← Back to writing

Can One Give a Purely Empirical Anti-Realist Account of Mathematical Truth?

In this essay, I will argue that one cannot give a purely empirical anti-realist account of why mathematical truths are so convincing, basing it on Kitcher's idealisation theory and Field's nominalisation. The steps in my argument for this thesis are: 1) I will present P. Kitcher's argument for an empirical account of mathematical truth, based on idealisation theory; 2) I will use S. Hoffman's argument to show why Kitcher's approach fails to achieve the truth required and hence fails to preserve mathematical truth in mathematical empiricism; 3) I will suggest that Kitcher's idealisation theory can be transformed into a nominalist theory and present an anti-realist account of why mathematical truths are convincing using Field's nominalization; 4) I will show that correct Field's nominalisation is impossible and hence, fails to show why mathematical truths are so convincing; 5) I will conclude that one cannot give a purely empirical anti-realist account of why mathematical truths are so convincing when building this account on either Kitcher's idealisation theory or Field's nominalisation programme.

I first explain the two main theories of how to understand mathematical truth. These are realism, or its analogue in philosophy of mathematics - Platonism, and empiricism. The first implies that "mathematical entities exist, that they are abstract, and that they are independent of all our rational activities". The second denies the existence of mathematical reality and states that all knowledge of mathematical facts comes from our sensory experience and observations.

1) Kitcher's Idealisation Theory

In The Nature of Mathematical Knowledge Kitcher argues for the empirical account of mathematical knowledge and mathematical truth. By adopting the view that mathematics is rooted in empirical observations, he rejects Platonism. Moreover, Kitcher gives a strong argument for preserving the mathematical truth and avoiding the Benacerraf's dilemma. I summarise his argument as follows:

P1 Mathematics is the product of "generalized human empirical operations performed on physical objects"

P2 Mathematics is an idealised science of human operations

P3 Statements of idealised science are vacuously true

C Mathematical statements are true

The second premise shows that Kitcher views mathematics as an idealisng theory. This means that to develop mathematical statements, which concern infinity, one can imagine an ideal agent who is able to give an empirical account of infinities through performing an infinite number of empirical operations. The "operations" refer to activities of such an agent, which are not abstract objects and do not entail epistemological difficulties. Kitcher develops a set of axioms in Mill Arithmetic; he simply stipulates these axioms to be true in a world where an ideal agent performs them (Kitcher, 1984, pp. 113-177). To illustrate his idea and to secure the third premise, I give Kitcher's example of idealisation theory, which is a reference to ideal gases in science:

The ideal-gas law:

(1) P·V =nR·T.

(1) is false, as ideal gases do not exist. But formulating this law using Mill Arithmetic gives a true ideal-gas law:

(1a) (x)(Gx → P(x) · V (x) = nR · T(x)),where Gx means 'x is an ideal gas'.(Kitcher, p. 116).

As (1a) is a conditional statement with a false antecedent it is always vacuously true (by definition of vacuous truth), so it is true when universally quantified as well. For Kitcher truth is achieved in the same way when mathematical statements are formulated in Mill Arithmetic.

2) Argument Against Kitcher's Idealisation

I argue that Kitcher's idealisation theory does not achieve the mathematical truth required. I use an argument by Hoffman to attack premise 3.

Axiom 14 states that every operation has a successor operation:

(2) (x) (x)(∃y)Syx

With the "ideal agent" the axiom transforms into:

(2a) (w)[Iw → (x)(∃y)Syx], where Iw is the ideal agent(Hoffman,2004, p.5)

The structure is the same as in the (1) and (1a) pair. As (2) is the antecedent of (2a) and it is always false, (2a) is vacuously true. However, even with an ideal agent it is not convincing how infinite numbers of operations (each operation is followed by another one) can be carried out in any possible world. "There will be a point at which there is a false axiom, namely the one in which the antecedent is the 'largest' existing operation, which by hypothesis will not have a successor operation. All after this one will be vacuously true"(Resnik, 1997 p.65). The problem with Kitcher's formulation then is that the ideal agent can create vacuous truth out of any conditional with a false antecedent! Such kind of truth, however, does not succeed to explain why mathematical truths are convincing. The only way to secure this truth would be to prove the existence of an ideal agent (due to its definition), which is impossible given Kitcher's account. Therefore, Kitcher's idealisation theory fails to achieve the mathematical truth required.

3) Field's Nominalism

I now suggest that Kitcher's idealisation can be powerful if it is transformed into a nominalist theory. Whereas nominalism is not a purely empirical view on mathematics, it is still an anti-realistic one. "Nominalism about mathematics is the view according to which mathematical objects, relations, and structures do not exist at all". Now, the key norm of mathematics is that "it can be interpreted instrumentally and remain physicalistically acceptable" (Irvine, 1990, p.12). In nominalism mathematical statements are false. Nevertheless, nominalist perspectives could still satisfy the purpose of my essay if it could show why mathematical "truths" are convincing. I consider H. Field's nominalisation programme; it requires the reformulation of mathematical theories to avoid the commitment to mathematical objects. This is where I see a parallel to Kitcher's "idealization theory" and his formulation of mathematical statements using Mill's Arithmetic. Thus, if Kitcher's ideal agent is fictional and does not aim to prove the mathematical statements to be true it can obtain all advantages of the nominalist approach. I now give the argument for a nominalist antirealist account of why mathematical truths are convincing:

P1 Mathematical theories can be nominalised

P2 Nominalist perspective on mathematical theories explains why their "truths" are convincing

C Nominalisation can give an anti-realist account of why mathematical truths are so convincing.

I use the "make-believe" theory by K. Walton to show that the second premise is true. According to Walton, true in mathematics does not have to entail true in the physical world to be convincing (or to have an effect). For example, horror stories do not entail real danger but create fear in the listener. Also, consider bet-worthiness of many fictional statements. A statement ''Luke Skywalker is a Jedi" is fictional but bet-worthy. Such statement convinces us, as it is true in the story of "Star Wars" although false in physical reality. This is due to the make-believe theory which accounts for bet-sensitivity of some propositions in terms of the games of make-believe (Hoffman, 2004, p.10). This account of fiction gives us a model of truth in mathematics as truth in fiction and explains why mathematical statements can convince us even though such statements are false. Therefore, a nominalist view of mathematics can explain why we are convinced by mathematical truths.

4) Argument Against Field's Nominalisation

I now argue that premise 1 of the argument for a nominalist account of mathematical truth is false, as mathematical statements fail to be reducible to particulars.I show that taking premise 1 to be true leads to absurdity using this example:

A. Six is a number

It can be possible to nominalise statement A equating it with

B. For any collection of particulars, X, if the members of X are six, then the members of X are numbers

But then it follows that:

C. For any collection of particulars, X, if the members of X are six, then the members of X are distinct, which is equivalent to

D. Six is distinctness.

We get an absurd outcome that a nominalised true statement becomes equivalent to a false statement. Therefore, A cannot be nominalised with B. Although this is just a single example, it outlines the traditional problems with nominalism: when mathematical statements are nominalised they allow a false analogy fallacy which leads to false outcomes. From this follows that mathematical statements cannot be nominalised (or at least easily nominalised). Correct reformulation without commitment to mathematical objects cannot be made using Field's nominalisation. Therefore, an anti-realist nominalist account of why mathematical truths are so convincing cannot be given.

5) Conclusion

I argued that a purely empirical mathematical anti-realist account of why mathematical truths are so convincing cannot be given on the basis of either Kitcher's idealisation theory or Field's nominalisation. I used Hoffman's example of Kitcher's idealisation to show that his theory is unable to achieve convincing truth. I suggested that Kitcher's idealisation could be transformed into a nominalist theory if objection to proving mathematical statements to be true is put aside. I shown that Fields nominalisation if successful could provide the explanation for why mathematical statements are convincing and can be called true in mathematics whilst they can be false in physical reality. I then argued that Field's nominalisation leads to absurdity and cannot avoid problems of traditional nominalism. Hence a nominalist anti-realist account of why mathematical truths are so convincing is impossible. I concluded that one cannot give a purely empirical anti-realist account of why mathematical truths are so convincing when building this account on either Kitcher's idealisation theory or Field's nominalisation programme.

Bibliography

  1. Kitcher, Philip. The Nature of Mathematical Knowledge (1984): 3-11,101-117.
  2. Hoffman, Sarah. "Kitcher, Ideal Agents, and Fictionalism" Philosophia Mathematica Vol. 12,No.1 (2004): 3-17.
  3. Irvine, A. "Physicalism in Mathematics"(1990): 10-17.
  4. Lavers, Gregory. "Benacerraf's Dilemma and Informal Mathematics" The Review of Symbolic Logic Vol. 2, No. 4, 2009: 769-772.
  5. Colyvan, Mark and Zlata, Edvard "Mathematics: Truth and Fiction? Review of Mark Balaguer's Platonism and Anti-Platonism in Mathematics" Philosophia Mathematica, Vol.7, No.3 (1999): 336-349.
  6. Benacerraf, Paul. "Mathematical Truth" The Journal of Philosophy, Vol. 70,No.19 (1973): 661-679.
  7. Bostock, David. "Philosophy of Mathematics" (2009): 190-191.
  8. Parsons, Charles. "The Nature of Mathematical Knowledge by Philip Kitcher. Review" The Philosophical Review, Vol. 95, No. 1(1986): 129-137.
  9. George, Alexander and Velleman, David J. "Introduction to Philosophies of Mathematics"(2002): 1-13.
  10. Shapiro, Stewart. "Thinking about Mathematics": 22-33.