Tensed Second Thoughts
Tensed Second Thoughts
Comments on Richard
Abstract and Keywords
This chapter presents the author's comments on Mark Richard's interpretation of his views about the state of relief. It argues that the account of the truth conditions of utterances expressing relief that some painful episode lies in one's past carries over to the thought itself, that one is relieved. The state of being relieved involves the state itself as a constituent of the object of relief, and demands also the location of that state as present; that is, as coincident with one's own self-consciousness of the state.
In Chapter 3 above I discussed an issue raised by A.N. Prior, concerning the objects of states such as relief, regret, and anticipation, as expressed by ordinary English tensed sentences. For a subject to be in any of these states involves having a conception of the position of other states or events in one's temporal experience. In the case, for example, of relief that some painful episode is concluded, it is crucial that one conceive the episode as lying, not merely in a time that is in fact past, but in one's own past, or its being over and done within one's own present. To bring out this feature of the state of relief, I proposed that it necessarily involved cross‐reference between the state and a constituent of its object, thus making relief what I called a reflexive state. On the view I suggested, what the semantics for English delivers as the truth conditions of my utterance u of ‘I am relieved that my root canal is over’ is (1):
(1) (∃s′) [relieved(I,that (∃s) (over(my root canal,s) & s includes s′),s′) & s′ includes u}
or in paraphrase, there is a state of my being relieved, whose time includes that of my utterance u, and whose content is that my root canal's being over includes that state. The reflexivity of the state of relief is thus shown in the cross‐reference marked by ‘s′’ Since the time of my root canal's being over includes that of my state, which in turn includes the time of my speech, my root canal itself precedes that time, and thus lies in my own past.
In a critical discussion of the article on which this chapter is based, Mark Richard (2003) proposes as one interpretation of my view that, disliking as I do root canals, his (9′), reproduced below as (2), should express a sufficient condition for me to be relieved:
(2) (∃s′) [believe(I,s′,that (∃s) (over(my root canal,s) & s includes s′)]
that is, it should be sufficient that I am in a state of belief whose content is that my root canal is over as of my being in that state. Relief should follow that (p.77) belief. But Richard then constructs a scenario on which (2) appears true, indeed one in which I might even be said to know its content, but I am not relieved, being at the time in the dentist's chair and actually undergoing the root canal.
Now, (2) exhibits the kind of cross‐reference that I conjectured was essential to states such as relief and anticipation. But neither its truth, nor, where s * is a state, the truth of (3), is a sufficient condition for me to be relieved; even obviously so, because the object s * has not been located with respect to my position in time:
(3) Believe(I,that (∃s) (over(my root canal,s) & s includes s *),s *)
Hence any commitment to the sufficiency of (2) or (3) to bring about the state of relief that my root canal is over would constitute a fatal defect for my account. As it turns out, there is no such commitment; but the discussion helps to emphasize the critical features of my or any similar cross‐referential account of the content of some of our states and attitudes.
To his objection Richard conjectures one response that he finds implausible, and then a more complex view, which involves a considerable elaboration of the rather elementary apparatus that I gave. I think that the issue to which Richard calls attention begins farther back. Distinguish first of all the thought at which one is relieved (that a painful or distasteful episode lies in one's past) from the emotional state of relief, which has that thought for its object. Having the thought, or believing or knowing it, is insufficient for the emotional state, as Richard rightly observes. But then what state of belief or cognition would be sufficient? Evidently, one that locates that state itself in one's present, and so locates the state of the painful episode's being over also in one's present, and the painful episode therefore in one's past. On the assumptions in force, for the case of saying, ‘I believe that my root canal is over’, the location of the state is given by its relation to (simultaneity with, or temporal inclusion of) the utterance itself. Belief is then sufficient for relief. But for the case of merely thinking that my root canal is over, the location is not automatic; or so I take Richard as suggesting. Thus he gives his (9), reproduced below as (4), as sufficient for being in the state of relief:
(4) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s includes s′),s′) & s′ includes u]
But he then proposes that ‘since I might have the belief without speaking’, his (9′) (=(2) above) alone should be sufficient (Richard 2003), which it isn't.
(p.78) What Richard's considerations bring out, I think, is that there must, in the occurrence of the belief or cognition that is sufficient for relief, be an element that plays the role of the utterance u in the avowal of that state. In my article I had glossed over this point, remarking only that the presentness of a present mental state was, in Shoemaker's formulation, ‘immune to error through misidentification’, so that given (5) in analogy to (4), where e is a mental particular, there was no question of the falsehood of ‘s′ includes e’:
(5) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s includes s′),s′) & s′ includes e]
What then is e? By analogy to the case of asserting, ‘I think my root canal is over’, where my own consciousness of myself as making the assertion through the utterance u locates my state of belief in my present, we may take e to be the event of my affirming the content of (5) itself. Then the thesis becomes: although the content of (2) is not sufficient for relief, that of (5) is sufficient. In affirming (2) I merely affirm, in effect, ‘There is such a thing as my believing that my root canal is over’; but in (5) that there is such a thing temporally coincident with my affirmation. The latter implicates the feeling of relief, the former not.
If only (5), which incorporates tense in the clause ‘s′ includes e’, will suffice, does it follow that we cannot but think in tenses, or as Richard puts it are incapable of untensed thoughts? No; but it does follow that the feeling of relief presupposes an element of self‐consciousness, an element that is masked when one considers only public utterances, which are by their nature necessarily self‐conscious acts. Evidently, the states of relief, anticipation, and regret presuppose belief if not knowledge. If I am right, however, they presuppose more than this, namely the capacity to locate one's own belief states with respect to one's current affirmations or other mental events.
I have argued above that the account of the truth conditions of utterances expressing relief that some painful episode lies in one's past carries over to the thought itself, that one is relieved. The state of being relieved involves the state itself as a constituent of the object of relief, and demands also the location of that state as present; that is, as coincident with one's own self‐consciousness of the state. But now that the latter point has been made explicit, it may be questioned whether the apparatus of cross‐reference was essential to begin with. Why not say simply that the object when I am relieved is just (6):
(6) My root canal is over now.
rendered as (7):
(7) (∃s) [over(my root canal,s) & s includes now]?
(p.79) The word ‘now’ contributes to (7) nothing but its reference, the time of utterance or, in the scenario lately envisaged, of inner affirmation. If, however, I am affirming it, then we can let that affirmation e, or its time τ(e), serve without cross‐reference between the state of belief and its object, obtaining instead of (5) the formulation (8):
(8) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s includes e),s′) & s′ includes e)
For the particular case of present avowals, indeed, there is nothing much to choose between (5) and (8): on either rendition, we have the time of the critical state including that of my avowal, directly in the case of (8), by swift implication in the case of (5). But notice that, even if (8) is chosen, my state of belief, hence of relief, continues to be reflexive, although so to speak at second remove, through the cross‐reference marked by ‘e’ between my affirmation and the object of my belief.
When we turn to embeddings under the past tense, however, the situation changes. For interpretations in which a subordinate past tense is taken as expressing the past relative to the past state in the superordinate tense, or where as in classical consecutio temporum it is taken as non‐past but expressing simultaneity with the superordinate state, there is an anaphoric relation between the state arguments, or so I suppose. Thus (9) may be taken either as (10) (asserting the existence of a past belief state whose content was that my root canal's being over preceded that state), or as (11) (asserting the existence of a past belief state whose content was that my root canal's being over included that state):
(9) I believed (then) that my root canal was over.
(10) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s<s′),s′) & s′<u]
(11) (∃s′) [relieved(I,that (∃s) (over(my root canal,s) & s includes s′),s′) & s′<u]
If so, then we should recognize as a general principle that the anaphoric connections run from the coordinate s′, the first coordinate of the superordinate tense, to the second coordinate of the subordinate tense, as in the original formulation (5). The crucial case is that represented in (11), where (9) is taken as asserting the existence of a past, intuitively present‐tensed, state, my belief in which was sufficient for me to have then felt relieved.
Throughout this chapter I have assumed that indexical words contribute to what is expressed nothing but their reference; but it may be proposed (as by Mark Balaguer (1997), in his contribution to the Santa Barbara conference) (p.80) that since (12) and (13) below may differ, in that assertions of the latter, but not the former, will be accompanied by the feeling of relief, their objects express different thoughts, even under circumstances guaranteeing that they will be true or false together:
(12) I believe that my root canal will be over then.
(13) I believe that my root canal is over now.
These examples are useful in expounding more fully the view I defend here. If u 1 and u 2 are my utterances of (12) and (13) respectively, then (12) is as in (14), and (13) is as in (15):
(14) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s>u 1 & then(s)),s′) & s′ includes u 1]
(15) (∃s′) [believe(I,that (∃s) (over(my root canal,s) & s includes s′ & now(s)],s′) & s′ includes u 2]
If s * is a state that satisfies what follows the outer existential quantifier in (15), then the objects of belief, namely the contents
(16) that (∃s) (over(my root canal,s) & s>u 1 & then(s))
of (14), and
(17) that (∃s) (over(my root canal,s) & s includes s * & now(s))
of (15) respectively, are certainly going to be different in their structure. That my belief in the content of (16) does not bring relief follows at once from the fact that the state s follows my self‐conscious utterance u 1; but to bring out that (17) does suffice, the information that s is located in my present must be recoverable from my use of ‘now’. How is this to be done? Evidently, the adverb or the present‐tense morpheme in (13) must not only refer to the time of my speech, but also place the state of my root canal's being over in temporal coincidence with it. Let this be so.
But then the modal statement (18):
(18) My root canal might not have been over now.
must ignore this very feature of the adverb and the tense; for its truth is obviously independent of what I may happen to utter or to think. Could there be a conception of indexical meanings or concepts that can look both ways, on the one hand incorporating information about the agent's temporal position, on the other allowing the simple semantics for the modal statement, to which this information is irrelevant? It is not clear (p.81) that there could: explication of ‘now’ as the present time simply pushes the problem back a step, and indeed (19), like (18), is indifferent to all but the value of that time, and so indifferent to whether the speaker thinks of it as ‘present’:
(19) My root canal might not have been over as of the present time.
In my original article I left the dual behavior of indexical expressions, as sometimes importing into content the principles that govern their use, but sometimes banishing those very principles from content, in the indeterminate state where the data seemed to place it. More recently (Chapter 10 below) I have elaborated the possibility that the proper form of a theory of truth for indexical expressions will systematically imply the banishment of the content‐determining rules from content. On this view, indexicals are governed by rules of use, whose content figures in the antecedents of conditional truth conditions of whole sentences. The rule of use for ‘now’, for instance, is that it is to be deployed as a predicate expressing the simultaneity or inclusion of one's own utterance (or its time) in another state as given by the sentence. As applied to (6), this account gives conditional truth conditions in (20):
(20) If u is an utterance of (19), and the speaker s of u uses the utterance of ‘now’ therein as a predicate P true of just those states that temporally include u, then u is true iff: (∃s) [A(s) & P(s)].
where ‘A’ abbreviates the contribution of the linguistic material apart from the adverb (and for simplicity I have abstracted away from the first person and the present tense). Given an utterance of (6) satisfying the antecedent of (20), and instantiating so as to detach the biconditional consequent
u is true iff (∃s) [A(s) & P(s)]
the right‐hand side contains nothing of the application of the rule of use for ‘now’ but the value of the predicate P. Thus we capture the modal behavior of (6), or the proper truth conditions of (19), but lose the information necessary to set up a content that, apprehended by the speaker, becomes an object of relief.
Suppose that, as far as semantics goes, the relevant parts of the whole story about (6) are as in (20). What consequences are there for a general account of the relations between the words we utter and the thoughts that we express? If I am right about the general form that semantic theory should take, and also about the structure of thought that is necessary to bring out the circumstances under which we are in states such as relief, then we should (p.82) construe the representations of thought expressed as extending beyond the narrowest truth‐conditional conception of semantics, to include the circumstances of our deployment of indexical concepts. The words we use do robustly express our thoughts; but to see them as so doing, we should take account of the principles that set up our assertions, and not just what we assert.