Wittgenstein’s Philosophy of Mathematics

André Porto
UFG/CNPq1
[email protected]

1. Introduction

Wittgenstein’s philosophy of mathematics is certainly one of the biggest

interpretative challenges facing any scholar interested in the thought of that great
philosopher. Not only more than half of Wittgenstein’s fabulous textual production
between his return to philosophical activity in 1929 up to almost the end of his
live (1944) was devoted to this topic,2 but we have even reasons to believe that
Wittgenstein considered his ideas on mathematics to be his most important
contribution to philosophy.3 Yet, despite all the importance Wittgenstein attached
to his own ideas on mathematics, the reception of his proposals within the
mathematical community has been anything but enthusiastic, laudatory.4 Even
otherwise highly sympathetic commentors have been quite reticent with respect
to the very quality of his work on the subject.5 In fact, we can today safely say

1
This research was supported by the grants 402705/2022-0 and 304548/2022-7, CNPq,
Brazil.
2
Cf. (Rodych, 2024, p. 1)
3
Cf. (Monk, 1971, p. 466)
4
Cf. (Monk, 2007, p. 273)
5
Thus, just to take a rather notorious example, Michael Dummett, who writes: “Many of
the thoughts are expressed in a manner which the author recognized as inaccurate or obscure;
some passages contradict others; some are quite inconclusive; some raise objections to ideas
which Wittgenstein held or had held which are not themselves stated clearly in the volume; other
passages again, particularly those on consistency and on Gödel's theorem, are of poor quality or
contain definite errors.” (Dummett, 1978, p. 166)

1
that Wittgenstein’s ideas on the topic have been largely relegated to a small
ghetto comprising just a handful of devoted scholars.6

And yet, it is the opinion of this writer that all these highly dismissive
evaluations regarding the quality of Wittgenstein’s ideas on the philosophy of
mathematics are profoundly misguided and are the result, not only of a still rather
poor understanding of the philosopher’s ideas, but also of a strongly biased (i.e.,
set theoretical, classical) view of mathematics as a whole, but also an inadequate
appraisal of the importance of alternative contemporary mathematical and logical
movements such as constructive intuitionistic type theory, category theory and
even theoretical computer science and proof-verifiers.

Contrary to the ordinary views on contributions to the philosophy of

mathematics, we believe that there are at least two central topics on the
foundations of mathematics in which Wittgenstein’s contributions loom very large.
The first one has to do with his complex and subtle discussions of the idea of a
“purely mechanical manipulation of symbols according to formal inferential rules”
(exactly what a computer is supposed to do) and the related issue of how to
construe the notion of “mathematical proof” (either a text, or a computational
process) as “demonstrating” a given mathematical assertion. The key idea here
is Wittgenstein’s insistence on the importance of the distinction between the
normative notion of a “formal specification” – determining how a procedure ought
to unfold (if it ever comes to be implemented) – and the descriptive idea of an
actual implementation process,7 – i.e., some concrete computational process,

6
Ray Monk writes: “Wittgenstein’s philosophy of mathematics has now become a
specialist scholarly field in its own right. Researchers in this field are now writing, not for logicians
or mathematicians, nor even for philosophers of mathematics. They cannot even assume they
will be read by other, more general Wittgenstein scholars. The only people they can assume will
read their work are other researchers working on Wittgenstein’s philosophy of mathematics.”
(Monk, 2007, p. 293)
7
Concerning this very distinction, we encounter the Canadian software engineering Eric
Hanner writing: “A program is a description or specification of computer behavior. A computer
executes a program by behaving according to the program, by satisfying the program. People
often confuse programs with computer behavior. They talk about what a program “does”; of

2
carried out by some computing agent (either a person, or a machine) within some
determined interval in of time.

The second topic which we believe is central to Wittgenstein’s entire

philosophy of mathematics is, of course, his notion of a “grammatical rule”. As we
will see, according to the philosopher, all accepted (i.e., “proved”) mathematical
and logical assertions8 should be viewed as being “grammatical rules”, i.e.,
statements which demarcate the limits of rational disagreement, segregating
senseful empirical propositions (i.e., propositions which for all we know could still
turn out to be true) from absurd, senseless assertions (which could never come
to describe any aspect of empirical reality). As we will also argue, there is a
surprising affinity between Wittgenstein construal of the semantics of
“grammatical rules” (particularly of mathematical assertions) and Paul Cohen’s
and Saul Kripke’s notion of “forcing”.

Before we begin our discussion, let us demarcate our intended scope. It

lies completely beyond the dimensions of this brief chapter to provide even a
cursory discussion of these two crucial components of Wittgenstein’s philosophy
of mathematics – his handling of the notion “computational process” and of
“proof”, and his construal of the notion of “(mathematical) grammatical rules”.9
Here we will have to content ourselves with a much less ambitious goal. Along

course, it just sits there on the page or screen; it is the computer that does something. … A
program is not behavior, but a specification of behavior.” (Hehner, 2004, p. 41). Compare those
remarks with rather congenial ideas made by Wittgenstein himself in (Wittgenstein, 2001, pp.
§193, §194), for example.
8
About the idea of an “accepted mathematical statement”, the American mathematician
Reuben Hersh writes: “In mathematical practice, in the real life of living mathematicians, proof is
convincing argument, as judged by qualified judges. How does this notion of proof differ from
proof in the sense of formal logic? Firstly, formal proof can exist only within a formalized theory.
… But the passage from an informal, intuitive theory to a formalized theory inevitably entails some
loss of meaning or change of meaning. … Secondly, for many mathematical investigations, full
formalization and complete formal proof, even if possible in principle, may be impossible in
practice. (Hersh. Proving is Convincing and Explaining. Pg. 147)
9
For a more comprehensive discussion of those two crucial issues in Wittgenstein’s
philosophy of mathematics cf. (Porto, forthcoming)

3
this text we will try to provide but a panoramic presentation of one of those two
central contributions Wittgenstein has made to the philosophy of mathematics:
his profoundly modal construal “(mathematical) grammatical rules”, i.e., of
(“already proved”) mathematical assertions.

2. Rules are not Assertions about Objects

Before we can even begin our sketchy presentation of the topic

“grammatical rules” in Wittgenstein’s mature philosophy, let us declare outright
an initial interpretative option which we will adopt with respect to a rather
controversial issue within his philosophy of mathematics. We are talking here
about his alleged “non-revisionism”, the idea that his philosophy of mathematics
should “leave [classical, set-theoretical] mathematics as it is”.10 We will not adopt
this “non-revisionistic” approach to his philosophy of mathematics. In fact, quite
the contrary, we believe that Wittgenstein’s philosophy of mathematics would be
simply unintelligible if we were to refuse to view it as including a profound criticism
of platonism and of traditional, set-theoretical, i.e., extensional mathematics. As
we will argue below, Wittgenstein’s philosophy of mathematics was
constructivistic – i.e., it rejected the classical notion of a “function” as an “abstract
existence of a relation” and rather identified it with an “actual possession of an
algorithm”. And his philosophy was also intuitionistic – i.e., it rejected the law of
excluded middle.

Let us step back a little and consider more broadly the family of
philosophical approaches to mathematics generally called “platonism”. If we
adopt Marco Panza and Andrea Sereni’s characterization of this family of
approaches as the idea in platonism “mathematical statements… are about
abstract objects forming a domain that those theorems describe” (Panza &
Sereni, 2013, p. 1), than, as we shall see, Wittgenstein’s construal of “rules” can
be viewed as a direct opposition to such construal of “platonism” and to its

10
Juliet Floyd goes as far as saying that non-revisionism represents “the main divide in
interpretations of Wittgenstein’s proposals” for mathematics (Floyd, 2005, p. 77). For recent
discussions on this topic, see (Panjvani, 2006) (Frascolla, 2014) (Dawson, 2016) (Nakano, 2020)

4
postulation of “objects” which could qualified as being “abstract”, “mathematical”.
According to the philosopher, unlike empirical propositions such as “There are
three windows in this room” or “Prince has blue trousers”, grammatical rules
should not be viewed as assertions about “objects”, as statements which can be
“made-true” by events in realities (i.e., the objects involved in these events do
possess the properties attributed to them). The textual evidence is quite
impressive on this point.

...mathematical propositions do not treat of numbers. Whereas

a proposition like “There are three windows in this room” does treat of the
number 3.
...a mathematical proposition is not about its constituents in the
sense in which “The sofa is in this room” is about the sofa.
Is 2 + 2 = 4 a proposition about 2 and about 4? Compare this
proposition with “There are no other men in this room than Jack and John”.
... if “There are two men here” is taken to be about 2, then it is misleading to
say 2 + 2 = 4 is about 2; for it is “about” it in a different sense. (Wittgenstein,
1976, pp. 250, 254,255)
Suppose I say “Prince has blue trousers”; that is a proposition
about trousers. (....) What about “two”? “2 + 2 = 4” ¾ but this isn’t about 2; it
is grammatical.
Turing: Isn’t it merely a question of how one extends the use
of the word “about”?
Wittgenstein: That is a most important mistake. ¾ Of course,
you can say mathematical propositions are about numbers. But if you do, you
are almost sure to be in a muddle. (Wittgenstein, 1976, p. 251)

Still according to Wittgenstein, instead of being assertions about

“purported (abstract) objects”, grammatical rules do stablish necessary
(“definitional”) connections between concepts:

I should like to say: the proof shows me a new connection, and hence
it also gives me a new concept. (RFM, V, §45, 297)
If I conceive a proof as my movement from one concept to another,
then I shall not want also to say that it is a new concept. (RFM, V, §42-3 296-
7)
An equation links two concepts; so that I can now pass from one to
the other. An equation constructs a conceptual path. But is a conceptual path
a concept? … It is surely a movement. It seems to be a movement between
two stationary points; these are the concepts. (RFM, V, §42-3 296-7)

5
 Thus, just to give a simple, illustrative example. According to the
philosopher’s interpretation of the grammatical rule “25 × 25 = 625”, an assertion
like that would state something like: there is a necessary link between the concept
“25 rows of 25 things” and the concept “625 things in total”. Thus, whenever one
was to describe an (empirical) situation as involving, say, 25 rows of 25 objects,
one would be bound to accept that that same situation could also be described
as one “comprising 625 things altogether”. The philosopher writes:

The fact that I have 25 × 25 nuts can be verified by my counting

625 nuts, but it can also be discovered in another way which is closer to the
form of expression ‘25 × 25’. And, of course, it is in the linking of these two
ways of determining a number that one point of multiplying lies”. (Wittgenstein,
1983, pp. 357, VII, § 3)

Just as in the case of those other passages above regarding demonstrated

propositions, an equation like “25 × 25 = 625” establishes a new, necessary link
between two concepts:

The equating of 252 and 625 could be said to give me a new

concept. And the proof shows what the position is regarding this equality. ¾
“To give a new concept” can only mean to introduce a new employment of a
concept, a new practice. (Wittgenstein, 1983, pp. 432, VII, §70)

3. Non-Predicativity of Rules

As we’ve said above, we believe the main point we’ve introduced in

our last section – rules are not assertions about objects – is crucial to
Wittgenstein’s entire philosophy of mathematics. Grammatical Rules are never
“made-true” by purported “(mathematical) facts”. Rules are not (veridical or
unveridical) descriptions of any “parallel abstract reality”. Any such description
would already have to presuppose those very rules of grammar (i.e., of sense)
which would made such descriptions possible to begin with and could not thus
be justified by them. Grammatical Rules are completely antecedent to the very
possibility of any empirical description, i.e., any assertions about reality (the only
“reality” Wittgenstein accepts, “empirical reality”, of course). Once again, textual
evidence is quite impressive also on this point:

6
 The conventions of grammar can’t be justified by a description
of what is represented. Any description of that kind already presupposes the
rules of grammar. That is, what counts as nonsense within the grammar that
is to be justified can’t count as making sense in the grammar of the justifying
propositions, and so on.
Grammar is not answerable to any reality.
(Grammar is not accountable to reality.) (Wittgenstein, 2005, pp. §56,
Pg 188, 184)

In fact, Wittgenstein is ready to go as far as suggesting that we should

altogether avoid using the terms “true” and “false” in connection to mathematical
rules:

The terms “sense” and “nonsense”, rather than the terms “true”
and “false”, bring out the relation of mathematical propositions to non-
mathematical propositions. (Wittgenstein, 1979, p. 152)

According to the interpretation we are sketching here,

Wittgenstein’s suggestions should thus be viewed as a direct and radical rejection
of platonism – construed as above, as the idea that “mathematical statements
should be viewed as predications on ‘(abstract, mathematical) objects’”.
Mathematical rules should not be construed as predications at all. Rather, they
should be viewed as determinations of the very meanings of the concepts
involved, of what could ever count as being an instance of this or that empirical
concept. In a very congenial formulation to the ideas we’ve being sketching here,
Alberto Coffa wrote:

[In genuine propositions] Recognition of the meaning of the

constituent phrases must be independent of and prior to the acceptance of
the statement in question. … Since geometric primitives do not acquire their
meaning prior to their incorporation into the axiomatic claims, such axioms do
not express propositions in Frege's or Russell’s sense. … it is a mistake to
conceive of the axioms of geometry as bona fide propositions (Coffa, 1991,
pp. 131, 133).

4. The Connection with the Semantics of “Forcing”

Albeit the exiguity of our discussion here – a single chapter to cover

what amounted to about half of Wittgenstein’s philosophical efforts – we are going

7
to alter our expository strategy thus far and attempt a rather ambitious
argumentative move. We are going to try to set up connections between
Wittgenstein’s ideas and some important and radical contemporary logical
advancements. As we’ve already insinuated before, the idea is to try to free
Wittgenstein scholarship from that ghetto we’ve mentioned before.11 Instead of
restricting our discussion to the usual themes carried out by that small “parish of
devoted apostles”, we will try to open his ideas up to a direct dialogue with some
important aspects of the contemporary foundational debate. Wittgenstein
proposals will be approximated to some of the most powerful and far-reaching
achievements within the current discussion on the foundations of mathematics
and of logic. We are referring here to Paul Cohen’s and Saul Kripke’s notorious
ideas on a new, non-classical semantics for logic based on their revolutionary
notion of “forcing”.12

We will not be interested here, of course, in covering the “technical

intricacies” involved in the formulation of these family of alterative semantics, as
they were, say, initially put forth by Cohen and by Kripke in the early sixties. What
matters to us here are but the general semantic intuitions. And we have to
acknowledge that those proposals are both powerful and radical. A first initial
precept is to reject a Tarskian, traditional correspondence theory of truth based
on the idea of “satisfaction”,13 an approach in which we construe “truth” as the
fact that some tuple of objects does satisfy some given property or relation. In a
characteristically uncompromising way, instead of that fixed, “eternal” ontology of

11
Cf. note 5 above.

12
Concerning Cohen’s notion of “forcing”, Raymond Smullyan and Melvin
Fitting wrote: “… a non-classical logic was invented for the occasion – its negation
behaved differently that in classical logic. A semantics was created … and used
to establish the classical independence of the continuum hypothesis. Later, other
more familiar non-classical logics were used for the purpose. Intuitionistic logic
was the basis for (Fitting, 1969)”. (Smullyan & Fitting, 1996, p. 190)
13
Cf. for example (Moreno, 2001)

8
objects determining, sub specie aeternitatis, which propositions are true, and
which are false, the opposite is the suggested. According to this new,
revolutionary semantics, we don’t have a “fixed universal domain of objects”
(which makes some assertions true, and others, false). Rather than that fixed,
eternal domain, a varying, growing ontology is boldly put forth instead. Regarding
these ideas, the category theorists William Lawvere and John Bell explain:

The interpretation of these variable sets was in terms of

subjective variation of knowledge; for any set 𝐸 [the “varying ontology”] we
have, at any given stage 𝐵, constructed certain elements of 𝐸 and proved
certain equalities between pairs of elements constructed, giving an abstract
set 𝐸(𝐵) [the ontology at 𝐵]; if 𝐴 < 𝐵 is a deeper stage of knowledge, the
transition map 𝐸(𝐵) → 𝐸(𝐴) reflects that no constructed elements are ever
lost and no proven equations are ever disproved, but the map is neither
surjective nor injective since new elements may be constructed and new
equalities proved at stage 𝐴. (Lawvere, 1976, p. 105)
…[thus] a mathematical concept may possess a fixed sense,
but a variable reference. The sense of the concept “real number” may be
taken as fixed by its definition within a local set theory, but its reference varies
with the framework of interpretation. (Bell, 2022, p. 33)

Each new assertion – each new mathematical advancement – alters not

only the available ontology (new objects are “constructed”, and new equations
are demonstrated, collapsing two singular expressions as denotations of one and
the same entity), but also alters the very extensions of the general concepts
involved, the various “sets” themselves. As we’ve anticipated before, instead of
a fixed, universal domain of objects, we have the idea of a growing ontology for
mathematics.

But, as we’ve argued elsewhere (Porto, 2015), (Porto, forthcoming, p.

Chapt. II) these ideas are surprisingly congenial to some of Wittgenstein’s most
central and ambitious proposals: Wittgenstein’s extravagant and notorious
“Semantical Mutation Thesis”. According to that thesis each new rule, each novel
mathematical result, cannot but to alter the very meanings of the various concepts
involved. As we’ve pointed out before, within the various interpretative challenges
facing the scholar who is trying to make sense of Wittgenstein’s ideas for the
philosophy of mathematics, the Semantical Mutation Thesis certainly looms quite

9
high. And, as we’ve also argued in those texts, we do not think that the option of
simply dismissing that thesis as “some transitory conceptual alternatives” –
explored by Wittgenstein during his intermediate but later rejected in his mature
years – is really adequate.14 As we’ve stressed before, the textual evidence does
not appear to bear that out at all.

When I said that a proof introduces a new concept, I meant

something like: the proof puts a new paradigm among the paradigms of the
language …One would like to say: the proof changes the grammar of our
language, changes our concepts. (Wittgenstein, 1983, pp. 166, III §31)
I should like to say: the proof shows me a new connection, and
hence it also gives me a new concept. (Wittgenstein, 1983, pp. 297, V §45)
Mathematics teaches us to operate with concepts in a new way. And
hence it can be said to change the way we work with concepts. (Wittgenstein,
1983, pp. 413, VII § 45)
What is the transition that I make from “It will be like this” to “it must
be like this”? I form a different concept. One involving something that was not
there before. (Wittgenstein, 1983, pp. 237, IV §29)
A new form has been found, constructed. But it is used to give a new
concept together with the old one. The concept is altered so that this had to
be the result. (Wittgenstein, 1983, pp. 248, IV §47)
But in that way every proof, each individual calculation makes new
connections! (Wittgenstein, 1983, pp. 181, III §47)
A mathematical proof molds our language. (Wittgenstein, 1983, pp.
196, III §71)

If we are correct, the core semantical ideas in the two proposals –

Wittgenstein’s and Cohen’s and Kripke’s semantics of “forcing” – are strikingly
similar. According to both these approaches, newly obtained mathematical
assertions are not “made-true” by any “mathematical reality” whatsoever. There
is no “prior, and independently given, fixed domain of objects” which are
supposed to either “satisfy” or “fail to satisfy” various kinds of predications
(properties, relations, etc.). Thus, in an utterly unconventional alternative

14
Cf. (Porto, 2015, p. 4). For authors who advocate that Wittgenstein’s had altogether
dropped that rather “troublesome thesis” in his mature years, cf. for example (Gerrard, 1991, p.
132) (Frascolla, 2004, pp. 180-1) (Panjvani, 2006, pp. 420-1).

10
suggestion, the very traditional way of viewing the connection between rules and
reality is reversed. Instead of the idea of assertions being valid because they
correctly describe a logically prior “reality” (which would thus “make those
assertions valid”),15 the opposite is suggested.16 Each new theorem forces some
new condition, lays down a further constraint on the range of possible “models”
of some mathematical theory.

We could also formulate the same point predicatively as saying that it is

not some previously fixed ontology of objects which determines which
predications are satisfied (or not satisfied). Rather, we should view the assertions
themselves as determining which “objects” are supposed to “exist” and under
which concepts these objects could possibly fall. Thus, according to both
Wittgenstein’s, Cohen and Kripkes’ ideas, it is not the “object” which determine
which assertions are true. Quite the opposite, it is the accepted assertions which
collectively determine what “objects” are supposed to exist and what properties
they could possibly have.

Those semantical ideas are, of course, not our usual semantic

formulations of the notions of “objects”, “ontology” and “truth”. They are quite
strange and might even sound somewhat outrageous in the beginning. But they
are not subscribed by Wittgenstein alone. The important point which we want to
bring out here is that these ideas are surprisingly shared by a vast array of
contemporary logicians, especially those connected to the category theoretical
tradition, such as William Lawvere and John Bell, above. Wittgenstein was thus
far from being alone in formulating his Semantical Mutation Thesis and the idea
that each mathematical advancement alter the very meanings of the concepts
involved.

15
We are following Wittgenstein here and avoiding the use of the terms “true” and “false”
in connection with mathematical assertions.
16
Cf. (Chateaubriand, 2001, p. 232). Chateaubriand cites Aristotle: “It is not because we
think truly that you are pale, that you are pale, but because you are pale, we who say this have
the truth” (Metaphysics 1051b7).

11
5. Two Routes to Intuitionism

We’ve stated above that Wittgenstein’s philosophy of mathematics

was not classical, extensional. Rather, it was, not only constructivistic, but also
intuitionistic. As we’ve anticipated there, we do believe Wittgenstein was a
constructivist because of his characteristic association of the notion of “function”
to that of an actual “possession of an algorithm” (to compute, or approximate, its
values). Thus, Wittgenstein rejected the traditional, set-theoretical notion of a
“function” as a “purely abstract existence of a relation”, whether or not we would
ever be aware of that relation, and even whether or not that relation could ever
be “cognizable by us”.17

As we’ve stressed before, we believe that Wittgenstein was also an

intuitionist, for he rejected the law of excluded middle. But, as we’ve argued in
detail in (Porto, forthcoming, p. Chatper II), it is important for us to realize here
that there is not a single route leading to intuitionism. Quite the contrary, there
are at least two independent itineraries which drives us towards the adoption of
the intuitionistic logic and the rejection of the law of excluded middle. First, there
is the route usually associated with Brouwer-Heyting-Kolmogorov interpretation
of the logical particles (the famous “BHK-interpretation”) and its characteristic
preference for Natural Deduction as a system which directly embody the crucial
idea of “assertibility conditions” underlaying that whole family of interpretations.
This is, of course, the route most commonly associated with intuitionistic logic.

17
We believe there can be no doubt that Wittgenstein adopted this identification of
“function” and “possession of an algorithm”. For example, in his intermediate period the
philosopher writes: “The theory of aggregates [i.e., Set Theory] attempts to grasp the infinite at a
more general level than a theory of prescriptions (Vorschriften). It says that you can’t grasp the
actual infinite by means of arithmetical symbolism at all and that therefore it can only be described
and not presented.” (Wittgenstein, 1983, pp. 206, §170), and also “But can I be in doubt whether
all the points of a line can actually be represented by arithmetical prescriptions (Vorschriften).
Can I then ever find a point for which I can show that this is not the case? If it is given by means
of a construction, then I can translate this into an arithmetical prescription (Vorschrift), … It is
clear that a point corresponds to a prescription (Vorschrift).” (Wittgenstein, 1983, pp. 222, §180)
(PR. § 180, Pg. 222).

12
But one should not forget that there is also another alternative, a second path to
that logic, and that path which is associated with the notion of “forcing” and with
the famous Brouwer-Beth-Kripke’s semantical diagrams we’ve been discussing
above.18 It is important for us to stress here that those two routes, both leading
to the same logical system of laws are, as we’ve already anticipated before, not
conceptually dependent, one from the other. A mathematician, philosopher or
logician can be an intuitionist (because of Brouwer-Beth-Kripke’s Semantics)
without ever being a constructivist. The same category-theorist William Lawvere
stresses just this point:

I would like to emphasize that recognizing the central

importance for mathematics of the Heyting predicate calculus in no way
depends on accepting a subjective idealist philosophy such as constructivism;
objectively variable sets occur (at least implicitly) every day in geometry and
physics and the fact that this variation is reflected in our minds in no way
means that it is “freely created” by our minds; but it seems to have been the
intuitionists who first succeeded in formulating the logic that holds for at least
a certain definite portion of variation in general. (Lawvere, 1976, p. 106)

Thus, let us take the very reason category-theorists like Lawvere explain
the idea of rejecting the law of excluded-middle using Brouwer-Beth-Kripke’s
diagrams.19 According to those diagrams, there is an asymmetry between
(positive) “demonstrations” and (negative) “refutations”. The opposite of a
demonstration of “¬𝑃” (i.e., the of refutation “¬𝑃”) is not demonstration of “𝑃”.
There is still a semantical distinction that the intuitionist wants to make (and which
the classical insists on ignoring), a sort of possible “intermediary step” between

18
We follow (Van Atten, 2018) and suggest adding the Brouwer’s name to that family of
semantical proposals.
19
This distinction is especially clear in the case of “identity”. A refutation of negation of
identity, “¬(𝑎 = 𝑏)”, is not equivalent to the assertion of an identity, “𝑎 = 𝑏”, but rather to the
assertion of the indistinguishability of “𝑎” and “𝑏”. Regarding this point, John Bell writes: “…
intuitionistic logic has allowed new features of the logico-mathematical landscape – invisible
through the lens of classical logic – to be discerned. Intuitionistic logic has proved to be a subtle
instrument, more delicate than classical logic, for investigating the mathematical world”. (Bell,
2001, p. 209)

13
refuting a negative assertion and demonstrating a positive one. That possible
intermediary step is, of course, a refutation “¬𝑃” (i.e., the demonstration that
“¬¬𝑃”):

𝑃𝑃
¬𝑃 ¬¬𝑃
𝑃 ¬¬𝑃

The important point which we would like to emphasize here is that we

believe all these ideas to be quite in syntony with some of Wittgenstein’s remarks
concerning the usage of “negation” in mathematics and on the law the excluded
middle. The philosopher, for example, writes:

The opposite of “there exists a law that 𝑃” is not: “there

exists a law that ~𝑃”. But if one expresses the first by means of 𝑃, and the
second by means of ~𝑃, one will get into difficulties. (Wright, 1980, pp. 272,
V §13)20

The basic semantical insight – shared by both the semantics of

“forcing” and by Wittgenstein – is the idea that mathematical statements do not
describe (either truly, of falsely) any reality, but rather that they prescribe (i.e.,
“force”) new semantical criteria all genuine (i.e., empirical) propositions have to
abide by in order to be senseful – the idea that grammatical rules are norms. That
very idea seems to be present in another remark by the philosopher concerning
the famous logical law:

When someone sets up the law of excluded middle, he is as it

were putting two pictures before us to choose from and saying that one must

20
This is of course precisely what is expressed by the contextual bar “¾” in the
intuitionistic tableaux rule for negation:
? ¬𝑃 ü
𝑃
(Bell, De Vidi, & Graham, 2001, p. 197)
As we’ve stressed elsewhere thought (Porto, forthcoming, p. Section III.16), Wittgenstein
had a purely modal (i.e., not epistemic) interpretation of these rules.

14
 correspond to the fact. But what if it is questionable whether the pictures can
be applied here? (Wittgenstein, 1983, pp. 268, V §10)

Rules are normative, they don’t describe anything, rather they

prescribe what is to count as being still “conceivable” and what should be
promptly rejected as being “absurdities”:

How it the law of excluded middle applied?

“Either there is a rule that prescribes it, or one that
forbids it.”
Assuming that there is no rule forbidding the
occurrence, ¾ why is there then supposed to be one that prescribes it?
(Wittgenstein, 1983, pp. 277, V §18) (RFM, pg. 277) [My emphasis] 21

6. Concluding Remarks

As we’ve said before, it is quite impossible do survey the immensely

rich variety of logical and foundational themes in Wittgenstein’s philosophy of
mathematics within the boundaries of a short introductory chapter like this one.
Our expository option was to concentrate our attention in but one (albeit crucial)
aspect of that philosophy: Wittgenstein’s notion of a “grammatical rule” in
connection with mathematical statements. Our effort was to argue that despite all
the apparent strangeness of some of his ideas (such as, for example, his
Semantical Mutation Thesis), Wittgenstein is surprisingly not alone, even when
advancing such extraordinary ideas. Thus, if we are right, Wittgenstein is not, as
it were, an isolated prophet preaching in a desert. If we are correct, his philosophy
of mathematics has strong conceptual affinities with some rather powerful, albeit
radical proposals within the contemporary debate surrounding mathematics and
logic, such as the idea of “forcing” and Brouwer-Beth-Kripke’s Semantics.22 But

21
Wittgenstein is ready to go as far as proposing that all (accepted) mathematical
assertions should be formulated as commandments. He writes: “Can we imagine all mathematical
propositions expressed in the imperative? For example: “Let 10 × 10 be 100” (Wittgenstein,
1983, pp. 276, V §17).
22
There is also an important similarity between Wittgenstein’s handling of the notions of
“computational process” and of “proof” and contemporary software engineering. Cf. (Porto, 2021)

15
in our opinion, in order to be able to do that we have to reject the rather
widespread idea that there is but “one acceptable way to conceive mathematics”
– i.e., as being the traditional, set theoretical mathematics – and realize that there
are several (albeit “strange” and “radical”) “alternatives” to that mainstream
approach. If by Wittgenstein’s assertion stating that “[philosophy] leaves
mathematics as it is”23 we should necessarily understand that his philosophy of
mathematics could never be critical of traditional, extensional, set theoretical
mathematics, then it is the opinion of this author that there would be no way out
for an adequate defense of his philosophical proposals regarding the foundations
of that discipline.

7. Bibliography

Bell, J. (2001). The Art of the Intelligible - An Elementary Survey of

Mathematics in its Conceptual Development. Dordrecht: Kluwer Academic
Publishers.

Bell, J. (2022). Lectures on the Foundations of Mathematics. Fonte: Home

Page: Professor John Lane Bell:
https://publish.uwo.ca/~jbell/foundations%20of%20mathematics.pdf

Bell, J., De Vidi, D., & Graham, S. (2001). Logical Options, An Introduction
to Classical and Alternative Logics. Ontario: Broadview Press.

Chateaubriand, O. (2001). Logical Forms, Part I - Truth and Description.

São Paulo: Coleção CLE.

Coffa, A. (1991). The Semantic Tradition from Kant to Carnap. Cambridge:

Cambridge University Press.

Dawson, R. (2016). Was Wittgenstein really a Constructivist about

Mathematics? Wittgenstein-Studien, 7(1), pp. 81-104.

23
(Wittgenstein L. , 2001, pp. 42, §24)

16
 Dummett, M. (1978). Truth and Other Enigmas. Cambridge: Harvard
University Press.

Fitting, M. (1969). Intuitionistic Logic, Model Theory and Forcing.

Amsterdam: North-Holland.

Floyd, J. (2005). Wittgenstein on the Philosphy of Mathematics. Em S.

Shapiro, The Oxford Hanbook of Philosophy of Mathematics and Logic (pp. 75-
128). Oxford: Oxford University Press.

Frascolla, P. (2004). Wittgenstein on Mathematical Proof. Em A. Coliva, &

E. Picardi, Wittgenstein Today (pp. 167-184). Padova: Poligrafo.

Frascolla, P. (2014). Realism, Anti-Realism, Quietism: Wittgenstein's

Stance. Grazer Philosophische Studien, 89, pp. 11-21.

Gerrard, S. (1991). Wittgenstein's Philosophies of mathematics. Synthese,

87, pp. 125-142.

Hehner, E. (2004). A Practical Theory of Programming. Nova Iorque:

Springer-Verlag.

Lawvere, W. (1976). Variable Quantities and Variable Structures in Topoi.

Em A. Heller, & M. Tierney, Algebra, Topology, and Category Theory A Collection
of Papers in Honor of Samuel Eilenberg (pp. 101-31). Cambridge: Academic
Press.

Monk, R. (1971). The Duty of Genius. New York: Penquin Books.

Monk, R. (2007). Bourgeois, Bolshevist or Anarchist? The Reception of

Wittgenstein's Philosophy of Mathematics. Em G. Kahane, E. Kanterian, & O.
Kuusela, Wittgenstein and his Interpreters - Essays in Memory of Gordon Baker
(pp. 269-295). Oxford: Blackwell Publishing Ltd.

Moreno, L. (2001). Tarskian Truth and the Correspondence Theory.

Synthese, 126, pp. 123-147.

17
 Nakano, A. (2020). Anti-Realism and Anti-Revisionism in Wittgenstein's
Philosophy of Mathematics. Grazer Philosphische Studien, 97, pp. 451-474.

Panjvani, C. (2006). Wittgenstein and Strong Mathematical

Verificationism. The Philosophical Quartely., 56, pp. 406-425.

Panza, M., & Sereni, A. (2013). Plato's Problem. Basingstoke: Palgrave

Macmillan.

Porto, A. (2015). Semantical Mutation, Algorithms and Programs.

Dissertatio, Volume Suplementar, pp. 44-76.

Porto, A. (2021). A concepção estândar de prova e o Problema de Kant.

Em G. Secco, F. Sauter, O. Esquisabel, & W. Sanz, De Mathematicae atque
Philosophiae Elegantia (pp. 150-168). College Publications.

Porto, A. (forthcoming). Mathematics and Time: A Reconstruction of

Wittgenstein's Philosophy of Mathematics.

Rodych, V. (2024). Wittgenstein's Philosophy of Mathematics. Fonte:

Stanford Encyclopedia: https://plato.stanford.edu/entries/wittgenstein-
mathematics/

Smullyan, R., & Fitting, M. (1996). Set Theory and the Continuum Problem.
Oxford: Claredon Press.

Wittgenstein, L. (1976). Lectures no the Foundations of Mathematics,

1939. Chicago: The Universtiy of Chicago Press.

Wittgenstein, L. (1979). Wittgenstein's Lectures: Cambridge 1932-1935.

Totowa: Rowman & Littlefield.

Wittgenstein, L. (1983). Remarks on the Foundations of Mathematics.

Cambridge: Mit.

Wittgenstein, L. (1983). Remarks on the Foundations of Mathematics.

Cambridge: Mit.

18
 Wittgenstein, L. (2001). Philosophical Investigations. Oxford: Blacwell
Publishing.

Wittgenstein, L. (2005). The Big Typescript. Malden: Blackwell Publishing.

Wright, C. (1980). Wittgenstein on the Foundations of Mathematics.

Cambridge: Harvard.

19