3.2.4 Locality Restrictions on movement

The restrictions on movement we have mentioned so far have concerned its structure preserving nature. Not all structures which conform to these restrictions are grammatical, however, indicating that other restrictions are in operation. Most of the movements we have looked at have involved something moving within a single clause. Passivisation, for example, moves the object of a predicate to the subject of that predicate and obviously both the extraction site and the landing site are within the same clause. There are movements however, that move elements from one clause to another. Consider the following:

Given the near synonymy of these two sentences and the fact that the subject of seem in (105b) does not appear to be semantically related to this verb (Fiona is not the one who ‘seems’) we might assume that the latter is formed by a movement of the lower clause subject into the higher clause subject position:

This movement is known as raising as the subject of the lower clause raises to the subject of a higher clause.

Raising can apparently happen out of a number of clauses:

Thus, at first sight it would seem that movement is unrestricted in terms of how far an element can be moved. But on closer inspection this might not be an accurate description of what is going on here. For example, note that in (107b) and (c) all the clauses that the subject is raised out of are non-finite and none of them seem to have subjects.

Suppose we try to move out of a finite clause instead:

As we can see, a subject can be raised out of a non-finite clause into the subject position of a finite clause, but it cannot be raised out of a finite clause. Note that the finite clause in (108a) has a subject of its own: it. It is a fact about English finite clauses that they must have subjects and hence the sentence would be ungrammatical if the subject were missing for independent reasons. So this case differs from the grammatical movement in (107c) in two ways: the moved subject is moved out of a finite clause and it is moved out of a clause with a subject.

To control for these variables, let us consider a case where the movement is out of a non-finite clause with a subject:

Again the result is ungrammatical, demonstrating that movement over a subject is itself enough to cause an ungrammaticality. But why would moving over a subject be a problem? If long distance movements are possible, it is hard to understand why the presence or absence of a subject should make any difference at all. However, if we suppose that long distance movements are not possible, though an element can move a long way via a series of short movements, we can come to an understanding of these observations. Consider the grammatical case of (107). As each subject position is empty it is possible for the moved subject to move into each one in turn, moving from one clause to the next each time:

If there is a subject in one of these positions the moving subject would be forced to make a longer movement and if long movements are not allowed, we predict the result to be ungrammatical, which it is:

We call this phenomenon, the boundedness of movement. For now it is enough to note that movement is bounded. We will look in more detail at the phenomenon in a subsequent chapter.