AW: Increasing interest in the HPSG conference

[Andrea Dauer]

Hi,

some remarks on Carl's recent posting.

> A more important distinction than whether the tree is gussied up into a
> feature structure is whether the tree is viewed as (a) a model
> (structural representation) of a sign, or (b) as a proof (derivation)
> of a sign. I believe this is a (maybe even THE) fundamental issue that
> has to be clarified

This should indeed be a center of debate, and I found it somewhat surprising
that the recent discussion did not even touch the following question: Can an
empirical justification for derivations be given? (Ok, we have to specify
what an empirical justification can be, it definitely will contain *some*
theory, as e.g. what we accept as a tenable analysis, whether competing
analyses can be compared, whether analyses can be compared at all ...)

My personal answer to Carl's question is as follows: quite obviously, trees
are what McCawley conceived them to be, i.e. structural representations. But
returning from the personal to the general, I would like to add the
following.

In the 1980s, it was somewhat vague whether GB should be considered
derivational, and explicitly representational variants like Koster's
existed, but then Chomsky and Pollock argued that certain well-known facts
about adverb position, negation and the like could not be handled in a
representational framework. An actual 'proof' in the broadest sense of this
term has never been given (and actually cannot be given, cf. Kim and Sag).
What is more, Lasnik (2000, cf. my short notice in JL) has pointed out that
the derivational analyses offered by Chomsky and others do not work!
Strangely, this point has been neglected inside *and* outside MP. Why is
that?

> Enemies are authoritarianism, narrow-mindedness, vagueness, to name a few.


Agreed, but let's try to clean up our own closet first. If your interested,
I can name some nice instances of the aforementioned faults in our small
parish. They only show how seriously Chomsky is taken in HPSG.

> [GPSG] wasn't, as Uszkoreit and Peters proved. (The context-
> freeness of GKPS was bought at the cost of imposing a linguistically
> unmotivated prohibition on application of a given metarule more than
> once in a derivation.)

You seem to assume that the price to be paid for the restrictiveness was to
high. I flatly reject this conclusion. In any case, restricting metarules is
much more restrictive than a bag of completely unrestricted lexical rules,
inheritance hierarchies, features and types of dubious origin, let alone
list- and set-valued features.

> More specifically (in connection with HPSG vs. GPSG) it has been
> suggested that moving from trees to feature structures was a
> misstep. Maybe so, but it is a short step, if trees can have their
> nodes labelled by AVM's and some of the features of those AVM's have
> values (such as indices) that can show up at more than one node.

If I remember correctly, there was no question of type vs. token-identity
cropping up in GPSG.
There weren't even AVMs in GPSG.

Best

T.