d. It occurs with a predicate featuring special verbal morphology (attribu-
tive form).
e. A nominative NP with adverbial particle attached to it cannot be a target.Numerous analyses have been proposed to account for this peculiarity in NGC.
The central issue is to determine what is responsible for genitive case marking.
There are two different views on this. One attributes genitive case marking to the
nominal head outside the prenominal clauses on the basis of the fact that genitive
case generally appears within noun phrases in Japanese. Adherents of this view
have advanced various means of associating a clause-internal element with an
external nominal head: the optional application of the restructuring rule (Bedel
1972); overt or covert movement of the relevant element into the domain of the
external head (Miyagawa 1993, Sakai 1994); feature-checking/Agree (Ochi 2001,
Maki and Uchibori 2008; Miyagawa 2012, 2013); or scrambling of the relevant
element to the clause initial position, allowing the external head to govern and
case license it (Fukui and Nishigauchi 1992).
When considering (8d) rather than (8b) as a fundamental property of NGC,
an alternative view argues that the genitive case marking is irrelevant to the
presence of the external nominal head. Thus, Watanabe (1994, 1996) claims
that genitive case is a consequence of wh-agreement within the attributive clauses.
Hiraiwa (2001) argues that the attributive form is a realization of the Agree
relation of C-T-v-V, and that genitive case marking is a consequence of the
Agree relation.
Given that genitive case is assigned to any nominal elements within Japanese
noun phases, attributing the genitive case to the external nominal head is more
natural and appropriate. In addition, the latter view does not clarify why no
case other than genitive emerges. Therefore, we develop the former view in our
proposed theory of case valuation. However, Watanabe and Hiraiwas’s insight
into genitive case marking that is associated with the attributive form is also
important for understanding the nature of NGC. To see the relevance of the
attributive form to NGC, it is instructive, following Hiraiwa (2001), to glance
at the historical development of attributive forms.
Old Japanese featured a clear distinction between attributive and conclusive
forms: the former was used only in certain types of subordinate clauses and
the latter only in matrix clauses. Examples of subordinate clauses in which
predicates take an attributive form are given below: (10a) is an example of a
relative clause, (10b) an internally headed relative clause, and (10c) a (nominal)
complement.
(10) a. [tubakurame-no motitaru] koyasugai
swallow-GEN have cowry
‘a cowry that the swallow has’ (Taketori-monogatari, cited in Kinsui
1995)
b.[Kogimi tikau husi-taru]-o okosi-tamahe-ba...
Kogimi near lay=asleep-ACC wake-HON-asCase and predicate-argument relations 57