Abstract
In this article we examine a Welsh adjectival construction which superficially looks simple but on closer examination proves to be somewhat challenging. The construction contains an NP constituent whose GF status is far from clear. We consider various analyses of this NP, as SUBJ, OBJ, and ADJ, and suggest that on balance the evidence favours the OBJ analysis. Beyond the purely parochial Welsh or Celtic interest, it may provide a useful case study of how difficult it is to determine the correct identification of grammatical functions beyond core cases.