🚀 go-pugleaf

>Posted on 14 Dec 1992 at 15:14:53 by Randall Holmes

>>>No, Randall showed that a conviction\belief can be true even if
>>>non-checkable. (It does make asignment of truth value in those
>>>cases rather arbitrary, though.)
>>
>>Huh?
>>I have no idea why you think he showed that.
>
>Neither do I.

Oh, good.

> I'm on Jamie's side of this argument; truth (of
>sentences in a certain language) is not defined in terms of observers
>or interpreters.  Take my argument and apply it, further, to the truth
>or falsehood of a sentence which has never in fact been uttered by
>anyone (it would be self-defeating for me to give a counterexample,
>but counting arguments show that there are such).

Of course, we can DESCRIBE a sentence which has never been uttered, and
it STILL will never have been uttered. So you CAN give a counterexample.

I'll take a shot!

Consider the sentence which would come first on an alphabetized
list of 100-word sentences of English.

(A small prize to whoever can utter it first.)

It is either a true sentence of English or a false one. No one has
ever interpreted it (I'll wager). With the possible exception of
one of us, no one ever will.

> I don't agree with
>Jamie on "types and tokens" (sentences are not equivalence classes of
>tokens), precisely because the canons which determine truth or
>falsehood of a sentence could in fact be applied to the so far
>unuttered sentence which has no "tokens" at all; I think of sentences
>as abstract objects of the same order as numbers.

I'm not sure what the disagreement is supposed to be. I think I agree
with the above.
I allowed that Jimbo could think of *propositions* as equivalence
classes of sentences, for most purposes. I was then thinking of
sentences as abstract objects. I usually think of types as
abstract objects.

Jamie