On Sat, 18 Mar 2006 01:16:46 +0000, Matthew Barnett <[log in to unmask]> wrote: >Rex May wrote: >> I know of two methods of self-segregating morphology. The Loglan/Lojban >> method, where morphemes are in the form (C)V(V) or CVCCV/CCVCV, and my >> Tceqli method, where everything is in the form nCnV, a method also used in >> Guaspi. >> >Are they really different? They both specify that morphemes have a fixed >'shape'. Well, they're variants of the shape rule, but Loglan is really fuzzy about it here and there, depending on pauses sometimes. There could be other methods — one might be to just end (or begin) all morphemes with a given sound that doesn't occur elsewhere. But that would seem very unnatural. I guess that's my question. What other ways can it be done? > >> My question is, what other possible methods are there? And what advantages >> and disadvantages do they have? This is all to the end of answering the >> big question - is SSM worth it? >> >When you're compounding or inflecting, SSM makes analysis easier. > >But how many natlangs have SSM? Mandarin pretty much does, a functions of virtually all morphemes being monosyllabic. I thought of that for Tceqli, and may take that direction in another project, but that really makes borrowing a challenge, and does severely cut back on the number of possible words. Let's see, suppose you have five vowels and three diphthongs (Call them all 'V'). And twenty consonants. You then have 8 morphemes of the shape V, 160 of the shape CV, And if you can end with a consonant, you have3200 CVS's. Is that math right? If so, monosyllabic roots are just fine.