The links between philosophy and
mathematics are ancient and complex.
The two disciplines are in a sense coeval:
for both, the Ancient Greeks were the first to
introduce systematicity, rigor, and the
centrality of justification to their practice.
Indeed, Plato (428-348(7) BCE) had it
inscribed on the gates of his Academy that
no one should enter who
knew no mathematics.
Let us turn now to a more positive characterization of a philosophical
approach to mathematics. It will be helpful to focus on instances of
actual mathematics, so let us consider a few theorems and their proofs,
and then survey the kinds of typically philosophical issues they raise.
The first two both involve the distinction between rational and irrational
numbers. A rational number is one that can be expressed as a fraction;
for example, 3/5, —19/12, and 8/1 are all rational numbers.
An irrational number - phi, for instance - is one that cannot be expressed as
such a fraction.
The rationals and irrationals together make up the real numbers.
Our first theorem dates from Ancient Greece; it is usually atttributed to a
member of the school of Pythagoras, although precisely who first proved it
is not known.
The English mathematician G. H. Hardy (1877-1947) called it a theorem "of
the highest class. [It] is as fresh and significant as when it was discovered -
two thousand years have not written a wrinkle" on it.
This demonstrates that not all magnitudes - in particular, not the
length of the hypotenuse of a right-angled triangle of unit base and height
- can be treated by the theory of numerical proportion upon which the
mathematics of Ancient Greece was based. It must have thus constituted
something of a revolution.
These examples are in a way paradigmatic mathematics.
To be sure, mathematics is filled with proofs that are much longer and
more complicated, and with theorems that involve concepts far more
intricate than those appearing above. But most philosophical questions
about mathematics can already be raised with regard to such simple
examples.
We shall briefly examine a few of these in turn.
To begin with, note that (assuming you had not seen these proofs
before) you now know three more truths than you did a few moments
ago. How did you acquire this knowledge? One way you did not acquire
it is through observation. It is true that you used your eyes to read the
sentences of the proofs.
platonism insists that mathematics is mind-independent, in the sense that
whether a mathematical statement holds is quite independent of what we
think.
We can imagine certain realms in which the beliefs of observers in effect
settle what is true and what is not. But mathematics, according to the
platonist, is not like this: the truth or falsity of a mathematical claim is
not determined by what anyone believes about its truth value.
This, too, is a plausible position with regard to the theorems above. For
instance, the square root of 2 is irrational regardless of whether anyone
believes or wants it to be; indeed, its irrationality is not contingent on
anyone's having beliefs about it at all. This result obtains, Hardy insisted,
"not because we think so, or because our minds are shaped in one way
rather than another, but because it is so, because mathematical reality is
built that way."
The period in the foundations of mathematics that started in
1879 with the publication of Frege’s Begriffsschrift [18] and
ended in 1931 with Go¨del’s [24] U¨ ber formal
unentscheidbare S¨atze der Principia Mathematica und
verwandter Systeme can reasonably be called the classical
period. It saw the development of three major
foundational programmes: the logicism of Frege, Russell and
Whitehead, the intuitionism of Brouwer, and Hilbert’s formalist
and proof-theoretic programme.
Kant claimed that our knowledge of mathematics is synthetic
apriori and based on a faculty of intuition. Frege accepted
Kant’s claim in the case of geometry, i.e., he thought that our
knowledge of Euclidian geometry is based on pure intuition of
space. But he could not accept Kant’s explanation of our
knowledge of statements about numbers.
Frege thought of numerical statements as being objectively true or false.
Moreover, he interpreted these statements as literally being about abstract
mathematical objects that do not exist in space or time. Now the question
arose: How can we have knowledge about numbers and their properties, if
numbers are abstract objects?
Clearly we cannot interact causally with abstract entities.
In order to show that apriori knowledge of arithmetic is possible, Frege thought
it necessary and sufficient to establish the logicist thesis that arithmetic is
reducible to logic. More precisely, he wanted to show that:
(i) the concepts of arithmetic can be explicitly defined in
terms of logical concepts;
(ii) the truths of arithmetic can be derived from logical
axioms (and definitions) by purely logical rules of inference.
The following four claims are implicit in Frege’s logicist programme:
(a) Logic is (or can be presented as) an interpreted formal system (a
Begriffsschrift);
(b) It can be known apriori that the axioms of logic are true and that the
logical rules of inference preserve truth;
(c) the concepts of arithmetic are logical concepts; and
(d) the truths of arithmetic are provable in logic.
From (a) and (b) it follows that the theorems of logic are true. Since a
contradiction cannot be true, it follows that logic is consistent.
Moreover, it seems to follow from (b) that we can gain apriori
knowledge of the theorems of logic by proving them. In virtue of (d) then,
arithmetic must be consistent and its truths knowable apriori.
Hume’s
principle says that two concepts F and G have the same
cardinal number iff they are equinumerous, i.e., iff there is
a one-to-one correspondence between the objects
falling under F and the objects falling under G. In
symbols:
where F ≈ G means that there exists a one-to-one
correspondence between the objects that fall under F
and G respectively.
According to Hume’s principle, the concept of (cardinal)
number is obtained by (Fregean) abstraction from the
concept of equinumerosity between concepts (or
properties).
A milestone in mathematics is Hilbert’s Grundlagen der Geometrie [33]
from 1899.
Its importance for the conceptual development of modern mathematics
is difficult to overstate. Here Hilbert gave, for the first time, a fully
precise axiomatization of Euclidean geometry. The entities like point,
line and plane are defined only implicitly by their mutual relations.
Generalising this method of implicit definitions it became possible to
work also with complicated mathematical systems characterised
axiomatically up to structural equivalence or isomorphisms. Hilbert’s
structuralist approach, of course, goes back to Dedekind’s
characterisation in [12] of the natural number system in terms of simply
infinite systems. It was also foreshadowed by Felix Klein’s classification
of geometries using group invariants (the Erlangen programme).
Hilbert proposed his finitist consistency programme: consider a formal
system T in which all of classical mathematics can be formalised and prove
by finitistic means the consistency of T .
In this way, Hilbert wanted to prove the consistency of classical
mathematics in a particularly elementary part: “finitistic mathematics”.
When Hilbert formulated his programme, he had two significant facts
available:
(i) Classical mathematics can be represented in formal systems of set
theory or type theory.
(ii) These formal systems can be described in a finitistic manner.
