Part I

Introduction

division of mathematics into algebra, geometry, and

I.1 What Is Mathematics About? analysis. So let us begin with this, and then qualify it
later.
It is notoriously hard to give a satisfactory answer to
the question, “What is mathematics?” The approach of 1.1 Algebra versus Geometry
this book is not to try. Rather than giving a definition
Most people who have done some high school mathe-
of mathematics, the intention is to give a good idea of
matics will think of algebra as the sort of mathemat-
what mathematics is by describing many of its most
ics that results when you substitute letters for num-
important concepts, theorems, and applications. Nev-
bers. Algebra will often be contrasted with arithmetic,
ertheless, to make sense of all this information it is
which is a more direct study of the numbers them-
useful to be able to classify it somehow.
selves. So, for example, the question, “What is 3 × 7?”
The most obvious way of classifying mathematics is
will be thought of as belonging to arithmetic, while the
by its subject matter, and that will be the approach of
question, “If x + y = 10 and xy = 21, then what is the
this brief introductory section and the longer section
value of the larger of x and y?” will be regarded as a
entitled some fundamental mathematical defini-
piece of algebra. This contrast is less apparent in more
tions [I.3]. However, it is not the only way, and not
advanced mathematics for the simple reason that it is
even obviously the best way. Another approach is to
very rare for numbers to appear without letters to keep
try to classify the kinds of questions that mathemati-
them company.
cians like to think about. This gives a usefully different
There is, however, a different contrast, between alge-
view of the subject: it often happens that two areas of
bra and geometry, which is much more important at an
mathematics that appear very different if you pay atten-
advanced level. The high school conception of geom-
tion to their subject matter are much more similar if
etry is that it is the study of shapes such as circles, tri-
you look at the kinds of questions that are being asked.
angles, cubes, and spheres together with concepts such
The last section of part I, entitled the general goals
as rotations, reflections, symmetries, and so on. Thus,
of mathematical research [I.4], looks at the subject
the objects of geometry, and the processes that they
from this point of view. At the end of that article there
undergo, have a much more visual character than the
is a brief discussion of what one might regard as a third
equations of algebra.
classification, not so much of mathematics itself but of
This contrast persists right up to the frontiers of
the content of a typical article in a mathematics jour-
modern mathematical research. Some parts of mathe-
nal. As well as theorems and proofs, such an article will
matics involve manipulating symbols according to cer-
contain definitions, examples, lemmas, formulas, con-
tain rules: for example, a true equation remains true
jectures, and so on. The point of that discussion will
if you “do the same to both sides.” These parts would
be to say what these words mean and why the different
typically be thought of as algebraic, whereas other parts
kinds of mathematical output are important.
are concerned with concepts that can be visualized, and
these are typically thought of as geometrical.
1 Algebra, Geometry, and Analysis
However, a distinction like this is never simple. If
Although any classification of the subject matter of you look at a typical research paper in geometry, will
mathematics must immediately be hedged around with it be full of pictures? Almost certainly not. In fact, the
qualifications, there is a crude division that undoubt- methods used to solve geometrical problems very often
edly works well as a first approximation, namely the involve a great deal of symbolic manipulation, although
2 I. Introduction

good powers of visualization may be needed to find and ces together, and it is designed to have the property
use these methods and pictures will typically under- that if matrix A represents a transformation R (such
lie what is going on. As for algebra, is it “mere” sym- as a reflection) and matrix B represents a transforma-
bolic manipulation? Not at all: very often one solves an tion T , then the product AB represents the transforma-
algebraic problem by finding a way to visualize it. tion that results when you first do T and then R. There-
As an example of visualizing an algebraic problem, fore, one can solve the problem above by writing down
consider how one might justify the rule that if a and the matrices that correspond to the transformations,
b are positive integers then ab = ba. It is possible to multiplying them together, and seeing what transfor-
approach the problem as a pure piece of algebra (per- mation corresponds to the product. In this way, the
haps proving it by induction), but the easiest way to geometrical problem has been converted into algebra
convince yourself that it is true is to imagine a rectangu- and solved algebraically.
lar array that consists of a rows with b objects in each Thus, while one can draw a useful distinction be-
row. The total number of objects can be thought of as tween algebra and geometry, one should not imagine
a lots of b, if you count it row by row, or as b lots of a, that the boundary between the two is sharply defined.
if you count it column by column. Therefore, ab = ba. In fact, one of the major branches of mathematics is
Similar justifications can be given for other basic rules even called algebraic geometry [IV.4]. And as the
such as a(b + c) = ab + ac and a(bc) = (ab)c. above examples illustrate, it is often possible to trans-
In the other direction, it turns out that a good way of late a piece of mathematics from algebra into geometry
solving many geometrical problems is to “convert them or vice versa. Nevertheless, there is a definite differ-
into algebra.” The most famous way of doing this is to ence between algebraic and geometric methods of think-
use Cartesian coordinates. For example, suppose that ing—one more symbolic and one more pictorial—and
you want to know what happens if you reflect a circle this can have a profound influence on which subjects a
about a line L through its center, then rotate it through mathematician chooses to pursue.
40◦ counterclockwise, and then reflect it once more
1.2 Algebra versus Analysis
about the same line L. One approach is to visualize the
situation as follows. The word “analysis,” used to denote a branch of math-
Imagine that the circle is made of a thin piece of ematics, is not one that features at high school level.
wood. Then instead of reflecting it about the line you However, the word “calculus” is much more familiar,
can rotate it through 180◦ about L (using the third and differentiation and integration are good examples
dimension). The result will be upside down, but this of mathematics that would be classified as analysis
does not matter if you simply ignore the thickness of rather than algebra or geometry. The reason for this
the wood. Now if you look up at the circle from below is that they involve limiting processes. For example, the
while it is rotated counterclockwise through 40◦ , what derivative of a function f at a point x is the limit of the
you will see is a circle being rotated clockwise through gradients of a sequence of chords of the graph of f , and
40◦ . Therefore, if you then turn it back the right way the area of a shape with a curved boundary is defined
up, by rotating about L once again, the total effect will to be the limit of the areas of rectilinear regions that
have been a clockwise rotation through 40◦ . fill up more and more of the shape. (These concepts
Mathematicians vary widely in their ability and will- are discussed in much more detail in [I.3 §5].)
ingness to follow an argument like that one. If you Thus, as a first approximation, one might say that a
cannot quite visualize it well enough to see that it is branch of mathematics belongs to analysis if it involves
definitely correct, then you may prefer an algebraic limiting processes, whereas it belongs to algebra if you
approach, using the theory of linear algebra and matri- can get to the answer after just a finite sequence of
ces (which will be discussed in more detail in [I.3 §3.2]). steps. However, here again the first approximation is
To begin with, one thinks of the circle as the set of all so crude as to be misleading, and for a similar reason:
pairs of numbers (x, y) such that x 2 + y 2  1. The two if one looks more closely one finds that it is not so much
transformations, reflection in a line through the center branches of mathematics that should be classified into
of the circle and rotation through an angle θ, can both analysis or algebra, but mathematical techniques.
be represented by 2 × 2 matrices, which are arrays of Given that we cannot write out infinitely long proofs,
numbers of the form ( ac d b
). There is a slightly compli- how can we hope to prove anything about limiting pro-
cated, but purely algebraic, rule for multiplying matri- cesses? To answer this, let us look at the justification
I.1. What Is Mathematics About? 3

for the simple statement that the derivative of x 3 is simply points out that x 4 − x 2 − 6x + 10 is equal to
3x 2 . The usual reasoning is that the gradient of the (x 2 − 1)2 + (x − 3)2 , and is therefore always positive.
chord of the line joining the two points (x, x 3 ) and This may make it seem as though, given the choice
((x + h), (x + h)3 ) is between analysis and algebra, one should go for alge-
(x + h)3 − x 3 bra. After all, the algebraic proof was much shorter,
, and makes it obvious that the function is always pos-
x+h−x
itive. However, although there were several steps to
which works out as 3x 2 + 3xh + h2 . As h “tends to
the analyst’s proof, they were all easy, and the brevity
zero,” this gradient “tends to 3x 2 ,” so we say that the
of the algebraic proof is misleading since no clue has
gradient at x is 3x 2 . But what if we wanted to be a bit
been given about how the equivalent expression for
more careful? For instance, if x is very large, are we
x 4 − x 2 − 6x + 10 was found. And in fact, the gen-
really justified in ignoring the term 3xh?
eral question of when a polynomial can be written as
To reassure ourselves on this point, we do a small cal-
a sum of squares of other polynomials turns out to be
culation to show that, whatever x is, the error 3xh + h2
an interesting and difficult one (particularly when the
can be made arbitrarily small, provided only that h is
polynomials have more than one variable).
sufficiently small. Here is one way of going about it.
There is also a third, hybrid approach to the prob-
Suppose we fix a small positive number , which rep-
lem, which is to use calculus to find the points where
resents the error we are prepared to tolerate. Then if
x 4 −x 2 −6x+10 is minimized. The idea would be to cal-
|h|  /6x, we know that |3xh| is at most /2. If in
 culate the derivative 4x 3 − 2x − 6 (an algebraic process,
addition we know that |h|  /2, then we also know
justified by an analytic argument), find its roots (alge-
that h2  /2. So, provided that |h| is smaller than
 bra), and check that the values of x 4 −x 2 −6x+10 at the
the minimum of the two numbers /6x and /2, the
roots of the derivative are positive. However, though
difference between 3x 2 + 3xh + h2 and 3x 2 will be at
the method is a good one for many problems, in this
most .
case it is tricky because the cubic 4x 3 − 2x − 6 does not
There are two features of the above argument that
have integer roots. But one could use an analytic argu-
are typical of analysis. First, although the statement we
ment to find small intervals inside which the minimum
wished to prove was about a limiting process, and was
must occur, and that would then reduce the number
therefore “infinitary,” the actual work that we needed to
of cases that had to be considered in the first, purely
do to prove it was entirely finite. Second, the nature of
analytic, argument.
that work was to find sufficient conditions for a certain
As this example suggests, although analysis often
fairly simple inequality (the inequality |3xh + h2 |  )
involves limiting processes and algebra usually does
to be true.
not, a more significant distinction is that algebraists
Let us illustrate this second feature with another
like to work with exact formulas and analysts use esti-
example: a proof that x 4 − x 2 − 6x + 10 is positive
mates. Or, to put it even more succinctly, algebraists
for every real number x. Here is an “analyst’s argu-
like equalities and analysts like inequalities.
ment.” Note first that if x  −1 then x 4  x 2 and
10−6x  0, so the result is certainly true in this case. If
2 The Main Branches of Mathematics
−1  x  1, then |x 4 − x 2 − 6x| cannot be greater than
x 4 +x 2 +6|x|, which is at most 8, so x 4 −x 2 −6x  −8, Now that we have discussed the differences between
which implies that x 4 − x 2 − 6x + 10  2. If 1  x  32 , algebraic, geometrical, and analytical thinking, we are
then x 4  x 2 and 6x  9, so x 4 − x 2 − 6x + 10  1. ready for a crude classification of the subject matter of
If 32  x  2, then x 2  94 , so x 4 − x 2 = x 2 (x 2 − 1)  mathematics. We face a potential confusion, because
9 5
4 · 4 > 2. Also, 6x  12, so 10 − 6x  −2. There- the words “algebra,” “geometry,” and “analysis” refer
fore, x 4 − x 2 − 6x + 10 > 0. Finally, if x  2, then both to specific branches of mathematics and to ways
x 4 −x 2 = x 2 (x 2 −1)  3x 2  6x, from which it follows of thinking that cut across many different branches.
that x 4 − x 2 − 6x + 10  10. Thus, it makes sense to say (and it is true) that some
The above argument is somewhat long, but each step branches of analysis are more algebraic (or geometri-
consists in proving a rather simple inequality—this cal) than others; similarly, there is no paradox in the
is the sense in which the proof is typical of analy- fact that algebraic topology is almost entirely algebraic
sis. Here, for contrast, is an “algebraist’s proof.” One and geometrical in character, even though the objects
4 I. Introduction

it studies, topological spaces, are part of analysis. In which describes the basic building blocks out of which
this section, we shall think primarily in terms of subject any finite group can be built.
matter, but it is important to keep in mind the distinc- Algebraic structures appear throughout mathemat-
tions of the previous section and be aware that they are ics, and there are many applications of algebra to other
in some ways more fundamental. Our descriptions will areas, such as number theory, geometry, and even
be very brief: further reading about the main branches mathematical physics.
of mathematics can be found in parts II and IV, and
more specific points are discussed in parts III and V. 2.2 Number Theory

2.1 Algebra Number theory is largely concerned with properties of

the set of positive integers, and as such has a consid-
The word “algebra,” when it denotes a branch of math- erable overlap with algebra. But a simple example that
ematics, means something more specific than manipu- illustrates the difference between a typical question in
lation of symbols and a preference for equalities over algebra and a typical question in number theory is pro-
inequalities. Algebraists are concerned with number vided by the equation 13x − 7y = 1. An algebraist
systems, polynomials, and more abstract structures would simply note that there is a one-parameter fam-
such as groups, fields, vector spaces, and rings (dis- ily of solutions: if y = λ then x = (1 + 7λ)/13, so the
cussed in some detail in some fundamental math- general solution is (x, y) = ((1 + 7λ)/13, λ). A num-
ematical definitions [I.3]). Historically, the abstract ber theorist would be interested in integer solutions,
structures emerged as generalizations from concrete and would therefore work out for which integers λ the
instances. For instance, there are important analo- number 1 + 7λ is a multiple of 13. (The answer is that
gies between the set of all integers and the set of all 1 + 7λ is a multiple of 13 if and only if λ has the form
polynomials with rational (for example) coefficients, 13m + 11 for some integer m.)
which are brought out by the fact that both sets are However, this description does not do full justice
examples of algebraic structures known as Euclidean to modern number theory, which has developed into
domains. If one has a good understanding of Euclidean a highly sophisticated subject. Most number theorists
domains, one can apply this understanding to integers are not directly trying to solve equations in integers;
and polynomials.
instead they are trying to understand structures that
This highlights a contrast that appears in many
were originally developed to study such equations but
branches of mathematics, namely the distinction be-
which then took on a life of their own and became
tween general, abstract statements and particular, con-
objects of study in their own right. In some cases,
crete ones. One algebraist might be thinking about
this process has happened several times, so the phrase
groups, say, in order to understand a particular rather
“number theory” gives a very misleading picture of
complicated group of symmetries, while another might
what some number theorists do. Nevertheless, even the
be interested in the general theory of groups on the
most abstract parts of the subject can have down-to-
grounds that they are a fundamental class of math-
earth applications: a notable example is Andrew Wiles’s
ematical objects. The development of abstract alge-
famous proof of fermat’s last theorem [V.10].
bra from its concrete beginnings is discussed in the
Interestingly, in view of the discussion earlier, num-
origins of modern algebra [II.3].
ber theory has two fairly distinct subbranches, known
A supreme example of a theorem of the first kind is
as algebraic number theory [IV.1] and analytic
the insolubility of the quintic [V.21]—the result
number theory [IV.2]. As a rough rule of thumb, the
that there is no formula for the roots of a quintic poly-
study of equations in integers leads to algebraic num-
nomial in terms of its coefficients. One proves this
ber theory, while analytic number theory has its roots
theorem by analyzing symmetries associated with the
in the study of prime numbers, but the true picture is
roots of a polynomial, and understanding the group
of course more complicated.
that these symmetries form. This concrete example of
a group (or rather, class of groups, one for each polyno-
2.3 Geometry
mial) played a very important part in the development
of the abstract theory of groups. A central object of study is the manifold, which is dis-
As for the second kind of theorem, a good example cussed in [I.3 §6.9]. Manifolds are higher-dimensional
is the classification of finite simple groups [V.7], generalizations of shapes like the surface of a sphere: a
I.1. What Is Mathematics About? 5

small portion of a manifold looks flat, but the manifold vertex of the cone. Thus, (0, 0, 0) is a singularity. (This
as a whole may be curved in complicated ways. Most means that the cone is not actually a manifold, but a
people who call themselves geometers are studying “manifold with a singularity.”)
manifolds in one way or another. As with algebra, some The interplay between algebra and geometry is part
will be interested in particular manifolds and others in of what gives algebraic geometry its fascination. A fur-
the more general theory. ther impetus to the subject comes from its connections
Within the study of manifolds, one can attempt a to other branches of mathematics. There is a particu-
further classification, according to when two mani- larly close connection with number theory, explained in
folds are regarded as “genuinely distinct.” A topolo- arithmetic geometry [IV.5]. More surprisingly, there
gist regards two objects as the same if one can be are important connections between algebraic geom-
continuously deformed, or “morphed,” into the other; etry and mathematical physics. See mirror symmetry
thus, for example, an apple and a pear would count [IV.16] for an account of some of these.
as the same for a topologist. This means that rela-
tive distances are not important to topologists, since 2.5 Analysis
one can change them by suitable continuous stretches.
A differential topologist asks for the deformations to Analysis comes in many different flavors. A major
be “smooth” (which means “sufficiently differentiable”). topic is the study of partial differential equations
This results in a finer classification of manifolds and a [IV.12]. This began because partial differential equa-
different set of problems. At the other, more “geomet- tions were found to govern many physical processes,
rical,” end of the spectrum are mathematicians who are such as motion in a gravitational field, for example.
much more interested in the precise nature of the dis- But partial differential equations arise in purely mathe-
tances between points on a manifold (a concept that matical contexts as well—particularly in geometry—so
would not make sense to a topologist) and in auxiliary they give rise to a big branch of mathematics with many
structures that one can associate with a manifold. See subbranches and links to many other areas.
riemannian metrics [I.3 §6.10] and ricci flow [III.78] Like algebra, analysis has an abstract side as well. In
for some indication of what the more geometrical side particular, certain abstract structures, such as banach
of geometry is like. spaces [III.62], hilbert spaces [III.37], C ∗ -algebras
[IV.15 §3], and von neumann algebras [IV.15 §2], are
2.4 Algebraic Geometry central objects of study. These four structures are all
As its name suggests, algebraic geometry does not have infinite-dimensional vector spaces [I.3 §2.3], and the
an obvious place in the above classification, so it is eas- last two are “algebras,” which means that one can multi-
ier to discuss it separately. Algebraic geometers also ply their elements together as well as adding them and
study manifolds, but with the important difference that multiplying them by scalars. Because these structures
their manifolds are defined using polynomials. (A sim- are infinite dimensional, studying them involves limit-
ple example of this is the surface of a sphere, which ing arguments, which is why they belong to analysis.
can be defined as the set of all (x, y, z) such that However, the extra algebraic structure of C ∗ -algebras
x 2 +y 2 +z2 = 1.) This means that algebraic geometry is and von Neumann algebras means that in those areas
algebraic in the sense that it is “all about polynomials” substantial use is made of algebraic tools as well. And
but geometric in the sense that the set of solutions of as the word “space” suggests, geometry also has a very
a polynomial in several variables is a geometric object. important role.
An important part of algebraic geometry is the study dynamics [IV.14] is another significant branch of
of singularities. Often the set of solutions to a system of analysis. It is concerned with what happens when you
polynomial equations is similar to a manifold, but has a take a simple process and do it over and over again.
few exceptional, singular points. For example, the equa- For example, if you take a complex number z0 , then
tion x 2 = y 2 + z 2 defines a (double) cone, which has let z1 = z02 + 2, and then let z2 = z12 + 2, and so
its vertex at the origin (0, 0, 0). If you look at a small on, then what is the limiting behavior of the sequence
enough neighborhood of a point x on the cone, then, z0 , z1 , z2 , . . . ? Does it head off to infinity or stay in some
provided x is not (0, 0, 0), the neighborhood will resem- bounded region? The answer turns out to depend in
ble a flat plane. However, if x is (0, 0, 0), then no mat- a complicated way on the original number z0 . Exactly
ter how small the neighborhood is, you will still see the how it depends on z0 is a question in dynamics.
6 I. Introduction

Sometimes the process to be repeated is an “infinites- applicable and has led to important discoveries in fields
imal” one. For example, if you are told the positions, well outside set theory.
velocities, and masses of all the planets in the solar
system at a particular moment (as well as the mass of 2.7 Combinatorics
the Sun), then there is a simple rule that tells you how
There are various ways in which one can try to define
the positions and velocities will be different an instant
combinatorics. None is satisfactory on its own, but
later. Later, the positions and velocities have changed,
together they give some idea of what the subject is like.
so the calculation changes; but the basic rule is the
A first definition is that combinatorics is about count-
same, so one can regard the whole process as applying
ing things. For example, how many ways are there of
the same simple infinitesimal process infinitely many
filling an n × n square grid with 0s and 1s if you are
times. The correct way to formulate this is by means
allowed at most two 1s in each row and at most two 1s
of partial differential equations and therefore much of
in each column? Because this problem asks us to count
dynamics is concerned with the long-term behavior of
something, it is, in a rather simple sense, combinatorial.
solutions to these.
Combinatorics is sometimes called “discrete math-
2.6 Logic ematics” because it is concerned with “discrete” struc-
tures as opposed to “continuous” ones. Roughly speak-
The word “logic” is sometimes used as a shorthand ing, an object is discrete if it consists of points that
for all branches of mathematics that are concerned are isolated from each other, and continuous if you
with fundamental questions about mathematics itself, can move from one point to another without making
notably set theory [IV.22], category theory [III.8], sudden jumps. (A good example of a discrete struc-
model theory [IV.23], and logic in the narrower sense ture is the integer lattice Z2 , which is the grid con-
of “rules of deduction.” Among the triumphs of set sisting of all points in the plane with integer coordin-
theory are gödel’s incompleteness theorems [V.15] ates, and a good example of a continuous one is the
and Paul Cohen’s proof of the independence of the surface of a sphere.) There is a close affinity between
continuum hypothesis [V.18]. Gödel’s theorems in combinatorics and theoretical computer science (which
particular had a dramatic effect on philosophical per- deals with the quintessentially discrete structure of
ceptions of mathematics, though now that it is under-
sequences of 0s and 1s), and combinatorics is some-
stood that not every mathematical statement has a
times contrasted with analysis, though in fact there are
proof or disproof most mathematicians carry on much
several connections between the two.
as before, since most statements they encounter do
A third view of combinatorics is that it is con-
tend to be decidable. However, set theorists are a dif-
cerned with mathematical structures that have “few
ferent breed. Since Gödel and Cohen, many further
constraints.” This idea helps to explain why number
statements have been shown to be undecidable, and
theory, despite the fact that it studies (among other
many new axioms have been proposed that would make
things) the distinctly discrete set of all positive inte-
them decidable. Thus, decidability is now studied for
gers, is not considered a branch of combinatorics.
mathematical rather than philosophical reasons.
In order to illustrate this last contrast, here are
Category theory is another subject that began as
two somewhat similar problems, both about positive
a study of the processes of mathematics and then
integers.
became a mathematical subject in its own right. It dif-
fers from set theory in that its focus is less on math- (i) Is there a positive integer that can be written in a
ematical objects themselves than on what is done to thousand different ways as a sum of two squares?
those objects—in particular, the maps that transform (ii) Let a1 , a2 , a3 , . . . be a sequence of positive inte-
one to another. gers, and suppose that each an lies between n2
A model for a collection of axioms is a mathematical and (n+1)2 . Will there always be a positive integer
structure for which those axioms, suitably interpreted, that can be written in a thousand different ways as
are true. For example, any concrete example of a group a sum of two numbers from the sequence?
is a model for the axioms of group theory. Set theorists
study models of set-theoretic axioms, and these are The first question counts as number theory, since it
essential to the proofs of the famous theorems men- concerns a very specific sequence—the sequence of
tioned above, but the notion of a model is more widely squares—and one would expect to use properties of
I.1. What Is Mathematics About? 7

this special set of numbers in order to determine the unexpectedly interesting behavior with direct practical
answer, which turns out to be yes.1 relevance. For example, it may happen that there is a
The second question concerns a far less structured “critical probability” p with the following property: if
sequence. All we know about an is its rough size—it is the probability of infection after contact of a certain
fairly close to n2 —but we know nothing about its more kind is above p then an epidemic may very well result,
detailed properties, such as whether it is a prime, or a whereas if it is below p then the disease will almost
perfect cube, or a power of 2, etc. For this reason, the certainly die out. A dramatic difference in behavior
second problem belongs to combinatorics. The answer like this is called a phase transition. (See probabilis-
is not known. If the answer turns out to be yes, then tic models of critical phenomena [IV.25] for further
it will show that, in a sense, the number theory in the discussion.)
first problem was an illusion and that all that really Setting up an appropriate mathematical model can
mattered was the rough rate of growth of the sequence be surprisingly difficult. For example, there are physical
of squares. circumstances where particles travel in what appears to
be a completely random manner. Can one make sense
2.8 Theoretical Computer Science
of the notion of a random continuous path? It turns
This branch of mathematics is described at consider- out that one can—the result is the elegant theory of
able length in part IV, so we shall be brief here. Broadly brownian motion [IV.24]—but the proof that one can
speaking, theoretical computer science is concerned is highly sophisticated, roughly speaking because the
with efficiency of computation, meaning the amounts set of all possible paths is so complex.
of various resources, such as time and computer mem-
ory, needed to perform given computational tasks. 2.10 Mathematical Physics
There are mathematical models of computation that
allow one to study questions about computational effi- The relationship between mathematics and physics has
ciency in great generality without having to worry about changed profoundly over the centuries. Up to the eigh-
precise details of how algorithms are implemented. teenth century there was no sharp distinction drawn
Thus, theoretical computer science is a genuine branch between mathematics and physics, and many famous
of pure mathematics: in theory, one could be an excel- mathematicians could also be regarded as physicists,
lent theoretical computer scientist and be unable to at least some of the time. During the nineteenth cen-
program a computer. However, it has had many notable tury and the beginning of the twentieth century this
applications as well, especially to cryptography (see situation gradually changed, until by the middle of the
mathematics and cryptography [VII.7] for more on twentieth century the two disciplines were very sepa-
this). rate. And then, toward the end of the twentieth cen-
tury, mathematicians started to find that ideas that had
2.9 Probability been discovered by physicists had huge mathematical
significance.
There are many phenomena, from biology and eco-
There is still a big cultural difference between the
nomics to computer science and physics, that are so
complicated that instead of trying to understand them two subjects: mathematicians are far more interested
in complete detail one tries to make probabilistic state- in finding rigorous proofs, whereas physicists, who use
ments instead. For example, if you wish to analyze how mathematics as a tool, are usually happy with a con-
a disease is likely to spread, you cannot hope to take vincing argument for the truth of a mathematical state-
account of all the relevant information (such as who will ment, even if that argument is not actually a proof. The
come into contact with whom) but you can build a math- result is that physicists, operating under less stringent
ematical model and analyze it. Such models can have constraints, often discover fascinating mathematical
phenomena long before mathematicians do.
1. Here is a quick hint at a proof. At the beginning of analytic Finding rigorous proofs to back up these discoveries
number theory [IV.2] you will find a condition that tells you pre- is often extremely hard: it is far more than a pedan-
cisely which numbers can be written as sums of two squares. From
this criterion it follows that “most” numbers cannot. A careful count tic exercise in certifying the truth of statements that
shows that if N is a large integer, then there are many more expres- no physicist seriously doubted. Indeed, it often leads
sions of the form m2 +n2 with both m2 and n2 less than N than there
are numbers less than 2N that can be written as a sum of two squares.
to further mathematical discoveries. The articles ver-
Therefore there is a lot of duplication. tex operator algebras [IV.17], mirror symmetry
8 I. Introduction

[IV.16], general relativity and the einstein equa- To illustrate the sort of clarity and simplicity that is
tions [IV.13], and operator algebras [IV.15] describe needed in mathematical discourse, let us consider the
some fascinating examples of how mathematics and famous mathematical sentence “Two plus two equals
physics have enriched each other. four” as a sentence of English rather than of mathemat-
ics, and try to analyze it grammatically. On the face of it,
I.2 The Language and Grammar of it contains three nouns (“two,” “two,” and “four”), a verb
Mathematics (“equals”) and a conjunction (“plus”). However, looking
more carefully we may begin to notice some oddities.
1 Introduction For example, although the word “plus” resembles the
word “and,” the most obvious example of a conjunc-
It is a remarkable phenomenon that children can learn tion, it does not behave in quite the same way, as is
to speak without ever being consciously aware of the shown by the sentence “Mary and Peter love Paris.” The
sophisticated grammar they are using. Indeed, adults verb in this sentence, “love,” is plural, whereas the verb
too can live a perfectly satisfactory life without ever in the previous sentence, “equals,” was singular. So the
thinking about ideas such as parts of speech, subjects, word “plus” seems to take two objects (which happen
predicates, or subordinate clauses. Both children and to be numbers) and produce out of them a new, sin-
adults can easily recognize ungrammatical sentences, gle object, while “and” conjoins “Mary” and “Peter” in
at least if the mistake is not too subtle, and to do this a looser way, leaving them as distinct people.
it is not necessary to be able to explain the rules that
Reflecting on the word “and” a bit more, one finds
have been violated. Nevertheless, there is no doubt that
that it has two very different uses. One, as above, is to
one’s understanding of language is hugely enhanced by
link two nouns, whereas the other is to join two whole
a knowledge of basic grammar, and this understanding
sentences together, as in “Mary likes Paris and Peter
is essential for anybody who wants to do more with
likes New York.” If we want the basics of our language
language than use it unreflectingly as a means to a
to be absolutely clear, then it will be important to be
nonlinguistic end.
aware of this distinction. (When mathematicians are at
The same is true of mathematical language. Up to
their most formal, they simply outlaw the noun-linking
a point, one can do and speak mathematics without
use of “and”—a sentence such as “3 and 5 are prime
knowing how to classify the different sorts of words
numbers” is then paraphrased as “3 is a prime number
one is using, but many of the sentences of advanced
and 5 is a prime number.”)
mathematics have a complicated structure that is much
easier to understand if one knows a few basic terms This is but one of many similar questions: anybody
of mathematical grammar. The object of this section who has tried to classify all words into the standard
is to explain the most important mathematical “parts eight parts of speech will know that the classification is
of speech,” some of which are similar to those of nat- hopelessly inadequate. What, for example, is the role of
ural languages and others quite different. These are the word “six” in the sentence “This section has six sub-
normally taught right at the beginning of a university sections”? Unlike “two” and “four” earlier, it is certainly
course in mathematics. Much of The Companion can be not a noun. Since it modifies the noun “subsection” it
understood without a precise knowledge of mathemat- would traditionally be classified as an adjective, but
ical grammar, but a careful reading of this article will it does not behave like most adjectives: the sentences
help the reader who wishes to follow some of the later, “My car is not very fast” and “Look at that tall build-
more advanced parts of the book. ing” are perfectly grammatical, whereas the sentences
The main reason for using mathematical grammar is “My car is not very six” and “Look at that six building”
that the statements of mathematics are supposed to are not just nonsense but ungrammatical nonsense. So
be completely precise, and it is not possible to achieve do we classify adjectives further into numerical adjec-
complete precision unless the language one uses is free tives and nonnumerical adjectives? Perhaps we do, but
of many of the vaguenesses and ambiguities of ordinary then our troubles will be only just beginning. For exam-
speech. Mathematical sentences can also be highly com- ple, what about possessive adjectives such as “my” and
plex: if the parts that made them up were not clear and “your”? In general, the more one tries to refine the clas-
simple, then the unclarities would rapidly accumulate sification of English words, the more one realizes how
and render the sentences unintelligible. many different grammatical roles there are.