Talk:Specifier

From Wikipedia, the free encyclopedia

Two comments: First, specifiers only precede heads in certain languages. A specifier is in "phrase-lead" position, but the phrase-lead may be after the head, or whatever. Second, the examples "the, no, some, every, John's and my mother's" are all what I would call heads of determiner phrases rather than specifiers to noun phrases. Maybe the model I'm operating off of is different from the author(s) of this article. Honestly, all I know about this is what I've learned in Dr. Alan Manning's Theoretical Syntax class, and I'm not quite sure how the theories he subscribes to relate to other theories out there.

Anyway, using the notation of <specifier> head (complement) {modifier} [sub-clause] t == head trace <o> == operator trace, here would be my example of a specifier:

Spec of an Inflection Phrase (IP-spec): <Those folks> have ( eaten (the food) {already} )
Spec of a Verb Phrase: <Those folks> have (<all> eaten (the food) {already} )
Spec of a Complementizer Phrase: <Why> did [ <you> t eat (the food) {already} <o> ] ?

Any comments on how this fits in with where the article currently stands? --RockRockOn 15:58, 14 April 2006 (UTC)

I've fixed the problem about the order of specifiers (to my knowledge, no-one has ever defined specifiers in terms of linear order, it's always been a structurally defined position).

Regarding the, some, etc., these were originally assumed to be specifiers of NP. More recently, the hypothesis that they are heads has become very popular. Maybe it would be better to stick to less controversial examples like [Spec,IP]. On the other hand, the reasons for analyzing subjects as specifiers are pretty theory internal -- in the original X-bar theory, S (i.e. IP) was not an X-bar phrase and subjects weren't specifiers. The same sort of problem arises for [Spec,VP], since the evidence that this position actually exists is not easy to summarize in an article like this. In some ways, determiners are more "accessible" examples of specifiers, even if they're not analyzed as specifiers these days. Cadr 14:54, 23 April 2007 (UTC)

Perhaps the article ought to explain both usages? And, why not enter a discussion on the existence of Spec-VP? This seems like exactly the place for that. I don't know how it's explained in more generally-available iterations of the theory, but I could summarize my professor's rationale for hypothesizing such a position: he relies on a "double reduction" as follows: for the IP 'The chicken had pecked the corn' we can have a structural reduction to only an inflection plus a specifier: 'The chicken had', or we can have a picture reduction to the verb with the same specifier: 'The chicken pecking the corn'. We can also discuss theory about the function of verbs in originating theta roles versus the function of inflections in nominative case assignment, VIS-traces, etc. --RockRockOn 07:47, 25 April 2007 (UTC)

I don't follow your professor's argument from "The chicken had". This would just be an IP lacking a VP complement whether or not [Spec,VP] contained a subject. In fact, in this case, the VP-internal subject hypothesis would force you to assume that the subject "the chicken" originally raised from [Spec,VP] to [Spec,IP] with subsequent VP deletion. In the case of "the chicken pecking the corn", this is not a VP but a reduced relative clause in most contexts (e.g. "The chicken pecking the corn looked up at Bill"). Is your prof. referring to constructions like "I saw [the chicken pecking the corn]", where the bracketed phrase seems to describe an event?

As you say, the standard conceptual motivation for postulating VP-internal subjects is theta-theoretic. However, theta-theory and X-bar theory and historically and conceptually separate. X-bar theory originated ~10 years earlier. For this reason, I don't think it's appropriate to bring too much discussion of theta-theory into the article (especially since theta-theory is a controversial area in its own right).

One of the (now) standard bits of evidence for VP-internal subjects is the possibility of "stranding" quantifiers like "all" in this position. For example:

(1) All the boys went to school

(2) The boys all went to school

In (2), all seems to have been left behind when "all the boys" raised from [Spec,VP] to [Spec,IP].

Anyway, after all that blathering on, I agree I was wrong to say that we should exclude the VP-internal subject hypothesis from the article, since it's now very widely assumed. Maybe we could say "...it is now commonly assumed that..." and then give the floating quantifier evidence, which has the advantage of not requiring any additional theoretical discussion. Cadr 15:59, 25 April 2007 (UTC)

Oops, I just noticed that you had a floating quantifier in one of your own examples. Cadr 17:37, 25 April 2007 (UTC)