No subject

[Stefan Knoob]

Dear Roger,

I think we should also consider processing issues of incomplete vs complete
low level constituency.
Thus, in German, I think one must distinguish two cases:

1. The case where only low-level components of the verb phrase intervene.
This is the case in the above present perfect sentences, where 'hat' has
already begun a complex node that remains incomplete until the object (if
there is one), adverbs and finally the participle have been added. In such
cases, at least according to my native intuition, the "extraposition" seems
to be the statistically (distributionally) unmarked alternative, in the
sense that I feel I am more likely to utter the extraposed construction
Er hat das Buch schon gelesen, das Lisa ihm gestern gekauft hat.
he has the book already read that Lisa him yesterday bought has
than the presumedly canonical
Er hat das Buch, das Lisa ihm gestern gekauft hat, schon gelesen.
he has the book that lisa him yesterday bought has already read

The same goes for sentences with negativ predicates such as
Ich mag das Buch nicht, das Lisa mir gestern gekauft hat. (statistically
unmarked)
I like the book not that lisa me yesterday bought has
Ich mag das Buch, das Lisa mir gestern gekauft hat, nicht. (statistically
marked)
I like the book that lisa me yesterday bought has not

2. The case where the intervening elements include other constituents as
well.

Er hat das Buch seiner Mutter gegeben, das Lisa ihm gestern gekauft hat.
(statistically marked)
Er hat das Buch, das Lisa ihm gestern gekauft hat, seiner Mutter gegeben.
(statistically unmarked)

Also, I think one must consider the diachronic development that has lead to
the peculiar position of the direct object in the German perfect between
auxiliary and participle. Thus, it originates in the structural reanalysis
of constructions with resultative adverbials such as the English:

He eats tomatoes raw.
He eats tomatoes cooked.

German
Er isst Tomaten roh.
Er [isst [Tomaten]] gekocht.
Er [hat Tomaten] gekocht.
Er [hat [Tomaten] gekocht].

In fact can still find the original construction in sentences such as:
Wir [[haben Tomaten] nur gekocht].
'We have Tomatoes only cooked.' (We don't have them raw)

Although diachrony is usually not considered relevant for issues of
synchronic structure, I wonder whether there is not a remnant competing or
secondary attributive relation between participle and the noun whose state
it expresses (cf also the now disappearing French gender marking on the
participle)
Note here that the unaccusative perfect shows similar markedness effects for
relative clause position

Ich bin auf dieselbe Schule gegangen, auf die schon meine Mutter gegangen
ist. (statistically unmarked)
I am on the.same school gone, on which already my mother gone is.
'I have been (or went) to the same school that my mother already went to.'

Ich bin auf dieselbe Schule, auf die schon meine Mutter gegangen ist,
gegangen. (statistically marked)

In this context, you might consider the following type of Korean and
Japanese data.
Both Korean and Japanese don't have relative clauses in a narrow sense.
Instead they use what is for Korean usually called a "modifier clause" which
always occur before the noun. In Korean then the verb is in a so-called
"modifier form" or "prenominal form", in Japanese prenominal forms are
identical to the simple (informal/non-polite) verb forms. The modifier
clause structurally resembles Indo-European participial clauses such as "a
forgotten author", but they may contain any number and kinds of nominal
constituents, and the modifier form does not change 'argument structure'.

Korean:
chayk-ul ilk-nun salam
book-OBJ reads-PREN.PROGESSIVE person
'person reading book'

chayk-ul sa-n salam
book-OBJ buys-PREN.PERF person
'person who bought book'

Interestingly, these also can have intervening material as long as it also
modifies the noun

Korean
chayk-ul sa-n ku salam
book-OBJ buys-PREN.PERF that person

chayk-ul sa-n Cinhuy tongsayng
book-OBJ buys-PREN.PERF Chinhee younger.sibling
Chinhee's younger sibling who had bought a book

namca chinkwu-lul kitali-nun yeppun akassi-tul
man friend-OBJ waits-PREN.PROGR is.pretty-PREN.PERF young.woman-PLUR
'pretty girls waiting for their boyfriends'

Japanese
hon-o kat-ta hito
book-OBJ buys-PERF that person

hon-o kat-ta Hanako-no imooto
book-OBJ buys-PERF Hanako-POSS younger.sister
Hanako's younger sister who had bought a book

karesi-o mat-teiru kawaii onna-no hito
boyfriend-OBJ waits-PROGR is.cute-PRES woman-POSS person
'pretty girls waiting for their boyfriends'

Here, different sequences show different scope effects

Chayk-ul san Cinhuy tongsayng
Chinhee's younger sibling, who has/had bought a book

Cinhuy-uy chayk-ul sa-n tongsayng (statistically marked)
Chinhee-POSS that book-OBJ buy-PREN.PERF person
the younger sibling of Chinhee who has/had bought a book (and not the one
who hadn't)

In the light of these Korean/Japanese fact one should consider possible
dependency relations of participle to noun vs relative clause to noun in
German as well

_________________________________________________________________
Use MSN Messenger to send music and pics to your friends
http://www.msn.co.uk/messenger