Re: Trigger language?

In a message dated 1/20/03 10:54:44 AM, hsteoh@QUICKFUR.ATH.CX writes:

>On Sun, Jan 19, 2003 at 10:20:01PM -0500, Josh Roth wrote:
>> >How would you describe such an odd system? :-)
>>
>> Perhaps 'locational' - for the case system that is, not all the other
>things
>> that go with it! Rather than focusing on who performed an action and
>who felt
>> it, E's cases are, as I am seeing it, based on where they start and stop,
>and
>> what has its location changed (the conveyant).
>
>OK, maybe my examples weren't very well chosen. :-) You are right that
>much of it is 'locational'; however, the 'locations' involved are
>sometimes metaphoric rather than physical. One example is:
I realized that when I said 'locational'; I meant it in a metaphorical as
well as physical way. I don't think the two examples below contradict such a
label. Of course, you have to take any label for a semantic-based system with
a grain of salt, because even two different languages with, say,
location-based systems will interpret some nouns' roles differently. I mean,
you could see 'beauty' as moving, or as staying still. But if an Ebisedian
speaker sees it as moving more than as not moving, and consequently uses the
conveyant case, well there you go.

>        I(rcp) see(v) beauty(cvy) her(org)
>        "I see her beauty."
>
>        I(org) look(v) her(rcp)
>        "I look at her."
>
>Not only so, but entire sentences or even paragraphs can be "subordinated"
>into a "noun":
>
>I(org) say(v) <cvy-subclause>go away</subclause> her(rcp)
>"I said to her, 'go away!'."
>
><org-subclause>he(org) injure(v) past(loc) me(rcp)</subclause> city(org)
>I(cvy) go(v)
>"Because he injured me, I left the city."
>
>He(org) insult(v) her(rcp) <rcp-subclause>she(cvy) shame(rcp)</subclause>
>"He insulted her so that she was in shame/embarrassment."
>
>In each of these examples, the subclause (or embedded sentence, if you
>will) behaves like a noun of the indicated case in the containing
>sentence.
I think most languages do this, at least to some degree. Take this English
example: "Go away!"'s not a nice thing to say.

Or, a sentence can be made into a noun phrase: [His injuring me] led me to
leave the city.

>Also, the name of the instrumental case is possibly a misnomer; the
>originative case is used for the source or origin of an event, and the
>instrumental is used for the "dynamo" or "propellent" that drive the event
>to its completion. So while you do use the instrumental in cases like "I
>rode the horse to town" (translated as "I(cvy) horse(instr) ride(v)
>town(rcp)"), you also have the following:
>
>I(org) give(v) gift(cvy) her(rcp) --> I gave a gift to her
>I(instr) give(v) gift(cvy) her(rcp) --> I delivered the gift to her.
><instr-subclause>he(org) injure(v) past(loc) her(rcp)</subclause> she(cvy)
>        anger(rcp) --> the fact the he had injured her (continually) drove
>        her to anger.
>
>> If you imagine a hierarchical classification, with several broad
>> categories at the top, 'locational' could be one, along with a system
>> where nouns are divided up as mentioned above, i.e., according to who
>> performed an action and who felt it (e.g., Kar Marinam, active
>> languages), and a system where syntactic positions determine case (this
>> could include nom-acc and ergative as well as tripartite langs).  Of
>> course the lines are a bit blurry here and there....
>
>Refresh my memory, what's a tripartite lang again? :-)
One where subjects of intransitive verbs (S), subjects of transitive verbs
(A), and objects (O) are all treated differently. Thus you have three
categories for what nom-acc and erg langs only divide into two.

I didn't state my hierarchy so well. The last part shouldn't be "where
syntactic positions determine case," of course, because many nom-acc, erg,
and tri langs have somewhat free word order. It's really more like "where
case is dependent on status as a subject or object," i.e., correlated in some
way with the structure of the sentence. Actually you could have two main
groupings, like this:

Structure-based:
   Nom-Acc (English)
   Ergative (Basque [fully or partially?])
   Tripartite (Yimas [a Papuan lang], partially)
   A, S & O alike (?Tagalog, when trigger is on an NP other than these)
   A & S alike, O different (Rushan [a wacko Iranian lang], partially)

Semantics-based:
   Active:
      Split-S [based more on semantics of verb] (?)
      Fluid-S [based more on semantics of noun] (Kar Marinam)
   Locational [or whatever else you'd like to call it] (Ebisedian)
   Animacy-based (some Native Am. langs)

I'm not claiming that this is completely right. For example, it could be
argued that Split-S and animacy-based systems are really structure-based,
since they depend on a specific property of a given verb or noun,
respectively, which may prescribe something different than an accurate
semantic description. That's probably not very clear. Example: In a Split-S
lang, the verb 'fall' may require the noun that falls to be in some sort of
Patient case, even if in a specific situation, the noun fell intentionally
and another lang would call it an agent, or an agent AND a patient ("I fell
myself"). Another language may call a leaf animate and a slave inanimate, so
if the slave hits a leaf, and it is a rule that the grammatically animate
noun is always acting on the inanimate one, you'll have to put some kind of
Reverse marking in, even though the true 'semantics' that happen when people
hit plant products usually aren't expressed with a Reverse marker.

Maybe instead of having two groupings like this, they could be two poles at
ends of a continuum, and Split-S and Animate-based could be somewhere between
them.

>> I have a question. You gave the example: "I see a man" --> "I(rcp) see(v)
>> man(org)"
>> KM would do this "I(senser) see man(focus)"
>> (Senser, remember, might as well be Patient. The word order would be
>reversed
>> too, but never mind that.)
>> To say "He shows me the man," I would add "He(agent)"
>> How would E render this, since the originative is already used? I imagine
>> you'd have to use a different verb, or add in another sentence or "because
>> of..." or something. The instrumental is somewhat tempting, but if we
>> happened to add "with the binoculars"="binoculars(inst)", that would
>already
>> be used too.
>[snip]
>
>Depending on context, this could be variously rendered:
>
>1) he(org) cause(v) man(cvy) <rcp-subclause>see(v) I(rcp)</subclause>
>
>Literally, "he causes the man to be seen by me".
If it is here, why wouldn't 'man' always be in the conveyant? Whether you
think of him as the *origin* of a certain instance of seeing, or an image
being *conveyed* to someone's eyes, he's really doing the same thing in the
causative and non-causative sentences! Perhaps I was wrong and E is not
purely semantically based. That's fine. What I personally might do though, is
always have the object of sight in the conveyant, and express the sentences
like this:

I(rcp) see man(cvy) = I see the man
I(rcp) see man(cvy) he(org) = He shows me the man; he causes the man to be
seen by me

>2) he(instr) cause(v) man(cvy) <rcp-subclause>see(v) I(rcp)</subclause>
>
>Same as (1), except that this also implies some effort on the part of
>"he".
According to what you wrote way up above, I'd think this would imply that the
source of the action was elswhere, and 'he(instr)' was just helping to propel
the image of the man somehow to its destination, willfully or not.

>3) he(org) make-appear(v) man(cvy) I(rcp)
>
>He shows the man to me. Basically, using a different verb as you
>suggested.
OK, but I still wonder, why 'man(cvy)' here and 'man(org)' elsewhere? Unless,
noun case is a function of sentence structure as well as semantics.

>For the binoculars part, it's a totally different construct depending on
>the context. If he is showing me the man by giving me the binoculars, it
>would be:
>        man(org) see(v) <subclause...>binoculars(instr) I(rcp)
>
>where the subclause would have the effect of "he(org) give(v)
>binoculars(cvy)". I.e., "I see the man through the binoculars he gave me."
>(I'm omitting the details so as not to bore you with the intricate
>workings of subclauses in Ebisedian.)
Would the subclause be considered a noun phrase, with a case of its own, or
it is considered a modifier of 'binoculars' (i.e., a relative clause), or is
simply a separate clause, which happens to be in the middle of this one?

>Of course, this sentence can be rearranged in many ways, one of the
>perhaps more interesting of which is "adjoinment", which is basically two
>sentences "rammed" together at a common word:
>
>he(org) give(v) I(rcp) binoculars(cvy)-(instr) see(v) man(org) I(rcp).
>
>Basically, this is "ramming" together "he(org) give(v) I(rcp)
>binoculars(cvy)" and "binoculars(instr) see(v) man(org) I(rcp)".
>Morphologically speaking, the instr case marker *prefixes* the word for
>binoculars in the first sentence. A rough translation of this would be
>"he
>give me the binoculars through which I see the man".
Sounds like a relative clause. Kar Marinam does the same thing for relative
clauses; the subject of the rc and the head noun must be the same, and this
noun takes both cases. And you can't always tell from the noun which case is
meant for which clause, which makes it extra fun.

>Note, however, that it is perfectly legal to have more than one noun per
>noun case, which may or may not have the same semantic function. Word
>order may become important in this case. Examples:
>        I(cvy) go(v) city(rcp) house(rcp) woman(rcp)
>        "I go to the city, to the woman's house."
>
>Literally, this reads "I go to the city, to the house, to the woman."
>
>        woman(org) city(org) go(v) I(cvy)
>        "I left the woman and went out from the city."
>
>Literally, "From the woman, from the city, I left."
I'd say the semantic functions are close enough in both of those to be
considered the same.

>T
>
>--
>Only boring people get bored. -- JM

Josh Roth

Reply