Mathematics and the Real World

Zvi Artstein
Prometheus Books, 2014, ISBN 978-1-61614-091-5

Zvi Artstein is a distinguished Professor of Mathematics, specialising in Control

Theory and Game Theory, at The Weizmann Institute of Science in Israel. This
book, subtitled ‘The remarkable role of Evolution in the making of Mathematics’, is
a translation from the Hebrew, so we may assume that its original target audience
was lay persons and mathematics students and teachers in Israel. Its intent is to
describe the relationship between mathematics and the physical and social world,
based on the concept that intellectual advances are driven by evolutionary pressure.
The book is a historical and philosophical account
of the development of mathematics with emphasis on
aspects such as simple arithmetic, spatial visualisa-
tion and pattern recognition which can be related
to evolution by natural selection. This is an inter-
esting approach, well worth serious consideration.
Unfortunately, the author adduces no evidence from
evolutionary anthropology, biological genetics, DNA
analysis or cognitive science to support his theories,
relying rather on folk psychology and ‘Just–So’
stories. Unsurprisingly, he deduces that most applied
mathematics but precious little pure mathematics has
an evolutionary advantage.
Interspersed in the text, which generally avoids
technicalities, are paragraphs in a different font, which
the author invites the uninterested reader to skip, and which are supposed to
contain more sophisticated mathematical arguments. Unfortunately, these sections
themselves are often historically inaccurate and mathematically misleading. Here
for example in its entirety, is Artstein’s explanation of how Newton verified
Galileo’s observation on the path of a falling object:
Newton showed that the derivative of the function a(t) = αtn , where
α is a constant, is αntn−1 . In particular, if the second derivative has
a fixed value g, its integral is gt, and the integral of the latter is 21 gt2 .
This shows that as the Earth’s gravitational pull g is constant over short
distances, the parabolas that Galileo observed when he dropped bodies
from the top of a tower fulfilled Newton’s second law of motion.
 Book Reviews 233

The historical account, ranging from tallying by pre-historic humans, astronomy

and mensuration in bronze-age agricultural societies, Greek mathematics and me-
dieval astronomy, through Galileo and Newton to the early moderns such as Euler
and the Bernoullis, contains nothing novel and is riddled with errors, of which the
most egregious are: the Babylonians inscribed calculations on potsherds; Euclid
developed axiomatics in order to avoid optical illusions in diagrams; Greek math-
ematicians studied the brachistochrone problem; Fermat’s Last Theorem occurs
among Hilbert’s list of unsolved problems presented at the 1900 International
Mathematical Congress. The author is on firmer ground when dealing with the
history of the mathematics with which he is most familiar: the principle of least
action, calculus of variations, statistics and the social sciences and computation.
He also has a perceptive chapter on the nature of research in mathematics.
The book contains frequent observations on how mathematics should be taught
and learned. However, the author’s remarks on mathematics education are, in the
reviewer’s opinion, one-sided and extreme.
Some of the foregoing negative comments must have filtered through to the author
or his publisher during the process of publication, because Artstein concludes with
a poignant Afterword, imploring readers to forgive his errors, since Evolution did
not prepare us for error-free rigorous analysis!

Phill Schultz
School of Mathematics and Statistics, The University of Western Australia.
Email address: [email protected]

     

Neverending Fractions
An introduction to Continued Fractions∗

Jonathan Borwein, Alf van der Poorten, Jeffrey Shallit and Wadim Zudilin
Australian Mathematical Society Lecture Series, number 23
Cambridge University Press, 2014, ISBN 978-1107-00665-2
Also available in eBook (ISBN 978-0-511-90265-9)

Continued fractions form a classical area within number theory, the roots of which
can be traced back to Euclid’s algorithm for the greatest common divisor of two
integers (300 BC). Several centuries ago, Rafael Bombelli (1579), Pietro Cataldi
(1613), and John Wallis (1695) developed the method of continued fractions for
rational approximations of irrational numbers (such as square roots), and later
on great mathematicians like Leonhard Euler (1737 and 1748), Johann Lambert
(1761), Joseph L. Lagrange (1768 and 1770), Carl Friedrich Gauss (1813), and
others discovered various fundamental properties and important applications of

∗ The Gazette thanks Zentralblatt MATH and Werner Kleinert for their permission to republish

this review, which originally appeared as Zbl 1307.11001.

234 Book Reviews

continued fractions. In fact, these fascinating objects have been a very active field
of research ever since, and the vast contemporary literature on continued fractions
evidently shows that this topic is still far from being exhausted.
The book under review grew out of many lectures that the four authors delivered
independently on different occasions to students of different levels. Its main goal
is to provide an introduction to continued fractions for a wide audience of readers,
including graduate students, postgraduates, researchers as well as teachers and
even amateurs in mathematics. As the authors point out in the preface, their
intention is to demonstrate that continued fractions represent a neverending re-
search field, with a wealth of results elementary enough to be explained to this
target readership.
Regarding the precise contents, the book comprises nine chapters, each of which
is divided into several sections. While the first three chapters are devoted to a
general introduction to continued fractions, the subsequent six chapters deal with
more special topics and applications of the theory.
Chapter 1 presents the necessary prerequisites from
elementary number theory with full proofs. These
concern the following themes: divisibility of integers
and the Euclidean algorithm, prime numbers and
the fundamental theorem of arithmetic, Fibonacci
numbers and the complexity of the Euclidean algo-
rithm, approximation of real numbers by rationals
and Farey sequences. Chapter 2 begins the study of
continued fractions and their algebraic theory, thereby
explaining the continued fraction of a real number in
general, the principle of Diophantine approximation,
the continued fraction of a quadratic irrational and
the Euler–Lagrange theorem in this context, the
construction of real numbers with bounded partial
quotients, and other results on rational approxima-
tion.
Chapter 3 touches upon the metric theory of continued fractions, with emphasis on
the growth of partial quotients of a continued fraction of a real number, the approx-
imation of almost all real numbers by rationals, and the classical Gauss–Kuzmin
statistics in metric number theory. Chapters 4, 5 and 6 originate in lectures that
one of the authors, the late Alf van der Poorten (1942–2010), gave in the last few
years before his untimely death.
Chapter 4 is titled ‘Quadratic irrationals through a magnifier’ and contains some
informal lectures on continued fractions of algebraic numbers, Pell’s equation, and
some concrete examples.
Chapter 5 is a survey of aspects of continued fractions in function fields, with a view
toward some so-called (recursively defined) Somos sequences, pseudo-elliptic inte-
grals, and hyperelliptic curves, whereas Chapter 6 briefly discusses the relationship
between neverending paper foldings and continued fractions. Chapter 7 provides
the study of a class of generating functions that are connected to remarkable
 Book Reviews 235

continued fractions and rational approximations. Lambert series expansions of

generating functions and an inhomogeneous Diophantine approximation algorithm
are the main tools applied here.
Chapter 8 treats the Erdős–Moser equation

1k + 2k + · · · + (m − 2)k + (m − 1)k = mk

and its possible integer solutions for m ≥ 2 and k ≥ 2.

A conjecture by P. Erdős states that such solutions do not exist, and L. Moser
proved in 1953 that only for even exponents k and rather large integers m such
solutions could be expected at all.
In this chapter, both the arithmetic and the analysis of the Erdős–Moser equa-
tion are outlined, where efficient ways of computing certain associated continued
fractions as well as explicit bounds for solutions are presented. The basic reference
for this chapter is the recent paper by Y. Gallot, P. Moree and W. Zudilin (Math.
Comput. 80, No. 274, 1221–1237 (2011; Zbl 1231.11038)).
The concluding Chapter 9 finally turns to irregular continued fractions by survey-
ing their general theory as well as some important examples, including Gauss’ irreg-
ular continued fraction for the hypergeometric function, Ramanujan’s arithmetic-
geometric mean (AGM) continued fraction (from his second notebook) and related
developments by one of the authors of the present book (J. Borwein) and his collab-
orators, an irregular continued fraction for the zeta value ζ(2) = π 2 /6, and a new
proof of R. Apéry’s theorem on the irrationality of ζ(3) as a striking application
of the foregoing discussion.
There is an appendix to the main text containing a collection of interesting contin-
ued fractions, both regular and irregular, where most of those represent special real
numbers, values of special functions, particular infinite series, and some q-series,
respectively. As one can see, the book is a combination of formal and informal
styles of expository writing, and a mixture of introductory textbook and topical
surveys likewise. Many of the special topics discussed in the later chapters are
not to be found in other books but only in scattered articles and lectures. As
for full details with regard to these topics chapters, the reader is referred to the
original research papers listed in the rich bibliography. In fact, each chapter ends
with a set of notes providing additional remarks and hints for further reading, and
a few exercises invite the reader to acquire complementary knowledge through
independent work.
All together, the present book gives a beautiful panoramic view of the ‘neverending
story of neverending fractions’ by making apparent their naturalness, their ubiq-
uity, and their wide-range of applications in very lucid and inspiring a manner.

Werner Kleinert
Berlin

     
236 Book Reviews

An Aristotelian Realist Philosophy of Mathematics

James Franklin
Palgrave Macmillan, 2014, ISBN 978-1-137-40072-7

What are the current trends in philosophy of mathematics, and what relevance
do they have to the practice of mathematics? Those are some of the questions
I asked myself when I began reading this book by James Franklin, Professor of
Mathematics at UNSW and founder of the ‘Sydney School’ in the philosophy of
mathematics.
There are a few hurdles to overcome for a lay person
attempting to read a book or paper in philosophy. The
first is the jargon. Some phrases, such as ‘epistemol-
ogy and ontology’, ‘mereology’, ‘potentially infinite’
(which just means unbounded) are clearly defined in
dictionaries and Wikipedia and pose no real problem.
But others, such as ‘necessary’, ‘contingent’, and
‘uninstantiated universal’ are ambiguous or depend
on individual psychology. Even more troubling are
words such as ‘random’ which have different meanings
in mathematics and philosophy. A striking example is
the claim by the author that the sequence of decimal
digits of π and even a finite initial segment of this
sequence, is random. Random, it seems, is in the eye
of the beholder.
A second feature of the philosophical literature is its polemical tone. Unlike the
polite euphemisms one finds in mathematical papers which point out errors or
incomplete proofs, philosophical papers bristle with words like ‘mistaken’, ‘ill-
informed’ and ‘falsity’. In mathematics I have only seen something approaching this
in papers on the foundation of probability espousing the frequentist or Bayesian
approach.
So what are the principal opposing schools in mathematical philosophy today?
According to Franklin they are Platonism, which holds that mathematics is about
the real world, and Nominalism which claims it is about words. They correspond
roughly to what mathematicians call platonism and formalism. (I use lower case
for the mathematical concepts to distinguish them from the slightly different philo-
sophical meaning). But whereas mathematicians see no contradiction in embracing
platonism when doing research and formalism when writing it up, philosophers
view this as akin to treason.
Within Platonism there are again two competing views of mathematics. One is
Platonic Idealism, which holds that the real world is an approximation to math-
ematics, and the other is Aristotelian Realism which holds that mathematics is
an approximation to the real world. The author is firmly in the latter school,
 Book Reviews 237

proclaiming mathematics to be the science of quantity and structure. For example,

the number 6 is the property of a heap of six apples which distinguishes it from a
heap of five apples. Similarly, 1 meter is the length of a chalk line measured with
a standard rod, and a sphere is the shape of a bronze ball. Franklin is well aware
that a straight line drawn on paper is neither straight nor a line, a 1000-sided
regular polygon is indistinguishable from a circle and a bronze ball is not a sphere.
He deals with this problem by associating with each real entity a tolerance, in the
engineering sense.
For Aristotelian Realists, the space in which we live is locally Euclidean and its
properties are not postulated, but observed and verified. The arithmetic of the
integers and their ratios make sense, and we can define an action of the rationals
on the continuum. Since there is no natural choice of a unit of measure, there
is no embedding of the rationals in the real line, so it is not obvious if and how
the arithmetic of the real and complex numbers is related to that of the integers
and rationals. The topology of space and hence continuous real functions can be
defined intrinsically, but Franklin does not address the question of smoothness and
the existence of pathological functions.
The claim that mathematics concerns the real world extends also to structures.
Franklin defines a mathematical property to be structural if it can be defined in
terms of the concepts ‘same and different’ and ‘part and whole’. For example an
entity is symmetrical if it consists of two parts which are the same in some respect.
What ‘in some respect’ means is apparently a property of human cognition. Among
the problems with this concept are the cognitive errors of apophenia (the human
tendency to perceive patterns in random or meaningless information), and the
opposite error of failing to recognise a pattern that exists. Terry Tao, for example,
sees relations between the eigenvalues of the random matrix and the distribution
of zeroes of the Riemann zeta function that I do not.
A major claim of the author is that while pure mathematicians may be more com-
fortable in the Platonic universe, Aristotelian Realism is especially appropriate for
applied mathematics. One breathtaking suggestion is that the applied mathemati-
cian need not be concerned with infinity because applied mathematics only deals
with finite objects. This concentration on small concrete cases has the unfortunate
consequence that the author fails to take account of the importance of general-
isation. For example, according to Franklin, the purpose of Euler’s Königsberg
Bridges paper was to prove that the burgers of Königsberg could not stroll over
all their bridges just once.
A major problem associated with Realist mathematics is the question of infinity,
both discrete and continuous. The author recognises that it must be faced in any
coherent mathematical system, and his partial solution is the notion of ‘uninstan-
tiated universals’, that is, entities that could exist but do not, the philosopher’s
favourite example being the golden mountain. By invoking uninstantiated univer-
sals, Franklin allows Realist mathematics to admit large cardinals when required.
But when explaining the reliance of mathematics on logical reasoning, he fails
to address the problem of the internal consistency of Aristotelian Realism. There
should be no need for this, because if mathematics is only about the real world, it
238 Book Reviews

has a model. But as soon as uninstantiated universals enter the picture, so does the
possibility of logical contradiction. For example using arguments valid in Realism
with uninstantiated universals, Gödel constructed a consistent universe in which
2ℵ0 = ℵ1 , and Cohen one in which 2ℵ0 > ℵ1 . Whereas formalist mathematicians
are happy to pursue the consequences of accepting or denying the continuum
hypothesis, to the Realist their truth or falsity is a property of the real world.
It is easy to ridicule such attempts to rewrite mathematics. It is more useful to
consider which aspects of mathematics can be developed using an Aristotelian
Realistic foundation when we allow a cautious interpretation of uninstantiated
universals. Firstly ZF Set Theory, including the axiom of infinity but without
the unrestricted axiom of choice, is realistic and can be used to define functions
and relations. Furthermore, the arithmetic of the integers and their ratios make
sense. We can embed the field of rationals in a complete Archimedean ordered
field and so construct real and complex numbers and vector spaces with their
associated topology. However, notions of compactness that require choice are not
allowed, so that classical analytic and harmonic analysis remain out of reach.
Elementary number theory makes sense and so does finite combinatorics. We can
define algebraic structures, but lacking the maximum principle, we cannot prove
the existence or properties of maximal normal substructures except in the finite
case.
To my mind, too much is lost without a platonic and formalistic approach.

Phill Schultz
School of Mathematics and Statistics, The University of Western Australia.
Email address: [email protected]