Martha McGinnis: Non-Lexical-semantic PS building (reply to Mark Volpe)

[Martha McGinnis]

Dear Mark,

I'll have a go at this.  I think the issue you're raising isn't
special to DM; it's an issue that also arises in Lexicalist theories.
For example, consider GB, where a clause structure is projected from
the lexical entry of the main predicate (say, a verb), in accordance
with X-bar Theory and the Extended Projection Principle.  How does GB
ensure that an unaccusative verb is ever selected from the Lexicon?
Recall that in GB, semantic interpretation doesn't take place until
LF, so there's presumably no conceptual/intentional influence on the
choice of items from the Lexicon. The same issue arises in the other
theories I'm familiar with -- Minimalism, LFG, RG, Categorial
Grammar, G/HPSG. I'm not familiar enough with Beard's LMBM to know
for sure whether it addresses this issue differently, but I'd be
surprised.  The usual response to this issue is that linguistic
theories are not theories of production; the production system is a
separate component that makes use of the language faculty in ways
that we don't yet understand.  No doubt Jerry Fodor has talked about
this at some point -- maybe someone else on the list can suggest a
source.

I've actually never seen any story about lexical selection that makes
more sense to me than your "want of variety" story.  It seems clear
that the issue can't be resolved by introspection alone. You might be
interested in looking at W. Levelt's book _Speaking_ -- his model of
speech production involves funnelling conceptual intentions into the
syntax via the Lexicon. It's not quite clear to me how that should
work, or what role LF should play, but at least his claims are fairly
explicit.

There's a separate issue about the 'selection' of unaccusative verbs
that I believe *is* special to DM.  Namely, how does DM ensure that
the Vocabulary item corresponding to the root of an unaccusative verb
(say, 'arrive') is inserted into a tree that has unaccusative syntax?
If, as DM assumes, all roots compete for insertion, then this can't
be the result of a contextual restriction on the item 'arrive' that
requires it to be inserted in an unaccusative structure. A
contextually restricted/specified item for 'arrive' would then always
block the insertion of the less restricted/specified item 'grow',
which is only optionally unaccusative.  Instead, I believe, DM holds
that any root Vocabulary item can be inserted into any root node, and
if the result makes sense to the conceptual system, it converges;
otherwise, it crashes.  So 'I arrived the plane' is no good because
the Encyclopedic content of 'arrive' is incompatible with a
transitive structure (see Harley & Noyer's MITWPL 32 paper for more
details).

Hope this helps.

Regards,
Martha



>Dear DM-listers,
>   I've been interested in how specific syntactic
>structures are derived in a post-syntax lexical
>Vocabulary Insertion framework; that is, where the
>former-Projection Principle assumed that the lexical
>semantics of the Verb, unergative vs unaccusative,
>etc., would determine the phrase sructure of the
>syntax, other frameworks, e.g., LMBM, which have
>Lexicons, allow for early selection, though insertion
>and closed-class realization would presumably be
>post-syntactic (and this has its own pitfalls
>requiring a story). How does DM insure that
>Unaccusative structures are built to accompany
>Unaccusative Verbs despite post-syntactic Vocab.
>Insertion? It seems to me that the building of an
>Unaccusative syntax pre-selection would be
>unmotivated, except perhaps by the language module's
>want of variety, "Well, the last 3 structures we built
>had Agents in Subject position, let's do an
>Unaccusative for the hell of it".
>   Needless to say, I'm seriously missing something
>here. Could you tell me how late lexical selection
>works?
>                   Thanks in advance,
>                   Mark Volpe, SUNY-Stony Brook
>
>__________________________________________________
>Do You Yahoo!?
>Yahoo! Auctions - buy the things you want at great prices
>http://auctions.yahoo.com/


mcginnis at ucalgary.ca