Re: Functions of Classifiers (in a conlang)

> Hmm, that is a dilemma then.
>
> Maybe if the origin of CLS + VERB => NOUN constructions are relative
> clauses, we could allow a complementizer to intervene in those cases
> where the speaker feels it's needed?  So if the diachronic origin of
> the construction is in phrases like "CLS.MAN who walks", "CLS.STONE
> that speaks", "CLS.PLANT which burns", etc., maybe we can allow the
> who/that/which to show up when needed?
>
I'm not a fan of that solution since it not only scuppers the whole
noun/verb indeterminacy thing but places a heavy burden on a language
that was going to use nominalizations quite a lot.

>
> As a side question, in these constructions, what is it that the
> classifier modifies?  Is it a verb, a verb phrase, an inflectional
> phrase?  Are the following legal?
>
>     CLS.MAN walk quickly = "runner"
>     CLS.MAN not eat = "anorexic"
>     CLS.MAN not eat meat = "vegetarian"
>     CLS.MAN will marry = "fiance"
>
Yes to all. The classifier is intented to nominalized a complete clause
or series of clauses, with arguments, including serial verb
constructions. :) Because serial verbs can be nominalized, there will
still of course be ambiguity about where a nominalization ends, but it
seems to me that it's better to know if anything explicit is being
nominalized or not even if you do still need context. I'm beginning to
think the best solution is some form of marking to turn the classifiers
into independent pronouns, as we were talking about earlier..