🚀 go-pugleaf

In article <SCHILLER.92Jul21162632@tira.uchicago.edu> schiller@tira.uchicago.edu (Eric Schiller) writes:
>Oh, dear the Chomsky: A Minimalist Program for Linguistic Theory,
>the sanitized MIT Working Papers version (not the samizdat draft
>we all came to know and love during the winter of '92.

The "samizdat draft" was not in fact samizdat, but ukradizdat.  Ask a
Russian speaker for translation, if necessary. It was a genuine rough
draft, which somehow ended up getting circulated.  The legal version is
not at all sanitized, but is in fact longer and corrected.


>Well, let's see...
>	End of Projection Principle
>	End of DS and SS (as we know it...)
>	Many new notions, like Procrastination and Greed
>	X-bar is now fundamental
>
>I kid you not.

Accurate enough, as far as it goes, though I don't know why we might
think you are kidding.  However, in case there are sci.lang readers out
there who are actually interested in syntax, a few notes are in order.

The abandonment of D-structure is the consequence of Chomsky's accepting
Richard Larson's ideas about thematic role assignment, by which a verb
with multiple arguments is not necessarily in a structural position to
assign all its theta roles at the initial level of structure.  Since
D-structure was defined/described in LGB as a level that is (a) initial
and (b) a locus for all theta-role assignment, this notion must
consequently be abandoned.  Since the Projection Principle as stated by
Chomsky in LGB required the satisfaction of lexical properties at
D-structure (as well as S-structure and LF), this too must be abandoned.

However, all the empirical *effects* of the Projection Principle are
retained, and in fact something exceedingly close to the Projection
Principle is still around, since the satisfaction of lexical properties
is immediate once the requisite structure is built under Chomsky's new
system, and perseveres to LF, as in previous systems.  This makes the
following remark inaccurate:

>Seriously, I think this is a radical departure from GB and does make much
>of the work of the last ten years moot, especially all the mental
>acrobatics reconciling the facts of language with the now-discarded
>projection principle. 

What is so striking is how few *analyses* seem to be affected even if
Chomsky is on the right track, though these analyses might be couchable
in slightly different terms.  Furthermore, the spur for this corner of
Chomsky's work, Larson's shells, are rather undermotivated in the first
place.  Therefore, if one discards this, one should take Chomsky's
rethinking of D-structure and its consequences with a grain of salt.

>Those stil unconvinced of the radical nature of the paper consider this:
>Raising to Object is back!

Not really, except as a figure of speech.  In the old theory and the
new, accusative subjects of infinitives (in languages like English)
never occupy the base position of direct objects.  Furthermore, in the
old theory and the new, accusative subjects and accusative direct
objects get case in the same way.  The new theory differs from the old
in positing a distinct position -- related, it is claimed, to object
agreement -- in which accusative case is assigned, a position which is
neither the underlying object position nor the underlying
subject-of-infinitive position, but another position higher than V.

-David Pesetsky