The Comp-trace effect is standardly taken to describe the outcome of certain movement operations that displace subjects across overtly headed clause boundaries (i.e., complementizers). In this light, the ungrammaticality of the sentence *'Who do you think that t left?', for example, is blamed on the fact that long subject extraction across a complementizer is somehow syntactically illicit, whereas the output 'Who do you think t left?' is not deficient in this way given the absence of a complementizer. For over thirty years, the phenomenon has played a pivotal role in the development of various avenues of the theory of syntax. This article advances a theory of Comp-trace phenomena that reduces the effect to one of prosodic ill-formedness at PF rather than to syntactic impropriety, however formulated. Drawing on novel data from English and Nupe, a Benue-Congo language of central Nigeria, the article motivates the somewhat bleak, yet interesting hypothesis that the Comp-trace effect does not represent a homogeneous linguistic phenomenon, nor is a grand unified analysis of the phenomenon available.