30
The grammaticalization of nominalizers in Japanese and Korean: A contrastive study * Kaoru Horie Tohoku University 1. Introduction Nominalization in Japanese, as in many other languages, is an intricate phenomenon which exhibits morphosyntactic, semantic and pragmatic dimensions. What makes nominalization in Japanese further intriguing as an object of inquiry are (i) its diachrony, in particular its grammaticalization process (Horie 1997, 1998a; Simpson and Wu 2001; Simpson 2003), and (ii) its typological and areal linguistic properties (cf. Horie 1998b, 2000; Noonan 1997). The present study analyses nominalization in Japanese as a prominent grammatical feature of the language which has evolved since ancient times into its current complex system, shared with other languages of East Asia. Specifically, this study addresses the following two questions: (a) What kind of grammaticalization patterns do Japanese nominalizers exhibit? (b) How do the grammaticalization patterns observed in Japanese contrast with those found in Korean, an East Asian language with similar typological characteristics to Japanese? * This study was partly supported by a grant from the Japan Society for the Promotion of Science (#15520242). Special thanks are due to the two anonymous reviewers for their penetrating and very helpful comments. Thanks also go to the participants of the workshop on the grammaticalization of nominalizers in East Asian languages at New Reflections on Grammaticalization 3 (July 2005, Santiago de Compostela, Spain), particularly Foong Ha Yap and Mickey Noonan, and the participants of the 4th Meeting of the Japanese-Korean Contrastive Study Seminar (November 2005, University of Tokyo, Japan), particularly Naoki Ogoshi and Kim Joung-Min, for constructive criticism and comments.

HORIE Inpress Japanese

Embed Size (px)

DESCRIPTION

HORIE Inpress Japanese

Citation preview

Page 1: HORIE Inpress Japanese

The grammaticalization of nominalizers in Japanese and Korean: A contrastive study*

Kaoru Horie

Tohoku University

1. Introduction

Nominalization in Japanese, as in many other languages, is an intricate phenomenon which

exhibits morphosyntactic, semantic and pragmatic dimensions. What makes nominalization

in Japanese further intriguing as an object of inquiry are (i) its diachrony, in particular its

grammaticalization process (Horie 1997, 1998a; Simpson and Wu 2001; Simpson 2003),

and (ii) its typological and areal linguistic properties (cf. Horie 1998b, 2000; Noonan 1997).

The present study analyses nominalization in Japanese as a prominent grammatical

feature of the language which has evolved since ancient times into its current complex

system, shared with other languages of East Asia. Specifically, this study addresses the

following two questions:

(a) What kind of grammaticalization patterns do Japanese nominalizers exhibit?

(b) How do the grammaticalization patterns observed in Japanese contrast with those found

in Korean, an East Asian language with similar typological characteristics to Japanese?

* This study was partly supported by a grant from the Japan Society for the Promotion of Science

(#15520242). Special thanks are due to the two anonymous reviewers for their penetrating and very helpful

comments. Thanks also go to the participants of the workshop on the grammaticalization of nominalizers in

East Asian languages at New Reflections on Grammaticalization 3 (July 2005, Santiago de Compostela,

Spain), particularly Foong Ha Yap and Mickey Noonan, and the participants of the 4th Meeting of the

Japanese-Korean Contrastive Study Seminar (November 2005, University of Tokyo, Japan), particularly

Naoki Ogoshi and Kim Joung-Min, for constructive criticism and comments.

Page 2: HORIE Inpress Japanese

The organization of the paper is as follows. Section 2 is devoted to a brief historical

background of nominalization in Japanese. Section 3, in turn, presents an analysis of the

grammaticalization of Japanese nominalizers as contrasted with that of Korean

nominalizers and it further explores theoretical implications of the crosslinguistic findings.

Finally, Section 4 provides the conclusions.

2. Nominalization in Japanese: Historical background

One of the noticeable features of Japanese grammar, arguably on a par with, or even more

prominent than, well-known grammatical phenomena like topicalization and honorification,

is the extensive use of sentential nominalizers. Sentential nominalizers in Japanese appear

not only in intra-sentential position, like argument positions, where they mark complement

clauses, but also in sentence-final position, either independently or together with the

copulas da / dearu / desu ‘to be’ (plain style / formal style / polite style), where they

convey a variety of modal and aspectual meanings.

Maynard (1997:113), drawing attention to a Japanese sentence from Abe (1968:6)

involving nominalization (cf. (1) below) and its English non-nominalized counterpart (1’)

from Saunders’ (1969:6) translation, makes the following insightful observation: “There is

a distinct preference for nominalized expressions, at least in some Japanese discourse.

When original Japanese and English translations are compared, there are many cases where

the Japanese writer uses nominal clauses, but the English translator does not” [emphasis

added].

(1) Modern Japanese

Kono ame-ga agare-ba moo sugu natu na no daroo

Page 3: HORIE Inpress Japanese

this rain-NOM let up-COND already soon summer be NMLZ would be

‘(lit.) It would be that, when the rains let up, it will be summer soon.’

(1’) Soon, when the rains let up, it would be summer

In fact, the non-nominalized counterpart to (1) is also possible in Japanese, as shown in

(1’’).

(1’’) Modern Japanese

Kono ame-ga agareba moo sugu natu daroo

As extensively discussed by many Japanese grammarians (for example, Kuno 1973,

among many others), there are several semantic / pragmatic differences between (1) and

(1’’), such as presence (nominalized version) versus absence (non-nominalized version) of

a preceding linguistic or non-linguistic context leading the speaker to conclude that

summer will come soon. Although I will not go into details regarding such semantico-

pragmatic differences between sentences with and without nominalization, it is important to

note that nominalization by overt sentential nominalizers like no is not a marked rhetorical

strategy; on the contrary, it can be extensively observed across genres, both in speech and

writing.

The preference for nominalization, noted by Maynard (1997), is not an innovation in

Modern Japanese, but has its roots in the history of the language. This is briefly discussed

in Section 2.1 below.

2.1. Nominalization in Classical Japanese

Page 4: HORIE Inpress Japanese

Some morphosyntactic characteristics of classical Japanese have been completely lost.

Notable among them is the distinction between attributive (nominal) predicate form and

conclusive (sentence-final) predicate form, as shown in (2a) and (2b), respectively.

(2) Classical Japanese

a. Oturu tori

fall:ATTR bird

‘a falling bird’

b. Tori otu

bird fall:CONC

‘a bird falls’

Besides modifying nouns, as in (2a), attributive (nominal) forms in Classical Japanese,

were employed as a means of forming a nominalized clause in argument position

(complement clause) or adjunct position (adverbial clause) without any overt nominalizer,

as in (3).

(3) Classical Japanese

[Tori oturu]-wo mi-tari

bird fall:ATTR-ACC see-PERF

‘I saw a bird falling.’

Attributive forms were also employed in sentence-final position when the sentence showed

the emphatic or focus particles zo or namu, or the interrogative particles ya or ka in pre-

final position, as in (4).1

1 Examples (4) and (5) are from Iwasaki (2000); glosses are partially modified.

Page 5: HORIE Inpress Japanese

(4) Classical Japanese

Ware-nomi-zo kimi-ni-wa kouru

I-only-FP you-DAT-TOP yearn

‘It is I alone who yearn for thee.’

These sentences ending in attributive forms in Classical Japanese served some rhetorical or

pragmatic functions and are referred to as kakari musubi bun (‘focus-concord sentence’) in

traditional Japanese grammar.

Furthermore, attributive forms were employed in sentence-final position even when

no focus or interrogative particle was present, as in (5).

(5) Classical Japanese

Tiru-to mite / aru-beki mono-wo / ume-no-hana // utate

scatter-QT see:GER exist-should thing-PRT plum-GEN- flower very

nioi-no / sode-ni tomare-ru

scent-GEN sleeve-LOC lay-PERF:ATTR

‘It would have been best simply to watch them scatter – now, alas, the scent from the

blossoms of the plum still lingers upon my sleeve.’

Sentences of this kind ending in attributive forms without any overt particle served to

convey the writer’s emotive or affective stance toward the proposition expressed in the

sentence, such as vivid recollection of a past event experienced by the writer. These are

referred to as rentai syuusi bun (‘attributive-final sentences’) in traditional Japanese

grammar.

Page 6: HORIE Inpress Japanese

The existence of examples like (4) and (5) ending in attributive forms suggests that,

even in Classical Japanese, it was not uncommon for sentences to end with some kind of

nominalization, parallel to nominalized sentences in Modern Japanese, as in (1) above. In

fact, sentences ending in attributive forms in Classical Japanese became the norm rather

than the exception because of a morphosyntactic change whereby conclusive (sentence-

final) forms (e.g. (2b)) were replaced by attributive forms (e.g. (2a)). Iwasaki (2000:243)

notes that “[t]he distinction between Conclusive and Attributive forms gradually

disappeared, and by the 14th century, the old Conclusive form of some verbs had been

replaced by the Attributive form. In other words, the same (Attributive) form now functions

both as the Conclusive and as the Attributive” [emphasis added]. Attributive forms in

Modern Japanese have thus become indistinguishable from conclusive forms, as shown in

(2a’) and (2b’).

(2) Modern Japanese

a’. otiru tori

fall:ATTR bird

‘a falling bird’

b’. Tori-ga otiru

bird-NOM fall

‘A bird falls.’

From the perspective of grammaticalization, the replacement of conclusive forms by

attributive forms suggests that attributive forms extended their functional domain over time.

Some of the functions served by Classical Japanese attributive forms, particularly

Page 7: HORIE Inpress Japanese

complement-marking function (cf. example (3) above) and pragmatic or rhetorical

functions (cf. examples (4) and (5)) have thus come to need more overt nominalization

marking, that is, overt nominalizers such as no. Section 3 below presents a contrastive

analysis of the grammaticalization of overt nominalizers in Japanese and Korean in order to

address the questions posed in Section 1 above.

3. Grammaticalization of nominalizers in Japanese and Korean

The sentential nominalizing function of Classical Japanese attributive forms, which

replaced conclusive forms (cf. Section 2.1 above), has its direct descendent in Modern

Japanese, a ‘non-overt’ nominalization process referred to as “direct nominalization” (cf.

Martin 1975), as in (6).

(6) Modern Japanese

[Tometa] -ni mo kakawarazu, dete it-ta

stop:PAST: ATTR-DAT also concern:NEG leave:GER go-PAST

‘Though I stopped, he departed.’

This nominalization process is, however, no longer productive in Modern Japanese and is

limited to “relics” (Harris and Campbell 1995), like conventionalized idiomatic expressions,

such as -ni mo kakawarazu ‘notwithstanding’ in example (6) above. Barring such relics,

Modern Japanese needs to employ overt sentential nominalizers like no, as in (3’), the

Modern Japanese counterpart to (3).

(3’) Modern Japanese

[Tori-ga otiru no]-wo mi-ta

bird-NOM fall:ATTR NMLZ-ACC see-PAST

Page 8: HORIE Inpress Japanese

‘I saw a bird falling.’

3.1. Grammaticalization of overt nominalizers in Japanese

Overt sentential nominalizers in Modern Japanese, referred to as keisiki meisi (lit. ‘formal

noun’) in traditional Japanese grammar, either have their lexical origin not confirmed,

which is the case with no (its earliest documented function being that of genitive), or have

originated from lexical nouns, as is the case with koto (< ‘matter, event’), tokoro (< ‘place’),

mono (< ‘thing’) or wake (< ‘reason’). All these overt nominalizers constitute a finite set of

forms which are stratified in three groups according to their different degrees of

grammaticalization, that is, their degrees of functional extension and semantic

generalization:

(a) Lexical meaning virtually absent; functional extension widest of all the nominalizers: no

(cf. also Horie 1998b; Yap et al. 2004; Nishi 2006).

(b) Very general or abstract lexical meaning; rather wide functional extension: mono (<

‘thing’), koto (< ‘event, matter’), tokoro (< ‘place’).

(c) More specific lexical meaning; limited functional extension: wake (< ‘reason’), yoo (<

‘appearance’), moyoo (< ‘design, likelihood’), kanzi (< ‘feeling’), etc.

What seems to be rather special about the grammaticalization of overt nominalizers

in Japanese is their multiple grammaticalization pathways. As shown below, these

nominalizers have developed a variety of grammatical uses, namely (i) complementizer; (ii)

cleft construction marker; (iii) marker of internally headed relative clauses; (iv)

conjunction; (v) modal and aspectual constructions marker; and (vi) sentence final particle.

In what follows, I will provide illustrative examples for each grammatical use, and specify

Page 9: HORIE Inpress Japanese

which of the three groups of overt nominalizers above have developed the uses in question,

since, as already mentioned, not all nominalizers behave alike in terms of their functional

extension.

(i) Complementizer. Three nominalizers, no in group (a) and koto and tokoro in (b), have

evolved into complementizers, as shown in (7a-b) below.

(7) Modern Japanese

a. [Kodomo-ga nai-te i-ru no]-o mi-ta

child-NOM cry-GER exist-PRES NMLZ-ACC see-PAST

‘I saw a child crying.’

b. [Kodomo-ga nai-te i-ru tokoro]-o mi-ta

child-NOM cry-GER exist-PRES NMLZ-ACC see-PAST

‘I saw a child crying.’

(ii) Cleft construction marker. Only one nominalizer, namely no in group (a), has evolved

into a cleft construction marker. Consider example (8).

(8) Modern Japanese

[Souru-ni tootyakusi-ta] no-wa sanzi desi-ta

Seoul- to arrive-PAST NMLZ-TOP three o’clock COP-PAST

‘It was at three o’clock that I arrived in Seoul.’

(iii) Marker of internally headed relative clauses. This is a particular type of relative clause

which, unlike regular externally headed relative clauses, shows its head occurring

internal to the clause. Only one nominalizer, no in group (a), has fully developed into a

marker of internally headed relative clauses. Another nominalizer, tokoro in group (b),

Page 10: HORIE Inpress Japanese

can serve this function only with a limited number of matrix verbs, like tukamaeru

(‘catch, arrest’). Thus, for instance, it would be infelicitous in an example such as (9a).

This co-occurrence restriction is absent with no, which could appear in (9b). This

suggests that the lexical meaning of tokoro, though fairly abstract, has not been

completely bleached.

(9) Modern Japanese

a. [Tukue-no ue-ni ringo-ga aru no]-o tabe-ta

desk-GEN top-LOC apple-NOM exist NMLZ-ACC eat-PAST

‘I ate the apple, which / as it was on the desk.’ (lit. ‘I ate that an apple was on the

desk.’)

b. [Doroboo-ga heya-kara dete ki-ta tokoro]-o tukamae-ta

thief-NOM room-from come out-PAST NMLZ-ACC catch-PAST

‘I caught a thief who / as he was coming out of the room.’

(iv) Conjunction. The nominalizers no, mono, koto and tokoro have evolved into

conjunctions by either coalescing with co-occurring case particles, like nominative ga,

accusative o, dative / locative ni and instrumental / locative de, or independently, which

is the case of tokoro only. The list of conjunctions includes the following forms: in

group (a), no-de (‘because’), no-ni (‘though’), etc.; in group (b), mono-o (‘though’),

mono-no (‘though’), koto-de (‘because’), tokoro-de (‘no matter how, incidentally’),

tokoro-ga (‘contrary to one’s expectation, however’), tokoro (‘in a situation where, as it

happens’), etc. Illustrative examples are given in (10a-b).

(10) Modern Japanese

Page 11: HORIE Inpress Japanese

a. [Yohukasi si-ta no-de], neboosi-te simat-ta

sit up late-PAST NMZL-INST oversleep-GER end up-PAST

‘Because I sat up late, I overslept [and I could not help it].’

b. [Kyoositu-ni it-ta tokoro], kyuukoo dat-ta

classroom-to go-PAST NMLZ cancellation COP-PAST

‘When I went to a classroom, I found [to my surprise] that the class was

cancelled.’

(v) Marker of modal and aspectual constructions. Nominalizers of all three groups (a)-(c)

have developed this use by coalescing with the copula da: in group (a), no-da (causality,

‘it is that / because …’; cf. Kim and Horie, in press); in group (b), mono-da (moral

obligation, past habit), koto-da (moral obligation), tokoro-da (progressive / proximative

/ immediate past), etc.; finally, in group (c), wake-da (explanation), yoo-da (inferential),

moyoo-da (inferential), etc. Consider the examples in (11a-c) below.

(11) Modern Japanese

a. A, ame-ga hut-ta n da

oh rain-NOM fall-PAST NMLZ COP

‘(I infer from some evidence that) it rained.’ (lit. ‘It is that it rained.’)

b. Yoku mukasi-wa umi-ni oyogi-ni it-ta mono-da

well past-TOP sea-to swim:CONJ-to go-PAST NMLZ-COP

‘In the past I would go to sea for swimming.’

c. Ame-ga hut-ta yoo-da

rain-NOM fall-PAST NMLZ-COP

Page 12: HORIE Inpress Japanese

‘It appears to have rained.’

(vi) Sentence-final particle (with discourse-pragmatic meaning). Nominalizers of all three

groups (a)-(c) have developed this use: (a) … no (offering explanation or confirmation);

(b) … mono (conveying one’s opinion), … koto (strong advice); (c) … wake (offering

explanation or justification).

(12) Modern Japanese

a. Hayaku iku no

quickly go NMLZ

‘Will you leave early?’ (lit. Is it that you leave early?) or ‘Please leave soon.’ (lit.

It is that you leave quickly.)

b. Sira-nai mon

know-NEG NMLZ

‘I don’t know [I assert this no matter what you say].’

c. Asoko-ni ikoo tte itta wake

that place-to go:INT QT said NMLZ

‘I said why not go there, you know.’

An interesting question which arises here is whether the grammaticalization patterns

(i)-(vi) found in Japanese nominalizers are typologically common or, on the contrary,

language-specific. As noted by Heine and Kuteva (2002:326) and other researchers in

grammaticalization studies, it is not uncommon for languages to evolve complementizers

from lexical nouns with generic meaning like ‘thing.’ However, this grammaticalization

pathway does not straightforwardly apply to Japanese, as the most relevant nominalizer

Page 13: HORIE Inpress Japanese

mono (< ‘thing’) has not developed the complementizer function. Instead, Japanese shows

some grammaticalization pathways not widely documented in other languages, like ‘place’

> complementizer (see 7b) and ‘place’ > conjunction (see 10b), which apply to the

nominalizer tokoro, or ‘thing’ > sentence-final particle (see 12b), which applies to the

nominalizer mono (cf. also Horie and Sassa 2000).

Moreover, it is not uncommon for some languages, particularly those in East Asia, to

exhibit a grammaticalization pathway from nominalizers to pragmatic markers (cf. Yap et

al. 2004). In fact, as shown in the overview of nominalizers in East Asian languages by

Yap and Matthews (this volume), many of the pathways documented in Japanese have also

been observed in other languages in the same region, which points to the areal nature of

versatile nominalization. However, it appears that the multiple grammaticalization

pathways (i)-(vi) observed in Japanese are not necessarily common even in East Asian

languages where such phenomena would be expected. In order to correctly assess the

grammaticalization pathways of Japanese nominalizers, I will compare them with their

Korean counterparts, a language known to exhibit remarkable structural similarity to

Japanese relative to other languages (cf. Horie 1998a, 1998b, 2000, 2002a, 2002b, 2003;

Horie and Taira 2002).

3.2. Grammaticalization of overt nominalizers in Korean

Similarly to Japanese, Korean has a group of nominalizers (or dependent nouns)

historically derived from lexical nouns (cf. Rhee this volume). Like the Japanese

nominalizers presented in Section 3.1 above, the membership of Korean nominalizers is

stratified in terms of differential degrees of grammaticalization, that is degrees of

Page 14: HORIE Inpress Japanese

functional extension and semantic generalization. One noticeable difference between the

overt nominalizers in Korean and their Japanese counterparts is that Korean does not

appear to have any nominalizer which has reached the same stage of grammaticalization as

Japanese no (i.e. lexical meaning virtually absent, functional extension widest of all the

nominalizers). The stratification of overt Korean nominalizers is then restricted to groups

(b) and (c): (b) very general or abstract lexical meaning and rather wide functional

extension: kes (< ‘thing’); (c) more specific lexical meaning and / or limited functional

extension: tey (< ‘place’), il (< ‘event, matter’), cek (< ‘event’), moyang (< ‘appearance’),

ttaymwun (< ‘reason’), etc.

Let us consider now the different grammatical uses of overt nominalizers in Korean.

(i) Complementizer. Only one Korean nominalizer, namely kes, from group (b), has

evolved into a complementizer (cf. example (13) below).

(13) Korean

[Phiano-lul chi-nun kes]-ul tul-ess-ta

piano-ACC play-ATTR:PRES NMLZ-ACC hear-PAST-DECL

‘I heard him playing the piano.’

(ii) Cleft construction marker. Once again, only the nominalizer kes has evolved into the

marker of cleft constructions, as shown in (14).

(14) Korean

[Chayksang wi-ey sakwa-ka iss-nun kes]-ul

desk top-LOC apple-NOM exist-ATTR:PRES NMLZ-TOP

‘It is an apple that is on the desk.’

Page 15: HORIE Inpress Japanese

However, kes is not felicitous in a sentence parallel to (8) in Japanese.

(8’) Korean

? [Taroo-ka sewul-ey tochakha-n kes]-un seysi i-ess-ta

NOM Seoul-to arrive-ATTR:PAST NMLZ-TOP three o’clock be-PAST-DECL

‘It was three o’clock that I arrived in Seoul.’ (Intended meaning: It was at three

o’clock that I arrived in Seoul.)

This indicates that Korean kes has not reached the same stage of semantic generalization as

its closest counterpart no in Japanese. Hence, its remaining lexical meaning prevents kes

from serving as a generalized cleft construction marker unlike Japanese no or English that.

(iii) Marker of internally headed relative clauses. Unlike its Japanese counterpart no in

example (9) above, kes is not fully acceptable in this construction (9’), as was the case

with the low acceptability of its use in cleft constructions in (8’).2

(9’) Korean

? [Chayksang wi-ey sakwa-ka iss-nun kes]-ul mek-ess-ta

desk top-LOC apple-NOM exist-ATTR:PRES NMLZ-ACC eat-PAST-DECL

‘I ate the apple, which / as it was on the desk.’ (lit. ‘I ate that an apple was on the

desk.’)

(iv) Conjunction. Unlike their Japanese counterparts, not many Korean overt nominalizers

with lexical origin have evolved into conjunctions, either by coalescing with co-

occurring morphemes or independently. A possible exception is tey ( < ‘place’), from

2 Cf. Horie (1993) for a more extensive analysis.

Page 16: HORIE Inpress Japanese

group (c), which has formed conjunctions by coalescing with the present attributive

ending -nun, as in nuntey (‘but, because, as, etc.’), or with the attributive form of the

stative predicate kuleta (‘to be so’), giving rise to kulentey (‘however’), illustrated in

example (15) below.

(15) Korean

Kulentey, imi syatha-ka tathye iss-ess-ta

NMLZ already shutter-NOM be closed exist-PAST-DECL

‘But the shutter (of the bank) was already closed.’

(v) Marker of modal and aspectual constructions. Like their Japanese counterparts, many

Korean nominalizers have evolved this use by coalescing with the copula –ita, such as

kes-ita (explanation, inference, suggestion), from group (b), or moyang-ita (inferential),

cham-ita (proximative), etc., from group (c). Consider the illustrative examples in (16a-

b) below.

(16) Korean

a. Hoyuy-ey chamsekha-l kes-ita

meeting-to attend-ATTR:FUT NMLZ-COP:DECL

‘He will attend the meeting.’

b. Kim sensayngnim-kkey cenhwaha-l cham-i-ess-e-yo

professor-DAT:HON telephone-ATTR:FUT verge-COP-PAST-CONJ-SFP:POL

‘I was just about to call Professor Kim.’

(vi) Sentence-final particle (with discourse-pragmatic meaning). Some Korean

nominalizers with lexical origin have evolved into sentence-final particles, among them

Page 17: HORIE Inpress Japanese

the following: … kel (< kes + lul ‘NMLZ-ACC’) and … nuntey (< nun + tey ‘Attributive

Present ending + place’), from groups (b) and (c) respectively, as in (17a-b).

(17) Korean

a. Ne-to kathi ka-l kel

you-also together go-ATTR:FUT NMLZ

‘You should have gone with them.’

b. Ku chayk-un yeki eps-nuntey-yo

that book-TOP here not exist- NMLZ:POL

‘We don’t carry that book here (I am sorry to say).’

The crosslinguistic comparison between Japanese nominalizers and their Korean

counterparts presented above has revealed that some pathways found in Japanese are not

attested in Korean or are attested to a lesser extent. Conversely, no grammaticalization

pathway has been observed only with Korean nominalizers. More specifically, the

crosslinguistic comparison between the Japanese nominalizer no and its Korean counterpart

kes presented in this section has shown that, in spite of the extensive parallelism between

the two forms, kes has not been as fully grammaticalized as no, since some pathways found

with Japanese no are not manifested at all for Korean kes, or manifested only to a lesser

extent.

Though less noteworthy than the contrast between no and kes, another pair of nominalizers,

tokoro and tey, both originating as lexical nouns meaning ‘place,’ manifest a similar con-

Page 18: HORIE Inpress Japanese

trast to that between no and kes in terms of their degree of grammaticalization.3 Japanese

tokoro has developed more grammatical uses (complementizer, internally headed relative

clause marker, conjunction and marker of aspectual constructions) than Korean tey, which

can only function as a conjunction and a sentence-final particle.

Table 1 summarizes the crosslinguistic comparison of the grammaticalization of

Japanese and Korean nominalizers discussed in this paper:

Table 1. Grammaticalization of Japanese and Korean Nominalizers Contrasted

Japanese Korean

(i) Complementizer no, koto, tokoro kes (ii) Cleft construction no (kes) (iii) Internally headed relative clause no, tokoro (kes) (iv) Conjunction no, mono, koto, tokoro tey (v) Modal and aspectual constructions no, mono, koto, tokoro etc. kes etc. (vi) (Utterance) final particles no, mono etc. kes, tey etc.

(‘( )’ indicates that the item in question not having been fully grammaticalized in that func-

tion)

3.3. Theoretical implications of the Japanese-Korean contrast

The grammaticalization patterns of overt nominalizers presented in Sections 3.1 and 3.2,

particularly those exhibited by the Japanese nominalizers no and tokoro and their Korean

3 See Horie and Sassa (2000) for a more extensive analysis of the grammaticalization pathways of tokoro and

tey.

Page 19: HORIE Inpress Japanese

counterparts kes and tey, suggest that the former instantiate a higher degree of grammati-

calization than the latter. This contrast leads to an interesting theoretical question, inspired

by Bybee et al. (1994), regarding the degree of grammaticalization in languages of a similar

morphological type. Bybee et al. (1994:118) propose the following hypothesis:

The existence of (morphological) types gives evidence for typological constraints on

grammaticization: in some languages grammaticization does not proceed as far as it does in

others. In particular, isolating languages do not carry grammaticization as far as fusional or

agglutinating languages do. Not only do they not affix, they also do not have grams with

meaning as abstract as synthetic languages do. The stability of certain isolating languages,

such as Chinese, over time further attests to typological constraints on grammaticization.

Bybee et al.’s (1994) hypothesis makes it possible to predict that grammaticalization in an

isolating language like Chinese does not proceed as far as in an agglutinating language like

Japanese. The contrast between the grammaticalization patterns of Japanese and Korean

nominalizers shown here suggests that the rate at which grammaticalization proceeds may

vary even between languages of the same morphological type.

Crucially, the tendency for Korean to exhibit a lesser degree of grammaticalization

than Japanese is not only manifested in the grammaticalization of overt nominalizers, as

seen in Sections 3.1 and 3.2, but is also observable in the grammaticalization of verbal

elements, such as periphrastic aspectual constructions. According to Bybee et al. (1994),

grammatical morphemes (or grams in their terminology) encoding tense, aspect and

modality do not evolve randomly from any lexical item, but manifest common

grammaticalization pathways crosslinguistically from lexical items with specific lexical

meaning, among them the following: (i) ‘be / have’ > resultative > anterior (= perfect) >

Page 20: HORIE Inpress Japanese

perfective / simple past; (ii) ‘finish’ > completive > anterior (= perfect) > perfective /

simple past. The pathway in (i) is attested in both Japanese and Korean:4

(a) Japanese: -te ari (conjunctive affix-‘be’ (existential) verb) > -tari (affix: resultative /

perfect) > -ta (affix: perfective / simple past).

(b) Korean: -e is- / -e isi- (conjunctive affix-‘be’ (existential) verb) > -eys- / -eysi- (affix:

resultative / perfective) > -ess- (affix: perfective / simple past).

The affixes -ta and -ess- have already reached the final stage of the pathway since they can

encode simple past tense.

Japanese and Korean also have less grammaticalized periphrastic aspectual

constructions whose constituent lexical items (in this case verbs) have not yet been

semantically bleached completely. This is the case with -te iru in Japanese and its Korean

counterparts -ko issta and -a / -e issta, as well as with -te simau in Japanese and its Korean

counterpart -a / e pelita:5

(a) Japanese: -te iru (progressive / resultative / perfect), -te simau (completive)6

(b) Korean: -ko issta (progressive), -a / e issta (resultative), -a / e pelita (completive)

These periphrastic aspectual constructions still have formally identifiable lexical

constituents, namely the ‘be’-verbs iru and issta and the ‘put away / finish’-verbs simau

4 Constituent lexical item indicated by bold type.

5 The verbs included in the periphrastic aspectual constructions, in bold type, are given in citation form.

6 The completive aspect, a less familiar term, refers to the aspectual meaning “to do something thoroughly and

to completion” (Bybee et al. 1994:318).

Page 21: HORIE Inpress Japanese

and pelita, and have not reached the final stage of grammaticalization, unlike -ta and -ess-

above.

What is interesting from a comparative perspective is that, parallel to the contrast

in grammaticalization patterns between Japanese no / tokoro and Korean kes / tey,

Japanese aspectual constructions attest a more advanced stage of grammaticalization than

their Korean counterparts.

Japanese -te iru is known to have developed multiple aspectual meanings

including progressive (18), resultative (19a) and perfect (‘anterior’ in Bybee et al.’s (1994)

terminology) (19b), partially according with the aforementioned grammaticalization

pathway: ‘be / have’ > resultative > anterior (= perfect) > perfective / simple past.

(18) Modern Japanese

Dareka-ga nokku si-te ru yo

someone-NOM knock-PROG SFP

‘Someone is knocking on the door.’

(19) Modern Japanese

a. Titi-wa sono T syatu-o ki-te iru /

father-TOP that T shirt-ACC wear-RES

‘My father wears that T-shirt.’

b. Titi-wa nando ka sono T syatu-o ki-te iru

father-TOP several times that T shirt-ACC wear-PERF

‘My father has worn that T-shirt several times.’

Page 22: HORIE Inpress Japanese

In contrast, Korean does not have a single periphrastic aspectual construction covering the

multiple meanings expressible by -te iru. A dynamic progressive meaning is encoded by -

ko issta (20) and a stative resultative meaning is encoded by -a / -e issta (21), the latter co-

occurring only with “a limited number of intransitive verbs” (Chang 1996:124).

(20) Korean

Bongsik-i cikum talli-ko iss-ta

Bongsik-NOM now run-PROG-DECL

‘Bongsik is now running.’

(21) Korean

Bongsik-i uyca-ey anc-a iss-ta

Bongsik-NOM chair-LOC sit-RES-DECL

‘Bongsik is seated in a chair.’

However, neither of these periphrastic constructions in Korean has developed a general

perfect meaning similar to that expressed by Japanese -te iru, which suggests that -te iru

has reached a more advanced stage of grammaticalization than its Korean counterparts.

Interestingly, as extensively discussed by Wako et al. (2003), Korean -ko issta has started

to take on perfective aspectual meaning when it co-occurs with reporting verbs, as in (22),

and attainment verbs, as in (23), particularly in newspaper articles and editorials.

(22) Korean

Kath-un kwunin chwulsin-ulo amithici pwucangkwan-kwa twul-to

same-ATTR soldier background-as Armitage vice-secretary-with two-also

eps-nun sai-i-n khollin phawel kwukmwucangkwan-un

Page 23: HORIE Inpress Japanese

not exist-ATTR relation-COP-ATTR Collin Powell State Secretary -TOP

pimanglok-eyse ku-ey tayha-n chesinsang-ul ilehkey

memoire-LOC he-regarding-ATTR first impression-ACC like this

cek-ko iss-ta

note-PERF–DECL

‘In his memoir, State Secretary Collin Powell has noted his first impression of Vice-

Secretary Armitage, also an ex-soldier whose presence is unlike any other to him, like

this.’ (from Wako et al. 2003: 69, partially modified)

(23) Korean

Yecapaksa-uy yenphyengkyun cungkayul-nun namcapaksa-uy

female PhD-GEN annual average increase rate-TOP male PhD-GEN

twupay-ey talha-ko iss-ta

twice-to reach-PERF-DECL

‘The annual average increase in the number of female Ph.D. recipients has almost

doubled [lit. ‘has reached the double size of’] that of male Ph.D. recipients.’ (from

Wako et al. 2003: 69, partially modified)

A grammaticalization pathway from progressive to perfect has not been reported in

grammaticalization literature and deserves further research. For our present concern,

however, it suffices to note that such genre-sensitive perfective meaning of -ko issta is at

best in its incipient stage of grammaticalization and has not yet generalized to other

semantic classes of verbs and genres, unlike -te iru in Japanese.

Page 24: HORIE Inpress Japanese

Finally, completive aspectual constructions in Japanese and Korean, -te simau and

-a / -e pelita respectively, show a similar contrast in terms of their degree of

grammaticalization. As shown in the Table 2 below from Strauss and Sohn (1998:221,

partially modified), -te simau has reached a more advanced stage of grammaticalization

than -a / -e pelita by having evolved from a lexical verb meaning ‘finish, put away’ through

the stage of a (completive) aspectual marker into the stage of an emphatic or affective

marker. The latter stage, characterized by formal reduction (-te simau > -tyau), can further

lead to the loss of such emphatic or affective meaning. In contrast, Korean -a / -e pelita has

not advanced to these latter developmental stages.

Table 2. Grammaticalization paths for Japanese -te simau and Korean -a / -e pelita

I II III IV

Physical domain Aspectual marker Emphatic or affective marker Light or no emphasis

simau > ‘put away’ -te simau > -te simau / -tyau (semi-

productive)

> -tyau

pelita > ‘throw

away’

-a / -e pelita > -a / -e pelita (rare) > limited cases (not productive)

The two case studies of grammaticalization of periphrastic aspectual constructions in

Japanese and Korean presented in this section demonstrate that, similarly to the case of

overt nominalizers discussed in Sections 3.1 and 3.2 above, periphrastic aspectual

constructions also show a tendency for grammaticalization to proceed more extensively in

Japanese than in Korean. This suggests that the degree or rate of grammaticalization can

Page 25: HORIE Inpress Japanese

differ between languages of a similar morphological type and the differing

grammaticalization patterns are manifested across different grammatical domains.

4. Conclusion

This paper has presented an analysis of the grammaticalization patterns of sentential

nominalizers in Japanese as contrasted with those of Korean sentential nominalizers. In

spite of the morphosyntactic typological similarities between the two languages, a striking

crosslinguistic contrast emerges in that Japanese nominalizers attest a more advanced stage

of grammaticalization than their Korean counterparts. This suggests that the degree of

grammaticalization can vary even between languages of a similar morphological type.

Furthermore, this paper also relates the contrastive grammaticalization patterns observed in

the nominal domain to a similar contrast observed in the verbal domain, namely the

grammaticalization patterns of periphrastic aspectual constructions in Japanese and Korean.

It has thus shown that the contrastive patterns of grammaticalization between these two

languages are observable across different grammatical domains. Whence such a

crosslinguistic contrast in grammaticalization patterns comes is next in our agenda.

Abbreviations:

ACC Accusative

ATTR Attributive

CONC Conclusive

COND Conditional

CONJ Conjunctive

COP Copula

DAT Dative

DECL Declarative

FP Focus Particle

FUT Future

GEN Genitive

GER Gerundive

Page 26: HORIE Inpress Japanese

HON Honorific

INT Intention

INST Instrumental

LOC Locative

NEG Negative

NOM Nominative

NMLZ Nominalizer

PAST Past

PERF Perfect

POL Polite

PRES Present

PROG Progressive

PRT Particle

QT Quotative

RES Resultative

SFP Sentence Final Particle

TOP Topic

References

Abe, Kooboo. 1968. Tanin no Kao [Face of another]. Tokyo: Sintyoosya.

Bybee, Joan L., Revere Perkins and William Pagliuca. 1994. The evolution of grammar:

Tense, aspect and modality in the languages of the world. Chicago: University of

Chicago Press.

Chang, Suk-Jin. 1996. Korean. Amsterdam and Philadelphia: John Benjamins.

Harris, Alice C. and Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective.

Cambridge: Cambridge University Press.

Heine, Bernd and Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge:

Cambridge University Press.

Horie, Kaoru. 1993. “Internally headed relative clauses in Korean and Japanese: Where do

the differences come from?” In Susumu Kuno, John Whitman and Young-Se Kang

(eds.), Harvard studies in Korean linguistics 5. Seoul: Hanshin Publishers: 449-458.

Page 27: HORIE Inpress Japanese

Horie, Kaoru. 1997. “Three types of nominalization in Modern Japanese: No, koto and

zero.” Linguistics 35/5: 879-894.

Horie, Kaoru. 1998a. “Functional duality of case-marking particles in Japanese and its im-

plications for grammaticalization: A contrastive study with Korean.” In David Silva

(ed.), Japanese / Korean linguistics 8. Stanford: Center for the Study of Language

and Information (CSLI): 147-159.

Horie, Kaoru. 1998b. “On the polyfunctionality of the Japanese particle no: From the per-

spectives of ontology and grammaticalization.” In Toshio Ohori (ed.), Studies in

Japanese grammaticalization: Cognitive and discourse perspectives. Tokyo: Kurosio

Publishers: 169-192.

Horie, Kaoru. 2000. “Complementation in Japanese and Korean: A contrastive and cogni-

tive linguistic approach.” In Kaoru Horie (ed.), Complementation: Cognitive and

functional perspectives. Amsterdam and Philadelphia: John Benjamins: 11-31.

Horie, Kaoru. 2002a. “Verbal nouns in Japanese and Korean: Cognitive typological impli-

cations.” In Kuniyoshi Kataoka and Sachi Ide (eds.), Culture, interaction, language.

Tokyo: Hituzi Syoboo: 77-101.

Horie, Kaoru. 2002b. “A comparative typological account of Japanese and Korean morpho-

syntactic contrasts.” Eonehag 32: 9-32.

Horie, Kaoru. 2003. “What cognitive linguistics can reveal about complementation in non-

IE languages: Case studies from Japanese and Korean.” In Eugene H. Casad and

Gary B. Palmer (eds.), Cognitive linguistics and non-Indo European languages. Ber-

lin and New York: Mouton de Gruyter: 363-388.

Page 28: HORIE Inpress Japanese

Horie, Kaoru and Yuko Sassa. 2000. “From place to space to discourse: A contrastive lin-

guistic analysis of tokoro and tey.” In Mineharu Nakayama and Charles J. Quinn

(eds.), Japanese / Korean linguistics 9. Stanford: Center for the Study of Language

and Information (CSLI): 181-194.

Horie, Kaoru and Kaori Taira. 2002. “Where Korean and Japanese differ: Modality vs. dis-

course modality.” In Noriko Akatsuka and Susan Strauss (eds.), Japanese / Korean

linguistics 10. Stanford: Center for the Study of Language and Information (CSLI):

178-191.

Iwasaki, Shoichi. 2000. “Suppressed assertion and the functions of the final-attributive in

prose and poetry of Heian Japanese.” In Susan C. Herring, Pieter van Reenen and

Lene Schøsler (eds.), Textual parameters in older languages. Amsterdam and Phila-

delphia: John Benjamins: 237-272.

Kim, Joungmin and Kaoru Horie. In press. “Intersubjectification and textual functions of

Japanese noda and Korean kes-ita. In Takubo, Yukinori (ed.) Japanese/Korean

Linguistics 16. Stanford: Center for the Study of Language and Information (CSLI).

Kuno, Susumu. 1973. The structure of the Japanese language. Cambridge: MIT Press.

Martin, Samuel E. 1975. A reference grammar of Japanese. New Haven: Yale University

Press.

Maynard, Senko K. 1997. Japanese communication: Language and thought in context.

Honolulu: University of Hawaii Press.

Page 29: HORIE Inpress Japanese

Nishi, Yumiko. 2006. “The emergence of the complementizer no in Japanese revisited: The

dual path approach.” In Timothy Vance (ed.), Japanese / Korean linguistics 14. Stan-

ford: Center for the Study of Language and Information (CSLI): 127-137.

Noonan, Michael. 1997. “Versatile nominalizations.” In Joan L. Bybee, John Haiman and

Sandra A. Thompson (eds.), Essays on language function and language type. Am-

sterdam and Philadelphia: John Benjamins: 373-394.

Rhee, Seongha. This volume. “On the rise and fall of Korean nominalizers.”

Saunders, Ernest Dale. 1969. Face of another. New York: Weidenfeld and N.

Simpson, Andrew. 2003. “On the re-analysis of nominalizers in Chinese, Japanese and Ko-

rean.” In Yen-hui Audrey Li and Andrew Simpson (eds.), Functional structure(s),

form and interpretation. London: Curzon: 131-160.

Simpson, Andrew and Zoe Wu. 2001. “The grammaticalization of formal nouns and nomi-

nalizers in Chinese, Japanese and Korean.” In Thomas E. McAuley (ed.), Language

change in East Asia. London: Curzon: 250-283.

Strauss, Susan and Sung-Ock Sohn. 1998. “Grammaticalization, aspect and emotion: The

case of Japanese –te shimau and Korean –a/e pelita.” In David Silva (ed.), Japanese /

Korean linguistics 8. Center for the Study of Language and Information (CSLI): 217-

230.

Wako, Masakazu, Shigeru Sato and Kaoru Horie. 2003. “From progressive to perfect: A

corpus-based study of the “perfect” meaning of the Korean progressive form -ko iss.”

In William McClure (ed.), Japanese / Korean linguistics 12. Stanford: Center for the

Study of Language and Information (CSLI): 64-74.

Page 30: HORIE Inpress Japanese

Yap, Foong Ha, Stephen Matthews and Kaoru Horie. 2004. “From pronominalizer to

pragmatic marker: Implications for unidirectionality from a crosslinguistic perspec-

tive.” In Olga Fischer, Muriel Norde and Harry Perridon (eds.), Up and down the

cline: The nature of grammaticalization (Typological Studies in Language 59). Am-

sterdam and Philadelphia: John Benjamins: 137-168.

Yap, Foong Ha and Stephen Matthews. This volume. “The development of nominalizers in

East Asian languages.”