A Direct and Entailment-based Linking Theory

advertisement
A Direct and Entailment-based Linking Theory
All the tests for unaccusativity are probably subsumed by the ergative test (Levin &
Rappaport 1988, Farrell 1994) and the past participial test of Bresnan 1982. The ergative test
provides evidence that subjects of (di)transitive verbs and of intransitive ones are truly agentive
(initial 1s in RG or the ‘external subject’) if they allow the formation of –er nouns. I have
adapted this test as also allowing the paraphrase ‘x is the verb-er’, with x being the participant
whose role is at issue. If such a paraphrase is semantically sound and if it yields a true
proposition as derived from the original sentence, the x participant is an initial 1 (= the verber).
From now on, I will refer to the ergative test as the ‘verber’ test since this test picks out the
agentive role linked with a given predicate. On the other hand, Bresnan’s past participial test
provides evidence that an x participant is an object (an initial 2 in RG or an ‘internal object’) if
the paraphrase ‘the verbed x’ is semantically sound. I will refer to Bresnan’s past participial test
as the ‘verbed’ test. The notation ‘#’ indicates that the string is uninterpretable or it is
semantically odd, at least as a paraphrase derived from the sentence in question.
(1)
a. Margarita is the returner.
b. #The returned Margarita.
c. #The blouse is the returner.
d. The returned blouse.
The verber is cast as the subject and the verbed is cast as the object. Observe that the
returned Margarita is a semantically sound string derived from a predicate like Margarita
returned to town, but it is not a correct entailment derived from Margarita returned the blouse,
because in this predicate she is the returner, not the returned.
The ‘verber’ and ‘verbed’ tests also give correct results when applied to intransitive or
intransitivized predicates:
(2)
a. Rose works.
b. Rose is the worker.
c. #The worked Rose.
(3)
a. The flowers wilted.
b. #The flowers are the wilter.
c. The wilted flowers.
Notice that 2b is semantically sound. Nobody disputes the existence of the noun worker. As for
2c, no native speaker will probably utter the string ‘the worked Rose.’ The ergative test (=the
verber test) shows that Rose is the verber in 2a. In 3b, the string ‘the flowers are the wilter’ is not
a semantically sound proposition derived from 3a, but the string ‘the wilted flowers’ is. The
grammatically and interpretability of 3c indicate that the flowers are the verbed. ‘The flowers
wilted’ is an unaccusative or unaccusativized predicate and the verbed test shows a subject that is
not a verber but a verbed. Thus, all verbers are subjects but not all subjects are verbers, and the
labels ‘verber’ and ‘verbed’ do not always coincide with subject and direct object, respectively,
the same way that ACTOR and UNDERGOER do not necessarily coincide with subject and
object (Foley & Van Valin 1984, Van Valin & LaPolla 1997). The same reasoning applies to
unaccusativized predicates like the ice melted (from the heat melted the ice), the river froze
(from a drop in temperature froze the river), taxes increased (from the government increased
taxes), and of course the truly unaccusative the mail arrived (cf. the arrived mail has not been
processed vs. #the mail is the arriver).
Consider this famous and troublesome pair for linking theories:
(4)
a. John feared the article.
b. The article frightened John.
In both of these sentences, John is assigned the role of Experiencer and the article the role of
Experienced or Theme. However, since John is the subject in the first sentence but the direct
object in the second, at least one problem relevant to the issue at hand arises. Any thematic role
hierarchy must recognize a misalignment of roles: if Experiencer precedes Theme for (a), it will
follow it for (b). This misalignment manifests itself in the famous split known as ‘object of
emotion’ vs. ‘cause of emotion’ (Pesetsky 1987, 1995) or ‘Experiencer-subject’ vs. ‘stimulus
subject’ (Croft 1993, Dowty 1991). By applying the verber and verbed tests, we can see that in
4a John is the fearer (=verber) and the article is the feared (=verbed). In 4b, the article is the
frightener and John is the frightened. The roles of verber and verbed are aligned with subject and
object in both sentences.
This paper will present evidence to show that the verber and verbed are the only roles
needed for the linking of intransitive and monotransitive predicates, without the need for a
thematic role hierarchy or the computation implied in assigning proto-roles (Dowty 1991) or
macro-roles (Foley & Van Valin 1984, Van Valin & LaPolla 1997). The linking theory proposed
will show a more direct, predictive, and straightforward semantics/syntax interface. This is an
entailment-based theory of linking (Ladusaw & Dowty 1988), able to account, without adding
any stipulations, for many verbs listed as exceptions (undergo, suffer, inherit, receive, comprise)
in all mainstream current theories of linking, including that of Dowty 1991, Levin and Rappaport
Hovav 1995, Wechsler 1995, Van Valin & LaPolla 1997, and Davis 2001. This proposal also
suggests that it is predicates rather than verbs that a theory of linking should be concerned with.
In this way, we can avoid the misleading labeling of break, freeze, and melt as intransitive verbs.
Those are transitive verbs that can readily appear in intransitive predicates. Those verbs are
unaccusativized. This proposal will also show the need to recognize not only that while only a
handful of verbs are unaccusative (truly intransitive), a countless number of verbs can be
unaccusativized.
References
Bresnan, Joan. 1982. The passive in lexical theory. The mental representation of grammatical
relations, ed. by Joan Bresnan, 3-86. Cambridge, Mass: MIT Press.
Croft, William.1993. Case marking and the semantics of mental verbs. Semantics and the lexicon, ed. by James
Pustejovsky, 55-72. Dordrecht: Kluwer.
Davis, Anthony. 2001. Linking by types in the hierarchical lexicon. Stanford: CSLI.
Dowty, David. 1991. Thematic proto-roles and argument selection. Language 67.547-619.
Farrell, Patrick. 1994. Grammatical relations and thematic roles. New York: Garland.
Foley, William A. and Robert D. Van Valin, Jr. 1984. Functional syntax and universal grammar. Cambridge UP.
Ladusaw, William A., and David R. Dowty. 1988. Towards a nongrammatical account of thematic roles.
In Syntax and semantics 21: Thematic relations, ed. Wendy Wilkins. San Diego, Academic Press.
Levin, B., and M. Rappaport. 1988. Non-event –er nominals: A probe into argument structure.
Linguistics 26.1067-83.
Levin, B., and Malka Rappaport Hovav. 1995. Unaccusativity. At the syntax-lexical semantics interface.
Cambridge, Mass: MIT Press.
Pesetsky, David. 1987. Binding problems with Experiencer verbs. Linguistic Inquiry 18.126-40.
____. 1995. Zero Syntax. Experiencers and cascades. Cambridge, Mass: MIT Press.
Van Valin, Robert D., Jr., and Randy J. LaPolla 1997. Syntax. Structure, meaning and function.
Cambridge, UK: Cambridge UP.
Wechsler, Stephen. 1995. The Semantic Basis of Argument Structure. Stanford: CSLI.
Download