April 22, 2005

Request for Advice: A Gricean Epicycle on Assertion

I'm making some changes to "Must We Know What We Say?" and I have a question. The question has two parts: What's going on here? and Should I put something about it in the paper?

This has to do with my account of the following case (from Williamson's "Knowing and Asserting," names supplied by me): Alice has a lottery ticket. The drawing has been held but the winner has not been publicly announced. Sarah, having no inside information, tells Alice

(1) Your ticket did not win

which turns out to be true. Nevertheless, when Alice finds out that Sarah's assertion of (1) was based entirely on the low probability of the ticket's winning, she's entitled to be annoyed. Why is that? Williamson says, "Because Sarah didn't know that (1) was true, and knowledge is the norm of assertion," but I don't think knowledge is the norm of assertion, so I need another account.

In the paper, after establishing that Alice is entitled to assume that Sarah has some warrant for asserting (1), I argue thus (last full paragraph on p. 15):

Why should Alice infer that Sarah's warrant is not merely probabilistic? The answer lies in Grice's maxim of Manner. If Sarah wants to make clear that her warrant is merely probabilistic, she has the option of saying
(6) Your ticket is almost certain not to have won.
Since Sarah did not say (6), Alice is entitled to assume that her warrant is not merely probabilistic. Since (1) is ambiguous between two kinds of warrant that Sarah may have, and (6) is appropriate if and only if Sarah has a probabilistic warrant (by the first maxim of Quantity), the assertion of (1) implicates that Sarah has a non-probabilistic warrant. [underlined words added due to referee's comment]

In response to a referee's comment, I've just added this footnote to the paragraph:

An anonymous referee suggests that a parallel argument would entitle Alice to conclude that Sarah has a probabilistic warrant. Sarah did not assert (6*):
(6*) Your ticket is absolutely certain not to have won
which entails that the speaker has non-probabilistic warrant. So, the suggestion goes, if asserting (1) instead of (6) implicates that Sarah has non-probabilistic warrant for her assertion, asserting (1) rather than (6*) implicates that Sarah lacks non-probabilistic warrant for her assertion. I do not think that the argument is truly parallel. Though (6*) entails that Sarah has non-probabilistic warrant, (6*) cannot be felicitously asserted in every case in which she has non-probabilistic warrant. If Sarah has merely heard from a usually reliable source that Alice's ticket did not win, she cannot assert (6*), because she is not absolutely certain that the ticket did not win. So in asserting (1) rather than (6*), Sarah does not implicate that she has probabilistic warrant rather than a non-probabilistic warrant that falls short of absolute certainty. (I thank the referee for making me see that "if and only if" rather than "only if" was required in the text.)

But now I'm thinking about this problem: Why, if Sarah had a non-probabilistic warrant, couldn't she say (6')?

(6') I know your ticket didn't win.

If (6') were assertable whenever Sarah had a non-probabilistic warrant, then the referee's parallel argument would go through.

In one way it seems obvious to me that (6') is inappropriate even if Sarah has a non-probabilistic warrant. It's just an odd thing to say. Somehow it draws attention to Sarah rather than to the ticket, and so (by the maxim of Manner?) it suggests something more than merely that Sarah has a non-probabilistic warrant for believing that Alice's ticket didn't win. So the parallel argument doesn't go through--the fact that Sarah asserts (1) rather than (6') doesn't implicate that she lacks non-probabilistic warrant, and the fact that she asserts (1) rather than (6) still does implicate that she has probabilistic warrant.

But you'll notice that I don't exactly have a good account of why (6') is inappropriate even if Sarah has non-probabilistic warrant.

Another issue is this: It's not absolutely obvious that (6') is assertable whenever Sarah has a non-probabilistic warrant. It seems as though most of the non-probabilistic warrants she might plausibly have in this situation would be enough to give her knowledge, and I guess that's enough in this case. But it would take a little development to show that (6') is threatening in a way that (6*) isn't.

So: Should I extend the footnote to mention my worry about (6')? In posting about it here, have I fulfilled my duties of intellectual honesty? And do you have a theory about why (6') really would be inappropriate? I'm leaning toward "No," "Yes," and "Not as much as I'd like," respectively.

Posted by Matt Weiner at April 22, 2005 10:24 AM

Well, to give an example which I think is related: in your paper, you spell the name of Sherlock Holmes' arch-enemy "Moriarity". A preponderance of Googled websites, as well as one Sherlock Holmes fan website, give the name as "Moriarty". So I have a probabilistic and non-probabilistic warrant for asserting:

1) Matt, that's spelled "Moriarty".

But in contrast to the example, I find I can say neither

2) That's almost certainly spelled "Moriarty"


3) That's absolutely certainly spelled "Moriarty".

Now, can I say

4) I know that's spelled "Moriarty"

without violating the conversational norms of Standard American English?

The idiomatic objection to this is that I don't "know know" it: it functions as a decomposed complex of information for me to pass on to you, and if I were to be confronted with the same predicament as you the best I could mentally essay would be "I pretty definitely think that's spelled 'Moriarty'". (This, in keeping with your paper, suggests to me that knowledge cannot be the norm of assertion because it is too *idiosyncratic*: we use knowledge for our own purposes, whereas assertion is for the benefit of others.)

The non-idiomatic support for this is that I appear to have a suitably justified true belief, and it seems to me that 4) is simply appropriate given the speaker's meeting appropriate conditions of knowledge: even if we wouldn't say 4) for all non-probabilistic knowledge that's no argument that it isn't. (Consider a miniature theory modeled on Davidson: "It's spelled 'Moriarty'. Jeff knew that.") So it seems that if we turn our attention away from the myriad ways in which a sentence can be conversationally inappropriate, it seems that it simply would be an adequate thing to say given a non-probabilistically justified belief and inappropriate for the reason given in the paper, that the language-game being played must equivocate.

Posted by: Jeff Rubard at April 23, 2005 08:12 PM

Aw crap. I'm totally busted. Thanks for pointing that out, Jeff. I'll correct that in the version I send off (and you'll get an acknowledgment for it).

My intuition is that you can say (1), (2), and (4); I'd even think you'd be OK to say (3), though I guess I'd look it up in the stories themselves before saying (3) myself. But I'm kind of loose on my fact-checking here....

Posted by: Matt Weiner at April 24, 2005 06:58 PM