Theiling Online    Sitemap    Conlang Mailing List HQ   

Noimi Parts of Speech?

From:Jeffrey Jones <jsjonesmiami@...>
Date:Sunday, July 3, 2005, 19:56
I don't remember where the parts-of-speech thread was, but this may be

The conlang project I've been working on the most lately uses an animacy
hierarchy, with inverse marking. The genders distinguished are "animate",
"inanimate", and a 3rd gender for situations. Note that these are terms of
convenience; I'm using them without any philosophical or religious intent.
The basic hierarchy is:

        animate > inanimate > situation

There are different ways of classifying the words, the ways being not
completely dependent on each other.

The highest level of distinction is between particles and inflected words.
Variations of the same inflectional system are used for all inflected
words, so there's no firm distinction between verbal morphology and nominal

Another way of classifying words is by their syntax. The inflected words
are broken down into determiners, quantifiers, and everything else. The
particles can also be subdivided.

The remaining classifications involve the inflectional system: stem-
formation distinctions and actant marking (if this is the right term)

The stem-formation distinctions have to do with which aspects and tenses
apply to a given class and how they're marked. The "nominoids" (normally)
use only the tenseless aspect. The "verboids" are divided into actionals
and statals. Those statal verboids that can be used as criteria for
comparisons are called "adjectoids".

Now for the interesting part: actant marking.

First, words are classified according to the number of arguments the word
has (its _valence_), which is either 1, 2, or 3. BTW, I'm using the
following sets of terms:

        V=1 (intransitive), V=2 (transitive), V=3 (ditransitive)
        All words have an _A1_ argument. V=2 add an _A2_ argument, and V=3
        add _A3_.

There's a set of actant markers for each argument, so, in some instances, a
word-form can be a complete clause.

The 3 valence classes are subdivided according to what genders are possible
for each of the arguments. Here's an incomplete set of rules:

 *  The most general one involves the animacy hierarchy: A2 >= A3 >= A1
    (if V=2, this is A2 >= A1, of course).
 *  Only A1 can refer to a situation.
 *  If V=3, A1 can't be animate, and A2 must be animate.
 *  A nominoid can't be V=3, and its A1 will normally be a specific gender
    (although this can be overridden).

I suspect this is another incomprehensible post.