Syntax Lecture 4 Specifiers The specifier The specifier

Syntax Lecture 4: Specifiers

The specifier • The specifier is a phrase • It appears as the first element in the phrase that contains it • There is only one specifier • It is not restricted by the head – i. e. different heads do not select different specifiers

The specifiers of thematic heads • Thematic heads usually do not appear with specifiers – [PP -- on [DP the table]] – [AP -- fond [PP of chocolate]] – [NP -- students [PP of linguistics]] – [VP -- go [PP to London]] • This is something that needs an explanation

Specifier of the VP • One interesting potential candidate for the specifier of the VP is all: – The contestants will [all throw the dice] • This has some of the properties we expect: – It must come first • * the contestants will [throw the dice all] – There can only be one of them: • the contestants will [all/both throw the dice] • But it doesn’t appear to be a phrase

Floating Quantifiers • The really interesting thing about these elements is that they do appear to be associated with a phrase – The DP in the subject position: • The audience will all leave • In fact, this quantifier can appear as part of the subject with no difference in meaning – All the audience will leave = the audience will all leave

Floating Quantifiers • One suggestion is that the two sentences are related and a formed by movement:

Conclusion • If this analysis can be maintained – The floating quantifier is part of a phrase – This phrase starts in the specifier of the VP – The phrase, or part of it, moves to the specifier of IP (the subject position) – Nothing is in the subject position before the movement – General conclusion: things move!

Things Move • Once one starts to look for things that move, we can see them everywhere: – In questions: • You will have a cup of tea will you have a cup of tea? • You will meet him who will you meet? – Objects: • I don’t like John, I don’t like • I know John was known • I gave him the money I gave the money to him – PPs • They met in the park, they met – VPs • I thought he might run away and run away, he did – APs • Though he is tall, he still can’t reach tall though he is, he still can’t reach – Particles • He took down the jar he took the jar down – Negatives • I saw nobody did I see • So movement is an extremely common grammatical device

Subjects inside VPs • Sometimes we can actually find a subject inside a VP – I made [John wash the dishes] • The bracketed part of this sentence has most of the components of a sentence, except for an inflection – A subject: John – A predicate: wash the dishes – No inflection • * I made [John will wash the dishes] • * I made [John washed the dishes] • * I made [John to wash the dishes] • So this does not look to be an IP • It seems to be a VP with a subject

Subjects inside VPs • But if there is a subject position in the VP and one in the IP, why do we not find more clauses with two subjects? – * John might [VP Mary win the race]? • Given that we have evidence that subjects move out of VP into IP, it seems that all subjects originate in the VP and move to IP – In this way there can only be one subject – But we will not usually see it inside VP because it moves

The underlying emptiness of the IP specifier • That the IP specifier is underlyingly empty is supported by a number of observations • First it is not only subjects which can end up there – In passive constructions, it is the object which moves to the subject position: • --- was killed John was killed --- – In other constructions it seems that the subject of another clause can move to a subject position • It seems [John is dead] • John seems [ --- to be dead] – In some cases the subject is filled by a meaningless element (it) • It seems [John is dead] this indicates that the subject position in the IP does not have to be associated with any meaningful element, which is explained if it is underlyingly empty: it gets associated with meaningful elements which move into it

The specifier of the NP and the DP • Last week we hinted at the connection between the IP and the DP • If this is more than just coincidence we might expect conditions on specifiers to be similar too • Specifically – The specifier of the DP is underlyingly empty – It can be filled by things moving to it from within the NP

What goes in the specifier of DP? • Specifiers are phrases which come first in the phrases that contain them – A phrase which comes at the beginning of the DP is the possessor: – [DP the artist’s] picture of Bill]

Evidence for the emptiness of the DP specifier • First, DP specifiers can be empty – If there is no possessor, there is no specifier • [DP --- the [NP horse]] • [DP [John’s] – [NP horse]] • The possessor can be the object of the noun – [ [Picasso’s] [painting of John]] – [ [John’s] [painting (by Picasso)]] • This is very similar to passive in the sentence • It suggest that things become possessors by moving to that position

Evidence for NP internal possessors • We have mentioned that DPs and IPs have a lot in common, so we might expect that if subjects start inside VP, possessors will start inside NPs • The VP internal subject will move to the empty specifier of IP and the NP internal possessor will move to the empty specifier of DP

Evidence for NP internal possessors • Unfortunately, we have no direct evidence from English that the possessor originates inside NP – They always seem to have to move • However, we do have direct evidence from other languages that a similar process happens to their possessors

Possessors in Hungarian • There are two types of Hungarian possessor – Az én rajzom – Nekem a rajzom • Not only are the two possessor in a different case (nominative in the first and dative in the second), but they stand in different places with respect to the determiner – The nominative possessor is after the determiner – The dative possessor is before the possessor • This suggests that the nominative possessor is in the specifier of the NP while the dative possessor is in the specifier of DP

Possessors in Hungarian • Nominative possessor • Dative possessor

Conclusions • The specifiers of thematic heads seem to be elements semantically related to those heads – Subjects or possessors • The specifier of functional heads seem to be underlyingly empty • The specifier of functional heads seem to act as landing sites for moved elements