Hostname: page-component-7b7567c8c-tnzcf Total loading time: 0 Render date: 2026-11-14T00:58:40.970Z Has data issue: false hasContentIssue false
  • English
  • Français

Defective Contexts, Accommodation, and Normalization

Published online by Cambridge University Press:  01 January 2020

Mark Richard
Mark Richard*
Affiliation:
Tufts University, Medford, MA 02155, USA
Get accessRights & Permissions [Opens in a new window]

Extract

Propositional Attitudes defends an account of ‘believes’ on which the verb is contextually sensitive. x believes that S says (quite roughly) that x has a belief which is ‘well rendered’ or acceptably translated by S; since contextually variable information about what makes for a good translation helps determine the extension of ‘believes,’ the verb is contextually sensitive. Sider and Soames criticize this account. They say it has unacceptable consequences in cases in which we make multiple ascriptions of belief to a single individual - as happens, for example, when we say that Odile believes such and such, that the woman in the corner believes so and so, but are ignorant of the identity of Odile and the woman in the corner.

I will distinguish two objections along these lines, and argue that neither is forceful. The objections differ as to whether or not the speaker mistakenly presupposes that the believers under discussion are distinct.

Information

Type
Research Article
Information
Canadian Journal of Philosophy , Volume 25 , Issue 4 , December 1995, pp. 551 - 570
Copyright
Copyright © The Authors 1995

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

1 Richard, Mark Propositional Attitudes: An Essay on Thoughts and How We Ascribe Them (Cambridge and New York: Cambridge University Press 1990).CrossRefGoogle Scholar Henceforth, I use 'PA’ to refer to that book.

An ancestor of the first third of the present paper was read at the APA symposium on PA in LosAngles in April, 1994; Scott Soames and Mark Crimmins were the other symposiasts. I thank Soames and Crimmins for their comments. I'm indebted to Ed Gettier and Terry Parsons for discussion of a draft of the current paper.

2 PA's account thus presupposes a weak version of the view that our beliefs are realized by psychological states which ‘are sententially structured.’ The book's first chapter argues that the psychological sententialism presupposed is quite benign.

3 Sider, Ted ‘Three Problems for Richard's Theory of Belief Ascription,’ Canadian Journal of Philosophy 25 (1995) 487-513CrossRefGoogle Scholar; Soames, Scott ‘Beyond Singular Propositions?’ Canadian Journal of Philosophy 25 (1995) 515–549CrossRefGoogle Scholar

4 Factors besides the current intentions of speaker and audience may provide the class in question: Prior conversation is an example.

5 Many examples make this point. Some predicates- ‘heap,’ ‘dirty,’ ‘round’ for example - shift extensions across contexts depending, in part, on what count as good or acceptable examples of satisfiers of the predicate. The contextual provision of exemplars for such predicates may be defective. Misidentifications, as in the Kerrigan case, may cause this. Or conversants may be committed to using a single 'paradigm’ for a predicate (or a set of paradigms that closely resemble each other relevant respects), but, unknown to one another, have quite different paradigms in mind. (My paradigm for ‘cooked’ may resemble your paradigm for ‘raw’ more than it does your paradigm for’ cooked.’) Surely this does not speak against the view that such expressions have contextually varying extensions which are determined in good part by the intentions, interests, and focus of conversants.

6 In fact, I will ignore the modal dimensions of Soames's objection, to keep this response to a manageable length. How to generalize what I say to give modal truth conditions is in any event fairly obvious.

7 If we identify sentences as uttered by a person with the articulated propositions (the RAMs, in the jargon I used in PA), such utterances would identify, we may then say that what is presupposed is a set of (articulated) propositions, thus achieving a unification of the objects of belief with the objects of presupposition.

8 ‘Supervaluating’ in such cases (that is, adapting the procedure suggested above for dealing with defective contexts) is not sufficient to guarantee that both sentences in (7) are true. For Pynchon could fail to accept ‘Twain is Clemens.’ If so, he might accept ‘Twain wrote Huck Finn’ but not ‘Clemens wrote Huck Finn.’ In such a case, if the context's restrictions include (8), ‘Pynchon believes that Twain wrote Huck Finn’ comes out false on a ‘supervaluational’ approach.

9 I do not, in fact, think that this is the only factor determining what restrictions are in play. What is presupposed in the context also may contribute to determining restrictions. I ignore this for simplicity's sake.

10 Examples which are naturally interpreted in terms of agglomeration are easy to construct. For example

When I read your book, I thought you were at least as talented as Jones. And the first time I heard you speak, I thought that you were more talented than he. I also thought, when I read your book, that you'd be a boring speaker. But after having heard you, I know that you aren't.

11 To a response akin to (but somewhat different from) that I give to the Pynchon objection, Soames objects that it ‘create[s] … a Russell-like problem for every apparently non-Russellian example’ which motivates the account of PA. If we merge restrictions, he suggests, then in a situation in which (i) I utter

(E) Thomas believes that Twain is dead

intending it to be interpreted in a Russellian fashion; (ii) utter

(F) Pynchon believes that Twain is happy

intending ‘Twain’ to translate Pynchon's use of ‘Twain,’ and (iii) the subjects of (E) and (F) name the same person, merging requires us to say that both restrictions must be given a Russellian interpretation.

Soames apparently draws this conclusion because he assumes that if (E) has Russellian truth conditions, this is because it is accompanied by a restriction

For Thomas, ‘Twain’ may translate absolutely any name of Twain. But this is not how the system of PA works: In the sort of case of Soames apparently has in mind, (E) should be understood as having no restrictions on translation accompanying it.

Furthermore, the force of the intuitions Soames is attempting to marshal with this comment are, I suspect, mitigated by the fact that sentences such as (E) and (F) will, in the sort of case Soames has in mind, occur in both a local and a global context. (See the discussion below for an explanation.) Space limitations prevent me from pursuing this last point.

12 ‘Scorekeeping in a Language Game’ in Lewis, David Collected Papers, Volume 1 (Oxford: Oxford University Press 1983), 245Google Scholar

13 See Lewis, 240, for a more precise formulation.

14 We should expect accommodation to apply only cateris paribus. In particular, if a context is defective, containing a false presupposition about the identity of those under discussion, it will not apply.

15 Part of the argument for the existence of semantic normalization above required that A's utterance u2 of ‘I shouldn't have said that France wasn't hexagonal’ at the end of the conversation, says (inter alia) that A said p, where p is the proposition he expressed at the beginning of the conversation with the utterance u1 of ‘France isn't hexagonal.’ What I intend in the argument is that u2, taken relative to the global context of the conversation, is a correct report of what Ul says, taken relative to the global context.

16 As, for instance, in the case of the argument ‘Mary has a brother, so she has a sibling.’

17 I assume a definition of validity along the lines of Kaplan's- on which validity is a matter of (at least) transmission of truth from premisses to conclusion within any context- is correct.

18 Actually, Dc will typically contain elements determined by intentions and beliefs of the audience, by earlier remarks in the conversation, and so on. I ignore this.

19 I assume that Dc and DPc are always finite.

20 more precisely: which is, in the jargon of PA, in its representational system,