We discussed here how in sentences like (1) the wh-word kuda is not extracted but is in situ:
(1) Ivan kuda idet?
Ivan where goes
Where is Ivan going?
Now how about sentence (2):
(2) Ivan skolko chitaet knig?
Ivan how.many reads books
How many books is Ivan reading?
The extraction is now not from the verb, but the head daughter of the filler-gap rule, if we were to apply it here, is still the VP reads books, so it doesn’t seem possible to apply it, just like in (1)? But this is a determiner, and so it is hard to imagine that it is in situ here?
Example (2) suggests that the generalization that the extracted element must be left-most is not correct. I guess the next questions are whether (a) Ivan and (b) skolko are in fact involved in long-distance dependencies. Can you construct examples where the thing before skolko comes from a lower clause? If not, is it always necessarily the subject of the same clause, or can it be a non-subject argument? Can you construct examples where an extracted determiner belongs to a lower clause (with or without something before it)?
Like this? (Punctuation is obligatory in all complex clauses in Russian, including in ``I know what Ivan reads’’)
(6) ??Ivan skolko, ty znaesh, (chto) prochital knig?
Ivan.NOM how.many 2SG know.2SG.PRES (that) read.MASC.PAST books.PARTITIVE
how many books do you know that Ivan read?
It’s pretty weird but I think is possible. It is slightly better with the that, I think. (7) is a pretty normal sentence.
(7) Ivan, ty znaesh, skolko prochital knig?
Ivan.NOM 2SG know.2SG.PRES how.many read.MASC.PAST books.PARTITIVE
How many books do you know Ivan read?
(I was wondering whether I should translate this as “As for Ivan, do you know how many books he read?” but decided that not necessarily; there is not necessarily such a special emphasis on Ivan here. It can even be deemphasized.)
Hmm I think that (7) actually sounds better with the translation: ``Do you know how many books Ivan read?’’ rather than “How many do you know that he read books”. For (6), that translation is not available, and it is noticeably weirder.
(7) looks like it topic/focus fronting, with “Ivan” extracted out of the lower clause. Given that that’s possible, perhaps (2) could be analysed as having both topic/focus fronting and wh fronting.
I understand that this would be consistent with some Minimalist work on Russian (Baylin, p.c.). It sounded to me that in general, whenever there is a wh-word, they want everything in front of it to have had moved also, perhaps as a result of topicalization.
I’d like to clarify the (somewhat squishy) data.
First, I’d like to get rid of the possible parenthetical reading (especially because the Russian punctuation makes it look like English parenthetical).
Suppose the situation is that Ivan visited many countries. So many that it is kind of hard to believe. Now, maybe previously Sandy said that Ivan visited 140 countries. But later Sandy added that in fact Ivan himself had said he only visited 135, and later Kim said something about 140 but that is less reliable. Now we are talking to Sandy and we want to establish how many it is that Ivan visited for sure.
(8) How many countries do you know for sure that Ivan visited? [eng]
In Russian, I can maybe say this in the following way:
(9) ?Ivan skolko, ty tochno znaesh, chto posetil stran?
Ivan.NOM how.many 2SG for.sure know.2SG that visited countries.PARTITIVE
"How many countries do you know for sure that Ivan visited?" [rus]
I think that (9) may look strange out of context but I also believe that it is entirely possible (and hopefully it is less confusing than (6) because it is clear that the “you know” thing is not parenthetical).