The general section then is 8.
----
He has a caveat that:
"the task of providing a semantics" for a formal system, "might ..." be "discharged".
"the procedure" which he suggests will be "intuitive" -- and thus in no need for explicitness.
----
and it won't "contravene [any] philosophical ideas underlying the construction" of the system.
-----
The first sub-section of this section VIII then, concerns
"Interpretation".
------ Grice provides four steps in "the provision of an interpretation Z"
1. The specification of a non-empty domain D.
2. The assignment of each propositional letter either to 1 or to 0.
3. The ASSIGNMENT of each n-ary predicate constant eta to a SET ... of ordered n-tuples, each of which has, as its elements, elements of D.
---
4. The assignment for individual constants.
----
In 4 he distinguishes between the correlatum and a designatum. Designatum he restricts for members of the domain D.
----
At this point, it is I think that R. B. Jones noted a BIG TYPO, where the text reads "description" rather than 'designation', but I will check if that is the point Jones made. I think it was.
---
In the B section, he deals with
1-correlata and 0-correlata.
He entitles the thing, "Truth and Validity". This should shed light then on Grice's rather standard views on 'entailment', etc.
----
In this section he provides the truth-tables for the connectives.
The Megarian Grice for example.
---
--- At this point, for the semantics of quantified formula, Grice is explicit that he is "closely following Mates's procedure in Elementary Logic".
---
It is at this point that Grice introduces 'valid':
"Let us also suppose that we shall define validity in Q by stipulating that phi is valid if Q iff, for any interpretation Z, phi is Corr(1) on Z."
----
He provides for specific "semantical rules" for his vacuous-names cases, and concludes that section:
"Validity ay be defined as follows: phi is valid in Q iff, for any interpretation Z, phi is Corr(1) on Z. Finally, we may, if we like, say that phi is true on Z iff phi is Corr(1) on Z."
Subscribe to:
Post Comments (Atom)
Posts
I think its correct to describe this as a "truth conditional" semantics (though more often it would be called "model theoretic", which usually is but I think need not necessarily be truth conditional).
ReplyDeleteIn the light of the questions raised on the matter of compositionality which cast doubt on whether the meaning of a proposition can be merely its truth conditions, it is of interest whether Grice commented on any possible shortfall in this kind of semantics.
In his retrospective epilogue, where he talks about central meaning, we see that the truth conditions are considered to be part of the central meaning, though not necessarily the whole.
So far as I recall, there is not much said about what else there might be, though there are these other criteria which he mentions "dictiveness" or "what is said" is one.
Do we know whether Grice's conception of "formal semantics" is confined to truth conditions, whether he construed it as potentially encompassing the whole of "central meaning" or even as going beyond central meaning?
Roger Jones