revisions in Tepa number marking

On Wed, 16 Aug 2000, Raymond Brown wrote:

> I haven't learnt any artlang basically for the reasons the Grey Wizard gave
> in his email. But if I had to choose one to learn, it would be Tepa.
<BLINK> Why, thank you! I suppose that you might be disappointed to
learn that Tepa is undergoing a rather large makeover then ...
Actually the size of the remake gets smaller and smaller the more I
think about it; the amount of work involved is daunting.

It all started when I was looking through old issues of Language and I
came across an article by Dick Demers and Eloise Jelinek about Straits
Salish. What I thought was really interesting about their analysis was
that they maintain that the language has no meaningful distinction
between 'noun' and 'verb' on the surface level (there may in fact be
lexical items which correspond to 'nouns' and 'verbs' but they are
roots and roots never appear without inflection). Now there have been
similar claims made for neighboring languages (Nootka, other Salish
languages, etc), but I hadn't really paid attention. This was an
article that I paid attention to. Basically, roots undergo morphology
to create predicates; clitics are attached to predicates to make
sentences. Arguments are indicated solely by affixes and clitics (the
Pronominal Argument Hypothesis), and referring expressions are
completely optional. (Now I'm not a synctactician by training, but
this is the kind of thing that almost makes me want to change my
mind.)

Tepa already has leanings towards a language of this type, so I
thought about moving it all the way there. But I'd lose some nifty
features by doing so.

Something I did need to do in any case was to work out how verbs show
number; the system presently in place for nouns was designed for CVCV
roots; an example is the word for 'moth'

        sape       '(a) moth' (INDEFINITE)
        sapee      'the moth' (DEFINITE)

        saspe      'some moths' (PAUCAL INDEFINITE)
        sasapee    'the several moths' (PAUCAL DEFINITE)

        sapespe    'moths' (DISTRIBUTIVE)
        sapesapee  'every moth' (EVERY)

        sakpe      '(a group of) moths' (COLLECTIVE)
        sakapee    'all moths' (ALL)

This system includes categories which we normally express with
quantifiers. This system doesn't work right for the prosodically more
complex forms found with verbs, however; this is what I've been trying
to fix. I have a tentative solution which slightly alters the prosodic
realization of the different number categories. It also involves
eliminating PAUCAL as a category, and marking DISTRIBUTIVE with former
PAUCAL prosody (paradigms available upon request).

By doing this, I've essentially eliminated the need for the number
distinctions on nouns. Since the person prefixes also capture the
argument structure for any given predicate, the nouns become
superfluous in context--hence, a Pronominal Argument language. Here's
how it might work. Take a sentence with a plural subject in English:
'The sheep are running around.' In Tepa this becomes:

        lulpa toko
        0- RED-  lupa toko
        3- DIST- run  sheep

Notice that the number marking does not mark an entity, but rather an
event;  here it shows that the event of running is distributed over a
number of sheep--the implication is that each sheep is running around
independently of her neighbors. In the context of a story, the noun
'sheep' may be entirely dispensed with, leaving only _lulpa_ 'They
were running around' or even, 'Running around was going on all over.'

The situation becomes a little more complicated with transitive
predicates. Suppose you have the following Tepa sentence (my apologies
for the violent nature of the example!)

        wanpopti nema
        wa-  n-  RED-  poti nema
        1>3- TR- DIST- beat man

This has three different meanings in English. It could mean:

        i.   I beat up the men.
        ii.  We beat up the man.
        iii. We beat up the men.

Only context can determine which reading is intended since number is
not marked on the noun but only on the predicate. (However, the first
reading is not as likely since the implication of the English sentence
is that the event of beating up happened once to a group of men--that
might be better expressed with a collective verb form: _wanpokti_. It
may be possible to use the distributive for a series of similar events
so a more accurate version of i. would be 'I beat up one man after
another.')

If this post seems a bit scattered or unorganized then it's an
accurate reflection of my state of mind WRT Tepa. Comments? Questions?

Dirk

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