May Bes and Might Have Beens
My subject in this paper is particular possibilities: the may-bes and the might-have-beens that relate essentially to particular individuals or situations. The topic is one which is apt to evoke very different responses from different philosophers. Some detect, or think they detect, an intoxicating scent of something more metaphysically interesting than either merely epistemic possibilities on the one hand or merely de dicto possibilities or necessities on the other. My remarks will not give much satisfaction to them. Some, on the other hand, are suspicious of modalities in general as being dubiously coherent notions. But it is clear that whether we believe there are such things as particular possibilities or not, we are in practice bound to take account of them.
KeywordsBlow Down Rational Inference Past Situation Electoral Committee Intelligence Officer
Unable to display preview. Download preview PDF.
- 1.At the limit of this process we can consider might-have-beens in relation to a period at which no one envisaged any may-bes at all because there was no one to envisage them. In such a case it is, of course, later knowledge which supplies all the facts of which we hold that they did not make it certain that such-and-such (which did not in fact happen) would not subsequently happen, i.e. the ‘available facts’ in relation to which we may say that such-and-such might have happened. But nothing in our saying this implies any indeterminism in what did happen.Google Scholar
- 2.Objection 1. Suppose a’s Oing, or not piling, is the sort of outco me which can reasonably be regarded as exclusively subject, at least from the reference-time onwards, to deterministic laws. In that case our knowledge that a did not in fact (or did in fact) (ß entitles us to infer, regarding the reference-time, that circumstances then obtained which were sufficient to ensure a’s not aping (or thing). So the condition which the analysis requires for an acceptable ‘might have been’ in such a case, viz. that knowledge now available regarding the reference-time is insufficient to warrant the rational inference that a did not 4 (or that a did 0), is never fulfilled. Indeed a convinced universal determinist would have no use at all for the idiom in the sense expounded.Google Scholar
- Reply. The objection construes ‘now available knowledge’ too widely. What is required for denying a reasonably supported assertion that there was a chance that a would di (or that a would not tp) is some more specific knowledge of conditions obtaining at the relevant time than can be derived from the premise that a did not in fact cp (or did in fact 0) coupled with a general conviction that a’s (ping or not dying is subject to deterministic laws. There is no reason to think that convinced universal determinists would (or do) eschew the idiom in the sense expounded or confine themselves to denying others’ uses of it.Google Scholar
- Objection 2. Suppose the reference-time is t. Then knowledge available at one time (say 1970) of the relevant circumstances of a at t may seem to warrant the assertion that a might have did, whereas additional knowledge available at a later time (say 1975) may seem to warrant its contradiction. The analysis then requires us to say that the proposition that a might have 4 d, relativized to t,is true in 1970 but false in 1975. This conclusion is unattractive.Google Scholar
- Replies: (a) There is not one proposition but two; for the analysis of the idiom brings out a concealed indexical element in it, viz. a reference to the state of knowledge at the time of utterance.Google Scholar
- b) The ‘analysis’ may be seen as giving not truth conditions but ‘justified assertibility’ conditions. The notion of truth conditions is inappropriate to the idiom itself, though not, of course, to the statement of justifying conditions.Google Scholar
- Evidently replies (a) and (b) are mutually exclusive alternatives. I leave open the choice between them.Google Scholar
- In formulating these objections and replies, I have been greatly helped by the contributions of Professors S. Kripke and H. Gaifman to the discussion of this paper.)Google Scholar