LINGUIST List 5.347

Thu 24 Mar 1994

Sum: Japanese NP or PP

Editor for this issue: <>

Directory

1. , sum: Japanese np or pp

In February I posted a question wether to define Japanese NP plus P as a NP
constituent or as a PP constituent: While reading about Japanese syntax I
found that some researchers classify Japanese nominal phrase plus particle
as a marked NP and some as a PP. Of course, the difference is, wether one
assumes the nominal part or the particle as the head. Does anyone know
about the arguments in this discussion? Are there any papers on the topic,
which category must be assumed as the head? Here is the summary of the
answers. Thank you very much:

- Detmar Meurers wrote, that there is a list of publications concerning
work on HPSG and an address list in:

julius.ling.ohio-state.edu (128.146.166.200)

- Carl Pollard anwered: I am not sure there is much of an empirical
difference between NP and PP for Japanese particle-marked nominals.
But technically, if we say the particles are markers (nonheads) rather
than postpositions (heads), some things are easier to say. First,
we can explain the impossibility of stranded particles. Second, we
can explain why the particles can often be deleted when semantic role
assignment is clear from context. That is, for example, we can
let unmarked nouns be unspecified for CASE (i.e. indeterminate among
nominative, accusative, or dative), but stipulate lexically that,
say, the marker -GA marks only an unmarked nominative NP. That way,
an intransitive verb can select NP[nom] rather than the disjunction
of NP and PP[GA].

- Mayumi Masuko wrote: Most of
transformational grammar works have used NPs, rather than PPs. If one wants
to push the claim that Japanese is a head-final language strongly, however,
then it is better to have PPs, especially if one wishes to use a postposion
to specify subcategorizaion, etc. One good example of using PPs is Takao
Gunji's _Japanese Phrase Structure Grammar_, published by Reidel in 1986
(or possibly 1987? sorry I'm not sure at the moment); this is rather
similar to HPSG.

- Tsutomu Fujinami also recommended Gunji's book.

- JJ. Nakayama answered: To answer yor question as to the status of particl
s (treating them equally or case and postpositions), you want to refer t
o Shigeru Miyagawa's 1989 Academic Press book (Structure and Case Markin
g). My student wrote an MA thesis about this (on the status of NI). Basi
 cally, if it is Case, there is no categorical projection. Floating nume
ral quantifiers, cleft sentences (if it's case, it can appear at the foc
us position without the particle), and Topicalization (again if it is ca
se, it can appear without the particle). There seems to be a clear diffe
rence between Case particles and postpositions. Miyagawa discusses Float
ing quantifier and argues that NP and NQ need to be in the "mutual c-com
and" relationship in order for them to be interpreted. Of course, if the
re is the category postposition (NP P NQ), this relationship is not estab
lished. Therefore, NP and NQ cannot be interpreted together. So far I kn
ow only these three arguements. (but this has a consequence. For instanc
e, in "John went to school - John-ga gakkoo-ni itta, NI is Case not a po
stposition. So there is a difference between English and Japanese.)

- Robert Westmoreland also recommended Miyagawa and Peter Sells: "Korean
and Japanese morphology from a lexical perspective" (unpublished)

- Johanna Nicols wrote: For heads of phrases in general there are some
useful papers in G. Corbett
et al., eds., _Heads in Grammatical Theory_, Cambridge UP, 1993. The index
to the book contains some references to Japanese, but I don't know whether
the phrases you inquire about are covered. None of the papers in the book
deals primarily with Japanese, but it's a good general source for analysis
and bibliography.

- Dale Russel answered: Pollard and Sag discuss this some in Volume 2 of
their HPSG book,
"Agreement, Binding and Control" (which is supposed to be published
any day now, if it hasn't been already).

If I remember correctly, in their analysis, there are two different
kinds of phrases with the structure PP --> P NP. If the P is
contentful, then it subcategorizes for the NP, and is the head of a
head-complement structure. But if the P is purely syntactic, like the
English "to" marking the indirect object of a ditransitive, or the
"by" marking the agent of a passive, then it is a marker. In that
case, the NP specifies for the P, and the NP is the head of a
marker-head structure.


Melanie Siegel
Universitaet Bielefeld
Fak.f.Linguistik und Literaturwiss.
Postfach 100131
33501 Bielefeld
e-mail: siegellili7.uni-bielefeld.de