tlhIngan-Hol Archive: Wed Oct 07 14:51:02 2009

Back to archive top level

To this year's listing



[Date Prev][Date Next][Thread Prev][Thread Next]

Re: The meaning of -moH

Brent Kesler (brent.of.all.people@gmail.com) [KLI Member]



On Tue, Oct 6, 2009 at 3:00 PM, David Trimboli <david@trimboli.name> wrote:
> Brent Kesler wrote:
>> Some verbs are monovalent (one argument):
>> 1. tuH yaS
>> - The officer is ashamed.
>
> But is it monovalent because of some syntactic rule, or is it monovalent
> because of some semantic rule? The evidence suggests that there is
> nothing *syntactically* wrong with saying {puq tuH yaS}; it just doesn't
> make any *semantic* sense.

First of all, what evidence suggests this? Second, how can you tell
the difference? You have a bunch of verbs that never take an object.
How can you tell which ones are intransitive because of semantics, and
which ones are intransitive because of syntax? Also, if only syntax
makes a verb intransitive, how could giving it an object make any
sense semantically? In other words, what makes semantics and syntax
different from each other under your analysis?

Using our semantic intuitions to determine whether or not a verb can
take an object doesn't work well. Take {jeq}, which means "protrude".
In English, "protrude" is intransitive.

1. The knife protrudes from the corpse.
2. *The knife protrudes the corpse.

But in Klingon {jeq} is transitive.

3. lom jeq taj.
4. *lomvo' jeq taj.

Would you say that there is nothing syntactically wrong with sentence
2, it just doesn't make semantic sense? Then why does sentence 3 make
semantic sense? There is little difference between "protrude" and
{jeq}, except {jeq} has a syntactic rule defining a semantic role for
its object. "Protrude" does not have a syntactic rule defining a
semantic role for an object, so it takes no object. The semantics does
not determine the syntax; the syntax determines the semantics.

> There is no list of verbs wherein, when you
> don't know their meanings, you can still declare that they don't take an
> object.

But if you don't know their meanings, how can you declare that they
*do* take an object?

> When you add {-moH} you're not changing the fundamental syntax
> of the sentence; you're just adding a meaning that lets an object make
> semantic sense.

Then why does {yaS} switch places between {tuH yaS} and {yaS tuHmoH
puq}? If I ask {tuH 'Iv} after either sentence, the answer would be
the same: {yaS}. {-moH} hasn't just given {tuH} the means to take an
object, it has also changed what it means to be the subject.

I agree the syntax hasn't changed; the sentence is still OVS. But the
semantic roles assigned by the syntax does change. Valency change
explains this. Lexical semantics does not.

>> The problem with applying the causative to transitive verbs is that we
>> end up with three arguments with only two core slots to put them in,
>> so we have to resort to a non-core marking, {-vaD}, for one of them.
>
> We're not "resorting" to anything. The {-vaD} noun of, say, {ghojmoH}
> "cause to learn" is the legitimate beneficiary or recipient of the
> causing to learn. It never would have been a core argument.

Why not? Without observing how the language works in practice, how are
we supposed to know how {-moH} assigns semantic roles to its various
syntactic slots? The semantic approach cannot tell us whether:

(iv) New causer becomes the new A, the original A becomes the new O,
and the original O becomes a peripheral (non-core) argument.

or

(v) New causer becomes the new A, the original A becomes a non-core
argument, and the original O remains the O.

Klingon appears to behave like (v). But, without evidence, it could
just as easily behave as (iv): *{QeD'e' puq vIghojmoH}. Nothing about
the semantics of {ghoj} and {-moH} makes it obvious which sentence is
correct. Only real sentences show us which model to use.

bI'reng






Back to archive top level