>Mattathias Persona scripsit:
>> (2) Compositionality:  Expressions of the language are composed of
>> smaller expressions belonging to the same language, and can in turn be
>> combined into larger expressions.  Expressions are formed according to
>> a finite set of rules.  An expression is interpreted by knowing the
>> meanings of its parts and how those parts are combined.
>Well, obviously no natural human language is *radically* non-compositional,
>but I think it's uncontroversial to say that the amount of
>non-compositionality is variable.

True.  But my point, of course, is that compositionality is a highly
salient property of language - central enough to be a defining property,
I think.  That and creativity are, as far as I can tell, the only two
features of language which set it apart from animal communication.

>What worries me is that the competence-performance distinction is
>often exploited so as to shove most of the non-compositionality
>under the rug (e.g., "Would you pass the salt?" is a muted command, not
>a request for the listener to evaluate his capabilities or intentions).

A better example of non-compositionality would be an idiom like "kick
the bucket" or a formula like "how's it going?" or "far be it from me".
It's true that "Would you pass the salt?" is not interpreted as a yes/no
question, but I think that the use of the yes/no question formula to
express muted commands is prevalent enough to be considered a part of
the grammar.  Even if that's not the case, it's certainly true that
"Would you pass the salt?" contains subparts which are expressions of
the language, and which can be replaced with other expressions of the
language to create well-formed strings with different meanings:  "Would
you pass the milk?" "Would you put away the salt?" "Would you close the
window", etc..  So compositionality certainly plays a role here...

But I agree with you that generativists often downplay the prevalence
of non-compositional utterances in actual language use.  I suppose it's
because of our fascination with that other defining property of human
language, creativity, for which compositionality is responsible.

>Per contra, many sentences that are highly compositional are also
>extremely remote from actual language use (e.g., the notorious
>"Every sailor loves any sailor").

Agreed, which is why I prefer to work with consultants who are not
linguists.  :-)  Nevertheless, I do occasionally inflict mostrosities
like "Every sailor loves any sailor" on my Malagasy speaker, just to see
what happens.  While it's important to focus on how a given language is
actually used, I think it's also useful to test the 'design limits' of
language by subjecting native speakers to highly atypical utterances and
seeing whether they accept them, and if so, how they interpret them.
A strict functionalist would argue that grammaticality judgements for
sentences like these do not constitute real data, that they're not
accessing the grammar but something else, but I don't agree (although I
do treat such judgements with a grain of salt).

>> Are we really so rare as to be a pleasure?  :-)  In my experience it's
>> the functionalists who are the rude ones, always referring to us formalists
>> as if Generative Grammar were some sort of religion with Chomsky as supreme
>> and infallible pontiff!  :-)
>I think the blind hostility is equally distributed among members of
>both camps, but functionalists have for some years felt that their
>viewpoint is not taken seriously (hence Nick Nicholas' characterization
>of Chomsky as "Manufacturer of Consent"), and thus may have felt it
>necessary to bellow in order to be heard at all.

Perhaps.  Again, I resist the characterisation of Chomsky as intellectual
dictator:  While he garners more respect than anyone else in the formalist
camp, he also receives more of a pummeling than anyone else.  Maybe they
worship him at MIT, but here at UCLA every new paper he sends out is greeted
with an extremely high level of skepticism (and not a little annoyance, since
he's a *very* bad writer and extremely hard to read).

But I see what you mean about the functionalists feeling as though they
haven't been taken seriously.  While I don't think they're subjected to
overt hostility (at least around here), it's true that their work is
generally ignored by formalists.  What we have, perhaps, is a situation
of 'transitivity of deprecation':  The hard scientists (physicists,
biologists) tend to look down on us formal linguists, and claim that what
we do is "unempirical" and "not real science".  In turn, many formalists
seem to regard their functionalist cousins in much the same light...

>Hopefully this is changing.

Indeed - if only because the formalists and functionalists could benefit
greatly from studying each other's methodologies, and reading up on
each other's descriptive literature (functionalists seem to have unearthed
all sorts of interesting facts about languages that have completely eluded
the formalists, and vice versa).

Myself, I have a mixed background.  I started as an undergrad in an
anthropologically oriented linguistics program, where we were taught lots
and lots of interesting facts but very little theory.  As a grad student I
came to UCLA, where Chomskyan generative theory reigns supreme, but where
rigourous descriptive work is treated with some respect.

As someone who works on the syntax of 'exotic languages' from a theory-
specific perspective, I find myself reading rather heavily in both the
formalist and the functionalist literatures.  In my case it's inevitable,
given my combination of interests:  Most of the serious descriptive and
analytical work on Austronesian languages (my speciality) has been done by
functionalists, while most of the work on the particular theoretical issues
I'm interested in (having to do with the interaction of verbal morphology
and event structure, and their impact on word order) has been done by


Matt Pearson
[log in to unmask]
UCLA Linguistics Department
405 Hilgard Avenue
Los Angeles, CA 90095-1543