The Reality of Sentences

[Rob Freeman]

On Sunday 13 April 2003 6:27 pm, Salinas17 at aol.com wrote:
> In a message dated 4/12/03 11:01:31 PM, rjfreeman at email.com writes:
> << Is there anyone else who didn't agree with my characterization of TG's
> "pragmatic middle" argument as a confusion between Chomskian competence and
> performance? >>
>
> Let me ask how in the world one can conclude anything about "competence"
> without observing "performance"?

No debate on concluding competence without observing performance.

> And -- think about it -- when exactly is performance possible without
> competence?  Can a subject without any language competence exhibit any
> language performance?

Right, the two occur together.

> How can these two concepts be seen as operational or theoretical opposites?
> This seems to be the equivalent of making two legs the opposite of walking.
> Or, for that matter, trying to understand the biological origins and
> structure of human legs without understanding walking.

Exactly, that is why I say Tom was confused. He was opposing these two
different things as if they conflicted. Generative grammar rules are rules
describing competence. Why then does he compare them with a system (emergent
generalization) specifying performance. Why indeed does he compare them to
performance. Clearly any performance will have a competence (100%), and any
competence will relate to some performance (100%).

Now it is true that while any set of rules describing competence should be
100% true, they need not be 100% complete. But let alone 90% of all
sentences, if Tom can show me a rule describing _all_ the structure of any
one sentence I would like to see it. The 90% figure would just be the
percentage of sentences for which any single consistent system of rules
describes _some_ of the structure.

We don't need to find a "pragmatic middle" between generative grammar rules
and emergent generalization, one is a model for the competence of language,
the other for its performance. They are just different things.

To be fair to Tom it is a common confusion, but he should not use it to
dismiss emergent generalization as the only candidate for a model of language
performance.

> The contrast that was drawn was actually between "generative" and
> "emergent". From my individual point of view, it is the difference between
> stasis and change.

This might be the core source of confusion. I keep getting this perspective
that sees emergence (what I call emergent generalization) only as an
explanation for how general categories change. I see it as an explanation for
the language performance underlying the perception of categories as well.

This is something I am trying to clear up.

> From my point of view, it's plain that the "pragmatic middle" is not a
> compromise, but a recognition that generative analysis captures a certain
> facet of the phenomenon, but does not capture it all or explain its
> existence.

I'll accept that. Though I can't help feeling that calling theoretical
duality a "pragmatic middle" is doing something of a violence to the language.

Anyway, that still leaves us looking for a model which "explains its
existence". I think emergent generalization is such a model.

-Rob