Re: revisions in Tepa number marking

On Thu, 17 Aug 2000, J Matthew Pearson wrote:

> I like it!  I can envision two possible ways to reduce the ambiguity generated by
> this system without losing the spirit of what you're trying to do.  I played
> around with both of these strategies when I was working on a sketch for a conlang
> a while back, and I came up with a couple ideas that I was fiddling with (the
> conlang itself didn't go anywhere, but I might return to some of its major design
> features at a later date):
>
> The first thing would be two expand the number of categories in the agreement
> system.  I recall that Tepa agreement prefixes only mark person of subject and
> object, not number.  This is a cool feature, and I presume you would want to keep
> it in the revised Tepa, but you might consider adding some finer person
> distinctions than just the standard 1, 2, and 3.  For example:
>
>   1      me (1st singular)
>   1+2    me and you, us two (1st dual inclusive)
>   1+2+3  me and you and other(s) (1st plural inclusive)
>   1+3    me and other(s) (1st dual/plural exclusive)
>   2      you (2nd singular, or plural when addressing a group)
>   2+3    you and other(s) (2nd plural, when addressing part
>             of a group)
>   3      other(s) (3rd singular/plural)
I remember this ... this was Amashkii, right? I still think that your
system of adjunction there was very cool and worth pursuing somewhere,
if not in Amashkii.

The number prefixes for Tepa are as follows:

             intr   tr
        0-     3     -
        wa-    1    1>3
        ku-    2    2>3
        ne-    3    3>3'
        le-    -    2>1

(The intransitive _ne-_ is a kind of antipassive, if I understand it
right; it will require some more thought.)

The transitive person prefixes don't so much indicate the arguments of
a predicate as much as the _direction_ along which the event flows.
Hence, _wa-_ is the direction from first person to third person, &c.

The way I see it, adding more person distinctions would create
exponential explosion of person prefixes. Since persons aren't
indicated separately in the prefixes, I would need about 7 x 7 = 49
prefixes to encode the transitivity relations for the distinctions you
suggested. Even if I only use a subset of those distinctions (thus
reducing the number of prefixes needed), I still think it would be
more suited to a system which marks subjects and objects separately.
I'll tuck that one away for my next project :-).

> A second strategy would be to break up the number marking on verbs into two
> domains.  Verbal predicates can be grouped into four different categories
> according to their eventivity and punctuality:
>
>   state ("alive", "happy")
>   activity ("sing", "eat apples", "hunt for deer")
>   activity leading to change-of-state
>         ("build a house", "eat the apple")
>
> You could have two parallel systems of plural marking, which operate separately
> or together, depending on the semantics of the verb.  One marking set indicates
> singular/plural of states (i.e., multiple states, or a single state shared by a
> group of entities), while the other set indicates singular/plural of activities
> (i.e., multiple activities, or a single activity shared by a group of actors).
> Verbs which denote states or changes of state would take state singular/plural
> marking (abbreviated SSg/SPl), verbs which denote non-terminal activities would
> take activity singular/plural marking (abbreviated ASg/APl), and verbs which
> denote an activity terminating in a change of state in the patient would take
> both:
>
>   3-happy-SSg     "s/he is happy"
>   3-happy-SPl     "they are happy"
>                    or "s/he is happy (on different occasions)"
>
>   3-sing-ASg      "s/he sings (once)"
>   3-sing-APl      "they sing", "s/he sings (iteratively)"
>
>   3>3-eat-SSg-ASg   "s/he eats it"
>   3>3-eat-SPl-ASg   "s/he eats them (collective)"
>   3>3-eat-SSg-APl   "they eat it"
>   3>3-eat-SPl-APl   "they eat them", "s/he eats them (on
>                         different occasions)"
This looks like number agreement on an ergative pattern, with the
absolutive argument marked SSg/SPl and the ergative argument marked
ASg/APl. Or am I not understanding it correctly?

> Just a couple ideas...
Thanks! I should say, though, that the potential ambiguity in the
Tepa system as it now stands doesn't bother me over much. But you have
given me some other things to mull over (and in the course of my reply
made me realize that I just might have antipassives in Tepa!).

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu