Re Personal pronouns

[Larry Trask]

[ moderator re-formatted ]

Pat Ryan writes:

>  I have no problem with Larry's definition of 'determiner' as recorded in his
>  dictionary; and I have no problem acknowledging the usefulness of
>  categorizing several different kinds of words
>  ("articles...demonstratives...quantifiers...", etal.) as 'determiners'.

>  But Larry seems to me to be denying the validity and usefulness of the term
>  'possessive pronoun' altogether though, of course, "possessive" is included
>  as one of the various categories of 'determiners'.

>  Larry's definition of pronoun contains all the classses traditionally
>  considered pronouns *except* possessive pronouns though it seems odd in the
>  extreme that, in  spite of the rejection of possessive pronouns,
>  demonstrative pronouns are included *as well as* being listed under
>  'determiner' as "demonstrative".

>  Perhaps you could explain to me why you think "demonstrative" may occur in
>  both classifications but "possesive" may not.

OK; I'll have a shot.

The labels 'pronoun' and 'determiner' denote *syntactic* categories -- that is,
categories established on the basis of grammatical behavior, such as
distribution and inflection.  Such categories are set up without regard for the
meanings or functions of their members.

But the labels 'possessive' and 'demonstrative' are quite different.  These are
not syntactic categories at all, but rather semantico-functional categories,
set up on the basis of the meanings or functions of their members, regardless
of their grammatical behavior.

Now, there is no requirement that any syntactic category should exactly
correspond to any semantico-functional category.  Often, of course, we observe
a high degree of overlap, but perhaps never perfect overlap, and sometimes the
match is not good at all.

Now, in English, we can set up the syntactic categories 'pronoun' and
'determiner' by the usual grammatical criteria.  When we do this, we find that
the class of pronouns includes, among others, the items 'she', 'mine' and
'this'.

And we find that the class of determiners includes, among others, 'the', 'my'
and 'this'.

Simple test for pronouns: ___ is nice.

Simple test for determiners: ___ book is nice.

So that's that: 'my' is strictly a determiner, while 'mine' is strictly a
pronoun, and 'this' can be either a pronoun or a determiner.

Now to the semantico-functional categories.  We can set up a class of
'demonstratives' on the basis of deictic properties, and we find that both
determiner 'this' and pronoun 'this' must be included in the demonstratives.
And we can perhaps set up a class of 'possessives' on some basis or other (this
is not so easy), and find that both determiner 'my' and pronoun 'mine' deserve
to be included in this possessive class.

But the two very different types of class are independent in nature and
independent in membership.  The observation that the pronoun 'mine' is a
pronoun does not entail that the determiner 'my' is also a pronoun, merely
because both are possessives -- just as noting that 'my' is a determiner does
not make 'mine' a determiner.

In the same way, the observation that the event-word 'destruction' belongs to
the syntactic category 'noun' does not entail that the event-word 'destroy'
must also be a noun.  Syntactic categories are independent of
semantico-functional categories.  That's just the way languages are.

>  Finally, I think it useful to retain the term "possessive pronoun" because
>  it overtly identifies the fact that the pronoun stands for a noun in the
>  possessive --- in addition to its function of deteremination.

No; I'm afraid I can't agree.

Whatever the intended meaning of "stands for" might be, a pronoun does not
"stand for" a noun.  A pronoun doesn't even belong to the category 'noun'.
Instead, it belongs to the category 'noun phrase', and, in many cases, it takes
its reference from another noun phrase overtly present in the discourse.

Example:

Q: "Where's the woman we're supposed to meet?"
A: "She's over there."

And *not*:

* "The she we're supposed to meet is over there."

A pronoun does *not* "stand for" a noun in any coherent sense.

Now, of course, 'my' is somehow related to 'I', at least semantically.
Depending on your theoretical tastes, you might like to state a rule -- say, a
lexical rule -- of the following approximate form:

	[I] + [Poss] --> [my]

This rule is comparable to other conceivable lexical rules; for example:

	[arrive] + [-al] --> [arrival]

But 'arrival' is not a verb because 'arrive' is a verb: it's a noun, as shown
by its grammatical behavior.  Likewise, 'my' is not a pronoun because 'I' is a
pronoun: it's a determiner, as shown by its grammatical behavior.

Finally, I might add that 'possessive' is a rather elusive category anyway.
Whatever 'possessive' might mean -- and this is far from being a trivial
question -- it has no uniform grammatical expression in English.  In fact,
English has *at least five* different sets of forms and constructions for
expressing relations we might reasonably call 'possessive':

	(1) Determiners ('my', 'your', 'their')
	(2) Pronouns ('mine', 'yours', 'theirs')
	(3) The Saxon genitive: NP + 's ('Bill Clinton's affairs)
	(4) Prepositions ('the capital of France', 'the man with the gun')
	(5) Juxtaposition ('the Microsoft story')

These highly varied forms and constructions have nothing in common
syntactically, and they cannot be coherently squeezed into any single syntactic
category -- even though we *might* like to say that they all express
possession, whatever that is.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk