LINGUIST List 3.459

Tue 02 Jun 1992

Disc: Rules

Editor for this issue: <>

Directory

1. , rules
2. , Normative Rules

Re Alexis Manaster-Ramer (Linguist 3-371)

>The point that is being made is that a sequence like "man the"
>would not constitute an NP even if it occurred in the speech
>of an English speaker.

There seems to be some need for clarification here. First we
havee claimed that a sequence like "man the" or "dog the" or
whatever may very wll occur in English but that would not
necessarily tell us anything about the structure of NPs. Take for
instance an utterance like "He gave the dog the bone". Here we do
indeed find the sequence "dog the", but no linguist we know of would
take that as a counterexample to the rule that the article precedes
the noun in English NPs/ This practice of "sifting" your data in
linguistics is, we believe, no different from the practice in other
empirical disciplines.

Our second point is this: the main issue is not that a sequence like
`dog the' *couldn't* be an NP in some sense (intended as one,
interpreted as one by someone who was clued in on the word-order game
being played, parsed as one by someone whose parser has relaxed the
constraint that the SPEC precedes the head, etc.), but simply that the
mere fact of the appearance of such a sequence in someone's speech
wouldn't in and of itself be evidence about the grammar internalized
by English-speakers (or, more precisely, the mental structure that the
grammar is an attempted description of). Isolated events, especially
without a rich description of the surrounding context, simply aren't
useful information, since you can't make sensible judgements about
their possible causes.

Hence we are not Itkonen-style normativists in principle, although
we do believe that grammarians pretty much have to describe linguistic norms
in practice, given that there don't seem to be reasonable methods for
getting sufficiently extensive and repeatable results about idiolects.
And we do agree with Alexis that linguistics needs to distinguish
between `normal' and various other kinds of language use, where the
constraints of the grammar won't necessarily apply.

And, re Nyman (3-441), another way of looking at it is this. In order
for the speakers of a language to distinguish the `normal' from the
`non-normal', they have to have some `circuit' in their heads that gives
a different reading in the two kinds of cases. We regard the nature of
this `circuit' as the fundamental question of grammatical theory, with
grammatical norms arising as a consequence of its nature and various
other things (such as the whatever it is that tends to cause members
of a speech-community to assimilate their grammars to each other).

I enjoyed Alexis Manaster Ramer's recent posting. And I think he's right
that there is a problem. But I would argue that the problem is not
limited to linguistics. And this suggests that the solution shouldn't
be phrased in terms specific to linguistics.

By parity of reasoning, much of psychology would be "normative" in
Itkonen's sense: By mistake or as a joke, people can *do* all sorts
of things. (Generalization: if deprived of water for 48 hours, subjects
will drink upon presentation of water -- unless they don't want to!!)
Reflexes might be invariant; but what about everything else?

What's more, the predictions of more familiar sciences are subject to
"ceteris paribus" clauses. Small physical objects, upon release, will
approach the earth -- unless a strong wind comes along, or some bozo
catches it, or...

Of course when doing experiments, scientists don't list all the things
which could have gone wrong, but didn't. (Surely they couldn't give such
a list.) They merely leave as background that all else *is* equal.

True enough, statements about the behaviour predicted given a particular
competence are also subject to ceteris paribus clauses. (For instance,
we might have to stipulate that these are "normal" utterances.) But that's a
problem which shows up all over the scientific map. And it should be
understood as such.

Just as water might boil at odd temperatures if it contains impurities,
or if the air pressure is changed, speakers might say "Box the". We all
agree that this is a problem. But it doesn't single out linguistics.
(Nor do I think they differ in a matter of degree, say degree of precision.)
And it doesn't, so far as I can see, call into question the notion
of competence.

Best,
Rob Stainton
MIT