A defense of temporalism

Philosophical Studies

Volumn 81, Issue 1, 1996, Pages 71-95

  Aronszajn, Mark ^a  

^a WEST VIRGINIA UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 34248777257     PISSN: 00318116     EISSN: 15730883     Source Type: Journal    
DOI: 10.1007/BF00354479     Document Type: Article

References (12)

1. Various accounts of the concept of proposition have been offered in the recent literature. Some of these presuppose one side or the other of the very issue that will be in dispute in this paper; I do not wish to employ any such characterization. As a rough guideline, I can say that when I speak of propositions here and in what follows, I shall mean things of a sort that fit the following descriptions: i) things expressed or asserted with literal utterances of declarative sentences ("what's said", to use David Kaplan's phrase), ii) objects of belief (and other prepositional attitudes), iii) bearers of truth-value. Though these formulations are informal and loose, I hope that they suffice to give the reader a workable idea of what I shall mean by 'proposition'. I believe that most participants in the dispute between Temporalists and Eternalists would accept this rough characterization, so far as it goes.
2. M. Richard, "Temporalism and Eternalism", Philosophical Studies 39 (1981); pp. 1-13.
3. A notable example of the influence is in Frege's Puzzle Nathon Salmon, 1986, The MIT Press; see especially, p. 24ff. There, Salmon appeals to Richard's argument in order to support his (Salmon's) view that the information content of such a sentence as (1), at any context of utterance, should be understood to be "eternal" and should be understood to be so because that information content concerns the time of the context in question.
4. In place of (ii), Richard has (ii) Mary still believes everything she once believed. But it is clear that Richard intends (ii) to be interpreted in the way I propose to interpret (ii); moreover, (ii) avoids an ambiguity attaching to (ii) that introduces an irrelevant worry concerning Richard's case against Temporalism. (ii) has a reading on which it says that there is some point in the past such that Mary still believes everything she believed then. On this reading of Richard's sentence, Temporalists will not be committed to the validity of the inference expressed in A. But since this is not a reading Richard intended, it seems just as well to use (ii), a sentence that has the reading Richard intends and eliminates this unintended interpretation available for (ii).
5. One might see problems with (TB) stemming from substitutivity failures of coreferential terms in the imbedded sentences of belief ascriptions. Some would wish to hold that (a) Mark Twain was a famous author and (b) Sam Clemens was a famous author express the same proposition, yet also hold that from (c) Jones believes that Mark Twain was a famous author we may not infer (d) Jones believes that Sam Clemens was a famous author. If one has such a view, then one had better give up (TB). For according to (TB), if (a) and (b) do express the same proposition, then the inference from (c) to (d) is valid. Even if there are problems along these lines for (TB), however, they are not relevant, as far as I can see, to Richard's case against Temporalism.
6. I am indebted to Larry Hohm, Dan Kervick and especially Ned Markosian for persuading me that it is worth considering whether the alternative interpretations I discuss here influence our intuitions as to the validity of the inference we take to be expressed in A.
7. Richard, op. cit., p. 6.
8. Eternalists might naturally rejoin at this point that indeterminateness is not something they intended to count as a truth-value. Or they might respond by amending the definition of "eternal". They might suggest that the term should be defined so as to concern specifically the two truth-values, true and false: (Dl') p is eternal =df necessarily, there are no times, t and t', such that p is true at t, and p is false at t′ If this definition is accepted, then a proposition may change from indeterminate to true, or indeterminate to false, and yet still count as eternal. There is, however, a view about future-contingents that temporalists might accept that would still lead them to deny that the proposition expressed by (3) is eternal. On this view, indeterminacy of future-contingents is eschewed. Instead it is supposed that future-contingents that will be true, but aren't yet, are currently false. Presumably such theorists would count the proposition expressed by (3) as an example of such a futurecontingent. So they would deny that that proposition is eternal, even with 'eternal' defined according to (Dl'). My main point, as I remark in the text, is that Temporalism is actually not the real target of the argument Richard has offered; rather it is the idea that sentences like (1) are invariant.
9. The example and analogy are due to Dan Kervick, though I am not sure that his assessment of the two cases, A and A, is the same as my own.
10. Richard considers (op. cit., pp. 7-9) a response that a Temporalist might have to his argument, a response that in some respects is akin to the one being considered here. The second premise in A may be cast as a claim involving the notion of belief retention, to the effect: (ii') For every proposition, p, if there's a time earlier than now when Mary believed p, then Mary presently retains a belief in p. Richard considers some proposals that a Temporalist might make as to what it is for a belief to be retained. With each of these proposals, Richard acknowledges that A(i) together with (ii') - according to the proposal at hand - do not commit the Temporalist to the validity of A. On one of these proposals, (ii' ) expresses a claim equivalent to that expressed by A(ii) with the quantifier restricted to eternal propositions. Richard then proceeds to argue that none of these proposals can be accepted as a general account of belief retention. I find that Richard's arguments are persuasive, yet I wish to stress that they do not effect the response I am now considering which takes A(ii) to admit a contextually determined restriction on its universal quantifier. No general account of any notion of belief retention is presupposed in this response. It is widely acknowledged that quantifiers in natural language do admit contextually determined restrictions. The response I am considering here will fail if the following situation obtains: no contextually determined restriction on the quantifier in A(ii) that results in A's expressing an invalid inference (according to Standard Temporalists) will apply to A(ii) in contexts where we intuitively find that an invalid argument is getting expressed. Nothing in Richard's criticisms of accounts of belief retention suggests that this is the case.
11. Frege's Puzzle, p. 27.
12. I wish to thank John Culler, Ted Drange, Greg Fitch, Dan Kervick, Ned Markosian, Sharon Ryan, Ted Sider and especially Larry Hohm, for helpful discussion of earlier drafts of this paper. I also received useful criticism from the anonymous referee for this journal. The paper was presented at a Reunion Conference at the University of Massachusetts/Amherst, and I am grateful to various participants for their remarks, especially to Carl Woolf, who was my commentator, and to Philip Bricker, Barbara Partee and (again) Ted Sider.