Don't Move!

All versions of Transformational Grammar assume that movement is a central feature of the syntax of human languages. However, frameworks which make no use of movement processes have existed for thirty years, and there has been very little attempt to show that movement analyses are superior to the analyses proposed within these frameworks. The strongest evidence for movement comes from filler-gap dependencies, where there is an extra clause-initial constituent of some kind and a gap somewhere later in the clause. Wh-questions are a typical example. The assumption that the filler has moved from the position of the gap accounts for the appearance of both the filler and the gap. However, consideration of a broader range of data casts doubt on the movement approach. There are (i) cases which look like filler-gap dependencies where there is no visible filler, (ii) cases with two gaps, (iii) cases where filler and gap do not match, and (iv) cases in various languages which look like filler-gap dependencies but where there is not a gap but a resumptive pronoun (RP). The alternative to movement that has been developed within Head-driven Phrase Structure Grammar involves the feature SLASH, which makes certain kinds of information available higher and lower in the structure than would normally be the case. There is no reason (i) why this information should always be associated with a filler, (ii) why it should not be associated with more than one gap, (iii) why it should not be associated with a gap with rather different properties, and (iv) why it should not be associated with an RP. For all these reasons, it seems that the SLASH-based approach is superior to a movement approach.