One standard assumption is that the subject of a clause originates from the specifier of its predicate, following the predicate internal subject hypothesis. For example, the subject of a simple transitive sentence would originate from [Spec, VP] (or [Spec, vP]) and the subject of a copular sentence with an adjectival predicate would originate from [Spec, AP].

One question arises when DP predicates are involved. In a sentence such as "this is Mary's watch", the specifier of the DP is occupied by "Mary".

enter image description here

The subject is, therefore, unable to be base-generated in [Spec, DP]. Where is it base-generated then? How does the hypothesis account for this?

  • 1
    It's pretty standard in generative syntax these days to analyse copula constructions as involving raising out of a small clause structure (PredP) headed by a null functional head. In other words, "This" originates in specPredP, and "Mary's watch" is in the complement position of pred, then "This" raises to SpecIP to fulfill the EPP. Line Mikkelsen discusses this analysis in her book on copulas. Not sure if this is the kind of thing you're looking for.
    – P Elliott
    May 29, 2016 at 16:32
  • @PElliott That sounds cool and all, and it is what I'm looking for, but wouldn't assuming functional projections that have no semantic impact go against minimalism? PredP would only exist for the sake of facilitating structural description and perhaps assignment of certain features. I guess that's the reason why AgrPs were dispensed with in the MP. Jun 1, 2016 at 11:37

3 Answers 3


In your example, Mary's watch is not a DP 'predicate', but the second argument of an 'identificational' verb be that expresses identity between its two arguments, not class-membership. Since 'identificational' be is a two-place predicate - not a copula linking a monadic nominal predicate to its subject - its two arguments can be initially projected, under X-bar assumptions, in its 'complement' and 'specifier' position, respectively, which means that the 'XP-internal subject hypothesis' can account for cases like this without the slightest difficulty. You should, instead, have used an example like John is an engineer, where is does express class-membership and an engineer is, indeed, a nominal one-place 'predicate', but note that, in that case, the specifier of D (= a) is empty, and, again, there is no problem whatsoever in initially projecting the eventual subject John in Spec D in case you should want to adopt that analysis. (You need not, though, because there are other, and arguably more attractive, possibilities, such as the 'predication hypothesis' that P Elliott refers to).


The two solutions that I can see are:
A) That such complex DPs have another phrasal constituent which contains them and has a null head. Though this seems a little far-fetched. Or...
B) That the subject originating in spec-AP only applies when it is indeed an AP and not to existentials with a DP. Thus it would presumably be generated in spec-VP.

Apologies for a fairly weak answer but I thought I should contribute as I can.


The subject of the sentence is This in IP,spec. In the DP Mary's watch, Mary specifies the whole DP, ergo she's kinda the subject of this complex DP.

  • 2
    This is missing the point of the question.
    – P Elliott
    May 28, 2016 at 14:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.