Carnap and Grice on "logical" versus "non-logical"/"descriptive"

By J. L. Speranza, of the Grice Club

for the Carnap Corner


----

I have shared this info elsewhere, but it amuses me and it comes straight from the pages of the OED -- with which Grice SHOULD have been more familiar with! ("I don't care what the dictionary says!" "And that's where you make your big mistake", got the rebuke from Austin)

-- A pirot, the OED has as follows

-- begin cited text:



pirot.
[Apparently < French pirot (1611 in Cotgrave: see quot. 1611 at sense 1),
although this is apparently not recorded elsewhere, and is of unknown
origin. Compare PIDDOCK n.]

1. A razor shell.

1611 R. COTGRAVE Dict. French & Eng. Tongues, Pirot, the Pirot, or
Hag-fish; a kind of long shell-fish.

2. A piddock.

1686 R. PLOT Nat. Hist. Staffs. vii. 250 A sort of Solenes (which the
Venetians call Cape longe, and the English Pirot)..a kind of Shell-fish deep
bedded in a solid rock.

-- end cited text.


---

In "Meaning and Necessity" Carnap uses the terminology (loose, i.e. he is not doing 'carnaps' here -- (a "carnap" defined in the philosopher's lexicon as "a formally defined symbol, operator, bit of notation")),

Pirots karulise elatically
A is a pirot
____________________________

A karulises elatically

----

He would have (as per "Meaning and Necessity" and intuitively enough):

"is" and "a" as 'logical words' -- especially in their formal counterparts.

'pirot' (as well as 'karulise' and 'elatic') as 'descriptive' words.

----

This has a slight connection with Grice, as misunderstood by Cohen.

In an infamous paper, Cohen ("Grice and the logical particles of natural language"), argues against Grice's thesis for things like 'if' -- which will feature in the logical form of

'pirots karulise elatically' (on the standard reading).

The meaning of the horseshoe -- the meaning of 'if'?

"if" as a logical (non descriptive) word. Yet not a variable, so a "logical constant" word.

When it comes to 'truth-functional operators' in the bivalent system that both Carnap adn Grice abode by, the thing is easy enough: the truth-table will do it for us.

When it comes to "(x)", for every x, for all xs, if x is a pirot, x karulises elatically, the idea of talking of the 'meaning' of '(x)' is somewhat trickier, in that it ceases to be algorithmically decided by a truth-table, but reference to general guidelines (what guideline is not general?) for standard models in standard interpretations -- in the semantics sub-domain of the formal system -- are required. Nothing too fancy: (x)(Px --> Kx) will be true iff all items falling under the extension of P are and none of them can be shown to fail to be K.

Etc.