Re: Unresolved problems in Lyanjen morphosyntactic nomenclature

> From: Matt M.
>
> or, How pretentious a title can Matt come up with for his email?
>
> Here are some problems I've discussed on-list before (largely
> last autumn);
> I'd like to see if we can come up with any fresh insight on them. Please
> bear in mind that I'm only a second-year Lx major and my morphosyntax is
> entirely self-taught.
>
> PROBLEM ONE: Noun cases. Lyanjen has three non-oblique cases. The
> first one
> is used for the subject of intransitive clauses:
>
> The second one is used to mark agents, such as the agent of a transitive
> verb, whether or not there's a patient. The third one is used to mark
> patients, whether or not there's an agent.
>
> So. In Chapter 7 of _Describing Morphosyntax_, Payne describes a three-way
> division of noun roles between subjects, agents, and patients. Most
> languages either group subjects and agents in the same case
> (nominative, vs.
> accusative) or subjects and patients (absolutive, vs. ergative). None of
> these terms work for me since mine are all separate. What do you think of
> the case names "subjective", "agentive," and "patientive"?
The case names seem fine, but do realize that tripartite case distributions
are very rare. Cases generally exist to discriminate the syntactic relations
of a verbs NP arguments. In transitive constructions there are two NP
arguments (A- and P-function) requiring discrimination to avoid ambiguity.
In intransitive constructions there is only one NP argument (S-function)
and, therefore, from a functional viewpoint, no need to discriminate this NP
from any other. Since it is never necessary to distinguish morphologically
between the S-function and the A-function arguments, nor between the
S-function and the P-function arguments, (they never co-occur), either case
could be, and usually is, used to mark S-function arguments.  Thus, two
cases are necessary and sufficient to discriminate the syntactic roles.  I
believe the number of natlangs that use a tripartite case system can be
counted on one hand.

> PROBLEM 2. Verbs in uncomposed moods (participles, infinitives, volitives,
> and imperatives) in Lyanjen usually have a prefix indicating the
> case of the
> noun they "replace". This is kinda difficult to describe, so I'll
> just give
> an example:
[examples snipped]

> I don't know what to call the class of prefixes in question. I've
> tentatively settled on "valency", but I don't know if that makes sense.
> Suggestions?
Hmmm.  I'm not sure I fully understand your examples, but it doesn't look
like a valency operator to me.  Valency typically refers to the "number" of
the predicate arguments.  AFMCL, amman iar requires a valency marker on the
lexical verb that specifies both the number of NP arguments as well as their
semantic role.

pat -a- used in intransitive constructions where the argument in S-function
is semantically patientive.

agt -e- used in intransitive constructions where the argument in S-function
is semantically agentive.

agt/pat -ae- used in transitive constructions where the argument in
A-function is semantically agentive and the argument in P-function is
semantically patientive.

pat/thm -ia- used in transitive constructions where the argument in
A-function is semantically patientive and the argument in P-function is a
semantic theme.

agt/thm -ie- used in transitive constructions where the argument in
A-function is semantically agentive and the argument in P-function is a
semantic theme.

Stay curious,
David

David E. Bell
The Gray Wizard
www.graywizard.net

Wisdom begins in wonder.