object-like adverbials

[bingfu]

in Dec. 17, Krifa's wrote:
It appears that the motivation for the object-like marking of delimiting
adverbials is motivated by the fact that they, too, "measure out" an event
predicate.

I agree with Krifka's explanation why the durative tends to pattern with
the object.  This is particlarly clearly illustrated in Chinese.  In Chinese, the
durative can be incorporated into object and behaves exactly like
the numeral in objedt NP.

	ta kan-le san-ben shu
	he read   three-Cl book
	"he read three books"

	ta kan-le san-tian (de) shu
	he read  three-day Mod. book.  (Mod. modification marker)
	"he read books for three days."
	lit. "he did three-day's book-reading."

Further, it seems that the numeral in the indefinite object and the durative
are distributionally complimentary to each other.  The two can be
not be co-occur, as shown below.

	*ta kan-le san-tian san-ben shu
	 he read  three-day three-Cl book
	 "he read three books for three days."
	(Is this sentence acceptable in English?)

However, if the object is definite, the sentence would be fine:

	ta kan-le zhe san-ben shu san-tian
	he read  the  three-Cl book three-day
	"He read these three books for three days."

When the time expression is definite, it would be OK for
the object to be indefinite.  But, now the time expression have to
precede the verb, behaving like an subject (as many Chinese
grammarians claim).

	ta zhe san-tian kan-le san-ben shu.
	he the three-day read three-Cl book.
	"He read three books in these three days."

Interestingly, though Chinese duratives can take
an modification marker 'de' and behave like NP
internal modifiers, 'frequency" expressions like
"three times" cannot do this way.

	ta kan-le san-ci      (*de) shu
	he read three-time Mod book
	"He read books three times."

This is I wonder the marking of frequency in other languages in
my last posting.

For detailed descriptions of Chinese duratives, see
Yen-Hui Audrey Li 1990 "Order and constituency in Mandarin Chinese"
Kluwer Academic Pub.

Bingfu Lu




Date: Thu, 17 Dec 1998 11:29:05 +0100
From: Krifka <krifka at MAIL.UTEXAS.EDU>
To: LINGTYP at LINGUIST.LDC.UPENN.EDU
Subject: object-like adverbials

1. What is remarkable about the cases of object-like adverbials mentioned in
the discussion so far is that they are all adverbials that specify the
length of an event (sometimes called "durative" adverbials). One can add to
these cases German (the example given by Sasse does not show case markering,
but the following examples do):

(i)  Ich arbeitete den ganzen Tag
      I      worked    the.ACC whole.ACC day

(ii) Ich schob den ganzen Tag einen Wagen.
     I     pushed the.ACC whole.ACC day a.ACC cart

Accusative-marked adverbials in Korean, with the same interpretation, were
the subject of a recent paper by Wechsler and Lee, in Natural Language and
Linguistic Theory 1997.

2. It is tempting to speculate that durative adverbials and object NPs have
a common property that motivates why they are marked in the same way.
Durative adverbials always delimit an event predicate (i.e. form a telic
predicate out of an atelic predicate), and object NPs very often do that (if
the object NP itself expresses a delimited quantity). Taking time-frame
adverbials ("in"-adverbials) as a test for delimitation, we have:

(iii)   eat an apple in ten minutes (vs. *eat in ten minutes)
        write a letter in ten minutes
        read a letter in ten minutes
        (but, e.g., *push a cart in ten minutes)

Verkuyl, Tenny, Krifka and others have analyzed the role of the object NP in
these cases (e.g., Tenny says that the object "measures out" the event).
Dowty (in Language 1991) has a property called "incremental theme" that is a
factor for the realization of an event participant as an object NP. It
appears that the motivation for the object-like marking of delimiting
adverbials is motivated by the fact that they, too, "measure out" an event
predicate. This can be made precise in the framework I proposed (e.g., in
Lexical Matters, ed. Sag/Szabolcsi, Stanford 1992): In a case like "den
ganzen Tag arbeiten", 'work the whole day', there is a correspondence (a
homomorphism) between the parts of the working event and the parts of the day.

3. This said, I should point out that colloquial German also has
accusative-marked adverbials that just locate the event in a time interval,
without measuring it out.

(iv)    Diesen Monat haben sie geheiratet.
         this.ACC month have they married
        "they married this month"

Best,