One part of A nice derangement that we have not directly addressed so far is that about Donnellan: the referential and attributive uses of definite descriptions. I enjoyed the droll interplay, and the link to modal logic; and the observation that one can say something that is true, using a sentence that is false. — Banno
Interesting that you should mention this, because that particular section has definitely captured my attention. Due to the overwhelming compelling interest that I have in the general subject matter, I'm very carefully studying this part at the moment. I'm reminded of a recurring issue that rears it's head elsewhere. It relates to meaning and/or reference, and seems quite relevant to the Donnellan distinction between two uses of definite descriptions and the MacKay objection to that distinction as discussed by Davidson.
I suspect that the conclusion that one can say something true by using a sentence that is false is based upon conflating the very nuanced but quite remarkable differences in both, the meaning and the truth conditions of what are otherwise identical
looking statements. Neglecting that distinction is the recurring issue I'm seeing here and elsewhere(Gettier in particular). I'll try to explain concisely showing the relevance to the paper.
Identical looking statements can and do mean very different things, particularly when examined in a more general sense, such as when completely isolated from the individual speaker(what words mean according to Davidson) as compared/contrasted to what the exact same words mean when they are examined as a belief statement(what a speaker means according to Davidson). The same sentence can and does have very different meaning/reference, and thus very different truth conditions depending upon whether or not we're examining what the words mean(in general) or what the words mean when used by a specific individual speaker.
For example, and in at least partial agreement with Donnellan, I too find that "Smith's murderer is insane" is true when and if "Smith's murderer" refers to an insane individual, regardless of whether or not that individual murdered Smith. I disagree that that is a false sentence when we examine it as an individual belief statement, for the referent of "Smith's murderer" does not need to have murdered Smith. Rather, the referent need only to be insane. That holds good because "Smith's murderer" is doing the work of a rigid designator(pace Kripke in Naming and Necessity). It picks an individual out of this world, to the exclusion of all others. If the individual picked out is insane, then "Smith's murderer is insane" is true because the individual picked out is insane.
It is only when we examine the statement as a general one(divorced from the individual, and such not as a belief statement) that problems arise. That becomes a problem, because "Smith's murderer" no longer refers to the same individual that Jones picked out. Rather, when divorced from Jones and taken in general; the referent, the meaning, and the truth conditions all change accordingly. In such a general case, "Smith's murderer" only picks out the individual that murdered Smith, to the exclusion of all others. In this example, the person that murdered Smith is not insane. There is no problem however, because Jone's most certainly referred to someone else, and that person was insane, despite the fact that a misnomer was used. What we have here are two identical
looking statements with very different referents, meaning, and thus truth conditions. The only problem I see is neglecting the differences between them.
That same neglect is also true of Gettier, and it is
the very foundation that gives rise to 'Gettier problems'. For example, "The man with ten coins in his pocket will get the job" refers to completely different people depending upon whether we examine it in general(any man with ten coins will do here) or as a particular belief statement of Smith about himself(only Smith himself will do here). Gettier conflates the two. Smith was, of course, talking about himself. He did not believe that anyone else would get the job. Hence, the recurring problem mentioned earlier. That's worth mention, but I digress...
Contrary to Donnellan, and perhaps in some agreement with Davidson(based upon his mention of the disconnect between MacKay's objection and Donnellan's answer)I also do not think MacKay's Humpty Dumpty objection is valid as a result of all this. Rather, I think Donnellan granted far too much, because it is not at all uncommon to use language like that. Perhaps the underlying importance of intentions for Donnellan represents the bulk of the actual problem(s). It could also be the result of the distinction drawn between what words mean and what a speaker means, for that distinction does not successfully do the job needed, as shown above. Davidson also denies that the truth conditions of the statement change according to the meaning, which is very odd to me. He said the following...
Jones’ belief about who murdered Smith cannot change the truth of the sentence he uses (and for the same reason cannot change the reference of the words in the sentence).
That is a point of contention. I would strongly disagree, in part based upon what's above. In addition, I would also charge Davidson with having this the wrong way around. The truth conditions of a statement are determined by the referent and what's being said about the referent, not the other way around. It's not so much that Jones' belief about who murdered Smith
changes the truth of the sentence. To quite the contrary, the referent of "Smith's murderer" and Jones' belief about the referent wholly
determines it's truth conditions, in the exact same way that Smith's belief determines those things in Gettier's Case I.
Tangentially:The distinction I've set out above may pose insurmountable problems for the T sentence, because the T sentence is incapable of drawing the aforementioned remarkable distinction between the meaning and truth conditions of identical statements(in general as compared/contrasted with individual belief statements). I see no way for the T sentence to disambiguate these remarkably different truth conditions and/or meanings of identical looking statements. Perhaps, Davidson knows this and as a result denies the distinction I've set out here as a means to salvage his project(save the T sentence).