Matt: > > BTW, if some clitics don't trigger stress shift, why write them as > > part of the preceding word at all, rather than as separate orthographic > > words? > > I guess because (a) native speakers [i.e., me] feel that enclitics form a > tight unit with the preceding word, and (b) encliticisation triggers > certain sound changes which are otherwise confined to word-level domains. > For example, adding a possessive enclitic to a noun ending in a nasal > triggers nasal assimilation: > > konom "hammer" [ko.'nom] > kononko "your hammer" [ko.'noN.ko] > > Also, failure to add an enclitic causes the allative suffix "-ni" to lower > and become "-ne" word-finally: > > totsat "table" [to.'tsat] > totsatne "towards a table" [to.'tsa?.nE] > totsatnima "towards my table" [to.'tsa?.ni.ma] > > (Note that the enclitics which *fail* to trigger these kind of changes are > the ones that I'm separating from the preceding word by an apostrophe. I > guess I need to posit level 1 clitics and level 2 clitics, where the latter > are more tightly bound to their host that the former.) If reason (b) applies only to enclitics not written with the apostrophe, then clitics that are written with the apostrophe could instead be written as separate words, no? If you were not so much guided by your personal aesthetics as deriving a practical roman orthography for Tokana, what would your strategy be? --And.