tlhIngan-Hol Archive: Thu Feb 05 08:38:41 1998

Back to archive top level

To this year's listing

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

Re: Locatives and {-bogh} (was Re: KLBC Poetry)

According to Terrence Donnelly:
> 
> At 02:03 PM 2/4/98 -0800, ~mark wrote:
> >>From: Terrence Donnelly <[email protected]>
> >>3. The head noun of the {-bogh} clause can also take those roles in
> >>the matrix clause which are marked by the other Type 5 verb suffixes.
> >>
> >>     |-------------|                         |--------------|
> >>     meQtaHbogh qachvo' vIHaw'       jor pengvo' Haw'bogh Duj
> >>                    |--------|       |------|
> >>     I flee the burning building     The torpedo from which the ship is
> >>                                     fleeing explodes.
> >
> >Ack...
> >
> >>4. You can't say things like "the ship in which I fled" because
> >>the locative (or other Type 5-suffixed noun) is not the subject or
> >>object of the {-bogh} verb, and so can't be the head noun of the clause.
> >
> >You just did say "the ship in which I fled" (or "the torpedo from which
> >the ship fled", which is the same grammar).  I agree that you can't.  But
> >you just did.
> >
> 
> I know, I know.  charghwI' has already pointed out to me that I did exactly
> the opposite of what I was attempting to say.  HIvqa' veqlargh jay'!
> 
> As an attempt at a Type-5 suffix on the _object_ of the relative clause
> (Which is what I was trying to do above and failed at), I proposed
> 
> {pengvo' baHta'bogh toQDuj Haw' veSDuj}  'the warship flees from the
>                                           torpedo launched by the BoP'
 
The problem with this example is that it can very easily be
interpreted to mean, "From the torpedo, the warship flees the
Bird-of-Prey which has successfully fired." We don't know that
{Haw'} can't take a direct object. If it can, {toQDuj} appears
to be that object, especially since it is available to be the
head noun of {baHta'bogh}.

This is NOT like the "At the edge of the known galaxy" example
(or was it "At the edge of the unknown universe"? I don't
remember). That example, like the {meQtaHbogh qach} example did
not have any other noun available to be head noun except the
one with the Type 5 suffix.

I don't like this construction, even when it is as
exceptionally (and accidentally) clear as Okrand's example.
Yours is worse because it includes a second candidate to be
head noun of the relative clause.

> >
> >>     qeylISvaD lIjlaHbe'bogh vay' maSuv
> >>         We fight for Kahless the Unforgettable
> >
> >If my analysis is right, this would have to be "?qeylIS lIjlaHbe'bogh
> >vay'vaD maSuv."  Assuming this can be done at all, as charghwI' quite
> >rightly points out (as did I).

This doesn't work for me. I would have marked {qeylIS} with
{-'e'}, but there is no canon to support this and we have canon
to contradict your suggestion.
 
> charghwI's guiding principle (he tells us) for considering innovations
> is how it plays out as it is heard one word at a time. (I think he
> gives too little attention to the possibility that the spoken and written
> registers are not necessarily identical, also tends to downplay the role
> that body language, intonation and speech patterns could play in
> making sentence structure clear, but that's another topic.)

In my defense, I have this guiding principle simply because in
all good examples of well written, clear sentences in Klingon,
this analysis works quite well. In confusing attempts at
Klingon sentences, this analysis usually points out why it is
confusing.

~mark talks about the "stack" in terms of holding one bit of
grammar/symantics in mind while taking in more of a sentence
from a different grammatical context. That analysis is good,
but mine focusses on another phenominon. Sometimes, as you hear
a sentence, you are naturally drawn toward interpreting the
grammatical setting of the words as you hear them. When things
get confusing, you find that you've been misled and you have
to, in memory, reassign words to different grammatical contexts.

When you have to do this, it gets confusing. If you want to
write clearly, you should avoid this kind of confusion. Since
word order interacting with affixes is at the core of Klingon
grammar, the vast majority of the time, you can tell the
grammatical function of words while you hear them instead of
having to lay them out and analyze them after you have heard
the entire sentence. This is one of the strengths of Klingon
which compensates for its "lack" of helping words, which
English has in such a huge volume and variety.

> My guiding principle, I guess, is how universally applicable a rule is.
> I like rules that can be applied in most cases without exceptions.  This
> is why I don't like your proposed solution.  If this is the proper way
> to put a Type-5 suffix on a relative clause, by considering the entire
> clause as if it were a single noun, then this is the only place in
> Klingon grammar where this is allowed, where a collection of separate
> words functions as a unit.  

Qe' QaQDaq jISop.

{Qe' QaQDaq} is a collection of separate words that function as
one unit.

HoD puqvaD taj vInob.

{HoD puqvaD} is collection of separate words that function as one
unit.

Relative clauses, adjectival verbs and noun-noun constructions
all modify nouns. From the context of the verbs for which these
modified nouns serve as subject or object or for which they
have a Type 5 noun suffix function, the entire clause or phrase
acts like a single noun. Moving Type 5 noun suffixes to the
adjectival verbs and disallowing Type 5 noun suffixes on the
first noun of a noun-noun construction are both examples of the
entire phrase being treated as a single noun.

If you think there are not already examples of collections of
separate words functioning as one grammatical unit, you aren't
looking very closely. As it happens, ~mark is wrong here, but
not because of your guiding principle. He is wrong because
Okrand's example shows him to be wrong. Meanwhile, what he
proposes adheres to existing grammatical principles as much as
Okrand's counterexample does.

> It also makes (keeps?) {-'e'} unique
> among Type-5 suffixes, and requires that you make an exception and use it
> whenever the head noun is _not_ the noun to which the other Type-5 suffix
> applies.

That would be true if I had been right about this contruction.
I wasn't, so this is not true. The fact is, if any noun other
than the one to which the Type 5 has been attached is available
to be head noun, you are not following the examples Okrand has
given us. In both his examples, the Type 5 noun is the only
available noun to act as head noun of the relative clause.

> Also, to me, understanding that the {-vaD} in the above example refers back
> to {qeylIs} and not the {vay'} is a big exception to our usual
> interpretation of nouns with Type-5s on them.

I would not argue with that. This is why I suggested we use
{-'e'} on {qeylIS}, but I was wrong.

> I'd be more likely to accept the form where the suffix is appended to
> the {-bogh} verb and the head noun, subject or object, marked with
> {-'e'}, except that this is totally unsupported by canon.

My sentiments exactly.

> But I see the idea that {-'e'} or any other Type-5 suffix can be appended
> to the head noun (subject or object) of the relative clause as being
> much more regular.  It simply takes what we already know about the
> {-'e'} suffix and its role in relative clauses, and extends it to the other
> Type-5 suffixes (and its role in the matrix clause; after all, {-'e'} on a
> head noun has no grammatical effect on the head noun within the relative 
> clause.  

I only see this as applicable if there is no other available
head noun. I do not see it as awarding head noun status to a
Type 5 noun which may or may not be head noun. You ignore the
possiblities that the Type 5 noun might be outside of the
relative clause altogether, or that it may be functioning in
its Type 5 role within the relative clause:

Qe'Daq Sopbogh HoD vIlegh.

This can mean, "While I was in the restaurant, I saw the
captain who was eating." He may have been across the street
buying a newspaper while I watched him through the window." The
action of seeing occurred in the restaurant. The locative
applies to the main clause and is not relating to the relative
clause at all.

It could also mean, "I saw the captain who was eating in the
restaurant." Now the locative is acting as a locative within
the relative clause.

By your "regular" rule, it would have to mean one of two
things:

1. "I saw it in the restaurant which the captain ate." 

2. "I saw the captain in the restaurant which is being eaten."

It just depends on whether {HoD} is object of {legh} or
subject, but not head noun of {Sopbogh}. Since {Qe'Daq is right
there next to {Sopbogh} and it has a Type 5 suffix, your
regular rule insists that it must be the head noun of {Sopbogh}.

~mark is wrong. I was wrong. You are wrong. We could spend time
trying to figure out what is right, though I doubt it is worth
it. As I see it, all we have examples for is the use of Type 5
noun suffixes on nouns which have a Type 5 role in the main
clause and can unambigously serve as head nouns for relative
clauses because no other noun is available for that function.
We ignore the Type 5 suffix within the context of the relative
clause and apply it to the noun in the context of the main
clause.

This is weak and likely not worthwhile, but if you insist on
exploring it, you need to work on it more before you get it
right. You are not there yet.

> It's purpose is to mark emphasis for the matrix clause).  All
> Type-5 suffixes then behave the same way, with both types of head nouns, and no 
> special rules or interpretations need to be done. 

I insist that it does not work to suggest that any Type 5
suffix in a relative clause suggests that the noun with that
suffix is the head noun and the entire clause has the Type 5
function within the main clause. See the above counterexample.
You want things that always work. Notice that your rule does
not always work.

> Besides, we do have a canon example for each case of head noun as subject 
> and as object.  Without that object example, I'd say you might be right
> about the Type-5 suffix modifying the entire relative clause, but MO's
> {'u' SepmeyDaq Sovlu'be'bogh...} example clinches it for me.  Yes, it
> introduces some ambiguity into the language, but this is not necessarily
> a Bad Thing to me; most languages have some ambiguous structures.

If you try to stretch these examples to also cover places where
relative clauses have a second noun which might serve as head
noun, you will have gone as much too far as ~mark and I ever
did.

> I did like it back in the Good Old Days when we thought that head nouns
> could only be subjects or objects in the matrix phrase, too.  I liked
> the simplicity and regularity of it.  But I've been made aware of
> these counter-examples, also of the fact that opinion is not universal
> on ignoring them as anomalies.  Like the rest of you, I'm trying to 
> interpret them in a way that satisfies my instincts and leaves us with 
> workable rules for understanding them.

Okay, then interpret them for what they are: Examples of the
only available noun as head noun for a relative clause wearing
an inappropriate Type 5 suffix for the context of that relative
clause. That is the one common link between these two examples.
Don't try to carry it farther than that.

> -- ter'eS

charghwI'

Back to archive top level