The Three Crises in Mathematics:

Logicism, Intuitionism and Formalism

Crises in classical philosophy reveal doubts

about mathematical and philosophical criteria
for a satisfactory foundation for mathematics.

ERNST SNAPPER
Dartmouth College
Hanover, NH 03755

The three schools, mentioned in the title, all tried to give a firm foundation to mathematics.
The three crises are the failures of these schools to complete their tasks. This article looks at
these crises "through modem eyes," using whatever mathematics is available today and not just
the mathematics which was available to the pioneers who created these schools. Hence, this
article does not approach the three crises in a strictly historical way. This article also does not
discuss the large volume of current, technical mathematics which has arisen out of the
techniques introduced by the three schools in question. One reason is that such a discussion
would take a book and not a short article. Another one is that all this technical mathematics has
very little to do with the philosophy of mathematics, and in this article I want to stress those
aspects of logicism, intuitionism, and formalism which show clearly that these schools are
founded in philosophy.
Logicism
This school was started in about 1884 by the German philosopher, logician and mathemati-
cian, Gottlob Frege (1848-1925). The school was rediscovered about eighteen years later by
Bertrand Russell. Other early logicists were Peano and Russell's coauthor of Principia Mathe-
matica, A. N. Whitehead. The purpose of logicism was to show that classical mathematics is part
of logic. If the logicists had been able to carry out their program successfully, such questions as
"Why is classical mathematics free of contradictions?" would have become "Why is logic free of
contradictions?". This latter question is one on which philosophers have at least a thorough
handle and one may say in general that the successful completion of the logicists' program
would have given classical mathematics a firm foundation in terms of logic.
Clearly, in order to carry out this program of the logicists, one must first, somehow, define
what "classical mathematics" is and what "logic" is. Otherwise, what are we supposed to show is
part of what? It is precisely at these two definitions that we want to look through modem eyes,
imagining that the pioneers of logicism had all of present-day mathematics available to them.
We begin with classical mathematics.
In order to carry out their program, Russell and Whitehead created Principia Mathematica
[10] which was published in 1910. (The first volume of this classic can be bought for $3.45!
Thank heaven, only modem books and not the classics have become too expensive for the
average reader.) Principia, as we will refer to Principia Mathematica, may be considered as a
formal set theory. Although the formalization was not entirely complete, Russell and Whitehead
thought that it was and planned to use it to show that mathematics can be reduced to logic.
They showed that all classical mathematics, known in their time, can be derived from set theory
and hence from the axioms of Principia. Consequently, what remained to be done, was to show
that all the axioms of Principia belong to logic.

VOL. 52, NO. 4, SEPTEMBER 1979 207

 Of course, instead of Principia, one can use any other formal set theory just as well. Since
today the formal set theory developed by Zermelo and Fraenkel (ZF) is so much better known
than Principia, we shall from now on refer to ZF instead of Principia. ZF has only nine axioms
and, although several of them are actually axiom schemas, we shall refer to all of them as
"axioms." The formulation of the logicists' program now becomes: Show that all nine axioms of
ZF belong to logic.
This formulation of logicism is based on the thesis that classical mathematics can be defined
as the set of theorems which can be proved within ZF. This definition of classical mathematics is
far from perfect, as is discussed in [12]. However, the above formulation of logicism is
satisfactory for the purpose of showing that this school was not able to carry out its program.
We now turn to the definition of logic.

In order to understand logicism, it is very important to see clearly what the logicists meant by
"logic." The reason is that, whatever they meant, they certainly meant more than classical logic.
Nowadays, one can define classical logic as consistingof all those theorems which can be proven
in first order languages (discussed below in the section on formalism) without the use of
nonlogical axioms. We are hence restricting ourselves to first order logic and use the deduction
rules and logical axioms of that logic. An example of such a theorem is the law of the excluded
middle which says that, if p is a proposition, then either p or its negation •P is true; in other
words, the propositionpV •P is always true where Vis the usual symbol for the inclusive "or."
If this definition of classical logic had also been the logicists' definition of logic, it would be a
folly to think for even one second that all of ZF can be reduced to logic. However, the logicists'
definition was more extensive. They had a general concept as to when a proposition belongs to
logic, that is, when a proposition should be called a "logical proposition." They said: A logical
proposition is a proposition which has complete generality and is true in virtue of its form rather
than its content. Here, the word "proposition" is used as synonymous with "theorem."
For example, the above law of the excluded middle "pV •P" is a logical proposition.
Namely, this law does not hold because of any special content of the propositionp; it does not
matter whether p is a proposition of mathematics or physics or what have you. On the contrary,
this law holds with "complete generality," that is, for any proposition p whatsoever. Why then
does it hold? The logicists answer: "Because of its form." Here they mean by form "syntactical
form," the form of pV •P being given by the two connectives of everyday speech, the inclusive
"or" and the negation "not" (denoted by V and ...,, respectively).
On the one hand, it is not difficult to argue that all theorems of classical logic, as defined
above, are logical propositions in the sense of logicism. On the other hand, there is no a priori
reason to believe that there could not be logical propositions which lie outside of classical logic.
This is why we said that the logicists' definition of logic is more extensive than the definition of
classical logic. And now the logicists' task becomes clearer: It consists in showing that all nine
axioms of ZF are logical propositions in the sense of logicism.
The only way to assess the success or failure of logicism in carrying out this task is by going
through all nine axioms of ZF and determining for each of them whether it falls under the
logicists' concept of a logical proposition. This would take a separate article and would be of
interest only to readers who are thoroughly familiar with ZF. Hence, instead, we simply state
that at least two of these axioms, namely, the axiom of infinity and the axiom of choice, cannot
possibly be considered as logical propositions. For example, the axiom of infinity says that there
exist infinite sets. Why do we accept this axiom as being true? The reason is that everyone is
familiar with so many infinite sets, say, the set of the natural numbers or the set of points in
Euclidean 3-space. Hence, we accept this axiom on grounds of our everyday experience with
sets, and this clearly shows that we accept it in virtue of its content and not in virtue of its
syntactical form. In general, when an axiom claims the existence of objects with which we are
familiar on grounds of our common everyday experience, it is pretty certain that this axiom is
not a logical proposition in the sense of logicism.

208 MATHEMATICS MAGAZINE

 And here then is the first crisis in mathematics: Since at least two out of the nine axioms of
ZF are not logical propositions in the sense of logicism, it is fair to say that this school failed by
about 20% in its effort to give mathematics a firm foundation. However, logicism has been of the
greatest importance for the development of modem mathematical logic. In fact, it was logicism
which started mathematical logic in a serious way. The two quantifiers, the "for all" quantifier V
and the "there exists" quantifier 3 were introduced into logic by Frege [5], and the influence of
Principia on the development of mathematical logic is history.
It is important to realize that logicism is founded in philosophy. For example, when the
logicists tell us what they mean by a logical proposition (above), they use philosophical and not
mathematical language. They have to use philosophical language for that purpose since mathe-
matics simply cannot handle definitions of so wide a scope.
The philosophy of logicism is sometimes said to be based on the philosophical school called
"realism." In medieval philosophy "realism" stood for the Platonic doctrine that abstract entities
have an existence independent of the human mind. Mathematics is, of course, full of abstract
entities such as numbers, functions, sets, etc., and according to Plato all such entities exist
outside our mind. The mind can discover them but does not create them. This doctrine has the
advantage that one can accept such a concept as "set" without worrying about how the mind
can construct a set. According to realism, sets are there for us to discover, not to be constructed,
and the same holds for all other abstract entities. In short, realism allows us to accept many
more abstract entities in mathematics than a philosophy which had limited us to accepting only
those entities the human mind can construct. Russell was a realist and accepted the abstract
entities which occur in classical mathematics without questioning whether our own minds can
construct them. This is the fundamental difference between logicism and intuitionism, since in
intuitionism abstract entities are admitted only if they are man made.
Excellent expositions of logicism can be found in Russell's writing, for example [9], [10] and
[11].

VOL. 52, NO. 4, SEPTEMBER 1979 209

Intuitionism
This school was begun about I908 by the Dutch mathematician, L. E. J. Brouwer (I88I-
1966). The intuitionists went about the foundations of mathematics in a radically different way
from the logicists. The logicists never thought that there was anything wrong with classical
mathematics; they simply wanted to show that classical mathematics is part of logic. The
intuitionists, on the contrary, felt that there was plenty wrong with classical mathematics.
By 1908, several paradoxes had arisen in Cantor's set theory. Here, the word "paradox" is
used as synonymous with "contradiction." Georg Cantor created set theory, starting around
1870, and he did his work "naively," meaning nonaxiomatically. Consequently, he formed sets
with such abandon that he himself, Russell and others found several paradoxes within his
theory. The logicists considered these paradoxes as common errors, caused by erring mathemati-
cians and not by a faulty mathematics. The intuitionists, on the other hand, considered these
paradoxes as clear indications that classical mathematics itself is far from perfect. They felt that
mathematics had to be rebuilt from the bottom on up.
The "bottom," that is, the beginning of mathematics for the intuitionists, is their explanation
of what the natural numbers 1, 2, 3,... are. (Observe that we do not include the number zero
among the natural numbers.) According to intuitionistic philosophy, all human beings have a
primordial intuition for the natural numbers within them. This means in the first place that we
have an immediate certainty as to what is meant by the number I and, secondly, that the mental
process which goes into the formation of the number I can be repeated. When we do repeat it,
we obtain the concept of the number 2; when we repeat it again, the concept of the number 3; in
this way, human beings can construct any finite initial segment I, 2, ... , n for any natural number
n. This mental con~truction of one natural number after the other would never have been
possible if we did not have an awareness of time within us. "After" refers to time and Brouwer
agrees with the philosopher Immanuel Kant (1724--1804) that human beings have an immediate
awareness of time. Kant used the word "intuition" for "immediate awareness" and this is where
the name "intuitionism" comes from. (See Chapter IV of [4] for more information about this
intuitionistic concept of natural numbers.)
It is important to observe that the intuitionistic construction of natural numbers allows one to
construct only arbitrarily long finite initial segments I, 2, ... , n. It does not allow us to construct
that whole closed set of all the natural numbers which is so familiar from classical mathematics.
It is equally important to observe that this construction is both "inductive" and "effective." It is
inductive in the sense that, if one wants to construct, say, the number 3, one has to go through
all the mental steps of first constructing the I, then the 2, and finally the 3; one cannot just grab
the number 3 out of the sky. It is effective in the sense that, once the construction of a natural
number has been finished, that natural number has been constructed in its entirety. It stands
before us as a completely finished mental construct, ready for our study of it. When someone
says, "I have finished the mental construction of the number 3," it is like a bricklayer saying, "I
have finished that wall," which he can say only after he has laid every stone in place.
We now turn to the intuitionistic definition of mathematics. According to intuitionistic
philosophy, mathematics should be defined as a mental activity and not as a set of theorems (as
was done above in the section on logicism). It is the activity which consists in carrying out, one
after the other, those mental constructions which are inductive and effective in the sense in
which the intuitionistic construction of the natural numbers is inductive and effective. In-
tuitionism maintains that human beings are able to recognize whether a given mental construc-
tion has these two properties. We shall refer to a mental construction which has these two
properties as a construct and hence the intuitionistic definition of mathematics says: Mathemat-
ics is the mental activity which consists in carrying out constructs one after the other.
A major consequence of this definition is that all of intuitionistic mathematics is effective or
"constructive" as one usually says. We shall use the adjective "constructive" as synonymous
with "effective" from now on. Namely, every construct is constructive, and intuitionistic
mathematics is nothing but carrying out constructs over and over. For instance, if a real number

210 MATHEMATICS MAGAZINE

r occurs in an intuitionistic proof or theorem, it never occurs there merely on grounds of an
existence proof. It occurs there because it has been constructed from top to bottom. This implies
for example that each decimal place in the decimal expansion of r can in principle be computed.
In short, all intuitionistic proofs, theorems, definitions, etc., are entirely constructive.
Another major consequence of the intuitionistic definition of mathematics is that mathemat-
ics cannot be reduced to any other science such as, for instance, logic. This definition comprises
too many mental processes for such a reduction. And here, then, we see a radical difference
between logicism and intuitionism. In fact, the intuitionistic attitude toward logic is precisely the
opposite from the logicists' attitude: According to the intuitionists, whatever valid logical
processes there are, they are all constructs; hence, the valid part of classical logic is part of
mathematics! Any law of classical logic which is not composed of constructs is for the
intuitionist a meaningless combination of words. It was, of course, shocking that the classical
law of the excluded middle turned out to be such a meaningless combination of words. This
implies that this law cannot be used indiscriminately in intuitionistic mathematics; it can often
be used, but not always.
Once the intuitionistic definition of mathematics has been understood and accepted, all there
remains to be done is to do mathematics the intuitionistic way. Indeed, the intuitionists have
developed intuitionistic arithmetic, algebra, analysis, set theory, etc. However, in each of these
branches of mathematics, there occur classical theorems which are not composed of constructs
and, hence, are meaningless combinations of words for the intuitionists. Consequently, one
cannot say that the intuitionists have reconstructed all of classical mathematics. This does not
bother the intuitionists since whatever parts of classical mathematics they cannot obtain are
meaningless for them anyway. Intuitionism does not have as its purpose the justification of
classical mathematics. Its purpose is to give a valid definition of mathematics and then to "wait
and see" what mathematics comes out of it. Whatever classical mathematics cannot be done
intuitionistically simply is not mathematics for the intuitionist. We observe here another
fundamental difference between logicism and intuitionism: The logicists wanted to justify all of
classical mathematics. (An excellent introduction to the actual techniques of intuitionism is [8].)
Let us now ask how successful the intuitionistic school has been in giving us a good
foundation for mathematics, acceptable to the majority of mathematicians. Again, there is a
sharp difference between the way this question has to be answered in the present case and in the
case of logicism. Even hard-nosed logicists have to admit that their school so far has failed to
give mathematics a firm foundation by about 20%. However, a hard-nosed intuitionist has every
right in the world to claim that intuitionism has given mathematics an entirely satisfactory
foundation. There is the meaningful definition of intuitionistic mathematics, discussed above;
there is the intuitionistic philosophy which tells us why constructs can never give rise to
contradictions and, hence, that intuitionistic mathematics is free of contradictions. In fact, not
only this problem (of freedom from contradiction) but all other problems of a foundational
nature as well receive perfectly satisfactory solutions in intuitionism.
Yet if one looks at intuitionism from the outside, namely, from the viewpoint of the classical
mathematician, one has to say that intuitionism has failed to give mathematics an adequate
foundation. In fact, the mathematical community has almost universally rejected intuitionism.
Why has the mathematical community done this, in spite of the many very attractive features of
intuitionism, some of which have just been mentioned?
One reason is that classical mathematicians flatly refuse to do away with the many beautiful
theorems that are meaningless combinations of words for the intuitionists. An example is the
Brouwer fixed point theorem of topology which the intuitionists reject because the fixed point
cannot be constructed, but can only be shown to exist on grounds of an existence proof. This, by
the way, is the same Brouwer who created intuitionism; he is equally famous for his work in
(nonintuitionistic) topology.
A second reason comes from theorems which can be proven both classically and intuitionisti-
cally. It often happens that the classical proof of such a theorem is short, elegant, and devilishly

VOL. 52, NO. 4, SEPTEMBER 1979 211

 clever, but not constructive. The intuitionists will of course reject such a proof and replace it by
their own constructive proof of the same theorem. However, this constructive proof frequently
turns out to be about ten times as long as the classical proof and often seems, at least to the
classical mathematician, to have lost all of its elegance. An example is the fundamental theorem
of algebra which in classical mathematics is proved in about half a page, but takes about ten
pages of proof in intuitionistic mathematics. Again, classical mathematicians refuse to believe
that their clever proofs are meaningless whenever such proofs are not constructive.
Finally, there are the theorems which hold in intuitionism but are false in classical mathemat-
ics. An example is the intuitionistic theorem which says that every real-valued function which is
defined for all real numbers is continuous. This theorem is not as strange as it sounds since it
depends on the intuitionistic concept of a function: A real-valued function f is defined in
intuitionism for all real numbers only if, for every real number r whose intuitionistic construc-
tion has been completed, the real number f(r) can be constructed. Any obviously discontinuous
function a classical mathematician may mention does not satisfy this constructive criterion.
Even so, theorems such as this one seem so far out to classical mathematicians that they reject
any mathematics which accepts them.
These three reasons for the rejection of intuitionism by classical mathematicians are neither
rational nor scientific. Nor are they pragmatic reasons, based on a conviction that classical
mathematics is better for applications to physics or other sciences than is intuitionism. They are
all emotional reasons, grounded in a deep sense as to what mathematics is all about. (If one of
the readers knows of a truly scientific rejection of intuitionism, the author would be grateful to
hear about it.) We now have the second crisis in mathematics in front of us: It consists in the
failure of the intuitionistic school to make intuitionism acceptable to at least the majority of
mathematicians.
It is important to realize that, like logicism, intuitionism is rooted in philosophy. When, for
instance, the intuitionists state their definition of mathematics, given earlier, they use strictly
philosophical and not mathematical language. It would, in fact, be quite impossible for them to
use mathematics for such a definition. The mental activity which is mathematics can be defined
in philosophical terms but this definition must, by necessity, use some terms which do not
belong to the activity it is trying to define.
Just as logicism is related to realism, intuitionism is related to the philosophy called
"conceptualism." This is the philosophy which maintains that abstract entities exist only insofar
as they are constructed by the human mind. This is very much the attitude of intuitionism which
holds that the abstract entities which occur in mathematics, whether sequences or order-relations
or what have you, are all mental constructions. This is precisely why one does not find in
intuitionism the staggering collection of abstract entities which occur in classical mathematics
and hence in logicism. The contrast between logicism and intuitionism is very similar to the
contrast between realism and conceptualism.
A very good way to get into intuitionism is by studying [8], Chapter IV of [4], [2] and [13], in
this order.

Formalism
This school was created in about 1910 by the German mathematician David Hilbert
(1862-1943). True, one might say that there were already formalists in the nineteenth century
since Frege argued against them in the second volume of his Grundgesetze der Arithmetik (see
the book by Geach and Black under [5], pages 182-233); the first volume of the Grundgesetze
appeared in 1893 and the second one in 1903. Nevertheless, the modern concept of formalism,
which includes finitary reasoning, must be credited to Hilbert. Since modern books and courses
in mathematical logic usually deal with formalism, this school is much better known today than
either logicism or intuitionism. We will hence discuss only the highlights of formalism and begin
by asking, "What is it that we formalize when we formalize something?"

212 MATHEMATICS MAGAZINE

 LIS

The answer is that we formalize some given axiomatized theory. One should guard against
confusing axiomatization and formalization. Euclid axiomatized geometry in about 300 B.C., but
formalization started only about 2200 years later with the logicists and formalists. Examples of
axiomatized theories are Euclidean plane geometry with the usual Euclidean axioms, arithmetic
with the Peano axioms, ZF with its nine axioms, etc. The next question is: "How do we
formalize a given axiomatized theory?"
Suppose then that some axiomatized theory T is given. Restricting ourselves to first order
logic, "to formalize T" means to choose an appropriate first order language for T. The
vocabulary of a first order language consists of five items, four of which are always the same
and are not dependent on the given theory T. These four items are the following: (I) A list of
denumerably many variables-who can talk about mathematics without using variables? (2)
Symbols for the connectives of everyday speech, say -, for "not," A for "and," V for the
inclusive "or," ~ for "if then," and~ for "if and only if"-who can talk about anything at all
without using connectives? (3). The equality sign =; again, no one can talk about mathematics
without using this sign. (4) The two quantifiers, the "for all" quantifier V and the "there exist"
quantifier 3; the first one is used to say such things as "all complex numbers have a square
root," the second one to say things like "there exist irrational numbers." One can do without
some of the above symbols, but there is no reason to go into that. Instead, we turn to the fifth
item.
Since T is an axiomatized theory, it has so called "undefined terms." One has to choose an
appropriate symbol for every undefined term of T and these symbols make up the fifth item. For

VOL. 52, NO. 4, SEPTEMBER 1979 213

instance, among the undefined terms of plane Euclidean geometry, occur "point," "line," and
"incidence," and for each one of them an appropriate symbol must be entered into the
vocabulary of the first order language. Among the undefined terms of arithmetic occur "zero,"
"addition," and "multiplication," and the symbols one chooses for them are of course 0, +,and
X, respectively. The easiest theory of all to formalize is ZF since this theory has only one
undefined term, namely, the membership relation. One chooses, of course, the usual symbol E
for that relation. These symbols, one for each undefined term of the axiomatized theory T, are
often called the "parameters" of the first order language and hence the parameters make up the
fifth item.
Since the parameters are the only symbols in the vocabulary of a first order language which
depend on the given axiomatized theory T, one formalizes T simply by choosing these
parameters. Once this choice has been made, the whole theory T has been completely for-
malized. One can now express in the resulting first order language L not only all axioms,
definitions, and theorems of T, but more! One can also express in L all axioms of classical logic
and, consequently, also all proofs one uses to prove theorems of T. In short, one can now
proceed entirely within L, that is, entirely "formally."
But now a third question presents itself: "Why in the world would anyone want to formalize
a given axiomatized theory?" After all, Euclid never saw a need to formalize his axiomatized
geometry. It is important to ask this question, since even the great Peano had mistaken ideas
about the real purpose of formalization. He published one of his most important discoveries in
differential equations in a formalized language (very similar to a first order language) with the
result that nobody read it until some charitable soul translated the article into common German.
Let us now try to answer the third question. If mathematicians do technical research in a
certain branch of mathematics, say, plane Euclidean geometry, they are interested in discovering
and proving the important theorems of the branch of mathematics. For that kind of technical
work, formalization is usually not only no help but a definite hindrance. If, however, one asks
such foundational questions as, for instance, "Why is this branch of mathematics free of
contradictions?", then formalization is not just a help but an absolute necessity.
It was really Hilbert's stroke of genius to understand that formalization is the proper
technique to tackle such foundational questions. What he taught us can be put roughly as
follows. Suppose that Tis an axiomatized theory which has been formalized in terms of the first
order language L. This language has such a precise syntax that it itself can be studied as a
mathematical object. One can ask for instance: "Can one possibly run into contradictions if one
proceeds entirely formally within L, using only the axioms of T and those of classical logic, all
of which have been expressed in L ?" If one can prove mathematically that the answer to this
question is "no," one has there a mathematical proof that the theory Tis free of contradictions!
This is basically what the famous "Hilbert program" was all about. The idea was to formalize
the various branches of mathematics and then to prove mathematically that each one of them is
free of contradictions. In fact if, by means of this technique, the formalists could have just
shown that ZF is free of contradictions, they would thereby already have shown that all of
classical mathematics is free of contradictions, since classical mathematics can be done
axiomatically in terms of the nine axioms of ZF. In short, the formalists tried to create a
mathematical technique by means of which one could prove that mathematics is free of
contradictions. This was the original purpose of formalism.

It is interesting to observe that both logicists and formalists formalized the various branches
of mathematics, but for entirely different reasons. The logicists wanted to use such a formaliza-
tion to show that the branch of mathematics in question belongs to logic; the formalists wanted
to use it to prove mathematically that that branch is free of contradictions. Since both schools
"formalized," they are sometimes confused.
Did the formalists complete their program successfully? No! In 1931, Kurt Godel showed in
[6] that formalization cannot be considered as a mathematical technique by means of which one

214 MATHEMATICS MAGAZINE

can prove that mathematics is free of contradictions. The theorem in that paper which rang the
death bell for the Hilbert program .concerns axiomatized theories which are free of contradic-
tions and whose axioms are strong enough so that arithmetic can be done in terms of them.
Examples of theories whose axioms are that strong are, of course, Peano arithmetic and ZF.
Suppose now that T is such a theory and that T has been formalized by means of the first order
language L. Then Godel's theorem says, in nontechnical language, "No sentence of L which can
be interpreted as asserting that T is free of contradictions can be provei}. formally within the
language L." Although the interpretation of this theorem is somewhat controversial, most
mathematicians have concluded from it that the Hilbert program cannot be carried out:
Mathematics is not able to prove its own freedom of contradictions. Here, then, is the third crisis
in mathematics.
Of course, the tremendous importance of the formalist school for present-day mathematics is
well known. It was in this school that modem mathematical logic and its various offshoots, such
as model theory, recursive function theory, etc., really came into bloom.
Formalism, as logicism and intuitionism, is founded in philosophy, but the philosophical
roots of formalism are somewhat more hidden than those of the other two schools. One can find
them, though, by reflecting a little on the Hilbert program.
Let again T be an axiomatized theory which has been formalized in terms of the first order
language L. In carrying out Hilbert's program, one has to talk about the language L as one
object, and while doing this, one is not talking within that safe language L itself. On the
contrary, one is talking about Lin ordinary, everyday language, be it English or French or what
have you. While using our natural language and not the formal language L, there is of course
every danger that contradictions, in fact, any kind of error, may slip in. Hilbert said that the way
to avoid this danger is by making absolutely certain that, while one is talking in one's natural
language about L, one uses only reasonings which are absolutely safe and beyond any kind of
suspicion. He called such reasonings "finitary reasonings," but had, of course, to give a
definition of them. The most explicit definition of finitary reasoning known to the author was
given by the French formalist Herbrand ([7], the footnote on page 622). It says, if we replace
"intuitionistic" by "finitary":
By a finitary argument we understand an argument satisfying the following conditions: In
it we never consider anything but a given finite number of objects and of functions; these
functions are well defined, their definition allowing the computation of their values in a
univocal way; we never state that an object exists without giving the means of constructing it;
we never consider the totality of all the objects x of an infinite collection; and when we say
that an argument (or a theorem) is ·true for all these x, we mean that, for each x taken by
itself, it is possible to repeat the general argument in question, which should be considered to
be merely the prototype of these particular arguments.

Observe that this definition uses philosophical and not mathematical language. Even so, no
one can claim to understand the Hilbert program without an understanding of what finitary
reasoning amounts to. The philosophical roots of formalism come out into the open when the
formalists define what they mean by finitary reasoning.
We have already compared logicism with realism, and intuitionism with conceptualism. The
philosophy which is closest to formalism is "nominalism." This is the philosophy which claims
that abstract entities have no existence of any kind, neither outside the human mind as
maintained by realism, nor as mental constructions within the human mind as maintained by
conceptualism. For nominalism, abstract entities are mere vocal utterances or written lines, mere
names. This is where the word "nominalism" comes from, since in Latin nomina/is means
"belonging to a name." Similarly, when formalists try to prove that a certain axiomatized theory
T is free of contradictions, they do not study the abstract entities which occur in T but, instead,
study that first order language L which was used to formalize T. That is, they study how one can
form sentences in L by the proper use of the vocabulary of L; how certain of these sentences
can be proven by the proper use of those special sentences of L which were singled out as

VOL. 52, NO. 4, SEPTEMBER 1979 215

axioms; and, in particular, they try to show that no sentence of L can be proven and disproven
at the same time, since they would thereby have established that the original theory T is free of
contradictions. The important point is that this whole study of L is a strictly syntactical study,
since no meanings or abstract entities are associated with the sentences of L. This language is
investigated by considering the sentences of L as meaningless expressions which are manipu-
lated according to explicit, syntactical rules, just as the pieces of a chess game are meaningless
figures which are pushed around according to the rules of the game. For the strict formalist "to
do mathematics" is "to manipulate the meaningless symbols of a first order language according
to explicit, syntactical rules." Hence, the strict formalist does not work with abstract entities,
such as infinite series or cardinals, but only with their meaningless names which are the
appropriate expressions in a· first order language. Both formalists and nominalists avoid the
direct use of abstract entities, and this is why formalism should be compared with nominalism.
The fact that logicism, intuitionism, and formalism correspond to realism, conceptualism, and
nominalism, respectively, was brought to light in Quine's article, "On What There Is" ([1], pages
183-196). Formalism can be learned from any modem book in mathematical logic, for instance
[3].
Epilogue
Where do the three crises in mathematics leave us? They leave us without a firm foundation
for mathematics. After Godel's paper [6] appeared in 1931, mathematicians on the whole threw
up their hands in frustration and turned away from the philosophy of mathematics. Neverthe-
less, the influence of the three schools discussed in this article has remained strong, since they
have given us much new and beautiful mathematics. This mathematics concerns mainly set
theory, intuitionism and its various constructivist modifications, and mathematical logic with its
many offshoots. However, although this kind of mathematics is often referred to as "foundations
of mathematics," one cannot claim to be advancing the philosophy of mathematics just because
one is working in one of these areas. Modem mathematical logic, set theory, and intuitionism
with its modifications are nowadays technical branches of mathematics, just as algebra or
analysis, and unless we return directly to the philosophy of mathematics, we cannot expect to
find a firm foundation for our science. It is evident that such a foundation is not necessary for
technical mathematical research, but there are still those among us who yearn for it. The author
believes that the key to the foundations of mathematics lies hidden somewhere among the
philosophical roots of logicism, intuitionism, and formalism and this is why he has uncovered
these roots, three times over.
Excellent literature on the foundations of mathematics is contained in [1] and [7].

References
[1] P. Benacerraf and H. Putnam, Philosophy of Mathematics, Prentice-Hall, 1964.
[2] M. Dummett, Elements of Intuitionism, Clarendon Press, Oxford, England, 1977.
[3] H. B. Enderton, A Mathematical Introduction to Logic, Academic Press, 1972.
[4] A. A. Fraenkel, Y. Bar-Hillel, and A. Levy, Foundations of Set Theory, North-Holland, Amsterdam,
Netherlands, 1973.
[5] G. Frege, Begriffschrift, in Translations from the Philosophical Writings of Gottlob Frege by P. Geach and
M. Black, Basil Blackwell, Oxford, England, 1970. Also in [7] pp. 1-82.
[6] K. Godel, On formally undecidable propositions of Principia Mathematica and related systems, in [7] pp.
596-616.
[7] J. van Heijenoort, From Frege to GOdel, Harvard Univ. Press, Cambridge. Available in paperback.
[8] A. Heyting, Intuitionism, An Introduction, North-Holland, Amsterdam, Netherlands, 1966.
[9] B. Russell, Principles of Mathematics, 1st ed. (1903) W. W. Norton, New York. Available in paperback.
[10] B. Russell and A. N. Whitehead, Principia Mathematicia, 1st ed. (1910) Cambridge Univ. Press, Cam-
bridge, England. Available in paperback.
[II] B. Russell, Introduction to Mathematical Philosophy, Simon and Schuster, New York, 1920. Available in
paperback.
[12] E. Snapper, What is mathematics?, Amer. Math. Monthly, no. 7, 86 (1979) 551-557.
[13] A. S. Troelstra, Choice Sequences, Oxford Univ. Press, Oxford, England, 1977.

216 MATHEMATICS MAGAZINE