In a phrase like CEFFYL FFERM, is FFERM thought of as genitival or attributive (in the abstract, or where context does not force the issue)? This might be a silly (i.e. intrinsically unanswerable) question, were it not for the possibility -- at least in principle -- of distinguishing (an X's) Y from an (X's Y) where Y is feminine singular and X begins with a mutatable consonant. I'm thinking of pairs like these: BISGEDEN GI 'a dog biscuit' -- a biscuit of the kind manufactured for dogs vs BISGEDEN CI 'a dog's biscuit' -- a biscuit of whatever kind belonging to an unspecified dog (e.g. a custard cream found in a kennel) or HET BLISMON -- a hat of the kind worn by policemen vs HET PLISMON -- a hat of whatever kind in the possession of some policeman or other (e.g. a trilby found in a police- station locker room) I take it that the first members of these pairs (so far as they are possible), but not the second, could appear with the article: Y FISGEDEN GI, etc. But it may be that there are restrictions on the attributive use of nouns that rule them out anyway. Is this possibility of differentiation exploited at all systematically? If it isn't, what happens in practice? If it is, is there then an answer to my question about CEFFYL FFERM? (The question might perhaps be reformulated as: would it be more natural to think of CEFFYL FFERM as the indefinite counterpart of Y CEFFYL FFERM, or of CEFFYL Y FFERM?) And where do proper compounds come in here? My impression is that proper-compounding is rather high-flown, and therefore not altogether productive. What would the native speaker make of coinages like CIFISGEDEN or FFERMGEFFYL? ************************ Nigel Love Linguistics Cape Town NLOVE@BEATTIE.UCT.AC.ZA ************************