SIbI' jIjangbe' 'e' vIpay 'a qaStaHvIS wej jajvam Seng vIDIghtaH.

jIjatlh:

the locative (or for that matter causative) noun of the relative clause can be the object of the main clause, couldn't it be the subject as well? Then one could say Saq DujDaq jIHaw'bogh The ship on which I fled landed. What do you make of this?

jatlh SuStel:

If this were correct, then yes, it could do that.

This reverses the peculiarity of meQtaHbogh qachDaq Suv qoH neH in which a noun with a Type 5 suffix other than -'e' can be the subject of a verb+bogh. I don't expect anyone to feel particularly comfortable with this, though that doesn't argue against it.

jIjatlh:

I see another major problem, besides the main verb being able to "see" into the relative clause, and one which is not totally unrelated to the problem of whether a preceding modifier acts on a relative clause preceding the main clause or on the main clause. (Example: vengDaq jIlwI' ghaHpu' loD'e' vIlegh Does it mean I saw the man who had been my neighbor in the city or In the city, I saw the man who had been my neighbor?) The example you use has only one possible candidate for head noun. If there's only one candidate for head noun, it could conceivably fill some other slot in the main clause, but if there are more than one, which gets priority? (Alan's post I referred to above suggests a way around this dilemma in at least some cases.)

jatlh SuStel:

You forgot the -bogh in the example. This is a poor example, because as a "to be" sentence the final noun requires an -'e', automatically making it the head noun. You have no choice.

HIvqa' veqlargh!

jatlh SuStel:

A more interesting example would be how to say I see the ship in which the captain fled. If you try DujDaq Haw'pu'bogh HoD vIlegh, how would you know that DujDaq, and not HoD, is supposed to be the head noun? You wouldn't.

And head nouns are normally subjects and objects, so in the above I'd assume the speaker sees the captain. Changing the above to DujDaq Haw'pu'bogh HoD vI'or causes HoD to make no sense as the head noun, but on the other hand, why should the identification of head noun depend on the wording and not on sentence construction? So I would not expect that to be the way to say "I pilot the ship the captain fled in." It should always to be possible to construct a sentence with a relative clause so that the head noun is clearly indicated, even if all the all the root words are blanked out.

The more unusual relative clauses in Klingon (canon or speculative) remind me of nonstandard relative clauses in some English dialects, e.g. "He left the shop to his son, which all he wanted was drink." This makes me envision a relative clause as an appended sentence with -bogh, but of course why bother with the -bogh? Just tack a short sentence on, as you have suggested.

Your example in which you translate the first sentence of War of the Worlds illustrates that there is no need to translate sentence-for-sentence, though I don't understand your use of Dor in DIS poH wa'maH Hut Dor "in the last years of the nineteenth century;" I'd say DortaHvIS DIS poH wa'maH Hut.

jIjatlh:

Lastly, (and I am confident that you won't like this, so please do not let this derail the thread) but your argument that a noun can have another syntactic role in the main clause because it's the only candidate for head noun of a relative clause reminds me of an argument I made for the combination of {-lu'} and {-wI'} to nominalize on the object because it is the only candidate, the subject being pointedly moot.

jatlh SuStel:

My argument is not that the locative noun of the relative clause must be the head noun because no other noun can be; my argument is that the locative noun can be RECOGNIZED as the head noun because there is no other noun distracting you from that conclusion. You have to first accept the notion that some noun other than the subject or object of the relative clause can be the head noun.

I don't accept the argument that -lu' + -wI' must nominalize the object because there's no subject available. -wI' never nominalizes the object. There is no rule, no example, that supports any evidence at all that -wI' can nominalize objects. IF there were such a rule, then I might consider the idea that, lacking a subject, it would nominalize the object. But there is no such rule, and no evidence that any such rule exists.

I agree wholeheartedly that the evidence does not lead to the conclusion that -lu' + -wI' nominalizes the object. I merely like the idea, since it would be handy.
If it were a horse race, I'd put my $2 down on it, but for all I know, someone else might win big betting on the longshot one who does ___ while wearing a pink tutu.

~mIp'av