Nominal repetitive constructions in Japanese: The ‘tautology’ controversy revisited

Nominal repetitive constructions in Japanese: The ‘tautology’ controversy revisited

433 Journal of Pragmatics 20 (1993) 433-466 North-Holland Nominal repetitive constructions in Japanese The ‘tautology’ controversy revisited : Shi...

2MB Sizes 0 Downloads 60 Views

433

Journal of Pragmatics 20 (1993) 433-466 North-Holland

Nominal repetitive constructions in Japanese The ‘tautology’ controversy revisited

:

Shigeko Okamoto* Received September 1991; revised version January 1993

The usage of the Japanese nominal ‘tautological’ construction X wa X is compared with the usages of two other formally similar constructions: X ga X and X mo X. Despite their apparent similarities, the three constructions are shown to possess distinct core meanings. X wa X emphasizes the category immutability or the autonomy of an item. Both X ga X and X mo X indicate that the referent of X has some undesirable quality; the two constructions differ in their presuppositions. The meanings of all three constructions are motivated rather than arbitrary: the particular particle and the repetition of the noun phrase have a certain bearing on the meaning of each construction. Yet each core meaning is not fully analyzable by regular compositional semantics. Rather, it is to be regarded as conventionally associated with each morphosyntactic pattern, which can be adequately analyzed only by recourse to irregular phrase structures. It is further argued that given their core meanings, ‘tautologies’ may convey additional meanings which are only inferable pragmatically. Hence the meanings of ‘tautologies’ are best described in terms of both conventionality and pragmatic calculability. Accordingly, ‘radical semantics’ and ‘radical pragmatics’ are both found to be inadequate for the interpretation of ‘tautologies’.

1. Introduction The interpretation of nominal ‘tautologies’l (e.g. War is war, Business is business) illustrates the intricate relationship between semantics and pragmatics in natural language. In fact, this topic has given rise to considerable * Correspondence to: S. Okamoto, Department of Linguistics, California State University, Fresno, CA 93740, USA. The present study is an expansion of my earlier paper, ‘Nominal “tautologies” in Japanese: X wa X, X ga X, and X mo x’, published in the Proceedings of the 17th Annual Meeting of the Berkeley Linguistics Society. 1991. I would like to thank many people who provided me with valuable comments and suggestions, in particular Orin Gensler, Yo Matsumoto, Yoshiko Matsumoto, Kimiko Nishimura, Graham Thurgood, and the anonymous referees for the Journal of Pragmatics. 1 Strictly speaking, the use of the term ‘tautology’ is not appropriate, since expressions such as War is war and Business is business convey non-redundant information. But as a convenient label,

I use the term ‘tautology’ enclosed in quotation marks. 0378-2166/93/$06.00 0 1993 -

Elsevier Science Publishers B.V. All rights reserved

434

S. Okamoro / Japanese nominal repetitive

constructions

controversy in recent years (see below). However, almost all previous studies of nominal ‘tautologies’ have been based on European languages, especially English.* Furthermore, most of these studies have examined ‘tautological’ sentences in isolation, and not their uses in discourse context.3 The present paper examines the Japenese ‘tautological’ construction X wa X, where X is a noun phrase. To understand the meanings of X wa X adequately, I will investigate closely its usage in actual discourse. Additionally, to elucidate the conventionality of meaning, the construction will be compared with other formally similar constructions: X ga X and X mo X. I believe that a comparison of these three Japanese constructions will provide a new perspective on the study of the interpretation of nominal ‘tautologies’ in general. This paper is organized as follows: section 2 summarizes previous studies of nominal ‘tautologies’. Section 3 introduces the three nominal repetitive constructions to be investigated: X wa X, X ga X, and X mo A’. Sections 4-6 investigate the usages of these three constructions. Section 7 discusses the conventionality of the meanings of the three constructions. Section 8 reexamines the ‘tautology’ controversy outlined in section 2. Section 9 provides concluding remarks.

2. Previous studies on nominal ‘tautologies’:

Three approaches

From the standpoint of truth-conditional semantics, nominal ‘tautologies’ ought to lack informational content because they are redundant expressions, true by virtue of their logical form alone. Yet it is well recognized that these expressions are commonly used in ordinary discourse to convey meanings which are not readily transparent from their literal meanings. The question is, What are these meanings conveyed by ‘tautological’ expressions? What is the best way to describe them? Three main approaches to these issues have been proposed. 2.1. The Gricean radical pragmatic 4 approach In this approach, ‘tautological’ expressions are considered uninformative by themselves, but meaningful in context; their meanings are regarded as conversational implicatures which are calculable from general, language-independent conversational principles (i.e. the Gricean Cooperative Principle combined with the flouting of the maxims of Quantity). Levinson (1983: 125), for example, explains that Boys are boys is meaningless in the abstract, but could

* 3 4

An exception is Farghal (1992), which examines colloquial Jordanian Arabic tautologies. With the exception of Gibbs and McCarrell (1990) and Ward and Hirschberg (1991). The terms ‘radical pragmatics’ and ‘radical semantics’ are taken from Wierzbicka (1987).

S. Okamoto / Japanese nominal repetitive constructions

435

be used to “implicate something like ‘that’s the kind of unruly behaviour you would expect from boys”‘. For Levinson, the implicature of a ‘tautological’ utterance is particularized: it depends on the particular context of utterance. Presenting a new Gricean approach, Ward and Hirschberg (1991) argue that ‘tautological’ utterances be treated as instances of generalized, rather than particularized, conversational implicatures. They state that ‘tautological utterances of the form ‘a is a’ (e.g. Terrorism is terrorism) are used to convey the implicature that alternative utterances of the form ‘a is b’ (e.g. Terrorism is sometimes just$able) or ‘some a is b’ are not relevant. This implicature, they explain, is effected through an exploitation of the Gricean maxims of Quantity and Relevance. 2.2. The radical semantic approach

In the radical semantic approach advanced by Wierzbicka (1987, 1988), the meanings of ‘tautologies’, particularly attitudinal meanings, are assumed to be partly conventional and language-specific. Further, Wierzbicka contends that nominal ‘tautologies’ in a language may have a number of different morphosyntactic patterns, each with a specific meaning; thus the meaning of each pattern cannot be fully predicted in terms of any universal pragmatic maxims and must be spelled out in rigorous semantic representations such as the following: “(i) The form ‘Nabstract is Nabstract’ (e.g. War is war) expresses a ‘sober’ attitude toward complex human activities. (ii) The form ‘Npl-human are Npl-human’ (e.g. Boys are boys) expresses tolerance for human nature. (iii) The form ‘(art) N is (art) N’ (e.g. A rule is a rule) expresses obligation.” (Wierzbicka 1987: 105-108)

2.3. The non-radical approach

The third, non-radical, approach both semantics and pragmatics: pendent core meanings, and they tures. Regarding core meanings, scribes the conventional meaning “An English nominal (i) that the speaker (ii) that the speaker (iii) that this view is

argues that nominal ‘tautologies’ partake of nominal ‘tautologies’ possess context-indemay convey additional meanings as implicaFraser (1988: 217-218), for example, deof nominal ‘tautologies’ as follows:

tautology signals that the speaker intends that the hearer recognize: holds some view towards all objects referenced by the NP; believes that the hearer can recognize this particular view; relevant to the conversation.”

Basing her claims on an analysis of Spanish data, Escandell-Vidal asserts that a nominal ‘tautology’ has an abstract meaning:

436

S. Okamoto

“In the sentence NPI = NPl (i) NPI NPI is a qualitative features’); and (ii) this is an unquestionable

/ Japanese nominal repetitive

intensification truth.”

of NPl

(Escandell-Vidal

constructions

(to be read as ‘NPI with its prototypical 1990: 7)

Farghal (1992) classifies ‘tautological’ expressions in colloquial Jordanian Arabic (including verbal ‘tautologies’ as well as nominal) into five groups, each of which is assigned a different core meaning: (1) assessment (tolerance, admiration, or condemnation), (2) absolute generalization, (3) fatalism, (4) obligation, and (5) indifference. All three authors argue that in addition to the core meanings, ‘tautologies’ may convey implicatures that are derivable only from the context of utterance. These are the three approaches which have been taken in the recent controversy over the interpretation of nominal ‘tautologies’. In the present study, through an analysis of Japanese data, I will evaluate the three approaches, arguing for a revised version of the third, non-radical approach.

3. Nominal repetitive constructions in Japanese: X wu X, Xga X, and X mo X As is well known, Japanese has a so-called topic marker wa and a subject marker ga. For example, sentences (1) and (2) are formally identical except for the particles wa and ga: (1) Okusan wife (2) Okasun

nihon-zin wa TM 5 Japanese nihon-zin ga SM

da. COP da.

Both (1) and (2) can be translated into English as ‘(His) wife is Japanese’, although the two sentences differ pragmatically in the way the proposition is presented.6 Sentences (3) and (4) are formally the same as (1) and (2) respectively, except that in the second noun phrase the word okusan is repeated :

5 The abbreviations used in this paper are as follows: TM (topic marker), SM (subject marker). COP (copula), Q (question), SFP (sentence-final particle). 6 Sentence (1) is a so-called thematic sentence; it represents a categorical judgment. That is, okusan wa (the topic NP) is the center of attention, representing the old information; and nihonzin da provides new information about the topic. Sentence (2) represents a neutral description or a thetic judgment. That is, okusan is not the center of attention; it is part of the new information conveyed by the sentence as a whole. (Another possible interpretation of (2) is to regard okusan ga as being used for exhaustive listing.) See in particular Kuroda (1972) and Kuno (1973) for further discussion of the particles wa and ga.

S. Okamoto / Japanese nominal repetitive constructions

431

(3) Okusan wa okusan da. (4) Okusan ga okusan da . . . ’ However, unlike (1) and (2), which convey identical propositional content, the meanings of (3) and (4) are quite different from each other. And it is only (3) that can be regarded as a so-called tautological expression, corresponding to the English sentence A wife is a wife (or Wives are wives). At the level of word-by-word interpretation, sentence (4) may appear to be similar to (3), but taken as a whole (4) conveys a very different meaning: it indicates that the subject-referent has some undesirable quality. Moreover, there is another sentence that is formally identical to (3) and (4) except for the particle mo : (5) Okusan mo okusan da. The meaning of (5) is similar to that of (4); the two differ with regard to presupposition (see section 6). Thus, despite their apparent similarities, the three types of nominal repetitive expressions - X wa X (e.g. (3)), X ga X (e.g. (4)), and X mo _I’ (e.g. (5)) - are distinct from each other and must be treated as different ‘grammatical constructions’ in the sense of Fillmore et al. (1987) - as particular morphosyntactic patterns bearing specific semantic and pragmatic properties (see section 7 for further discussion). In the following three sections (46) I will investigate closely the usages of these three nominal repetitive constructionss Although the so-called tautological construction X wa X remains the main focus of this study, a comparison of the three constructions is instructive. It helps clarify a number of intricate questions regarding the semantic and pragmatic nature of X wa X and of the other two constructions: To what extent is the meaning of each construction analyzable? How is the choice of particle related to the meaning of each construction? What effect is created by the repetition of a noun phrase in each construction? To what extent is the meaning of each construction idiomatic or arbitrary? To what extent is it pragmatically inferable? These questions will be examined in sections 7 and 8.

’ Sentence (4) is left incomplete because X ga X is typically used in a dependent clause (see subsection 7.2). a X wa A’, A’ga A’, and X mo X are not the only nominal repetitive constructions in Japanese. Other constructions include: X to yuu X and X no X. X to yuu X, literally ‘X called X, means ‘all the X (e.g. ie to yuu ie means ‘all the houses’); X no A’, literally ‘Ys X, means ‘the very X’ (e.g. saigo no saigo means ‘the very end’). These constructions, unlike the three constructions in question, are phrasal and do not represent propositions, hence they will not be examined in this study.

438

S. Okamoto / Japanese nominal repetitive

constructions

4. x wa x

The functions of the X wa X construction are twofold: (a) to emphasize the category immutability of an item (subsection 4.1) and (b) to emphasize the discreteness or autonomy of an item (subssection 4.2). The X wa X construction may take various forms: X wa X alone, or X wa X followed by the declarative copula da or its variants (e.g. datta, desu, na (no da), de). The form X wa X may be used as a sentence-final form, or preposed to another clause (e.g. (6)). X wa X da (or datta, desu, na (no da), etc.)g may be used as a sentence-final form (e.g. (3), (7)), or placed before various conjunctive particles (e.g. kara ‘because’). (Although the difference is subtle, X wa X da seems more assertive than X wa Xl0 because of the presence of the declarative copula da.) X wa Xde is the continuative form of X wa X da (e.g.(7)). (Since de in X wa X de signals continuation, the absence of de after Kimi was kimi in (6) may be compared to the absence of the conjunction and in the corresponding English sentence.) (6) Kimi wa kimi, boku wa boku. TM I you TM you I ‘You are you, I am I.’ (7) Kimi wa kimi de, boku wa boku da. TM I you TM you COP I COP ‘You are you and I am I.’ In addition to the continuative usage illustrated in (7), X wa X de has another usage, which will be discussed in subsection 4.3. 4.1. X wa X for emphasizing

category immutability

The first function of X wa X is to emphasize immutability of categorization i.e., to insist that the categorization of an item identified as X cannot be changed under any circumstances, and hence must be accepted. (8) (from Croissant, a women’s magazine) koi Nan-sai demo, donna syokugyoo demo, no matter what age (one) is, no matter what kind of job (one) has, love wa

koi.

TM love ‘No matter what age one is, no matter what kind of job one has, love is love.’ 9 Which of these variants is chosen depends on tense, level of formality, and evidentiality. lo Even though the form X wa X does not contain the so-called copula da, it is interpreted as a copulative sentence like X wa A’ da. The absence of da in X wa X is parallel to that in a regular copulative sentence of the form X wa Y (e.g. Kore wa ring0 ‘This is (an) apple’).

S. Okamoto / Japanese nominal repetiiive constructions

439

In (8), for example, koi wa koi is used to insist that no matter how marginal the particular instantiation of the category may be, what is identified as koi ‘love’ cannot change its categorial attribute and must be accepted as an instance of koi. That is, by uttering X wa X, all argument over categorization is refused, and category immutability is simply insisted on with no logical explanation. As indicated by example (8) X wa X in this usage concerns any item identified as X; that is, X in X W(Irepresents any non-specific X and category immutability is claimed with regard to any X. In this respect, X wa X in this usage can be regarded as a kind of generic statement (see subsection 7.1 for further discussion). However, X wa X as a whole may be further applied to, or predicated of, a specific item, as in the following examples: (9) (from the novel Tozai-Nanboku Satsujin Jiken by Akagawa Jiroo) Heeki-bumon e itte, misairu no yuudoo-sooti o tukur-as-areru forced to make go to the arms division, and, missile guiding systems koto ni natte mo, sigoto wa sigoto to warikiru yo. (I) will accept, (thinking) that even if (it) turns out that, job TM job ‘Even if it turns out that I will be forced to make missile guiding systems in the arms division, I will accept it (thinking) that a job is a job.’ (10) [The speaker is talking about her parent.] (from a TV drama) wa oya da Hanarete ite mo, kara oya even though (he) is away, parent TM parent COP because mendoo-mi-nakutya. (I) must look after (him) ‘Even though he is away, because a parent is a parent, I must look after him.’ (11) [A woman writes about how much she cried when her husband passed away.] (from the newspaper Asuhi Shinbun) Watasi-tati wa, kessite, naka no yoi huuhu de wa arimasen desita. by no means, (an) amicable couple were not we, Sore demo, huuhu wa huuhu datta no da couple TM couple COP it is that yet, to tuukan-s-ase-rareta koto desu. (I) was made to feel strongly that it is that ‘We were by no means an amicable couple, yet a married couple was a married couple, I was made to feel this strongly.’ Examples (9)--(11) illustrate the usage of X wa X in a situation similar to (8) - a situation in which the given instantiation of the category may not necessarily be a prototypical example of X. But in these examples, the application of X wa X is twofold: As a generic statement, X wa X concerns any non-specific X, which

S. Okamoto / Japanese nominal repetitive constructions

440

is represented by X MU; then X wu X as a whole is further applied to a specific item, which is an instance of X. This specific item is something that has been evoked in the preceding discourse as a discourse topic. That is, category immutability is claimed first with regard to any X (e.g. any job in (9)) then applied to a specific item (e.g. the speaker’s job in (9)) as an instance of X. The categorization of any item identified as X cannot be changed, hence the specific X in question must be also accepted as an X, even though it is not an ideal example of X. It is to be emphasized that in (9Hll) X wa in X wa X represents any X rather than a specific X. For example, sigoto wu in (9) represents any job, not one specific job or the speaker’s job, which is the discourse topic. If sigoto wu is interpreted as the speaker’s job, sigoto wu sigoto must be synonymous with konolboku no sigoto wu sigoto, which can be interpreted only as a redundant statement ‘This/My job is a job’.ll Rather, sigoto wu sigoto is a general statement regarding any non-specific job: Any job is a job and this unis changeable categorization must be accepted. l2 Then the same immutability applied to the specific job in question, which has been mentioned in the preceding discourse. (See subsection 7.1 for further discussion of the application of X wu X to specific items.) Example (12) is similar to (8):

(12) [A

high school girl is complaining about boys who change their attitude toward her depending on her weight.] (from the newspaper Asuhi Shinbun) Gaiken wa doo de are, watusi wu wutusi nu no desu. no matter how (I) look, I TM I COP it is that ‘No matter how I look, I am I.’

Here, however, the noun phrase is a pronominal NP. Example (12) shows that the category in question in X wa X does not have to be a generic category. A pronominal NP can be used in X wu X if it suggests unique attributive properties of the referent. I1 Here, Boku no sigoto wa sigoto is parsed as [Boku no sigoto waJ [sigotoJ, which is a redundant statement. Boku no sigoto wa sigoto, however, can be a non-redundant statement, if boku-no is interpreted as a modifier of [sigoto wa sigofo] as a whole: i.e. (Boku no/ /sigofo wa sigotoJ (see subsection 7.1 for further discussion). In any event, sigoto wa in sigofo wa sigoto does not refer to a specific job. Note also that sigoro wa sigoro in (9) cannot be paraphrased as konolboku no sigoto wa konolboku no sigoto ‘this/my job is this/my job’, which would emphasize the discreteness of the speaker’s job. This meaning corresponds to the second usage of X wa X, to be discussed in subsection 4.2. Such a meaning, though possible in principle, is not appropriate in the context of (9). I* The analysis of X wa A’ in the first usage as a statement about any X is also supported by the fact that X wa X in the examples in this subsection (4.1) can be rephrased by statements such as Donna X demo X ni tigainai (e.g. Donna sigoto demo sigoro ni tigainai ‘Any job must be a job) and X to yuu MONOwa X ni rigainai (e.g. Oya to yuu mono wa oya tigainai ‘What is called/identified as a parent must be a parent.‘).

S. Okamoto / Japanese nominal repetitive constructions

441

Another situation in which X wa X is often used for emphasizing category immutability is when there is an attempt to change the category in question. (13) [A man is talking to his ex-mistress, who, contrary to his wishes, wants to resume the relationship.] (from a TV drama) moo osimai tte itta n desyo. Soo yuu kankee ga iya de, such a relationship (you) didn’t like, and now (it)% over, (you) said, right? n zya nai no. Osimai wa osimai na TM end COP isn’t (it)? end ‘You didn’t like such a relationship, so you said it’s over now, didn’t you? The end is the end, isn’t it? (14) [A mother and her daughter are talking about the mother’s love affair.] (from a TV drama) Daughter: Kawaii wa ne, okaasan. Moo honto ni kekkon sitara. (it) is charming, mother why don’t (you) get married for real? ‘It’s charming, Mother. Why don’t you get married for real?’ Mother: Ii no yo. Mamagoto wa mamagoto da kara, (it)% all right playing house TM playing house COP because kawaii no yo. (it)‘s charming ‘It’s all right. It’s charming because playing house is (just) playing house.’ (15) [A driver stopped by a police officer for speeding tries to explain that he had reasons for speeding. The police officer then says:] (from a TV drama) Donna zizyoo ga atte mo, kisoku wa kisoku desu kara. no matter what the reasons are, rule TM rule COP because ‘No matter what the reasons are, because a rule is a rule, . . .’ In these three instances, the addressee tries to change the category of the specific item in question. But the speaker insists on its immutability by using X wa X: One cannot change the category membership of any item identified as X, including the specific X in question. A slightly different situation is illustrated in example (16) : (16) [The speaker, after criticizing the condition of a house she has inspected for possible purchase, concludes:] (attested in conversation) yasui uti wa yasui uti ne. Yappari, as expected, cheap house TM cheap house isn’t it? ‘As one would expect, a cheap house is a cheap house, isn’t it?

S. Okamoto / Japanese nominal repetitive constructions

442

In (16) there is no reference to the inappropriateness of the ascribed category. Nor is there any explicit attempt to change the category. But one gathers that before seeing the house, the speaker had hoped that the house, though inexpensive, might be in good condition. Her use of X wa X emphasizes that like for any other cheap house, it is, after all, impossible to change the attribute of this particular house as a cheap house. In sum, examples (8)-(16) all demonstrate the use of X WCIX to emphasize category immutability. That is, in situations in which the categorization is (explicitly or implicitly) called into question, X wa X is employed to insist that it is absolutely immutable under all conditions, whether the categorization is reasonable or not. In other words, the use of X wa X can be said to involve a kind of speech act - the speech act of insistence or emphasis. Accordingly, when there is no need for such a speech act, the use of X wa X becomes awkward. For example, in response to A’s utterance in (17), a straightforward category ascription such as Bl is sufficient, while B2 is nonsensical: (17) A:

Ano tatemono, nani? that building, what (is) ‘What is that building?’ Bl : Aa, are wa bizyutukan da. oh that TM (a) museum COP ‘Oh, that’s a museum.’ B2: #Aa, bizyutukan wa bizyutukan TM museum oh museum ‘Oh, a museum is a museum.’

Furthermore, X wa X cannot as shown in (18) :

normally

da.13 COP

be employed

in a yes-no

question,

desu ka. wu oya (18) # Hanarete ite mo, oya Even though (he) is away parent TM parent COP Q ‘Even though he is away, is a parent a parent?’ (18) is unacceptable as a yes-no question (with a rising intonation on the particle ka) because one cannot insist on something as absolutely true while at the same time inquiring about its validity, and because it is strange to question one’s own current speech act (e.g., ‘Am I insisting that . ..?‘). Note, however, that (18) is acceptable if used as kind of a tag-question (with a falling intonation on the particle ka). In this case, the particle ka is used to mitigate the force of the speech act or the speaker’s insistence (see also example (13)). This usage is thus congruous with the function of X wa X. I3 Whereas an asterisk (*) traditionally is used to indicate ungrammaticality, for semantically and/or pragmatically unacceptable sentences.

I use the symbol

#

S. Okamoto / Japanese nominal repetitive constructions

443

4.2. X wa X for emphasizing autonomy The second function of X wa X is to emphasize the discreteness or autonomy of an item. In this usage, there are two or more contrastive items in the discourse world and each is marked as distinct from the other(s). This is illustrated in (19) and (20): (19) (from a TV drama) Daughter-in-law: Konya, uti de, ikaga desu ka, tonight, at our place, how is (it)? okaasan no taiin-iwai. celebration of Mother’s return from the hospital ‘Tonight at our house, how about celebrating your return from the hospital?’ Tui kono aida, Mother-in-law: Sonna, ii wa yo. such not necessary just the other day, site moratta bakkari tanzyoo-iwai (a) birthday-celebration (I) just received (from you), zya nai no. isn’t (it the case)? ‘That’s not necessary. You just gave me a birthday party the other day, didn’t you? wa kondo desu. Daughter-in-law: Iie, are wa are, kondo no, that TM that this timeTM this timeCOP ‘No, that is that, this time is this time.’ (20) (from the newspaper Asahi Shinbun) Kimi wa kimi, ware wa ware nari, saredo nakayosi. COP yet, (we) are good friends you TM you I TM I ‘You are you, I am I, yet we are good friends.’ In both (19) and (20), the speaker uses a pair of contrastive clauses of the form A’ wa X to draw the distinction. Sometimes, however, only one clause is used and the contrast is implicit, as illustrated in (21): (21) (from Syuukan Asahi, a weekly magazine) Wagaya de wa, kakubetu no kimari-goto to itta mono wa such things as in my home, particular rules nai no desu ga, tada, kazoku minna o aisi, do not exist, but, just, love every family member, and, sontyoo-si-au koto sorezore no tyoosyo 0 each person’s strengths to respect each other gurai desyoo ka . . . Dakara, musume wa musume, perhaps, (that)‘s about it therefore, daughter TM daughter

444

S. Okamoto 1 Japanese nominal repetitive

constructions

kanozyo-nari no kangae ya konomi ga atte toozen desu. her own (she) has (it) is natural thoughts and taste ‘In my house, no particular (formal) rules exist, but, just, the family members (are expected) to love each other and respect each other’s strengths; perhaps, that’s about it. Therefore, my daughter is my daughter, it is natural that she has her own thoughts and taste.’ Here, the speaker emphasizes the autonomy of his daughter as a consequence of the mutual respect among his family members. It is also implicit that not only the daughter but each of the other members of the family is autonomous. By way of comparison, (22) uses the same expression musume wa musume to illustrate the first function of X wa X - to emphasize category immutability:

(22) Hanarete

ite mo, even though (she) is nani ka to one way or another ‘Even though she is one way or another.’

yappari musume wa musume, away after all daughter TM daughter ki ni kakaru. (I) worry about (her) away, a daughter is a daughter, I worry

about

her

The expression hanarete ite mo in (22) suggests that the speaker, though he does not consider his daughter an ideal daughter, does concede that she is still his daughter. Thus, in (22), it is the categorization of the daughter which is at issue, not her autonomy.14 As seen in (19) and (20), pronominal NPs often appear in the second usage of X wa X. (23) is a similar example in common use: (23) Sore wa sore, kore wa kore da. that TM that this TM this COP ‘That is that, this is this.’ The frequency of pronominal NPs in the second usage of X wa X seems due to the fact that what is in focus here is the discreteness of the item in question rather than its categorial attributes. Hence a generic name, a conveyor of categorial attributes, does not have to be used. The difference between the two functions of X wa X, however, is not always clear-cut.

I4 Note that in (22) musume wa musume concerns both non-specific daughters and the speaker’s own daughter. In (21) on the other hand, musume wa mwume concerns only the speaker’s daughter because in (21) (unlike (22)) it is not necessary to reaffirm the categorization of the speaker’s daughter based on the category immutability of any daughter. Rather, her identity is given; only her autonomy is emphasized.

S. Okamoto

/ Japanese nominal

repetitive

constructions

445

(24) [A career woman expresses her feelings about turning 40.1 (from Croissant, a women’s magazine) yoo ni natta. Iiwake de naku, watasi wa watasi to i-eru able to say that (I) became not as an excuse, I TM I ‘Not as an excuse, I became able to say “I am I”.’ In (24), watasi wa watasi ‘I am I’ may be interpreted as an insistence on the immutable identity of the speaker as well as an assertion of her autonomy. (Similarly, example (14) allows both interpretations.) As example (24) suggests, the two functions of X wa X are closely related. It may well be the case that the second function of X wa X is based on, or presupposes, the first function, and that it becomes particularly relevant when there exist contrastive categories in the discourse world. Here, immutability of the categories of two (or more) items is interpreted as discreteness of the items. wa otoko, onna wa onna, rikai-si-a-eru koto wa nai. man TM man woman TM woman can never understand each other ‘Men are men, women are women, they can never understand each other.’

(25) Otoko

In (25), for example, men and women are presented as absolutely discrete from each other, which in turn must depend on the impossibility of changing the category membership of men and women. 4.3. X wa X de

As mentioned earlier, the form X wa X de is used as a non-final form of X wa X da. (26) Hanarete ite mo,

(7)

yappari Musume wa musume de, even though (she) is away, after all daughter TM daughter COP nani ka to ki ni kakaru. one way or another (I) worry about (her) ‘Even though she is away, after all, a daughter is a daughter, and I worry about her one way or another’ Kimi wa kimi de boku wa boku da. you TM you COP I TM I COP ‘You are you and I am I.’

Example (26) illustrates the first function of X wa X in the form of X wa X de; example (7) (reproduced above) illustrates the second function. In addition to these two usages, however, X wa X de has a third usage, which is restricted to

446

S. Okamoto / Japanese nominal repetitive consrrucrions

this form alone. Here X wa X de does not constitute a clause representing a proposition. Rather, it is placed within a clause and functions as a phrase; as such, it simultaneously sets up a frame for the predication that follows it and emphasizes the discreteness or autonomy of this frame. (27) Arai-owat-tara, when (you) finish washing (them), koko ni oite tyoodai. here please put (them) ‘When you finish washing them, separately (from other things) .’

yasai vegetable

wa yasai TM vegetable

the vegetables,

please

de COP

put them

(28) [A nurse is talking to a patient about the patient’s mistress.] (from the novel Fushin no Toki by Ariyoshi Sawako) Okusan ni wa sir-ase-nai de, uti wa uti de without telling (one’s) wife, home TM home COP daizi ni sit-oku no ga hontoo desu nee, keeping (it) carefully is the right (way), isn’t (it)? ‘(One should) carefully keep his home intact without telling (about his affair) - that’s the right way, isn’t it?’

here

his wife

In (27), for example, yusai wa yasai de does not represent the independent proposition ‘Vegetables are vegetables’. Rather, it is part of a clause and specifies the frame (i.e. yasai) for the following predication (i.e. koko ni oite tyoodai), while emphasizing the discreteness of this frame. Example (28) illustrates the same principle. A similar example is given in (29), this time with a pronominal noun phrase: (29) (from a TV drama) yoi tokoro ga aru no desu. Aliata wa anata de COP have strenghts it is that TM you You ‘You have your own strengths.’ The same phrasal use of X wa X de often occurs when two contrastive frames produce the same kind of consequence, as shown in (30) and (31): (30) [The speaker is talking about heavy traffic on the Bay Bridge.] (attested in conversation) moo itu demo mono-sugoku konde ru si, uiiku-endo Heezitu wa (on) weekdays indeed always (it) is terribly crowded, and, weekend mata sugoi no. wa uiiku-endo de TM weekend COP (it) is again formidable ‘On weekdays, it (the bridge) is indeed always terribly crowded, and on weekends, for other reasons it is again terrible.’

S. Okamoto / Japanese nominal repetitive constructions

441

(31) [A nurse is talking to a patient.] (from the novel F&in no Toki by Ariyoshi Sawako) desu yo. nan te tatte, sue ga tanosimi Otoko no ko wa, can look forward to (their) future (in the case of) boys, surely, Tanomosii mono desu yo. Demo, musume wa musume de girls TM girls COP but, (they) are reliable Ryoohoo aru no ga itiban desu yo. kawaii mono desu nee. is the best (they) are cute, aren’t (they)? to have both ‘In the case of boys, you can surely look forward to their future and rely on them. But, girls are also cute fir other reasons, aren’t they? To have both is the best.’ In (30) and (31) X wa X de sets up a frame (e.g. uiikuendo ‘weekends’ in (30)) which stands in contrast to another frame (e.g. heezitu ‘weekdays’ in (30)). But both independently bring about the same kind of consequence (e.g. terrible crowdedness in (30)). If the two contrastive frames result in contrastive consequences, however, the use of X wa X de is inappropriate. In such a case, X wa by itself is appropriate in its contrastive use, as in (32): moo itu demo mono-sugoku konde ru kedo, (32) Heezitu wa (on) weekdays indeed always (it) is terribly crowded, but, wariai suite iru. uiiku-endo wa ( # uiiku-endo de) weekend TM weekend COP (it) is relatively empty ‘On weekdays, it (the bridge) is indeed always terribly crowded, but on weekends, (for other reasons) it is not so crowded.’ Note that in the clausal usage of X wa X de (e.g. (7)) the connective form de is deletable, as shown in (6) (reproduced below): (6)

Kimi wa kimi, boku wa boku. you TM you I TM I ‘You are you, I am I.’

This deletion of de, however, is not possible with the phrasal usage of X wa X de.

(33) # Arai-owat-tara, yasai wa yasai when (you) finish washing (them), vegetable TM vegetable koko ni oite tyoodai. here please put (them) ‘When you finish washing them, vegetables are vegetables. Please put them here.’

448

S. Okamoro / Japanese nominal repetitive constructiom

In (33), for example, the deletion of de renders the remainder a clause representing a proposition (i.e. Vegetables are vegetables). Such a proposition is inappropriate in the context of (33), however. Thus, despite the resemblance, the two usages in question are distinct from each other.

5. xgax

As mentioned earlier, despite the apparent similarities, the function of X ga X is quite different from that of X wa X. The construction X ga X is used to indicate that the referent of X has some undesirable quality.

(34) [A. man and his wife are talking about their son’s poor academic record.] (from a TV drama) Wife: Anata, seeseki doo datta. how were (your) grades? YOU, ‘How were your grades?’ Husband: Un, maa maa. Kimi wa. well, so-so you? ‘Well, they were so-so. How about you?’ Wife : Maa maa. Oya ga oya da nee. kara so-so parent SM parent COP because right? ‘So-so. Because the parents are (not very smart) parents, right?’ (35) Wife: Maa maa. Demo, Oya wa oya da kara nee. parent TM parent COP because right? so-so but, ‘So-so. But, because parents are parents, . . . right?’

In (34), the couple move on from their son’s bad academic record and proceed to talk about their own grades. The wife says Oya ga oya da km-a nee, which means that the parents are not very desirable or smart. This comment could easily imply that it is understandable that the son is not very smart, either. If, instead, the wife had said Oya wa oya da kara nee, as in (35) it would mean that the problem of the parents should be considered separately from the son’s problem, that it is not impossible for the son to do well. The undesirability of the weather is at issue in the following constructed example : (36) A: Nee, kyoo pikunikku iku? Hey, today shall (we) go on a picnic? ‘Hey, shall we go on a picnic today?’

S. Okamoto 1 Japanese nominal repetitive constructions

449

kara B: Soo nee, Otenki ga otenki da well, weather SM weather COP because pikunikku wa yameyoo yo. let’s cancel (the) picnic ‘Well, because the weather is (not good) weather, let’s cancel the picnic.’ pikunikku ni ikoo yo. (36a) B: #Otenki ga otenki da kara weather SM weather COP because let’s go on (a) picnic ‘Because the weather is (not good) weather, let’s go on a picnic.’ kedo, pikunikku ni ikoo yo. (36b) B: Otenki ga otenki da weather SM weather COP but, let’s go on (a) picnic.’ ‘The weather is (not good) weather, but let’s go on a picnic.’ pikunikku wa yameyoo yo. kara, (36~) B: #Otenki wa otenki da weather TM weather COP because let’s cancel (the) picnic ‘Because weather is weather, let’s cancel the picnic.’ B’s utterance in (36) is an appropriate

answer to A’s question: The expression

Otenki ga otenki, which indicates that the weather is not so good, is presented

as a reason for cancelling the picnic, Sentence (36a), on the other hand, is infelicitous because the bad weather is presented as a reason for going on a picnic. Sentence (36b) contains the same two propositions as (36a), but now the response is appropriate, since the two propositions are presented as contradictory by the use of the conjunction kedo ‘but’. Sentence (36c), in which X wa X is used, does not make sense. Similarly, in the following three examples (37)-(39), X ga X is again used to convey the undesirable nature of the item in question: (37) [A man is talking about how badly he is treated by his wife because of his affairs.] (from the novel Fushin no Toki by Ariyoshi Sawako) Iya, watasi nado wa, motto hidoi mono desu. Uti ni kaer-eba, no, in my case, (it)‘s much worse when (I) go home, gesyuku-nin desu yo. Nanisiro, zisseki ga zisseki de, (1)‘m a boarder anyhow, record SM record COP ‘No, in my case, it’s much worse. When I go home, I’m treated like a boarder. Anyhow, my record (of affairs) is (a bad) record, so . ..’ (38) [A and B are police officers; B is A’s boss. B is talking about a murder case.] (from a TV drama) A: Nani ka. what? ‘What (did you want to say)?’

450

S. Okamoro / Japanese nominal repetitive constructiom

B: Un, ree no ken, sono-go, doo natta ka to omotte. yes, that case, since then, what became of (I) am wondering Nanisiro ziken ga ziken dake ni, syakai no kansin mo at any rate case SM case since, society’s attention, also atumete iru koto da si (it) is drawing, so ‘Yes, I’m wondering what has become of that case since then. At any rate, since the case is (an unusually bad) case, it is also drawing the attention of the public, so . ..’ (39) [The narrator worries about his 70-year-old friend who lives in northern Japan.] (from the novel Uma no Hanashi by Mizukami Tsutomu) samui toti de no Tosi ga tosi da kara, age SM age COP because, in (that) cold place nangi daroo. huyu-gosi wa, getting through (a) winter, is probably difficult ‘Because his age is (not a desirable) age, it is probably difficult to get through a winter in that cold place.’ Note that in the context of (39) the age in question is old age. However, tosi ga tosi could also be used to suggest a young age that would be considered undesirable for some different activity (e.g. a teenager raising a child). In the above examples, X ga X has been used to suggest some negative quality of the referent. But it may also be used when the item in question has an unusally high quality. Contrast (40) and (41): (40) [The speaker is talking about the very formal wedding to which she is invited.] tyan to site ik-anakutya. Basyo ga basyo da kara, place SM place COP because, in a proper (dress) (I) must go ‘Because the place is (not an ordinary) place, I must go in a proper dress.’ (41) [The speaker is talking about a place that is considered dangerous.] ki o tuke-nasai. Basyo ga basyo da kara, place SM place COP because, be careful ‘Because the place is (not a safe) place, be careful.’ In the context of (40) basyo ga is not ordinary, but very fancy. same expression basyo ga basyo (42) also illustrates the use of referent of X:

basyo indicates that the place of the wedding In the context of (41) on the other hand, the suggests that the place is dangerous. Example X ga X to indicate the high quality of the

S. Okamoto 1 Japanese nominal repetitive constructions

451

(42) [The speaker is talking about the tennis match in which he is facing a

strong opponent.] da kara, Nanisiro aite ga aite at any rate, opponent SM opponent COP because, katu zisin wa nai yo. self-confidence in winning (I) don’t have ‘At any rate, because the opponent is a (very strong) opponent, I’m not confident if I can beat him.’ The contexts of (40) and (42) suggest that when X ga X concerns high quality, the quality is considered almost too high or too threatening for the person(s) involved. Accordingly, we can conclude that whether the referent’s quality is negative or positive in the abstract, by using X ga X the speaker presents the quality as undesirable or threatening for those who are involved - as something which it makes the speaker uneasy to describe explicitly. The noun phrase is thus simply repeated to avoid explicit mention of the undesirable quality.15

6. XmoX The function of X mo X is similar, but not identical, to that of X ga X: X mo X is also used to indicate that the item in question has some undesirable quality. But X mo X, unlike X ga X, presupposes the existence of another item that is equally undesirable. (43) [An elderly man criticizes the casual way young married men and single women have affairs.] (from the novel Fushin no Toki by Ariyoshi Sawako) Otoko mo otoko nara, onna mo onna da, ittai men also men if women also women COP how in the world I5 Though very restricted, there is another usage of the form X ga X, as illustrated below: (i) (from Croissant, a women’s magazine) Huto mi-mawas-eba, yo-nin ga yo-nin, isu no ue ni when (I) look around suddenly, four people SM four people on a chair hiza tatete terebi o mite-tari suru. raising (their) knees do things like watching T.V. ‘When I look around suddenly, (I find) things, such as all four of them sitting on a chair with their knees up and watching T.V.’ (ii) Mina ga mina dame da to yuu wake de wa nai. all SM all bad COP it is not that ‘It is not that each and every one of them is bad.’ In this usage, X ga X as a whole functions as a subject noun phrase and X is restricted to a quantified noun phrase. The repetition of X emphasizes that the predicate applies, without exception, to every member of the group referred to by X.

452

S. Okamoto / Japanese nominal repetitive constructions

kanozyo-tati wa yome ni itte kara wa, donna katee o what kind of family after they get married, tukuru no daroo. (they) will make (I) wonder ‘(lit.) If men are also men, then women are also women. (I find both the men and the women undesirable.) I wonder what kind of families they will make when they get married.’ kodomo mo kodomo da. oya da kara (44) Oya ga parent SM parent COP because child also child COP ‘Because the parents are (bad) parents, the child is also (a bad) child.’ (45) Anata mo anata yo. SFP YOU aslo you ‘(lit.) You are also you. (You, too, are bad/undesirable.)’ In (43), both the men and the women italicized expression indicates that the able. Example (49, a set phrase often poses that besides the addressee, there

are considered undesirable. In (44) the child, like the parents, is not commendused to criticize the addressee, presupis someone else to be criticized.16

7. Conventionality of Japanese nominal repetitive constructions By investigating the usages of the three nominal repetitive constructions X wa X, X ga X, and X mo X, I have demonstrated that despite their apparent similarities, the meanings (or functions) of the three constructions are distinct

I6 Though not directly pertinent to the current discussion, the form X MO X has another usage. Here, X MO X as a whole functions as a noun phrase and is used for intensifying the quality associated with X. (i) (from the short story Hana yori Kekkon Kibi-dango by Hayashi Mariko) “ . tonikaku, kakko ii kuruma o te ni ireru no da” to honki de omotte simau no da. (I) started thinking seriously that at any rate, nice-looking car (I) will get Ori yuuzin no hitori ga makka na arufa romeo o MO ori, occasion also occasion one of (my) friends bright red Alfa Romeo to moti-kakete kita. kawanai ka won’t (you) buy? came to suggest that ‘ “At any rate, I will get a nice-looking car,” I started thinking seriously. That very moment, one of my friends came to suggest, “Won’t you buy a bright red Alfa Romeo.“’ (ii) A: Kanozyo bizin na n datte ne. she is (a) beauty I hear ‘She is a beauty, I hear.’ B: Un, bizin mo bizin, at-tara, bikkuri-suru yo. yes beauty also beauty if (you) meet (her), will be surprised ‘Yes, she is a (real) beauty, if you meet her, you will bc surprised.’ In (i), ori mo ori stresses the aptness of the timing; in (ii), bizin mo bizin indicates that the person in question is extremely beautiful.

S. Okamoto / Japanese nominal repetitive constructions

453

from each other. These meanings cannot be regarded exclusively as a matter of pragmatic inference - as something entirely contingent upon the context of utterance. Rather, as will be shown below, they are to be considered the conventionalized core meanings of the three constructions. However, characterizing the core meanings of these constructions as conventional does not mean that they are totally arbitrary or idiomatic. Nor does it mean that they are entirely compositional - analyzable based on conventional semantic interpretation rules. In what follows, then, I will discuss the nature of the conventionality of meanings associated with X wa X (subsection 7.1) and X ga X and X mo X (subsection 7.2). 7.1. X wa X With regard to X wa X, its core meanings - i.e., the speaker’s emphasis on category immutability or on the autonomy of an item - are not totally arbitrary because the components of the construction, namely, the particle wa and the repetition of the noun phrase, do have a certain bearing on the meaning. First of all, the particle wa in general is used in making a categorical judgment rather than a thetic judgment (for which the particle gu is used) (Kuroda, 1972): In X wu Y, X represents the topic or the center of attention, and Y represents the comment (or new information) about the topic. Generic statements (e.g. definitional sentences) represent a particular kind of categorical judgment, and the particle wu is normally used in such statements (Kuroda, 1972). For example, English sentences such as Dogs are animals and Dogs run, as generic statements, are translated into Japanese with the particle wu, not gu, as in (46): (46) Inu wa/*ga doobutu da. dog TM/SM animal COP ‘Dogs are animals.’ And ‘tautological’ expressions such as (47) can be naturally taken as analogous to generic statements like (46): (47) Inu wa inu da. dog TM dog COP ‘Dogs are dogs./A dog is a dog.’ That is, X wu X can properly be considered a statement about 2’. Even if it conveys no attributes of X, it does tell something about X: it highlights the special and particular aspect of X by pointing out its category immutability or autonomy. In this way, the use of wu in X wu X can be linked to its use in marking the topic in a generic statement or a categorical judgment.

454

S. Okamoto

/ Japanese nominal repetitive

It is also the case that the topic NP marked focus of contrast (e.g. (48)): (48) Hanako

wa gakusee TM student ‘Hanako is a student,

constructions

by wa may in general

da ga Taroo wa gakusee COP but TM student but Taroo is not a student.’

serve as a

de wa nai. COP not

This contrastive nature of wa is evident in one of the meanings of X wa X mentioned above, whereby the discreteness of an item is emphasized. The repetition of X is also germane to the meanings of X wa X. Conventional compositional semantics would treat X wa X as a regular topiccomment sentence; the repeated X would then be only a redundant comment. However, in fact the repetition is not redundant, but meaningful. The repetition of exactly the same NP in X wn X, I argue, indicates iconically the speaker’s insistence on sameness or immutability. In particular, with the particle wa, this iconic repetition signifies the speaker’s insistence on the category immutability or the autonomy of an item. Thus, the meanings of X wa X are not arbitrary, but motivated. On the other hand, they are not totally motivated, for they are not entirely analyzable by regular compositional semantics. Rather, they are best regarded as associated with the morphosyntactic pattern NPZ wa NPI as a whole. If any part of this pattern is altered, the meaning changes. As we saw with regard to X ga X and X mo X, the morphological difference produces entirely different meanings, even though the identical NP repetition is used. It is, moreover, essential to repeat exactly the same NP in X wa X, a fact which cannot be accounted for with regular compositional semantics. For example, compare (15) (reproduced below) with (49)-(51) : (15) [A driver stopped by a police officer for speeding tries to explain that he had reasons for speeding. The police officer then says:] (from a TV drama) Donna zizyoo ga atte mo, kisoku wa kisoku desu kara. no matter what the reasons are, rule TM rule COP because ‘No matter what the reasons are, because a rule is a rule, . ..’ (49) Donna zizyoo ga atte mo, kisoku desu kara. no matter what the reasons are, rule COP because ‘(lit.) No matter what the reasons are, because (it) is a rule, . ..’ (50) Donna zizyoo ga atte mo, kore wa kisoku desu kara. no matter what the reasons are, this TM rule COP because ‘No matter what the reasons are, because this is a rule, . ..’ (5 1) #Donna zizyoo ga atte mo, kono kisoku wa kisoku desu kara. TM rule COP since no matter what the reasons are, this rule ‘No matter what the reasons are, because this rule is a rule, . . .’

S. Okamoto / Japanese nominal repetitive constructions

455

The topic NP kisoku wa is replaced by an empty pronominal NP in (49) and by the pronominal NP kore wu in (50). Both (49) and (50) simply assert that the thing under discussion is a rule. In (15) by contrast, the speaker insists that the thing under discussion is a rule by dejinition, and not something he can alter at his discretion. In (51) kisoku wu is modified by the prenominal demonstrative kono ‘this’. (51) is excluded because Kono kisoku wa kisoku can be interpreted only as a redundant expression ‘This rule is a rule’. Kisoku wa in (15) represents any non-specific rule, not the specific rule under discussion, hence it cannot be replaced by kono kisoku wa, as in (51). Since X wu X emphasizes the immutability of the category, it might be thought that kisoku wu kisoku and oyu wu oyu, for example, should be paraphrasable, as in (52) and (53) respectively: kisoku wu yuuzuu no kiku (52) Donna zizyoo ga atte mo, no matter what the reasons are, rule TM flexible mono zyu urimusen kara. thing is not because ‘No matter what the reasons are, because a rule is not a flexible thing, ...’ kara (53) Hanarete ite mo, ZYU nui OYU wu tunin even though (he) is away, parent TM non-kin COP NEG because mendoo-mi-nakutya. (I) must look after (him) ‘Even though he is away, because a parent is not non-kin, I must look after him.’ However, the italicized clauses in (52) and (53) are more explicitly explanatory and argumentative than the corresponding ‘tautological’ expressions: they do provide logical explanations. Compared to these sentences, kisoku wu kisoku and oyu wu oyu are blunt assertions: With no logical explanation offered, the speaker insists that the categorization cannot be changed by any means, thus must be accepted. Hence, none of the italicized clauses in (49~(53) which are regular topic-comment sentences, are synonymous with the corresponding ‘tautological’ sentences. These comparisons show the importance of repeating exactly the same NP in the X wu X construction. Furthermore, the sequence of the two NPs in X wu X cannot be interrupted. Consider first what happens with a normal X wu Y construction, as in (54): (54) Nan-sai demo, koi donna syokugyoo demo, no matter what age (one) is, no matter what kind of job (one) has love wu tunosii mono da. TM enjoyable thing COP ‘(lit.) No matter what age (one) is, no matter what kind of job (one) has, love is an enjoyable thing.’

456

S. Okamoto / Japanese nominal repetitive

constructions

Here the italicized sentence has a regular topiccomment structure. Accordingly, the topic koi wa can be moved to the beginning of the sentence with no change in meaning, as shown in (55): (55) Koi wa, nan-sai demo, love TM no matter what age (one) is, donna syokugyoo demo, tanosii mono da. no matter what kind of job (one) has enjoyable thing COP ‘(lit.) Love, no matter what age (one) is, no matter what kind of job (one) has, is an enjoyable thing.’ This is not so in the X wa X construction, however. In the case of (8) (reproduced below), koi wa cannot be moved to sentence-initial position without changing the meaning: (8) (from Croissant, a women’s magazine) Nan-sai demo, donna syokugyoo demo, no matter what age (one) is, no matter what kind of job (one) has koi wa koi. love TM love ‘(lit.) No matter what age (one) is, no matter what kind of job (one) has, love is love’. (56) #Koi wa, nan-sai demo, 1oveTM no matter what age (one) is, donna syokugyoo demo, koi. no matter what kind of job (one) has love ‘(lit.) Love, no matter what age (one) is, no matter what kind of job (one) has, is love.’ (56) is awkward because the second koi is construed as an independent predicate phrase which expresses a redundant comment about the topic koi wa. That is, the nuance of insistence on category immutability is lost in (56). In the same way, the italicized sentence in (57) cannot be considered a paraphrase of oya wa oya da: wa hanarete ite mo, oya da kara (57) #Oya parent TM even though (he) is away, parent COP because mendoo-mi-nakutya. (I) must look after (him) ‘(lit.) A parent, even though (he) is away, because (he) is a parent, must look after (him).’

(I)

S. Okamoto / Japanese nominal repetitive constructions

451

From the foregoing discussion, it is evident that the pattern NPZ wa NPl is used as a whole to convey specific meanings, and that it cannot be analyzed as a regular topic-comment sentence (i.e. /X wu/ [X (&)A like the italicized clauses in (49)-(54); that is, category immutability or the item’s autonomy can be expressed only by the construction as a whole. Yet at the same time, as discussed earlier, it is also the case that X wa X can be related to a topiccomment sentence pattern, representing a categorical judgment about X. To capture this ambiguous situation, I propose a dual analysis for X wu, though it is an unconventional analysis: In the interpretation of X wa X (da), X wa is taken as the topic (JX wan, and simultaneously reanalyzed as part of the predicate (IlX wa X] (da)J). That is, [X WCIXl (da)] is construed as being predicated of [X wu/, this predicate emphasizing the categorical immutability of an item identified as X or the discreteness of the referent of X. For example, in (8), koi wa koi as a whole can be considered a comment about the topic koi wa. Or in Kore wa kore ‘This is this’, kore wa kore as a whole is the predicate of the topic kore wa. From the viewpoint of syntactic structure, the analysis of [X wa X/ as a predicate of [X wal is irregular. But I argue that this analysis appropriately captures the nature of the construction, whose meanings exhibit partial compositionality (i.e., /X wa/ as topic) and non-compositionality (i.e., [X wa Xl as predicate, expressing category immutability or item’s autonomy). The following observation provides further support for this analysis of the structure of the X wa X construction. As we saw in section 4 (e.g. (9x1 l)), X wa X may involve not only a non-specific item, but also a specific item in the discourse world. (10) [The speaker is talking about her (from a TV drama) Hanarete ite mo, oya even though (he) is away, parent mendoo-mi-nakutya. (I) must look after (him) ‘Even though he is away, because him.’

parent.] wa oya da kara TM parent COP because

a parent is a parent, I must look after

In (10) (reproduced above), for example, oya wa oya da is used to emphasize the category immutability of any parent; here ‘any parent’ is represented by the phrase [oya wu] as the topic, and [[oya wa oyu] (da)] is predicated of this topic. The same category immutability is then applied to the speaker’s parent as an instance of oya. In this application, [[oya wa oyu/ (da)] can be regarded as being predicated of the speaker’s parent, which is the (implicit) topic of the whole sentence. In fact, this topic can be mentioned explicitly, as shown in (58):

458

S. Okamoto / Japanese nominal reperilive construclions

(58) Hanarete ite mo,

kare, oya wa oya da kara even though (he) is away, he parent TM parent COP because mendoo-mi-nakutya. (I) must look after (him) ‘(lit.) Even though (he) is away, because he, (a) parent is (a) parent (i.e., because he cannot be denied as my parent), (I) must look after (him).’

The italicized sentence in (58) can be parsed as follows: [kare] [[oya wa oyaJ da]; that is, kare is the topic NP17 and oya wa oya da is predicated of kare, in which oya wa oya as a whole functions like a predicate nominal. A parallel example is given in (59), which corresponds to example (11) (reproduced below) : de wa arimasen desita. wa, kessite, naka no yoi huuhu were not by no means, (an) amicable couple we, Sore demo, watasitati, huuhu wa huuhu datta no da couple TM couple COP it is that yet, we, to tuukan-s-ase-rareta koto desu. (I) was made to feel strongly that it is that ‘(lit.) We were by no means (an) amicable couple, yet, we, (a) married couple was (a) married couple (i.e., it cannot be denied that we were in fact a married couple), (I) was made to feel this strongly.’ (11) [A woman writes about how much she cried when her husband passed away.] (from the newspaper Asahi Shinbun) de wa arimasen desita. Watasi-tati wa, kessite, naka no yoi huuhu by no means, (an) amicable couple were not we, Sore demo, huuhu wa huuhu datta no da couple TM couple COP it is that yet, koto desu. to tuukan-s-ase-rareta (I) was made to feel strongly that it is that ‘We were by no means an amicable couple, yet a married couple was a married couple, I was made to feel this strongly.’ (59) Watasi-tati

Note that in (59) and (11) X wa X is in the past tense - datta - not in the unmarked, ‘timeless’ present tense. This temporal specificity provides an additional argument that [[X wa XJ datta] in fact applies to a specific item (i.e. watasi-tati), whose category immutability is presented as a fact in the past. It is also significant that X wa X as a whole may be modified by a possessive NP, as in (60): I7 Kare in (58) can be. regarded as topic, even though topic is common particularly in speech.

it is not marked

by wa. Deletion

of wa for

S. Okamoto

1Japanese nominal repetitive constructions

(60) Hanarete ite mo, kare, boku no oya wa even though (he) is away, he, my parent TM mendoo-mi-nakutya. (I) must look after (him). ‘(lit.) Even though (he) is away, because he, my, (i.e., he cannot be denied as my parent), (I) must

459

da kara oya parent COP because

(a) parent is (a) parent look after (him).’

Here, the phrase boku no ‘my’ modifies oya wa oya as a whole: [[[boku no/ [oya wa oya]] da]. The phrase [[[boku no/ [oya wa oya]J da] is in turn predicated of the topic Fare]: Fare] [[[boku no] [oya wa oya]] da]. In this usage, a regular structural analysis is not applicable; that is, boku no cannot be construed as a modifier of only the first oya (i.e. [[boku no oya wa] [oya da]‘. In such an analysis, [boku no oya wa] would be interpreted as referring to a specific parent, yielding a mere redundant statement ‘my parent is a parent’, which in turn makes the italicized sentence anomalous - ‘he, my parent is a parent’. Note that when a specific item is mentioned explicitly as the topic of [[X wa X] (da)] (e.g. kare in (58)), the usual interpretation of X wa X (da) as a general statement concerning any X is curtailed, and category immutability is claimed only for the specific item. This is probably because the presence of the ‘extra’ topic syntactically overrides [X wa] as topic NP, which in turn precludes the reading of X wa X as a generic statement. On the other hand, when nothing is mentioned explicitly as the topic of X wa X (da), as in (lo), [X wa] as topic is not overridden, and hence the interpretation of X wa X (da) as a generic statement about X is not blocked; at the same time, however, it is also construed as involving the specific item. These observations thus support the suggested structural analysis of X wa X involving a dual treatment of X wa, as topic and as part of the predicate. 7.2. XgaXandXmoX As in the case of X wa X, the meaning of X gaX is not fully compositional: No word in X ga X denotes the undesirable quality of the referent of X. However, the association between the pattern X ga X and its meaning is not entirely arbitrary, either. As discussed in section 5, the repetition here serves as a device for avoiding an explicit description of something undesirable, namely, the undesirable quality of the referent in question. To this extent, the meaning of X ga X is motivated, though not predictable. The choice of the specific particle ga in X ga X is not arbitrary, either: it can be related to the general use of ga for marking the subject of a sentence representing a thetic judgment. Unlike the topic (NP wa) in a categorical judgment, which is the center of attention, the subject of a thetic judgment is

460

S. Okamoto / Japanese nominal repetitive

constructions

not the main focus of attention but simply a constituent (Kuroda, 1972).18 Observe the following two examples:

of the proposition

(61) Otenki ga warm kara, pikunikku wa yameyoo yo. weather SM bad because let’s cancel (the) picnic ‘(lit.) Because (the) weather is bad, the picnic, let’s cancel (it).’ (62) Otenki ga otenki, da kara, pikunikku wa yameyoo yo. weather SM weather COP because let’s cancel (the) picnic ‘(lit.) Because (the) weather is (bad) weather, the picnic, let’s cancel (it).’ In (61) otenki ga is the subject and warui is its predicate. Likewise, in (62) otenki ga can be considered the subject because the italicized sentence is about the weather. (But otenki da by itself cannot be regarded as the predicate. See below.) In both (61) and (62) otenki ga is not the center of attention, hence cannot be marked by the particle wa: The center of attention or topic in (61) and (62) is pikunikku wa. As seen in (62), X ga X is typically used in a dependent clause, particularly in a causal clause; the use of X ga X in a main clause is awkward (e.g. (63)),l” which suggests that the referent of X in X ga X is not the center of attention in the given sentence as a whole. (63) # Saikin,

seeseki da. benkyoo-site-nai kara, seeseki ga recently, because (I)‘m not studying grade SM grade COP ‘Because I’m not studying recently, my grades are (bad) grades.’

Furthermore, X ga X as a whole is normally used as a comment about the topic which (implicitly or explicitly) precedes Xga X. For example, in (65) tosi ga tosi da is the comment about kare, which is the topic or the center of attention: kara . . . tosi ga tosi da he TM age SM age COP because ‘(lit.) Because he, (his) age is (not a desirable) age, . ..’

(64) Kare wa

Or, in (62) otenki ga otenki can be regarded as a comment about an implicit topic (e.g. kyoo ‘today’), which could be explicitly mentioned (e.g. Kyoo wa otenki ga otenki da kara). Example (64) is similar to a sentence like (65) which represents a particularly common sentence-type, containing a topic NP (kare wa) as well as a subject NP (tosi ga): I8 The notion of thetic judgment is similar to that of neutral description as characterized by Kuno (1973). I9 This does not, however, mean that a thetic judgment or NP ga in main clauses is impossible.

S. Okamoto / Japanese nominal repetitive

constructions

461

(65) Kare wa

tosi ga wakai. he TM age SM young ‘(lit.) He, (his) age is young.’

By analogy to the structure of (65), we can properly regard kare wa in (64) as the topic or the center of attention and tosi ga as the subject; that is, tosi ga in (64) is simply a constituent in the comment-clause, not the center of attention. (66) Kare he ‘(lit.) after

memdoo-minakutya. kara, wa oya ga oya da TM parent SM parent COP because (he) must look after (him) Because he, (his) parent is (not a desirable) parent, (he) must look (him).’

Similarly, in (66), kare wa and oya ga refer to two different individuals, and the former represents the topic and the latter the subject. Thus, the particle ga in X ga X can be related to the function of ga for marking the subject of a sentence representing a thetic judgment. It might be further argued that X ga X is in fact analyzable, in that X ga is the subject and the second X is its predicate meaning ‘undesirable X. However, such an analysis is inappropriate because the meaning ‘undesirable’ cannot be effected by the second X alone; the repetition of X in the form of X ga X is required. In other words, the meaning ‘undesirable’ can be conveyed only by X ga X as a whole. In this respect, X ga X cannot be analyzed as a regular subjectpredicate sentence (IX ga] /X (da)j). I therefore employ a dual analysis for X ga in X ga X, analogous to my analysis of X wa X. That is, in the interpretation of X ga X, X ga is both construed as the subject (IX gafi, and at the same time reanalyzed as part of the predicate UX ga Xfi. Though unconventional, this dual analysis enables us to appropriately account for the meaning of X ga X, which is partly analyzable and partly idiomatic. Similarly, the meaning of X mo X is motivated rather than entirely arbitrary. The repetition of X serves the same function as in X ga X: to avoid an explicit mention of the undesirable property. Further, the appearance of the particle mo is itself non-arbitrary, but can be understood in terms of one of its regular meanings, ‘also’: In X mo X, the use of mo presupposes another item, in addition to X, that has the same undesirable quality. Yet the meaning of X mo X is not fully analyzable by regular compositional semantics because no word in X mo X denotes the meaning ‘undesirable’. As in the case of X ga X, the interpretation of X mo X requires a reanalysis of X mo: first as subject,20 then as part of the predicate. 2o X MO should be regarded as the subject, since X mo X as construction is based on Xga A’, as I have shown. Note, however, that unlike X ga A’, X MO A’ may be used in either a dependent or a main clause (e.g. (43)).

462

S. Okamoto / Japanese nominal repedtive

constructions

To summarize, the core meanings of the three repetitive constructions are motivated rather than totally arbitrary: The repetition of a noun phrase and the particular particle have a certain bearing on the meaning(s) of each construction. Yet these core meanings are not fully analyzable by regular compositional semantics based on a regular phrase structure analysis (i.e., a topiccomment or subject-predicate structure). Rather, each core meaning is to be regarded as conventionally associated with each construction as a whole; and the construction can be appropriately analyzed only by recourse to irregular phrase structures based on a dual analysis of a subcomponent of the construction. That is, in order to describe the meanings of the three constructions adequately, we need to resort to syntactic configurations larger and more complex than those definable by regular single phrase structure rules, as is the case with many ‘grammatical constructions’ (e.g. Fillmore et al., 1987). It is, therefore, most appropriate to treat each of the three constructions as a ‘grammatical construction’ - as a particular morphosyntactic pattern bearing specific semantic and pragmatic properties.

8. The ‘tautology’ controversy revisited At the beginning of this paper (section 2), I referred to the three main approaches to nominal ‘tautologies’ that have been proposed in the literature: (1) radical semantics, (2) radical pragmatics, and (3) non-radical semantics and pragmatics. Now we may examine each of these approaches in relation to my analysis of the Japanese construction. 8. I. The radical semantic approach The radical semantic approach has been criticized by a number of researchers (e.g. Fraser, 1988; Ward and Hirschberg, 1991) as being too specific to be able to provide context-independent semantic representations in a systematic way. My analysis of X wa X, too, suggests that the radical semantic approach cannot be maintained. As we have seen, X wa X is used to emphasize the category immutability of the autonomy of an item. However, it could be argued that in addition to these meanings, X wa X may plausibly be used to convey other meanings, in particular attitudinal meanings, such as resignation, tolerance, defiance, obligation, criticism, or soberness. (67) Hanarete ite even though ‘Even though (68) Zyooken ga even though ‘Even though

wa oya. mo, oya (he) is away, parent TM parent (he) is away, a parent is a parent.’ sigoto wa sigoto. warukute mo, TM job (the) conditions are bad, job the conditions are bad, a job is a job.’

S. Okamoto / Japanese nominal repetitive constructions

463

For example, (67) may suggest certain social obligations and (68) may indicate resignation. According to Wierzbicka’s radical semantics, these attitudinal meanings are not inferable from any pragmatic principles, but rather must be explicitly included in the semantic representation of each type of tautological expression (see section 2). I would argue, however, that these additional meanings are inferable pragmatically, and not part of the core meanings of ‘tautological’ expressions. By way of evidence, we may note that these meanings can be explicitly stated separately in the same discourse: wa oya. Mendoo-mi-nakutya. (69) Hanarete ite mo, oya even though (he) is away, parent TM parent (I) must look after (him) ‘Even though he is away, a parent is a parent. I must look after him.’ sigoto wa sigoto. (70) Zyooken ga warukute mo, TM job even though (the) conditions are bad, job Damatte yaru yo.

(I) will do (it) quietly ‘Even though the conditions are bad, a job is a job. I will do it without complaining.’ Thus in (69) Mendoo-mi-nakutya explicitly refers to the feeling of obligation, and in (70) Damatte yaru yo expresses the speaker’s resignation. More importantly, however, these attitudinal meanings may vary considerably, depending on the context. For example, in (67) the expression oya wa oya emphasizes the immutability of the category ‘parent’. From this, various context-dependent inferences may be drawn. For example: the person in question is indeed my parent. Therefore, I must take care of him even though he cannot be considered an ideal parent (obligation); or, it is understandable that he should behave like a typical parent (tolerance); or, I must appreciate what he does for me as my parent (appreciation); etc. Similarly, sigoto was sigoto in (68) could imply obligation, resignation, appreciation, tolerance, defiance, etc. It is difficult to see how these numerous meanings could be represented in rigorous semantic, or context-independent, representations. (This argument is equally applicable to ‘tautologies’ such as A job is a job in English and other languages. See also Fraser, 1988, and Ward and Hirschberg, 1991, for similar discussion.) I contend that the attitudinal meanings suggested by these ‘tautological’ expressions are inferable, as illustrated above, from the construction’s core meanings in combination with discourse context and general knowledge about the category X, including sociocultural expectations.

464

8.2.

S. Okamoto / Japanese nominal repetitive constructions

The radical pragmatic approach

The second approach, that of Gricean radical pragmatics, assumes that ‘tautological’ expressions are uninformative by themselves, but that they convey meanings in the form of inferred implicatures based on Gricean pragmatic principles. However, as Levinson (1983 : 111) himself recognizes, it remains quite unclear how such implicatures are to be worked out. That is, it is difficult to see how recognizing that the Quantity maxims are intentionally violated can make an uninformative sentence (e.g. Boys are boys) informative or meaningful in any specific way (e.g. ‘That’s the kind of unruly behavior you would expect from boys’). As I have demonstrated, associated with the ‘tautological’ construction X wa X we can recognize specific core meanings, which are not contingent upon the context of utterance. The conventionality of the meanings of X wa X becomes evident particularly when we compare X wa X with X ga X and X mo X; despite the apparent similarities, each morphosyntactic pattern has its own distinct core meaning or meanings. The interpretation of X wa X cannot, therefore, be totally relegated to the realm of pragmatic inference. Furthermore, given the core meanings of X wa X, the derivation of particularized implicatures such as those discussed by Levinson can be accounted for in a more coherent manner. For example, by recognizing that one cannot change the categorial attribute of the given person as a parent, one may draw a particular conclusion that is both generally expected of the category ‘parent’ and at the same time relevant to the current discourse world (e.g. obligation, tolerance). This inference process is in principle the same as in the case of deriving attitudinal meanings, which was illustrated in subsection 8.1. As mentioned in subsection 2.1, Ward and Hirschberg (199 l), proposing a new Gricean approach, claim that tautological utterances are used (universally) to convey a generalized implicature, namely, a rejection of the relevancy of other, alternative propositions (i.e. ‘a is b’ as opposed to ‘a is a’). This is tantamount to saying that tautological expressions are used to imply ‘uninformative’, is that only the proposition ‘a is a’, which is considered relevant to the current discourse. As such, however, Ward and Hirschberg’s characterization does not fully capture the function of ‘tautological’ expressions. 8.3. The non-radical approach My study thus supports the third approach, non-radical semantics and pragmatics. That is, the interpretation of X wa X is best accounted for in terms of both conventionality and pragmatic calculability: X wa X is conventionally used to emphasize the immutability of the category or the autonomy of an

S. Okamoto / Japanese nominal repetitive constructions

465

item,21 and additional meanings may be conveyed as implicatures, which are inferable from the core meanings of the construction in combination with general knowledge about the category X and the context of utterance. As mentioned in section 2, this third approach to nominal ‘tautologies’ is also that proposed by Fraser (1988), Escadell-Vidal (1990), and Farghal (1992). However, as pointed out by Ward and Hirschberg (1991), Fraser’s definition of the meaning of nominal ‘tautologies’ is so broad that it can be applied not only to nominal ‘tautologies’ but also to many other expressions. Escandell-Vidal’s definition is inadequate in that the repetition seen in ‘tautological’ expressions need not intensify the quality of the NP, as she claims. The core meanings of ‘tautological’ expressions given by Farghal are mostly attitudinal or evaluative; however, such meanings are highly contingent upon the discourse context, as discussed in the previous subsections. My own analysis of X wa X, as a device for emphasizing the category immutability or the item’s autonomy, would seem applicable to nominal ‘tautologies’ in other languages as well, although details of usage may differ from language to language; further cross-linguistic investigation is called for.

9. Conclusion Most studies on nominal tautologies, whether radical or non-radical, have been concerned with implicatures and attitudinal meanings. In part, at least, this focus has resulted from the analysis of single tautological sentences in isolation. In the present study, I have investigated how the Japanese ‘tautological’ construction X wa X is used in discourse, and compared the results with the usages of two other formally similar constructions: X ga X and X mo X. Through these analyses, I have demonstrated that the three constructions have distinct core meanings: X wa X emphasizes the category immutability or the autonomy of an item; X ga X indicates that the referent of X has some undesirable quality; and X mo X indicates that the referent of X as well as another item (present in the context) has an undesirable quality. I have argued that the core meanings of the three constructions are motivated rather than totally arbitrary: the particular particle and the repetition of the noun have a certain bearing on the meaning(s) of each construction. However, it has also been shown that each core meaning is not fully analyzable by regular semantic interpretation rules. Rather, it is to be regarded as conventionally associated with the particular morphosyntactic pattern as a whole, whose structure can

*I Wierzbicka (1987: 109) points out that nominal tautologies may have a semantic invariant of their own, which she paraphrases as follows: “An X is not different from other X’s (all x’s are the same). This cannot change”. However, as noted by Wierzbicka herself, there are many uses of nominal tautologies to which this definition cannot be applied (e.g. East is East: you are you).

466

S. Okamoto

/ Japanese nominal repetitive constructions

be adequately analyzed only by recourse to irregular phrase structures. Thus, it is most appropriate to treat each construction as a ‘grammatical construction’ in the sense of Fillmore et al. (1987). With regard to the ‘tautology’ controversy, I have argued that the interpretation of X wa X is best accounted for in terms of both conventionality and pragmatic calculability: X wa X has conventionalized core meanings, and it may evoke additional meanings as implicatures, which are inferable from the core meanings of the construction in combination with general knowledge about the category X and the context of utterance. Consequently, the two radical approaches ~ radical semantics and radical pragmatics - are both found to be inadequate for the interpretation of nominal ‘tautologies’.

References Escandell-Vidal, Victoria, 1990. Nominal tautologies in Spanish. Paper presented at the 1990 International Conference on Pragmatics, Barcelona, Spain. Farghal, Mohammed, 1992. Colloquial Jordanian Arabic tautologies. Journal of Pragmatics 17: 223-240. Fillmore, Charles, Paul Kay and Mary C. O’Connor, 1988. Regularity and idiomaticity in grammatical constructions. Language 64(3): 501-538. Fraser, Bruce, 1988. Motor oil is motor oil: An account of English nominal tautologies, Journal of Pragmatics 12(2): 215-220. Gibbs, Raymond W. and Nancy S. McCarrell, 1990. Why boys will be boys and girls will be girls: Understanding colloquial tautologies. Journal of Psycholinguistic Research 19(2): 125-145. Grice, Paul H., 1975. Logic and conversation. In: P. Cole and J.L. Morgan, eds., Syntax and semantics, Vol. 3, 41-58. New York: Academic Press. Kuno, Susumu, 1973. The structure of the Japanese language. Cambridge, MA: MIT Press. Kuroda, S.-Y., 1972. The categorical and the thetic judgment: Evidence from Japanese syntax. Foundations of Language 9(2): 153-185. Levinson, Stephen C., 1983. Pragmatics. Cambridge: Cambridge University Press. Okamoto, Shigeko, 1991. Nominal ‘tautologies’ in Japanese: X wa A’. X ga A’, and X lylo X. Proceedings of the 17th Annual Meeting of the Berkeley Linguistics Society, 2 I S-229. Ward, Gregory and Julia Hirschberg, 1991. A pragmatic analysis of tautological utterances. Journal of Pragmatics 15(6): 507-520. Wierzbicka, Anna, 1987. Boys will be boys: ‘Radical semantics’ vs. ‘radical pragmatics’. Language 63(l): 95-l 14. Wierzbicka, Anna, 1988. Boys will be boys: A rejoinder to Bruce Fraser. Journal of Pragmatics 12(2): 221-224.