On some non-evidence for the cycle in syntax

The principle of cyclic application of rules has been important since Chomsky's Aspects. However, this principle has been called into question; and Grinder 1972 has argued that many examples thought to support the cycle do not, in fact, do so. Grinder discusses cases where rules must re-apply in certain derivations (e.g., Passive must apply, then Raising, then Passive again in the derivation of Alice was believed by us to have been kidnapped by Hector ), and he argues that these cases can be explained without the use of a cycle. However, a new class of arguments has recently appeared, based on a theory without extrinsic ordering (cf. Koutsoudas 1973a): it has been claimed that the cycle must be used to explain why, in some derivations, a given rule must apply before another rule. I show that these arguments do not support the cycle, for in each instance some principle of grammar other than the cycle can explain the precedence of application of one of the rules.