Motto:

"There are none so blind as those who will not see." --

Google+ Badge

Friday, January 18, 2008

Dialetheism and Hume's Principle

Suppose the dialetheic treatment of the logical and / or semantic paradoxes is correct; in particular, that sentences of the liar family express propositions which are both true and false. Then consider the following list, which I will call List A:

(1) All whales are mammals.

(2) Nothing can escape a black hole.

(3) (3) is not true.

How many statements on List A are true? Well, both (1) and (2) are true, and under our assumption of dialetheism so is (3). Since these statements are all distinct from each other, it follows that there are three true statements on the list. But (3) is also false, so what it says is not true; hence it is also true that there are only two true statements on the list. So the number of true statements on List A is both two and three. By a parity of reasoning, we can conclude that the number of false statements on the list is both zero and one.

What if we bring Hume’s Principle to bear on this case? According to Hume’s Principle, the number of F’s equals the number of G’s if and only if there is a one-to-one correspondence between them. Suppose, now, that Jones has three books on his coffee table, which we will call Book A, Book B and Book C. Is the number of books on Jones’ coffee table the same as the number of true statements on List A? Invoking Hume’s principle tells us that the number of books is the same as the number of true statements if and only if each book can be paired off with exactly one true statement and vice versa. Can this be done? On the assumption of dialetheism, the answer is “Yes and no”: Book A can be paired off with (1) and Book B with (2), but can Book C be paired with (3)? Insofar as (3) is merely a statement Book C can indeed be paired with it. But in being paired with (3), is Book C paired with a true statement? Since (3) is both true and false, it follows that Book C both is and is not paired with a true statement. Because of this, while it is true that Book A, Book B and Book C are each paired with a different true statement in List A, it is also false that they are thus paired. Hence, if Hume’s Principle holds, we get the result that the number of books on Jones’ table both does and does not equal the number of true statements on List A. What follows from this? If the number of books on the table is three, and the number of books on the table both does and does not equal the number of true statements on List A, shouldn’t it follow that three does not equal three? After all, if there are three books on the table and the number of true statements on List A equals this, there must be three true statements on List A. If the number of true statements on List A is also not equal to three, how can it fail to hold that three, which does in fact number the true statements, is not equal to three? Now, as equality does not depend upon context, if three is not equal to three in this case, it is not equal to three in any case, and I take it that this would be a Very Bad Thing. However, a dialetheist need not embrace this result. Even though, on Hume’s Principle, the number of books on the table is both equal and not equal to three, it does not follow that “the number of books on the table” picks out some one thing that is not equal to itself. Instead, the dialetheist should hold that the phrases “the number of books on the table” and “the number of true statements on List A” are really descriptions which falsely presuppose uniqueness because they contain the definite article ‘the’. In truth, what’s going on here is that there is more than one number which exhaustively numbers the true statements on List A, and this in no way entails that there is some one number which is unequal to itself. All that follows is that the numbers involved are not equal to each other. Yet it remains true that there both is and is not a one-to-one mapping from the books on Jones’ table to the true statements on List A. From this, I think dialetheists should conclude that the number of F’s can be different from the number of G’s even if there is a one-to-one mapping between them, and can be the same even if there is not. They should hold that Hume’s Principle, in dialetheic contexts, is a biconditional which fails in both directions, and as such cannot be used to provide a criterion of identity for numbers.

9 comments:

Carlos said...

Wow, very interesting!

Quirinius_Quine said...

Hi Carlos,

Thanks. Are you familiar with dialetheism already? If not, I explain what it is in this post:

http://philosophicalpontifications.blogspot.com/2007/09/limitation-on-dialetheism.html

This (relatively short) article in the online Stanford Encyclopedia of Philosophy also explains, and gives some arguments for it:

http://plato.stanford.edu/entries/dialetheism/

Quirinius_Quine said...

Oops, the first link got cut off. Here's the first half:

http://philosophicalpontifications.blogspot.com/

And here's the second half:

2007/09/limitation-on-dialetheism.html

Aaron Cotnoir said...

Your post in grounded in a misunderstanding of the dialetheist position.

A dialetheist will say there are exactly three true statements on the list and only one false one.

Simply because the dialetheist endorses that (3) is false he can still 'count' (3) among the truths. Especially since the dialetheist contends that sentences like (3) are both true and false.

There is a question in the vicinity about the adequate treatment of quantifiers like "there are 3 ...", etc. much of this is made clear in Priest's _In Contradiction_ (which you should read).

Quirinius_Quine said...

Hi Aaron,

Thanks for the recommendation; I've had that on my list of books to get for a while, though I haven't gotten around to buying it yet.

You say, "Simply because the dialetheist endorses that (3) is false he can still 'count' (3) among the truths. Especially since the dialetheist contends that sentences like (3) are both true and false." Maybe I'm just confused (and I hope Priest's book will clear this up), but it seems to me that if (3) is both true and false, then (3) both can and can't be counted among the truths. Based on what I've read so far, Priest seems to hold that if p is a dialetheia, then "p is a dialetheia" is also a dialetheia (that is, if p is both true and false, then "p is both true and false" is also both true and false). And if that's so, wouldn't "(3) can be counted among the truths" also be both true and false?

Aaron Cotnoir said...

Well, I won't rehearse all the arguments here, since I can't off the top of my head anyway. But you should look in particular at the sections about "minimally inconsistent LP". Priest argues that we shouldn't `multiply contradictions beyond necessity. In keeping with that principle, he denies that T<~A> implies ~T<A>. If it did, every contradiction of the form T<A> & T<~A> would also imply another contradiction, namely, T<A> & ~T<A>. All this is to say that a sentence may be false (and true) but that needn't force it to be *untrue*.

As to your point about whether "(3) can be counted among the truths" is a glut if (3) is a glut. I don't think that's right. First, the modal `can' will mess things up for your argument. Secondly, there are likely lots of ways to treat the quantifiers to get things right. Again, refer to IC for the official position.

BTW, you should check out some of the stuff on my blog: http://cotnoir.wordpress.com

Quirinius_Quine said...

Aaron,

I suppose you're right; if Priest's claim that T(~A) doesn't imply ~T(A) holds up, then the statement that some proposition is a glut needn't be a glut itself. However, this seems to run counter to what he says on pp. 4-5 of "Beyond the Limits of Thought" (Hardback, 2nd edition; and changing the symbols to ones I can type):

"Similarly, a & b is true iff both conjuncts are true, and false if at least one conjunct is false. In particular, if a is both true and false, so is ~a, and so is a & ~a. Hence, a contradiction can be true (if false as well)."

I think this is where I got the wrong impression of Priest's position. Anyway, thanks for helping to clear things up; I'll check out "In Contradiction" and your blog when I have the time.

Aaron Cotnoir said...

You're right about the semantics for conjunction and negation. But his truth predicate behaves in such a way that it invalidates the inference from T>~A< to ~T>A<. I'll get a page reference when I can locate my copy of IC.

(I should note, not all dialetheists agree with this -- in particular, JC Beall's dialetheic theory has a truth predicate for which A and T>A< are intersubstitutable in all non-opaque contexts.)

Aaron Cotnoir said...

Here's that reference:

"There seems to be no reason why, in general, if A is a dialetheia, T>A< is too. If A is a dialetheia, T>A< is certainly true, but it might be simply true and not also false."

In Contradiction, 2nd ed. p. 79

Arguments to this end can be found in section 4.9, pp. 69--72