Antonia Rothmayr - Why measure verbs are stative

The paper addresses the question why measure verbs such as German ‘wiegen’ (weigh), ‘kosten’ (cost) or ‘dauern’ (last) (and verbs that select for a PP complement such as ‘folgen aus’ (follow from)) do not allow for eventive readings. In other words, these classes of verbs always refer to a state and never denote an event.
   The answer I want to pursue in the talk is that measure verbs do not contain CAUSE in their lexical semantics. Rather, they denote a single state only. While the presence of CAUSE allows for a systematic stative/eventive ambiguity (as with verbs that undergo the instrumental alternation such as ‘obstruct’), verbs lacking the CAUSE operator receive the stative reading only.
   First, I will show that stative measure verbs have a special kind of stative reading - the Kimian stative reading (as defined by Maienborn (2003)), i.e., they do not have a Davidsonian event argument but a Kimian state variable. This fact can be detected along the following lines: Kimian statives do not allow for event related locative modifiers, nor for event related manner adverbials. In contrast, verbs like ‘obstruct’ may have both the Kimian stative and the eventive reading. The latter is created by the insertion of the DO or the BECOME operator (in the agentive or the inchoative variant, respectively). It is the CAUSE operator that enables the insertion of the two aspectual operators.
   Second, I will adress some peculiarities of the measure phrase. The measure phrase, in contrast to ordinary direct objects, does not allow for strong determiners to show up.

(1)    *The movie lasts every/the hour.

In addition, according to Rizzi (1990) extraction from wh-islands is not possible with measure phrases.

(2)    ?What did John wonder how to weigh?
    --> potatoes
       --> *200 lbs

Therefore, the measure phrase clearly has a different status than ordinary direct objects (such as ‘house’ in ‘build a house’), although both bear accusative case in German.
   In order to adequately capture the behavior of the measure phrase, I suggest that it is part of the predicate. In other words, the verbal predicate does not select for a measure argument, but contains a degree argument which must be specified explicitely. This gives the following lexical-semantic entry for a measure verb like ‘last two hours’:

(3) L(x) L(s) Ex(d) [LAST(x, d) and d = ‘2 hours’] (s)

   The fact that the measure phrase is part of the predicate is reflected syntactically, as the measure phrase, despite being carrying accusative case and occupying the complement position of the verb, is no real direct argument, hence the prohibition of strong determiners and of extraction. This contradicts the solution of Rizzi (1990) who claims that the peculiarities are due to the fact that the verb does not assign a theta-role to the measure phrase.
   In sum, the paper provides a new view regarding how stativity can be captured, and it adds a new perspective on the status of the measure phrase.

References

Maienborn, Claudia. (2003). Die logische Form von Kopula-Sätzen. Berlin:Akademie.

Rizzi, Luigi. (1990). Relativized Minimality. Cambridge Mass.: MIT Press.