Biblical Hebrew Forum

[peter_kirk AT SIL.ORG]

At risk of seeming to show off my knowledge of languages (not much 
wider than in this post, in fact!), let me answer some of Dave's and 
Doug's points.

I don't know any Korean. But I do know that it is structurally very 
similar to Turkic languages. Turkish has no relative pronoun, and thus 
I can believe that Korean has none either. Instead, Turkish has a 
variety of constructions using embedded phrases with non-finite or 
participle verb forms. These can be used to do almost everything you 
can do with relative clauses in Indo-European languages, though 
because of the head-last structure it gets very confusing for us 
westerners. (Koreans learn Turkish incredibly fast! Everything is the 
right way round for them.)

Azerbaijani is closely related to Turkish, but its structure has been 
more influenced by Farsi, and it has borrowed a complementiser "ki" 
(probably related to the Hebrew KIY via Arabic) introducing real 
subordinate clauses with finite verbs, which is widely used in 
colloquial speech though deprecated in writing. Relative clauses 
closer to the Indo-European pattern can be formed using "ki", but 
closer inspection reveals a pattern closer to the Hebrew one e.g.

ev    ki,    adam ondadir
house which, man  in-it-is

ev    ki,    biz onda  yashayiriq
house which, we  in-it (we-)live

The comma indicates the phonological break (and is required in 
writing) and suggests also the clause boundary. The subordinate clause 
is in fact a fully independent clause with no shift of components, 
trace theory does not apply. "ki" could be replaced by a full stop 
(period) with no other changes and in written language it often is.

Russian also offers a phenomenon comparable with another side of the 
Hebrew one. Alongside the regular relative "kotoryy" there is a 
construction "tot, chto" or "tot, kto", equivalent to English "that 
which" or "he who". This construction consists of two elements, which 
must always be separated by a comma (indicating the phonological 
break). Each element is a separately declinable pronoun; one is part 
of the main clause and the other part of the subordinate clause. The 
main clause element "tot" may be adjectival or nominal, when 
adjectival it is shifted before the noun. But in this case the element 
in the subordinate clause is fronted, only a preposition may come 
before it.

tot  dom,   v  chom  chelovek
that house, in which person (is)

tot,        v  chom  chelovek
that (one), in which person (is)

u  togo,       v  chom  chelovek
at that (one), in which person (is)

Interestingly, the form "potomu, chto" "according to that, which" can 
be used, but this generally becomes merged into an expression in which 
the comma and phonological break is shifted forwards, ", potomu chto" 
(pronounced as one word) "because". This clear case of shift of a word 
from the main clause to the subordinate might interest you 
generativists.

Then to get back to Hebrew: It seems to me that Hebrew combines 
elements of the Azerbaijani and the Russian constructions. ):A$ER 
corresponds to the Azerbaijani "ki" or the Russian "tot" in that it 
goes with the main clause, and with the Russian "tot" in that it can 
be nominal or adjectival and (when nominal) can be governed by 
prepositions which relate to the main clause. The subordinate clause 
(excluding ):A$ER ), like the Azerbaijani subordinate clause following 
"ki", is a sentence in itself with no shift of ordering. Hence my 
questions about verb forms, which might help to confirm my view. By 
the way, I would not expect wayyiqtol forms here for two reasons: one 
is that the conjunction is superfluous, and the second is that there 
is no sequentiality. Doug's first data suggests that ):A$ER + qatal is 
pluperfect like X + qatal, but then there are few examples of sentence 
initial qatal to compare with. I would be interested to see if ):A$ER 
+ yiqtol is more likely to be jussive (like sentence initial yiqtol) 
than ):A$ER + X + yiqtol - that could help to confirm my hypothesis.

I hope that helps to explain where my idea is coming from. I will leave 
it to experts like Dave to fit this data into particular theoretical 
models. I am just supplying some data which may confirm or break those 
models.

Peter Kirk

PS to Dave: You quoted the sentence

(9) HABBAYIT BA):A$ER HF)IY$ (the house in which the man (is) )

as if it is a valid sentence. I had taken this to be invalid, or at 
least to have the rather odd meaning "the house in the one which the 
man is". Does this type of construction actually occur with the 
meaning "the house in which the man is"? Can you give references? You 
originally asked, "So I'm wondering why an author would sometimes 
choose to move the preposition with the 'asher, and sometimes choose 
not to and put the preposition with an anaphor instead." To rephrase 
my previous question, are there actual cases in the Hebrew Bible in 
which you would say that the preposition has moved with ):A$ER, and 
the preposition clearly relates to the subordinate clause and not to 
the main clause? This is important, because the existence of such 
sentences is the only justification for your assertion: "Of course, a 
side-effect of all this is that Peter's idea of ):$ER being part of 
the main clause is unlikely because either way it's a moved 
constituent from the "lower" clause". My argument would be that ):A$ER 
is not a moved constituent at all. In English we can write "the house 
which the man is in" or (pedantically) "the house in which the man 
is", because "which" is a moved constituent; you are suggesting that 
the equivalent two possibilities exist in Hebrew, but I doubt if the 
second one actually exists.