Re: THEORY: Gender in verbs

On Sun, 30 Sep 2007 14:57:22 -0400, John Vertical
<johnvertical@...> wrote:
>So, everyone here knoes that nouns sometimes have gender, ie. come in
>classes, according to their morphology. Most, if not all of the time, this
>also affects other words in the noun phrase, so there are different forms of
>adjectivs or pronouns or articles etc. (autc? :), that are used together
>with a certain class of nouns. Sometimes, in Bantu frex, the verb phrase or
>parts of it may also be affected.
Actually I believe in order to be referred to as "gender" the noun-classes
_must_ be concordial, that is, sometimes inspire agreement in other words.

>Most likely, everyone has also come across verbs having different inflection
>classes, aka. conjugations. However, AFAIK they generally do not affect
>anything else?
I can't remember hearing of such a thing.
Except things like, in some split-ergative languages, past-tense clauses are
ergative and future-tense clauses are accusative, or perfective clauses are
ergative and imperfective clauses are accusative, or realis clauses are ergative
and irrealis clauses are negative; so one could say the case of the
noun "agreed" with the tense or aspect or mood of the verb; but this is
not "agreement" or "concord" in the sense usually meant.
Likewise voice, or applicatives, could almost as easily be thought of as the
noun's case "agreeing" with the voice of the verb as vice-versa; but either
way this still wouldn't be what is usually meant by "concord" or "agreement".
There are languages where one form of verb is used when all participants are
third-person and a suppletively different form is used when all participants are
speech-act-participants; and some languages where one form of verb is used
when all participants are singular and suppletively different form is used whin
all participants are non-singular; and some languages where one form of verb
is used when all participants are animate and a suppletively different form is
uses when all participants are inanimate.  I suppose this could be thought of
as the person or number or gender of one-or-more of the
participants "agreeing" with the verb; but that's also backwards from what's
usually meant by "agreement".

>But if in some language they did affect, say, the forms of
>the adverbs, or auxiliaries, or objects, would you think it'd make sense to
>say that verbs have gender in such a language?
What you are looking for, then, are "concordial verb-classes".  You want a
psychologically or lexically "inherent" property of the verb, not necessarily
marked on the verb itself, which requires "concord" or "agreement" with (an)
other word(s) in the clause.

Verbal nouns such as infinitives and gerunds do affect the case of their
participants; frequently one or more such participants have to be in the
genitive or possessive case when the verb is an infinitive or gerund or other
verbal noun.  Again that's probably not what you meant.

>What other parts-of-speech
>might plausibly agree with verbal gender?
The most obvious one is adverbs.  The prototypical examples of concord are
adjectives "agreeing" with their head noun(s) in gender and/or number and/or
case.

If there are "pro-verbs" (proforms for verbs), they might agree as well;
pronouns frequently "agree" with the gender of their referent.

Some semantic classes of verbs make requirements (some of which are
sometimes called "subcategorization") on one or another or more of their
participants; some of these "subcategories" could be called "genders", or at
least, "noun-classes".  We might think of this as concord; then the concordial
verb-class would be seen to require a particular noun-class as one of its
participants.

Some verbs require certain participants to be certain cases; sometimes these
cases are "quirky".  Also some of them sometimes have obligatory oblique
arguments; these might have to take the form of adpositional phrases.  So it
might be said that noun-cases and adpositions "agree" with the verb.

>One phenomenon that comes close would be split-S languages, but there the
>forms are AFAIK generally still chosen on the basis of the semantics of the
>verb; while in nouns, gender generally has a good splash of arbitraryness to
>it.
In concordial-noun-class systems, often one class is the "default" class to
which any noun belongs because it does not belong to one of the others.  In
your concordial-verb-class system you'd likely also want such a verb-class.
In concordial-noun-class systems, usually most of the noun-classes have
a "semantic core", one to which nouns belong because of their meaning.  But
usually most of the noun-classes have many members that belong because of
their morphology; because of the way they sound or because of the way they
decline.  Often one or a few of the classes have _only_ such members; these
are morpologically-based genders without a "semantic core".
In your concordial-verb-class system, there'd be no harm in having all-but-one
of your verb-classes have "semantic centers" that verbs belong to because of
their meaning; and have other verbs also belong to them not because of their
meaning, but because of how they sound or because of the way they
conjugate.  Indeed, for perfect parallelism with concordial _noun_-classes,
you'd want most of your verb-classes to have such "semantic centers".

>Is anyone able to provide better ANADEW, or AFMCL of any kind?
Not me (unless quirky-case verbs and/or oblique-argument verbs qualify).

Quirky-case-governing etc. verbs have a quirky membership.  There are
semantic tendencies making it likelier that a verb will be one of the quirky
ones, but nothing is really hard-and-fast.

>I may be overlooking something obvious; smack me upside the head if I am. :)
If you are, my favorite candidate for what you are overlooking would be
the "quirky-subject"-governing verbs, the "quirky-object"-governing verbs, and
the "mandatory oblique argument"-requiring verbs.  But I don't know that that
qualifies as something you're overlooking.

>John Vertical
I hope that helps!

Reply