of graphs is called graph theory. The field was founded in 1736 by the Swiss mathematician LEONHARD EULER (1707–83) with his solution to the famous seven bridges of Königsberg problem: The old German city of Königsberg (now the Russian city of Kaliningrad) was built on the two banks of the river Pregel and on two islands in the river. The different parts of the city were connected by seven bridges. A debate ensued among the residents as to whether it was possible to walk a complete tour of the city crossing each and every bridge precisely once. (It wasn’t deemed necessary to return to one’s starting location.) Most people felt that
The seven bridges of Königsberg problem
this was impossible, but could this be proved to be the case? Euler’s insight into solving the problem came by reducing the city plan to a GRAPH, that is, a diagram of points (vertices) and edges, with each point representing a land mass, and each edge a bridge. Any stroll through the city thus corresponds to a journey along the edges of the graph. The Königsberg bridge problem is therefore equivalent to asking: Is it possible to draw the above graph without lifting pencil from page and without tracing over the same edge twice? Euler was able to solve the more general problem of classifying all those graphs that can be so traced. He observed that, in tracing such a path, any edge drawn entering a vertex must be matched by an edge exiting that vertex. Thus all the edges meeting at each vertex are matched in pairs unless one starts or ends a journey at a particular vertex, in which case one edge remains unmatched. Thus a graph that can be so traced must contain just two vertices of odd degree— these will be the start and end of the journey; or no vertices of odd degree—the path starts and ends at the same location. Any graph containing more than two vertices of odd degree cannot be drawn without lifting pencil from page. In particular, the seven bridges of Königsberg problem cannot be solved (all four vertices have odd degree). Any path that traces through a graph following each edge precisely once is known today as an Eulerian path. If the path starts and ends at the same vertex it is called an Eulerian circuit. Euler went further to show that all connected graphs possessing precisely two vertices of odd degree do indeed have Eulerian paths, and that graphs with all vertices of even degree do possess
236 greatest common divisor Eulerian circuits. If a graph represents transportation routes in a city or across a nation, for example, then seeking Eulerian circuits for the graph can have important practical use. In 1857 WILLIAM ROWAN HAMILTON explored the issue of finding paths through graphs that do not necessarily trace each edge, but instead visit each vertex in the graph once, and only once. Today such a path is called a Hamiltonian path, or a Hamiltonian circuit if it is a path that returns to the starting vertex. Although mathematicians have developed some theorems that give conditions under which Hamiltonian paths will exist, no one to this day knows a simple algorithm that enables us to find them. Each graph must still be examined individually, and finding a Hamiltonian path—if one exists—is usually a matter of inspired guessing. Simple counting can lead to important results in graph theory. For example, summing the degrees of all the vertices in a graph counts the total numbers of ends of edges. As every edge has two ends, we have: In any graph, the sum of the degrees of all the vertices equals twice the number of edges in the graph. In particular, the sum of all the degrees of vertices must be an even number. Consequently, there cannot be an odd number of odd numbers in this sum. In any graph, the number of vertices of odd degree is even. This result has an amusing interpretation. Thinking of edges in a graph as handshakes between people, we have established the socalled HANDSHAKE LEMMA: At any instant, the number of people on this planet, living or deceased, who have participated in an odd number of handshakes is even. A POLYHEDRON can be thought of as a graph drawn on a SPHERE—each corner of the polyhedron is a vertex, and each edge of the figure is an edge of the graph. If a polyhedron consists of only t triangular faces, then 3t counts the number of edges of the figure twice (each triangle has three edges, and each edge borders two triangles). Thus 3t = 2e, where e is the number of edges. This shows that t must be an even number. We have:
It is impossible to cover a sphere with an odd number of triangles. Euler also went on to show that for any graph drawn on a plane or a sphere (with no edges of the graph intersecting): v–e+r=2 where v is the number of vertices the graph possesses, e its number of edges, and r the number of regions defined by the graph—including the large “outer” region if the graph is drawn on a plane. This is called EULER’S FORMULA. If the graph is drawn instead on a TORUS, this formula is modified to read: v – e + r = 0. Graph theory is a discipline under intense continued study. Its many applications vary from the purely theoretical to the very concrete and practical. Routing problems, informationflow problems, and electronic circuit design, for example, can all be effectively analyzed and refined through the study of this field. See also EULERIAN PATH/CIRCUIT; HAMILTONIAN PATH/CIRCUIT; THREEUTILITIES PROBLEM; TOPOLOGY; TOURNAMENT; TRAVELINGSALESMAN PROBLEM.
greatest common divisor (greatest common factor, highest common factor) The largest FACTOR common to a given set of integers is called the greatest common divisor of those integers. For example, the numbers 72, 120, and 180 have factors 1, 2, 3, 4, 6, and 12 in common, with 12 being the greatest common divisor. We write gcd(72, 120, 180) = 12. One can find the greatest common divisor of two or more integers simply by listing the factors of each integer and identifying the largest factor they have in common. Alternatively, if the prime decompositions of the integers are known, then their greatest common divisor can be determined as the product of the primes they have in common. For example, noting that: 72 = 2 × 2 × 2 × 3× 3 120 = 2 × 2 × 2 × 3× 5 180 = 2 × 2 × 3 × 3× 5 we see that these three numbers share, as prime factors, two 2s and one 3. Their greatest common divisor is thus gcd (72,120,180) = 2 × 2 × 3 = 12. (This second
Greek mathematics approach shows that any common factor of a given set of integers is a divisor of the greatest common factor.) The EUCLIDEAN ALGORITHM can be used to find the greatest common divisor of two integers if either of these methods is infeasible. Repeated use of the Euclidean algorithm will find the greatest common divisor of more than two integers. The Euclidean algorithm also shows that it is always possible to write the greatest common divisor of two integers a and b as a linear combination of a and b, that is, it is always possible to find integers x and y so that: gcd(a,b) = ax + by Similarly, the greatest common factor of any finite set of integers a1,a2,…,an can be expressed as a linear combination of the form gcd(a1,a2,…,an) = a1x1 + a2x2+ …+anxn. In our example: 12 = gcd(72, 120, 180) = 72 × (1) + 120 × (1) + 180 × (–1) See also
FUNDAMENTAL THEOREM OF ARITHMETIC;
JUGFILLING PROBLEM; RELATIVELY PRIME.
Greek alphabet To honor the mathematical scholars of Greek antiquity, mathematicians today often use letters of the Greek alphabet to represent variables and symbols in equations. Typically, lowercase letters are used to represent variables (such as angles, COMPLEX NUMBERS, and quantities studied in STATISTICS), and uppercase letters are used for standard arithmetical and statistical operations. The uses can vary from author to author, however. The following table lists the letters of the Greek alphabet along with the common uses of some characters. See also GREEK MATHEMATICS.
Greek mathematics The ancient Greeks of ca. 600 B.C.E.
to ca. 480 C.E. set the current standards of logical rigor in mathematics. Although many ancient cultures practiced and developed mathematics, it was the Greeks who developed the explicit art of “proof” and explored the power of pure deductive reasoning to its fullest.
Upper Lower Case Case Name
Pronunciation
237
Use
A
α
alpha
ALfuh
α: often denotes an angle β: often denotes an angle γ: often denotes an angle Γ: the GAMMA
B
β
beta
BAYtuh
Γ
γ
gamma GAMuh
∆
δ
delta
DELtuh
Ε
ε
epsilon
EPsilon
Ζ
ζ
zeta
ZAYtuh
Η Θ
η θ
eta theta
AYtuh THAYtuh
Ι Κ Λ
ι κ λ
iota eyeOHtuh kappa KAPuh lambda LAMduh
Μ
µ
mu
MYOO
Ν
ν
nu
NYOO
Ξ Ο Π
ξ ο π
Ρ
ρ
Σ
σ
xi kuhSEYE omicron OMeeKRON pi PIE π: the ratio of a circle to its diameter Π: (infinite) product rho ROH ρ: radius of a sphere (SPHERICAL COORDINATES) sigma SIGma σ: STANDARD
FUNCTION
δ: a small quantity (EPSILONDELTA DEFINITION) ∆: denotes change ε: a small quantity (EPSILONDELTA DEFINITION) ζ: the ZETA FUNCTION
θ: often denotes an angle (POLAR COORDINATES) ι: a small quantity λ: wavelength (in physics) µ: denotes "microns”; Möbius function ν: frequency (physics)
DEVIATION
Σ: SUMMATION
Τ Υ Φ
τ υ ϕ
tau upsilon phi
TAU OOPsiLON FEE
Χ Ψ
χ ψ
chi psi
KEYE SIGH
Ω
ω
omega
ohMAYguh
ϕ: often denotes an angle (SPHERICAL COORDINATES) χ: CHISQUARED TEST ψ: wave function (physics) ω: a complex number; the first transfinite ORDINAL NUMBER.
238 Greek mathematics We should mention that when speaking of “Greek mathematics,” historians include any mathematician who wrote in the Greek language and followed the Greek tradition of mathematical thought. Greek was the common language of the Mediterranean world during ancient times, and many intellectuals from different parts of that region are today considered Greek scholars. For instance, the great Archimedes was from Syracuse, now a part of Italy, and EUCLID (ca. 300–260 B.C.E.) is believed to have lived in Alexandria, Egypt. There are very few original records of Greek work. Initially, knowledge was transmitted only orally from teacher to student. Around 450 B.C.E. the Greeks adopted the ancient Egyptian practice of writing on papyrus scrolls. Unfortunately, papyrus—a grasslike plant grown in the Nile Delta region—decays rapidly away from the exceptionally dry climate of Egypt. The Greeks combated this problem by repeatedly making copies of their works but, because of the effort involved, copied only those pieces they deemed of utmost importance. The first mathematical work preserved and honored this way was Euclid’s masterpiece THE ELEMENTS of ca. 300 B.C.E. Historians have had to rely on commentary made by later scholars to deduce what was accomplished mathematically before the time of Euclid. Greek scholars approached all of mathematics through the study of GEOMETRY. Even their work on the properties of whole numbers, ratios, and proportions, as well as mechanics and astronomy was done in a geometric style. A “number,” for instance, was literally a line segment, and a “ratio” was understood in terms of COMMENSURABLE segments. It is interesting to note that Greek scholars took careful steps to avoid speaking directly of the infinite. (The fifthcentury B.C.E. paradoxes on the nature of motion and the infinitely small developed by ZENO OF ELEA deeply affected Greek thinking.) For instance, Euclid stated that any line segment could be extended to any arbitrary length, but never spoke of lines that were infinitely long. In EUCLID’S PROOF OF THE INFINITUDE OF PRIMES, Euclid stated that from any finite list of PRIME numbers one can always construct one more, but never spoke of the set of primes as infinite. Many historians regard THALES OF MILETUS (ca. 625–547 B.C.E.) as the first Greek mathematician of note. Commentaries suggest that Thales identified, and proved, seven key geometric propositions, including
that the base angles of an ISOSCELES TRIANGLE are always equal and that the inscribed angle from the diameter of a CIRCLE is always a right angle, for instance. The great scholar and mystic PYTHAGORAS lived a century later, and he and his followers are credited with the discovery of the famous result about right triangles (today called PYTHAGORAS’S THEOREM) and the discovery of IRRATIONAL NUMBERS. A great deal of mystery surrounds the life and legend of Pythagoras. He founded a semireligious sect called the Pythagorean Brotherhood (women were equal members) based on certain mystic significances ascribed to whole numbers and their ratios. The great philosopher PLATO (428–348 B.C.E.) wrote a great deal about mathematics in his famous dialogues, demonstrating a deep personal respect for the subject. The five regular polyhedra—the PLATONIC SOLIDs—are named in his honor. In his philosophical treatises, Plato used the example of mathematics as something that cannot be discovered by the senses, but can nonetheless be discovered by the power of logical reasoning. He also believed mathematics to be an essential part of a cultured person’s education. Philosopher ARISTOTLE (384–322 B.C.E.) adopted the same view and used mathematics as examples in his development of FORMAL LOGIC and his analysis of ARGUMENTs. Today, the Greek scholar Euclid is considered to be the most influential mathematics scholar of all time. In his famous work The Elements, Euclid collated all mathematical knowledge known at his time into a single tome. Although an impressive feat, it was the organization of the text that had the greatest impact. Beginning with a small collection of “selfevident truths,” Euclid showed that all mathematical knowledge of his time could be deduced by pure logical reasoning alone. This work demonstrated the power of the mind and set the model for all mathematical research in the future. Mathematicians today still work to the standards of rigor as set by Euclid. Next to the Bible, Euclid’s The Elements is the most widely published book of all time. After producing The Elements, Euclid continued work on the CONIC SECTIONS, on optics, and on general problems in geometry. He continued interest in CONSTRUCTIBLE numbers and no doubt contemplated the classic Greek problem of SQUARING THE CIRCLE. (In The Elements Euclid had demonstrated general procedures for squaring arbitrary polygonal figures.) This
Green, George 239 challenge, as well as the problems of TRISECTING AN ANGLE and DUPLICATING THE CUBE, spurred a great deal of significant further research in mathematics for centuries to come. ARCHIMEDES OF SYRACUSE (ca. 287–212 B.C.E.) solved the problem of squaring the parabola, as well as made significant advances in computing the areas and volumes of curved figures and solids. (He also “solved” the problem of squaring the circle by making use of his ARCHIMEDEAN SPIRAL. Unfortunately, his method went beyond the use of a straightedge and compass alone, and so is not a permissible solution to the original problem.) APOLLONIUS OF PERGA (ca. 262–190 B.C.E.) continued work on conic sections and is credited for properly defining an ELLIPSE, a HYPERBOLA, and a PARABOLA. Around the same time, Greek astronomer Hipparchus wrote a table of “chord values” (the equivalent to a modern table of sine values), which he used to solve astronomical problems. This represented the beginning development of TRIGONOMETRY in Greek mathematics, but also marked an end of fervent mathematical development in the Greek tradition. For the five centuries that followed, new developments were limited to straightforward advances in astronomy, trigonometry, and algebra, with just a few notable exceptions. In the second century C.E., Greek astronomer CLAUDIUS PTOLEMY corrected and extended Hipparchus’s table and clarified the mathematics that is used to produce such a table. He is also known as one of the first scholars to make a serious attempt at proving Euclid’s PARALLEL POSTULATE. In the third century, DIOPHANTUS OF ALEXANDRIA produced his famous text Arithmetica (Arithmetic), from which the study of DIOPHANTINE EQUATIONs was born. In the midfourth century, the enthusiastic PAPPUS OF ALEXANDRIA attempted to revive interest in ardent mathematical research of the Greek style. He produced his treatise Synagoge (Collections) to act as a commentary and guide to all the geometric works of his time and included in it a significant number of original results, extensions of ideas, and innovative shifts of perspective. Unfortunately, he did not succeed in his general goal. After Pappus, of note is HYPATIA OF ALEXANDRIA (370–415), the first woman to be named in the history of mathematics, credited for writing insightful commentaries on the works of
Apollonius and Diophantus, and PROCLUS (ca. 410–485), who is noted for his detailed commentary on the work of Euclid and his own attempt to prove the parallel postulate. The beginning of the fifth century marks a clear end to the tradition of Greek mathematics. See also ARCHYTAS OF TARENTUM; DEDUCTIVE/ INDUCTIVE REASONING; ERATOSTHENES OF CYRENE; EUCLID’S POSTULATES; EUCLIDEAN ALGORITHM EUCLIDEAN GEOMETRY; EUDOXUS OF CNIDUS; HERON OF ALEXANDRIA; HIPPASUS OF METAPONTUM; HIPPOCRATES OF CHIOS; HISTORY OF EQUATONS AND ALGEBRA; MENELAUS OF ALEXANDRIA; PAPPUS’S THEOREMS; PTOLEMY’S THEOREM; PYTHAGOREAN TRIPLES; THEODORUS OF CYRENE; ZENO’S PARADOXES.
Green, George (1793–1841) British Calculus Born in July 1793 (his exact birth date is not known) in Nottingham, England, George Green is remembered today for his influential 1828 paper “Essay on the Application of Mathematical Analysis to the Theory of Electricity and Magnetism,” in which he developed the notion of “potential” and proved a fundamental mathematical result today known as Green’s theorem. Green left school at the age of nine and worked in his father’s bakery for the next 30 years. Historians do not know how Green developed an understanding of mathematics, nor how he had access to current work in the field, but in 1828 he published one of the most important scientific papers of the time. Apart from advancing the mathematical understanding of electromagnetism, important for physicists, Green also established a fundamental mathematical technique for computing CONTOUR INTEGRALs and DOUBLE INTEGRALs. The famous theorem that bears his name states that if a region R in the xyplane is bounded by a curve C, and if functions P(x,y) and Q(x,y) have continuous PARTIAL DERIVATIVEs, then:
∫ P(x, y) dx + Q(x, y) dy = ∫∫
C
R
∂Q ∂P − dA ∂x ∂y
This result appears in every multivariable calculus textbook of today. After reading the famous 1828 piece, mathematician Sir Edward Bromhead invited Green to continue
240 Gregory, David work in the field of mathematical physics, and over the next six years they together produced three papers on the topics of electricity and hydrodynamics. At the advice of Bromhead, Green entered Cambridge University in 1833, at age 40, to start an undergraduate degree in mathematics. Green published six more papers after completing the program. He died on May 31, 1841, in Nottingham.
Gregory, David (1659–1708) British Calculus, Geometry Born on June 3, 1659, in Aberdeen, Scotland, David Gregory is remembered for his expository writings. In 1684 he published Exercitatio geometrica (Exercises in geometry), outlining many of the results of JAMES GREGORY, his uncle, on infinite SERIES. He also published some of SIR ISAAC NEWTON’s results on mathematics and astronomy, and he was the first to teach the “modern” Newtonian theories at Cambridge. In 1703 he issued the firstever complete collection of all of EUCLID’s works. Gregory studied at Marischal College, part of the University of Aberdeen, and accepted a professorship of mathematics at the University of Edinburgh in 1683. He lectured on the topics of mathematics, mechanics, and hydrostatics. In 1691 he was elected Savilian Professor at Oxford, the same year he was awarded membership to the ROYAL SOCIETY of London. In 1702 Gregory published Astronomiae physicae et geometricae elementa (Astronomical physics and elements of geometry), which was a popular account of Newton’s work, and a year later, his edition of the collected works of Euclid. Gregory strongly supported Newton in the debate over whether it was he, or German mathematician GOTTFRIED WILHELM LEIBNIZ, who had first discovered CALCULUS. Gregory also completed his own work on the study of series, and published work on the topic of optics. He died on October 10, 1708, in Berkshire, England. Not noted as an outstanding mathematician, Gregory is remembered primarily for the role he played in preserving the papers and recording the verbal communications passed to him by his uncle James Gregory and by Sir Isaac Newton.
Gregory, James (1638–1675) British Calculus Born in November 1638 (the exact date is not known), Scottish mathematician James Gregory independently discov
ered and explained important concepts in CALCULUS before the subject was later fully developed by founders SIR ISAAC NEWTON and GOTTFREID WILHELM LEIBNIZ. Gregory was aware of the FUNDAMENTAL THEOREM OF CALCULUS, was the first to distinguish between convergent and divergent SERIES, and knew of TAYLOR SERIES 40 years before BROOK TAYLOR published his results. In mathematics he is remembered, in particular, for the series that bears his name. Outside of mathematics, Gregory is best remembered for his theoretical description of a revolutionary type of reflecting telescope. Gregory was homeschooled in mathematics before undertaking studies at Marischal College in Aberdeen. He studied the mathematics of optics and in 1661 published a text on the subject, Optica promota (The advancement of optics), written in a purely mathematical style. Beginning with five postulates and 37 definitions, Gregory developed the theory of reflection and refraction of light in a systematic and rigorous manner, culminating with the description of a new principle for the construction of a telescope. In 1664 Gregory traveled to Italy to stay at the University of Padua to pursue interests in mathematics. During this time Gregory developed the foundations of “infinitesimal geometry,” the details of which he published as Geometriae pars universalis (The universal part of geometry). Today we would describe this book as a systematic treatment of topics in calculus. Gregory was the first to discuss the convergence and divergence of series, work which led him in 1671 to the discovery of series expansions of functions. He refrained from publishing his discoveries, however, having heard a rumor that Newton may have already developed similar results. (Actually Newton was not aware of the key theorem needed to put this theory on a sound footing and so was unable to develop the theory.) Brook Taylor later discovered the same key result and published his description of the topic in 1715. Throughout his short life, Gregory maintained an interest in astronomy. He died at the age of 36 in late October 1675, shortly after suffering a stroke he incurred while observing the moons of Jupiter with some students. Gregory did not achieve fame as a mathematician during his life. It is only in retrospect that historians today realize the important influence he had in helping Newton develop his ideas. See also GREGORY SERIES.
group 241
Gregory series (Leibniz’s series) Named after the Scottish astronomer and algebraist JAMES GREGORY (1638–75), the MACLAURIN SERIES for arctan(x) is sometimes called the Gregory series: arctan(x) = x −
x3 x 5 x 7 + − +L 3 5 7
This expression is valid for –1 ≤ x ≤ 1. Placing x = 1 into the formula yields the following remarkable formula for PI: π 1 1 1 = 1− + − +L 4 3 5 7 Often the name “Gregory series” is used to mean this particular expression. See also INVERSE TRIGONOMETRIC FUNCTIONS; MADHAVA OF SANGAMAGRAMMA; TAYLOR SERIES.
group Research in pure mathematics is motivated by one question: what makes mathematics work the way it does? By identifying the key principles that underpin one type of mathematical system—be it geometry and symmetry, or numbers and arithmetic— mathematicians establish connections between disparate fields: known facts about any one system satisfying a set of basic principles translate immediately to analogous facts about a second system satisfying analogous principles. For example, a study of arithmetic shows that the operation of addition satisfies the same four basic principles as multiplication. Thus any known fact about addition is accompanied by a known fact about multiplication (and this corresponding result about multiplication consequently requires no new proof.) Motivated by the workings of arithmetic, mathematicians define a group to be any set, often denoted G, whose elements can be combined in some way that mimics the addition or the multiplication of the integers. Specifically, if we denote the result of combining the elements a and b of G by the symbol a*b, then G is a group if the following four axioms hold: 1. Closure: For all a and b in G, the element a*b is also a member of G.
2. Associativity: For all a,b,c in G we have: a*(b*c) = (a*b)*c. 3. Existence of an identity: There is an element e in G so that a*e = a = e*a for all a in G. 4. Existence of inverses: For any a in G there is an element b in G such that a*b = e and b*a = e. To honor the work of Norwegian scholar NIELS HENRIK ABEL (1802–29), mathematicians call G “Abelian” if a fifth axiom also holds: 5. Commutativity: For all a and b in G we have: a*b = b*a. These axioms do indeed capture the working principles behind both addition and multiplication. For example, interpreting “*” as addition with “e” as the number zero, the set of all integers satisfies the axioms of an Abelian group. (In this setting the inverse of an integer a is usually denoted –a.) The set of all real numbers with zero removed forms an Abelian group under multiplication. In this context, * is interpreted as the product operation, “e” is the number 1, and the inverse of an element a is the number 1/a. A group could be abstract. For example, the set of four elements G = {e,a,b,c} is an Abelian group under an operation * given as follows: *
e
a
b
c
e a b c
e a b c
a e c b
b c e a
c b a e
The set of elements G = {1, –1, i, – i, j, – j, k, –k} with group operation *, given by multiplication as QUATERNIONs, is an example of a nonAbelian group. The set of all 2 × 2 invertible matrices with real entries is also a nonAbelian group under the operation of MATRIX multiplication. GROUP THEORY is the study of the general structure of groups and all the results that follow from the four (or five) basic axioms. The subject is incredibly rich, and many mathematicians today devote their entire research careers to the further development of this topic. Group theory has profound applications to physics and science. Any physical system that possesses
242 group theory symmetry of some kind, for instance, can be analyzed through the tools and techniques of this topic. For example, the six symmetries of an equilateral triangle form a (nonAbelian) group: there are two rotations, a clockwise and a counterclockwise rotation of 60°; three reflections, one about each altitude of the triangle; and an identity operation corresponding to conducting no action at all. By declaring the operation “*” to be the effect of performing one action followed by the other, one can check that all four group axioms hold for this system. (The fifth axiom, however, does not hold, since the performance of rotation followed by a reflection, for instance, gives a result different from the performance of that same reflection followed by the rotation.) The analysis of symmetry plays an important role in crystallography and quantum mechanics. Researchers in these fields deem group theory an essential tool in their work. The set of symmetries of any regular nsided POLYGON forms a group called the nth dihedral group. It has 2n members consisting of n rotations (including the identity element) and n reflections. The set of rotations in and unto itself forms a group of just n elements. Notice that n is a factor of 2n. In general, any subset H of a group G that itself satisfies the four axioms of a group is called a subgroup of G. French mathematician JOSEPHLOUIS LAGRANGE (1736–1813) proved that if H is a subgroup of a group G, then the number of elements in H evenly divides the count of elements in G.
This result has interesting consequences when applied to systems of symmetry, or to groups arising from the study of NUMBER THEORY and MODULAR ARITHMETIC. See also ABSTRACT ALGEBRA; FIELD; GENERAL LINEAR GROUP; HOMOMORPHISM; ISOMORPHISM; RING.
group theory The general study of
GROUPs and the results that follow from the basic axioms that define them is called group theory. Many of the key principles behind group theory were first identified by the German mathematician CARL FRIEDRICH GAUSS (1777–1855) in his studies of NUMBER THEORY and MODULAR ARITHMETIC. The development of a group theory as a subject in its own right, however, is usually attributed to the young French mathematician ÉVARISTE GALOIS (1811–32); who devised the innovative tools necessary to study solutions to algebraic equations in depth and from an abstract perspective. By identifying the abstract principles that make algebra and arithmetic work the way they do, group theory provides a powerful tool for analyzing any mathematical system that satisfies the same basic axioms. Applying group theory to the symmetries of a physical system, for example, can often lead to important consequences in physics. See also ABSTRACT ALGEBRA; FIELD; GROUP; RING.
H Hadamard,
Jacques (1865–1963) French Number
study of mathematics education. Hadamard wrote over 300 scientific papers. In 1906 he was elected president of the French Mathematical Society, and 3 years later was appointed chair of the Collège de France. He stayed in that position for only 3 years, before accepting the position as professor of analysis at the École Polytechnique. He was elected to the Paris Academy of Sciences in 1912. Hadamard died in Paris, France, on October 17, 1963. His proof of the prime number theorem is considered his most outstanding achievement.
theory, Analysis, Mathematical physics Born on December 8, 1865, in Versailles, France, Jacques Hadamard is remembered as one of the two mathematicians who independently proved the famous PRIME NUMBER THEOREM first conjectured by the German mathematician CARL FRIEDRICH GAUSS. Belgian scholar CHARLESJEAN DE LA VALLÉEPOUSSIN was the second mathematician to prove the result. Hadamard received a doctoral degree in 1892 from the École Normale Supérieure after completing a dissertation on complex functions and TAYLOR SERIES. The same year he also completed work on the Riemann ZETA FUNCTION and the Riemann hypothesis, earning him the Grand Prix des Sciences Mathématique from the institution. Hadamard soon came to also realize that, with the recent developments in the field of complex functions, all the necessary pieces were now in place to develop a proof of the famous outstanding prime number conjecture made by Gauss. (ValléePoussin independently made the same realization.) Hadamard presented his proof of the famous result in 1896. Later that same year Hadamard was appointed as professor of astronomy and rational mechanics at the University of Bordeaux. Throughout his career Hadamard also made significant contributions to the field of matrix theory by identifying a class of matrices that can be used in COMBINATORICS to create “block designs” and have applications to PROBABILITY theory. He also studied VECTOR SPACE theory (defining the term functional for a linear function on a vector space) and contributed to the
halfplane The plane that lies on one side of a given line is called a halfplane. If the points on the line itself are considered part of the region, then we say that the halfplane is “closed.” An “open” halfplane excludes the points on the line. If the equation of the given line is ax + by = c, then the set of points (x,y) that satisfy ax + by > c form an open halfplane on one side of the line, and those points (x,y) satisfying ax + by < c form an open halfplane on the other side. If c is positive, then this second inequality represents the halfplane that contains the origin (we have a · 0 + b · 0 = 0 < c), and if c is negative, then the first inequality contains the origin. If c equals zero, one must substitute in different values for x and y to determine which inequality represents which halfplane. The inequalities ax + by ≥ c and ax + by ≤ c represent closed halfplanes. In threedimensional space, a halfspace is the region of space that lies on one side of a plane. The 243
244 Hall’s matching theorem halfspace can be either closed or open according to whether or not points of the plane should be considered as part of this region.
Hall’s matching theorem (Hall’s marriage theorem) In 1935 English mathematician Philip Hall established the following result, now known as Hall’s matching theorem: Given n sets A1, A2, … , An with the property that the union of any k of them (1 ≤ k ≤ n) contains at least k distinct elements, it is always possible to select n distinct objects, one from each set. The condition placed on these sets is not trivial—some sets could be empty, or the same element could appear in more than one set, for example. The theorem is certainly true for the case n = 1: a single set satisfying the condition of the theorem contains at least one element. Two sets satisfying the conditions of the theorem together contain at least two distinct elements. As neither set is allowed to be empty, one set contains one element, and the other another distinct element. Thus the theorem also holds true for the case n = 2. One can build up a general proof of the theorem using a proof by INDUCTION. The validity of the theorem can be demonstrated with an amusing game of solitaire: divide a shuffled deck of cards into 13 piles of four cards. The challenge is to select an ace from one pile, a two from another, a three from a third, all the way down to king from a 13th pile. Hall’s matching theorem ensures that this game can always be won, no matter how the cards are shuffled. (Think of each pile as a set containing one, two, three, or four elements—the distinct denominations that appear in that pile. Among any k piles it must be the case that at least k distinct denominations appear.) Hall gave another interpretation to his theorem (explaining the alternative name to the result): Suppose n women each list the names of the men they would like to marry. As long as any k women mention at least k distinct names among them, 1 ≤ k ≤ n, then it is possible to make satisfactory matches for all. See also SEMIMAGIC SQUARE.
halting problem In 1936 computer theorist Alan Turing contemplated whether or not it would ever be possible to write a computer program that could read any other program and determine whether that program will come to a stop or will run forever (by falling into an infinite loop, for example). This question has since become known as the halting problem. Turing concluded that such a program could not possibly exist. He reasoned via a clever ARGUMENT of SELFREFERENCE: Suppose a program HALT(P) exists that can read a computer program P and print yes or no according to whether P will or will not halt. Consider then another program, which we will call TROUBLE, that takes a program P and does the following: TROUBLE (P): If HALT(P) = “yes” then perform an infinite loop. If HALT(P) = “no” then halt. That is, TROUBLE takes a program P and goes into an infinite loop if P is a program that halts, and it halts if P is a program that does not. Now ask: what does TROUBLE (TROUBLE) do? We have that TROUBLE halts if TROUBLE does not halt, and does not halt if it does! This absurdity shows that no such program HALT(P) could exist. There are technical difficulties with this argument (one must be careful to properly distinguish between the roles of a program and the input of a program, for example), but Turing was able to overcome these concerns and show that this argument is fundamentally sound.
Hamilton, Sir William Rowan (1805–1865) Irish Algebra, Graph theory, Number theory, Mathematical physics Born on August 4, 1805, in Dublin, Ireland, William Rowan is generally considered Ireland’s greatest mathematician of all time. He is remembered for his development of an entirely new algebraic system, the QUATERNIONS, which, seven decades later, proved to be crucial for the development of quantum mechanics and the mathematical physics of the internal structure of an atom. Hamilton was a child prodigy, mastering 12 different languages by the age of 13. During the teen years,
handshake lemma he read the works of EUCLID in the original Greek, the works of SIR ISAAC NEWTON in Latin, and the works of PIERRESIMON DE LAPLACE in French. He found a subtle error in Laplace’s classic text Mécanique céleste, and wrote a letter to the astronomer royal of Ireland, John Brinkley, explaining the error and how it should be corrected. Brinkley immediately recognized Hamilton’s genius as a rising mathematician and publicly dubbed him the “first mathematician of his age.” Hamilton entered Trinity College at the age of 18 to study optics and mathematics. His original work in these fields as an undergraduate, which included two papers “Systems of Right Lines in a Plane” and “Theory of Systems of Rays,” was regarded as so significant and innovative as to warrant his immediate appointment as a professor of astronomy at the college before the completion of his basic degree. Along with his work in optics, Hamilton made significant contributions to the field of GRAPH THEORY and to the algebra of COMPLEX NUMBERS, publishing results on the latter topic in his 1837 paper “Preliminary and Elementary Essay on Algebra as the Science of Pure Time.” In 1842 he took on the difficult challenge of trying to create an algebraic system for threedimensional space that had the same algebraic properties as the complex numbers in twodimensional space. [A point (x,y) in the plane can be matched with the complex number x + iy. This thus provides a means to “multiply” to two points in space: (x1, y1) · (x2, y2) = (x1x2 – y1y2, x1y2 + x2y1).] Although he was never able to find a solution to this “multiplication of triples” problem, his efforts did lead him to the remarkable discovery of a different type of number system suitable for fourdimensional space. He called this system the quarternions, and found some surprising connections to mathematical physics. In particular he observed that each quaternion corresponds naturally to a physical transformation in threedimensional space and that the multiplication of two quaternions matches perfectly with the composition of the two physical transformations they represent. Thus the geometry of threedimensional physical space can be reduced to the algebraic study of the algebra of quaternions. Hamilton was convinced his work would revolutionize mathematical physics. Although it does have applications to the field today, sadly, his work did not have the impact he hoped. Hamilton received many awards throughout his life, most notably a knighthood in 1835 and election to
245
the National Academy of Sciences in the United States as its first foreign member. He wrote poetry for solace throughout his life, and argued publicly that the language of mathematics is just as artistic as the language expressed through poetry. His close friend poet William Wordsworth (1772–1834) did not agree. Hamilton died near Dublin on September 2, 1865.
Hamiltonian path/circuit See GRAPH THEORY. hamsandwich theorem As a generalization of the INTERMEDIATEVALUE THEOREM
and the twopancake theorem that follows from it, mathematicians have proved the following result, called the hamsandwich theorem: Given any three objects sitting in threedimensional space, there exists a single plane that simultaneously slices the volume of each object exactly in half. For example, there is a single plane that simultaneously slices the Eiffel tower, the planet Neptune, and this book each precisely in half by volume. The theorem gains its name from the following interpretation: It is possible, in a single planar cut, to divide each of two pieces of bread and a slab of ham into two pieces of equal volume. This is possible no matter the shape of the food pieces and no matter where in space the three items are placed. The result generalizes to higher dimensions: Given any N objects sitting in Ndimensional space, it is always possible to find an (N – 1)dimensional “plane” that simultaneously slices the “volume” of each object in half. (With N = 2, this is the twopancake theorem.)
handshake lemma This amusing result states that, at any instant, the number of people on this planet, living or deceased, who have taken part in an odd total number of handshakes is necessarily even. This lemma can be proved with the aid of GRAPH THEORY.
246 Hardy, Godfrey Harold
Hardy, Godfrey Harold (1877–1947) British Analysis, Number theory Born on February 7, 1877, in Cranleigh, England, eminent mathematician Godfrey Hardy is remembered for his significant contributions to the fields of NUMBER THEORY, inequalities, and to the study of the Riemann ZETA FUNCTION and the Riemann hypothesis. Hardy also encouraged the Indian mathematician SRINIVASA AIYANGAR RAMANUJAN to come to England, and collaborated with him for five years to produce a number of significant results, most notably on the theory of PARTITIONs. In 1940 Hardy published A Mathematician’s Apology, which remains, to this day, one of the most vivid and eloquent descriptions of how a mathematician thinks. Hardy entered Trinity College, Cambridge, in 1896 to pursue an advanced degree in mathematics. After publishing a number of papers on the topics of INTEGRALs, SERIES, and general topics of ANALYSIS, Hardy published, in 1908, the undergraduate textbook A Course of Pure Mathematics, which explained, in a rigorous manner, the concepts of function, LIMIT, and the elements of analysis. This work was very influential and is said to have transformed the entire nature of university teaching. In 1911 Hardy began a 35yearlong collaboration with English mathematician John Littlewood (1885–1977), leading to a change of focus in the field of number theory. Together they produced over 100 joint publications. This work also tied in nicely with the research Hardy was conducting in 1914 with Ramanujan, also on topics in number theory. Hardy was recognized as an important figure in mathematics. In 1910 he was elected a fellow of the ROYAL SOCIETY of London, received the Royal Medal of the Society in 1920, the De Morgan Medal of the Society in 1929, and the Sylvester Medal of the Society in 1940 for his work in pure mathematics. Seven years later he was awarded the Copley Medal of the Society for his contributions to the field of analysis. Hardy died in Cambridge, England, on December 1, 1947. He is generally recognized as the leading English pure mathematician of his time. His 1932 book The Theory of Numbers, cowritten with E. M. Wright, is considered a classic text and is still used by graduate students and researchers in the field of number theory today.
harmonic function A function defined at a discrete number of locations is said to be harmonic if the value
of the function at any one location equals the average of the function values at its neighboring locations. For example, if 10 students sit in a circle, then their “age function” would be harmonic if the age of any one student equals the average age of his or her two neighbors. (A little thought shows that this is only possible if the age of each student is the same. A harmonic function for points arranged in a circle must be constant.) Harmonic functions play an important role in the study of RANDOM WALKs and calculating odds in gambling. Imagine, for example, a gambler playing a simple game of tossing a coin to either win a dollar or to lose a dollar. We ask: with $3 in hand, what are her chances of reaching the $10 mark before going broke? To compute this, let P(N) denote the PROBABILITY of achieving the goal starting with N dollars in hand. Clearly P(0) = 0 and P(10) = 1. We wish to compute P(3). The key is to note that P(N) is a harmonic function on its 11 values zero through 10. This follows because there are equal chances for the gambler to lose or win a dollar and thus to next play with either N – 1 or N + 1 dollars in hand. Consequently: P(N) =
1 1 P(N − 1) + P(N + 1) P(N − 1) + P(N + 1) = 2 2 2
and so each quantity P(N) is indeed the average of the values just preceding and succeeding it. Some thought shows that the values P(0) = 0 up to P(10) = 1 must be strictly increasing by equal intervals of onetenth. The N 3 values P(N) are thus P(N) = – . In particular, P(3) = – , 10 10 showing that there is only a 30 percent chance that the gambler will achieve her goal before losing all her cash.
harmonic mean See MEAN. harmonic sequence (harmonic progression) A
SE1 1 1 QUENCE of numbers of the form – , – , – ,… with a1 a2 a3 integers a1,a2,a3,… forming an ARITHMETIC SEQUENCE is called a harmonic sequence. The numbers a1,a2,a3,… have a constant difference between them. For example, 1 1 1 1 1 1 1, – , – , – ,… is a harmonic sequence, as is – , – , – 2 3 4 4 7 10 1 1 1 1 1 – ,… and – , – , – , ––– ,… . The word harmonic is 13 25 50 75 100
harmonic series 247 used because the nth harmonic produced by a violin string is the tone produced by the string that is 1/n times as long. ∞
The corresponding
SERIES
1
∑a
n=1 n
for any harmonic
sequence necessarily diverges. In the study of CONVERGENT the comparison test shows that this must be the ∞ 1 case. (Compare a series of the form ∑ a + dn with n =0 ∞ 1 ∑ , which we know diverges.) n=1 n See also HARMONIC SERIES.
1.53 × 1043 terms an answer larger than 100. Clearly the series diverges to infinity very slowly. In 1734 LEONHARD EULER showed that, for large values of n, the nth PARTIAL SUM of the harmonic series can be well approximated by a LOGARITHM: 1+
SERIES,
1
harmonic series The particular infinite sum 1 + – + 2 1 1 1 – + – + – + … is called the harmonic SERIES. The 3 4 5 word harmonic is used because the nth harmonic produced by a violin string is the tone produced by the string that is 1/n times as long. Even though the terms of this series approach zero, the series does not sum to a finite value. This can be seen by grouping the terms of the series into sections of length two, four, eight, 16, and so on, and making a simple comparison: 1+
1 1 1 1 1 1 1 1 + + + +L+ + +L+ + +L 2 3 4 5 16 17 32 33
1 1 1 1 1 1 1 = 1+ ( ) + ( + ) + ( +L+ ) + ( +L+ ) 2 3 4 5 16 17 32 +(
1 +L 33
1 1 1 1 1 1 1 > 1+ ( ) + ( + ) + ( +L+ ) + ( +L+ ) 2 4 4 16 16 32 32 +(
1 +L 64
1 1 1 1 1 = 1+ ( ) + ( ) + ( ) + ( ) + ( ) +L 2 2 2 2 2 =∞ That the series diverges means that summing sufficiently many initial terms of the series will produce answers arbitrarily large (although it may take a large number of terms to do this). For example, summing the first four terms produces an answer larger than 2, the first 11 terms an answer larger than 3, the first 13,671 terms an answer larger than 10, and the first
1 1 1 + + L + ≈ ln(n) + γ 2 3 n
where γ ≈ 0.577 is a constant (called EULER’S CONand the error in this approximation is no larger than 1/n. (Notice that ln(n) + γ → ∞ as n grows. This again shows that the series diverges.) The partial sums of the harmonic series are called the harmonic numbers, and are denoted Hn. The first 10 harmonic numbers are: STANT)
H1 = 1; H2 =
3 11 11 ; H3 = ; = 2 6 2 ⋅3
H4 =
25 137 49 363 ; H5 = ; H6 = ; H7 = ; 4 ⋅3 4 ⋅ 15 4⋅5 4 ⋅ 35
H8 =
761 7129 7381 ; H9 = ; H10 = . 8 ⋅ 35 8 ⋅ 315 8 ⋅ 315
One can use an induction argument to show that if 2k ≤ n < 2k+1, then the denominator of Hn (written in reduced form) is a multiple of 2k. Consequently, no denominator (except for the first) can be 1. This proves: No harmonic number, other than the first, is an integer. The divergence of the harmonic series solves the amusing rubberband problem: An infinitely tiny ant starts at one end of a rubber band, 1 ft long, and crawls a distance of 1 in. toward the other end. It then pauses, and the band is stretched 1 ft longer (to a total length of 2 ft), carrying the ant along with it to the 2in. position. The ant then crawls for another inch, to the 3in. position, and pauses while the band is stretched another foot longer. This process of walking an inch and pausing while the band stretches a foot continues indefinitely. (We assume the band is infinitely elastic.) Will the ant ever make it to the end of the rubber band?
248 helix Note that in the first leg of the journey, the ant covers 1/12th of the length of the band, and that this proportion remains the same as the band is stretched. During the second leg of the journey, the ant covers now only 1/24th of the length of the band (1 in. of 24 in.), 1/36th of the length during the third length, and so on. Thus after n legs of the journey, the ant covers the fraction 1 1 + 1 + 1 + L + 1 = H n of the band. The 12 1 2 3 n 12 ant reaches the finish only if Hn ever surpasses the value 12. As the harmonic series diverges, this must indeed be the case. Although the harmonic series diverges, the ALTERNATING SERIES test from the study of CONVERGENT SERIES shows that the alternating harmonic series, given by 1 1 1 1 1 1 – – + – – – + – – – +… 2 3 4 5 6 converges. We can use Euler’s approximation formula to find its value. First note: S2n = 1 −
1 1 1 1 1 + − +L+ − 2 3 4 2n − 1 2n
1 1 1 1 1 1 1 = 1 + + + L + + − 2 + + L + 2 3 2n − 1 2n 2n 2 4 1 1 1 1 1 1 1 = 1 + + + L + + − + +L+ 2 3 2n − 1 2n 1 2 n = ln(2n) + γ − ln(n) − γ + error = ln 2 + error where the error is no larger than 2/n and so converges to zero as n grows. Consequently, the even partial sums of the series approach the value In(2). The sum 1 = of an odd number of terms equals S2n+1 = S2n + 2n + 1 1 1 S2n+1 = S2n + = ln2 + error + and so too 2n + 1 2n + 1 approaches ln2 as n becomes large. Thus we have: 1 1 1 1 1 1 – – + – – – + – – – + … = ln(2) 2 3 4 5 6 This example is often used to illustrate the difference between conditional and ABSOLUTE CONVERGENCE of a series. In that setting, it can also be used to provide an amusing “proof” that 1 equals 2. See also MERCATOR’S EXPANSION.
helix A spiralshaped curve sitting in threedimensional space is called a helix. The name is the Greek word for a “spiral” or a “twist.” A cylindrical helix lies on a cylinder and cuts across straight lines drawn along the length of the cylinder at a constant angle α. A spiral staircase and the thread on a straight screw are examples of cylindrical helices. A conical helix is a spiral curve on a CONE, and a spherical helix is a spiral on a SPHERE that cuts lines of longitude at a constant angle. A cylindrical helix has PARAMETRIC EQUATIONS: x = a cost, y = a sint, and z = bt, where a and b are constants, and t is the parameter. A conical helix is given by: x = aet cost, y = aet sint, and z = et.
Heron of Alexandria (ca.10–75 C.E.) Greek Geometry, Number theory, Physics, Engineering Sometimes called Hero, Heron of Alexandria is remembered in mathematics for his threevolume text Metrica, rediscovered in 1896. The work discusses and develops in great detail the principles of geometry, number, and numerical approximation. It also contains the earliest known proof of the famous formula that bears Heron’s name. Outside of mathematics, Heron is best known for his contributions to mechanics and fluid mechanics. Demonstrating a wide range of scientific interests, Heron wrote studies in optics, pneumatics (the study and use of gas and fluid pressures), astronomy, surveying techniques, and planar and solid geometry, but it is the work presented in Metrica that proves his genius as a mathematical intellect. Book I of this famous piece computes the areas of triangles, quadrilaterals, and regular polygons, as well as the surface areas of cones, prisms, spheres, and other threedimensional shapes. His famous formula for the area of a triangle solely in terms of its sidelengths is presented in this section, along with a general method for computing the square root of a number to any prescribed degree of accuracy. (This procedure is today called HERON’S METHOD. Book II of Metrica derives formulae for the volume of each PLATONIC SOLID, as well as the volumes of cones and spherical segments, and Book III extends EUCLID’s study of geometric division. Heron also presents a method for determining the cube root of a number in this third volume. Heron also wrote a number of important treatises on mechanics, many of which survive today. His text Pneumatica represents a careful (but, in places, inaccu
Hilbert, David 249 rate) study of pressure in fluids, along with a description of a collection of trick gadgets and toys illustrating specific scientific principles. He also describes designs for over 100 practical machines, including pneumatic pulleys and lifts, wind organs, coinoperated machines, fire engines, and steampowered engines that operate in a way similar to today’s jet engine. Some of Heron’s texts have the appearance of draft lecture notes, leading some historians to suspect that he may have taught at the famous Museum of Alexandria. Little is actually known of Heron’s life.
Heron’s formula (Hero’s formula) In his work Metrica, HERON OF ALEXANDRIA (ca. 100 C.E.) presented a formula for the AREA of a TRIANGLE solely in terms of the sidelengths of the triangle. Today known as Heron’s formula, it reads: area = √s(s – a)(s – b)(s – c) where a, b, and c are the sides of the triangle, and 1 s = – (a + b + c) is its “semiperimeter.” The formula is a 2 special case of BRAHMAGUPTA’S FORMULA, discovered 500 years later. Heron’s formula can be proved as follows: if θ is the angle between the sides of length a and b, then the area 1 of the triangle is given by area = – ab sin(θ). The LAW 2 OF COSINES asserts that c2 = a2 + b2 –2ab cos(θ). Solving for sin(θ) and cos(θ), and substituting into the standard identity from TRIGONOMETRY, cos2(θ) + sin2(θ) = 1, yields, after some algebraic work, Heron’s result. See also BRETSCHNEIDER’S FORMULA; MEDIAN OF A TRIANGLE; QUADRILATERAL; TRIANGLE.
Heron’s method (Hero’s method) In Book I of his volume Metrica, HERON OF ALEXANDRIA gives a method for approximating the SQUARE ROOT of a number. It works as follows: Estimate the value of the square root. Divide this guess into the number under consideration, and take the average of the result and the initial estimate. This will produce a better approximation to the square root. Repeated application of this method produces an estimate to any desired degree of accuracy.
In symbols, Heron claims that if x approximates N x+ x is a better the square root of a number N, then 2 approximation. For example, taking 3 as an approxima10 3+ tion to the square root of 10, we obtain 3 ≈ 3.1667 2 as an improved estimate. Repeating the procedure yields 10 3.1667 + 3.1667 ≈ 3.1623 as an even better approxima2 – tion. (In fact, √10 ≈ 3.16227766.) N x+ x . To show why this method works, let y = 2 1 (x − N )2 , which shows that if the Then (y − N ) = 2 x – – error x – √N is small, then the error y – √N will be even smaller. (We are assuming here that x is a value greater than 1.) This method was known to the Babylonians of 2000 B.C.E. It is also equivalent to NEWTON’S METHOD when applied to the function f(x) = x2 – N. See also BABYLONIAN MATHEMATICS.
higher derivative Taking the DERIVATIVE of the same function more than once, if permissible, produces the higher derivatives of that function. The first, second and third derivatives of a function f(x) are denoted, respectively, f ′(x), f ′′(x) and f′′′(x), and for n ≥ 4, the nth derivative as f (n)(x). For example, the third derivative of f(x) = x4 + sin x is f′′′(x) = 24x – cos x. This notation for the repeated derivative is due to JOSEPHLOUIS LAGRANGE (1736–1813). GOTTFRIED WILHELM LEIBNIZ (1646–1716), coinventor of CALCULUS, used the dn f(x) notation –––– dxn for the higher derivatives, and French mathematician Louis Arbogast (1759–1803) wrote Dn f(x). All three notational systems are used today.
Hilbert, David (1862–1943) German Formal logic, Geometry, Mathematical physics, Algebraic number theory Born on January 23, 1862, in Königsberg, Prussia (now Kaliningrad, Russia), mathematician David Hilbert is remembered as one of the founding fathers of 20thcentury mathematics. In 1899 Hilbert
250 Hilbert’s infinite hotel published his famous Grundlagen der Geometrie (Foundations of geometry), in which he provided a completely rigorous axiomatic foundation of the subject clarifying the hidden assumptions that EUCLID had made in his development of the subject two millennia earlier. Moreover, Hilbert advanced the topic of FORMAL LOGIC and used his results to prove that his approach to geometry is CONSISTENT given that the arithmetic of the real numbers is free of contradictions. In 1900 Hilbert posed 23 problems to the mathematicians of the 20th century that he felt lay at the heart of vital mathematical research. Two of his problems were solved almost immediately, but the remaining 21 challenges did indeed stimulate important mathematical thinking. Many of his challenges are still unsolved today. Hilbert also greatly influenced the development of quantum theory in theoretical physics: his notion of a “Hilbert space” provided the right conceptual framework for the subject. Hilbert also made important contributions to the fields of special relativity and general relativity. Hilbert received a doctorate of mathematics from the University of Königsberg in 1885 after writing a thesis in ABSTRACT ALGEBRA. He was appointed to a faculty position at the university, where he remained for 10 years before accepting the position as chair of mathematics at the University of Göttingen in 1895. Hilbert taught and worked at Göttingen for the remainder of his career. His 1897 text Zahlbericht (Number theory) was hailed as a brilliant synthesis of current thinking in algebraic number theory, and the original results it contained were acknowledged as outstanding. Hilbert’s abilities to grasp the subtleties of a sophisticated mathematical theory, develop penetrating insights, and provide new innovative and stimulating perspectives on a subject were apparent. Throughout his career Hilbert worked on a wide variety of disparate subjects, making groundbreaking contributions to each before moving on to the next. He published his famous work on EUCLIDEAN GEOMETRY in 1899. In 1900 Hilbert was invited to address the Paris meeting of the International Congress of Mathematicians. During his speech he detailed 10 mathematical problems that he felt were of great importance. (He expanded the list to 23 problems when he published his address.) These problems include the CONTINUUM HYPOTHESIS, GOLDBACH’S CONJECTURE, a search for
the axiomatization of physics, and a search for a general algorithm for solving DIOPHANTINE EQUATIONs. Some important progress, and in many cases, complete solution, has been made on all the challenges posed except for one, the socalled Riemann hypothesis, which asks for the locations of the roots of the ZETA FUNCTION. This remains, perhaps, the most famous unsolved problem of today. Later in life Hilbert worked on formal logic and on the foundations of theoretical physics. Between 1934 and 1939 he published two volumes of Grundlagen der Mathematik (Foundations of mathematics), cowritten with Paul Bernays (1888–1977), which were intended to develop a proof of the consistency of mathematics. (GÖDEL’S INCOMPLETENESS THEOREMS showed, however, that such a goal is unattainable.) His development of functional analysis provided the correct mathematical framework for the theory of quantum mechanics. Hilbert received many honors throughout his career, including a special citation from the Hungarian Academy of Sciences in 1905. Upon his retirement in 1929, the city of Göttingen named a street after him, and the city of Königsberg, his birthplace, declared him an honorary citizen. He died in Göttingen on February 14, 1943. He is remembered for shaping the very nature of 20thcentury research in the pure mathematics.
Hilbert’s infinite hotel (Hilbert’s paradox) German mathematician DAVID HILBERT (1862–1943) observed that studies of the infinite can often lead to nonintuitive and surprising conclusions. His famous infinitehotel paradox illustrates some of his ideas: Imagine a hotel with an infinite number of rooms. Suppose every room is occupied. Is it possible for the hotel to accommodate one more guest? If the rooms are numbered 1, 2, 3, and so on, Hilbert pointed out that, despite the inconvenience, each existing guest can be moved from room n to room n + 1, thereby leaving room 1 free for a latecomer. There is indeed room for another single guest. One can go further: suppose a tourist bus, with an infinite number of tourists on board, arrives at the hotel. Moving each hotel guest from room n to room 2n, instead, leaves all the oddnumbered rooms vacant for the infinite number of new arrivals.
Hippocrates of Chios 251 If an infinite number of tourist buses, each containing an infinite number of tourists, arrive at the same time, one can still find room to accommodate the multitude of new guests: Move all the existing hotel guests to the evennumbered rooms. Place the tourists from the first bus into the rooms given by the powers of 3: 3, 9, 27,… Place the tourists from the second bus into the rooms given by the powers of 5: 5, 25, 125,… Next use all rooms numbered the powers of 7, of 11, of 13, and so on, down along the list of all the PRIME numbers. As, according to EUCLID’S PROOF OF THE INFINITUDE OF PRIMES, there are infinitely many primes, and no two powers of different primes are the same, this allocation scheme does the trick. See also CARDINALITY; FUNDAMENTAL THEOREM OF TRISTRAM SHANDY PARADOX.
ARITHMETIC; INFINITY; PARADOX;
HinduArabic numerals The numeral system we use today is called the HinduArabic numeral system. Using a base10 PLACEVALUE SYSTEM, numbers are expressed via combinations of the symbols 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9, organized so as to represent groupings of powers of 10. For instance, the number 574 represents the five groups of 100, seven groups of 10, and four single units. This numerical system originated from India around 600 C.E., almost in the exact same form as we use it today. The system was transmitted to the Arabs two centuries later as they worked to translate the Sanskrit works on astronomy into Arabic. The Arab math– –A ALKHW–ARIZM–ı ematician MUHAMMAD IBN MUS (ca. 800) wrote an influential treatise describing the Hindu numeral system, and used it in his famous book Hisab aljabr w’almuq¯abala (Calculation by restoration and reduction), from whose title the modern word algebra is derived. As Western scholars began translating the Arabic texts into Latin, word of the efficient numeration system spread across Western Europe. Italian scholar FIBONACCI (ca. 1170–1250) avidly promoted their use. By the end of the 17th century, the HinduArabic numeral system completely replaced the
cumbersome system of ROMAN NUMERALS that were the standard in Europe for over 1,500 years. See also BASE OF A NUMBER SYSTEM; DECIMAL REPRESENTATION; NUMBER; ZERO.
Hippasus of Metapontum See PYTHAGORAS. Hippocrates of Chios (ca. 470–410
Greek Geometry Born in Khíos (Chios), Greece, Hippocrates is remembered as the first mathematician to have found the area of a curved figure, namely, that of a LUNE. He also made first steps toward properly analyzing the problem of SQUARING THE CIRCLE. Very little is known of Hippocrates’ life. It is thought that Hippocrates may have first worked as a merchant before developing an interest and talent for geometry later in his life. Although he wrote only one text on the topic (unfortunately lost to us today), many scholars throughout history referred to his work in their own studies. Hippocrates was interested in the famous problem of squaring the circle, that is, finding a method of constructing a square of the same area as a given circle. Although he was unable to solve this problem, he did manage to show that, in many instances, it is possible to “square a lune.” For example, in the diagram below, figure ABCD is a square, and the shaded region is the lune formed by one circular arc with center D and a second circular arc with center O, the center of the square. Hippocrates, using geometric reasoning, showed that the area of the lune matches the area of the smaller square shown.
Hippocrates lune
B.C.E.)
252 histogram (One can check this algebraically by labeling the radius of the larger circular arc r and computing the sidelength of the large square, the area of the sector DAC, the area of the semicircle ABC, and consequently the area of the lune.) In applying this type of analysis to a number of different lunes, Hippocrates managed to prove that the ratio of the areas of two different circles is the same as the ratio of their radii squared. This new result was a significant achievement. Some historians believe that EUCLID may have drawn on the work of Hippocrates when he described and proved this result in his famous text THE ELEMENTS.
histogram See DISTRIBUTION; STATISTICS: DESCRIPTIVE.
homogeneous A
POLYNOMIAL in several variables p(x,y,z,…) is called homogeneous if all terms in the polynomial have the same total degree. For example, each term of the polynomial x3 + 5x2y – xy2, has total degree three, and so represents a homogeneous polynomial. More generally, a function of several variables f(x,y,z,…) is homogeneous if there is a natural number n such that the following relation holds for any nonzero constant k:
f(kx,ky,kz,…) = kn f(x,y,z,…) For example, the function x 2 sin(
y2 x+y ) + xy cos( 2 ) is y x
homogeneous. Identifying homogeneous functions can be helpful in solving DIFFERENTIAL EQUATIONs. Any formula that represents the MEAN of a set of numbers is required to be homogeneous. In physics, the term homogeneous describes a substance or an object whose properties do not vary with position. For example, an object of uniform density is sometimes described as homogeneous. In cooking, when creaming butter and sugar, for instance, one aims to produce a homogeneous mixture.
homomorphism A map f between two sets S and T, each equipped with some kind of algebraic structure, is called a homomorphism (Greek, homos, “same”; mor
phé, “form” or “shape”) if it preserves the algebraic relations within the sets. For example, if the algebraic operation is “multiplication” in the set S and “addition” in the set T, and if z is the product of two elements in S, z = x × y, then f(z) should correspond to the sum of the two corresponding elements in T: f(x × y) = f(x) + f(y) The LOGARITHM function, for example, is a homomorphism from the set of positive real numbers under multiplication to the set of all real numbers under addition. If f is the doubling function, f(x) = 2x, on real numbers, then f is a homomorphism under addition, but not under multiplication: f(x + y) = 2(x + y) = 2x + 2y = f(x) + f(y) f(x × y) = 2 × x × y ≠ f(x) × f(y) The squaring map f(x) = x2 preserves multiplication but not addition. A map between two GROUPs is called a homomorphism if it preserves the algebraic operations of the groups. A map f : S → T between two RINGs is called a homomorphism if it preserves both the additions and multiplications of the system: f(x + y) = f(x) + f(y) f(x × y) = f(x) × f(y) for all elements x and y of the ring S. See also ISOMORPHISM.
Hypatia of Alexandria (ca. 370–415 C.E.) Greek Philosophy, Commentary Born about 370 in Alexandria, Egypt, Hypatia is remembered as the first woman in the history of mathematics to have made significant contributions to the field, both as a scholar and as a teacher. She was the daughter of mathematician and astronomer Theon of Alexandria. Hypatia wrote influential commentaries on her father’s work and also on the work of APOLLONIUS OF PERGA (ca. 263–190 B.C.E.) and DIOPHANTUS OF ALEXANDRIA (ca. 200–284 C.E.). Little is known of Hypatia’s life. She likely received instruction in mathematics from her father, and certainly maintained a close working relationship with him throughout her life. It is known that she assisted
hyperbola 253 him in writing his 11part commentary on the mathematical work of CLAUDIUS PTOLEMY (85–165 C.E.) and on his production of a revised version of EUCLID’s The Elements. Hypatia produced her own commentaries on classical pieces, including Diophantus’s famous Arithmetica, Apollonius’s Conics, and astronomical works by Ptolemy. All of her work, however, is today lost, and we know of them only through references made by later scholars. Around 400 C.E. Hypatia headed the Platonist school at Alexandria, where she consulted on scientific matters and lectured on philosophy and mathematics. During this time, Christianity surfaced as the dominant religion of the region, and fanatics felt threatened by her intellect and scholarship. Around 415 C.E. Hypatia was brutally murdered by a group of religious followers who deemed her philosophical views pagan. Many historians suggest that the death of Hypatia marks the beginning of Alexandria’s decline as the great center of scholarship and learning of antiquity. It is worth mentioning that at least one other woman is known to have played an active role in mathematics during the Greek times. In his work Collection, PAPPUS OF ALEXANDRIA (300–350 C.E.) gives acknowledgment to a female scholar by the name of Pandrosion. Essentially nothing is known about her.
hyperbola As one of the CONIC SECTIONS, the hyperbola is the plane curve consisting of all points P whose distances from two given points F1 and F2 in the plane have a constant difference. The two fixed points F1 and F2 are called the foci of the hyperbola. The hyperbola also arises as the curve produced by the intersection of a plane with the two nappes of a right circular CONE. Using the notation PF1 and PF2 for the lengths of the line segments connecting P to F1 and F2, respectively, the defining condition of a hyperbola can be written as one of two equations: PF1 – PF2 = d or PF2 – PF1 = d where d denotes the constant difference. Each equation defines its own curve, or branch, of the same hyperbola. The equation of a hyperbola can be found by introducing a coordinate system in which the foci are located at positions F1 = (–c, 0) and F2 = (c,0), for some
Hyperbola
positive number c. It is convenient to write d = 2a, for some a > 0. If P = (x,y) is an arbitrary point on the hyperbola, then, according to the DISTANCE FORMULA, the defining conditions read: (x + c)2 + y 2 − (x − c)2 + y 2 = ± 2a Moving the second radical to the righthand side, squaring, and simplifying yields the equation: c (x − c)2 + y 2 = ± x − a a x2 y2 − = 1. a2 c 2 − a2 Noting that c is greater than a, we can set the positive quantity c2 – a2 as equal to b2, for some b > 0. Thus the equation of the hyperbola is: Squaring and simplifying again yields
x2 y2 − =1 a 2 b2 Conversely, one can show that any equation of this form does indeed yield a hyperbola with foci at positions
254 hyperbolic functions (± √a2 + b2 ,0), and whose points P have distances with common difference 2a from the foci. This equation also reveals that the hyperbola has slant ASYMPTOTEs b x. y = ±– a Hyperbolas have the following reflection property: any ray of light that approaches the convex side of one branch along a line pointing toward one focus is reflected directly toward the other focus. (This property can be proved in a similar way that the reflection property of a PARABOLA is proved.) Hyperbolas appear in the folding of a thin piece of paper. Draw a circle and a dot outside the circle on a sheet of paper. Fold the dot onto the circle and crease the paper. Open up the fold and do this again, this time folding the dot to a different point on the circle. As you do this many times, one branch of a hyperbola will emerge along the side of all the creases. The marked dot is one focus of the hyperbola, and the center of the circle is the other. The radius of the circle is the constant difference of distances of a point on the hyperbola from the two foci. In the process of deriving the equation of a hyperbola, we presented the equation: c (x − c)2 + y 2 = x − a a valid for one branch of the figure at least. Set e =
c = a
a 2 + b2 b2 = 1 + 2 . This is called the ECCENTRICITY a a of the hyperbola and has a value greater than 1. The above equation can be rewritten: (x − c)2 + y 2 =e a x− e
focus) to its distance from a fixed line (the directrix) equals a constant e > 1. See also APOLLONIUS’S CIRCLE; ELLIPSE.
hyperbolic functions In analogy to the fact that the sine and cosine functions of TRIGONOMETRY represent the coordinates of a point on the unit circle, x2 + y2 = 1, the hyperbolic functions represent the coordinates of a point on the right branch of a HYPERBOLA, x2 – y2 = 1. Specifically, the xcoordinate of a point on the curve, at an “angle” t, is called the hyperbolic cosine function, and is denoted cosh(t), and the ycoordinate is called the hyperbolic sine function, denoted sinh(t). We have: cosh2(t) – sinh2(t) = 1 One can also see that cosh(–t) = cosh(t) and sinh(–t) = –sinh(t). Mathematicians sometimes define the hyperbolic sine and cosine functions via the formulae: cosh(t) =
e t + e −t 2
sinh(t) =
e t − e −t 2
One can check that these formulae do indeed satisfy the equation of a hyperbola x2 – y2 = 1 with x > 0, and so must indeed be the coordinates of its points. (These equations are reminiscent of the formula cos(t) = e it + e − it e it + e − it and sin(t) = , which follow from 2i 2 EULER’S FORMULA for ordinary trigonometric functions.)
The numerator of the quantity on the left side is the distance of a given point P from a focus, and the denominator is the distance of the point P from the vera , called a directrix of the hyperbola. tical line x = – e This formulation provides an alternative characterization of the hyperbola: A hyperbola is the set of all points P such that the ratio of its distance from a fixed point (the
The trigonometric and the hyperbolic functions
hyperboloid 255 Taking the DERIVATIVE we see: d d e t + e −t e t − e −t cosh(t) = = = sinh(t) dt dt 2 2 and d sinh(t) = cosh(t) dt In analogy to the ordinary trigonometric functions, mathematicians define four additional hyperbolic functions: tanh(t) =
1 sinh(t) , coth(t) = , cosh(t) tanh(t)
sech(t) =
1 1 , csch(t) = cosh(t) sinh(t)
The hyperbolic cosine curve arises in nature as the shape of the curve formed by a chain hanging freely between two points. This curve is called a CATENARY. See also OSBORNE’S RULE.
hyperbolic geometry (Lobachevskian geometry) Independently discovered in 1823 by Hungarian mathematician JÁNOS BOLYAI (1802–60) and in 1829 by Russian mathematician NIKOLAI IVANOVICH LOBACHEVSKY (1792–1856), hyperbolic geometry is a NONEUCLIDEAN GEOMETRY in which the famous PARALLEL POSTULATE fails in the following manner: Through a given point not on a given line, there is more than one line parallel to that given line. French mathematician JULES HENRI POINCARÉ (1854–1912) later provided a simple model for this geometry and the means to easily visualize geometric results in this theory. The “Poincaré disk” consists of all the points in the interior of the UNIT CIRCLE. A “point” in geometry is any point inside this circle, and a “line” is to be interpreted as a circular arc within the circle with endpoints perpendicular to the boundary of the circle. Any diameter of the boundary circle is also considered a line. Distances are not measured with a traditional ruler: points on the boundary circle are considered to be infinitely far from the center of the circle.
Bolyai and Lobachevsky showed that all but the fifth of EUCLID’S POSTULATES hold in the hyperbolic geometry and, moreover, that this model of geometry is consistent (that is, free of CONTRADICTIONs). This establishes that the parallel postulate cannot be logically deduced as a consequence of the remaining axioms proposed by Euclid. In hyperbolic geometry, all angles in triangles sum to less than 180°, and the ratio of the circumference of any circle to its diameter is less than π. (Moreover, the value of this ratio is not the same for all circles.) Also, it is possible for two perpendicular lines to be parallel to the same line. Physicists, following the work of ALBERT EINSTEIN, suggest that the geometry of our universe is hyperbolic: that it appears to us as Euclidean is a result of the fact that we occupy such a small portion of it. (This is analogous to the fact that it is difficult to recognize the Earth as round when living on it.) See also EUCLIDEAN GEOMETRY; PLAYFAIR’S AXIOM; SPHERICAL GEOMETRY.
hyperboloid The
SOLID OF REVOLUTION obtained by rotating a HYPERBOLA about one of its axes is called a hyperboloid. If the rotation is performed about the axis that lies between the two branches of the curve, then the resulting surface is called a hyperboloid of one sheet. It resembles a cylinder “pinched” at the center so as to curve inward. Points on this surface satisfy an y2 x2 z2 equation of the form –2 + –2 = –2 + 1 for some a c b constants a, b, and c. If, instead, the rotation is performed about the axis that connects the two foci of the hyperbola, then the resulting surface has two distinct parts, each resembling a bowl. This surface is called a hyperboloid of two sheets. Points on this surface satisfy x2 y2 z2 the equation –2 + –2 = –2 – 1. a c b One can construct a model of a hyperboloid of one sheet by holding two metal rings, one directly above the other, and tying vertical strings from points on the lower ring to their corresponding points on the top ring. If one then rotates the top ring so that the vertical strings begin to tilt, the model begins to “constrict.” The resulting surface is a hyperboloid of one sheet. One can also see this surface by tying a string around a handful of uncooked spaghetti strands.
256 hypercube
By rotating the top disc of this 1872 string model, one can transform a cylinder of chords into the shape of a hyperboloid. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
When the strands tilt, the surface of a hyperboloid of one sheet appears. See also ELLIPSOID; PARABOLOID.
hypercube The higherdimensional analog of SQUARE in two dimensions and a CUBE in three dimensions is called a hypercube. A unit square in the plane has four vertices at locations (0,0),(1,0),(0,1) and (1,1). A unit cube in threedimensional space has eight vertices at locations (0,0,0), (1,0,0), (0,1,0), (1,1,0), (0,0,1), (1,0,1), (0,1,1) and (1,1,1). By analogy, a hypercube of sidelength 1 sitting in fourdimensional space has 16 vertices given by: (0,0,0,0), (1,0,0,0), (0,1,0,0),…,(1,1,1,1). Some people find it helpful to interpret the fourth coordinate as time. In this context, each particular value of the fourth coordinate represents a given instant in time, and the remaining three coordinates describe an ordinary object in threedimensional space. A fourdimensional hypercube can thus be thought of as an ordinary cube moving through time. (One can also apply this interpretation to ordinary squares and cubes of lower dimension. For example, a line segment is the result of a point sliding horizontally over time, a
square is the result of a horizontal line segment sliding in the direction of the positive yaxis over time, and a cube is the result of a twodimensional square sliding in the direction of the positive zaxis over time.) A fourdimensional hypercube is also called a tesseract. A square in twodimensional space has four vertices and four edges. A cube in threedimensional space has eight vertices and six square faces. One can show that a tesseract has eight vertices and six cubic “faces,” and that, in general, an ndimensional hypercube has 2n vertices and 2n faces, each itself a hypercube of one dimension less. Four line segments hinged endtoend fold to make a square. Six squares arranged in a “cross” fold to make a cube. Eight cubes arranged in an analogous way—four in a row with the four remaining cubes attached to the four exposed faces of the second cube—fold in fourdimensional space to make a hypercube. All in all, there are 261 different ways to arrange eight ordinary cubes that fold in fourdimensional space to make a hypercube. Each arrangement is called a NET.
hypotenuse The longest side of a triangle that contains a 90° angle, that is, a right triangle, is called the hypotenuse of the triangle. PYTHAGORAS’S THEOREM shows that this longest side lies opposite the right angle. (If the three sidelengths of a right triangle are a, b, and c, with c opposite the right angle, then c2 = a2 + b2. This establishes that c is indeed larger than both a and b.) The two remaining sides of a right triangle are called the legs of the triangle. The word hypotenuse is derived from the Greek term hypoteinousa meaning “under tension” (the prefix hypomeans “under” and teinein means “to stretch”). Mathematicians have shown that an integer c can be the length of the hypotenuse of a right triangle with integer sidelengths if, and only if, in the prime FACTORIZATION of c, no PRIME that appears an odd number of times is 3 more than a multiple of 4. See also PYTHAGOREAN TRIPLES. hypothesis testing See STATISTICS: INFERENTIAL.
I identity An equation that states that two mathemati
For example, the number zero is an identity element for the set of real numbers under the operation of addition: we have that a + 0 = a = 0 + a for all numbers a. With respect to multiplication, 1 is an identity element, since a × 1 = a = 1 × a for all numbers a. (The number zero is sometimes called an “additive identity” and the number 1 a “multiplicative identity.”) In matrix theory the IDENTITY MATRIX is an identity element under multiplication for the set of all square matrices of a particular size. In SET THEORY, the EMPTY SET is an identity under the operation of taking union. Any GROUP is required to have an identity element. It is impossible for a mathematical system to possess two different identity elements with respect to the same binary operation. If, for instance, e and f are both identity elements for a set S, then e*f equals e (since f is an identity element), and it also equals f (since e is an identity element). Thus e and f are the same. In some mathematical theories it is desirable to distinguish between a “left identity,” that is an element eL such that eL*a = a for all elements a, and a “right identity” eR that satisfies a*eR = a for all a. Of course, if the theory satisfies the COMMUTATIVE PROPERTY, then eL and eR are equal and represent an identity for the system. See also INVERSE ELEMENT.
cal expressions are equal for all possible values of any variables that occur in them is called an identity. For example, the DIFFERENCE OF TWO SQUARES formula x2 – y2 = (x – y)(x + y) and the trigonometric equation sin2θ + cos2θ = 1 are identities. Sometimes the symbol ≡ is used instead of = to indicate that the statement is an identity. It is read as “identically equal to.” A numerical statement, one involving no variables at all, might still be called an identity if it illuminates something interesting about the numbers involved. For example, the following relationship could be called an identity because of the unexpected appearance of the number π: 1 1 1 π 3 arctan + arctan + arctan = 4 20 1985 4 See also EQUATION; IDENTITY ELEMENT.
identity element (identity, neutral element) An element of a set that, combined with any other element of the set, leaves that element unchanged is called an identity element. More precisely, given a BINARY OPERATION “*” on a set S, we say that e in S is an identity element if:
identity matrix (unit matrix) A square MATRIX with
a*e = a e*a = a
each main diagonal element equal to 1 and all other entries equal to zero is called an identity matrix. Such a matrix is usually denoted I or Id, or even Idn if the
for all elements a of the set. 257
258 image ORDER OF A MATRIX
is n. For example, the 3 × 3 iden
tity matrix is: 1 0 0 I3 = 0 1 0 0 0 1 If A is a matrix with m rows and n columns, then matrix multiplication shows we have: AIn = A ImA = A that is, multiplication with the appropriate identity matrix leaves any other matrix unchanged. If all the matrices under consideration have the same number of rows as columns, say n of each, then In acts as an IDENTITY ELEMENT for that set under matrix multiplication. Any 2 × 2 matrix A of the form: 1 − d2 d c −d c or d 1 − d2 c
c −d
satisfies A2 = I2. Thus any such matrix can be thought of as a SQUARE ROOT of the 2 × 2 identity matrix. If one is willing to permit the use of COMPLEX NUMBERS, then the following matrix is an example of a cube root of the 3 × 3 identity matrix: 0 0 −i A = i 0 0 0 1 0
implication See CONDITIONAL. implicit differentiation When two variables x and y satisfy a single equation F(x,y) = 0, it may be possible to regard the equation nonetheless as defining y as a function of x, even though no explicit formula of this type may be apparent. (See IMPLICIT FUNCTION.) In such a case, one can go further and differentiate the equation as a whole, regarding y as a function of x and using the CHAIN RULE in the process to find a formula dy for the derivative –– dx . This process is known as implicit differentiation. For example, if xy3 + 7x2y = 1, then differentiating yields: y3 + x3y 2
dy dy + 14xy + 7 x 2 =0 dx dx
and so: dy y3 + 14xy =− dx 3xy 2 + 7 x 2 (assuming the denominator is not zero). Implicit differentiation is useful for finding the derivatives of inverse functions, for example. For instance, if y = sin–1(x), then sin(y) = x. Differentiating dy 1 yields cos(y) –– = 1, and so, dy = 1 = dx dx cos(y) 1 − sin 2 (y)
This matrix satisfies the relation A3 = I3.
image (range) The set of all values that a
all REAL NUMBERS is the set of all numbers greater than or equal to zero. The term is also used for the output of a specific input for the function. For instance, in the example above, the image of the number 3 is 9. In dealing with functions of real numbers, the term range is preferred over image. Often mathematicians will use the word image only when thinking of a problem geometrically. For example, if the function is a GEOMETRIC TRANSFORMATION such as a reflection in a line, then one would speak of the image of geometric figures under this transformation. In this example, the image of any circle is another circle, and the image of a straight line is another straight line.
=
FUNCTION
could adopt is called the image of the function. For example, the image of the function y = x2 defined for
1
. 1 − x2 Sometimes a function is more easily differentiated if one first applies a logarithm and then differentiates
improper integral 259 implicitly. This process is called logarithmic differentiation. For example, to differentiate y = xx, write ln(y) = dy 1 1 ln(x)x = xln(x). Then – ⋅ –– = ln(x) + x ⋅ – , and so, y x dx dy = y(ln(x) + 1) = x x (ln(x) + 1) . dx
implicit function A formula of the form F(x,y) = 0 is said to define the variable y implicitly as a function of x. For example, the equation xy = 1 implicitly defines 1 . However, matters the dependent variable y as y = – x can be confusing, for a single equation may define y as several possible functions of x (for example, the equation x2 + y2 = 1 suggests that either y = √1 – x2 or y = – √1 – x2 ), and in some instances, it might not even be possible to solve for an explicit formula for y (for example, it is not at all clear that x2y2 + ycos(xy) = 1 does indeed give a formula for y). Mathematicians have proved that as long as the function F(x,y) has continuous PARTIAL DERIVATIVEs, and that there is some point (a,b) for which F(a,b) = 0 ∂F and –– ∂y ≠ 0 at (a,b), then the equation F(x,y) = 0 does indeed define y as a function of x, at least for values of x within a small neighborhood of x = a. Moreover, the process of IMPLICIT DIFFERENTIATION is valid for these values of x. This result is known as the implicit function theorem.
bounded on the interval [L,1] for any 0 < L < 1, and the integral
1
1
∫L
dx is valid. We have:
1
∫L
1
1 1
dx = 2x 2
x x L – = 2 – 2√ L. The improper integral under consideration is then defined to be the LIMIT as L → 0, from 1 1 1 1 dx = lim L→0 + ∫ dx = above, of these values: ∫0 L x x lim L→0 + 2 − 2 L = 2. In general, if an integrand f(x) is infinite at an end b point a, then the improper integral ∫ a f(x)dx is defined as the limit: b
b
∫ a f(x) dx = limL→a+ ∫ L f(x)dx or, if infinite at the end point b, then the improper integral is defined as the limit: b
L
∫ a f(x)dx = limL→b– ∫ a f(x)dx If the limit exists, then the improper integral is said to 1 1 dx is a convergent converge. (Thus, for example, ∫0 x improper integral.) If the limit does not exist, then the improper integral is said to diverge. (One can check, 11 for example, that ∫0 dx diverges.) x To illustrate the second type of improper integral, ∞ 1 consider, for example, the quantity ∫1 2 dx . This is to x be interpreted as the limit:
improper integral (infinite integral, unrestricted integral) In the theory of INTEGRAL CALCULUS it is permissible to extend the notion of a definite integral b ∫ a f(x)dx to include functions f(x) that become infinite in the range under consideration, [a,b], or, alternatively, to consider integration over an infinite range: [a,∞), (– ∞,b], or even (–∞,∞). Such integrals are called improper integrals. One deals with such integrals by restricting the range of integration and taking the limit as the interval expands to the required size. 1 1 dx . The Consider, for example, the integral ∫0 x 1 integrand becomes unbounded as x approaches x the value zero, and so, in the normal way, the integral 1 1 ∫0 x dx is not well defined. However, the function is
∞
∫1 which
can
lim L→∞ ∫
L
1
L 1 1 dx = lim L→∞ ∫ 2 dx 2 1 x x
be
readily
computed:
∞
∫1
1 dx = x2
1 1 L 1 dx = lim − = lim L→∞ 1 − = 1 L →∞ 2 x L . (We x 1
interpret this as saying that the total area under the 1 curve y = – x2 to the right of x = 1 is 1.) In general, improper integrals of this type are computed as follows: ∞
L
∫ a f(x)dx = limL→∞ ∫ a f(x)dx b
b
∫ –∞ f(x)dx = limL→–∞ ∫ L f(x)dx ∞
L
∫ –∞ f(x)dx = limL→∞ ∫ –L f(x)dx
260 Incan mathematics If the limits exist, then the improper integrals are said to ∞ 1 converge. (Thus, for example, ∫1 2 dx is a convergent x improper integral.) If the limit does not exist, then the improper integral is said to diverge. (One can check, for ∞ 1 dx diverges.) The integral test used example, that ∫ 1 x in the study of CONVERGENT SERIES makes use of improper integrals of this type. It is possible to define even moregeneral improper integrals by considering integrands that take the value infinity at several places within the range of integration, possibly over an infinite range of integration. One interprets such improper integrals as the sum of several limits.
Incan mathematics The Incan Empire ruled the region of Peru and its surroundings before the Spanish conquest of the continent in 1532. It was an extraordinarily sophisticated civilization. The Inca people had highly developed systems of agriculture, transportation, road construction, textile production, and administration, yet, surprisingly, had no form of writing. Consequently, with absolutely no cultural records or documents available, very little is known of the mathematics of these people. It is known that the Inca recorded numerical information with a device known as a quipu, which consisted of a collection of strings all tied to a common central object, usually another rope. Administrators simply tied knots in the strings to record numbers, with each string representing a different count. Knots were grouped in clusters to represent quantities of powers of 10. Thus, for example, the number 283 was represented (backward) by a cluster of three knots placed near the free end of a string (three units), a space, a cluster of eight knots in the center of the string (eight 10s), another space, and finally two knots at the top of the string (two 100s). The Incas, in effect, were following the DECIMAL REPRESENTATION system we use today. A large space indicated a digit of zero (to distinguish 208 from 28, for instance). Strings of different colors were used to signify what the knots on that string were counting. For example, white strings may have represented the counts of cattle, while green strings may have represented the levels of punishment to be executed for committing a certain crime. Each town in the Incan Empire had its own quipcamayoc, a “keeper of the knots,” who recorded census
information about the population, the produce, and weapons of the town. Quipues recording this information were sent every year, by relay runners, to the capital of the empire, Cuzco. The extent to which the Incan people manipulated numbers and performed calculations is unclear. A Spanish priest by the name of José de Acosta, who lived among the Incas from 1571 to 1586, describes in his book, Historia natural y moral de las Indias, residents shuffling kernels of corn on the ground as though manipulating the beads of an ABACUS. Obviously some kind of numerical computation was being conducted, but Acosta did not understand the method and was unable to explain what exactly was being done. The Inca thought of numbers in very concrete terms. The concept of “twoness,” for instance, did not exist for the Inca, and the count of 2 applied to very specific things. This is evidenced by the fact that the Inca had several different words for two, which were used to distinguish different contexts—the two objects that make a pair, the 2 that came from one object divided in half, and the one object and its specific partner that make a pair, for instance. There was no single allencompassing word for the state of having two parts. It is likely that the Inca people kept astronomical records, a topic of interest to most ancient cultures, but there is no direct evidence of this.
incircle A circle that lies inside a triangle and touches all three sides is called the incircle of that triangle. The center of that circle is called the incenter of the triangle, and its radius is called the inradius of the triangle. It is relatively straightforward to see that for any triangle there is precisely one circle in the triangle that touches all three sides. (Erase one side of the triangle and note that there are infinitely many circles that fit into the wedge formed by the two remaining sides. In redrawing the third side, observe that precisely one of those circles lies tangent to that additional edge.) The incenter of a triangle is a point that lies the same distance from each of the three sides of the triangle. Note that any point that is EQUIDISTANT from just two intersecting lines lies on the angle bisector of those lines (that is, on a line that cuts the angle between the two given lines in half). Consequently, we have that the incenter of any triangle lies on each of the angle bisectors of the vertices of the triangle. This establishes:
increasing/decreasing 261 The three angle bisectors of a triangle are CONCURRENT.
compensate for this double count.) Similarly, the number of elements in the union of three sets is given by:
An excircle for a triangle is a circle that lies outside the triangle and is tangent to one side of the triangle and tangent to extensions of the remaining two sides. A triangle has three distinct excircles. In a more general context, if it is possible to draw a circle inside a POLYGON tangent to each side of the polygon, then that circle is called an incircle of the polygon. Every regular polygon has an incircle. There is no incircle for a nonsquare rectangle. See also CIRCUMCIRCLE.
n(A ∪ B ∪ C) = n(A) + n(B) + n(C) –n(A ∩ B) – n(B ∩ C) – n(A ∩ C + n(A ∩ B ∩ C)
inclination/declination The angle θ between a ray
inclusionexclusion principle If n(A) denotes the number of elements in a finite set A, then the number of elements in the union A ∪ B of two sets is given by: n(A ∪ B) = n(A) + n(B) – n(A ∩ B) (Counting the number of elements in each of A and B counts the elements that belong to both twice. One must
n((A ∪ B) ∪ C) = n(A ∪ B) + n(C) – n((A ∪ B) ∩ C) = n(A ∪ B) + n(C) – n((A ∩ C) ∪ (B ∩ C)) Two more applications of the formula for the union of two give the result.) In general, an INDUCTION argument shows that the number of elements in the union of n sets is given by: n(A1 ∪ A2 ∪…∪ Ak) = n(A1) + n(A2 +…+ n(Ak) –n(A1 ∩ A2) – n(A1 ∩ A3) –… – n(Ak–1 ∩ Ak) + n(A1 ∩ A2 ∩ A3) +… + n(Ak–2 ∩ Ak–1 ∩ Ak …
emanating from the origin of a CARTESIAN COORDINATE system and the positive xaxis as measured in an anticlockwise direction is called the inclination of the ray. An angle measured in the clockwise direction is called its declination. These terms are rarely used in mathematics today. A PLANE that is not horizontal is called an inclined plane, and the angle that the line of greatest slope within the plane makes with the horizontal is called the angle of inclination of the plane.
(One can establish this either by reasoning through which elements are counted multiple times, or by noting that A ∪ B ∪ C = (A ∪ B) ∪ C and applying the previous observation:
+ (–1)kn(A1 ∩ A2 ∩…∩ Ak) This formula is called the general inclusionexclusion principle. It can be interpreted as follows: The number of elements of a finite set that possess at least one of k possible properties is equal to the number possessing exactly one property, minus the number possessing exactly two properties, plus the number possessing precisely three properties, and so on, up to the count of those elements possessing all k properties. This powerful counting principle has important applications in PROBABILITY theory.
increasing/decreasing A
Angles of inclination and declination
of numbers {an} is said to be increasing if a1 ≤ a2 ≤ a3 ≤ …, that is, each term in the sequence is greater than or equal to the one that precedes it. It is called strictly increasing if a1 < a2 < a3 < … The constant sequence 1,1,1, …, for example, is considered increasing. SEQUENCE
262 increment Increasing sequences have the pleasing property that their convergence (or divergence) is easy to identify: An increasing sequence converges if, and only if, it is bounded above. (See BOUND and CONVERGENT SEQUENCE.) This seems clear geometrically. If on the real number line the sequence of numbers {an} moves steadily to the right yet does not penetrate a barrier at position B, then the numbers must “pile up” at some limit value L ≤ B. Conversely, if an increasing sequence converges to a value L, then all the terms of the sequence lie to the left of L on the number line and hence are bounded above by L. (These statements can be proved rigorously by making use of the “completeness” of the real number line via the notion of a DEDEKIND CUT.) A sequence {an} is said to be decreasing if a1 ≥ a2 ≥ a3 ≥ …, that is, each term in the sequence is less than or equal to the one that precedes it. It is called strictly decreasing if a1 > a2 > a3 >…. Similarly one can show that a decreasing sequence converges if, and only if, it is bounded below. A sequence that is either increasing or decreasing is called monotonic. A realvalued function f is said to be increasing over an interval if, over that interval, greater input values produce greater (or possibly equal) output values, that is, if a and b are two values in the interval with a < b, then f(a) ≤ f(b). If it is always the case that f(a) < f(b), then the function is said to be strictly increasing on the interval. The MEANVALUE THEOREM shows that a differentiable function f is increasing precisely on those intervals where the DERIVATIVE is nonnegative, f ′ (x) ≥ 0, and strictly increasing if f ′ (x) > 0. A realvalued function f is said to be decreasing over an interval if greater input values produces smaller (or possibly equal) output values of the function, that is, if a < b, then f(a) ≥ f(b). If it is always the case that f(a) > f(b), then the function is said to be strictly decreasing on the interval. The MEANVALUE THEOREM shows that a differentiable function f is decreasing precisely on those intervals for which f ′ (x) ≤ 0, and strictly decreasing if f ′ (x) < 0. A function that is either increasing or decreasing on an interval is called monotonic on that interval. See also DIFFERENTIAL CALCULUS; GRAPH OF A FUNCTION; MAXIMUM/MINIMUM.
increment A small finite change in the value of a variable or in the value of a function is called an increment. If the variable is x, then an increment of x is usually denoted ∆x or δx. If f is a function of x, then the corresponding increment of f, denoted ∆f, equals ∆f = f(x + ∆x) – f(x). The LIMIT of the ratio ∆f/∆x, as the increment ∆x decreases to zero, is called the DERIVATIVE of f. A negative increment is sometimes called a decrement. See also DIFFERENTIAL.
indefinite integral See
ANTIDIFFERENTIATION; INTE
GRAL CALCULUS.
independent axiom An AXIOM of a mathematical theory is said to be independent if it cannot be derived from the remaining axioms of a theory as a logical consequence. For instance, the existence of NONEUCLIDEAN GEOMETRIES establishes that Euclid’s famous PARALLEL POSTULATE is independent of his four remaining postulates. In general, this is the approach mathematicians take to show that an axiom A in a mathematical system is independent of the remaining axioms: present an example of another consistent mathematical theory proved to be free of contradictions in which all axioms except A hold and for which A is false. (Consequently, it cannot be the case that these axioms do imply that A is true, for then one has a system in which both A and not A hold.) See also EUCLID’S POSTULATES. independent events Two experiments run in succession are deemed independent if the outcomes obtained from one experiment do not affect the outcomes obtained in the second. For example, the results from casting a die do not influence the results obtained in later tossing a coin. These two actions are independent. In PROBABILITY theory we say two EVENTs A and B are independent if knowledge of A having occurred has no influence on the likelihood of B next occurring. For example, in two rolls of die casting, an even number on the first roll and casting a 6 on the second are independent events—the chances of rolling a six are 1/6, no matter the result of the first roll. Two events that are not independent are called dependent. For example, the event “Sally is wearing a
Indian mathematics 263 raincoat” is likely to be dependent on the event “it is raining.” CONDITIONAL PROBABILITY is used to analyze dependent events. Experiments that involve drawing objects from a finite source, such as balls from a bag, or cards from a deck, without replacing the objects withdrawn, often yield dependent events. For example, the events “the first card drawn from a deck of cards is red” and “the second card drawn is red” are dependent—the probability that the second card selected is the desired color could either be 26/51 or 25/51 , depending on whether the first card drawn is black or red. A careful study of PROBABILITY shows that two independent events A and B satisfy the probability relation: P(A and B) = P(A) × P(B) Mathematicians take this relation as the definition of two events A and B being independent.
indeterminate equation An equation, or a
SYSTEM
with an infinite number of solutions is called indeterminate. For instance, the equation:
example, one might say that the function f(x) = x5 has index 5 to indicate that quantities are being raised to 3– the fifth power. The radical function g(x) = √x, for example, has index 3, and the logarithmic function h(x) = log10 x has index 10. In statistics and business, an index is a figure used to show the variation in some quantity over a period of time, usually standardized relative to some base value. For example, a retail price index is used to measure changes in the cost of household items. For the base year, the index is usually set at 100. If, for instance, the cost of dish detergent in the base year is $1.60, and the price rises to $1.76 and to $2.00 in the two subsequent years, then the price index for this item would be 100, 110, and 125 in each of those years, respectively. It is possible to design an index that gives a general measure of the value of more than one commodity simultaneously. For example, the Dow Jones index of the New York Stock Exchange is an indicator of the worth of a representative selection of industrial, transportation, and utility stocks.
OF EQUATIONS,
x + 2y = 5 is indeterminate (for each real number t there is the solution x = 5 – 2t and y = t), as is the system of two equations: a+b+c=1 2a + 3b – c = 0 in variables a, b, and c. (All triples of the form (a,b,c) = 1 – 4t 2 – t —— (t, ——– 3 , 3 ) are solutions.) The single equation x2 + y2 + z2 + w2 = 0 is not indeterminate: it has the unique (real) solution x = y = z = w = 0. See also DEGREES OF FREEDOM.
index (plural, indices) A number that indicates a place or describes the characteristic or the nature of a function is called an index. For example, a SEQUENCE of numbers might be denoted {an}, with the index n denoting the place in the sequence. (Thus a7, for instance, would represent the seventh number in the list.) An EXPONENT is sometimes called an index. For
Indian mathematics The entire course of Western mathematics was profoundly affected by a single Indian invention, that of the placevalue decimal system. That every possible number can be expressed via a set of just 10 symbols, 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9, by making careful use of the place of each symbol, seems such a simple notion nowadays that it is hard to appreciate its profound importance. This elegant notation system provided the means for Indian scholars to perform complicated arithmetical computations with relative ease, which in turn led to significant developments in numerical techniques, approximation methods, and the theory of arithmetic. Only when other cultures adopted the placevalue decimal system from India could they accomplish the same mathematical feats that this culture had already developed. The earliest dated evidence of mathematical activity in the Indian subcontinent goes back to the Indus civilization of 2500 B.C.E. Bronze weights and graded rods (rulers) from the period show that these people were already working with a decimal system. The Indus people worked with a basic unit of length 1.32 in. long (today called the “Indus inch”), 10 of which make their version of a “foot.” Excavations show that the weights and graded rods were used extensively in construction.
264 Indian mathematics The earliest written records of Indian culture are the religious texts the Vedas, dating between 1500 B.C.E. and 800 B.C.E. Although not mathematical in content, appendices to the texts give specific rules for constructing altars, exhibiting a thorough understanding of the basic principles of geometry. Early versions of the digits 0 through 9 were used at this time. By 600 C.E., the Vedic religion had gone into decline, and Jainism came to the fore. During this period mathematics was driven by the needs of the religion and its demands for careful astronomical observations to pinpoint the exact times of religious observances and the development of an accurate calendar. The decimal representation system was now fully developed, and scholars were able to make precise and surprisingly accurate calculations. The math– ematician ARYABHATA (ca. 500 C.E.), for instance, had developed a theory of TRIGONOMETRY to aid astronomical calculations, had developed methods for extracting square roots, evaluated π to a high degree of accuracy, and was able to find integer solutions to a large class of equations that arose from astronomical theories. One written text from this period was discovered in 1881 in the town of Bakhshali, now in Pakistan. Written on birch bark, the Bakhshali manuscript shows that mathematicians were also comfortable with fractions, basic algebraic manipulations (they used a dot to represent an unknown quantity), and sophisticated approximation formulae. For example, the manuscript describes, in words, the following approximation for the square root of a number N: 2
b 2a b N = a2 + b ≈ a + − 2a b 2 a + 2a Here a2 is the largest perfect square smaller than N and b = N – a2. For example, we have:
50 = 7 2 + 1 ≈ 7 +
1 − 14
1 14
2
1 2(7 + ) 14
≈ 7.071067821
which is correct to seven decimal places.
Two mathematical research centers were formed in India during the era of Jainism, both astronomical – observatories. Aryabhata headed the first center at Kusumapura in the northeast of the Indian subcontinent, and mathematician Varahamihira, who also made contributions to astronomy and trigonometry, headed the second center at Ujjain, also in the north. Varahamihira was succeeded by the seventhcentury mathematician BRAHMAGUPTA, who, in his famous work Brahmasphutasiddhanta (The opening of the universe), introduced and explained the arithmetic of nonpositive numbers. He was the first mathematician in history to give zero the status of a number, defining it to be the result of subtracting a quantity from itself. Brahmagupta’s work also includes a formula for the area of a cyclic quadrilateral in terms of its sides (today called BRAHMAGUPTA’S FORMULA), and presents methods for solving linear and quadratic equations, as well as systems of equations. Brahmagupta also developed sophisticated interpolation techniques for computing sine values in trigonometry. For the next 200 years, Indian scholars worked to refine further methods of trigonometry and techniques of astronomical calculation. The mathematician BH –ASKARA II of the 12th century made advances in number theory, algebra, combinatorics, and astronomy, and wrote a comprehensive text summarizing the state of mathematics and astronomy in India at his time. Soon afterward, other Indian scholars developed these ideas further. Jaina mathematicians also clarified the standard EXPONENT rules and manipulated exponents in a manner that suggests today that they were also familiar with the basic principles of LOGARITHMs. The 14thcentury scholar MADHAVA OF SANGAMAGRAMMA made significant advances in ANALYSIS. He produced the infinite series expansions of trigonometric and inverse trigonometric functions (today called TAYLOR SERIES), discovered the BINOMIAL THEOREM, and even produced GREGORY’S SERIES for π, which he used to approximate its value to a considerably accurate degree. During the first millennium India had very little contact with the cultures of the West. News of the decimal representation system, however, did manage to spread to other countries relatively quickly. A manuscript written in Syria in 662 discusses the new method of calculation, and there is evidence that the decimal system was being used in Cambodia and other Asian countries soon afterward. By the ninth century,
induction 265 the decimal system was in common use in the Islamic world, and from there it was quickly transmitted to Europe. Arab scholars maintained a keen interest in the work of Indian mathematicians for the centuries that followed and took an active role in preserving and translating many Indian texts. See also ARABIC MATHEMATICS; BASE OF A NUMBER SYSTEM.
indirect proof (proof by contradiction, reductio ad absurdum) Most claims made in mathematics are statements of the form: If the premise A is true, then the conclusion B is true. An indirect proof of such a statement attempts to establish the validity of the claim by assuming that the premise A is true and showing, consequently, that the conclusion B cannot be false. One does this by exploring the logical consequences of assuming A and “not B” until, at some point, a contradiction (such as, 1 + 1 = 3, for instance) is reached. Based on the belief that mathematics should be free from absurdities, mathematicians generally accept this approach as sufficient for establishing the validity of B. As an example, we prove: for a natural number n, if n2 is even, then n is even. We base the proof on known facts about even and odd numbers, and the standard algebraic manipulations. Proof: assume that n2 is even, and assume, to the contrary, that n is not. Then it must be the case that n is odd. Consequently, n is one more than a multiple of 2 and can be written in the form 2k + 1, for some number k. This means that n2 is given by n2 = (2k + 2 1) = 4k2 + 4k + 1 = 2(2k2 + 2k) + 1 and so is also one more than a multiple of 2. We conclude that n2 is both even and odd—clearly a contradiction. It cannot be the case, then, that n is odd. It must therefore be even. Notice in this proof that we assumed premise A to be true, and arrived at the contradiction that “not A” also holds. An indirect proof that arrives at a contradiction of this type is usually called a “proof using the
contrapositive”: we established the validity of “A implies B” by demonstrating that the CONTRAPOSITIVE “not B implies not A” holds. EUCLID’S PROOF OF THE INFINITUDE OF PRIMES is an example of an indirect proof that does not rely on the contrapositive. EUCLID (300–260 B.C.E.) was the first mathematician to extensively employ the technique of indirect proof. A DIRECT PROOF attempts to establish the validity of a proposition “if A, then B” by assuming that the premise A is true and following its logical consequences until statement B is established. Not all propositions, however, are amenable to a direct approach. For example, given that a squared number n2 is even (that is, n2 = 2m, say) it is not immediate how one should proceed to establish directly that n is consequently even. Any indirect proof relies on the assumption that a statement that cannot be false must be true. Some philosophers and mathematicians who study the LAWS OF THOUGHT seriously question this assumption. See also CONTRADICTION; DEDUCTIVE/INDUCTIVE REASONING; PROOF; QED; THEOREM.
induction (complete induction, finite induction, mathematical induction) The method of proof known as mathematical induction has been used by scholars since the earliest times, but it was not until 1838, thanks to the work of English logician and mathematician AUGUSTUS DE MORGAN, that the principle was properly identified and described. In a formal context, the principle of mathematical induction asserts: If a set S of numbers satisfies the following two properties: i. The number 1 belongs to S. ii. If a number k belongs to S, then so does its successor k + 1. then it must be the case that S contains all the natural numbers 1, 2, 3,… (This principle appears as an AXIOM in GIUSEPPE PEANO’s set of postulates for the construction of the natural numbers.) Often the set S is taken to be a set of natural numbers n for which some property or formula P(n) is true: S = {n: P(n) holds}. To illustrate the principle of induction, we shall prove that, for every natural number n, we have that 1 1 + 2 + 3 +…+ n equals – n(n + 1). Let P(n) represent 2
266 induction
1. The number 1 belongs to S.
ing formulae and equations involving the natural numbers. One disadvantage, however, is that this method of proof gives no indication as to what the appropriate formula to be proved should be: the inductive method of proof offers no insight, for instance, as to why the formula for the sum of the first n counting numbers is 1 – n(n + 1). One must arrive at candidates for formulae 2 via other means. The inductive process is often likened to knocking down a chain of dominoes. To successfully topple a row of standing dominoes one must:
Entering n = 1 into the formula yields the patently 1 true statement 1 = – × 1 × (1 + 1). Thus P(1) is true 2 and 1 ∈ S.
1. Knock down the first domino 2. Be certain that the dominoes are appropriately spaced so that when any one domino falls, it is sure to knock down the next
2. Assume that the number k belongs to S. That is, assume that the formula
If one can establish that these two properties hold, then all dominoes in a chain (even an infinitely long one) will fall. Despite the elegance and simplicity of the principle, one must still apply care when utilizing the principle of mathematical induction. The following amusing argument, for instance, illustrates what can go wrong:
this assertion and set S to be the set of all natural numbers for which this assertion is true: 1 P(n): 1 + 2 + 3 +…+ n = – n(n + 1) 2 S = {n: the formula P(n) is true} Our aim is to show that all natural numbers belong to S (that is, that P(n) is true for all natural numbers n). We must establish two things:
1 1 + 2 +…+k = – k(k + 1) 2 is true. Does it necessarily follow that the corresponding formula for k + 1 is valid? That is, can we then deduce, under this assumption, that 1 1 + 2 +…+ k + (k + 1) = – (k + 1)((k + 1) + 1) 2 will hold? To achieve this, start with the equation P(k), assumed to be valid, and add the quantity k + 1 to both sides. This gives: 1 1 + 2 +…+ k + (k + 1) = – k(k + 1) + (k + 1) 2 After rearranging and regrouping terms on the right, this reads: 1 1 + 2 +…+ k + (k + 1) = – (k + 1)(k + 2) 2 which is the statement P(k + 1). By the principle of mathematical induction, it now follows that the set S does indeed contain all the natural numbers, that is, P(n) is a valid formula for all values of n. The principle of mathematical induction is a powerful technique that can establish the validity of an infinite number of statements in one fell swoop. As illustrated above, it is particularly useful for establish
Claim: all horses are the same color. We “prove” this as follows: Let S be the set of all natural numbers for which the following statement is true. P(n): if n horses stand in a field, then all horses in that field are the same color. Clearly P(1) is true: if only one horse is standing in a field, then all horses in that field are the same color. Now make the assumption, despite its absurdity, that P(k) is true, that is, any k horses in a field must be the same color. Now consider k+1 horses in a field. Remove one horse, Chester, say. This leaves k horses in the field, which, by our assumption, must all be the same color. Return Chester to the field and remove a different horse. This again leaves k horses in the field, which, by assumption, must all be the same color. This shows that Chester is the same color as the first k horses, and in fact that all k+1 horses are the same color. We have, from P(k), established that P(k+1) follows. By the principle of mathematical induction, it now follows that P(n) is true, no matter the value of n. In particular, all the horses in the world are the same color.
inequality 267 To examine this argument, note that it is indeed the case that P(1) is valid, and it is true that P(k) implies P(k+1)—for almost all values of k. The fault in the argument above is that although P(2) establishes P(3), and P(3) establishes P(4), and so forth, it is not the case that P(1) establishes P(2): removing one horse from a field containing just two horses does not, alas, establish that both horses are the same color. In our chain of dominoes, all but the first two dominoes are properly spaced: the first domino does not topple the second domino during its fall, and the chain remains standing. This illuminates that one must always be careful that any argument presented in a proof by induction is indeed valid for all values of k. See also PROOF; SUMS OF POWERS.
inductive
reasoning See
DEDUCTIVE/INDUCTIVE
REASONING.
inequality A mathematical statement that one quantity or expression is greater than or less than another is called an inequality. The following symbols are used: a > b, a is greater than b a < b, a is less than b a ≥ b, a is greater than or equal to b a ≤ b, a is less than or equal to b Inequalities satisfy a number of ORDER PROPERTIES. An inequality is called closed or unconditional if it holds for all values of the variables, if any, that appear in the equation. For instance, the inequalities 3 ≤ 5, x2 + 1 > x2, and 5 + y2 > 4y are closed inequalities for they are always true. Inequalities that are not closed are called open or conditional. The set of values of the variables that appear in the inequality that make the statement true is called the solution set of the inequality. For example, the open inequality 2x + 1 > 7 has as a solution set the set of all real numbers x for which x > 3. The solution set of the open inequality a2 + b2 < 0 is the EMPTY SET. Open inequalities can be solved in much the same manner as equations. As the order properties show, one can add or subtract the same quantity to both sides of the inequality and preserve the inequality, or can multiply the inequality through by a positive quantity. Multiplying through by a negative quantity changes the sense
of the inequality. (For example, if a < b, then subtracting the quantity a + b from both sides yields –b < –a. This shows that the effect of multiplying through by –1 is to reverse the sense of the inequality.) For example, one can solve the inequality 2x + 1 > 7, as follows: 2x + 1 > 7 2x > 6 1 1 – × 2x > – 2 2 ×6 x>3 indeed yielding the solution set {x:x > 3}. It is worth noting that if a · b > 0, then we can be sure that a and b are either both positive or both negative. If, on the other hand, a · b < 0, then we can be sure that a and b have opposite signs. These observations are essential for solving inequalities involving a single variable raised to the second power. For example, to solve the open inequality x2 + x – 2 > 0, factor the QUADRATIC to obtain (x + 1)(x – 2) > 0 and examine the two possible scenarios. Either x + 1 and x – 2 are both greater than zero (yielding that x must be greater than 2), or x + 1 and x – 2 are both less than zero (yielding that x must be less than –1). Thus the solution set to the inequality is the set of all real numbers x with x > 2 or x < –1. A single inequality in two variables defines a region in the plane. For example, the inequality 2x + y ≥ 3 is satisfied by the points (0,3), (5,0), and (2,2), for instance. In this example, the complete solution set is the closed HALFPLANE sitting above the line y = –2x + 3 in the plane. There are a number of standard inequalities in mathematics: Triangle inequality: For any triangle with sidelengths a,b, and c we have: a + b > c. Arithmeticgeometric mean inequality: For nonnegative numbers a1,a2,…,an we have: n
a1a2 ⋅ L ⋅ an ≤
a1 + a2 + L + an n
Bernoulli’s inequality: For any real number x greater than 1 and positive integer n: (1 + x) n > 1 + nx
268 inference (This can be proved by truncating the MACLAURIN SERIES: 1 (1 + x)n = 1 + nx + – n(n – 1)x2 2 1 + – n(n – 1)(n – 2)x3 + …, 6 One can also show that the inequality is also true for –1 < x < 0). CauchySchwarz inequality: For any two VECTORS a and b, their DOT PRODUCT satisfies: a · b ≤ a × b Thus, for any two lists of numbers a1,a2,…,an and b1,b2,…,bn we have: 2
n n 2 n 2 a b ≤ ∑ k k ∑ ak ∑ bk k=1 k=1 k=1 Weierstrass’s product inequality: For any list of numbers a1,a2,…,an with 0 ≤ ak ≤ 1 for all k, we have: (1 – a1)(1 – a2)…(1 – an) ≥ 1 + a1 + a2+…+an (This can be proved by INDUCTION on the number of elements in the list.) Napier’s inequality: For any two positive real numbers a and b we have: 1 ln b − ln a 1 < < b b−a a Exponential inequalities: For a positive real number x and a real number c, we have: c
x < 1 + c(x – 1) if 0 < c < 1 xc > 1 + c(x – 1) if c > 1 (This is a generalization of Bernoulli’s inequality.) Isoperimetric inequality: For any closed geometric figure in the plane with perimeter P, its area A is less than the area of a circle of the same perimeter: 2
P P2 A ≤ π = 2π 4π Equality holds if, and only if, that figure is a circle.
A mathematical statement that two quantities or expressions are never equal is called an inequation. For instance, the statement 3 ≠ 5 is an inequation. Mathematicians and physicists often write a >> b if a is significantly larger than b and a > 0. There is a joke among mathematicians to use the symbol ≤≥ to mean “less than, greater than, or possibly equal to” when one is not sure of the numerical answer to a problem. See also ARITHMETICGEOMETRIC MEAN INEQUALITY; ISOPERIMETRIC PROBLEM; TRIANGLE INEQUALITY.
inference In logic, the general process of developing an ARGUMENT to draw a conclusion from a set of premises is called inference. The process could be deductive or inductive. In statistics, inference is the process of coming to a conclusion about a population based on a study of a sample. Sometimes the conclusion itself is called an inference. Inferential statistics is the science of making inferences and predictions about a population based on numerical information gathered from a sample. See also DEDUCTIVE/INDUCTIVE REASONING; POPULATION AND SAMPLE; STATISTICS: INFERENTIAL.
infinite product The product of an infinite number of factors, a1 × a2 × a3 ×…, is called an infinite product. The nth number in the product is called the nth term of the product, and the product of the first n terms, Pn = a1 × a2 ×…× an is called the nth partial product. In 1812 CARL FRIEDRICH GAUSS introduced the notation ∞
∏ ai
n
for an infinite product (and
∏ ai
for the nth
i =1
i=1
partial product). An infinite product might have a value of zero (1 × 1 1 1 – × – – 2 3 × 4 ×…, for example), might be infinite in value (1 × 2 × 3 × 4 ×…, for example), or could oscillate in value (1 × (–1) × 1 × (–1) ×…, for instance). Only if the partial products Pn approach a finite nonzero value L as n → ∞ is the infinite product said to converge (to the value L). Otherwise the infinite product diverges. ∞
For example, the infinite product partial product:
1
∏ 1 − i 2 i=2
has nth
infinity 269 1 1 1 1 Pn = 1 − 2 1 − 2 1 − 2 L 1 − 2 4 2 3 n 1 1 1 1 1 1 = 1 − 1 + 1 − 1 + 1 − 1 + 2 2 3 3 4 4 1 1 L 1 − 1 + n n =
1 3 2 4 3 5 n −1 n +1 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅L⋅ ⋅ 2 2 3 3 4 4 n n
infinity In common usage, the word infinite is used to
n +1 = 2n 1 as n → ∞. Thus we can which approaches the value – 2 ∞ 1 1 write ∏ 1 − 2 = . i 2 i=2 By taking the
LOGARITHM,
we see that an infinite ∞
product with positive terms
∏ ai
converges if, and
i=1
∞
only if, the
SERIES
∑ ln ai
converges. (This is why
i=1
mathematicians do not deem convergence to zero as a valid form of convergence for an infinite product: the ∞
value of the sum
∑ ln ai
would then be –∞) Mathe
i=1
∞
maticians have shown that a series
∑ ai
converges
i=1
∞
absolutely exactly if
CAUCHY (1789–1857) introduced the notion of a LIMIT. Cauchy’s work put calculus on a sound theoretical setting and removed the need to ever speak in terms of infinitesimals. The notion is now considered obsolete. (Although mathematicians have recently developed a theory of nonstandard analysis that does, in some sense, include a valid concept of an infinitesimal, after all.) See also FLUXION; HISTORY OF CALCULUS (essay).
∏ (1+  ai ) converges. i =1
In 1656 English mathematician JOHN WALLIS discovered the following remarkable representation for π – 2 as an infinite product:
π 2 2 4 4 6 6 = ⋅ ⋅ ⋅ ⋅ ⋅ ⋅L 2 1 3 3 5 5 7 This expression is today called WALLIS’S PRODUCT. See also ABSOLUTE CONVERGENCE; ZETA FUNCTION.
infinitesimal A positive quantity, supposedly infinitely small yet not itself zero, is called an infinitesimal. The abstract existence of such quantities played an important, yet troublesome, role in the early development of CALCULUS. The theoretical difficulties incurred by them were later made moot when AUGUSTINLOUIS
denote something that is unbounded, limitless, and endless. For example, the set of counting numbers 1, 2, 3, … is unbounded (after any given counting number there is always another), and the set of these numbers is said to be infinite. In geometry, a straight line is usually perceived as without end, and so it is infinite in extent. In the physical world, however, there is no clear example of infinite quantity (or, if there were, we as humans would never be able to fully perceive it). Physicists have come to believe, for instance, that there are only a finite number of atoms in the universe (on the order of 1087 atoms) and that the universe is closed in shape and not of infinite extent. If this is indeed the case, then it would never be possible to draw an infinite line (there are not enough atoms for the ink), and if the universe is indeed bounded, the line may eventually loop back on itself. Nonetheless, the infinite is an abstract concept we feel we can, on some level, comprehend. It does, however, bring with it many paradoxical difficulties, as demonstrated by ZENO’S PARADOXES, HILBERT’S INFINITE HOTEL, and the TRISTRAM SHANDY PARADOX, for example. The notion of something that is unlimited arises in mathematics in a number of varied settings. For instance: 1. Limits: In 1655 English mathematician John Wallis introduced the symbol ∞ as shorthand for the phrase “becoming large and more positive.” This notation is used today, in particular, in the study of LIMITs. For example, the statement limn→∞
10n 2 − 3n + 2 = 5 is 2n 2 + n + 1
10n 2 − 3n + 2 approaches the 2n 2 + n + 1 value 5 as n becomes large and positive.” The equation 2 − x3 2 − x3 limn→∞ = –∞ reads: “the function becomes x x large and negative as x becomes large and positive.” This means that for any negative number –M one cares to be read: “the quantity
270 inflection to choose, one can always find a larger value of x for 2 − x3 which is, and remains, less than –M for ever x larger values of x. 2. Geometry: In DESARGUES’S THEOREM, French mathematician GIRARD DESARGUES (1591–1661) found it convenient to regard parallel lines as intersecting at some point of infinity. Thus in the theory of PROJECTIVE GEOMETRY, the notion of points “infinitely far away” is given meaning and context. 3. Set theory: Italian astronomer and physicist GALILEO GALILEI (1564–1642) observed that every counting number can be matched with its square, showing in some sense that the infinite set of counting numbers is no more infinite than the subset of square numbers: 1
2
3
4
b b b b 1 4
5
6
...
b b
9 16 25 36 ...
In the 19th century, British algebraist AUGUSTUS DE MORGAN (1806–71) and German mathematicians JULIUS WILHELM RICHARD DEDEKIND (1831–1916) and GEORG CANTOR (1845–1918) realized that this is a common property of infinite sets, and that it is appropriate to use this property as the definition of what it means for a set to be infinite: A set is infinite if its elements can be matched, without repetition, with the elements of a proper subset of itself. This definition has the advantage that it makes clear what it means for a set to be finite. A set is finite if it is not infinite. (A comment should be made on this point. Although we all have a clear intuitive understanding of what it means for a set to be FINITE, it is not at all easy to provide a direct mathematical description of this concept. However, it is possible to show that no set of the form {1,2,3,…,n} is infinite, that is, there is no means to match the elements of this set with the elements of a proper subset of itself without producing repetition. This is done with an INDUCTION argument on n. One can use this as a link between this indirect approach and our intuitive under
standing.) Cantor went further to develop an astounding theory of CARDINALITY that shows, among other things, that there are many different types of infinite sets, some deserving of being called “more infinite” than others. The notion of the infinite has been studied and used since antiquity. ARCHIMEDES OF SYRACUSE (ca.287–212 B.C.E.) used the notion of an infinitely small quantity to develop formulae for the areas and volumes of curved figures and solids. (In some vague sense, one can view a circle, for instance, as a regular polygon with infinitely many sides, all infinitely short in length. It is better, however, to view the circle as the LIMIT figure of a sequence of regular polygons.) The geometer EUCLID (ca.300–260 B.C.E.) proved that the set of PRIME numbers is infinite, and ZENO OF ELEA (ca. 490–425 B.C.E.) contemplated the infinite in his studies of time, space, and motion. Although the scholars of Greek antiquity utilized the infinite, they were wary of it. Euclid, for example, went to great pains to phrase matters in a way that never made mention of a quantity that was actually infinite. For instance, he proved that from any given finite collection of primes, one can always construct one more (rather than state that the set of prime numbers is infinite), and in his famous work THE ELEMENTS, he never made mention of lines that continue indefinitely; he only spoke of extending line segments further if needed. (This subtle turn of phrase proved to be important to GEORG FRIEDRICH BERNHARD RIEMANN with his 19thcentury invention of SPHERICAL GEOMETRY.) ARISTOTLE (384–322 B.C.E.) argued that the “actual infinite” did not exist, and that one can only argue in terms of potentiality: given a finite part, one can always provide more. This point of view held fast for almost two millennia. Cantor’s work on the actual infinite, inspired by the beginning ideas of Dedekind and De Morgan, was deemed revolutionary at its time.
inflection (inflexion) See CONCAVE UP/CONCAVE DOWN.
inflection point (point of inflection) A point on a curve at which the tangent line to the curve changes from rotating in one sense (clockwise or counterclockwise) to rotating in the opposite sense. The concavity of the curve changes at such a point. See also CONCAVE UP/CONCAVE DOWN.
integral calculus 271
information theory The branch of
theory that deals with the transmission, processing, and checking of messages sent electronically is called information theory. The field was established in 1948 by American mathematician CLAUDE EDWARD SHANNON, who first showed that it is possible to encode all types of information (words, sounds, pictures) into simple sequences of 0 and 1 bits that can then be transmitted along a wire as pulses. (Up to then, scientists thought it would be necessary to transmit electromagnetic waves along wires.) To adjust for erroneous noise that is often transmitted along with the message, mathematicians have since developed probability techniques that determine the likelihood that the message received is free from errors. Over the decades, mathematicians have also developed efficient redundancy checks to help detect and correct errors. These techniques are far more efficient than simply repeating the message for comparison. Information theory is also used to measure the amount of information a signal of any kind might contain. Transmitting and correctly receiving a single letter of the alphabet—that is, one of any 26 letters— contains more information, for instance, than the receipt of a single binary digit, 0 or 1. Mathematicians use logarithms to measure information content and say that receipt of a letter of the alphabet contains
called its instantaneous value. An instantaneous value is a DERIVATIVE.
log 2 26 ≈ 4.7 as much information as receipt of a log 2 2
(This series is valid for –1 < x < 1.) This shows, for 1 , the fraction 10 has the instance, that in setting x = – – 100 81 integers appearing in turn in each decimal place: 10 – = 0.12345.... (Unfortunately our practice of “carry81 ing digits” disguises the fact that the pattern we see continues.) See also FLOOR/CEILING/FRACTIONAL PART FUNCTIONS.
PROBABILITY
log 2 2 = 1 .) This assumes that log 2 2 each letter in the alphabet is equally likely to occur. In general, if the probability of the letter a being sent is p1, the letter b is p2, and so forth, then the measure of informational content is given by: binary digit. (We have
– p1 log2 p1 – p2 log2 p2 –…– p26 log2 p26 (This agrees with the previous value of 4.7 computed when each probability pi has value 1/26.) This quantity is a measure of the entropy of the initial data set. The field of information theory has obvious applications to work in telegraphy, radio transmission, and the like, but it has also recently been used to analyze human speech, in the study of languages, and in cybernetics.
instantaneous value The value of a varying quantity, such as
VELOCITY,
at a particular instant in time is
integer (directed number, signed number) Any of the positive or negative WHOLE NUMBERs, or ZERO, is called an integer: …–3,–2,–1,0,1,2,3,… More precisely, once the NATURAL NUMBERs have been defined via PEANO’S POSTULATES, say, one can define an integer to be any quantity that can be expressed as the sum or difference of two natural numbers. (For instance, the number –5 can be regarded, formally, as the pair of natural numbers 3 and 8 written in the form 3–8. It can be defined equally well by the pair 1–6 or the pair 12–17, for example.) The difference of any two integers is always another integer, which constitutes a mathematical RING. The set of integers is denoted Z (from the German word Zahlen for “numbers”). German mathematician GEORG CANTOR (1845–1918) showed that the set of integers is COUNTABLE and so has CARDINALITY ℵ0. x The TAYLOR SERIES of f (x) = has the (1 − x)2 integers appearing as coefficients: x
(1 − x)
integral See
2
= x + 2x 2 + 3x3 + 4x 4 + ...
ANTIDIFFERENTIATION; INTEGRAL CAL
CULUS.
integral calculus The calculation of sums of infinitely small quantities, INFINITESIMALs, is called integral calculus. For example, consider the problem of finding the length of a curved path using only a straight yardstick. One could approximate the distance along the curve by marking a number of points along the curve, measuring the straightline segments between them, and summing the lengths of these segments.
272 integral calculus
Approximating length and area
An even better approximation could be made using more points and consequently shorter line segments. The actual length of the curve would be the LIMIT value of these improved approximations as we use shorter and shorter line segments connecting more and more points on the curve (the sum of infinitely short line segments). Similarly, the area under a curve drawn in the plane can be approximated as the sum of the areas of a finite number of rectangles drawn under the curve. Using narrower and narrower rectangles will give better and better approximations. The actual area under the curve is the limit value of these approximations (the sum of infinitely narrow rectangles). Any process that involves segmenting a quantity into manageable pieces, summing, and taking the limit of these sums as the process is refined falls under the category of integral calculus. Traditionally, integral calculus is first taught as the process of finding the area under a curve y = f(x) over an interval [a,b]. The area denoted b
∫ a f(x) dx is called a definite integral, and is defined to be the limit, as h tends to zero, of the sums of the areas of rectangles of width at most h, used to approximate the area of the curve as described above. Such an approximation with rectangles is called a Riemann sum, in honor of the German mathematician BERNHARD RIEMANN (1826–66), whose work led mathematicians to show that this approach is indeed mathematically sound, in particular that, under reasonable conditions, all ways of approximating the area under the curve lead to the same limit value. GOTTFRIED WILHELM LEIBNIZ (1646–1716), one of the inventors of CALCULUS, introduced the symbol ∫ to represent an integral. He thought of it as an elongated
S denoting sum, and he called the theory of integration calculus summatorius. Swiss mathematician Johann Bernoulli (1667–1748) of the famous BERNOULLI FAMILY worked with Leibniz in developing the theory, but he preferred the name calculus integralis and the use of a capital letter I as the sign of integration. (In Latin, the word integralis means “making up a whole.”) The two gentlemen settled on a happy compromise of using Bernoulli’s name for the theory and Leibniz’s symbol for the integral. The idea of using a limit to calculate the areas of curved figures, or the lengths of curved paths, has been used by scholars from the time of ARCHIMEDES OF SYRACUSE in the third century B.C.E. to the time of PIERRE DE FERMAT in the middle of the 17thcentury. In practice, however, the techniques employed to perform these calculations have always been extremely difficult and complicated. The great achievement of Leibniz and ISAAC NEWTON (1642–1727), independent discoverers of calculus, was to recognize that integration is simply a process of reverse differentiation, today called ANTIDIFFERENTIATION. This result, known as the FUNDAMENTAL THEOREM OF CALCULUS, essentially states that to find the area under a curve y = f(x) over an interval [a,b], look for a function F(x) whose DERIVATIVE is f(x). Then: b
∫ a f (x)dx = F(b) – F(a) The function F(x) is called an antiderivative of f. The righthand side of this equation is often abbreviated b as F(x) a . Thus, for example, the area under the parabola y = x2 from x = 0 to x = 2, is given by 2 1 2 1 1 8 ∫0 x 2 dx = 3 x3 = 3 23 − 3 03 = 3 . This remarkable 0 result obviates all need to work with complicated limits. To highlight the interplay between integration and reverse differentiation, the antiderivative of a function f(x) is usually denoted ∫f(x)dx and is called the indefinite integral of f. It is defined up to a CONSTANT OF INTEGRATION. Thus ∫f(x)dx is a function whose derivab tive is f(x) (whereas the definite integral ∫ a f(x)dx is a number equal to the area under the curve y = f(x) over the interval [a,b)]). The thrust of integral calculus is the development of methods for finding the antiderivatives of functions. Since the derivative of the sum of two functions is the sum of the derivatives, and the derivative of a

integration by parts 273 function multiplied by a number k is just k times its derivative, we have:
as
INTEGRATION BY PARTS
See also ∫f (x) + g(x)dx = ∫f(x)dx + ∫g(x)dx ∫k · f (x)dx = k ∫f (x)dx
and
INTEGRATION BY SUB
STITUTION. ANTIDIFFERENTIATION; CALCULUS; DIF
FERENTIAL CALCULUS; DOUBLE INTEGRAL; HISTORY OF CALCULUS
(essay);
IMPROPER INTEGRAL; NUMERICAL
INTEGRATION.
The corresponding results for definite integrals also hold true. This follows from the fundamental theorem of calculus (since these results are true for the antiderivatives), but they can also be seen to be valid by geometric arguments. (For example, if the values of a graph are multiplied by a number k, that is, if the function y = f(x) is replaced by y = k · f(x), then the area under the graph increases by a factor of k.) The following table shows some common integrals: f (x)
∫f (x)dx
xr
xr +1 + C provided r ≠ −1 r +1
1 – x
ln x  + C
ex
ex + C
e ax
1 ax – a e +C
ax
ax –– lna + C
ln x
x lnx – x + C
sin x
–cos x + C
cos x
sinx + C
tan x
–ln cos x  + C
csc x
ln csc x – cot x  + C
sec x
ln sec x + tan x  + C
cot x
cos2 x
ln sin x  + C 1 1 – – 2 x – 4 sin 2x + C 1 1 – – 2 x + 4 sin 2x + C
sinh x
cosh x + C
cosh x
sinh x + C
f ′(x) –– f (x)
ln f (x) + C
sin2 x
There are a number of techniques for finding the indefinite integrals of morecomplicated functions such
integral test See CONVERGENT SERIES.
integrand The function that is to be integrated is called the integrand. For example, the expression f(x) in either of the integrals ∫f (x)dx or ∫ba f (x)dx is the integrand. See also ANTIDIFFERENTIATION; INTEGRAL CALCULUS.
integration by parts This technique of integration is useful for finding the integral of the product of two functions. It makes use of the PRODUCT RULE for differentiation in reverse. Specifically, the product rule reads: d –– (u(x) · v(x)) = u′(x)v(x) – u(x)v′(x) dx Integrating both sides and rearranging thus yields: ∫u(x)v′(x)dx = u(x)v(x) – ∫u′(x)v(x)dx (A CONSTANT OF INTEGRATION will appear when all integrals are finally computed.) Thus one can make effective use of this formula if the integral ∫u′(x)v(x)dx turns out to be much easier to compute. For example, to evaluate ∫x cos x dx, write: u(x) = x and v′(x) = cos(x), yielding u′(x) = 1 and v(x) = sin(x) (again ignoring a constant of integration for the moment) so that: ∫xcos(x)dx = xsin(x) – ∫1· sin(x)dx = xsin(x) + cos(x) + C One typically chooses the factor that is easy to differentiate to be u(x) and the factor that is straightforward to integrate for v′(x). The integrationbyparts formula can be used even if the integrand is composed of a single factor. One can “insert a 1” into the integrand to imagine that there is a second factor equal to the constant function 1. To illustrate, to compute ∫ln(x)dx write: ∫ln(x)dx = ∫ln(x) · 1dx and set:
274 integration by substitution u(x) = ln(x) v ′(x) = 1 u ′(x) =
1 x
v(x) = x
Noting that, since f is the antiderivative of f′, the righthand side of this formula can be thought of as the indefinite integral of f′ evaluated with u as the variable. Thus it is valid to write:
Then:
du
1
∫ ln(x) dx = x ln(x) − ∫ x ⋅ x dx = x ln(x) − x + C The integrationbyparts formula can be applied more than once to complete an integration problem. For example, to evaluate ∫ex cos(x)dx, one application of the technique yields: ∫ex cos(x)dx = ex sin(x) – ∫ex sin(x)dx
∫ f ′(u(x)) dx dx = ∫ f ′(u)du This equation is the changeofvariables equation for integration. For example, we can use it to evaluate du ∫2x(1 + x2)4 dx. Setting u(x) = 1 + x2, giving –– = 2x, dx the integral reads:
∫ 2x(1 + x 2 )
∫ex cos(x)dx = ex sin(x) – ∫ex sin(x)dx = ex sin(x) – [–ex cos(x) – ∫ex (–cos(x))dx] = ex (sin(x) + cos(x)) – ∫ex cos(x) dx
1
∫ e x cos(x)dx = 2 e x (sin(x) + cos(x)) + C The method of integration by parts was discovered by English mathematician BROOK TAYLOR (1685–1731). See also ANTIDIFFERENTIATION; INTEGRAL CALCULUS ; INTEGRATION BY SUBSTITUTION ; RECURRENCE RELATION .
integration by substitution (change of variable, substitution rule for integration) This technique of integration is used to find the integral of a function easily recognized as a COMPOSITION of two simpler functions. It is essentially the CHAIN RULE for differentiation employed in reverse. Specifically, the chain rule states: d du f (u(x)) = f ′(u(x)) dx dx Integrating both sides yields: du ∫ f ′(u(x)) dx dx = f (u(x)) + C
dx = ∫ u 4
du dx dx
= ∫ u 4 du
Applying the integrationbyparts technique to the second integral yields:
Algebra now shows that the integral we seek is given by:
4
=
1 5 u +C 5
=
1 1 + x2 5
(
)
4
+C
Notice that the notation we used here mimics the properties of fractions: we are permitted to replace a term du –– dx under an integral sign by the term dx. dx When using this technique, one typically chooses u(x) to be a function that simplifies a complicated part of the INTEGRAND, and then adjusts matters so that the du factor –– appears explicitly. For example, to evaluate dx ∫x2 √x3 –1 dx, it is natural to set u(x) = x3 – 1. Then du –– = 3x3, and we have: dx
∫ x2
x3 − 1 dx = ∫ x 2 u dx 1 = ∫ 3x 2 u dx 3 =
1 du u dx 3∫ dx
=
1 u du 3∫
=
1 2 32 ⋅ u +C 3 3
=
2 3 x −1 9
(
)
3 2
+C
interest 275 When computing a definite integral via this technique, it is advisable to compute the indefinite integral first and then, at the end of the process, work with the limits of integration. For example, the definite integral 1 5 – 5 ∫ 1 x2 √x3 –1 dx does not equal – 3 ∫ 1√ u du, since the values x = 1 to x = 5 refer to the xvariable and not to the new uvariable. We compute: 5
∫1 x 2 See also
x3 − 1 dx =
(
)
2 3 x −1 9
3 5 2 1
=
3 2 124 2
9
ANTIDIFFERENTIATION; INTEGRAL CALCU
LUS; INTEGRATION BY PARTS.
intercept A point at which two figures intersect is called an intercept. The term is usually reserved for the locations at which the GRAPH OF A FUNCTION crosses the xaxis and the yaxis in a system of CARTESIAN COORDINATES, to form what are called the x and yintercepts. If a straight line has xintercept (a, 0) and yintercept (0, b), then the equation of the x – y line is given by – a + b = 1. This is called the “intercept form” of its equation. The yintercept of an arbitrary function y = f(x) is the point (0, f (0)).
interest A fee paid for the use of money is called interest. For instance, mortgage companies charge home buyers a fee for borrowing money, and credit card companies charge customers a fee for the privilege of using company money to make purchases. In reverse, banks pay customers money for maintaining a balance in a savings account. (The customers have, in effect, lent money to the institution.) In any such arrangement, the amount borrowed or invested is called the principal or capital, and the fee, expressed as a percentage rate, that is, as a number of dollars per hundred to be paid each year of the loan, is called the interest rate. There are two types of interest. The first, simple interest, is computed only on the principal. For example, if a customer borrows $3,000 for 3 years under simple interest computed at 15 percent per annum, then the customer has agreed to pay 0.15 × 3,000 = $450 for each year of the loan, plus return the original $3,000 at
the end of the 3year period. The customer thus pays a total of $4,350 at the end of the 3 years. In general, the total interest I paid on a principal amount P at an annual interest rate R (expressed as a decimal) for T years is given by the formula: I=P×R×T (In our example, I = 3,000 × 0.15 × 3 = 1,350.) At the end of the loan, the amount A owed is: A = P + I = P + P × R × T = P(1 + RT) It is rare today, however, that a lending institution will provide loans with fees computed by simple interest. Typically, one is also expected to pay interest on any interest incurred as the loan progresses. Compound interest is calculated by adding the interest to the principal and recalculating the interest at the end of agreed “conversion periods.” For example, a credit card company may charge 18 percent interest per annum, but will compound the interest monthly (at a rate of 18/12 = 1.5% per month). Thus, after borrowing $1,000, say, a customer is expected to return the principal plus 0.015×1,000 = $15 in interest, that is, a total of $1,015, at the end of the first month. In failing to do so, the customer will then be expected to return this amount, plus an additional 1.5% on this amount, that is, a total of 1,015 × (1 + 0.015) = $1,030.23 at the end of the second month. At the end of each month passed, 0.18 the amount owed increases by a factor of 1 + 12 = (1 + 0.015) . After 1 year, that is 12 months, the balance 0.18 on the credit card will thus be 1, 000 × 1 + 12
12
=
$1,195.62. In general, if a principal amount of P dollars is borrowed at an interest rate of R percent per year (expressed as a decimal), and the interest is compounded n times a year for T years, then the amount A owed at the end of that time period is given by the formula: R A = P 1 + n
nT
276 interior angle Consider now a bank that offers its savings account customers an annual interest rate of 10 percent compounded weekly. A $1,000 investment after 1 52
0.10 year would thus yield 1, 000 × 1 + = $1,105.06 . 52 If the interest is instead computed daily, customers 0.10 receive a slightly better return: 1, 000 × 1 + 365
365
=
= $1,105.16, and an even better return if the interest is compounded hourly, or even every minute. The maximum possible return is achieved when interest is computed every instant. In the mid1700s Swiss mathematician LEONHARD EULER tackled the problem of computing continuously compounded interest and showed that as R )nT n becomes large, the quantity A = P(1 + – n approaches the value PeRT. (A study of the number e establishes this claim.) The formula: A = Pe RT thus represents the balance of an investment, or a loan, after T years, under the ideal state of interest compounded continuously. Thus our $1,000 investment, after 1 year, compounded continuously at a rate of 10 percent per annum, yields a return of 1,000 × e0.10×1 = $1,105.17. Banks today use this formula to compute interest on savings accounts. The practice of charging a fee for the use of money is an ancient one. Records show that many civilizations, the Hebrews, the Greeks, the Romans, and even the Babylonians of 2000 B.C.E., for instance, charged simple interest on loans. The Romans called the practice usury and often fees were as high as 60 percent. Some religious orders, including Christianity, Judaism, and Islam, questioned the ethics of the practice, arguing that one should only be charged a fee for the use of something that could be worn out or lose value due to wear and tear. (At the time, money was not seen to lose any value during the course of a loan.) For many centuries the practice of charging a fee for the use of money was forbidden by these orders. During the growth of industry and trade during the Middle Ages and the Renaissance, attitudes changed. More and more people requested cash loans to take part in new opportunities, and lenders felt it appropri
ate to be compensated for not taking part in those opportunities themselves. Lenders again started charging fees. The church relaxed its attitude toward usury, and new establishments, called banks, were formed to handle, store, and loan money. The term interest from the Latin phrase id quod interest meaning “that which is between” soon replaced the term usury. Today the word usury is used only in a negative context of charging illegally high interest rates. See also E.
interior angle The
formed by two sides of a POLYGON lying inside the polygon is called an interior angle. For example, all four interior angles of a RECTANGLE equal 90°. If an interior angle is greater than 180°, then that angle is called a “reentrant angle” and the polygon is concave. Any interior angle less than 180° is called “salient.” Each interior angle of an nsided regular polygon n–2 equals –– n × 180°. See also CONCAVE/CONVEX; TRANSVERSAL. ANGLE
intermediatevalue theorem (Bolzano’s theorem) Named after the Czech mathematician BERNHARD PLACIDUS BOLZANO (1781–1848), the intermediatevalue theorem asserts that if f(x) is a CONTINUOUS function defined on a closed INTERVAL, then this function assumes every value between f(a) and f(b); that is, if N is any number between f(a) and f(b), then there is at least one point c between a and b such that f(c) = N. For example, the function f(x) = x2 is continuous on the interval [3,4] and has f(3) = 9 and f(4) = 16. Since 11 is between 9 and 16, the intermediate value theorem ensures us of the existence of a number c, between 3 and 4, such that f(c) = c2 = 11, that is, it proves that the square root of 11 exists. The theorem is intuitively clear if we think of a continuous function on a closed interval as one whose graph consists of a single continuous piece with no gaps, jumps, or holes: in moving a pencil from the left endpoint (a, f(a)) to the right endpoint (b, f(b)), it seems obvious that the pencil tip adopts all “heights” between initial height f(a) and final height f(b). In climbing the face of a mountain, say, one must indeed pass through
interpolation 277 every intermediate height from the base to the apex. A rigorous proof of the theorem, however, relies on the notion of the “completeness” of the real numbers (meaning that no points, like the square root of 11, are missing from the real line). This is a subtle property, one that was not properly understood until the late 1800s with the construction of a DEDEKIND CUT. The intermediatevalue theorem is useful for locating roots of equations. For example, consider the function f(x) = x3 +2 x – 5. It is continuous on the interval [1,2] and satisfies f(1) < 0 and f(2) > 0. It follows then that the equation x3 + 2x – 5 = 0 has a root somewhere between 1 and 2. By working with smaller and smaller intervals, one can often use this method to determine the location of a root with a good degree of precision. The BISECTION METHOD, for example, employs this technique. The intermediatevalue theorem has a number of amusing consequences: In theory, it is always possible to slice a pancake, no matter how irregular its shape, exactly in half with a single straightline cut. Hold a knife to the left of the pancake so that 100 percent of the cake lies to its right. Now slide the knife, in parallel, across the pancake until it lies on the other side of the cake. At this location, zero percent of the pancake lies to the right of the knife. By the intermediatevalue theorem, there must be some intermediate location for the knife that yields the value of 50 percent lying to its right. That is, there is, in theory, a knife position that cuts the pancake exactly in half. Note that this result does not depend on the angle we initially hold the knife—vertically, horizontally, or diagonally. We have in fact shown that it is always possible to slice a pancake in half with a knife held at any previously set angle. In theory, it is always possible to simultaneously slice two pancakes each exactly in half with a single straightline cut, no matter the shapes of the pancakes nor their location on the table. This result is known as the twopancake theorem. The previous result assures us we are always able to slice the first pancake exactly in half, pointing the knife at any angle we care to choose. The concern is that the knife might or might not cut the second pancake. Sup
pose we find a direction, deem this angle zero degrees, that slices the first pancake in half, but misses the second pancake entirely, with 100 percent of the second pancake lying to the right of the knife, say. Turn the knife 180°. We now have a knife cut (same line, opposite direction) that slices the first pancake exactly in half, with the second pancake lying entirely to its left, that is, zero percent to the right. By the intermediatevalue theorem, there must be an intermediate angle between zero degrees and 180° that slices the first pancake exactly in half, and has 50 percent of the second pancake lying to its right—that is, one that simultaneously slices the second pancake in half as well. At any instant, there are two points on the Earth’s equator directly opposite each other with exactly the same air temperature. For each position θ degrees longitude on the equator, let f (θ) be the air temperature at this position minus the air temperature at the opposite point of the equator, at θ + 180°: f(θ) = temp(θ) – temp(θ + 180°) Notice that f (θ + 180°) equals the same value, but opposite in sign, to f(θ): f(θ + 180°) = temp(θ + 180°) – temp(θ) = –f(θ) Thus the function f(θ) moves from positive to negative values. By the intermediatevalue theorem, there must be an intermediate location where the function is zero. This is the desired position on the Earth’s equator. See also EXTREMEVALUE THEOREM; FIXED POINT; HAMSANDWICH THEOREM; SPHERE.
interpolation The process of estimating the value of a function between two values already known is called interpolation. For example, if the temperature of a cup of tea was initially 200°F, and two minutes later it was 180°F, one might guess that its temperature at the oneminute mark was 190°F, based in the assumption that the temperature decreases steadily over time. This represents that simplest method of interpolation, called linear interpolation: we presuppose that the variation of the function can be described as a straight line passing through the known values.
278 interval If a function f has values f(a) and f(b) at locations a and b, the linear interpolation estimates the value of the function at x, between a and b, as: f (x) ≈ f (a) +
x−a (f (b) − f (a)) b−a
If the graph of f is smooth and the interval [a,b] is small, then linear interpolation generally gives a good approximation of the true value. If, however, the interval is large, it is less likely that this will remain the case. For example, given that 32 = 9 and 34 = 81, linear interpolation suggests that 33 (which equals 27) is 3–2 approximately 9 + ––– 4 – 2 (81 –9) =45. Improved methods of interpolation take into account other data values. For example, JOSEPHLOUIS LAGRANGE’s interpolation method, called LAGRANGE’S FORMULA, estimates function values by fitting a POLYNOMIAL to all the known data values. Alternatively, if it is suspected, for example, that the function is exponential, that is, of the form y = bx, then linear interpolation works well for the logarithm of the function: the relationship log y = b log x is linear. EXTRAPOLATION is the process of estimating function values outside of the range of values observed. See also REGRESSION.
interval Any single segment of the number line is called an interval. More precisely, an interval is a set of numbers containing all real numbers between two given numbers. The given numbers are called endpoints, and they might, or might not, be regarded as part of the interval. For example, the set of all real numbers between, and including, 2 and 5 is an interval; as is the set of all real numbers strictly greater than 9 but less than or equal to 23. An interval is called closed if both endpoints are included. Square brackets are used to indicate that this is the case. For example, [a,b] denotes the set of all real numbers x that satisfy a ≤ x ≤ b. On a number line, the endpoints are marked by blackedin circles to indicate that they are included. An interval is open if neither endpoint is included. Parentheses are used to indicate that this is the intent. For example:
(a,b) denotes the set of all real numbers x that satisfy a < x < b. On a number line, the endpoints are marked by open circles to indicate that they are not included. If a variable x satisfies either a ≤ x < b or a < x ≤ b then it is said to be located in a halfopen interval denoted, respectively: [a,b) or (a,b] On a number line, a blackedin circle marks the endpoint that is included in the interval and an open circle the endpoint that is excluded. With this notational system, the two intervals described in the opening paragraph are thus denoted [2,5] and (9,23]. Some texts use reverse square brackets instead of parentheses to denote the exclusion of endpoints—thus an open interval is denoted ]a,b[, and a halfopen interval as [a,b[, for example. The length of an interval is the distance between its endpoints. For example, the interval (9, 23] has length 14, as do the intervals [9, 23], [9, 23), and (9, 23). Incorporating the symbol ∞ for infinity allows one to denote unbounded intervals. For example, [a, ∞) denotes the set of all real numbers x larger than or equal to a; (–∞,a) denotes the set of all real numbers x strictly less than a; and so on. Sometimes mathematicians write (–∞,∞) for the set of all real numbers. (One typically does not place a square bracket next to the infinity symbol, although it is possible to define a number system that allows for inclusion of ∞ as a valid number for consideration.) An interval can consist of a single point (the interval [1,1] for example) or can be the empty set (the interval [4,1] for example). The intersection of any two intervals is an interval, but the union of two intervals might or might not be an interval. See also CARDINALITY; INTERMEDIATEVALUE THEOREM.
invariant A property, quantity, or relationship that is not changed by collection of specific operations or transformations, or by how that property is observed, is called an invariant. For example, the distance between two points in a plane does not change under a ROTATION. Thus distance is invariant of rotations. (It is not, however, invariant of DILATIONs, for instance.) An
inverse element 279 application of EULER’S THEOREM shows that any soccer ball design that uses hexagons and pentagons, no matter how irregular those shapes may be, with three edges meeting at each vertex, must include precisely 12 pentagons. (The number of hexagons can vary.) The count of pentagons is thus an invariant in soccer ball design. Some invariants can be quite surprising. For example, consider the following theoretical exercise: Suppose 10 ft of length is added to a rope that was just long enough to wrap snugly around the equator of the Earth. Imagine that the rope is again wrapped around the equator, but this time—due to its extra length—it hovers just above the ground. How high is the gap between the ground and the suspended rope? To answer this question, let R denote the radius of the Earth. The length of the extended rope is 2π R + 10 = 5 2π R + feet, which corresponds to the circumference π 5 feet. This shows that the of a circle of radius R + – π 5 ≈ 1.6 ft off the ground. Notice that the rope hovers – π answer, apart from being surprisingly large in value, does not depend on the value of R. This means that the size of the planet is immaterial. Thus if 10 ft of length is added to a rope that fits snugly around the equator of any planet—Mars, Jupiter, or a planet the size of a 5 ft off the pea—the extended rope will always hover – π ground. This value is an invariant for the problem. As another example, consider the following famous “splitting game”: Write the number 12 at the top of a page and below it write a pair of positive whole numbers that sum to 12, say, 7 and 5. On the side of the paper record the product 7 × 5 = 35. Now write below 7 a pair of numbers that sum to 7, say, 3 and 4, and record the product 3 × 4 = 12. Continue in this manner, “splitting” each number that appears in the diagram into two and recording the product of the pair of numbers chosen. Do this until the number 1 appears 12 times. The following represents one possible such splitting diagram:
The splitting game
Now sum all the products recorded. What value is obtained? Surprisingly, no matter which splitting diagram one constructs, the sum of products is an invariant of the game and will always have value 66. (The number 66 happens to be the 11th TRIANGULAR NUMBER. In general, if one begins this game with a number N, then the invariant that arises in the game is the (N –1)th triangular number.)
inverse element An element of a set that, when combined with another element produces the IDENTITY ELEMENT of the set, is called an inverse element. More precisely, if a set S comes equipped with a BINARY OPERATION “*” and an identity element e, then an inverse for an element a of the set is another element b such that a*b = b*a = e. For example, for the set of numbers under the operation of addition, the inverse of any number a is its negative –a. In this context, the identity element is zero and we do indeed have: a + (–a) = (–a) + a = 0 Under the operation of multiplication, the identity element is 1, and the inverse of any (nonzero) number a is 1: its reciprocal – a 1 1 a× – = – ×a=1 a a Each and every element of a GROUP is required to have an inverse. The inverse of the identity element is itself.
280 inverse function It is not possible for a single element a to have two different inverses b1 and b2. This is established by noting that e = a*b2 so that b1 = b1*e = b1*(a*b2) = (b1*a)*b2 = e*b2 = b2. From the symmetry of the definition, we have that if b is the inverse of a, then a is also the inverse of b. Consequently, the inverse of an inverse is the original element. Phrased in terms of addition, this reads –(–a) = a and in terms of multiplication as: 1 =a 1 a The inverse of an element a is often denoted a–1, especially if the binary operation under consideration can be interpreted as a type of multiplication. For example, the inverse of a square MATRIX A, if it exists, is denoted A–1.
inverse function (inverse mapping, reverse function) A FUNCTION f with domain D and range R, f : D → R, is said to be invertible or to have an inverse function if, for each possible output y of the function, y ∈ R, there is one, and only one, input x ∈ D, that produces that output. We write x = f –1(y) for the input x that produces the given output y. (Thus x = f –1(y) if, and only if, f(x) = y.) This then defines a function f –1: R → D, called the inverse function to f. In some sense, the inverse function “undoes” the original function. For example, consider the function on real numbers that doubles an input and adds 3: f(x) = 2x + 3. The output 11 is produced from the input of 4, and so we have f –1(11) = 4. In general, an output of y is produced y–3 y–3 from an input x = ––– , and so f –1(y) = ––– . (This 2 2 formula is obtained by solving for x in the equation: y–3 y = 2x + 3 to yield x = ––– .) 2 Since f –1(y) is the input that produces the output y, and x is the input that produces the output f(x), the following relations hold: f(f –1(y)) = y for all values y in the range of f and f –1(f(x)) = x for all x in the domain of f This explains the awkward notation for the inverse function: In the study of the COMPOSITION of functions,
f m denotes the composite f0 f0…0 f (m times), and we have f m0 f n = f m+n. To give meaning to the quantity f 0, this rule states that f 0 f 0 = f 10 f 0 = f 1+0 = f, suggesting that we should set f 0(x) = x for all values x. Consequently, the statement f –1 0 f 1 = f 0 suggests that f –1(f(x)) = x for all x, indicating that f –1 is the appropriate notation for the inverse function. The superscript of –1 should not be confused with the operation of 1 –– , and leave inversion. (We write (f(x))–1 to denote f(x) f –1(x) to mean the inverse function of f.) It is customary to denote the input of a real function as the variable x and the output as the variable y. This can lead to some confusion. For instance, to compute the inverse function of y = f(x) = x3 + 2 we solve for the input x in the equation in terms of the output y to yield, 3 x = √y – 2 , but we interchange the x and y variables so that x denotes the new input and y the new output: y = 3 3 √x – 2 . This yields the formula f –1(x) = √x – 2 for the inverse function. As the formulae y = f(x) and x = f –1(y) represent exactly the same equation, the two formulae yield exactly the same curves when plotted against a pair of x ycoordinate axes. Following the convention to interchange the x and yvariables for the second equation to write y = f –1(x) is tantamount to interchanging the x and yaxes in the graph of the curve. Flipping the graph across the diagonal line y = x returns the yaxis to the vertical position and the xaxis to the horizontal position, but also flips the curve drawn across the diagonal line. Thus the graphs of y = f(x) and y = f –1(x) are mirror images of each other across a diagonal line. Not every function possesses an inverse function. For example, there is no inverse function to the squaring function y = x2: some outputs arise from more than one possible input. (The output of 4, for instance, arises from the two inputs 2 and –2.) However, it is often possible to restrict a function to a certain portion of its domain and define an inverse function for that restricted domain. For instance, for the squaring function, if we require that only nonnegative inputs are to be considered, then an inverse function does exist: we – have y = √ x (the positive square root) as inverse function. One defines the INVERSE TRIGONOMETRIC FUNCTIONS, for example, by restricting to a suitable portion of the domain. If the function y = f(x), then the derivative of the inverse function y = f –1(x) is given by:
inverse matrix 281 d −1 1 f (x) = dx f ′ f −1(x)
(
)
provided that the quantity in the denominator is not zero. This formula can be established by making use of the CHAIN RULE in the statement f(y) = x. (Differentiating 1 gives: f ′(y) · y′ = 1 and so y′ = –– , which is the above f ′(y) formula.) A simple version of a general inversefunction theorem states that if the derivative of a function y = f(x) is nonzero at a point x = a, then an inverse function exists, at least when the domain is restricted to a small interval about a. (Since the derivative of f(x) = x2 is zero at x = 0, it is not possible to define an inverse function to the squaring function about the point x = 0.)
inverse hyperbolic functions (area hyperbolic functions) Defined in an analogous way to the INVERSE TRIGONOMETRIC FUNCTIONS, the inverse hyperbolic functions are the inverse functions of the HYPERBOLIC FUNCTIONS. For instance, the inverse hyperbolic sine of a number x, written arc sinh x or sinh–1x, is a value a whose hyperbolic sine is x: sinh a = x. Similarly, the inverse hyperbolic cosine of x is a value a with cosh a = x, and the inverse hyperbolic tangent of x is a value a such that tanh a = x. (Technically, for a given value x there are two different values a for which cosh a = x, one positive and one negative. By convention, the positive value is always chosen.) Since the hyperbolic sine function is defined on all real values and yields all real values as possible outputs, the function sinh–1 x is defined for all real values of x. On the other hand, the hyperbolic cosine function only yields output values greater than or equal to 1, and consequently the inverse hyperbolic cosine function cosh–1 x is defined only for values of x ≥ 1. Similarly, the inverse hyperbolic tangent function tanh–1 x is defined only for –1 < x < 1. The inverse hyperbolic functions have the following DERIVATIVEs: d 1 sinh −1 x = dx 1 + x2 d 1 cosh −1 x = 2 dx x −1 d 1 tanh −1 x = dx 1 − x2
These can be established by making use of the relation cosh2 y – sinh2 y = 1. For instance, to compute the derivative of y = sinh–1 x, write sinh y = x and then differentiate this equation making use of the CHAIN RULE. dy This yields cosh y · –– = 1, thereby establishing: dx dy 1 1 1 = = = dx cosh y 1 + (sinh y)2 1 + x2 as claimed. It is possible to give alternative formulations of the inverse hyperbolic functions. Noting that cosh y = ey + e –y ey – e –y –––– and sinh y = –––– 2 2 , we have cosh y + sinh y = ey or y = ln(cosh y + sinh y). Set y = sinh–1 x. Then sinh y = x and cosh y = √1 + sinh2 y = √1 + x2 , yielding: sinh–1 x = ln(x + √x2 + 1 ) which is valid for all values of x. Similarly, cosh–1 x = ln(x + √x2 – 1 ) valid for x ≥ 1, and tanh −1 x = ln
1+ x 1− x
valid for – 1 < x < 1.
inverse matrix (matrix inverse) A square
MATRIX A is said to be invertible (or nonsingular) if there is a matrix B such that AB = BA = I, where I is the IDENTITY MATRIX. The matrix B is called the inverse matrix to A. There is at most one inverse matrix to given matrix A. (If B1 and B2 are both inverse matrices, then B1 = IB1 = B2AB1 = B2I = B2.) If an inverse matrix for a matrix A exists, then it is denoted A–1. A study of DETERMINANTs shows that a matrix is invertible if, and only if, its determinant is not zero. The matrix inverse of a 2 × 2 matrix
⎛ a b⎞ A=⎜ ⎟ ⎝ c d⎠
282 inverse of a statement with determinant det(A) = ad – bc is the matrix: A −1 =
1 d −b ad − bc −c a
The process of GAUSSIAN ELIMINATION provides a relatively straightforward method for computing the matrix inverse of any square matrix of a larger size. If, in a system of n SIMULTANEOUS LINEAR EQUATIONS Ax = b, the matrix A of coefficients is invertible, then the system has solution given by x = A–1b. In particular, if ej denotes the column vector whose only nonzero entry is a 1 in the jth position, then xj = A–1ej is the jth column of A–1. That is, the jth column of this inverse matrix is a solution to the system of equations Axj = ej. By CRAMER’S RULE, the ith entry of this column, that is the (i,j)th entry of the inverse matrix, is the ratio of determinants: (A −1)ij =
det(A ij ) det(A)
where A ij is the matrix A with the ith column replaced by ej. Computing det(A ij) is equivalent, up to a plus or minus sign, to computing the determinant of the matrix obtained from A by deleting its ith column and jth row. This value is sometimes called the (i,j)th cofactor of A. If A is invertible, then it is impossible to find a nonzero column vector x such that Ax = 0. (Otherwise, x = A–1 0 = 0.) This observation is important for the study of EIGENVECTORs and EIGENVALUEs. See also GENERAL LINEAR GROUP.
inverse of a statement See CONTRAPOSITIVE.
inverse square law Any relationship between two physical variables for which one is proportional to the RECIPROCAL of the square of the other is referred to as an inverse square law. For example, the law of gravitation as developed by SIR ISAAC NEWTON (1642–1727) asserts that the magnitude F of the gravitational force between two bodies of masses m and M is given by: F =G
mM r2
Here G is the gravitational constant (equal to 6.67 × 10–11 m3kg–1sec–2) and r is the distance between the two masses. This is an inverse square law. The illumination provided by a source of light decreases by the inverse of the square of the distance from the source and so too is an inverse square relationship.
inverse trigonometric functions An
INVERSE FUNCto any trigonometric function is called an inverse trigonometric function. For instance, the inverse sine of a number x, written arcsinx or sin–1x, is an ANGLE a for whose sine is x: sina = x. Since the sine curve adopts values only between –1 and 1, the inverse sine function is defined only for values –1 ≤ x ≤ 1. One should also note that for any value x there are infinitely many angles a with sin a = x. It is usually assumed then that the π π – angle a is chosen so that – – 2 ≤ a ≤ 2 . (This is called the TION
range of principal values for sine.) Similarly the inverse cosine of a number x with –1 ≤ x ≤ 1, written arccos x or cos–1 x, is an angle a, usually chosen in the principal range for cosine, 0 ≤ a ≤ π, with cos a = x. Since the tangent function adopts all real values, the inverse tangent function is defined for any real number x, and arctan x, or tan–1 x, is defined to be that angle a in the principal π π – range for tangent, – – 2 ≤ a ≤ 2 , such that tan a = x. The inverse trigonometric functions have the following DERIVATIVEs: d 1 sin −1 x = , for x ≠ ±1 dx 1 − x2 d 1 cos −1 x = − , for x ≠ ±1 dx 1 − x2 d 1 tan −1 x = dx 1 + x2 These can be established by making use of the relation sin2 y + cos2 y = 1. For instance, to compute the derivative of y = sin–1 x, write sin y = x and then differentiate this equation making use of the CHAIN RULE. This dy yields cos y –– dx = 1, thereby establishing: dy 1 = = dx cos y
1 1 − (sin y)
as claimed. The TAYLOR gives GREGORY’S SERIES.
SERIES
2
=
1 1 − x2
of the arctan function
isometry 283 One can establish a number of forwardinverse identities for the trigonometric functions. As examples we have: sin(cos −1 x) = 1 − x 2 sin(tan −1 x) =
x 1 + x2
cos(sin −1 x) = 1 − x 2 cos(tan −1 x) = tan(sin −1 x) = tan(cos −1 x) =
1 1 + x2 x 1 − x2 1 − x2 x
x For instance, if a = cos –1 x, then cos a = x = – . Thus 1 angle a appears in a right triangle with hypotenuse 1 and adjacent leg of length x. By PYTHAGORAS’S THEOREM, the length of the opposite leg is √1 – x2 and so – x2 = √1 – x2 , establishing the first relasin a = √1 ––––– 1 tion. The remaining identities are proved similarly. See also INVERSE HYPERBOLIC FUNCTIONS.
irrational number Any number that cannot be expressed as a RATIO of two integers is called an irrational number. As the study of RATIONAL NUMBERS shows, the irrational numbers are precisely those numbers whose decimal expansions do not terminate or fall into a repeating cycle of values. For example, the number with the decimal expansion 0.113133133313333133333133… is irrational. The study of rational numbers also shows that, in a very real sense, “most” numbers are irrational. A famous proof, often attributed to Hippasus of – Metapontum (ca. 470 B.C.E.), shows that √ 2 is irrational. THEODORUS OF CYRENE (ca. 465–398 B.C.E.) established the same result geometrically, and also – – showed that the numbers √ 3 through to √17 (exclud– – – ing √ 4, √ 9, and √16) are irrational. The FUNDAMENTAL THEOREM OF ARITHMETIC can be used to prove that the mth root of a positive integer n is rational if, and only if, n is already the mth power of an integer. m– a (If √n = – for some integers a and b, then am = nbm. b
Writing each of a, b, and n as a product of primes, and noting that the primes that consequently appear on the left side of this equation must match those that appear on the right, we conclude that each prime factor of n appears in n a multiple of m times. This establishes that n = cm for some integer c.) The same reasoning shows that a number such as log25 is irrational. (If a log25 = – , then 2a = 5b, contradicting the fundamental b theorem of arithmetic.) Truncating the decimal expansion of an irrational number produces a rational arbitrarily close to that irrational number. For example, 1.4, 1.41, 1.414, … is – a sequence of rational numbers converging to √ 2 = 1.41421356… In 1737 LEONHARD EULER established that the number e is irrational, and in 1761 JOHANN HEINRICH LAMBERT (1728–77) proved the irrationality of π. No one to this– day knows whether or not the numbers 2e, πe, and π√2 are irrational. (It is known that eπ and e · π are irrational.) Surprisingly, the rationality or irrationality of EULER’S CONSTANT γ is still not known. It is possible for an irrational number raised to an irrational power to be rational. For example, if x = – √–2 (√ 2) turns out to be rational, then we have an example of such a phenomenon. If x, on the–other hand, is – – – not rational, then it is irrational and x√2 = (√ 2)√2 )√2 = – (√ 2)2 = 2 is an example of what we seek. (Unfortunately this indirect line of reasoning does not indicate which of the two possibilities actually occurs.) See also ALGEBRAIC NUMBER; CONTINUED FRACTION; E; NUMBER; REAL NUMBERS; SURD; TRANSCENDENTAL NUMBER.
isolated point (acnode) A point that satisfies the equation of a curve but is not on the main arc of the curve is called an isolated point. For example, the curve has y2 = x3 – x2 has x = 0, y = 0 as a solution, with no other solution near this position. The point (0,0) is an isolated point for the equation. See also DOUBLE POINT.
isometry (congruence transformation) A GEOMETRIC TRANSFORMATION,
such as a translation, rotation, or a reflection, that preserves the distances between points in space is called an isometry. Isometries thus have the property of preserving the shape and size of geometric
284 isomorphism figures. (If, for instance, three points A, B, and C represent the vertices of an equilateral triangle of sidelength 1, then the images A′, B′, and C′ of the three points under an isometry again form an equilateral triangle of sidelength 1.) The FUNDAMENTAL THEOREM OF ISOMETRIES shows that any isometry in the plane is the result of at most three reflections. Any isometry that preserves the location of three planar points leaves all points in the plane unchanged. No isometry can preserve the location of only two points. If an isometry preserves the location of just one point, then that isometry must be a rotation about that point.
isomorphism A onetoone and onto correspondence between the elements of two sets equipped with BINARY OPERATIONs that preserved the operations on those sets is called an isomorphism. For example, the map f:n → 2n matches each natural number with an even number (and, in reverse, each even number is paired with a unique natural number) and, moreover, preserves the operation of addition between the two sets: We have f (n + m) = f(n) + f(m) for all natural numbers n and m. Thus, with respect to addition, f is an isomorphism between the two sets. (Note, however, that f does not preserve products and so is not an isomorphism with respect to multiplication.) The LOGARITHM function ln x provides a map between the set of all positive real numbers and the set of all real numbers. The logarithm converts multiplications in the first set into additions in the second set: ln(xy) = ln x + ln y. The set of positive reals under multiplication is thus isomorphic to the set of all reals under addition. The word isomorphism is derived from the Greek words isos meaning “same” or “equal” and morphos meaning “shape” or “structure.” See also HOMOMORPHISM.
isoperimetric problem A classical problem in mathematics, called the isoperimetric problem, asks: Of all figures drawn in the plane with a given perimeter, which encloses the largest area? The origin of this problem dates back to antiquity with the famous legend of Princess Dido and DIDO’S PROBLEM.
The topic of isoperimetrics (prefix iso: the same) was systematically studied by the 17th and 18thcentury Swiss family of mathematicians, the BERNOULLI FAMILY. They discovered and classified many curves satisfying certain maximum and minimum properties, but did not solve the isoperimetric problem. LEONHARD EULER (1707–83) also contributed to this field, using the techniques of CALCULUS. JAKOB STEINER (1796–1863) used purely geometric methods to establish the following partial answer: if there were a shape that offered maximal area for a given perimeter, then that shape must be a circle. Steiner thought that he had completely solved the isoperimetric problem, that the answer must be a circle, but several years later German mathematician KARL THEODOR WILHELM WEIERSTRASS (1815–97) pointed out that the assumption that problems always have solutions can lead to absurdities. The following amusing example illustrates his concern: Consider the question: of all positive integers, which is the largest? If there is an answer to this problem, then it cannot be a number n different from 1, for then n2 is an integer larger than n. This leaves only the integer 1 as the answer to the question. Weierstrass realized that it could very well be the case that the isoperimetric problem has no solution, despite Steiner’s work, and so perhaps even the circle can be replaced with a shape of greater area for the same perimeter. To completely solve the isoperimetric problem, Weierstrass realized one must also prove that a solution to the problem exists. Finally, in 1870, Weierstrass was able to do this by developing a new mathematical theory today called the “calculus of variations.” The solution to the problem, known by ancient scholars, was, at long last, established as mathematically correct. Simple variations of the isoperimetric problem are easier to handle. Consider, for example, the question: of all rectangles of perimeter 40 ft, which has the largest area? This can be answered via an exercise in algebra: Any rectangle has two long sides and two short sides. If the perimeter is 40 ft, then two sides are longer than 10 ft, and two are shorter than 10 ft (unless the figure is a square). Writing the dimensions of the rectangle as 10 + x ft long, and 10 – x ft wide, for some number x, its area is then given by the formula
iteration 285 (10 – x)(10 + x) = 100 – x2. We see now that the area is maximal precisely when x is zero, that is, the figure is a square. An exercise in geometry shows that of all triangles of a given perimeter, the equilateral triangle has largest area. See also SOAP BUBBLES; STEINER POINT.
isosceles trapezoid (isosceles trapezium) A trapezoid with the two nonparallel sides of equal length is called an isosceles trapezoid. The figure has a line of symmetry perpendicular to the parallel sides. Any isosceles trapezoid can be considered a truncated ISOSCELES TRIANGLE. See also TRAPEZOID/TRAPEZIUM.
isosceles triangle A triangle possessing two sides of equal length is called isosceles. The name is derived from the Greek terms iso meaning “same” and skelos meaning “leg.” Some scholars insist that a triangle be called isosceles only if two, but not all three, of its sides are equal in length, but this is really only a matter of taste. (A triangle with all three sides equal in length is called equilateral.) The base angles of an isosceles triangle are equal. This can be established as follows: Suppose triangle ABC is isosceles with sides AB and BC equal in length. Regard a diagram of this triangle as the picture of two triangles, ABC and CBA superimposed on top of one another. These two triangles share pairs of sides that match in length, and share a common angle (at B) between those two sides. Thus, by the SAS principle from the study of similarity, the two triangles are congruent. Consequently, all corresponding angles between the triangles match. In particular, the angle at vertex A must match the angle at vertex C. If one slides the vertex B in a direction parallel to the base AC, then the perimeter of the triangle changes 1 (although the AREA of the figure, A = – × base × height 2 does not). The study of OPTIMIZATION shows that the sum of distances AB + BC is at a minimum when the triangle is isosceles. We have: Of all triangles with a given base and a fixed height, the isosceles triangle has the least perimeter.
Isosceles triangle
To minimize the entire perimeter of a triangle without changing its area, one can adjust each vertex of the triangle (by sliding the vertex along a line parallel to the opposite side) so that it is an apex of an isosceles triangle. This produces an equilateral triangle and thereby establishes: Of all triangles of a fixed area, the equilateral triangle has the least perimeter. It is possible to arrange five points in a plane so that any three points chosen at random among them form the vertices of an isosceles triangle. (The points lie at the vertices of a regular pentagon.) It is impossible to accomplish the same feat with six or more points. See also AAA/AAS/ASA/SAS/SSS; CONGRUENT FIGURES; TRIANGLE.
iteration The repeated application of a mathematical procedure in which each step is applied to the output of the previous step is called iteration. For example, HERON’S METHOD for computing the square roots of numbers is an iterative procedure. The iterates of a function f, beginning with the value x0, are defined by the sequence of values: x1 = f(x0), x2 = f(x1) = f(f(x0)), x3 = f(x2) = f(f(f(x0))),… If this sequence converges to a value a, say, then: a = limn→∞ xn = limn→∞ f(xn –1) = f(a) (We are assuming here that f is a CONTINUOUS FUNCTION.) Thus iteration provides a means for solving
286 iteration equations of the form x = f(x). For example, repeatedly pushing the “cosine” button on a calculator will produce values that approach a solution to the equation x = cos x. (As the calculator can only display eight or 10 decimal places, this means that the values shown
on the screen will eventually stabilize.) NEWTON’S METHOD is often cited as the most widely used iterative procedure. See also CONVERGENT SEQUENCE; DYNAMICAL SYSTEM; RECURSIVE DEFINITION.
J Jacobi, Carl Gustav Jacob (1804–1851) German
b
d
du
∫a f (u) du = ∫c f (u(x)) dx dx
Analysis Born on December 10, 1804, in Potsdam, Prussia (now Germany), Carl Jacobi is remembered for his important work on the theory of elliptic functions and for applying his work in astonishing ways to the theory of numbers. He proved a famous conjecture of PIERRE DE FERMAT (1601–65) stating that every number can be written as the sum of four perfect squares. (Previously, CARL FRIEDRICH GAUSS had proved that every number is the sum of the three triangular FIGURATE NUMBERS.) He also made important contributions to the theory of dynamics in physics and made a careful study of the theory of DETERMINANTs. The generalized CHANGE OF VARIABLE formula for DOUBLE INTEGRALs (and highermultiple integrals) contains a determinant that today is named in his honor. Jacobi completed his entire secondary education within 1 year but was forced to wait several years before reaching the minimum of age of 16 to enter the University of Berlin. During this time Jacobi read advanced works in mathematics and conducted research work on polynomial equations. Jacobi received his doctorate from Berlin in 1825 and took a teaching position at the University of Königsberg a year later. By this time, Jacobi had already made some fundamental discoveries in the field of NUMBER THEORY. He commenced his work on elliptic functions soon afterward. The changeofvariables formula in INTEGRAL CALCULUS states that if f is a function of a variable u, which in turn is a variable of x, then the integral of f with respect to u can be computed as:
Here the limits of integration change to reflect the change of variable. In his study of determinants, Jacobi showed, in two dimensions, that if f(u,v) is a function of two variables, with u and v each functions of x and y, then the appropriate changeofvariable formula becomes: ∂u ∂x ∫∫ f (u, v) du dv = ∫∫ f (u(x, y), v(x, y)) ∂v R S ∂x
∂u ∂y dx dy ∂v ∂y
where R is a region in the uvplane, S is the correspond∂u ∂u ∂x ∂y ing region in the xyplane, and ∂v ∂v represents the ∂x ∂y determinant of the 2 × 2 MATRIX of PARTIAL DERIVATIVEs. An analogous result applies for triple and highermultiple integrals. Such determinants are today called Jacobians. Although French mathematician AUGUSTINLOUIS CAUCHY (1789–1857) had discovered these transformation formulae earlier, it was Jacobi who first developed the theory of functional determinants fully in his comprehensive 1841 publication De determinantibus functionalibus (Functional determinants). 287
288
Jiuzhang suanshu
Jacobi was passionate about theoretical research and once uttered the following defense for pursuing science for its own sake: “The sole aim of science is the honor of the human mind, and from this point of view a question about numbers is as important as a question about the system of the world.” Jacobi died of smallpox on February 18, 1851.
Jordan elimination, is not named after Camille Jordan, but rather surveyor Wilhelm Jordan (1842–99). W. Jordan used this elimination method to properly analyze measurement errors. (The “Jordan canonical form” of a matrix, however, is due to Camille Jordan.) See also JORDAN CURVE THEOREM.
Jordan curve theorem In 1893 French mathematiJiuzhang suanshu (Chiuchang Suanshu) See CHINESE MATHEMATICS.
Jordan, Marie Ennemond Camille (1838–1922) French Topology, Abstract algebra, Engineering Born on January 5, 1838, in Lyon, France, Camille Jordan is remembered for his work on PERMUTATION groups and the theory of equations. He revived interest in the important work of ÉVARISTE GALOIS and extended many of his ideas and offered substantial new contributions. Jordan’s later work on ANALYSIS includes the famous theorem that now bears his name. Jordan trained and worked throughout his life as an engineer. His interests in GROUP THEORY were motivated by the mathematical study of crystal structures. In many respects he defined this field of study with the publication of his 1870 text Traité des substitutions et des equations algebraique (Treatise on permutations and algebraic equations), the first book ever published on the topic. This work stimulated significant further mathematical research on the subject for the entire century that followed. Jordan was awarded the Poncelet Prize from the Académie des Science for the piece. After the release of his famous work Jordan continued work in algebra, establishing important results in the fields of LINEAR ALGEBRA, NUMBER THEORY, DIFFERENTIAL EQUATIONS, mechanics, and FOURIER SERIES. Work in these latter fields led him to the topics of TOPOLOGY and analysis, and the study of continuous groups, which occupied the final part of his career. Jordan died in Paris, France, on January 22, 1922. From 1961 to 1964 René Garnier and Jean Dieudonné collated all of Jordan’s work and released the fourvolume series Oeuvres de Camille Jordan (The work of Camille Jordan). One point of confusion should be settled. Although Jordan did work in the field of matrix algebra, the method of GAUSSIAN ELIMINATION, often called Gauss
cian MARIE ENNEMOND CAMILLE JORDAN stated a fundamental result: Any loop drawn in the plane that does not intersect itself divides the plane into two distinct parts: an inside and an outside. This utterly obvious theorem, now called the Jordan curve theorem, is surprisingly difficult to prove. Its validity says something significant about the geometry and structure of the surface on which the curve is drawn. For example, circles drawn on a plane or a sphere certainly divide those surfaces into two parts, but this need not be the case for a circle drawn on a TORUS, for example: a “loop” that circumnavigates the donut does not separate the surface into two distinct pieces. One must first develop a clear understanding of the geometry of space before results such as the Jordan curve theorem can be proved. This is a subtle and difficult issue. American mathematician Oswald Veblen (1880–1960) was the first to prove the theorem valid.
Josephus problem This famous mathematical puzzle is based on the difficult story of Josephus Flavius, a Jewish military leader fighting against the Romans in the town of Jotapata during the first century C.E. It is said that the battle, which took place in the spring of 67, was lost. Rather than submit their fates to the hands of the Romans, Josephus’s men took brief refuge in a cave, where they voted to commit group suicide. Although Josephus did not agree with this plan, he offered no open opposition to it. Instead, he suggested that they proceed in as orderly a fashion as possible. He proposed that they sit in a circle and humanely kill every third man until only one person was left. That person would then commit suicide. Josephus, it is said, cleverly positioned himself so that he would be the final survivor. Rather than commit suicide, he surrendered to the Romans.
jugfilling problem 289 This story leads to the following general mathematical problem: Suppose N participants, numbered 1, 2, …, N, sit in a circle and begin counting off every nth person. Each person so selected is called “out” and leaves the game. Counting continues until the last participant is declared the winner of the game. Given the size of the group, is it possible to predict beforehand who the winner will be? Assume the count begins with player 1. The following table shows the position of the winner, counting every third person, for games involving N = 1 through N = 15 people: N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 W(N) 1 2 2 1 4 1 4 7 1 4 7 10 13 2 5 The numbers arising can be analyzed as follows: A game with N players becomes a game with N–1 players as soon as the first person is called out, except that the start of the N–1 game is moved three places along. Thus, if W(N) denotes the winner of an Nplayer game, we have W(N) = W(N–1) + 3. One may need to adjust this formula, however, to take into account possible “counting around the full circle.” For example: Clearly W(1) = 1. Adding 3 gives W(2) = 4. But, counting around the circle, we see that position four in a twoplayer game is really player 2. Thus W(2) = 2. Adding three gives W(3) = 5. But, counting around the circle, we see that position five in a threeplayer game is really player 2. Thus W(3) = 2, and so on. In this way we can compute all the entries of the table without having to perform the game. Other versions of the game can be analyzed similarly.
Jourdain’s paradox In 1913 French mathematician Philip Jourdain proposed a variation of the famous LIAR’S PARADOX, now sometimes called the card paradox. On one side of a card is printed: “The statement on the other side of this card is true,” and on the
reverse side: “The statement on the other side of this card is false.” Neither sentence can be true or false, for the statement on the reverse side implies the opposite. It is worth pointing out that if, instead, both sentences read: “The statement on the other side of this card is false,” then no paradox occurs; it is simply the case that just one of the sentences is actually false, and the other is true. If, on the other hand, both sentences read: “The statement on the other side of this card is true,” then both statements could be true, or both could be false. See also SELFREFERENCE.
jugfilling problem A famous category of decanting problems contains puzzlers of the following ilk: Given a 3gallon jug and a 5gallon jug (without any markings), is it possible to draw exactly 1 gallon of water from a well? As one is not given the means to measure the exact contents of a partially filled jug, we are left with only three allowable maneuvers: 1. Completely fill an empty jug from the well 2. Completely empty a full jug into the well 3. Pour water from one jug into another, completely filling or emptying one jug in the process This particular problem has an easy solution: completely fill the 3gallon jug and pour its contents into the 5gallon jug. Refill the 3gallon jug and pour part of its contents to fill up the 5gallon jug. This leaves (and we are certain of this) precisely 1 gallon of water in the 3gallon jug and 5 gallons of water in the second. Now empty the 5gallon jug. In this solution the 3gallon jug was filled twice and the 5gallon jug emptied once. If we count +1 each time a jug is filled and –1 for each time it is emptied, the solution described above can thus be represented by the equation: 2 × 3 + (–1) × 5 = 1 Surprisingly any solution to the DIOPHANTINE EQUATION x × 3 + y × 5 = 1 corresponds to a solution to the jugfilling problem. (For example, (–3)×3 + 2×5 = 1 represents a solution in which the 5gallon jug is
290 Julia set completely filled twice and the 3gallon jug completely emptied three times.) Since 1 is the GREATEST COMMON FACTOR of 3 and 5, the EUCLIDEAN ALGORITHM provides solutions to the equation x × 3 + y × 5 = 1, and hence to the jugfilling problem. As a variation we ask: is it possible to obtain exactly 1 gallon of water using a 9gallon jug and a 15gallon jug? This would require finding a solution to the equation x × 9 + y × 15 = 1. If there were a solution to this problem, then 1 would be a combination of two
multiples of 3, and so itself a multiple of 3. This, of course, is absurd, and there is no solution to this problem. This type of argument can be used to show that d gallons of water can be obtained from an agallon jug and a bgallon jug if, and only if, d is a multiple of the greatest common factor of a and b. See also BICONDITIONAL.
Julia set See FRACTAL.
K alKashi, Jamshid Mas’ud (ca. 1380–1429) Iranian
The Treatise on the Chord and Sine, alKashi presented tables of sine values and listed a value of sin(1°) accurate to 16 decimal places. This allowed for extremely accurate astronomical calculations. On a purely mathematical note, alKashi also worked on the TRISECTING AN ANGLE problem, noting that the obstacle behind it reduces to one of solving a cubic equation. Applying iterative techniques, alKashi was able to provide an innovative approach to the problem, one that would find approximate solutions to any prescribed degree of accuracy. European scholars working on the same problem throughout the Middle Ages and the Renaissance were unaware of alKashi’s work.
Arithmetic, Astronomy Born in Kashan, Iran, alKashi is best remembered in mathematics for his early use of decimals to represent fractions and to approximate real numbers, two centuries before European scholars developed the same technique. His famous text, The Key to Arithmetic, completed March 2, 1427, describes all the mathematics necessary for those studying astronomy, surveying, architecture, and accounting. AlKashi developed iterative techniques for extracting nth roots and for solving CUBIC and QUARTIC EQUATIONs, and gave an estimate for the value of π correct to 16 decimal places. (This far surpasses the achievement of the ancient Greek and Chinese scholars who approximated π as a fraction correct, in decimal notation, to six decimal places.) That alKashi and his contemporaries chose to represent quantities as decimals indicates an important and sophisticated shift in the historical understanding of “number.” For instance, writing π explicitly as a decimal, just as any number can be so written, changed the status of π from a physical geometric construct (the ratio of the circumference of a circle to its diameter) to an ordinary member of the class of REAL NUMBERS. AlKashi adopted a perspective where all of mathematics was thus unified by the study of the realnumber system. AlKashi was very interested in astronomy and viewed much of his mathematical work as a tool for assisting astronomical calculations. Techniques of TRIGONOMETRY played a prominent role. In his text
Kendall’s method See RANK CORRELATION.
Kepler, Johannes (1571–1630) German Astronomy, Calculus Born on December 27, 1571, in Württemberg, Germany, Johannes Kepler is best remembered for his three laws of planetary motion: that planets travel in elliptical orbits with the Sun at one focus; that they do so at a rate so that the line joining the planet to the Sun sweeps out equal areas in equal units of time; and that the square of the period of a planetary orbit is proportional to the cube of the width of the ELLIPSE it traverses. These laws were all based on detailed observation of the orbit of Mars. As a mathematician, Kepler is noted for giving the first clear proof of how LOGARITHMs work, for his study of the 291
292 Kepler, Johannes geometry of the PLATONIC SOLIDs, for his discovery of two nonconvex regular polyhedra, and for his mathematical treatment of the closepacking properties of SPHEREs. (He also explained why the honeycomb shape is the most efficient design for dividing a planar region into separate cells.) More importantly, Kepler devised a method for computing the volumes of many SOLIDs OF REVOLUTION with the aid of INFINITESIMALs. Today this is seen as a significant contribution to the development of CALCULUS. Kepler studied astronomy and theology at the University of Tübingen, Germany. At the time, only six planets were known to astronomers, and all were assumed to be in circular orbit about the Sun. In 1596 Kepler published Mysterium cosmographicum (Mys
Johannes Kepler, an eminent astronomer of the 17th century, is noted for his introduction of three laws of planetary motion. These laws provided Sir Isaac Newton the inspiration to develop his inversesquare law of gravitation. (Photo courtesy of the Science Museum, London/TophamHIP/The Image Works)
tery of the cosmos), in which he presented a mathematical theory explaining the relative sizes of the planets’ orbits. Convinced that God had created the universe according to a mathematical plan, Kepler posed that if a sphere were drawn about the path of Saturn and a CUBE inscribed in this sphere, then the orbit of Jupiter lies on a sphere inscribed in this cube. Moreover, inscribing a TETRAHEDRON in this second sphere and a sphere within the tetrahedron captures the orbital path of Mars. Continuing in this way, with a dodecahedron between Mars and Earth, an icosahedron between Earth and Venus, and an octahedron between Venus and Mercury, Kepler produced a model for orbit sizes that is accurate to within 10 percent of observed values, well within experimental error. As there are only five Platonic solids, this model also explained why, supposedly, there were only six planets. Of course Kepler’s Platonic model of the solar system is not correct—three more planets were later discovered and, as Kepler himself later established, no orbit of a planet is circular. Kepler moved to Prague near the turn of the century to work with one of the foremost astronomers of the time, Tycho Brahe (1546–1601). When Brahe died, Kepler succeeded him as imperial mathematician. Brahe had kept extensive records on the orbit of Mars, and from them, Kepler was forced to conclude that its orbit was an ellipse. He also noted, from the data, that the velocity of the planet altered in such a way that the line connecting the Sun to the planet swept out equal areas in equal times. These two laws, when extended to all planets, are today called Kepler’s first two laws. He published them in his 1609 piece Astronomia nova (New astronomy). Ten years later he added a third law: the squares of the times taken by the planets to complete an orbit are proportional to the cubes of the lengths of the major axes of their elliptical orbits. Kepler had no explanation as to why these laws were true other than the compelling evidence of the data. It was not until SIR ISAAC NEWTON (1642–1727) formulated his famous law of gravitation that Kepler’s laws could be mathematically deduced. During his marriage ceremony in 1613 to his second wife Susanna (his first wife Barbara died in 1611), Kepler noticed that the servants would measure the volume of a wine barrel by slipping a rod diagonally through the bunghole and measuring the length that fit. He began to wonder why this method worked. This led him to his study of the solids of revolutions and the
alKhw–arizm–ı, Muhammad ibn M –us–a 293 publication of his famous 1615 piece Nova stereometria doliorum (New stereometry of wine barrels). This work was later developed further by BONAVENTURA CAVALIERI (1598–1647). In 1614 JOHN NAPIER published his account of logarithms, a mathematical device he invented with the aim of helping astronomers with their large numerical calculations. Kepler was delighted to learn of Napier’s work and helped provide a mathematical explanation of Napier’s method. Kepler also published tables of logarithmic values accurate to eight decimal places. Kepler died on November 15, 1630, in Regensburg, Germany. His second law of planetary motion played a crucial role in shaping Sir Isaac Newton’s thinking in the development of his theory of mechanics. Although Newton made no direct mention of Kepler’s work in his early writings, he credited Kepler as the source of inspiration for his ideas in a lecture he gave to the members of the ROYAL SOCIETY before the release of his famous 1687 work Principia.
– –ı, Muhammad ibn Mus – –a (ca. 780–850) alKhwarizm
– Arab Algebra Born in Khw arizm, (now Khiva), Uzbe– a– alKhwarizm – –ı kistan, Arab scholar Muhammad ibn Mus is remembered for writing two extraordinarily influential texts. The first, Aljam’ w’altafriq ib hisab alhind (Addition and subtraction in Indian arithmetic), introduced the Indian system of numerals to the West, and a second work that sparked the development of algebra. The title – of his second work, Hisab aljabr w’almuq abala (Calculation by restoration and compensation), when translated into Latin, produced the very word algebra we use today. AlKhw–arizm–ı was a scholar at the House of Wisdom in Baghdad, a learning academy and library founded by the Caliph alMamun in 813. (This institution was the first library constructed anywhere in the world since the destruction of the famous Library of Alexandria.) Housing a large collection of Greek philosophical and scientific texts, the library hired Arab scholars to translate manuscripts into Arabic and to conduct further research in the topics covered. – –ı was one such scholar who took particular AlKhwarizm interest in the works on mathematics and astronomy. In his first famous pieces, Aljam’ w’altafriq ib hisab alhind, alKhw–arizm–ı described and explained the advantages of the decimalplace system as used in India for writing numbers and doing arithmetic. This
work was translated into Latin some 300 years later and became the primary source for Europeans who wanted to learn the new numeration system. It is interesting to note that although this work uses a symbol for zero (in Arabic called sifr) alKhw–arizm–ı did not regard it as a number and used it only as a placeholder (to distinguish 203 from 23, for instance). AlKhw–arizm–ı intended that his second famous piece, Hisab aljabr w’almuq –abala, also be used as a teaching guide. It aimed to offer an array of techniques and methods for solving very practical problems in matters of trade, inheritance, law, surveying, and architecture. Beginning at an extraordinarily elementary level, alKhw–arizm–ı first defined the natural numbers and the act of counting to then move on to the main topic of the first section of the piece, namely, solving elementary linear and quadratic equations. It is important to note that all of alKhw–arizm–ı’s mathematics was done entirely in words, and no symbols were ever used. He called a number a “unit,” an unknown a “root,” and a quantity squared a “square.” To solve equations alKhw–arizm–ı used two operations which he called aljabr and almuq –abala. The first, “completion,” is the process of removing any negative terms from an equation. In modern notation, this converts x2 – 5x = 3 to x2 = 3 + 5x, for instance. The second, “balancing,” is the process of subtracting positive terms of the same power when they occur on each side of an equation. In modern notation, this means rewriting 5x2 + 7 = 3x2 + x, for instance, as 2x2 + 7 = x. With these two maneuvers, alKhw–arizm–ı reduces all linear and quadratic equations into six basic types, which he lists. The remainder of this part of the text is devoted to solving these equations. Today’s standard practice of COMPLETING THE SQUARE is described in this section. The next part of alKhw–arizm–ı ’s text consists of applications and worked examples. Because many of the problems were geometric in nature, alKhw–arizm–ı would dismiss negative solutions to problems and allow only positive answers. He recognized that quadratic equations may yield two solutions to a problem. AlKhw–arizm–ı also computed areas of simple geometric figures, as well as the volume of a sphere, the cone, and the pyramid, in this section. The final section of the text focuses on the complicated Islamic rules for inheritance and does not develop any new mathematical content.
294 Klein, Felix Christian By the 10th century, Spain was under Islamic control, and European scholars interested in the intellectual culture of the Islamic world traveled to Spain to study the Arabic texts. It was there that AlKhw–arizm–ı ’s works were discovered and translated into Latin. AlKhw–arizm–ı also wrote texts on astronomy, the sundial, the Jewish calendar, and on geography. He computed the latitudes and longitudes of over 2,400 specific localities in preparation for the construction of an accurate world map. The word algorithm is believed to be derived from alKhw–arizm–ı ’s name. Medieval European scholars, attempting to translate the Arab scholar’s name into Latin, called the practice of using HinduArabic numerals “algorism,” from which, in turn, any general practice or procedure became known as an “algorithm.” That a version of his name became part of our Western vocabulary illustrates the extent to which alKhw–arizm–ı ’s work influenced the development of arithmetic and algebra in Europe.
Klein, Felix Christian (1849–1925) German Abstract algebra, Geometry, Topology Born on April 25, 1849, scholar Felix Klein is remembered for uniting the disparate fields of GEOMETRY and ALGEBRA through the study of GEOMETRIC TRANSFORMATIONs. He showed that the abstract analysis of the algebra of transformations through GROUP THEORY leads to a clear understanding of both EUCLIDEAN GEOMETRY and NONEUCLIDEAN GEOMETRY. In particular, he proved that ordinary Euclidean geometry can be proved CONSISTENT, that is, free of contradiction, if, and only if, nonEuclidean geometry is consistent. This demonstrated, for the first time, that the controversial nonEuclidean theories of geometry were of equal importance to the theory of ordinary Euclidean geometry. In topology, the KLEIN BOTTLE is named in his honor. Klein received a doctorate in mathematics from the University of Bonn in 1868 after completing a thesis exploring applications of geometry to mechanics. After a short period of military service, Klein accepted a position as professor of mathematics at Erlangen, in Bavaria, in 1872. It was here that he began his work on the analysis of geometric transformations and the unification of Euclidean and nonEuclidean geometries. His approach to the subject with a focus on “group invariants” was extremely influential, and
work in this field continues today. (It is called Klein’s “Erlangen program.”) In 1886 Klein accepted the position as department chair at the University of Göttingen, where he remained until his retirement in 1913. He tutored students from his home in the years that followed. In addition to conducting highlevel research in mathematics, Klein also wrote mathematical pieces intended for the general public and founded a mathematical encyclopedia that he supervised until his death. Klein was honored with election to the ROYAL SOCIETY in 1885 and received the Copley Medal from the Society in 1912. He died on June 22, 1925, in Göttingen, Germany. See also KLEINFOUR GROUP.
Klein bottle The MÖBIUS BAND is a threedimensional object possessing just one surface and just one edge. A SPHERE and a TORUS (donut shape), on the other hand, are objects with two surfaces—an outside surface and an inside surface—possessing no edges. The Klein bottle is an alternative threedimensional object with only one surface, akin to that of a Möbius band, but possessing no edges. One can easily model these surfaces mentioned with a pair of trousers. For example, sewing together the two leg openings produces a circular tube with a hole for the waist. This hole can be patched with a piece of material to produce a complete model of a torus. A Klein bottle is produced by bringing one trouser leg up, over, and through the waist of the trousers and pushing it down through the tube of the second leg before sewing the two leg openings together. Again, the hole for the waistband represents a hole in the surface, but the object produced is nonetheless a (punctured) Klein bottle. Unfortunately it is no longer physically possible to patch the hole with a piece of material. This shows that the Klein bottle does not properly exist in a threedimensional universe. (One can imagine, however, that the material that would make a patch for the hole of the waist has been “plucked” up into the fourthdimension. In this sense, the Klein bottle is a valid mathematical object in fourdimensional space.) One can check with this model that the ideal surface of a Klein bottle would indeed be onesided: an ant crawling on one side of trouser material could reach any other part of the trousers, on either side of
knot theory 295 the material, without ever cheating by crossing over the edge of the hole given by the waist band. If one cuts a pair of trousers sewn to make a (punctured) Klein bottle in half along the line that follows the inner and outer seams of the trousers, then the model falls into two pieces, each of which is easily seen to have arisen from a Möbius band. This shows that a Klein bottle can also be thought of as the union of two Möbius bands adjoined along their edges. (If one physically attempts to sew together two Möbius bands, one soon finds it is necessary to puncture a band to produce a hole akin to the hole for the waist of a pair of trousers.)
Kleinfour group (viergruppe) There are essentially only two GROUPs with just four elements. The first is the “cyclic group” C4 given by the rotational symmetries of a square, C4 = {1, R, R2, R3}, where R represents a 90° rotation of a square in a clockwise direction and 1 is the IDENTITY ELEMENT, a rotation of zero degrees. As a rotation of 360° is equivalent to no action, we have that R4 = 1. Multiplication in this group is given by the product rule of exponents. We have, for example, that R2 × R3 = R5 = R. The second group with just four elements is called the Kleinfour group. Denoting its elements as 1, a, b, and c, with 1 being the identity element, it has a multiplication table given by: × 1 a b c
1 1 a b c
a a 1 c b
b b c 1 a
c c b a 1
This group can also be represented by a set of symmetries of a square. Set a to mean a reflection about a vertical axis, b a reflection about a horizontal axis, and c a rotation of 180°. We have, for instance, that a vertical reflection followed by a rotation of 180° is the equivalent of a horizontal reflection (c × a = b), and that two reflections about the same axis lead to the identity (a × a = b × b = 1).
knot theory The branch of
that studies the properties of closed loops embedded in threeTOPOLOGY
The two knots with fewer than five crossings
dimensional space is called knot theory. Each loop studied, called a knot, represents the path traced by a piece of string interlaced in space (without selfintersection) whose free ends have been joined together. If one tangled piece of string can be physically transformed into an exact copy of a second tangled string (or at least a mirror image of that second tangle), then we say that the two knots the strings represent are equivalent. At present, no one knows a general procedure guaranteed to determine with relative ease whether or not two given knots are equivalent. Knots can be depicted on a twodimensional page as a picture of a loop that crosses over and under itself. It is conventional to draw the picture of a given knot with as few crossings as possible. A picture with just one or two crossings is equivalent to an untangled loop (the “unknot”). There is only one knot (up to equivalence) with three crossings and only one with four crossings. Each are shown above. There are two distinct knots requiring a minimum of five crossings, three requiring a minimum of six crossings, and seven requiring a minimum of seven crossings. After this, the number of distinct knots with a given minimum number of crossings when drawn on a page grows rapidly. Knots were first studied by CARL FRIEDRICH GAUSS (1777–1855) and his student Johann Listing (1808–82). In 1877 Scottish physicist Peter Tait classified all knots with up to seven crossings. He also conjectured that no “alternating knot” (that is, one whose path alternately crosses over and under itself) is equivalent to the unknot. A century later, New Zealand mathematician Vaughn Jones proved him to be correct. In 1928 American mathematician James Waddell Alexander discovered a theoretical means to associate to each knot a POLYNOMIAL in such a manner that if two
296 Koch curve knots are equivalent, then the polynomials associated to them turn out to be the same. This thus provided an algebraic means to distinguish knots: if two knots have different Alexander polynomials, then they are not equivalent. In the 1960s British mathematician John Horton Conway discovered an explicit method for implementing Alexander’s polynomials, and in 1984 Vaughn Jones developed an alternative (simpler) choice of polynomial to associate with each knot. (It was with the “Jones polynomial” that Jones was able to establish Tait’s claim.) The same year, eight mathematicians independently discovered another polynomial that generalizes Jones’s approach. (This polynomial is called the HOMFLYPT polynomial—an acronym of their names.) Unfortunately, it is not yet known whether Jones’s technique provides a perfect correspondence between distinct knots and distinct polynomials—it is feasible that two nonequivalent knots could produce the same Jones’s polynomial. (No one has yet seen an example of this.) Knot theory is an active area of current research. It is used in the practical study of complicated molecules such as DNA, in the analysis of electrical circuits, and in the planning of street and highway networks. Although the mirror image of a knot is deemed equivalent to the original knot, mathematicians have proved that no knot (except the unknot) can be physically deformed into its mirror image. It is also interesting to note that no knots exist in twodimensional space or fourdimensional space (meaning that any knot in these spaces can be deformed into the unknot). See also BRAID.
Koch curve See FRACTAL.
Kolmogorov, Andrey Nikolaevich (1903–1987) Russian Probability theory, Analysis, Dynamical systems Born on April 25, 1903, in Tambov, Russia, scholar Andrey Kolmogorov is remembered for his significant contributions to the field of PROBABILITY theory and his groundbreaking work in the study of DYNAMICAL SYSTEMs. He also contributed to the study of FOURIER SERIES, ANALYSIS, and TOPOLOGY. Applying his work to planetary studies, Kolmogorov was also able to establish the mathematical stability of our solar system. Kolmogorov began his working career as a railway conductor. In his spare time, however, he studied math
ematics and physics, and managed to write a treatise on SIR ISAAC NEWTON’s laws of mechanics. In 1920 he entered Moscow State University. By the time he completed his undergraduate degree he had published eight influential papers on the topics of SET THEORY and analysis. Kolmogorov published another 10 papers before receiving his doctoral degree in 1929. Two years later Kolmogorov was appointed a professor of mathematics at Moscow State University. At this time, he began work on writing Grundbegriffe der Wahrscheinlichkeitsrechnung (Foundations on the theory of probability), in which he attempted, successfully, to build up the entire theory of probability from a finite set of axioms using nothing but logical rigor. Kolmogorov received national fame when the work was published in 1933. The piece was translated into the English language by Nathan Morrison in 1950. In 1938 Kolmogorov was appointed department head of probability and statistics at the newly established Steklov Mathematical Institute in Russia. Soon afterward Kolmogorov’s attention turned to the study of turbulence, and in 1941 he wrote two important papers on the nature of turbulent airflow from jet engines, laying down the founding principles of the theory of dynamical systems. He later applied this work to the study of planetary motion. Kolmogorov received many awards for his outstanding work. In 1939 he was elected to the U.S.S.R. Academy of Science, and over the following years, he received eight prizes from the state and the academy. He was also elected to a number of foreign academies, including the Royal Statistical Society of London in 1956, the American Academy of Arts and Sciences in 1959, the Netherlands Academy of Sciences in 1963, the ROYAL SOCIETY of London in 1964, and the French Academy of Sciences in 1968. Kolmogorov was also awarded the Balzan International Prize in 1962. Kolmogorov died in Moscow, Russia, on October 20, 1987. His work paved the way for continued research in the fields of Fourier analysis, “Markov chains” in probability theory, and topological analysis.
Kronecker, Leopold (1823–1891) German Number theory Born on December 7, 1823, in Liegitz, Prussia (now Legnica, Poland), mathematician and philosopher Leopold Kronecker is noted as a scholar for his considerable contributions to the field of NUMBER THEORY. As
Kruskal’s count 297 a philosopher, he is also remembered as the first to cast doubts on the existence of TRANSCENDENTAL NUMBERs and on hierarchies of infinite sets, notions first put forward by KARL THEODOR WILHELM WEIERSTRASS (1815–97) and GEORG CANTOR (1845–1918). At one point, Kronecker even went as far as to deny the existence of IRRATIONAL NUMBERs. Kronecker entered Berlin University in 1841 and received a doctoral degree in mathematics after completing a thesis on the topic of algebraic number theory 4 years later. Rather than pursue an academic career, Kronecker decided to return to Liegitz to help with the family banking business. Over the 10 years that followed, he became quite wealthy and then had the luxury to pursue mathematics on his own without the need to ever accept a university position. In 1855 he returned to Berlin and published a series of influential papers in quick succession, which garnered him national attention. In 1861 Kronecker was elected as a member of the prestigious Berlin Academy, which earned him the right to teach at Berlin University, even though he was not a faculty member. (In 1883, some 20 years later, Kronecker was awarded an official position with the university.) Kronecker was also elected to the Paris Academy in 1868, and to the ROYAL SOCIETY of London in 1884. During the 1870s Kronecker took a keen interest in the new, indirect approaches mathematicians were using to prove the existence of certain types of numbers. But he soon came to feel that such nonconstructive practices were philosophically flawed. For instance, Cantor had demonstrated that the set of real numbers is “more infinite” than the set of ALGEBRAIC NUMBERs, and hence, numbers that are not algebraic (socalled transcendental numbers) must exist. In Kronecker’s view, mathematics could only ever be based on finite quantities with a finite number of operations applied to them. An argument such as Cantor’s, therefore, in Kronecker’s thinking, had no meaning. Even when CARL LOUIS FERDINAND VON LINDEMANN (1852–1939) proved in 1882 that the specific number π is transcendental, Kronecker complimented Lindemann on his beautiful proof but added that, in fact, he had accomplished nothing, since transcendental numbers do not exist. Kronecker was adamant in his views and, in his 1887 piece Über den Zahlbergriff (On number theory), he attempted to persuade the mathematical community of the absolute necessity of only using direct and finite
techniques. He is famous for once having remarked that “God created the integers. All else is the work of man” as an attempt to bring mathematical thinking back to concrete principles. Both Weierstrass and Cantor felt under personal attack by Knonecker and thought that he was deliberately undermining their own research programs. Although many mathematicians today would not agree with Kronecker’s views, they were taken seriously at the time, and the ideas that Kronecker put forward were later expanded upon in the century that followed. The practice of using only direct and finite approaches to prove results in mathematics is today called constructivism. Kronecker died on December 29, 1891, in Berlin, Germany.
Kruskal’s count In the early 1980s, Princeton physicist Martin Kruskal discovered a remarkable mathematical property that all passages of written text seem to possess. This phenomenon is now referred to as Kruskal’s count. To illustrate the principle, review, for example, the familiar nursery rhyme: Twinkle twinkle little star, How I wonder what you are, Up above the world so high, Like a diamond in the sky. Twinkle twinkle little star, How I wonder what you are. Perform the following steps: 1. Select any word from the first or second line and count the number of letters it contains. 2. Count that many words forward through the passage to land on a new word. (For example, choosing the word star, with four letters, will transport you to the word what.) 3. Count the number of letters in the new word, and move forward again that many places. 4. Repeat this procedure until you can go no further (that is, counting forward will take you off the nursery rhyme). 5. Observe the final word on which you have landed. Surprisingly, no matter on which word you start this counting task, the procedure always takes you to the same word in the final line, namely, the word you.
298 Kruskal’s count Kruskal observed that this same phenomenon seems to occur with any sufficiently large piece of text—counting forward in this way from any choice of beginning word lands you at the same place at the end of the page. This provides an amusing activity for several people to perform simultaneously, all working with the same text, but starting with different choices of initial word. The phenomenon can be explained as follows: imagine that on a first run through the text, all the words encountered are circled. Landing on any one of these circled words in a subsequent run of the experi
ment will take the player to the same final word. What then is the likelihood that a second run through the experiment will not “hit” any of the previously circled words? If the passage is sufficiently large, say involving 20 steps, the chances of “missing” a circled word every time is likely to be very small, especially since the typical English passage contains reasonably short words. Estimates computing this PROBABILITY show that it likely has a value smaller than 0.1 percent. Thus, with about 99.9 percent certainty, two paths through a passage of text will coincide at some point.
L Lagrange, JosephLouis (1736–1813) ItalianFrench
Paris and identified himself as French. He was introduced to the topics of physics and mathematics at the College of Turin. He first studied the principles of the TAUTOCHRONE and, by 1754, had published some important discoveries about the topic. He shared the results with LEONHARD EULER (1707–83), who was suitably impressed. At the young age of 19, Lagrange was appointed professor of mathematics at the Royal Artillery School in Turin. There he made significant progress on the general study of mechanics and wrote a number of influential papers. In 1756 he was elected to the Berlin Academy under Euler’s recommendation. During the 20 years that Lagrange stayed at the Berlin Academy, he worked on a wide variety of topics. He lay down the foundations of a subject to become known as “dynamics” and made significant contributions to the studies of fluid mechanics, to the integration of differential equations, and to the solution of systems of differential equations, as well as to the study of vibrating strings and the propagation of sound. In 1766 he succeeded Euler as director of mathematics at the Berlin Academy, and in 1772 he shared one of the biennial prizes from the Paris Académie des Sciences with Euler for his work on the “threebody problem” from physics. He won a prize again from the Paris Académie in 1774 for his work on the motion of the moon, and again in 1780 for his mathematical study of planetary motion. In 1787 Lagrange left Berlin to take a position at the Académie des Sciences in Paris. He published his work Mécanique analytique the following year. He also
Analysis, Mechanics, Abstract algebra, Number theory Born on January 25, 1736, in Turin, Italy, JosephLouis Lagrange is remembered for, among other things, his definitive 1788 text Mécanique analytique (Analytical mechanics), in which he lays out a purely formal and completely rigorous exposition of how and why things move. In effect, the work summarizes the entire state of the postNewton mechanics of the 18th century. The publication is also famous for featuring no figures or diagrams. (Lagrange apparently felt that the formulae he developed captured physical reality sufficiently well that pictures were not necessary.) He also published two important memoirs in 1769 and 1770, which he combined into a single 1798 treatise, Traité de la résolution des équations numériques de tous les degrés (Treatise on the resolution of numeric equations of all degree), on the theory of equations, in which he developed general principles for solving POLYNOMIAL equations up to degree four via principles that we today would label GROUP THEORY. The famous “Lagrange’s theorem” from the study of groups appears in this work, as does the famous INTERPOLATION formula that bears his name. He also made significant contributions in the field of NUMBER THEORY, proving in 1770, for instance, that every positive integer is the sum of four square numbers. Also, in 1771, he was the first to prove what is today called “Wilson’s theorem,” that a number n is PRIME if, and only if, (n – 1)! + 1 is divisible by n. Although Lagrange was born in Turin and spent the early part of his life there, he eventually settled in 299
300 Lagrange’s formula worked with the Paris Académie to standardize the weights and measures of the day. The committee formulated the metric system and advocated the general use of a decimal base. Lagrange’s significant achievements were recognized by the emperor Napoleon in 1808, when he was named to the Legion of Honour and a count of the empire. Five years later he was also named Grand Croix of the Ordre Impérial de la Réunion. He died in Paris, France, on April 10, 1813. Lagrange’s impact in mathematics, especially in mathematical physics, is still felt today. Many fundamental concepts in mechanics and multivariable calculus—such as the Lagrangian (the difference between kinetic energy and potential energy of a set of particles), the Lagrangian description (a measure of deformation of a physical body), and Lagrange multipliers in calculus—play a vital role in the current study of these subjects. See also LAGRANGE’S FORMULA.
Lagrange’s formula (Lagrange’s interpolation formula) Given a collection of points on the plane, it is sometimes desired to find the formula for a function that passes through each of those points. For example, a scientist may seek a formula for a function that fits all the data values obtained from an experiment. In the late 1700s, ItalianFrench mathematician JOSEPHLOUIS LAGRANGE suggested the following INTERPOLATION formula: If (a1,b1), (a2,b2), …, (an,bn) are a collection of points in the plane, with the values ai distinct, then f (x) = b1
(x − a2 )(x − a3 )L(x − an ) (a1 − a2 )(a1 − a3 )L(a1 − an )
+ b2
(x − a1)(x − a3 )L(x − an ) (a2 − a1)(a2 − a3 )L(a2 − an )
+ L + bn
(x − a1)(x − a2 )L(x − an−1) (an − a1)(an − a2 )L(an − an−1)
is a POLYNOMIAL, of degree n – 1, that passes through each of the points. This formula is today known as Lagrange’s formula. One can check that it works by substituting x = a1 to see that f(a1) = b1, and so on.
As an example, consider the points (1,2), (2,5), and (3,1) in the plane. Lagrange’s formula shows that the quadratic f (x) = 2 =−
(x − 2)(x − 3) (x − 1)(x − 3) (x − 1)(x − 2) +5 +1 (1 − 2)(1 − 3) (2 − 1)(2 − 3) (3 − 1)(3 − 2) 7 2 27 x + x −8 2 2
passes through each of them. Many intelligence tests ask participants to identify “the next number in the sequence.” Lagrange’s formula provides a means for justifying absolutely any answer to such a question. For example, the next number in the sequence 2,4,6,… could well be 103. We can argue that the sequence follows the formula: 1057 95 3 an = n − 95n 2 + n − 95 . (Apply Lagrange’s 6 6 formula to the points (1,2), (2,4), (3,6), and (4,103).)
Lambert, Johann Heinrich (1728–1777) SwissGerman Geometry, Analysis, Number theory, Physics Born on August 26, 1728, scholar Johann Lambert is best remembered as the first to prove, in 1761, that π is an IRRATIONAL NUMBER. He also worked on Euclid’s PARALLEL POSTULATE and came close to the discovery of NONEUCLIDEAN GEOMETRY. Lambert also developed the notation and the theory of HYPERBOLIC FUNCTIONS. In 1766 Lambert wrote Theorie der Parellellinien (On the theory of parallel lines), in which he postulated the existence of surfaces on which triangles have angular sums less than 180°, thereby yielding an example of a geometry in which the parallel postulate would be false. (Such a surface was later discovered. It is called a pseudosphere.) Lambert proved that in this geometry, the sum of the angles of a triangle would not be constant, and in fact would increase (but never to equal 180°) as its area decreases. In 1737 LEONHARD EULER had proved that e and 2 e are both irrational. In the paper “Mémoire sur quelques propriétés remarquables des quantités transcendantes circulaires et logarithmiques” (Memoir on some remarkable properties of transcendental quantities circular and logarithmic) presented to the Berlin Academy of Sciences in 1761, Lambert provided a proof that if x is a rational number different from zero, then neither ex nor tan x can be rational. Thus Lambert
Laplace, PierreSimon, marquis de 301
Laplace, PierreSimon, marquis de (1749–1827)
In 1773 Laplace was elected to the Paris Académie des Sciences and was awarded a senior position there 12 years later. He was assigned to a special committtee of the Paris Académie in 1790 with the charge of standardizing all European weights and measures. Laplace’s work on celestial mechanics was revolutionary. In 1786 he had proved that the small perturbations observed in the orbital motion of the planets will forever remain small, constant, and selfcorrecting, and in 1796 he was the first to propose the idea that the solar system originated from the contraction and cooling of a large rotating, and consequently flattened, nebula of incandescent gas. In 1799 he published the first two volumes of his famous Mécanique céleste, in which he described the general laws of motion of solids and fluids, applied the universal law of gravity to studies of the solar system, and developed methods for analyzing the difference and differential
French Mechanics, Analysis, Differential equations, Probability theory Born on March 23, 1749, in Normandy, France, scholar PierreSimon Laplace is best remembered for his influential fivevolume treatise Traité de mécanique céleste (Celestial mechanics), published between the years 1799 and 1825. In this work Laplace tried to develop a rigorous mathematical understanding of the motion of the heavenly bodies, including the various anomalies and inequalities that were observed in their orbits. In doing so, Laplace made significant strides in the development of DIFFERENTIAL EQUATIONs, DIFFERENCE equations, PROBABILITY, and STATISTICS. He was the first to extend SIR ISAAC NEWTON’s theory of gravitation to the study of the whole solar system. In 1812 Laplace also published his Théorie analytique des probabilités (Analytic theory of probability), which advanced the topics of probability and statistics considerably. At the age of 16 Laplace entered Caen University and soon discovered a love for mathematics. Three years later, without completing his degree, Laplace moved to Paris to work with the mathematician JEAN LE ROND D’ALEMBERT (1717–83), supporting himself as a professor of mathematics at the École Militaire. By the time he was 24, Laplace had produced 13 highquality papers on the topics of INTEGRAL CALCULUS, mechanics, and physical astronomy, earning him national attention. He continued to produce fundamentally important results, and was soon regarded as one of the most influential scientists of his time.
PierreSimon Laplace, an eminent mathematical physicist of the 19th century, made fundamental contributions to the study of planetary motion by applying Sir Issac Newton’s theory of gravitation to the entire solar system. He also pioneered work in the theory of probability and statistics. (Photo courtesy of Topham/The Image Works)
had succeeded in extending Euler’s result in establishing p – that e raised to a fractional power, e q, is never rational, π π and, moreover, noting that tan – = 1, that – also cannot 4 4 be rational. Lambert also conjectured that both e and π are TRANSCENDENTAL NUMBERs. (It was not for another 100 years that, in 1873, French mathematician Charles Hermite proved the transcendence of e, and German mathematician CARL LOUIS FERDINAND VON LINDEMANN, in 1883, the transcendence of π.) Lambert also worked on the theory of PROBABILITY and, in physics, made important contributions to the study of heat and light. He died in Berlin, Germany, on September 25, 1777. Much of Lambert’s work, although significant at his time, can be seen today as having paved the way for others to achieve greater advances.
302 Latin square equations that arise in these studies. Later volumes studied the local forces acting between molecules and the largescale effect of these local interactions. He applied this work to the study of pressure and density, the nature of gravity, capillary action, sound, refraction, and the theory of heat. In 1812 Laplace published his famous piece on the topic of probability. Motivated by the analysis of errors in observation, Laplace developed the theory of probability from beginning principles and made clear the mathematical underpinnings of many aspects of the topic. He provided his own definition of probability and used it to justify the basic mathematical manipulations. The work discusses his famous LEAST SQUARES METHOD, the BUFFON NEEDLE PROBLEM, and BAYES’S THEOREM. In analyzing the distribution of errors in scientific observations, Laplace also applied this work to the accurate determination of the masses of Jupiter, Saturn, and Uranus, as well as to improving triangulation methods in surveying, and correct determination of longitude and latitude in geodesy. Under Napoleon, Laplace became a count of the empire in 1806 and was awarded the title of marquis in 1817. Laplace wrote that he believed that the nature of the universe is completely deterministic, meaning that the motions all objects in the solar system are predetermined by the initial conditions at the start of the universe. When Napoleon asked him where God fit into this view, Laplace is said to have replied: “I have no need of that hypothesis.” Laplace died in Paris, France, on March 5, 1827. It is not possible to exaggerate the influence Laplace had on the development of the mathematical theory of mechanics. A number of fundamental concepts, such as the Laplace operator in potential theory and the Laplace transform in the study of differential equations, are named in his honor.
Latin square An n × n Latin square is an arrangement of the first n Latin letters A, B, C, D, … in a square array in such a manner that no row or column of that array contains two identical letters. For example, the arrangement below is a 3 × 3 Latin square: A C B
B A C
C B A
A Latin square is called diagonal if, in addition, no letter is repeated in either diagonal of the square. The following, for example, is a diagonal 4 × 4 Latin square: A C D B
B D C A
C A B D
D B A C
No 3 × 3 Latin square is diagonal. There is just one 1 × 1 Latin square, two 2 × 2 Latin squares (neither of which are diagonal), 12 3 × 3 Latin squares, and 576 4 × 4 Latin squares. The number of Latin squares of a given order grows extremely rapidly, as the following table shows:
n
Number of n × n Latin Squares
1 2 3 4 5 6 7 8 9 10
1 2 12 576 161,280 812,851,200 61,479,419,904,000 108,776,032,459,082,956,800 5,524,751,496,156,892,842,531,225,600 9,982,437,658,213,039,871,725,064,756,920,320,000
There are more 15 × 15 Latin squares than there are atoms in the universe (which physicists calculate to be about 1081). Two Latin squares of the same order are called mutually orthogonal if all the pairs of letters that appear when the two squares are superimposed are different. For example, the two squares shown below are mutually orthogonal. Here we have used letters of the Greek alphabet for the second square. The resultant array of pairs of elements that appears is called a GraecoLatin square. A B C D
B A D C
C D A B
D α C γ + B δ A β
β δ γ α
Aα Bβ Cγ Dδ γ δ Bγ Aδ Dα Cβ α β = Cδ Dγ Aβ Bα β α Dβ Cα Bδ Aγ δ γ
This square solves the famous “officer problem” first posed by LEONHARD EULER (1707–83):
law of cosines/law of sines 303 Arrange 16 officers, each from one of four possible regiments and of one of four possible ranks, in a 4 × 4 array so that no two officers in any given row or column come from the same regiment, nor have the same rank. GraecoLatin squares are also sometimes called Euler squares. Mathematicians have proved that there is no solution to the officer problem for the case of 36 officers of six different ranks from six different regiments (nor for the case of four officers from two regiments of two ranks), but that all other versions of the officer problem do have solutions. In other words, n × n GraecoLatin squares exist for all values of n except n = 2 and n = 6. GraecoLatin squares are important in the design of experiments in scientific studies. For example, if four species of tomato A, B, C, and D are to be tested with four different fertilizers α, β, γ, and δ, the plots can be laid out according to a GraecoLatin square to be sure that each species of tomato and each fertilizer appears in each row and column.
law of averages This “law” refers to the incorrect belief that previous outcomes of independent runs of a random trial influence the outcomes of runs yet to occur. For example, after tossing seven “heads” in a row, the supposed law of averages would dictate that “tails” is now a more likely outcome. This of course is not the case. The chance of tossing tails on the eighth toss is still 50 percent. The law of averages is a common misinterpretation of the mathematically correct LAW OF LARGE NUMBERS. Gamblers often feel that after a long string of losses, the chances of winning a next hand must be considerably greater.
Establishing the law of cosines
The law can be proved by drawing an altitude from the apex of the triangle and applying Pythagoras’s theorem to the rightangled triangle containing the altitude and side c in the picture above. The law of sines asserts: a b c = = sin(A) sin(B) sin(C) This can be proved by drawing the altitudes of the triangle. For example, the altitude above is simultaneously of length asin(C) and csin(A), thereby establishing part of the law of sines. Drawing the CIRCUMCIRCLE to the triangle, and calculating the sine of angle A in the shaded triangle shown (with the same peripheral angle A) establishes: a =D sin(A)
law of cosines/law of sines (cosine rule, sine rule) Let a,b, and c be the three sidelengths of a triangle with interior angles as shown. The law of cosines asserts: c2 = a2 + b2 – 2abcos(C) (with analogous statements for angles A and B). When C is a right angle, this result reduces to a statement of PYTHAGORAS’S THEOREM. Hence the law of cosines can be regarded as a generalization of this famous result.
Establishing the law of sines
304 law of large numbers where D is the diameter of the circumcircle. (Here we make use of the fact that any angle subtended from a diameter is 90°.) Thus the three quantities expressed in the law of sines each equal the diameter of the circumcircle of the triangle, thereby offering an alternative proof of the law. See also CIRCLE THEOREMS.
law of large numbers If one tosses a fair coin 10 times, one would expect, on average, five of those tosses to be “heads.” Of course, in any single run of 10 tosses, any number of heads is possible, even a string of 10 heads in a row, but this is extremely unlikely. The number of heads actually observed in an experiment is likely to be four, five, or six, close to 50 percent. (The probability of any particular count of heads appearing is described by the BINOMIAL DISTRIBUTION.) In a run of 100 tosses we would expect the effects of excessive runs of heads, or tails, to “average out” and the proportion of heads obtained to be even closer, on average, to the “true value” of 50 percent—and closer still if we run an experiment of 1,000 or 10,000 tosses, or more. This is the law of large numbers in action. Precisely, this law states the following: The more times a random phenomenon is performed, the closer the proportion of trials in which a particular outcome occurs approximates the true probability of that outcome occurring. If, for example, a 1 never occurred when rolling a die 10 times, we can be assured, however, that the proportion of 1s appearing in another 100, 1,000, 10,000, … tosses will approach the value onesixth. Many gamblers incorrectly interpret the law of large numbers as a method for predicting outcomes of random events. (See LAW OF AVERAGES.) The law of large numbers is a mathematical consequence of CHEBYSHEV’S THEOREM. It can be interpreted as saying that if a random phenomenon produces numerical outcomes with mean value µ, then the mean of N observed values of the phenomenon approaches the value µ as N increases. Chebyshev’s theorem is related, for it gives measures of how values are distributed about the mean. See also MONTE CARLO METHOD; STATISTICS: DESCRIPTIVE.
law of sines (sine rule) See
LAW OF COSINES/LAW
OF SINES.
law of the lever ARCHIMEDES OF SYRACUSE (ca. 287–212 B.C.E.) recognized that two weights w1 and w2 placed at distances x1 and x2, respectively, from the fulcrum (pivot point) of a simple lever will balance when x1w1 = x2w2. This principle is called the law of the lever. For example, an adult weighting twice as much as a child will balance on a seesaw if she sits half the distance from the pivot point as the child. See also CENTER OF GRAVITY.
laws of thought The Greek philosopher ARISTOTLE (384–322 B.C.E.) identified three laws of logic, all tautologies (meaning that each has a TRUTH TABLE with constant value T) that have since been deemed fundamental descriptions of the way we think. His three laws of thought are: 1. Law of Noncontradiction: It is not the case that something can be both true and not true. ∨ Symbolically: ¬[p (¬p)] 2. Law of Excluded Middle: Each must either be true or not true. Symbolically: p ∨ (¬p) 3. Law of Identity: If something is true, then it is true. Symbolically: p → p Mathematicians often rely on the law of excluded middle to establish the validity of mathematical results: an INDIRECT PROOF or a PROOF BY CONTRADICTION proves that a statement p is true by showing that it cannot be false. However, not all philosophers (and mathematicians) agree with this approach and question the validity of this second law. For example, as the 20thcentury Austrian mathematician KURT GÖDEL showed, there are some statements in mathematics that can neither be proved nor disproved, and are consequently neither true nor false. The constructivist movement accepts results established by DIRECT PROOF only. To move beyond the law of the excluded middle, logicians have attempted to generalize FORMAL LOGIC to include three possible values of truthhood: true,
least squares method 305 false, and undecided. A further generalization, called FUZZY LOGIC, treats truth as a continuous quantity with values ranging from 0 (utterly false) to 1 (utterly true). As an example, in this theory the sentence: This sentence is false. is assigned a truth value of of 1/2, and is deemed half true and half false. (One arrives at this value as follows: first note that if a statement p is assigned a truthvalue v, which is either 0 or 1, then the statement ¬ p has opposite value 1 – v, 1 or 0. We assume that this remains true even if v is of fractional value. If p represents the sentence: “This sentence is false,” then its truth establishes its falsehood, p → ¬p, and its falsehood its truth, ¬p → p. The truthvalue v of the statement oscillates between the values v and 1 – v. The only stable value for v occurs when v = 1 – v, yielding the appropriate truthvalue v = 1/2.) See also ARGUMENT; SELFREFERENCE.
p(x) = anxn + an–1x +…+ a1x + a0, the coefficient an of the highest power of the variable is called the leading coefficient of the polynomial. For example, the leading coefficient of the polynomial 2x3 – 3x + 6 is 2, and that of x + 3 is 1. A polynomial is called monic if its leading coefficient is 1. In solving a polynomial equation anxn + an–1xn–1 +…+ a1x + a0 = 0, it is often convenient to assume that the polynomial in question is monic. One achieves this by dividing the equation through by an. Solving the equation 2x4 – 8x3 + 2x – 6 = 0, for instance, is equivalent to solving x4 – 4x3 + x – 3 = 0. If all the coefficients of a monic polynomial are integers, then any rational ROOT to the polynomial must itself be an integer. For example, if the fraction p x= – q (written in reduced form) were a solution to the
leading coefficient For a
POLYNOMIAL
n–1
polynomial equation x4 – 4x3 + x – 3 = 0, then, substituting in this value for x and multiplying through by q4 yields: p4 – 4p3q + pq3 – 3q4 = 0 This shows that p4 is a multiple of q. Since p and q share no common factors, this is only possible if q p equals 1. Thus the root x = – q = p is an integer.
A similar argument shows that if the CONSTANT term a0 of a polynomial is 1, then any rational root of the polynomial (if it has one) must be a fraction of the 1 form x = – q .