[YG Conlang Archives] > [ceqli group] > messages [Date Index] [Thread Index] >
>>> rmay@hidden.email 03/03/02 09:42pm >>> #on 3/3/02 9:47 AM, Rob Speer at rob@hidden.email wrote: #> On Sun, Mar 03, 2002 at 07:59:05AM -0700, Rex May - Baloo wrote: #>> Now, do we want to permit the voiced couterparts of the 's' series? zg, #>> zgr, zd, zb, none of them or some of them? #> #> I say certainly. There's no reason these shouldn't be pronounceable if #> the unvoiced ones are. # #I just got back from a trip to Borders' to do research on Italian, and, #it seems, all those clusters are allowable in Italian, provided that #sb- sd- sg- are really pronounced zb, zd, zg, etc. I just emailed Claudio #Gnoli and invited him to join us as an expert on Italian consonant clusters. They are indeed pronounced with [z]. I find them very difficult to say -- it's the feature of Italian I find hardest to pronounce (with long consonants being the feature I find hardest to hear). Rob is incorrect to suggest that the voiced counterpart of any voicless cluster is as easy to say; voiced syllable- initial clusters are considerably rarer than voiceless, cross-linguistically, though I gather that any sort of syllable initial cluster is rare cross- linguistically. Maybe at this juncture I should make a point I was planning to make later when I'd got to the end of the current thread: I seem to recall Rex saying that "schwa buffering" occurs between consonants. Translating that Loglan notion into naturalistic terms, that means that consonant clusters don't exist underlyingly. The cluster rules can simply specify the conditions under which the schwa can be unrealized, as is the case with Livagian. Since the cluster rules are then concerned with relatively trivial details of surface realization, they can be left to a future stage of development. Or -- better -- they can be done away with altogether, allowing individual speakers to delete the schwa wherever and whenever they wish. --And.