The root of ruthless

Linguistic variation between individuals must be linked to how linguistic material is mentally represented. Therefore, by examining individual variation, light can be shed on the nature of mental representation itself. This paper presents an individual differences study of semi-opaque derivations (e.g. ruthless) to establish whether their representations are mentally associated to those of fully segmentable forms with the same suffix (e.g. sightless). This way, a prediction of connectionist and exemplar models of morphology is tested, namely that to language users semi-opaque forms are likely to retain some degree of internal complexity, despite the fact that they are not segmentable. Using corpus data, it is demonstrated that individuals who rely more heavily on the segmentable forms are also more likely to use the semi-opaque forms. This pattern in the variation across individuals indicates that semi-opaque derivations are not represented independently of the derivational paradigm from which they historically derive.