Monday, January 23, 2006

A question from Joe Markus on Michael Martin

Hello Professor Reppert,
I recently came across some posts involving the notions of necessity and possibility on William Vallicella's blog and it reminded me of a debate I had years ago with someone on the SecularWeb. I'm hoping that you can offer your impression of the topic.
As I'm sure you're aware, in Chapter 2 of his Atheism:a philosophical justification, Michael Martin argues that theological assertions are neither true nor false. He offers a modest defense of a variation of a verificationist style criterion of factual meaningfulness.
It has been claimed by some critics that Martin's overall position in Atheism is inconsistent. They claim that despite arguing that theological assertions are neither true nor false he does however proceed to offer arguments for the falsity of theism. Specifically, he argues that there is no good reasons to believe theism is true and that there are quite a few good reasons to believe theism is false. The critics seem to be saying that he cannot argue BOTH that theism is neither true nor false and then make arguments regarding its truth value.
Martin has defended himself here:
Martin claims that his critics are confused about the structure of his argument. He thinks his critics attribute to him the follow argument structure:
  (a) p is neither true nor false.
  (b) p is true.
However, he goes on to state that the actual structure of his argument is:
  (a) p is neither true nor false.
  (b) if p is either true or false, then p is true.
He goes on to summarize his argument ---"God exists" is probably factually meaningless. But if it's meaningful then it's false.
In my discussion on the SecularWeb, I argued that if Martin believes that theism is likely neither true nor false then he must believe that it is POSSIBLY true or false. My reasoning was that if you believe p is likely then you must believe its denial is possible. (It may be possible to deny that.) And if you believe something is possibly true or false then you must believe that it is possible that p is true.
If p is "God does not exist." and p is possibly true or false then "God does not exist" is possibly true. In other words, Martin must presuppose that theism is factually meaningful in order to even formulate his "fall back position" argument.
Anyway, I was hoping that you could make some comments or offer some suggestions about the issue. Is my reasoning completely bogus?
I appreciate your time. Any comments would be greatly appreciated.
  Joe Markus

I'd want to make sure we have a clear distinction between epistemic and metaphysical possibility. Martin thinks that theism is probably meaningless but holds out the epistemic possibility that it is meaningful (in which case he says it is certainly false). He might be saying something like this: According to every theory of meaning that I know and think likely to be true, theism comes out meaningless. But just in case those theories turn out not to be true, then I suppose it might be meaningful, but if it is it is almost surely false. So I don't know if he can be tagged with an incoherence here, on the face of things. However, I think the charge of meaninglessness against theism has been refuted time and again, and people should have realized the jig was up when ardent atheist J. L. Mackie, in The Mirace of Theism, maintained that Swinburne had effectively refuted the charge of meaninglessness.


Hume's Ghost said...

An example that comes to mind for me is the issue of out of body experiences where a person describes seeing and hearing things out of body.

A philosopher might say that "seeing" without use of the body is an incoherent description and thus a meaningless statement, while a scientist would more probably argue its an empiracally false statement.

Jason Pratt said...

Actually, Joe's correct (in the quote the metaphysician referenced). A belief that p is _likely_, is a belief about _probability_. Probability assumes possibility, but probability isn't certainty.

Of course, if one substitutes for p a proposal believed to be necessarily true, then it isn't necessary (to say the least!) to believe that p may be impossible. But then, you've crucially altered the terms of Joe's claim, not provided a counterexample to it. A proposal believed to be necessarily true, is _NOT_ a proposal believed to be merely probably true; and Joe was talking about probability.

(e.g. if one substitutes for p the belief "this die must necessarily roll up a six, because there are six pips on every face", one is no longer talking about a situation similar to the belief "this typical six-sided die is likely to land more often on faces 1-4 on any series of rolls, than on faces 5-6.")

Put more simply: a belief that p is probably true, _does_ tacitly admit the possibility that p is not true after all.

While I generally agree with Victor's analysis (in principle), I think Joe has a point worth considering, too: Martin can only be rendering an opinion about the falsity of a claim p if he accepts some kind of meaning for p. He may decide the meaning involves mutually exclusive elements, but he could only identify them as such by recognizing a meaning to the elements.

In other words, if he goes on to do anything other than offer a completely ungrounded assertion that the proposal of God's existence is certainly (or even probably) false, then his argument(s) require accepting one or more meanings to the proposal of God's existence. Once he does that, he's tacitly answered his original claim (the probable meaninglessness of the proposal 'God exists') in the negative.

Put another way, he could coherently say something to the effect of (building from Victor's charitable paraphrase): 'According to every theory of meaning that I know and think likely to be true, theism comes out meaningless. But just in case I'm wrong and there is some meaning to the claim of theism after all, then I sheerly assert it to be certainly or probably false anyway.' This would look oddly prejudicial (to say the least), but it would be coherent (as far as it goes).

What Martin cannot coherently be doing, is maintain a provisional dismissal (on whatever grounds) of the meaningfulness of theism, and then proceed to a logical operation requiring some kind of acceptance to a meaning of theism as a claim. If he's doing that (as Joe reports), then yes, he's out of bounds. He can't have it both ways: he can't legitimately insist on a disbelief based on (probable or certain) meaninglessness, and _also_ insist on a disbelief that requires accepting one or more meanings.

I suspect what Martin is actually doing, in the first instance, is judging against claims of theism based on incommensurate meanings of elements of the claims. Incommensurate meanings don't render a claim meaningless, though (despite rhetorical or metaphorical convenience in calling it that); just incoherent.

6=16 is not a meaningless claim, for instance. ajsf opwer gashoui si is a meaningless claim; I know that, because it consists of letters I know I randomly typed. Someone who didn't know that fact might have a lot more difficulty making a legitimate judgment of its meaninglessness, though. (Maybe I'm writing in code?) Theistic claims, obviously, aren't the same as marks scribbled randomly at their elemental level on a page.


Jason Pratt said...

Fair enough; though when Joe uses the word 'likely' (twice), it seems probable (to me anyway {g}) that he's not really thinking in terms of a probability that technically includes 100% probability.

In any case, there would be no escape for Martin along this line. Whether or not he includes 100% probability in the notion '"God exists" is probably factually meaningless', the moment he makes use of a meaning (composite or otherwise) to the phrase--which he will have to do in order to do anything other than sheerly assert its falsity even if it happened not to be meaningless--then he has denied the first position of meaninglessness (whether merely probably or including certainty.)

Or, as Joe put it, "Martin must presuppose that theism is factually meaningful in order to even formulate his 'fall back position' argument."

Surely this is reasonably clear in Martin's own restructuring of {b}: he explicitly proceeds by hypothesizing p to be meaningful instead of (certainly or probably) not meaningful, in order to escape the charge that he's trying to make a mutually exclusive claim (p _is_ neither true nor false, _and_ p _is_ true--or false, rather, since Martin is certainly not arguing "God exists" to be true.)

Martin's move (as I admitted) isn't _strictly_ illegitimate, so long as he never attempts to make a practical use of it. Any attempt to make a practical use of it, though, requires importing some real meaning to the phrase "God exists", which either zorches a claim that the phrase certainly has no meaning, or else answers the claim of (non-certain) probability in favor of the improbable option.