Re: 'Yemls Cases - Comments?
| From: | Jeff Jones <jeffsjones@...> |
|---|
| Date: | Wednesday, February 19, 2003, 22:10 |
|---|
On Wed, 19 Feb 2003 20:59:17 +0100, Christophe Grandsire
<christophe.grandsire@...> wrote:
>En réponse à Jeff Jones <jeffsjones@...>:
>
>> This is a simplified description of how 'Yemls matches case-roles to
>> arguments. It's short on examples, partly due to vocabulary shortage.
>
>Don't worry, I'm an expert on grammatical discussions without examples too,
>since I just usually don't have any at hand :)) (or :(( here).
Good. Your reply counts as an expert comment, then!
>> There are 3 primary cases corresponding to the 3 primary arguments,
>> representing 3 of the various possible case-roles. These are called
>> the A-case, the P-case, and the C-case. There are also 3 basic
>> argument-to-case correspondence patterns used.
>
>I imagine A is for "Agent" and P for "Patient", but C?
More or less. Each case is "officially" named after the prefix that puts it
into the subject position. Extrahistorically, the prefixes were just
mnemonics for Agent, Patient, and Complement, respectively, but I ended up
using the identically written syllables for the orthographic forms
(pronounced [gV] or [gO], [bE], and [t_S_hE]). The term "complement" comes
from MNCL, where the 3rd argument is required wherever it's defined, thus
necessary to "complete" the verb phrase. In 'Yemls, C could also stand for
"clarifier" (similar to Rick Morneau's focus case). Intrahistorically, the
prefix {A} comes from a verb {AA} meaning to do or to cause, while the
prefix {P} comes from a verb {PP} meaning to undergo something. I'm not
sure where {C} might come from yet. Perhaps an inverse voice marker that
got extended to action words.
>> Head-marking is used for cases, meaning the head of the clause
>> determines which argument has which case. The presence of a
>> "grammatical voice" prefix on the head word, either {A}, {P}, or {C},
>> always selects one of the basic patterns:
>> A-case: {A}head P-case C-case
>> P-case: {P}head C-case A-case
>> C-case: {C}head P-case A-case
>
>It looks a little like Itakian in imperfective sentences, which are TVC,
>i.e. Trigger-Verb-Object word order, where the "Object" is a mandatory
>direct "object" (absent from sentences in the perfective aspect which take
>VT word order) whose function depends on the function of the trigger and
>the meaning of the verb. It's a bit more idiosyncratic than your system
>though.
Undoubtedly :) It sounds like Itakian combines trigger with a glance toward
ergativity. In 'Yemls, aspect and case-marking are completely independent
of each other, which is logical, but very un-indoeuropean.
>> When none of these prefixes is present, the pattern depends on the
>> head word's class. Words denoting actions generally take the 1st pattern,
>> while other words generally take the 2nd. Those in the 1st group don't
>> take the {A} prefix and those in the 2nd group don't take the {P}
>> prefix.
>
>Sounds pretty reasonable so far. I have the same with trigger affixes on
>the verb (the affixes which mark the function of the trigger in the
>sentence). IIRC, if the trigger is of class 1 or 2 (so-called human
>classes), the trigger affix for actor is actually the null affix, i.e. no
>affix at all. It's an idea of action hierarchy which considers that the
>higher in the hierarchy you are, the more common and normal it is for you
>to be an actor, and thus you needn't mark what's taken as a default.
I'm not sure I understand. Are you saying that the presence of the trigger
affix depends on the class of the trigger argument (my subject) rather than
on the class of the verb (my head word) as I have it?
>> ABSENT ARGUMENTS
>>
>> An argument may be absent for lexical reasons (because its case isn't
>> defined for that particular head word), for syntactical reasons (such
>> as the subject in participial and imperative clauses), because it's
>> implied, or because it's not specified. Some specifics:
>> The subject is always present except in
>> * "impersonal" clauses
>> * participial and imperative clauses (implied)
>> * when implied in adverbial and infinitive clauses
>> For typical "nouns" and "adjectives", the C-case will be undefined and
>> the A-case will normally not be specified. However, for relationship
>> words (including kinship and body part terms), the C-case will
>> normally be required.
>
>Which raises the question: how do you do when the absent argument would be
>the 2nd argument, while the 3rd would be present? The 3rd argument could
>be confused with a 2nd argument, hence a confusion in function of the
>argument. Or would you use the voice prefixes to ensure that the missing
>argument is never in 2nd position?
Good question. I've considered a lot of things, including additional
prefixes, but what I have been using is a rule requiring the 2nd argument
to be present if the 3rd argument is present. This rule doesn't apply if
the 2nd argument is undefined. Also, there are probably some situations
where it's clear which of the 2 cases is meant.
>OK, not much to comment, as it looked clear to me so far. Of course, if you
>could explain the usual role of each case, it would help me understand the
>thing a little better :) .
Let's see. For actions, the A-case argument is the agent who is directly
involved in the action. For other words, the A-case argument is an indirect
agent or cause. The P-case argument is what's affected by the action,
undergoes a change of state, perceives or receives something, or is simply
an intransitive subject. The C-case argument "clarifies" an action. It's
used as the object of perception, a required complement for relationship
words, or as a 2nd "object".
Does that help? I suppose I could come up with examples (using mostly
English words) if not.
Jeff
Reply