tlhIngan-Hol Archive: Mon Mar 08 10:09:13 1999
[Date Prev][Date Next][Thread Prev][Thread Next]
Re: deep structures and parsing
On Sat, 6 Mar 1999 16:19:27 -0800 (PST) Alan Anderson
<[email protected]> wrote:
> This thread has had an inappropriate KLBC on it for too long.
Agreed. Thanks for catching that.
> ja' charghwI':
> >Well, I just ran into a counterexample in TKD. Now that I see
> >it, I remember being bothered by it:
> >
> >{HIQoymoH} "let me hear (something)"
>
> That's the one! I knew I had seen another example of a possibly
> transitive verb actually used this way.
I try to stay honest. I would not withhold evidence counter to
my argument.
...
> >[sigh]
> >
> >I guess my argument is failing. I want to be honest. I'm not so
> >much interested in winning as I am in figuring out how this all
> >works. I thought this was clearcut, but as we progress, I see
> >that it is muddier than I thought.
>
> Yes, this is apparently a very fuzzy area of the grammar. When two
> people can make each other understood well yet disagree so strongly
> about exactly *why* they are understood, there's obviously something
> odd going on behind the scenes.
Agreed.
> >[...]
> >I tend to parse Klingon left to right, as if I were hearing it,...
>
> Meanwhile, I understand both written and spoken Klingon a bit more
> holistically. I don't try to "parse" it syllable by syllable. I get
> an entire phrase in my head as a chunk and let the pattern of roots
> and suffixes sort itself out into a meaningful idea based mostly on
> examples I have encountered. It's not often that I must concentrate
> explicitly on how a particular suffix affects the subject and object
> of a verb.
Interesting. I'm often fascinating to hear people describe how
they process language. We share language among people who use
vastly differing mechanisms to process that communication. I
can't read without hearing the sounds in my head, so I'll come
to a complete halt when I encounter a word I don't know how to
pronounce. I can't understand it until I can "hear" it. A friend
has often stopped others to ask the meaning of the word the
other person just said, even though she was very familiar with
the word from reading it. She makes no connection whatsoever
between what she reads and what she hears. It is like her
reading and speaking are two different languages for her.
And now your description... I definitely process language from
memory. We all have to. I just begin processing it while I hear
it. I don't wait for it all to come in before I start digesting
it.
> [deconstruction of targhvaD qaSopmoH snipped]
>
> >It doesn't matter that there is no explicit direct object. We
> >have an indirect object and an "indicated" direct object and
> >subject. These fit the grammar for transitive verbs with {-moH}.
>
> They also fit the grammar for indicating beneficiaries of a verb
> that does not suffer from the "ditransitivity" problem.
I need examples before I can agree.
...
> >Since we parse this in a different sequence, it is natural that
> >we'd differently conclude its meaning. {-moH} changes the
> >grammatical role of the nouns surrounding the verb and the
> >"indicated" pronouns in the prefix. I included the {-vaD} when I
> >interpreted that role modification and you didn't. I don't
> >consider {-moH}'s alteration to an indirect object's role to be
> >optional and you do.
>
> I tend to interpret {-moH} more as modifying the meaning of the verb
> itself than as reassigning roles of the surrounding nouns.
In that case, I think you are just going through some
subconscious reassigning of the nouns, because the "change" you
describe for the verbs really changes the relationship of those
verbs with their related nouns.
> Example:
> {nguv} "it is tinted"; {nguvmoH} "it tints, it paints". I don't pay
> much attention to what the subject and object of the un-moH'ed verb
> would be when I hear a -moH'ed one, unless the verb is one I am not
> used to encountering in an intransitive sense in the first place.
The net result is the same. Those nouns have reassigned roles.
> I
> often hear {Sop} used in ways similar to {qet}, for instance, so the
> word {SopmoH} quite naturally strikes me as a verb with its natural
> object as the entity being made to eat.
Meanwhile, if something pleases me, I see that as quite
different from something causing me to eat. Causation is so
imbedded into "pleasing" that I don't even see it as cause, but
it is so surficial on "cause to eat" that I can't ignore it.
> >> I'm just not seeing any reason to forbid interpreting {qa-V-moH} in
> >> a consistent way whether or not {N-vaD} is in front of it.
> >
> >My reason is that in our one example that includes both {-vaD}
> >and {-moH}, the {-moH} altered the grammatical function of the
> >noun with {-vaD} and I see no reason to consider that
> >modification to be optional.
>
> In that example, the concept underlying the words has to be expressed
> using something appropriate for a ditransitive meaning. We do have
> more than one example of ditransitive constructions -- {'oHvaD Qo'noS
> ponglu'}, for instance -- and I have no problem with it at all. I'm
> just not convinced that having a {-vaD} and a {-moH} at the same time
> *always* has to imply a ditransitive interpretation.
Give me an example that will make sense to me.
> >You and I both likely agree that {-moH} always modifies the
> >grammatical function of the subject and direct object of the
> >verb. You consider its effect on an indirect object to be
> >optional. I have no reason to agree. I know that my way works.
> >I've never seen an example that shows your way to work. If I see
> >one, I'll definitely change my mind and agree with you that this
> >is ambiguous.
>
> Hmm. I really don't see {-moH} affecting an indirect object's function
> much at all. If I say {vavwI'vaD nguv jan}, have I really changed the
> role of {vavwI'} when I say {vavwI'vaD nguvmoH jan}?
Yes. "The device is stained for my father." "The device causes
my father to stain (something)." These two are COMPLETELY
different, especially in the role of your father.
> >I hope this is not the case. I do accept ambiguity, but I really
> >feel a loss when new, unnecessary ambiguity is introduced into
> >the language.
>
> I've *always* seen the {ghaHvaD quHDaj qawmoH} example as potentially
> ambiguous. Because "heritage" is so unlikely to be able to "remember"
> anything, I never thought it was worth pointing out the existence of
> the unlikely alternate meaning. But {tIqwIjvaD qaDuQmoH} works in the
> other direction -- since "my heart" is so unlikely to be able to "stab"
> you, I decided it *was* worth pointing out what I think is the likely
> alternate meaning.
If {ghaHvaD quHDaj qawmoH} was not an example of new grammar and
was merely an ambiguous statement which can be decyphered by
context, then how did it even get its "secondary" meaning which
we would select from context? What explains an indirect object
acting as subject of the root verb? It had never happened before.
I don't think that there are words that are ditransitive and you
can arbitrarily assign one of the objects as direct and the
other as indirect and it is just ambiguous and you figure it out
from context. I think that is an absurd idea. As much as I think
it is wrong, I HOPE it is wrong.
Instead, I think that each verb has established relationships
with its objects. These are somewhat arbitrary, but they are
consistent. That's why it sounds natural to say "Orbit the
planet in your ship" and even with English's wildly varying word
order, it sounds awkward to say "Orbit your ship around the
planet." The word orbit implies "around" as its relationship
with its direct object.
When we say "Give me the apple," we are not arbitrarily
assigning direct and indirect object. We have established rules
for determining which is which. I honestly believe that Klingon
does the same thing. For {-moH} on transitive verbs, we have
only one example of the grammar and I think that prescribes how
we should determine what is direct and what is indirect object
IN ALL INSTANCES of {-moH} with transitive verbs. I also think
that given this example, all instances of verbs with {-moH} that
have indirect objects should be assumed to be transitive until
we are given some sort of example showing how that would work.
You still haven't given an example that looks to me like it
would work in anything but the grammar this example demonstrates.
> >I yield on the argument that I made too much of a big deal out
> >of transitive vs. intransitive verbs. I did it at the beginning
> >of this post and instead of editing it out so I'll appear to be
> >more correct, I'll leave it and admit that I was wrong.
>
> I will not gloat. While there indeed are examples that toss a lot of
> the basis for your argument out the window, we still haven't actually
> managed to resolve things clearly in my favor either.
>
> >I hold on the argument that the grammatical role-change for an
> >indirect object of a transitive verb root with {-moH} is
> >optional. I think it is a strong argument. I don't see the
> >weakness in it.
>
> The flaw in your argument I see is that you assume that a transitive
> verb root can not be used for an intransitive idea, and that {-moH}
> on such a verb root necessarily creates a ditransitivity issue.
No. That is not accurate. I do not accept (until shown
otherwise) that a verb with {-moH} can have an indirect object
that is not acting as subject of root verb with a sense of
transitivity even if the direct object is unstated and not a
point of focus, as is the case with {Sop}. You can't eat without
eating SOMETHING, so even as you use {Sop} "intransitively",
there is an implication that it actually is transitive with a
very vague indifference to what that direct object would be. It
is the direct object equivalent to what {-lu'} does to the
subject.
> >[...]
> >{targhvaD Qa'Hom vISopmoH}
> >
> >By my model, this means, "I caused the targ to eat the Qa'Hom."
> >Your model could also explain it as "I caused the Qa'Hom to eat
> >for the targ." Do we gain any ability to communicate useful
> >things clearly by adding this unnecessary ambiguity? I don't
> >think so. I don't just THINK I'm right. I HOPE I'm right. The
> >language will certainly survive if I'm wrong. We got along well
> >enough when we couldn't use {-moH} on a transitive verb at all.
> >I guess we'll do okay if every single such example will always
> >be ambiguous.
>
> I actually read it most literally as "I cause-to-eat the Qa'Hom
> for the targ." The ambiguity comes in when trying to decide if
> it should be "I cause-to-eat-the-Qa'Hom for the targ." Does the
> *direct object* get captured in the effect of the {-moH} or not?
> It's a similar sort of scope problem that you and I occasionally
> disagree on when considering how much of the sentence can get the
> effect of the suffix {-be'}, except this time it looks like you
> are arguing that the more limited range of action is not right.
You are introducing the infinitive in a way that creates this
ambiguity. In earlier, intransitive examples, the grammar made
the direct object become subject to the root verb. "I cause the
Qa'Hom to eat." In the later transitive example, the role
changes for the direct object. The indirect object usurps its
earlier function. "I cause the targ to eat the Qa'Hom."
By your current description, the indirect object's role never
changes. You just have to decide whether or not to include the
direct object as a direct object or leave it apart, in which
case it magically migrates to the subject of the root verb.
So, by your model, let's go for that non-inclusion of the direct
object. "I cause-to-eat the Qa'Hom for the targ." If it is not
included as direct object, why not just not include it? "I
cause-to-eat for the targ." While we are dropping off optional
nouns, let's drop the indirect object as well. "I-cause-to-eat".
I think I saw one weird example of a verb with {-moH} being used
intransitively and it looked like a mistake.
TKD p45: maghoSchoHmoHneS'a'
I think Okrand just popped a fuse trying to combine suffixes. It
is the only intransitive prefix I've ever seen on a verb with
{-moH} and I really don't think it works that well.
Meanwhile, for your construction...
I really think what is happening is that you are not paying much
attention to the relationship between the {Qa'Hom} and the
eating. You just hang the noun off to the side and let it slide
in whereever it makes the most sense. Is it the subject or the
object of the eating? You don't care much. You just let your
English property of letting words slide around and let that
slide into Klingon.
Meanwhile, in Klingon, the subject and object are always
defined. Other nouns can come first, especially with Type 5 or
Time Stamp references, but the subject and object are generally
clear in Klingon. Why do you consider that feature of Klingon to
be so unnecessary as soon as one adds {-moH} to a verb?
> The rest of this is about {-lu'} and should probably be in a separate
> thread, but I don't really think it calls out for more than a couple
> of sentences and I don't want to spawn another long discussion on it.
>
> >The section in TKD which describes {-lu'} spends a paragraph or
> >two describing the affect {-lu'} has on the proper
> >interpretation of the prefix. The indicated subject changes
> >roles so that it instead indicates the functional direct object,
> >and the indicated direct object is always third person singular
> >so that it maps to the grammatically singular indefinite
> >subject.
> >
> >None of this is valid if there is no direct object. There is no
> >such thing as a prefix for "no subject", so you can't reverse
> >roles for indicated subject and object. So, there is a role
> >versal for the indicated subject and object for all verbs with
> >{-lu'}, unless the verb is intransitive, in which case the
> >prefix remains unaltered.
> >
> >We had to learn that exception through example because the
> >grammatical description never touched on it. It was simply not
> >addressed. I learned a mechanism and the exception to that
> >mechanism was not described, so I was confused. Through canon
> >usage, I've learned this exception to the rule, but it continues
> >to be an exception. The roles of indicated subject and object do
> >not reverse for intransitive verbs the way it does for
> >transitive ones.
> >
> >You don't see that as exceptional?
>
> I see your whole focus on treating transitive and intransitive verbs
> differently as a major complication of the grammar. The grammatical
> description consistently says that the basic meaning of {-lu'} is
> merely "indefinite subject", and the prefix reversal is incidental.
It may be incidental, but it is necessary and it violates the
normal roles of subject and object in the grammar. {vI-} always
means first person subject and third person object UNLESS THE
VERB INCLUDES {-LU'}, IN WHICH CASE THE ROLES REVERSE. All of
the discussion in TKD about what the prefix does when {-lu'} is
present describes this reversal of roles. It is never mentioned
that if the verb is intransitive, that reversal does not occur.
> The only problem with that is what to do when there's a verb needing
> to indicate no object at all. I reconcile the usage with a simple
> observation that there indeed is no verb prefix given for "no subject"
> in the matrix, so the obvious way to deal with it is either not to use
> a prefix at all (radical, eh?) or simply to co-opt the null prefix to
> carry the required meaning.
Meanwhile, Okrand has explicitly explained that there is no such
thing as a verb with no prefix, hence the "null" prefix, so you
can toss that possibility. As for "co-opting", my whole point is
that you are making up new grammar not described in TKD when you
use {-lu'} on any verb and you DON'T reverse the subject and
object functions of the prefix. You do have to make up this new
grammar to explain the example given in the stock phrase section
of TKD, so clearly this is the right thing to do, but to try to
say that the example doesn't violate the described grammar is,
well dillusional. It simply does.
Show me the place in the grammar description where Okrand
explains the exceptional nature of handling the prefix for
intransitive verbs. I looked for it a lot and I never found it.
> That seems simpler to me than summoning
> into existence an exceptional grammatical treatment of intransitive
> verbs when the {-lu'} suffix is used.
Well, you see, there are rules to this grammar, and that case
violated those rules, so I had to make up new rules. So did you,
whether you are willing to admit it or not.
> -- ghunchu'wI'
charghwI' 'utlh