This has just occurred to me, I hope it's not totally off base. But it seems that the line of argument behind the Benacerraf dilemma is something like a case of a failure of theoretical identification. So we can imagine that Mr. Body gets killed. The detective shows up and he creates a theory:
1. The victim was killed with a pipe. The victim is a male. He is...
And we can later, given certain facts we know about the victim, we can identify him with Mr. Body.
But if we don't have much to go on with regards to his killer, we just have our theoretical killer in our detective story. We know that this killer killed the victim, and that the victim is to be identified as Mr. Body.
It would seem that the Benacerraf dilemma applies when we think that either Jones or Smith killed Mr. Body. If we are sure that there was only one killer, say because there was only one set of foot prints, or one intruder captured on camera, it seems that we cannot identify both Jones and Smith as the killer. Since we cannot conclusively say that Jones or Smith is the killer, and since both are the best possible candidates, neither can be identified as the killer. We are stuck and have to come up with a better picture.
It seems to me that the Benacerraf dilemma runs the same way with Platonic propositions. If we say that there is a real entity out there that our theoretically posited Platonic entity (proposition) links up to, we need a specific and conclusive entity. If we just say it is some entity or other, this is like saying that Mr. Body was killed by some person or other. This seems to be uninformative and unhelpful.
So when we say:
2. The sentence "There is a date" expresses the proposition that there is a date
it seems that we'd have to give some account of what this proposition is. If it is an entity, which one? If we just say: "Ya, it is some entity or other" or decide to name this entity "p" we don't help ourselves. It could literally be anything. It's like calling Mr. Body's killed "Ned." But who is Ned?
Do I understand this correctly?
This seems like Quine with Radical Translation, that since we can always offer different systems of interpretation for sentences, meanings cannot be nicely pinned down like we want. This seems like Kripkenstein as well, with the worry that past patterns of behavior cannot be linked to rules or generalizations which justify future patterns of behavior. Since we can always offer different systems of interpretation for behavior, dispositions cannot be used to nicely justify conduct.
It would seem that a middle ground position like King, and something similar in Sellars and I'm sure others, nicely avoids the problems for propositions from the Benacerraf dilemma, the Quinean paradox about meaning and the Kripkensteinean paradox of rule following, and the related Wittgensteinean paradox of rule following.
I don't know if this going to be informative to anyone, but it seems rather interesting to me.