Download Ergativity of Nouns and Case Assignment

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Malay grammar wikipedia , lookup

Udmurt grammar wikipedia , lookup

Lithuanian grammar wikipedia , lookup

Macedonian grammar wikipedia , lookup

English clause syntax wikipedia , lookup

Ukrainian grammar wikipedia , lookup

Arabic grammar wikipedia , lookup

Old Norse morphology wikipedia , lookup

Causative wikipedia , lookup

Modern Greek grammar wikipedia , lookup

Inflection wikipedia , lookup

Old Irish grammar wikipedia , lookup

Zulu grammar wikipedia , lookup

Chinese grammar wikipedia , lookup

Swedish grammar wikipedia , lookup

French grammar wikipedia , lookup

Kannada grammar wikipedia , lookup

Japanese grammar wikipedia , lookup

Navajo grammar wikipedia , lookup

Portuguese grammar wikipedia , lookup

Kagoshima verb conjugations wikipedia , lookup

Esperanto grammar wikipedia , lookup

Old English grammar wikipedia , lookup

Romanian nouns wikipedia , lookup

Scottish Gaelic grammar wikipedia , lookup

Modern Hebrew grammar wikipedia , lookup

Spanish grammar wikipedia , lookup

Georgian grammar wikipedia , lookup

Icelandic grammar wikipedia , lookup

Polish grammar wikipedia , lookup

Turkish grammar wikipedia , lookup

Lexical semantics wikipedia , lookup

Latin syntax wikipedia , lookup

Ancient Greek grammar wikipedia , lookup

Yiddish grammar wikipedia , lookup

Serbo-Croatian grammar wikipedia , lookup

Pipil grammar wikipedia , lookup

Transcript
Ergativity of Nouns and Case Assignment
Author(s): Natsuko Tsujimura
Reviewed work(s):
Source: Linguistic Inquiry, Vol. 21, No. 2 (Spring, 1990), pp. 277-287
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/4178672 .
Accessed: 25/05/2012 20:56
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.
http://www.jstor.org
Remarks
and
Replies
Ergativity of Nouns and Case Assignment
Natsuko Tsujimura
1. Introduction
Grimshawand Mester (1988) deal with the interestingissue of 0-markingin regardto
complex expressions consisting of a noun followed by the verb suru in Japanese.' A
representativeexample of what they discuss is in (1), in which KEEKOKUis a SinoJapanese compound noun. The noun KEEKOKUcan be assigned accusative case by
suru as in (la), or it can be incorporatedinto the verb suru as in (Ib), forminga complex
predicate.2
KEEKOKU-o sita.
John-wa murabito-niookami-gakuru-to
John-Topvillager-to wolf-Nom come-Compwarning-Acc suru-Past
'John warnedthe villagersthat the wolf was coming.'
b. John-wa murabito-niookami-gakuru-to
KEEKOKU-sita.
John-Topvillager-to wolf-Nom come-Compwarning-suru-Past
'John warnedthe villagersthat the wolf was coming.'
(1) a.
Grimshawand Mester identify the verb suru in (la) as a light verb and claim that the
light verb has an empty argumentstructurein its lexical entry. When the light verb
cooccurs with a nominalthat has argumentstructure,the 0-role of the nominalcan be
transferredto the light verb, whereby the verb acquires0-markingability. They call this
I would like to thank Beth Levin for her extremely valuablecomments and suggestionson an earlier
version of this article.The finalversionhas greatlybenefitedfromthe commentsprovidedby two anonymous
LI reviewers. The researchfor this articlewas partiallysupportedby IndianaUniversitythrougha Summer
Faculty Fellowshipawardedto the author.
' Many of the nouns that occur in this constructionare Sino-Japanesecompoundnouns (that is, compounds of Chinese origin).However, nouns of Japaneseorigin, such as hanasi 'talk', are not excludedfrom
appearingin the construction.
2 FollowingGrimshawand Mester, the 0-markingnoun will be capitalizedin the examples throughout
this article.
Linguistic Inquiry, Volume 21, Number 2, Spring 1990
277-287
g 1990 by The Massachusetts Institute of Technology
277
278
REMARKS
AND
REPLIES
process Argument Transfer.Another property of the light verb suru is that it assigns
accusative case. The lexical entry of suru is given in (2).
(2) suru, V; (
) (acc)
The empty space in the parenthesesindicatesthatthe verb's argumentstructureis empty,
and (acc) shows the verb's ability to assign accusative case.
Grimshawand Mester furtherdiscuss three crucialgeneralizationspertinentto Argument Transfer. Of the three generalizations,the following two are relevant to our
discussion: (i) at least one argumentapartfrom the subject must be transferred,and (ii)
the subject argumentmust always be transferred.Assuming that argumentstructureis
hierarchicallyorganized,the externalargumentbeing externaland the internalargument
being internal, Transferoccurs in a top-down fashion. It follows that if an internalargument is transferred,then the external argumentis transferredas well. The process of
ArgumentTransferinvolved in the formationof (la) is illustratedin (3).
(3) a. KEEKOKU(Agent (Goal (Theme)))
) (acc)
b. suru (
) + suru (Agent (Goal (Theme)))(acc)3
c. KEEKOKU(
Since the internalarguments(Goaland Theme)are transferredto the argumentstructure
of suru, the external argument(Agent) is also transferred.Suru now has the ability to
assign a 0-roleto the Agent, Goal, and Theme argumentsin additionto its originalability
to assign accusative case.
Miyagawa(1989) correctly points out that Grimshawand Mester have overlooked
the fact that not every noun plus suru displays the configurationsshown in (la) and (lb).
For example, compare (4) (= Grimshawand Mester's (19c)) and (5).
(4) ??John-wa butyoo-ni
SYOOSIN-o sita.
John-Topsection chief-to promotion-Accsuru-Past
'John obtaineda promotionto section chief.'
John-wa butyoo-ni
SYOOSIN-sita.
(5)
John-Topsection chief-to promotion-suru-Past
The sentence in (4), where SYOOSINappearswith accusative case, is extremely awkward, whereas (5), where the noun is incorporatedinto the verb suru, is perfectly acceptable. There is a class of noun plus suru complexes that allows only the type of
configurationshown in (lb). Miyagawaexplains the ungrammaticalityof (4) based on
the syntactic behavior of numeral quantifiers. He claims that the noun SYOOSIN is
ergative and accordingly bears a 0-role for the internal argument, which is the sole
argumentof the noun. Thus, the noun fails to bear a 0-role for its grammaticalfunction
3 It is not clear from Grimshaw and Mester's discussion whether the internal argument remains internal
and the external argument remains external after Transfer. In this article, I will assume that the status of the
argument (external or internal) is also inherited as a result of Transfer.
REMARKS
AND
279
REPLIES
(GF) subject. However, if the light verb assigns accusative case to the noun as in (4),
it clearly violates Burzio's Generalization.Burzio's Generalization(Burzio(1981;1986)),
which is shown in (6), states that if a given verb does not assign a 0-roleto its GF subject
slot, then it will not assign accusative case to its object.
(6) - Os-
-A
The complex predicatescomprisinga noun and suruthat exhibitthe configurationshown
in (5) but not the one shown in (4) (that is, those exactly like SYOOSIN-suru)include
the following:
(7) SEETYOO-suru
KAKUDAI-suru
TINBOTU-suru
SYUKUSYOO-suru
ZENSYOO-suru
'grow'
'enlarge'
'submerge'
'reduce'
'burndown'
KOOZYOO-suru
ZYOOSYOO-suru
GYOOKO-suru
KANSOO-suru
'improve'
'go up'
'solidify'
'dry'
Hence, Miyagawaattributesthe ill-formednessof (4) to the violation of Burzio's Generalization.The identificationof nouns such as SYOOSINas ergative is overlooked by
Grimshawand Mester, who instead account for the ill-formednessof (4) as a "lexical
gap." Nevertheless, Miyagawa'sanalysis is compatiblewith their process of Argument
Transferand moreover provides supportfor Burzio's Generalization.
In this articleI will demonstratethat the resultativeconstructionin Japanese, taken
together with argumentlinkinggeneralizationsand Miyagawa'sindependentdiscussion
on numeralquantifiers,constitutes a substantialbody of evidence for the ergativityof
the nominals in (7). Given the ergativity of the nominals, the ArgumentTransferthat
they undergois schematizedin (8).
(8) a. S YOOSIN(Theme)
b. suru (
) (acc)
c. SYOOSIN(
) + suru (Theme) (acc)
Since S YOOSINis ergative, it takes only an internalargument.Accordingto the abovementionedgeneralizationsrelevantto ArgumentTransfer,at least one internalargument
must be transferred.The only argumentavailablefor Transferis the sole argumentof
SYOOSIN, namely, Theme. Recall that when a noun's internalargumentis transferred
to a light verb, the external argumentis also transferredas a consequence of the topdown fashion of Transfer.However, in the case of (8) the external argumentcannot be
transferredsimply because the nominal does not have one in its argumentstructure.
That is, the light verb inherits the nominal's inability to assign a subject 0-role. Then
Burzio's Generalizationcomes into play and explains why the light verb should also be
incapableof assigningaccusative case. Hence, I will conclude that the analysis outlined
above provides evidence that both Burzio's Generalizationand ArgumentTransferplay
an importantrole in explainingthe 0-assigningand case-assigningpropertiesof ergative
nouns.
280
REMARKS
AND
REPLIES
2. The ResultativeConstruction
The resultativeconstructionis describedby Halliday(1967,63) as an attributethatresults
from a process. Simpson (1983) discusses the resultative constructionin English with
examples like those in (9).
(9) a.
b.
I paintedthe car yellow.
I painted the car a pale shade of yellow.
c. I cooked the meat to a cinder.
d. The boxer knocked John out.
The italicized expression is called a resultativeattributeand describes the state of an
argumentresulting from the action denoted by the verb. For example, (9a) means 'I
paintedthe car, andbecause I paintedthe car, the car becameyellow'. Simpsonobserves
that resultative phrases are always predicatedof the object function in English. Thus,
in the above examples the object of the verb (the car in (9a,b), the meat in (9c), and
John in (9d)) is the argumentwhose state is describedas resultingfrom the action.
Whatabout resultativeexpressions with intransitiveor passive verbs? Considerthe
intransitiveand passive examples with resultativeattributesin (10)-(11).
(10) a. The ice cream froze solid.
b. The butter melted to a liquid.
c.
The vase broke into little pieces.
(11) a. The car was painted red.
b. The ice cream was frozen solid.
The subjects in (10) are the argumentsof which the resultativeattributesare predicated.
Thus, in (lOa), for example, the ice cream, which is the subject of the sentence, froze,
and as a result, it became solid. The situationis the same for passive constructions.In
( lla) the car was painted, and as a result, it became red. Therefore, the sentences in
(10) and (11) seem to suggest that the controllerof a resultativeattributeis the subject
when the verb is intransitiveor passive.
Simpson, however, goes on to say that we can still maintainthe generalizationin
Englishthat the controllerof a resultativeattributeis always the object functionbecause
the subjects of the sentences in (10) and (11) are, in fact, underlyinglythe objects. For
verbs such as those in (10), if we focus on the Theme role assignment, assuming that
the Theme role is normallylinkedto the object, the underlyingobject statusof the subject
can be observed in the contrast in (12)-(14).
(12) a.
b.
(13) a.
b.
(14) a.
b.
I froze the ice cream solid.
The ice cream froze solid.
I melted the butter to a liquid.
The butter melted to a liquid.
I broke the vase into pieces.
The vase broke into pieces.
REMARKS
AND
REPLIES
281
The underlyingobjects in the (b)-sentencescan clearly be contrastedwith the subjects of unergativeverbs in regardto resultatives. Considerthe following examples.
(15) a. *1 danced tired.
b. *1 laughedtired.
c. *1 walked tired.
(1Sa),for example, does not meanthat I danced, and as a result, I got tired. The italicized
resultatives thus cannot be interpretedas being predicatedof the subject. These sentences sharply contrast with the (b)-sentences of (12)-(14), where such a predication
relationholds. Hence, if the surface subjectin the (b)-sentencesof (12)-(14) is identified
as the underlyingobject, neither case in (10)-(11) (ergative verbs and passive) would
constitute a counterexampleagainstthe generalizationthat the controllerof a resultative
attributeis always the object. The statementin (16) is the syntactic generalizationthat
Simpson claims to hold for resultativesin English (1983, 146).
(16) The controller of a resultative attributemust be an OBJECT, whether that
OBJECTis a surface OBJECT,as in transitiveverbs, or an underlyingOBJECT, as in passive and intransitiveverbs of the Unaccusative L= Ergative]
class.
Let us now examinewhetherthe generalizationin (16) also holds for Japaneseverbs
of Japanese origin.4First, consider the resultativeconstructionwith a transitiveverb.
(17) a.
b.
c.
Kuruma-oakakunutta.
car-Acc red painted
'(I) paintedthe car red.'
Hanmaa-de kinzoku-o hiratakuutta.
hammer-withmetal-Accflat
hit
'(I) hit the metal flat with a hammer.'
Syatu-o kiree-ni aratta.
shirt-Accclean-to washed
'(I) washed the shirt clean.'
d.
Pan-o
e.
bread-Accreally black-to toasted
'I burnedthe bread black.'
Hanako-wa kami-o nagaku nobasita.
Hanako-Tophair-Acclong lengthened
'Hanakolet her hair grow long.'
makkuro-ni
yaita.
4 Verbs of Japaneseorigindisplaya verbalinflectionalpattern.For example, they all inflect for past by
the suffixationof -ta, as in tabe 'to eat' + -ta (past)-- tabeta 'ate'. Sino-Japanesecompoundsby themselves
do not show such an inflectionalpatternsimplybecause they are categoricallynouns. To makethem function
as verbs, the light verb suru must be added. Since suru is a verb of Japaneseorigin, followingthe verbal
inflectionalpattern,a complexpredicatecomprisinga Sino-Japanesecompoundand suru patternslike a verb
of Japaneseorigin.
282
REMARKS
AND
REPLIES
The italicized expressions are resultativeattributes.In all the examples above, the resultativeattributesare predicatedof the accusative(-o) markednoun. (17c), for instance,
does not meanthat I washed the shirt, and because I washed it, I became clean. Instead,
it means that I washed the shirt, and because I washed it, the shirt became clean. Thus,
as far as transitive verbs are concerned, the controllerof a resultative attributeis the
object in Japanese.
Second, let us examine intransitiveverbs in the resultativeconstruction.
(18) a.
Hanako-no kami-ga nagaku nobita.
Hanako-Genhair-Nomlong lengthened
'Hanako's hair grew long.'
b. Pan-ga
makkuro-ni yaketa.
bread-Nomreally black-to toasted
'The bread burnedblack.'
c.
Hune-ga suityuu hukaku sizunda.
ship-Nom in water deep sank
'The ship sank deep in the water.'
Each italicized resultativeattributein (18) is predicatedof the noun in its sentence that
is marked with the nominativecase -ga-in other words, of the subject. In (18a), for
example, it is Hanako'shairthat became long as a result of the growthof the hair. Thus,
on the basis of the intransitiveexamples in (18), it appears that the controller of the
resultative with an intransitiveverb is its subject. The situation is, however, exactly
parallel to the English cases discussed above: the contrast between (18) and (19) corresponds to the one between the (a)-sentences and the (b)-sentences of (12)-(14).
(19) a.
Hanako-wa kami-o nagaku nobasita.
Hanako-Tophair-Acclong lengthened
'Hanako let her hair grow long.'
b. Taroo-wa pan-o
makkuro-ni yaita.
Taroo-Topbread-Accreally black-to toasted
'Taroo burnedthe bread black.'
c. Sobietogun-wahune-o suityuu hukakusizumeta.
Soviets-Top ship-Acc in water deep sank
'The Soviets sank the ship deep in the water.'
In (19) the same nouns as in (18) are the controllers of the resultative attributes.The
causative/inchoativepairs in Japaneseare not formallyidentical, as they are in English.
However, the forms in each pair are morphologicallyrelated, as (20) illustrates. They
are parallel to English in that the subject of the inchoative verb is the object of its
causative counterpart.5
' The causative/inchoative pairs in
Japanese are explored in detail in Jacobsen (1982). Each causative/
inchoative pair shares a verbal root, which is why the surface forms of the pair resemble each other. To the
verbal root, a certain morpheme is suffixed. There are more than a dozen morpheme sets for causative/
inchoative pairs. Which morpheme goes with which root appears to be a lexical property.
REMARKS
(20) Causative
taosu
kesu
tukeru
dasu
simeru
akeru
AND
REPLIES
283
Inchoative
taoreru
kieru
tuku
deru
simaru
aku
'fall'
'turnoff'
'turnon'
'come out'
'close'
'open'
It is well known that there is a systematicregularityin the way in which 0-roles are
linked to grammaticalfunctions (Fillmore(1968)and Jackendoff(1972), among others):
Agent is linked to subject, Theme is mappedonto object, and so on. Let us assume that
the underlyingobjecthoodof inchoative verbs in English and Japanese relies on such a
linkinggeneralization.That is, the sole argumentof inchoativeverbs is Theme and thus
is linked to object. This leads us to the conclusionthat the object, surfaceor underlying,
is the controllerof a resultativeattributein Japanese. Hence, the generalizationin (16)
is borne out not only in English but also in Japanese.
Assuming that the generalizationin (16) is a relevant condition on the resultative
constructionin Japanese,let us now investigatethe complex predicatesparticularlywith
respect to the resultativeconstruction.To beginwith, (21)illustratesa transitivecomplex
predicate with a resultativeattribute.6
(21) a.
Taroo-ga densen-o
mapputatu-niSETUDAN-sita.
Taroo-Nomelectric wire-Acc two-to
cutting-suru-Past
'Taroo cut the electric wire in two.'
b. Sobietogun-gadairiseki-nosiro-o
konagona-niHAKAI-sita.
Soviets-Nom marble-Gencastle-Accinto pieces destroying-suru-Past
'The Soviets destroyed the marblecastle into pieces.'
c. Kokka-wa sono arehatetatoti-o
utukusii heeya-ni
country-Topthat desolate land-Accbeautifulplain-to
KAITAKU-sita.
improvement-suru-Past
'The country improvedthat desolate land into a beautifulplain.'
The complex predicates in (21) are all transitive. As predictedby the generalizationin
(16), the italicized resultativeattributeis predicatedof the object, which is markedby
the accusative case -o. Thus, (21a) means that as a result of Taroo's cuttingthe electric
wire, the wire became two pieces, not that Taroo became two persons.
Next, observe the followingexamples, where some of the complex predicatesin (7)
are used.
6 Some complex predicatesconsisting of a noun and suru are impossible when trying to construct a
sentence with a resultativeattribute.This is because, as Simpsonobserves, a resultativeattributecooccurs
with a verbof contactanda verbthatdenotes changeof state. I assumesuch a semanticrestrictionis effective
in the Japaneseresultativeconstructionas well.
284
REMARKS
AND
REPLIES
(22) a.
Taroo-wa otona-nilookikuSEETYOO-sita.
Taroo-Topadult-to/big growth-suru-Past
'Taroo grew into an adult/big(person).'
b. Hune-ga suityuu hukaku TINBOTU-sita.
ship-Nomin waterdeep submersion-suru-Past
'The ship submergeddeep in the water.'
c. Saiboo-ganibai-ni KAKUDAI-sita.
cell-Nom double-toenlargement-suru-Past
'The cell doubled in size.'
sensinkoku-nami-ni
d.
Gizyutu-ga
e.
technology-Nomadvancedcountries-average-toimprovement-suru-Past
'The technology improvedto the level of the average advanced country.'
makkuro-ni ZENSYOO-sita.
Uti-ga
house-Nom really black-to burningdown-suru-Past
'The house got burnedblack.'
KOOZYOO-sita.
In every case in (22), the controller of the resultative attribute is the subject of the
sentence. Thus, the resultativeattributedescribesthe state of the subjectof the sentence.
These examples sharplycontrast with the following sentences.
(23) a. *John-ga kutakutaniSANPO-sita.
John-Nomdead tired takinga walk-suru-Past
'*Johntook a walk tired.'
b. *Mary-ga kanasikuBISYOO-sita.
Mary-Nomsad
smile-suru-Past
'*Marysmiled sad.'
The italicized phrases cannot be construed as resultative attributespredicated of the
subject. That is, (23a) cannot mean that John took a walk, and as a result, he got dead
tired. Likewise, the resultativereadingcannot be obtainedin (23b).
Notice that both sets of complex predicates-those in (22) and those in (23)-are
used intransitively,but the resultativeconstructionilluminatesa distinct differencebetween the two. The subject of each sentence in (22) serves as the controller of the
resultative whereas in (23) it does not. As in the English case of (10) versus (15), the
Japanese data from the resultatives together with the above-mentionedlinking generalizationslead us to the conclusionthatthe subjectof each sentence in (22)is underlyingly
its object whereas the subject of each sentence in (23) has always been its subject. This
conclusion, coupled with Miyagawa's(1989) independentsyntactic evidence, substantially supportsthe view that the nominalsin (22) as well as those in (7) are ergative.
It should be noted that the identificationof the nouns in (22) and those in (7) as
ergative is not as predictableas it is with the Japanesenative verbs in (18)-(20) or with
the Englishexpressionsin (12)-(14). This is because, unlikethe formsin (20), the complex
predicates (or the nouns themselves) in (7) do not exhibit causative counterpartsthat
REMARKS
AND
285
REPLIES
are morphologicallyidentical with or similar to them.7 Hence, it is only by syntactic
and/or semantictests, but not by forms, that we can recognize the ergative status of the
nouns, and ultimatelyof the complex predicates, in (7).
The ergativeanalysis of the complex predicatesin (7) is also supportedon semantic
grounds. The evidence comes from the parallelismin meaning between the complex
predicates in (7) and the Japanese native verbs of the ergative class, as in (20). Recall
that the complex predicatesin (7) are always intransitiveand have no morphologically
identical or related verbs that would constitute causative/inchoativepairs like the Japanese native verbs in (20). However, many of them find synonyms among the causative/
inchoative pairs belongingto the Japanese native verbs. When the complex predicates
in (7) and their synonyms from the Japanese native class are comparedwith regardto
their meanings,it is always the ergativecounterpartthat correspondsto the meaningof
the complex predicate. For example, the complex predicate SEETYOO-suru'grow' in
(7) has a synonymous verb of Japaneseorigin, namely, sodatu. Sodatu is an inchoative
(= ergative)verb whose causative counterpartis sodateru. As can be seen, sodatu and
sodateru are morphologicallyrelated. I list more examples of this relationshipbetween
the complex predicatesin (7) and causative/inchoativepairs of Japanese origin in (24).
(In parenthesesare the transitivecounterpartsof the synonyms.)
(24) Complex predicates in (7)
TINBOTU-suru
KAKUDAI-suru
ZENSYOO-suru
GYOOKO-suru
ZYOOSYOO-suru
KAKOO-suru
KAIMETU-suru
SYUKUSYOO-suru
'submerge'
'enlarge'
'burndown'
'solidify'
'go up'
'go down'
'be demolished'
'reduce'
Synonym-inchoative
sizumu
hirogaru
yakeru
katamaru
agaru
sagaru
kowareru
tizimaru
(sizumeru)
(hirogeru)
(yaku)
(katameru)
(ageru)
(sageru)
(kowasu)
(tizimeru)
(24) clearly demonstratesthat there is a tight semantic relationbetween the inchoative
verbs of Japanese origin and the complex predicates in (7). Furthermore,the sole argument of the complex predicates in (7) and of their inchoative synonyms is Theme.
Since the inchoative verbs in (24) are ergative, their synonymous complex predicates,
or, more specifically, the nouns themselves, may well be so analyzed, assumingthat the
linkinggeneralizationis in effect.
Before closing the discussion of ergative nouns, I would like to point out that the
investigation of the relationshipbetween noun types and accusative case assignment
leads to a classificationof nouns and complex predicatesthat parallelsthe classification
of verbs. For example, KEEKOKU-surubelongs to the transitive type since the noun
7By causative I mean a lexical causative rather than a morphologically productive causative form involving -sase.
286
REMARKS
AND
REPLIES
KEEKOKUis transitive. With a transitive noun, the external argumentnecessarily is
transferredbecause of the generalizationpertinentto ArgumentTransfer.This means
that suru inheritsthe noun's abilityto assign the subject 0-role. It follows from Burzio's
Generalization,then, that the verb can also assign accusative case. The result is (la).
An unergativenoun, exemplified by AISEKI 'table-sharing',is expected to have
only an external argument.The external argumentmust be transferredto the argument
structureof suru, and at the same time suru inheritsthe ability to assign the subject 0role. According to Burzio's Generalization,accusative assignmentshould be possible.
And indeed it is, as (25) shows.
sita.
John-wa Bill-to AISEKI-o
John-TopBill-withtable-sharing-Accsuru-Past
'John shareda table with Bill.'
b. John-wa Bill-to AISEKI-sita.
John-TopBill-withtable-sharing-suru-Past
(25) a.
Burzio (1986) claims that an unergativeverb such as laugh can assign accusative case
to its object if such an object NP is available. This is why laugh can appear with an
object as in John laughed a laugh.8 We can see the parallel situation in nouns like
AISEKI. It is a consequence of Burzio's Generalizationthat unergativenouns receive
accusative case.
An ergativenoun such as S YOOSIN'promotion',on the other hand, cannot involve
accusative case assignmentat all. As schematizedin (8), surureceives the sole argument
of the noun-that is, Theme-by ArgumentTransfer.Being ergative, SYOOSINis not
able to assign the subject 0-role. It is this inabilitythat the verb also inherits as a consequence of ArgumentTransfer.It follows from Burzio's Generalizationthat if the verb
cannot assign the subject 0-role, then it cannot assign accusative case. This is why the
configurationof the sentence in (4) is ungrammatical.
3. Conclusion
In this article I have shown that some nouns and inchoative verbs in Japanese pattern
in the same mannerwith respect to the resultativeconstruction. Further,adoptingthe
linking generalizationhas led us to the claim that the subjects of complex predicates
comprisingthe nouns in (7) and the light verb suru are underlyingobjects. This conclusion, put together with Miyagawa's independentargument, strongly supports the ergativity of the nouns in (7). Furthermore,the inabilityof ergativenouns to appearin the
configurationin (4) is straightforwardlyexplained, given Burzio's Generalizationand
the process of ArgumentTransfer.When the sole argumentof the noun is transferred
to the argumentstructureof the light verb, the verb also inheritsthe noun's inabilityto
assign a subject0-role. It then follows from Burzio's Generalizationthat the verb should
8
I would like to thank Beth Levin for pointing this out to me.
REMARKS
AND
REPLIES
287
not be able to assign accusative case. Hence, the above discussion presentsan argument
for Burzio's Generalizationand for ArgumentTransfer.
References
Burzio, L. (1981) IntransitiveVerbs and Italian Auxiliaries, Doctoral dissertation, MIT, Cambridge, Massachusetts.
Burzio, L. (1986)Italian Syntax, Reidel, Dordrecht.
Fillmore,C. (1968)"The Case for Case," in E. Bach and R. Harms,eds., Universalsin Linguistic
Theory,Holt, Rinehartand Winston,New York.
Grimshaw,J. and A. Mester (1988) "Light Verbs and 0-Marking,"LinguisticInquiry 19, 205232.
Halliday, M. A. K. (1967) "Notes on Transitivityand Theme in English, Part 1," Joulrnalof
Linguistics 3, 37-81.
Jackendoff,R. (1972) Semantic Interpretationin GenerativeGrammar,MIT Press, Cambridge,
Massachusetts.
Jacobsen, W. (1982) Transitivityin the Japanese VerbalSystem, IndianaUniversity Linguistics
Club, Bloomington.
Miyagawa,S. (1989)"LightVerbsand the ErgativeHypothesis," LinguisticInquiry20, 659-688.
Simpson, J. (1983) "Resultatives," in L. Levin, M. Rappaport,and A. Zaenen, eds., Papers in
Lexical-FunctionalGrammar,IndianaUniversity LinguisticsClub, Bloomington.
Department of East Asian
Languages and Cultures
Goodbody Hall 248
Indiana University
Bloomington, Indiana 47405