ABSTRACT
This chapter traces the main issues that have shaped the discussion about coordination in the last three decades. It explores whether coordinations should be generated directly, by means of some sort of PS-rule, or whether they should be derived from coordinated simplex sentences by means of some sort of deletion rule. The chapter addresses the transformations vs. PS-rule problem in Chomsky. Coordination is always coordination of constituents in Chomsky. Postulating such a rule leaves us with two competing hypotheses, both of which are equally adequate from a descriptive point of view. The grammatical models have become vastly more sophisticated, but the problems remain the same: nontransformational accounts cannot deal with obvious non-constituent coordinations which result from e.g. Gapping, while transformational accounts still have trouble with NP-coordinations in combination with synmmetrical predicates like be similar, be a nice pair.
