Evans2.letter.dvi

Let us call “tense logic” the programme of explaining tense in natural lan- guages by means of a model theory similar in structure to possible worlds seman-tics for modality. This programme would make the following claims.
1) Tense is best represented formally by intensional sentential operators which should be read as “it was the case that” and “it will be the case that”.
2) These operators should be treated similarly to modal logic’s “diamond”, or “it is possible that”, operator. In particular, they should be analysed quantifica-tionally — something like this: “it was the case that there are dinosaurs” is trueiff there is a past time t such that “there are dinosaurs” is true-at-t. Unmodifiedobject-language sentences also get this kind of treatment. They are treated as im-plicitly present tense: “there are dinosaurs” is true iff there is a present time t suchthat “there are dinosaurs” is true-at-t.
The analogy between time and modality that tense logic exploits is a contro- versial one. Gareth Evans (1985) has objected that it is weakest just where thetense logicians needs it to be strongest: in the analysis of truth in terms of truth-at-a-time. What I am going to do is offer the tense logician a knock-down rebuttalto any Evans-style argument. It will be up to the tense logicians to decide whetherthey want my answer.
Evans’s argument centers around the crucial concept (for the tense logician) of truth-at-a-time, and is an argument by alternatives: Evans considers three read-ings of “true-at-t”, T1, T2, and T3, which he considers to exhaust the field, and eachof which he believes to be problematic. In each case, the problematicity comesdown to the reading’s failing to preserve the analogy between time and modal-ity which motivates tense logic. T1 appears to be the reading that Evans himself believes to be analogous to truth-at-a-world; but he also believes that it cannotbe applied in the temporal case without becoming enmeshed in problematic rela-tivism about truth, and that this constitutes an important difference between timeand modality. T2 makes tense logic more like a supervaluational semantics forvague language than like modal logic. T3 involves a “hitherto unknown form ofembedding” (Evans 1985, p. 357), hence, a fortiori, one not known from the priorexample of modal logic.
The details of Evans’s trilemma are not important for my purposes, for Evans overlooks what seems to me to be an obvious way of reconstructing tense logicwhich cannot have any problem of this form. Let the tense logician say not thattime is like modality, but that time is modality. More precisely, let the tenselogician replace all talk of truth-at-a-time with talk of truth-at-a-world.
The simplest way of doing this is to imagine that there is a fact of the matter of what time it is at each world. Each world has, as it were, a perfectly objectiveclock associated with it. Or, to put it another way, each world represents (truly orfalsely) some time as being present. Now construct your model theory for tenselogic using the domain of not of all possible worlds, but of those that differ onlyin what time it is at those worlds. And replace all talk of S’s being true-at-t withtalk of there being a world w where the time is t, and S is true-at-w.
There can be no Evans-style complaint against this. Whatever else may be wrong with it, the problem is not that true-at-w is objectionably unlike the cor-responding semantic primitive of modal logic — for they are the same. PerhapsEvans would complain that “what time it is at a world” is no more intelligiblethan “true at a time”. But that is not the case. Given certain assumptions in themetaphysics of time, it is easy to explain.
In my (2002), I characterised the A-theory as the view that there are intrinsic “A-properties” of pastness, presentness, and futurity, and that it is in virtue ofhaving such properties that things in time are past, present, or future. I imagineda continuum of such properties, stretching from the distant past to the distantfuture, as if — as in Bigelow’s metaphor (1991, p. 6) — past times were colouredin deeper and deeper shades of blue, and future times in deeper and deeper shadesof red.
Possible worlds might differ in how the A-properties are distributed. It might be that there is a world much like ours, except the events of 50 BC have theproperty of presentness, and events thereafter are increasingly future (and eventsprior to 50 BC are increasingly past, though not to as great a degree as in our world). Such a world would be reasonably described as “a world in which thetime is 50 BC”. So, the tense logician can say that S is true-at-t iff there is a worldw where t has presentness, and S is true-at-w. This suffices to answer Evans: theA-theory is at least intelligible — and therefore, truth-at-at-time is too.
References
Bigelow, J. (1991). Worlds enough for time. Noˆus 25, 1–19.
Evans, G. (1985). Does tense logic rest upon a mistake? In Collected Papers.
Parsons, J. (2002). A-theory for B-theorists. Philosophical Quarterly 52, 1–20.

Source: http://otagouniversity.co.nz/philosophy/Staff/JoshParsons/papers/evans2.letter.pdf