74 / Kameyama

APPENDIX

shared vrovertv: examvles in this study: Susumu Kuno

Harvard University*
1) IDENT&SUBJ (3) (4) (l4)<zero-sbj> (6)a
(7)a (40)a (39)a<zero-sbj>
2) DENT only (10) (1 1) (39)bczero-obj2>
3) SUBJ only (42)a(may be [+DENTI) Reported speech in Japanese is always marked by the clause-final
particle to, regardlw of whether the discourse mode is direct or indirect.
4) nonIDENT&nonSUBJ (13) (14)<zero-obj> (39)aaero-obj2> For example, obsewe the following sentence:
5) nonDENT only (40)b (39)baero-obj> (Ll) Taroo wa ore wa oobaka da to itta
I ( d g . ) biggest-fool be that said
6) nonSUBJ only none a 'Taroo said that I am the biggest fwL'
no property shared unacceptable sentences: b. ' T m said, T am the biggest fooln.'
(1) (2) (9)b (6)b (7)b (47) (48)
Sentence (1.1) ie ambiguous between the indirect-discourse interpretation
acceptable IDENT-neutral sentences: (a), and the dired-diecourw interpretation in (b).
(12) (41) (42)b (43) (46) It is not the case that all to-marked claoses are ambiguous: in the
following, (a) allows only a direct-discourse interpretation, while (b) allows
only an indirect-discourse interpretation:

(1.2) a. Hanako wa wmimasen ga okane o k i t e kudasaimasen ha

feel-sorry but money lending give-not Q
to itta.
that said
'Hanako said, "1 am sorry, but would you please lend me
some money?"'
b. Hanako wa ore ga usotuki da to iihnrasite irn.
I liar am that saying is
'Hanako is telling everybody that I am a liar.'

(1.24, as a faithfnl reproduction of what Hanako must have said, clearly

signals that the to-marked clause is in direct-discourse repreeentation.
(1.2b), in which Hanako, a female, is unlikely to have used the vulgar
first-person pronoun ore, signals that the to-marked clause is in
indirect-diecourse reprentation, with ore referring to the external speaker of
the sentence. Furthermore, the use of the nominative particle go in the
76 / Kuno Blended Quasi-Direct Discourse in Japanese / 77

to-clause sign& that the c h is a subordiuate chum, I l r t k than a main In (1.6a) and (1.6b), dore 'who' is bound by the clause-mate ka, while in
clause, and reinforces the indired-diecouree interpretation of the cl.use.1 (I.&), it is bound by ka in a higher clam. Note that the embedded c l a w
Nor is it the csse that reported speech in Japan- is either all in (1.5~)ia not a question, but reporfed speech.
d i d or all indirect. There are sentences in which part of the reported We obtain the following patterns of embedded c l a w and questions:
speech is direct, while the rest is indire~t.~ For example, observe the
following sentence: (1.6) 8 [...[s'-to]...]s
H d o wa paroo ga kita] to itta
(1.3) Taroogayatu no uti ni eugu Loi came that said
him(*.) 's h o m to immediately come (Imp.) 'Hanako said that Taroo came.'
to denwa o M e t e kita b. [[...[S'-to] ..Is Qlsv
that phone placing came Hanako wa [Taroo ga kita] to itta ka
'Taroo called me up and said that (lit.) "Come right 'Did Hanako say that Taroo came?'
now" to his house.' C. I...[S-QlS. . .Is
Hanako wa [Taroo ga kita ka] iwanakatta
The third person pronoun yatr 'that guy' (vulgar form) is intended to refer 'Hanako didn't say whether T m came.'
to Taroo. Therefore, gats no uti cw' must be in i n d i i dkowse. On the
other hand, koi, which is an imperative verb form, must be in
d. I[...[S-QI, ... Is 9191
direct-disooorse mode. I win d e r to the above kind of mixed mported Hanako wa [ T m ga tita ka] itta k a
apeech as "blended disoouraen. 'Did Hanako say whether Taroo came?'
Before I examine some of the syntactic and disoourw-functii e. [... [[S-Qls. to] ...Is
characteristice of blended dbcowee, I should outline some bssic fade about Hanako wa [Taroo ga Lita kal to itta
reported speech and embedded questions. 'Hanako said, "Did T m come?"'
f. 11.- I[S-Qls bl ..-Is 01s-
(1.4) Questions, both main and embedded, are marked with the Hanako wa [Taroo ga kita ka] to itta La.
sentence-fmd particle ko: 'Did Hanako say, "Did Tanx, come?"'
a. Tarcx, wa kita ka. g. I...[Wa-S-Q]s, ..Is
came Q Hanako wa [dare ga kits ka] iwanakatta
'Did Taroo come?' who came didn't ery
b. [Taroo ga genki de irn ka] siritai. 'Hanako didn't say who came.'
well ia Q know-want h. [[...[wa-s-Q]s, ..I, Qls.
'I want to iind out if Taroo is well.' Hanako wa [dare ga kita Lo] iwanakatb ka
(1.6) WA-expreesione must be bound by a Q = ka that 'Didn't Hanako say who came?'
commanda them?
a D m ga kita t.
i. [...[[--S-Q], ...
to] Is
who came Q Hanako wa [dare ga kita ka] to i t t a
'Who came?' 'Hanako aaid, "Who came?"'
b. (Dm ga kit* ka] siritai. j. [[...I[--S-Ql, ...Y Is Qls
who came Q know-want Hanako wa [dare ga kita ka] to itta ka.
'I want to find out who came.' 'Did Hanako nay, "Who came?"'
c. Hanako wa [duega kita] to itta ka. k. [[...[[Wh-S]F to] ..Is QIs.
who came that said Q Hanako wa [dare ga kita] to itta ka
'Who did Henako say came?' 'Who did Hanako my came?'
 Blended Quasi-Direct Discourse in Japanese / 79

However, the following pattern is unacceptable becam the Idb 'Not a thing did the patiant say that he wanted.'
expmion in the embedded c l a m doea not have a Q to bind it.
As is well known, the particle sib 'only' and the indefinite pronoun nk mo
(1.7) a. *(...[[WL- S], to] ..I, are negative polarity items, and they require a clause-mate negative
*Hand0 wa [dare ga Lita] to itta morpheme. For example, compare the following sentences:
who came that said
b. *[... Wh...[[ S-Q], to] ..Is (2.3) a. Hanako wa nihongo rB. h a n d .
Japan- only speak-Neg-Rw.
*Dare p Hanako p kita h to itta 'Hanako can speak only Japaneee.'
who came Q that said b. Hanako wa [eigo riL. h a n d ] ameriksuii
Enghh only speak-can-Neg American
Sentence (1.7.) is unacceptable d t w the embedded &claw or the main to kekkomita
alGclawe is marked with a rising intonation to substitute for )a Likewise, with married
(1.7b) ia unacceptable u n h the main c l a w in marked with a rising 'Hanako married an American who a spsk only
intonation. Note that the d-expression in (1.7b) cannot be bound by the English.'
embedded Q, because the latter does not command the former. c. *Hand0 wa [eigo dL hamwru] a m a i n to
On the other hand, the following pattern is allowable, becawe more English only speak-can American
than one &-expreesion can be bound by a single Q: kebn d a m .
m P y can-Neg-Present
(1.8) [... WL ...
[(Wh-Sls QlSv 'Hanako cannot marry an American who can epeak
Dare ga Hanako ga nani o sita to itta (no) ka only Jhglbh.'
who what did that said Q (2.4) a Hanako wa ud mo d e b h t t a .
'(Lit.) Who said that Hanako did what?' anything can-do-Neg-P&
'Hanab couldn't do anything.'
In (1.8)' the main c l a m Q bids both the upstaim and downstaim b. Hanako wa [d alro d e k i ~ i ] otoko to
wh-expressions. smything can-do--Neg-Pres. man with
kdkO~it&
married
'Hanako has married a nun who can't do anything.'
c. *Haado wa [d mo d e b ] otoko to kekkonubdatta.
Observe the following sentences: anyone can-do man with marry-Neg-Past
'Hanako didn't marry a man who could do anything.
(2.1) a Kanzya wa mhu riL hosinui to itta
patient water only want-Neg that said Both (2.3a) and (2.3b) are rumptable becaw dko has a c h - m a t e
'The patient d that he didn't want anything other negative -no--, but (2.3~)is unacceptable: siko is in the ambedded clam,
than water.' while the negrrt'ie -ne- is in the main clam. L i k e , both (2.48) and
b. Kan7,ya wa mien dk hosii to iwaukatta. (2.4b) are acceptable becanse the negative polarity expmaion nosi mo han a
patient water only want that soy-Neg-Past clause mate negative in th- sentences, but (2.4) in unacceptable because
'Only water did the patient soy that he wanted.' nani mo is in the embedded clam, while the negative morpheme -no- is in
(2.2) a. Kanzya wa uli mo hoe- to itta the main cl.ase.
patient anything want-Neg that oaid This mggesb that the d)a in (2.1b) and noni mo in (2.2b) am
'The patient d that he didn't want .nything.' c l e w matee of the main clause negative morpheme. That ia, mizu sika
b. Kanzya wa rrul mo hooii to iwaukatta 'only water' and nani mo in these sentences are thematic elements that are
patient anything want that say-Nee-Pad
80 / Kuno Blended Quasi-Direct Discourse in Japanese / 81

outeide the embedded clam. The following sentences support this to itta
hypothesis: that said
'Taroo said (lit.) that "Don't call (me)" except at his office.'
(2.6) a. *W dka mno kanzya wa hods& to itta b. T m ga d mo yatu no office ni motte kuru-M
water only that patient want-Neg-Prse. that said anything his (vulg.) bringing come-Neg.Imp.
b. Mizu si& sono kanzya wa hosii to i w d t t a . to itta.
water only that patient want that say-Neg-Past that said
'Ody water did the patient say that he wanted.' 'Taroo said (lit.) that "Don't bring anything" to office.'
(2.6) a. * N d mo eono kanzya wa hoe- to itta.
anything that patient want-Neg-Pms. that said The fact that (2.9a) is perfectly acceptable shows that the indirect-discourse
b. N d mo eon0 kaneya wa haii to i w h t t a element yatu no office ni aka 'only to to office' and the direct-discom
anything that patient wants that say-Neg-Past element denma o k h - n o 'Don't call (me)' are in the same clause.
'Not a thing did the patient say that he wanted.' Likewise, the fact that (2.9b) ia acceptable shows that nani mo 'anything'
and motfc kum-na 'Don't bring' are c l a m mates, and that consequently, the
(2.6a), in which mito uika of (2.14 haa been fronted out of the embedded indirect-diecourse element vatu no oflice ni, which appeara between the
c l a w to wntence--initial position, in unacceptable, wbie (2.6b), in which above two c l a w mate elemenb, must a h be a clause mate of the
mim aka of (2.lb) has been fronted to sentence-initial position, m acceptable direct-discourse element moue h - n a This c o n f i i my initial hypotheaia
and synonymom with (2.lb). Likewise, (2.6a), in which nani mo of (2.24 that Japanese allows a mixture of direct and indirect diecourse elemente
has been fronted out of the embedded clause to sentence-initial position, m within a tu-marked clause.
unacceptable, while (2.6b), in which nani mo of (2.2b) hm been fronted to Now obeerve the following sentences:
sentence-initial position, is acceptable, and ia synonymow with (2.2b). This
firmly eetablbhea that (2.14 and (2.lb) have two totally different structures:' (2.10) a. Taroo ga (yatu no office ni sikal kite ii
his (vulg.) only coming alright
to i w d a t t a .
that say-Neg-Past
'Only to his office did Taroo say "it is alright to come",'
b. *Taroo ga [ore no office ni sika] kite ii to i w d a t t a
The question then arisee as to whether indirect-discourse elemenb my (vulg.) only coming alright said-Neg.-Past
in blended d i i a r s e are within the to-marked c l a m , or whether they are '(Lit.) "Only to my office" did Taroo say "it is alright to come".'
main c l a m elemenb; that in, whether (2.8) has the phraee-stmctwe
configuration shown in (a) or in (b): In both (2.104 and (2.10b), the negative morpheme -na- is attached to the
main c l a w verb. Therefore, the negative polarity expreesion yatu/on no
(2.8) Taroo ga yata no nti ni snga Loi to itta office ni silo 'only to b / m y office' must be a constituent of the main
his(*.) 's house to imrned. come that said clause. The fact that (2.10b) is unacceptable is explainable only by assuming
'Taroo said (lit.) that "Come right now" to hi house.' that the following constraint holds:
a [ T m ga (yatu no uti ni "mgu Loin] to itta]
b. [Taroo ga yatu no uti ni ["sugu koimjto itta] (2.11) Bam on Extrretbn from Dirset-Disorurc Ckwr: Elements in a
direct-discourse clause cannot be extracted out of that c l a w .
The test involving sika ...-no- and nani mo ...-na shows that (a)
repreeenb the correct stmctnn. Observe the following: (2.104 ia acceptable because gotu no office ni aika is an indirect-discourn
element in a blended d i i u n r e clause, but (2.10b) is unacceptable because
(2.9) a. Taroo ga yatu no office ni denwa o kakem-u ore no oj'jice ni (sika) is a direct-diecourse element in a wholly
his (vulg.) to only phone place-Neg.Imp.
82 / Kuno Blended Quasi-Direct Discourse in Japanese / 83

direct-discourse c h In (a), the &expression nanzi ni 'what time' is bound by the Q which is
part of Hanalro's direct discourse. This eentence is perfectly acceptable. On
(2.12) a. [Taroo ga yatm m a(llea d rP. [ 9 "kite iia] to iwp..l.tta] the other hand, in (b), the ad-expression, which is not part of Hanako's
b. *[Turn, ga "ore m odllcs m i dLw[" 9 kite ii"] to iwutlrIrtta] direct discourse, is bound by the main clause Q. Thie eentence is
unacceptable. But the unacceptabiity of (b) does not mean that a
Liewise, obeerve the f o h i n g sentence: &-expression bound by the main c l a m Q cannot appear in reported speech
containing a direct-discourse element. Observe the following sentence:
(2.13) a ' h m o wa yatu no tokom ni DUI mo motte kite hosii
his (valg.) place to anything bringing coming want (3.2) Tram wa Iyatu no uti ni nmsi nl koi] to itta ka.
to itte W t a his(vulg.) house to what-time come that said Q
that e g be-Neg-Past 'What time did Taroo say that (lit.) "Come 9" to his house?'
'Not a thing did Turn, say that (lit.) "(I) want (you)
to bring" to his plaoa.' The comparieon of unacceptable (3.lb) and acceptable (3.2) leads to the
b. *Tarw, wa "ore no tokom ni" rut mo motte kite hosii following hypothesis:
my (vulg.) p l a a to anything bringing coming want
to itte h.latta (3.3) Claw-Phul Conditbn: Dit-disoonnre elemente in blended
that saying was discourse must appear in claw-final position.
'Not a thing did Turn, my that (lit.) "I) want (you)
to bring 9 to my offbe.".' Then (3.lb) is unacceptable because the direct-diecourse element ore no uti
ni ia not claw-final. On the other hand, (3.2) is acceptable because it
In both (2.134 and (2.13b), the negative polarity expression mni mo mud does not contain any non-clause-final direct-discourse element. The
be a c l a w mate of the main c l a w negative morpheme. Therefore, yatu/orc ,sentences below demonstrate the difference between (3.lb) and (3.3):
no toke ni 'to his/my place' muet be a main c l a w constituent. The quotation marks indicate the direct-discourse elements:
ungnunmaticality of (2.13b) is attributable to the fact that the
direct-diecoume element ore no tokoro ni 'to my place' has b a n fronted out (3.4) a. *Tam0 wa ["ore no uti ni" nanai ni "koi"] to itta ka
of a wholly direct-discoruse clam. b. Taroo wa Iyatu no nti ni nanzi ni "koin] to itta h.

Now ( z k e the following bentences:

(3.6) a. Taroo wa Iyatu no uti ni sugu "koi",

O k w e the following sentences: hi(vulg.) house to immed. come (hp.)
sosite yatu no sigoto o "tetudae"] to itta
(3.1) a. T a m wa [ore no uti ni d d kom ka] and hie (vulg.) work help that said
my(+.) h o w to what-time can-come Q 'Taroo said that (lit.) "Come" right now to his
to itta house, and "Help" hie work.'
that #aid b. T a m wa kare no uti ni sugu "korenai ka"
'Taroo said, "What time can you come to my house?"' his house to immed. can't-come Q
b. * T m wa [ore no uti ni o.nri ni koi] sosite kare no sigoto o "tetudaenai ken] to itta.
my(vu1g.) house to what-time come and his work csn't-help Q that said
to itta (no) ka. 'Taroo eaid that (lit.) "Can't you come" to his
that said Q house right now, and "Can't you help" his work.'
'*What time did Taroo say, "Come to my house flu?'
In (3.6a), two imperative v e r b appear at the end of their respective clanses.
 Blended Quasi-Direct Discourse in Japanese / 85

In (3.6b), two interrogative v e r b appear at the end of their respective speaker's empathy be on y (i.e., E(z) < E(y))." An independently motivated
clauses. The acceptability status of these sentences varied from apeaker to meta-rule which dictates that the degree of the speaker's empathy with
speaker. For thoee speakem who accept these two sentencea as grammatical, himael must be larger than that with someone elae (i.e., Speech Act
it mu& be that the Clauoe-Final Condition applies individually to each of Empathy Hieruchy: E(Speaker) > E(0them)) accounts for the unacceptability
the conjoined clanees. On the other hand, for thoee speakera who consider of (3.7b) and (3.8a): both result from conflicting empathy requirements. In
these sentences less than acceptabk, it must be that the condition applies to (3.7b), the uoe of yatta q u i r e s the relationship E(z = Hanako) 2 E(y =
the entire coinjoined structure. The following sentence ie acceptable to all boku), while the Speech Act Empathy Hierarchy requirw E(boku) >
speakem: E(Hanako). The sentence is unacceptable because of the irreconcilable
conflict between these two qairements. Similarly, in (3.8a), the uoe of
(3.6) a. Taroo wa [yatu no uti ni sugn kite, yatu no sigoto o b e t a requires the relationship E(z = boku) < E(y = Hanako), which again
hi houee to immed. coming hi work contradicts the requirement of the Speech Act Empathy Hierarchy. Hence,
Vetudaen] to itta. the ~naccepta~lity of the sentence.
help (Imp.) that said The giving verbs yam and krm am a h used as supporting verbe
'Taroo said that (lit.) "Helpn hie work by coming to indicate the speaker's empathy vis-a-vis the agent (subject) and
to hie h o w right away.' nonagent (nonrmbject) of action. For example, observe the following
b. Taroo wa [kare no uti ni mgn kite, kare no sigoto o sentences:
his house to irnrned. coming his work
"tetudaenai kan] to itta (3.9) a H& ga Taroo ni tegami o kaita
can't-help Q that said to letter wrote
'Taroo said that (lit.) "Can't you helpn his work 'H.luLo wrote a letter to T m . '
by coming to his houm right away.' b. H d o ga Taroo ni tegami o k i t e yatta
c. Hanako ga Taroo ni tegami o kaite kumh.
Note that the direct-dkoume portion of the blended discourse in these
sentences is in clawe-final position under either interpretation of the (3.94 is a neatral sentence, and contains no overt signal as to whether the
constraint. apeaker is placing himaeK closer to HanaLo or to Taroo in describing the
In Kuno and Kaburuki (Pn),the Empathy Perspective was uaed to event ander diacuoaion. On the other hand, y d t a in (3.9b) indicates the
account for the uoe of the giving verbs yam and k r m in Japan-. relatiomhip E(z = Hanako) > E ( y = Taroo). Similarly, k & a in (3.94
Observe the following conbots shows that the @er has placed himoelf closer to Taroo than to Hanako
(i.e., E(z = Hanako) < E(y = Taroo)) in producing the eentencef
(3.7) a. Boku ga Hanako ni o h e o y d . The behavior of yam and h r m as supporting v e r b is more or lam
I to money gave identical to that of yam and )rmr as giving verbs. The acceptability statua
'I gave money to Hanako.' of the sentencea in (3.10) and (3.11) exactly paralleb that of (3.7) and (3.8):
b. *Hanako ga boku ni o h e o yatta.
'Hanako gave me .money.' (3.10) a. B o h ga Hanako ni tegami o kaite yatta.
(3.8) a. *Baku ga Hanako ni okane o LPrd.. b. *Haa.Lo ga boku ni tegami o kaite yatta.
to money gave (3.11) a. *Baku ga Hanako ni tegami o bite hrsta.
'I gave money to Hanako.' b. Hanako ga boku ni tegami o kaite hrst..
b. Hanako ga boku ni okane o hrst..
'Hanako gave me money.' Let as now examine how yam and kurcrr behave in nported speech.
First,observe the following sentences:
We can explain the above contwt by mourning that the expreusion z ga y
ni yam 'a given to j reqnires that the speaker's empathy be on z (i.e., E(z) (3.12) a H d o wa boku ga kanoayo ni b e o y a w to itte h.
2 E(y)), while the expression z ga y ni k r m 'z given to y' r e q h that the I her to money gave that oaying is
 Blended Quasi-Direct Discourse in Japanese / 87
86 / Kuno
y a r m h t h " ] to itte iru.
'Hansko ssp that I gave her money.' 'Taroo says, "I did not lend money to my meetheartm.'
b. Hanako wa boku ni okane o kmda to itte h. b. Tarw wq [ore no koibito ni okane o . b i t e
me to money gave that saying is yumhtt."] to itte in.
'Hanako says that (she) gave me money.' 'Taroo says that (lit.) "I did not lend" money
to my (i.e., the external speaker's) meetheart.'
The w of ydta and krrda in the above sentences ia consistent with the c. Taroo w q ["ore no koibito niR b e o b i t e
empathy requirements of these two verbs. y u a m h t t . ] to itte irn.
What nee& to be explained is the ~rooeptabiilityof sentences like 'Taroo eoys that I (i.e., the external @er) did
(3.1%) and (8.13b) --they appear to violate the above requirements on not lend money to "my sweetheart".'
yam and kureru
It is extremely diffiiult to mtarpret (3.14) as (3.16~). This fact,
(3.13) a. Hanako wa boku d okane o ~ 8 t h to iiiurssite ira. explained automatically by the Claw-Final Condition.
too, is
Note t h 3 (3.lb) 1
me to money gave that saying is violates this constraint.
' H d o is telling everybody that "(9 gave money" to me.' Observe next the following sentences
b. Hanako wa boku ga kanoyo d okane o hrst. to itte i n .
I her to money gave that saying is (3.10) a Yamado wa [ore ni kindokei o "trttc yath', sosite
'Hanako says that I "have given" money to her.' me (vulg.) gold-watch buying gave and
ore no waihu ni minLu no kooto o "tte yathm]to
Sentence (a) contains the expmion Qanogo go) boku m i okanc o ydta, my (vulg.) wife mink coat buying gave that
which is unacceptable as the independent sentence (3.7b). However, here it iihnrcrsite iru rasii.
is perfectly acceptable. S h d d y , sentence (b) contains the expmion boku saying ia it--
go kanozyo ni okans o b e t a , which is unacceptable as the independent 'It seems that Yamada is telling everybody that (lit.)
clause (3.8a). But this sentence, too, is perfectly acceptable. Kuno and "I have bought-e (i.e., the external speaker) a gold
K a b d accounted for this fact by d g that the empathy condition on watch, and "I have bought" my wife a mink cost.'
the uee of guru and k r r m appliea both to the surface structure and to the b. Yamada wa [ore no uti ni "tomete lruaui Ln,m i t e
direct discourse reprwentation of reported speech: acceptability rwulb if the my(*.) h o w at let-stay wouldn't-you and
condition is satisfied at either of the two levela. Now, with the ore no kuruma o "tukawusts hrslul k]to denwa o
blended-discourse pempective under diecussion available to us, we can my (vulg.)car let-uee wouldn't-you that phone
account for the acceptabiity of (3.13a) and (3.13b) by saying that gaita and kakete kita
b e t o in these eentencea are in direct-discourse mode, with the rest of the placing came
to-clause in indirect-dkourse mode. 'Yamada called me up and d that (lit.) "Won't you let
With this awumption in mind, observe the following: me stay" at my (i.e., the external speaker'^) h o w , and
"Won't you let me use" my (i.e., the external speaker'^) car.'
(3.14) Tarw wa, [ore no koibito ni okane o kasite yuamhtt.1
my(vu1g.) sweetheart money lending not-gave
to itte iru.
Here again, the acceptability status of these sentences varies from speaker to
speaker. I aamme that the sentences are acceptable to those speakem whose
!
that saying ie Clause-Final Condition appliea to each member of coqoined clauses, and
that they are unacceptable to those for whom the constraint appliea to the
The above sentence is potentially ambiguoue between (3.15a) and (3.15b) entire conjoined structure.
with respect to which elements of the to-claw are in direct-discourse
representation, and which are in indirect-discourse representation:
Now we should show that sentences of the pattern (3.13)' where the
empathy reqairement on gam and kurcrr is o k e d only if they am
I
interpreted as k t - d k o u m e expmiona in blended dkounre, .re limited
(3.15) a. Tanx, wa ["ore no koibito ni okane o Lasite
88 / Kuno Blended Quasi-Direct Discourse in Japanese / 89

to sentences containing to-marked reported spsech. O h the following

sentences: (4.2) Sugu h i .
'Come immediately'
(3.17) a. Hanako wa [boku ni ikwa lcasite brat. ka] This is an i n f o d vulgar imperative form whose uee is almost exclusively
me to how-much lending gave Q limited to male speakers. Hanako must have nsed one of the following
oboete inai.
expressions:
ramembering isn't
'Hanako doesn't remember how much she has lent me.'
b. *Hanako wa (boku ni ikwa h i t e yatta ka]
(4.3) a. Sugu kite kurenai?
coming won't-you-give
oboete inai. b. Sugu kite kudasai.
'Hanako doesn't remember how much (lit.) "I have coming give (Imperative)
lentw me.' c. Sugu kite kudaeaimesen La
(3.18) a. Hanako wa [boku ni okane o kasite -1 &to o give-Polite-not Q
me money lending gave that
oboete inai.
S i l y , observe the following sentence:
remembering isn't
'Hanako doesn't remember that she lent me money.' Musuxne ga [mgu kite kurd to denwa o M e t e kita.
b. *Hanako wa [boku ni okane o b i t e yatta] koto o (4.4)
daughter come give that phone placing came
oboete inai. 'My daughter d e d me and said "Come right now."
'Hanako doesn't remember that (lit.) "1 have lentw
me money.'
A daughter would never w e the i n f o d vulgar imperative form kite )rrc to
her parent. Themfore, the "direct-discowee" portion of the blended
(3.17) involves an embedded Q-clam, while (3.18) involves an embedded
koto '(the fact) that' c l a w . The unacceptabiity of (3.17b) and (3.18b)
diecourae is "quasi-directn -- it does not represent the quoted speech
verbatim. Let ua tentatively a s m e that what the qussi-direct d i o u m
shows that these types of c l a m do not allow blended discourae.
portion of (4.4) represenb is the h e r ' s rendition of the imperative that
maintains the original syntax of the imperative, but which uaea a level of
speech appropriate not to the quoted speaker, but to the external speaker of
the whole sentence.
The logical question, then, is whether there is true direct disco-
Observe the following sentence:
in Japanese. The anewer to this question is atrumative. Observe the
following sentence:
(4.1) Hanako go b o s y o no ie ni "sugu Loin] to denwa o
her h o w to immed. come (Imp.) that phone Hanako ga ["watalmsi no ie ni wa rmmimaesn ga moo
kakete kita (4.6)
placing came mY house to I-am-sorry but any-more
konaide kudasaim] to itta.
'Hanako called me up and eaid that (lit.) "Come right now" don't--come (Polite) that said
to her house.' 'Hanako said, "1 am sorry, but please don't come to my
home any more."'
Since argr koi is an imperative, it is the direct-disco- portion of the to reflect
The to-marked reported speech in the above sentence
to-marked blendid discom. But "direct-diiurse" is not an accurate
characterization of thie imperative expression. Hanako's female-style speech, and can be assumed to be true
Hanako, being a female direct-discourse speech. T h ~
is confirmed by the fact that the above style
speaker, is unlikely to have x t u d y said
 Blended Quasi-Direct Discourse in Japanese / 91

of quotation cannot be mixed with i n d i r e c t - d i m

below:
elements, as shown "Would you give me the favor of not comingm
"1 do not want (you) to comen I
to her house.'
(4.6) *Hanako ga [keaozyo no ie ni wa "sumhasen ga moo
her house to I-am-sorry but any-more Theee wntencea are likewise unacceptable because of the mismatch between
konaide kudasain] to itta. the level of speech in the quasi-direct disco- portion of the reported
don't-come (Polite) that Mid speech and the main clause verb. On the other hand,
'Hanako said that (lit.) "Please don't come any moren
to her house.' (4.10) Hanako ga [keaozyo no ie ni wa
(4.7) *Handso wa [ n w i ni "kite itadakemesen ka] her house to
what-time come wouldn't-you Q
to itte ita (no) ka.
that saying was Q
'What time did Hanako say "Wouldn't you pleaee come
(4.11)
r a "moo kite kureruna"]
coming give-Neg. Imp.
b. "moo kite hosikunai"]
coming want-Neg.-Pree.
to itta
that said

Hanako wa [nanzi ni "kite kurenai ha"] to itte

The fact that (4.6) and (4.7) are unacceptable shows that the following what-time come won't-you Q that saying
constrsint hob its (no) ka?
was Q
(4.8) BbnQd Dkowm C b d B h x Blended discourw is poseible 'What time did Hanako say that (lit.) "Won't you come g"?'
only when it involves quasi-direct-discourse. True direct-discourse
cannot appear in blended diwoume. are acceptable because both the quasi-direct discourse and the main clause
verb are in an informal level of speech.
This constraint makes it impossible to interpret the string within quotes of The above phenomenon does not seem to be any Merent fmm the
the lo-marked clause in (4.6) and (4.7) .s direct discoume. Themfore, it is lwel-of-speech phenomenon in nonreported speech. For example, compare
interpreted as quasi-direct diecourse, which, according to our tentative the following three sentences:
hypothesis, is the speaker's rendition of what hake said wing the original
syntax, but adjusted to the speaker's level of speech. The unacceptability of (4.12) a. Hanako wa ["dono sensei ni gomodansitars
the eentences is due to the fact that the highly polite level of the putative which teacher to consult-if (Honorif.)
quasi-direct disco- (as signaled by the polite expressions somimasen ga 'I ii desyoo ka"] to itta
am sorry but' and konaidc Ludasai ' P h don't come') in (4.6) and kite good is (Polite) Q that said
itadekemasen ka 'Wouldn't you please come?' in (4.7)) does not match the 'Hanako said, "Which teacher should I consult?"
speaker's informal lwel of speech signaled by the main clause verb itta 'said' b.??Hanako wa [dono seneei ni " g c m o o d m " to itte
(Informal). which teacher consult (Honor.) that saying
Liiewm, obarve the following sentences: ita (no) h.
was Q
(4.9) Hanako ga b o s y o no ia ni wa 'Which teacher did Hanako say that (lit.) "I will
her house to wmlt"?'
a "*moo konaide kudasaim~enka"] c.??Hanako wa [dono eensei ni g o m o d a ~ ~ ~ ~ka' ~itte
]

{ Neg-coming give-Neg-Polite Q
b. "*moo kite hosiku arimrrsennj
coming wanting not(Po1ite)
to itta
that said ita (no) k a
was Q
which teacher consult (Honor.) Q saying

'Hanako Mid that (lit.) 'Did Hanako say which teacher she would consult?'
a "Would you give me the favor of not comingN
92 / Kuno Blended Quasi-Direct Discourse in Japanese / 93

The to-clause of (4.12a) is entirely in d i r e c t - k n r s e representation, and (4.14) a *Wafakwi wa [Linoo kaimssita] hon o yomimasita
therefore, the lweh of speech between the reported clsnse and the main I y e a t d a y bought (Polite) book red (Polite)
clause do not have to be identical. Hence the acceptability of the sentence. 'I red the book that I bought yesterday.'
The to-clause of (4.12b) is blended discouree. The awkwardnew or b. Wataknsi wa b o o katta] hon o yomimasita
marginality of the sentence, according to our tentative hypothesis, derives I yesterday bought (Informal) book read (Polite)
from the fact that the quasi-direct portion of the to-clause retains the
non-subject honorifiic marking that represents Hanako's deference to the (4.14a) is unacceptable. The informal form kotta muat be used in the
teacher; hence, it is not hlly the speaker's rendition of what Hanako said. relative clall~e, as shown in (4.14b). Similarly, obeerve the following
The sentence is acceptable to those who can interpret the nombject sentences:
honorific verb gosoodamru 'consult' as representing the speaker's deference
to the teacher. The embedded clause of (4.12~)is not direct-, quasi-direct, (4.16) a.??Taroo wa [Hanako ga s e m i kara okarieita]
or indirect-diecourse at sll; it is simply an embedded Q-clam. The teacher from borrowed (Nonsbj Honor.)
sentence is awkward or marginal if goosoodamru 'consult' is interpreted as hon o yonda.
representing Hanako's deference to the teacher, and acceptable if it is book read
interpreted as repmnting the speaker's deference. ' T a m read the book that Hanako borrowed from
The following sentence, howwer, is unexplainable by our tentative the teacher.'
hypothesis: b. T w o wa [Hanako ga sansei kara harita] hon o yondo.
kacher from borrowed book read
(4.13) * H d o ga pima dattara, "ore no uti ni
free if-be my (vulg.) h o w The relative c l a m in (4.168) contains the nonrmbject honorific verb form'
mP' koi"] to denwa o kakete kik. okarisita 'borrowed', which makea the sentence awkward. The fact that the
immediately come(Imp.) that phone placing came following sentence is perfect shows that (4.16.) is c o ~ i d dto be marginal
'Hanako called me up and said that (lit.) if (I) am became it is ditrlcnlt to interpret the nombject honorific form as
free, "come right away to my house." representing the speaker's deference to the teacher

Theoretically, it should be poeeible to interpret hima dattara 'if fnx' as (4.16) Tsroo wa b k u ga sensei kara oksrisitd
indirect discourse, and the rest of the reported speech as quasi-direct. In I teacher from borrowed (Nonsbj Honor.)
the quasi-direct portion of the reported speech, the speaker is supposed to hon o yonda.
use his rendition of what has been said, while maintaining the original book read
syntax, and adjusting the leveb of speech. If the speaker usee ore, an 'Tamo =ad the book that I borrowed from the teacher.'
informal fvat person vulgar pronoun, it should replace Hanako's fvet person
pronoun watakwi. However, (4.13) is totally unacceptable on the intended Whatever the exact condition might be for the use of honorifii verbals in
reading, unless Hanako uses we to refer to hereeIt; that is, unless the whole embedded clauses, it seems that the same condition would account for the
reported speech is interpreted as direct. marginality of (4.12b, c). That is, blended disoouree claaees behave just like
The above fact makes it necessary to sharpen our concept of other embedded c l a m , and require that informal verbal forms be wed in
"speech-level adjustment" to apply only to verbal forms, and not to them.
pronoune. Furthermore, we need to ask why blended diecourse uaea informal Now that we have succeeded in characteris'ig the speech-level
nonhonorific verbal forms predominantly. We hypothesize that the obligatory adjustment of quasi-direct dieooume as belonging to a more general rule of
change from the polite konaide ktidauaimaucn ka and kite hosiku arimauen of speech-level adjustment in subordinate c l a w , we have a h accounted for
(4.10) to the informal moo kite Larcruna and moo kite hosikunai of (4.11) is the unacceptabiity of (4.13). This general speech-level adjustment rule doea
due to a general rule in Japanese that avoide the polite level of speech in not involve adjustment of pronouns:
embedded clauees. For example, obeerve the following sentences:
 Blended Quasi-Direct Discourse in Japanese / 95

The above hypothesis about speech-lwel adjnetment still leaves the

(4.17) a *Watakwi wa [waQ.Lari ga kinoo kaimasita] hon o following sentence unaccounted for:
I (Polite) I (Polite) yesterday bought (Polite) book
yonde simaimasita (4.20) *Handto ga [watakwi no uti ni sugu koi] to denwa o
read I-have (Polite) kakete kita
"I have read the book that I bought yesterday.' 'Hanako called me up and said that (lit.) "Come (male-style,
b. *Wataknsi wa [boku/ore ga kin00 katta] hon o vulgar) to my (polite) house right away".'
I (Informal)/I (Vulgar) bought
yonde simaimasita. It is clear that wga koi is in quasi-direct mode. Since w a t W no uti ni is
c. Watakusi wa [watakusl ga Linoo katta] hon o yonde simaimadta. intended not refer to the external speaker's house, but to Hanako's, it must
be either in direct discourse mode, or in quasi-direct mode. Thus,
The polite iirst-person pronoun wdekwi is not changed in the relative clause theoretically we obtain the following two interpretations of the to-clause of
-- unacceptabiity resulb if the informal fust-person pronoun b o k or the (4.20):
vulgar firat-person pronoun ore is used in the relative clause.
The above generalization makes it posaible to m n n t for the (4.21) ... [ w a h h d n o o t i d mugmkoi] ...
acceptability of sentences of the following pattern: a. Quasi-Direct - Quasi-Direct
(4.18) Hanako-san ga (kanozyo no ie ni mkwi ni
b. Direct - Quasi-Direct
her house to 6 o'clock at The f a d that (4.20) is unacceptable shows that in actuality, neither of these
"kite kure"] to denwa o kakete kimasita two interpretations of wmt& no d i ni is allowed. The unacceptability of
come (Imp.) that phone placing came Interpretation (b) is predicted by the Blended Disoourse Condition given in
'Hanako-san called me up and said that (lit.) (4.8), which states that direct dieconrae cannot be blended with indirect
"Come" to her honse at 6.' discourse. The unacceptability of Interpretation (b), on the other hand,
suggests that that the following constraint holde:
The main clause is in the polite lwel, while the blended diecourse is in the
informal level. This is as prediited by the speech-level adjustment rule for (4.22) Qlud-Direct h u m e Codtiom Quasi-direct discourse
subordinate clauses, and does not require any explanation. Obaerve, further, repreaentation of reported speech is allowable only in
the following sentence: clause-final verb position.
(4.19) a. 1Yamada-sensei ga [sensei no uti ni wa "moo konaide Thim hypothesis is oupported by the fact that it is not poglible to interpret
teacher teacher 'a house to not-coming koko ni 'here' in the following sentence as referring to Hanako's place of
kudssain] to oseyaimasita utterance:
give (Polite) said (Honorific-Polite)
'Teacher Yamada said that (lit.) "Please don't (4.23) Hanako ga (koko ni rmgn koi] to denwa o kakete kita
come" to the teacher's house.' here right-away come that phone placing came
b. Yarnada-sensei ga [sensei no uti ni wa "moo konaide 'Hanako called me up and oaid that "me right now"
kure"] to oeeyaimasita to this place."'
give (Informal) said (Honorific-Polite)
Koko ni in the above sentence rdem only to the place of utterance of the
(4.19a), in which the quasi-direct diacounre and the main verb match in the external speaker, and not to that of Hanako. Thii contrssts mukedly with
lwel of politeness, is at beat awkward. On the other hand, (4.19b), in which the interpretation of koko ni in the following direct discourse sentence:
the quasi-&it discourse is in the informal level, and the main clause verb
in the polite level, is perfect. Thia fact is sleo automatically explainable by
the general speech-level adjustment mle under discussion.
 Blended Quasi-Direct Discourse in Japanese / 97

(4.24) Hanako ga ["koko ni s u p kite kudasaimasen ka" to denwa o but not in colloquial speech, the syntax of the direct speech is preserved, as
here to right-away come won't-you that phone witneeeed by the fact that Subject-Aux Inversion applies. Howwer, it
kakete kita undergoes obligatory tense and pereon agreement.
placing came I showed in Section 1 that the Quasi-indirect" portion of blended
'Hanako called me up and said that "Won't you please discourse in Japanese can be an imperative clause. The following examplea
come here?"' show that the English quasi-indirect discourse pattern cannot be used for
imperati~ee:~
In the above sentence, koko ni readily refem to the went place, that b, to
Hanako's place of utterance. (5.3) a. *Take clrre of herself, firat, Maryi told herself,.
b. *Don't lose her, temper, Maryi told herselfi.

huthermore, sentences that involve quasi-indirect discoure must be decla-

rative -- they cannot be interrogative or imperative:
In this paper, I have examined a style of reported speech in
Japaneee that mixu direct-discourse representation with indirect-discourse (6.4) a. Would she, be punished, Mary, wondered. (= 6.2a)
repmentation, and I have arrived at the following conclusions:
b. *Would she, be punished, did Mary, wonder?
(5.1) a. Japanese has a blended diecoume pattern which wnsists (5.6) a. Have you been kind to her,, you ask Mary, wery day.
of indirect-dieconree and quasi-direct discourse portions. b. 'Have you been kind to her,, ask Mary, wery day.
b. The q u a s i - f i t portion b "direct" in that it retaine the
original syntax (imperative and interrogative) of the quoted The unacceptabiity of (6.4b) and (5.5b) contrasts interestingly with the fact
speech, but "quasi-" in that there ie an obligatory adjustment that sentences that involve blended diecourse in Japanese can freely be
of speech levels. interrogative or imperative:
c. This speech level adjustment follows a g e n e d rule of
Japanese that requires that the informal lwel of speech be used (5.6) a Taroo wa, hontoo ni yatu no uti ni moo
in subordinate clauses. truly hie(vu1g.) house to any-more
d. Only clause-final v e r b of reported speech can be in quasi- kuruna to itta no ha?
direct mode. come-Neg.(Imp.) that said Q
e. Elements in a direct-discourse clause cannot be extracted out 'Did Taroo d y say that (lit.) "Don't comewto his
of that clause. Extraction b poesible only from indirect- house any more?'
discourse or blended-discourse. b. Taroo ni, sugu yatu no uti ni kaere
to immed. his(*.) house to return (Imp.)
There are undoubtedly many other interesting characteristics of this peculiar to ie.
style of reported speech, but the exploration of these characteristics must that say (Imp.)
await future -arch. 'Tell Taroo that (lit.) "Go backw to hie house right away.'
I will end this paper by contrasting the Japanese blended discouree
pattern with the Engbh quasi-indirect discourn pattern. Observe the I showed in Section 4 that blended diswume can wntain wh-
following example: expressions:
(5.2) a. Would she, be punished, Mary, wondered. (5.7) Hanako wa [nanzi ni "kite kurenai ham]to itte
b. When would he come to see heri next, asked Mary,. what-time wme won't-you Q-that saying
ita (no) ka? (~4.11)
In thim mode of reported speech, which is wmmonly nsed in literary writing, 6
 Blended Quasi-Direct Discourse in Japanese / 99
98 / Kuno

very productive p a t h that does not resist application of various syntactic

'What time did Hanako say that (lit.) "Won't you come $P?' proceseee, English quasi-in&& discourse is a fiwen pattern that b
(Expected Answer: She said "Won't you come at six?") insulated from the application of syntactic proceseee. This makes the English
pattern much less interesting as a target of syntactic analysis.
Note that nand ni 'at what time' in the above sentence is bound by the There is the English contrnction that eomewhat resembles the
main clause ka There is no English counterpart to (5.7) because sentences Japanese blended diecourse pattern. For example, obeerve the following
that contain quasi-indiit discourse cannot be questions: sentence:'
(5.8) a. [Shei would return at six], Mary, said. (5.11) John yelled (said, screamed, etc.) to get out of hem right away.
b. *What time [she, would return 91, did Mary, say?
The embedded e x p d o n gei otd oj A m righi awag of the above sentence is
(5.8b) is totally unacceptable. indirect in that the verb is not in the imperative mode. At the same time,
Sentences that contain blended discourse in Japanese can be it is direct in that here refem to the event place, and not to the utterance
embedded, as shown in the following examples: place of the external speaker. It is interesting to see that what goes on in
(5.11) is the complete oppoeite of what goea on in blended discourse in
(5.9) a. [Tanw, ga [yatu no uti ni mgu kite k m ] to itta] Japanese. It was observed in Section 4 that clam-final v e r b are in
him (vulg.) house to irnmed. come (Imp.) that said "quasi-direct" mode in Japanese, maintaining, for example, the imperative
no wa, itizikan-mae da. mode syntactically, but that deictic expressions are in indirect mode, with
that one-hour-ago is koko 'here1, for example, referring to the utterance place of the external
'It was one hour ago that Tanw, said that (lit.) "Comen speaker. In contrast, in the English pattern under diecumion, v e r b are
right away to his house.' indirect, and deictic expressions are in "quasi-directWmode. However, the
b. [Taroo ga (yatu no uti ni turete koi] to itta] following examples show that this pattern is also of extremely limited scope:
his (vulg.) h o w to bring (Imp.) that said
zyaeei wa Hanako d a (6.12) a. *John, yelled to get me, out of here right away.
woman is b. *Johni yelled to get him, out of here [= event place]
'The girl that Tanw, said that (lit.) "Bring" to his house right away.
right away was Hanako.'
As shown by the unacceptability of (5.12a)' pereon agreement is obligatory.
In (5.94, a blended discourse clause is embedded in a cleft construction, At the same time, (5.12b) is unacceptable if here is understood to refer to
whiie in (5.9b), it is embedded in a relative clauae construction. In contrast, the event place. That is, once person agreement is applied to the embedded
sentences that contain quasi-indirect discourse in English cannot be infinitival expression, here laws its "quasi-direct" nature, and can no longer
embedded: refer to the event place. It mxms that the pattern illustrated in (6.1 1) is
limited to a few idiomatic expressions, and cannot be considered a productive
(6.10) a. She, should never forgive that man, Mary, told herselfi' one.
b. *This is the man whom she, should never forgive, Mary,
told herself,.
FOOTNOTES
(5.10b) might be acceptable if Thk is the man whom she shoald never forgive
as a whole ie interpreted as representing Mary's internal utterance to herself, Research reported on in this paper has been supported in part by a
but it is totally unacceptable as a sentence in which only she shodd never grant from the National Science Foundation to Harvard University (Grant
forgive constitutes Mary's internal utterance. The latter interpretation should No. BNS 82 14193). I am greatly indebted to Iioes King, who has given
be poasible if Mary told herself were a relative clause constituent. me numerous invaluable comments on the previous version of the paper.
The above contrasts show that while Japanese blended discourse is a
100 / Kuno Blended Quasi-Direct Discourse in Japanese / 101

[I] In main clauses in Japanese, the wa-marked subject and the the pattern of (i), in which ...mka and -no- do not appear to be clauee
ga-marked subject show the following semantic contrast: matea, are acceptable:

(i) a. Taroo wa oobaka d a (i) Taroo ni wa eigo sika hanasu koto ga d e w .

biggest-fool is to English only speak to able-Neg-Pres.
(i) Thematic Interpretation: 'Only English can Taroo speak.'
'Taroo is the biggest fool.'
(i) Contrastive Interpretation: He hypotheeieed that Predicate W i g applies to hanow (koto go) and
'Aa for Taroo, he b the biggest fooL1 makes it a part of the main-clause verb complex:
b. Taroo ga oobaka d a
Exhaustive-listing Interpretation: (ii) Taroo ni wa eigo s h ,[ hanasu koto ga d e k i ~ i ] .
'It's Taro0 that is the biggest fool.'
It would be difficult to extend this analysb to sentence8 of the pattern of
The exhaustive listing interpretation is obligatory for the ga-marked subject (2.lb). Note that the verb of reported speech does not have to be in the
of stative predicates. This contrast is neutralized in subordinate clauses - Preaent Tense form:
the subject of subordinate clauses is regularly marked with ga without
necessitating an exhaustive-Wig interpretation: (iii) Taroo wa sono tokei sika nusunda to hakuzyooeinakatta.
the watch only steal-Past that confepl-Neg-Past
(ii) Hanako wa [Taroo g oobaka na] koto o siranai. 'Only the watch did Taroo confese that he had stolen.'
biggest-fool is that know-not
'Hanako doew't know that Taroo is the biggest fool.' Also, adverbial elements can be inserted between the to-clause and the
reporting verb, albeit with some difficulty:
It is not poasible to mark the subject of the embedded clause with mo unleea
it is wed contrastively, with the eecond argument of contrast present in the (i) ?Taroo wa sono tokei sika n u m d a to keisatu ni
clause: the watch only atole that police to
hakuayoo8imkatta
(iii) *Hanako wa [Taroo wa oobaka na] koto o siranai. cod--Neg-Past
(iv) Hanako wa [Taroo wa oobaka da ga, Z h wa tensai de aru] 'Only the watch did Taroo contess to the police that he had
biggest-fool is but genius is stolen.'
koto o sitte irn.
that knowing is If Predicate Raising had taken place eo as to make nvarnda to 'that ...
'Hanako knows that Taroo is the biggest fool, but that Ziroo stole' and hakutl(oosinakatta 'confese-Neg-Past' a complex verbal, sentence
is a genius.' (iv), in which ktisatu ni appears between these two elements, would be
totally unacceptable. But (iv) is only awkward, and is much better than
The fact that (1.2b) does not necessitate an exhaustive listing interpretation expected.
of Taroo ga shows that it is in a subordinate clause. Furthermore, it ie poeaible to conjoin "V + ...
ton using to La ... to
La '...and...and...':
[ t ] As far as I know, this was first obeerved in Kuno (19788).
(v) Taroo wa rock music dL. utaitai to La hikitai
[S] It is irrelevant whether the command concept needed here is that of only sing-want-Pres. and play-want-%.
S(entence)-command, c(onstituent)-command, or k-command. to ka kikitai to i w d a t t a
and hear-want that say-Neg.-Past
[ I ] As far as I know, Muraki (1978) was the first to note that sentences of 'Only rock music did Taroo say that (lit.) "(I) want to sing,'
102 / Kuno

that "(I) want to play", and that "(I) want to hear".' . Whether we or not:
a comparative syntax of English and Japanese
Obeerve that to ka in the above sentence is a contraction of the reporting S.-Y.Kuroda
particle to and the coordinating e x p m i o n to k a The acceptability of (v) UCSD
makes Muraki's Predicate Raisiig analysis even lese plaueible.
The ah-raising analysis propoeed in the present paper is free from 0. Introduction.
all the above difficulties.
English has visible wh-movement; Japanese doesn't. Japanese scrambles and
word order is free; English doesn't scramble and has an orderly word order.
(61 Note that while yam defines a "larger than or equal ton relationship, The topic is prominent in Japanese; it is not in Ehglish. Japanese has double-
k r e m defies only a nsmaller thanw relationship, and doea not include a or multiple-subject structures; English does not. Such are the major
"equal to" relationship. See Kuno (1978b) for the justification of this typological differences between bglish and Japanese, and sane linguistic
difference. theoreticians entertain the idea that parametric differences concerning Deep
Structure exist between English and Japanese which are responsible for these
differences. It has been proposed that English is configurational while
(61 The following sentence might appear to be a counterexample to the Japanese is nonconfigurational; cf: Hale (1980), Chomsky ( 1981), m n g others.
claim that imperatives cannot be used in quasi-indirect discourse: Or it has been suggested that Japanese clauses are Max(V), while English ones
are Max(1); for example, Chomsky in a lecture at UCSD, 1985. I would like to
(i) Do this immediately, Mary ordered Jane. sketch in this paper a claim to the contrary that there is no parametric
difference between English and Japanese that results in essentially different
However, there does not seem to be any compelling reason to assume that deep strudure configurations. Instead, the parametric difference between
(i) is in quasi-indirect discourse mode rather than in direct discourse mode. English and Japanese consists simply of the follow5.ng: Agreement is forced
English; it is not in Japanese.1
[I] I am indebted to Susan Fisher and Peter Sells for sentences of the 1. Completed X-bar theory
pattern in (5.11).
Chomsky proposes to generalize X-bar theory to the categories C
(Complementizer) and I (INFL), a welcome generalization. (Chomsky 1986)
REFERENCES Thus, we have a schema of D-structure well-fonnedness conditions:

Kuno, S. (1978a), "Japanese: A Characteristic OV Languagen, in Lehmann,

W . (Ed.), Syntactic Typology, University of Texas h, pp. 57-138.
where Max(W) (W = X, Y, Z) is the maximal projection, W' the bar-level
Kuno, S. (1978b), Danwa no Bunpoo, Tsiehnkan Publ. Co., Tokyo. category, and W the zero-level (basic) ca-ory, of categor~rtype W, and where
W ranges over lexical categories (N, A, V, P) as well as over two clausal
augmentizers, C(omp1ementizer) and I(NFL). The linear order of the
Kmo, S. and E. Kabnraki (1977), "Empathy and Syntax", Linguistic Inquiry constituents is language specific. (1) represents the linear order for
- - -

8, pp. 626-672. English.

Muraki, M. (1978), "The Sika Nai Construction and Predicate Restructuring",
Max(Y) , which is a sister of Xes in the above schema, is usually called
Spec(X) (the Specifier of X) , and Max(Z) in the above schema, the Complement
1
I

in Hinds, J. and I. Howard (Eds.),Problems in Japanese Syntax and of X. Instead, I will call them Ext(X) and Int(X) (External and Internal
Semantics, Kaitakuaha, Tokyo, pp. 156-177. Complement of X) , respectively. 2
I have departed from Chomsky, however, on the formulation and construal of
the schema of the X-bar theory. C h m k y formulates the schema in terms of
1
I
Kleene's star, Max(X)*, indicating an arbitrary number of concatenated copies. I
Instead I ass- that concatenation of a Max to a nonmaximal c a w o w in the
schema is construed as possibly involving Chomsky-adjunction. I assume, given