Theiling Online    Sitemap    Conlang Mailing List HQ   

Re: What do you call this suffix?

From:Eric Christopherson <rakko@...>
Date:Sunday, April 13, 2008, 22:11
On Apr 13, 2008, at 6:09 AM, JR wrote:
> In Khafos, nouns take a certain suffix (-l/-el/-ol/) when preceded > by at > least one adjective. Ex: > > kevre jetse vato-l > big green house-? > > I don't recall seeing anything quite like this in a natlang. In > Tagalog and > Persian, for example, there are "linkers" or "ligatures" or > "ezafe", but as > I understand, those are clitics, they come *between* noun *phrases* > and > (certain) modifiers, and they're used recursively. Ex: > > Persian: > khaane=ye sabz=e bozorg > house=LK green=LK big > > In Khafos though, the marker is clearly a suffix to the noun (it > changes for > vowel harmony, and does not attach to anything else), and does not > come > between between the noun and modifier at all. The one modifier that > does > follow a noun, a relative clause, does not trigger it (unless it's > fronted > to adjectival position! [this structure is marked]). > > I've been calling this the "modified suffix" for lack of a better > term, > since it is used on nouns that are modified, but that sounds kinda > funny. Is > there another word I should be using? And did ANADEW? Or even a > conlang? > > Josh Roth
It reminds me of the construct state (status constructus) in Afro- Asiatic languages, although as I understand it, that always involves two NPs in a possessive/genitive relationship. Also, in those languages, the construct state morpheme is a suffix on the head, and the head comes before the dependent, so in effect the affix comes between two nouns. However, Ainu also has a construction where possessors (heads) are marked while their possessa (dependents) are not, and the heads come after the dependents, so at least in terms of word order, it resembles your example. The remaining difference is that the relationship is still a possessive one, not an adj+noun one. However... ...if you said that <kevre> and <jetse> were (historically at least) abstract *nouns*, you could have a pattern like some phrases in Semitic where e.g. you use "house of bigness" to mean "big house", i.e. <vato-l> would be the possessor of <kevre>.