On the role of segmental contrasts in the acquisition of clusters

Abstract

The acquisition of an underlying contrast between /l/ and /r/ has been claimed to be a necessary prerequisite to the acquisition of clusters (Archibald 1998). To evaluate this claim, an archival database including more than a hundred children with phonological delays, ages 3;0 to 8;6, was consulted. A number of apparent counterexamples were identified. All problematic cases reliably produced consonant + /l/ clusters but lacked an underlying contrast between /l/ and /r/. In an effort to reconcile these (apparent) counterexamples with the many compliant cases, these data were further reanalyzed within optimality theory (McCarthy and Prince 1995). The analyses revealed that the apparent clusters were more properly understood as complex segments similar to affricates. Thus, while such cases do not contradict Archibald's proposal, they do provide a richer account of the development of clusters. The paper concludes with a discussion of the implications of OT accounts for the learnability of structure and for clinical treatment.