Subcategorisation

[Colin Fine <C.J.Fine@BRADFORD.AC.UK>]

John answers me:
+++++++>
 > I now claim that the distinction between set, masses and individuals is an
> obligatory grammatical SUBcategory. The level of abstraction is another such
> (I'm still wondering about this idea of 2nd level abstractions).

What is the distinction between a "grammatical category" and a "grammatical
subcategory"?
>++++++++

I'm not sure if 'grammatical category' is well-defined or not, but what I would
understand by it is something like a selma'o: getting the wrong one will deliver
a text that will not parse as a sentence.
By 'subcategory'  I mean a classification which will not vitiate the syntactic
parse, but would be rejected by a hypothetical analyser which considers
subcategorisation frames. An important feature of subcategories is that
the distinction can be neutralised, as I suggested in the example of xamgu at
the end of the mail.

++++++++>
 I take your point, although it is a confusing one.  I believe, however,
that it is sound to say things like:

        le cmima cu cmima le se cmima
        The member-I-have-in-mind is-a-member-of the set-I-have-in-mind.

even though the x2 of the outermost "cmima" (the second one in surface order)
normally would be followed by a description beginning with "le'i" or the like.
So although "le se cmima" is one or more sets viewed as individuals, this
view does not make them other than sets, and sets at the "se cmima" level
are the same sets as those at the "le'i" level.
>+++++++


The point *is* subtle, but I believe it is important.
It is certainly sound to say what you have said, but remember that it
actually means

The member-I-have-in-mind is-a-member-of ALL the sets-I-have-in-mind.

I would be cautious about your statement 'normally would be followed by a
description beginning with "le'i" or the like'. This is not incorrect, because
of the 'normally', but I think it needs more explication.

I've been thinking it out further, and I have an Essay on Subcategorisation
which I shall send in the following mail.

        Colin