Is there really movement?
There are a number of problems with the idea that things are moving around in the grammar. First of all, it seems strange. It is not easy to believe that our language ability involves generating syntactic transformations. Cognitive Linguists, who do not believe in Universal Grammar and the Language Faculty, claim that their numbers are growing because the Chomskyan approach to Generative Grammar is simply not believable.
Subcategorization
One problem is that movement solutions often seem to be deliberately ignoring simple facts related to the lexicon. For example take the example offered earlier:
1. Taro seems to like Hanako.
2. Taro/i seems [t/i to like Hanako]
A movement explanation is going to suggest that the NP/DP Taro moves from the subject position of like.
So let’s say we have the subject Taro starting off in the [-tense] lower clause as the subject of like. Then it moves, in order to be in a tensed clause, for example. That seems alright. However, this kind of sentence is only grammatical for a limited class of verbs. For example, you can’t say 2 below:
2. *Taro thinks to like Hanako.
Obviously, a verb such as thinks takes (subcategorizes for) a finite clause. So 3 is fine.
3. Taro thinks Ichiro loves Hanako.
The thing about a verb such as seem is that it takes a non-finite [-tense] complement with the subject of seem and the subject of the complement verb matched.
The problem is, if we need subcategorization information in the lexicon, why do you ALSO need movement? Look at the following simple subcategorization information.
The point is that we have to have subcategorization and it is quite easy to encode a simple matching of different sorts. Should we refuse to allow this matching information just so we need movement?
Anyway, check out the PowerPoint slides here.
Goldberg, A. E. (2011). General Introduction. In Cognitive Linguistics. London & New York: Routledge.