In
222 and in most classes on transformational syntax if not all,
it is assumed that the subject of the sentence is defined as:
The NP that is immediately dominated by S.
This definition assumes that S does not immediately dominate any other node, an assumption which we will claim is correct. In some sense the above definition may be true, but it misses something. One problem that arises is what is S and what is its head? 'S' as a node was abandonned in the early 80s. It was replaced with INFL (inflection) or TP (tense phrase). The former we do not adopt here, the latter we do, albeit in a slightly modified form.
What we will assume here is that the subject is a head-position (S), and that its complement is TENSE (or VP, whose head is V and must be modified by all required operators including TENSE). As in all other cases a head plus its complement forms a phrase. Hence:
SP --> S Comp-S.
While all other nodes seem to have some connection to conceptual forms in one way or another, S does not. S is a grammatical form that seems to occur in all languages.
What function does
S have? Although it is not clear and is certainly controversial,
the subject seems to bring a node into prominence. This is the
point of view that we shall adopt here. Furthermore, most languages
seem to have a restriction that one NP must be prominent in the
active voice. In some languages (Polish, Ukrainian, Lesotho, e,g,)
there is a passive construction where there is no prominent NP.
English is not one of them.
Is
the subject position filled at A-structure (argument structure)?
No, since the subject has no connection to conceptual structure
which feeds A-structure. Initially, we will assume that one NP
must raise to S, in which case the NP is prominent.
There is also another factor associated with the subject. We will claim that C (the head of mood, otherwise called a complementizer) assigns Case to the subject NP. This may be a more important feature of grammar, though there are some problems with this. We will not cover them here.
We
have not defended or explained why we assume the node S as subject.
We will do so in sections that follow which deal with raising
to S.
![]()