Abstract
The study investigates how speakers use ‘nai-expressions’ (a verb root + the negative suffix (a)nai as in shabere-nai ‘cannot speak’ and ik-anai ‘will/do not go’) in naturally occurring conversation. The data demonstrate that although negative utterances have been considered to be ‘grammatical’ constructions that simply negate the truth value of a proposition, nai-expressions show formulaic tendencies and serve not only to express a speaker’s emotional personal stance on a particular story/event but also to create interpersonal space with other conversation participant(s) and to involve them in the story/event. The patterns which emerged from the data are quite similar to those in Ono and Thompson’s study on Japanese adjectives. As they pointed out, some of nai-expressions in the data may also be re-analyzed as adjectives.
Keywords
Introduction
Japanese negatives have been extensively investigated with attention given specific to syntactic and semantic dimensions (e.g. Ikeya and Kawamori, 1998; Iwakura, 1974; Kato, 1985; Kuno, 1980; Takubo, 1985). While these studies were based on constructed data, some studies are grounded on natural language use (McGloin, 1986; Yamada, 2000, 2003; Ono and Thompson, 2017) and have highlighted examples where ‘negatives are not simply a reversal of the truth value of the corresponding affirmatives’ (McGloin, 1986: 99) but also encode pragmatic notions, such as markedness, speaker-/discourse-/hearer-motivated negation, and textual/contextual expectations, as part of lexical meanings (McGloin, 1986; Yamada, 2000, 2003).
The Japanese negative nai can be categorized into two types for its syntactic and semantic roles (Ikeya and Kawamori, 1998): (a) as jiritsugo or an independent word treated as an adjective meaning ‘non-existent, absent’ and (b) as fuzokugo or a dependent word. Nai, as an independent word, is the plain nonpast negative form of the existential and possessive verb aru as shown in example (1) from my conversational data: (1) a. . . .shikke-ga humidity-GA exist:NEG FP
1
‘there’s no humidity, is there?’ b. . . . okane-nanka hotondo money-NANKA almost possess:NEG FP ‘(he) hardly has things like money’
Example (1a) is an example of nai describing the non-existence of ‘humidity’ and (1b) shows nai as the non-possession of ‘money’. In most contemporary Japanese dictionaries, nai as an independent word is listed as an ‘adjective’ like taka-i ‘expensive’ and chiisa-i ‘small’ since nai ends with -i and shares the same inflectional patterns as those adjectives: Nonpast affirmative Past affirmative Past negative -------------------------------------------------------------------------------------- na-i ‘nonexistent’ na-katta na-kunakatta taka-i ‘expensive’ taka-katta taka-kunakatta
Nai also functions as a dependent word, specifically, a negative suffix. According to the online Nihon Kokugo Daijiten (n.d.), the largest Japanese-language dictionary, this particular usage began to appear in the Kanto area (the current Greater Tokyo Area) during the late Muromachi Period (1336–1573). It is added to the root of verbs, as in tabe- (2) M and K are housemates. M tries to recall the name of the very expensive sports car her younger brother bought. 1 M: . . . n nantoka .. Zekkusu da ka, such and such Zex COP Q 2 nan da [ka shiranai kedo], What COP Q know:NEG FP 3 K: [Aaru Ekkusu Se]bun? RX7 4 M: . . . nan= da ka wakannai. What COP Q know:NEG 5 shiranai atashi .. mada-, know:NEG I still 6 .. oboekirenai. memorize completely:POT:NEG --------------------------------------------------------------------- M: (I) K: (you mean) RX7? M: (I)
A series of clauses with the negative suffix nai appears in (2) where speaker M is trying to recall the name of the car her brother purchased: shiranai ‘don’t know’ in lines 2 and 5, wakannai ‘don’t know’ in line 4, and oboekirenai ‘cannot memorize (it) completely’ in line 6.
As pointed out in Ono and Thompson (2009), there are some expressions with nai which are structurally fixed to the point where the affirmative counterparts of those expressions are non-existent: (3) a. tondemo otonage shooga dooshiyoomo tamara b. ~nakyaike ~chaike ~kamoshire ~taku
The expressions in (3a) are listed in the online Nihon Kokugo Daijiten (n.d.) as either adjectives or compounds, and those in (3b) serve as modal phrases and express obligation, prohibition, uncertainty or undesired activity. Both of them are more likely to be holistically stored and retrieved as one lexical unit (e.g. tondemonai ‘far from it’) rather than multiple morpheme units (e.g. to(n) ‘reason’ + the copula de + mo ‘also’ + nai ‘exist’).
The current study focuses on examples with the negative suffix nai as shown in (2) and calls them ‘nai-expressions’ hereafter. This study is firmly in keeping with a usage-based perspective on language (e.g. Barlow and Kemmer, 2000; Bybee, 2007), and offers new insights into the interactional and performative nature of language by addressing how language speakers use nai-expressions in everyday talk from a viewpoint of discourse-pragmatics. Although traditionally Japanese negation had been treated as simply the reversal of the truth value of a proposition (e.g. Iwakura, 1974; Kato, 1985; Kuno, 1980; Takubo, 1985), based on constructed examples like Watashi-wa Pari-de tokei-o kawa-na-katta I-TOP in-Paris watches did-not-buy, ‘In Paris, I didn’t buy watches’ (Takubo, 1985: 40), where the negative form of an activity verb (‘bought’) occurs with two lexically expressed core arguments (‘I’ and ‘watches’), the discourse-based studies pointed out that many Japanese negatives are, in fact, structurally fixed (Ono and Thompson, 2017) and used to express linguistic subjectivity (i.e. a speaker’s/writer’s voice, emotion, feeling, attitude and point of view in discourse) (McGloin, 1986), as well as intersubjectivity (i.e. a speaker’s/writer’s attention to the hearer/reader) (Yamada, 2000, 2003). While these discourse-based studies revealed important findings on Japanese negation in general, there have been no studies specifically on the negative suffix nai in everyday talk which investigates the interaction between grammar and pragmatics. By examining the use of nai-expressions in terms of formulaicity and intersubjectivity in naturally occurring conversation, this study provides a more comprehensive account, demonstrates how their occurrences in everyday talk are different from the constructed examples, and reveals that although negative utterances have been considered to be ‘grammatical’ constructions that simply negate the truth value of a proposition, their use is actually pragmatically motivated, showing speakers attention to not just the content of information but also to their conversation participants as well as the participants’ involvement in ongoing talk.
Background: Formulaicity and intersubjectivity
The current study takes a usage-based view as its starting point that ‘the speaker’s linguistic system is fundamentally grounded in “usage events”: instances of a speaker’s producing and understanding language’ (Barlow and Kemmer, 2000: viii). More specifically, the study conforms to the critical role of frequency on representation that frequency of a particular usage pattern is ‘both a result and a shaping force of the system’ (Barlow and Kemmer, 2000: x). This section lays out the two key factors which are of great relevance to the current analysis of nai-expressions in everyday conversation: formulaicity and intersubjectivity.
Formulaicity
Formulaicity or formulaic language has been extensively discussed and researched in a wide array of themes and features such as discourse function, human cognition, interaction, pragmatics, L1 and L2 acquisition, and intercultural communication, and these studies have reported how prevalent formulaic expressions are in human language (e.g. Bardovi-Harlig, 2006, 2012; Cameron-Faulkner et al., 2003; Corrigan et al., 2009; Erman and Warren, 2000; Wood, 2012; Wray, 2002, 2008).
According to Corrigan et al. (2009), what makes formulaic language different from non-formulaic language is its restriction in form and distribution. Formulaic expressions, such as the check is in the mail in response to an inquiry and your call is important to us in a voice mail message when the caller is put on hold, are structurally more rigid, non-compositional, and thus are ‘more like words and morphemes than ordinary sentences’ (e.g. We have placed the check in the mail; Your telephone call has great importance to us) (p. xiv). Structural fixedness is also addressed in Ono and Thompson’s (2017) study on English and Japanese negatives: ‘a significant portion of the entries in dictionaries of Japanese fixed expressions actually involve negation’ (pp. 560–561). Furthermore, Japanese negatives tend to appear in fixed expressions and often co-occur with ‘negative-anticipating forms’, such as zenzen ‘at all’ and amari/anmari ‘not very’ (Ono and Thompson, 2017: 562) as in the following example: (4) at all day off exist:NEG QT say:STAT:PAST FP ‘(she) was saying that (he could) take no day off at all’
According to Ono and Thompson (2017), in (4), a negative-anticipating form zenzen occurs at the beginning of the clause and signals that a semi-fixed form with yasumi nai ‘no day off’ is coming up.
In terms of distribution, formulaic expressions are also more likely to ‘occur in particular styles of language tied to particular communicative situations’ (Corrigan et al., 2009: xiv). A number of studies such as Biber et al. (1999) and Biber and Conrad (1999) reported heavy use of formulaic language in spontaneous spoken language. For example, Calude (2009) reported the frequent use of demonstrative clefts such as that’s what I think and that’s what they say in informal conversation. In their study on Japanese adjectives in everyday talk, Ono and Thompson (2009) demonstrate higher occurrence of predicate adjectives (e.g. ii nee ‘good!’; sugoi jan ‘terrific, isn’t it!’) and fully or semi-lexicalized expressions (e.g. kimochiii ‘feel good’; isogashii-hoo-ga ii ‘it’s better to be busy’) in their conversational data.
Another frequently mentioned characteristic is that formulaic expressions share some properties as single words in that they are also learned, stored and retrieved as chunks or single units. The idea of holistic storage or chunks is indicated by the prosodic patterns of formulaic language in terms of ‘phonological cohesiveness’ (Bardovi-Harlig, 2009; Hickey, 1993) or phonological fluency (Calude, 2009; cf. ‘nativelike fluency’ in Pawley and Syder, 1983). For example, formulaic expressions such as Believe you me!, That’s what I think and That’s what they say are more likely to be pronounced faster with no pauses or breaks and with fewer changes in pitch and intonation contour (Calude, 2009; Wray, 2002). In her examination of English demonstrative cleft sentences, Calude (2009) found only six examples out of 205 which had any pauses during the course of the cleft (p. 70). It has also been reported that code-switching is likely to occur not during the course of formulaic expressions but at the boundary of them (Backus, 1999). Furthermore, Bybee (2002, 2006) pointed out phonological reduction as another characteristic of formulaic language: ‘high-frequency words and phrases undergo phonetic reduction at a faster rate than low- and mid-frequency sequences’ (Bybee, 2006: 714). The strong connection between high-frequency words and phonetic reductions is supported by such phenomena as the reduction of don’t in sequences like I don’t know, I don’t think and I don’t have in American English (Bybee and Scheibman, 1999); English final t/d deletion (Bybee, 2002); and the palatalization in sequences like did you, don’t you, would you and that you (Bush, 2001).
In the case of Japanese negatives, Ono and Thompson (2017) pointed out a tendency that Japanese negative forms are very short and most of those short negatives are produced within the same intonation contour (p. 20). Uehara and Fukushima (2008), in their study on the two negative forms – the polite negative form -masen (e.g. tabemasen ‘don’t eat’) and the semi-polite style ‘verb-nai desu’ (e.g. tabenai desu ‘don’t eat’) in naturally occurring conversation – noted a close association of the latter form with contractions as in wakkannai desu kedo (non-contracted form wakaranai desu kedo) ‘(I) don’t understand, though’ (pp. 180–181).
While these characteristics might be useful to distinguish a formulaic expression from a non-formulaic one, Corrigan et al. (2009) caution that the restrictions in form and distribution of formulaic language are not absolute but ‘merely probabilistic’ since formulaic expressions may occur in varied forms, or allow other linguistic elements within the expressions, and be used in situations other than their preferred ones (p. xiv). In order to capture differing degrees of formulaicity, the current study, following Calude (2009), uses the term ‘formulaic tendencies’ to describe what speakers actually do using formulaic expressions in everyday conversation. The term ‘formulaic tendencies’ seems quite justifiable since ‘the inventory of formulaic expressions is a dynamic system which undergoes constant change to meet the needs of language users’ (Wray, 2002: 101). It also allows us to perceive the notion of formulaicity as a continuum between more formulaic and less formulaic, as opposed to the dichotomy of formulaic and non-formulaic, and thus to include different degrees of formulaicity, such as semi-fixed phrases and situation-based expressions.
Intersubjectivity
Another essential aspect of formulaic language is its context- and situation-dependent feature. Formulaic expressions are used for specific social and pragmatic purposes in that they ‘often succinctly capture the illocutionary force of a contribution by virtue of the fact that the speech community in which they are used has tacitly agreed on their form, meaning, and use’ (Bardovi-Harlig, 2012: 207). Intersubjectivity is one such key notion that is pertinent to the current analysis of expressions with the negative suffix nai in informal conversation; it directs the analysis from the language users’ perspectives.
While linguistic subjectivity generally refers to the expressions that encode a speaker/writer’s voice, emotion, feeling, attitude and point of view in discourse, intersubjectivity denotes the expressions of a speaker/writer’s attention to the hearer/reader. A number of studies have shown how a speaker’s attention is given not just to the content of information but also to the addressee in interactive contexts. Shinzato (2006), for example, provided a succinct summary of various cross-linguistic and cross-disciplinary research which addressed distinct roles of the speaker and the addressee to classify modality (Benveniste, 1971; Haga, 1954; Maynard, 2002; Sweetser, 1990). Citing Haga (1954), Shinzato explained that modality (chinjutsu in Japanese) consists of juttei ‘judgment’ or ‘the speaker’s attitude toward the proposition’ and dentatsu ‘communication’ or ‘the illocutionary force directed toward the addressee’ (p. 17), and pointed out a striking similarity between Haga’s proposal and Benveniste’s notion of subjectivity and intersubjectivity.
While in Shinzato (2006), subjectivity and intersubjectivity were presented in terms of their binary relationship, Scheibman (2007), in her examination of generalization types (i.e. grammatical subjects that refer to classes or groups, such as all these teachers, worse things and the French), took a view that a particular expression is subjective and intersubjective: ‘these are expressions of position and attitude that are relevant both to individual speakers (their subjective uses) and to relational activities among participants (their intersubjective uses)’ (p. 111). In line with Scheibman, Sadler (2010) examined two commonly used Japanese verbs of cognition – wakaru and shiru ‘to know’ in everyday conversation – and demonstrated that a speaker’s lexical choice is intended for the sharing of his or her subjective and intersubjective stance with other conversation participant(s) in interactive contexts. Specifically, the use of wakaru is closely associated with such features as experiencer perspective, speaker empathy and involvement, directness, and immediacy whereas the use of shiru is characterized as observer/reporter perspective, speaker detachment and indirectness.
In the case of Japanese negatives, McGloin (1986), based on written discourse data, pointed out the markedness and highly evaluative use of Japanese negatives as pragmatic functions: Negatives are presuppositionally more marked than the corresponding affirmatives, and hence are used not only to negate the corresponding affirmatives but to express the speaker’s stronger wish or to give the speaker’s value judgement. (p. 122)
With particular regard to the suffix nai, McGloin pointed out that adding nai gives ‘emotional emphasis’, as in messoomonai ‘it couldn’t be; don’t be absurd’ – consisting of the adjectival noun messoo ‘absurd’, mo as the particle of additionalness, and the negative suffix nai – in contrast to its affirmative counterpart messoona ‘absurd’ (p. 109).
The intersubjective nature of Japanese negatives was extensively investigated by Yamada (2000, 2003). Based on the detailed discourse-level analyses of his 71 narratives of personal experience collected using the sociolinguistic interview technique, he provided a list of pragmatic functions of negatives such as denial of expectations, contrast and evaluation. Of particular note is his characterization of negatives as evaluation, similarly to McGloin (1986). According to Labov and Waletzky (1967: 37, cited in Yamada, 2003: 217), evaluation ‘reveals the attitude of the narrator towards the narrative by emphasizing the relative importance of some narrative units as compared to others’. Going one step further, Yamada (2003) addressed the expressive and interactional functions of evaluation. Namely, the evaluative use of negatives shows not only a speaker’s personal involvement with a story/event itself but also his or her attention to the audience (cf. an involvement strategy in Tannen, 1989) to ‘create interpersonal involvement between narrator and audience, and also the audience’s involvement with the narrative’ (Yamada, 2003: 393). This sensitivity to the audience was also a focal point in Takano’s (2008) study. His close examination of the ways in which negation is prosodically realized in the three various social situations (i.e. casual conversation, televised debates and broadcast news read by newscasters on televised news programs) demonstrated how a speaker utilizes prosodic variation (i.e. prosodic focus on a negative utterance or non-prosodic focus) to communicate information effectively and/or express his or her emotion and attitude.
The intersubjective use of nai is also discussed in some studies of stylistic variation. Uehara and Fukushima (2008) examined the two negative forms – the polite negative form -masen (e.g. tabemasen ‘don’t eat’) and the semi-polite style verb -nai desu (e.g. tabenai desu ‘don’t eat’) – in naturally occurring conversation. They discovered that while the polite negative form -masen occurs at the beginning of a conversation and/or after the conversation came to a halt, the semi-polite style verb -nai desu is more likely to appear ‘as the speaker’s familiarity, and possibly comfort, with the interlocutor increases in the course of their “first-encounter” conversation’ (p. 180), suggesting that the shift from -masen to -nai desu signals the speaker’s attention to the different level of politeness. While both forms are considered to be ‘polite’ forms, -nai desu is the less formal, semi-polite sub-level than -masen. However, Hudson (2008), based on her analysis of a contemporary Japanese novel, suggested the purpose of using verb-nai desu is ‘to convey to the addressee the meta-linguistic message’ (p. 151) by signaling ‘the speaker’s psychological distance from the hearer’ and imparting ‘the meaning that the speaker is explaining or evaluating a situation’ (p. 156). While it seems that in Uehara and Fukushima (2008), the use of -nai desu demonstrates the speaker’s effort to involve conversation participants in their first-encounter conversation, -nai desu in a contemporary Japanese novel (Hudson, 2008) that has the opposite effect – the speaker’s distancing from the addressee.
The current study
Goals
As addressed in the previous section, the current study takes a usage-based perspective on language – that language use shapes language form and meaning (e.g. Barlow and Kemmer, 2000; Bybee, 2007). Specifically, this study captures how language speakers use nai-expressions in everyday talk and examines two dimensions: formulaic tendencies and intersubjectivity. As pointed out in previous sections, most Japanese negation studies were based on constructed data and from the view that negation is simply the opposite of its affirmative counterpart, and the pragmatic nature of negation were absent or neglected (e.g. Ikeya and Kawamori, 1998; Iwakura, 1974; Kato, 1985; Kuno, 1980; Takubo, 1985). Even among a handful of discourse-based studies, naturally occurring conversation was not their data source (i.e. written discourse in McGloin, 1986; narrative in Yamada, 2000, 2003). To my knowledge, Ono and Thompson (2017) was the only study that utilized naturally occurring conversation as the object of an analysis, but the pragmatic nature of negation was not the focus of their study. Although a number of previous studies (e.g. Iwakura, 1974; Kato, 1985; Kuno, 1980; Takubo, 1985) and language textbooks used constructed examples like Watashi-wa Pari-de tokei-o kawa-na-katta I-TOP in-Paris watches did-not-buy, ‘In Paris, I didn’t buy watches’ (Takubo, 1985: 40), where the negative form of an activity verb ‘bought’ occurs with two lexically expressed core arguments (‘I’ and ‘watches’), it is unclear whether speakers use such an utterance in everyday talk. Even a casual observation of nai-expressions in conversation reveals the patterns of its structure markedly different from constructed examples in prior literature; many nai-expressions in conversation seem to be non-activity verbs occurring without any overt core arguments, as demonstrated in example (2) (e.g. wakannai ‘(I) don’t know (that)’). The current study investigates what predicate types of negative utterances (i.e. activity verb vs non-activity verb) are more commonly used in natural language and whether negative utterances occur with overtly expressed arguments. In addition, this study explores the pragmatic nature of negative utterances in naturally occurring conversation, which has not been sufficiently investigated.
The current study provides a comprehensive account for how Japanese language speakers use nai-expressions in everyday conversation and aims to accomplish three goals. The initial objective is to investigate how frequently nai-expressions are used and what shape(s) they take during use. Specifically, we examined the number of tokens of nai-expressions, predicate types (i.e. activity vs non-activity types, such as cognition and potentials), the number of predicative and attributive nai-expressions, and the number of overtly expressed core arguments occurring with nai-expressions. The data provide evidence that the patterns observed in everyday talk are significantly different from those observed in the constructed examples highlighted in prior literature.
The second goal is to examine how nai-expressions show formulaic tendencies (i.e. structural fixedness). The data demonstrate the varying degrees of structural fixedness involving (semi-)construction-like sequences as well as phonological coherence and reduction. The third goal is to demonstrate how speakers use nai-expressions to express their evaluative stance and to involve other conversation participants.
Data
The data for this current study are a private collection of naturally occurring conversations, consisting of 26 audio-taped casual conversations recorded in private homes and restaurants in Japan and the United States. These conversations include single-sex and mixed-sex groups with two to five participants ranging in age from 15 to 50 years old. The participants are mostly friends, couples and family, and all use standard Japanese. Each conversation is 3–18 minutes long. The total amount of data analyzed is approximately 150 minutes, or 278 pages of transcripted talk amounting to 4968 clauses.
Methodology
In order to capture a more comprehensive picture of how language speakers use nai-expressions in everyday talk, we first identified all utterances containing the negative suffix nai out of 4968 clauses. We identified 311 nai-expressions in total and categorized them all depending on the following predicate types: <Predicate types> (i) Non-activity verbs: Cognition wakannai ‘don’t know/understand’ omowanai ‘don’t think’ Potentials taberenai ‘can’t eat’ dekinai ‘can’t do’ Stative (including both ongoing and resultant actions consisting of a verb gerund and ~te(i)nai ) kiitenai ‘(I) am not listening; (I) haven’t heard’ yattenai ‘(I) am not doing; (I) haven’t done’ Others hairanai ‘won’t/don’t fit’ kakaranai ‘won’t/don’t cost money or take time’ (ii) Activity verbs: shinai/yannai ‘won’t/don’t do’ awanai ‘won’t/don’t meet’
Nai-expressions can be either predicative or attributive as shown in (5). Out of 311 nai-expressions, 277 were predicative and 34 were attributive.
(5) a. <Predicative nai-expression> ne shinjirarenai ne moo FP believe:POT:NEG FP any more ‘(I) can’t believe (it) any more/(it)’s unbelievable’ b. <Attributive nai-expression> [ shinjirarenai yoona shoosetsu no sekai no yoona] believe:POT:NEG like novel of world of like hanashi da story COP ‘(it)’s a story that is, like unbelievable, like a world of novel’
A potential nai-expression, shinjirarenai ‘can’t believe; unbelievable’, appears in both (5a) and (5b); the nai-expression is predicative in (5a) and attributive, occurring in a clause modifying the noun hanashi ‘a story’ in (5b). As Yamada (2003) stated, only the negations in main clauses determine the overall negativity of the utterances (resulting in sentence negation), while the negations in other parts, such as dependent clauses, adverbials and noun phrases, may only affect the local negation of these parts (resulting in constituent negation). (pp. 81–82)
Yamada’s ‘negation in main clauses’ roughly corresponds to predicative negatives in this study, therefore, following Yamada, the focus of this study is predicative nai-expressions as in (5a).
Furthermore, the number of overtly expressed core arguments were identified. Out of 277 predicative nai-expressions, 205 are identified as ‘2-argument nai’ or nai-expressions which can occur with two core arguments (traditionally called ‘transitive’) as shown in example (6), and 72 as ‘1-argument nai’ or those which can occur with one core argument (traditionally called ‘intransitive’) as in (7): (6) 2-argument nai-expressions: a. <two overtly expressed arguments> atashi goshujin-no-koto-wa zennzen shiranai kedo I husband-of-thing-WA at all know:NEG but ‘I don’t know things about (her) husband at all but . . .’ b. <one argument> soryaa shiranai that know:NEG ‘(I) don’t know that’ c. <zero argument> shiranai know:NEG ‘(I) don’t know (that)’ (7) 1-argument nai-expressions: a. <one overtly expressed argument>
hatachi ika niwa mienai desho ore sa
20 years old below NIWA look:NEG TEN I FP ‘I probably don’t look younger than 20 years old’ b. <zero argument> ookiku mienai n da yone big look:NEG NOM COP FP ‘(she) doesn’t look big, don’t you think?’
All examples in (6), contain the nai-expression shiranai ‘(I) don’t know (something)’, which is associated with two arguments (i.e.
Results
Overall frequencies
Predicate types
Among 4968 clauses examined, we identified 277 nai-expressions. Although most constructed negative sentences consist of an activity verb such as kawanai ‘don’t buy’, the predicate types of nai-expressions in the data show otherwise. Observe Figure 1:
Predicate types of nai-expressions.
As demonstrated in Figure 1, the majority of predicate types occurring with the negative suffix nai are
The number of overtly expressed core arguments
It is a well-established phenomenon that obligatory or core arguments are often left unstated in Japanese conversation (e.g. Clancy, 1980; Hinds, 1982; Maynard, 1989; Ono and Sadler, 2001; Ono and Thompson, 1997, 2017; Sadler, 2007). Ono and Thompson (2017) argue that because of that, Japanese clauses are relatively short. The data also support these claims.
The results demonstrate how rarely nai-expressions occur with two overtly expressed arguments (5.8%; 16 out of 277). Much more common are those with no argument (57.4%; 159 out of 277). A closer examination for which predicate type is more likely to occur with zero argument shows that Potential yielded the highest likeliness (63.2%; 43/69) followed by Stative (54.5%; 24/44) and Cognition (52.9%; 36/68).
The overall frequencies show that nai-expressions are quite frequent in everyday talk, and unlike the constructed examples observed in prior literature (e.g. Kuno, 1980; Takubo, 1985), the majority of nai-expressions are non-activity verbs, such as verbs of potentials (norenai ‘can’t ride’) and cognition (wakannai ‘don’t know/understand’). The results are also in accordance with studies done by Clancy (1980) and Hinds (1982) among others, and demonstrate that clauses with two overtly expressed arguments are extremely uncommon and thus more marked in everyday talk.
Do nai-expression show formulaic tendencies? If so, how?
It has been pointed out that Japanese negatives have a strong association with structural fixedness (Ono and Thompson, 2017). The nai-expressions in the data confirm this claim demonstrating the varying degrees of formulaicity.
The first evidence of formulaic tendencies in nai-expressions is their co-occurrence with ‘negative-anticipating’ forms (Ono and Thompson, 2017) such as zenzen ‘at all’ and amari/anmari ‘not very’, and noun phrases marked by the topic/contrastive marker wa, zettai(ni) ‘absolutely’, nanimo ‘anything’ (consisting of an interrogative nani ‘what’ and mo ‘also’) and daremo ‘anyone’(consisting of an interrogative dare ‘who’ and mo ‘also’) as shown in (8). As Ono and Thompson (2017) discussed, since negative predicates tend to occur with such negative-anticipating forms, whole sequences can be viewed as a type of fixed expressions, and they are often produced in the same intonational contour with no pauses or breaks. Of all 277 nai-expressions, approximately one-third of them (92 tokens) account for these types. Below are examples from the data: (8) a. Tookyoo no hito Tokyo of people WA such thing absolutely do:NEG FP ‘there’s no way Tokyoite would do such a thing, is there?’ b. atashi I anything know:NEG FP ‘I know nothing’
In (8a), the nai-expression shinai ‘don’t/won’t do’ occurs with the noun phrase Tokyoo no hito ‘Tokyoite’ marked with the topic/contrastive marker wa, and with zettai ‘absolutely’; (8b) is an example of shiranai ‘don’t know’ occurring with nanimo ‘anything’.
Another piece of evidence for the formulaic tendencies in nai-expressions comes from a phonological perspective; 37.5% of token (104 out of 277) are phonologically reduced as in (9): (9) wakannai (cf. wakaranai) ‘(I) don’t know/understand (that)’ shinnai (cf. shiranai) ‘(I) don’t know (that)’ ukannai (cf. ukaranai) ‘(I) won’t pass (that)’ attenai (cf. atteinai) ‘(I) haven’t seen (someone)’
It should be mentioned that over half of these phonologically reduced nai-expressions occur alone with no overtly expressed arguments as shown in (10): (10) K shares his experience with a homeless person while at the dinner table with his family. 1 K: . . . sooiu hito mo iru. such people also exist 2 M: . . . [sooiu hito mo iru] kedo, such people also exist but 3 E: [n= n=]. uh-huh uh-huh 4 M: .. [2 soo-janai 2] hito mo iru yo. so-COP:NEG people also exist FP 5 K: [2 n= 2]. uh-huh 6 E: . . . n=. uh-huh 7 K: . . . da .. honto, so really 8 . . . koo=, this 9 .. ironna hito iru kara ne=. various people exist FP FP 10 wakannai kedo. know:NEG FP 11 E: . . . n==. uh-huh ---------------------------------------------------------------------------- K: there are such people, too M: there are such people, too, but. . . E: uh-huh uh-huh M: there are people who are not like that, too K: uh-huh E: uh-huh K: so, really, there are all sorts of people. E: uh-huh
Example (10) shows another example of a phonologically reduced nai-expression. Prior to this excerpt, speaker K shared a story with his wife M and their friend E that he once had a drink with a homeless person. K discovered that this homeless person used to be an architect and became homeless by choice. The excerpt starts with K’s statement, sooiu hito mo iru ‘there are people like that, too’, in line 1, which invites M’s comment in lines 2 and 4 that ‘there are people like that but there are people who are not like that, too’. M’s comment was partially overlapped with E’s acknowledgment of K’s story in line 3 and K’s acknowledgment for M’s comment in line 5. E also expresses her acknowledgment of M’s comment in line 6. K then gives a recap of the story from line 7 to 9, ‘So, really, there are all sorts of people’. The phonologically reduced form wakannai of wakaranai appears immediately after his recap but in a different intonation contour in line 10. All 42 instances of wakanarai appear in its contracted form, wakannai, in the data.
It is also worth mentioning that this particular example also demonstrates structural fixedness of nai-expressions. In the data, wakannai as well as other Cognition types often co-occur with kedo (e.g. zenzen
In sum, I have demonstrated that nai-expressions have formulaic tendencies showing varying degrees of structural fixedness in which: (a) nai-expressions occur with negative-anticipating forms; (b) nai-expressions are produced in phonologically reduced forms; and (c) Cognition nai-expressions occur with the clause-ending kedo without main clauses.
Do nai-expressions show intersubjectivity? If so, how?
As discussed in the preceding section, the semi-fixed sequence consisting of a Cognition nai-expression and the clause-ending kedo (e.g. wakannai kedo ‘(I) don’t know . . . ’) is not just a negative form of the corresponding affirmative (i.e. wakaru ‘understand; know’) but serves as an involvement strategy, encoding speaker’s personal attitudes and emotions toward the content of the information as well as the other conversation participant(s). In (10), for example, wakannai kedo ‘(I) don’t know . . . ’ shows K’s level of certainty about what he’s just said regarding people who become homeless by choice, and at the same time, since this was produced after M countered his view in line 4 (‘there are people who are not like that’), it might be understood to show his deliberate effort to negotiate his stance with other speakers and involve them to share their opinions and/or feelings.
This intersubjective nature of nai-expressions is also observed in the nai-expressions occurring without the clause-ending kedo. In the next example, a mother (M), father (P) and two daughters (R and K) are eating dinner in the living room. Prior to this segment, K (younger sister) talked about her friend who unintentionally made a fool of herself by reading one of the ramen noodle menu items wrong, and the worker who was taking her order didn’t know what to do, probably because she or he couldn’t tell what the customer wanted. The story made everyone burst out in laughter.
(11) 1 R: . . . <@ wara[tcha=u @> laugh-end up 2 P: @@@ 3 M: fu=[2 == 2]= [3 =n 3]. huh 4 H: [2 aji da na 2]. taste COP FP 5 K: [3 ja= atashi-wa 3] ka nanka well I-WA Q something itte it[4 chau no 4]. say:STAT say-end up FP 6 R: [4 @@@@ 4]@@@ 7 P: @@[5 @@ 5] [6 @@@ 6]@@@ 8 K: [5 (H) 5] [6 shinjirannai 6]. believe:POT:NEG 9 R: [6 hazukashii= 6]. embarrassing ---------------------------------------------------------------------------- R: can’t help laughing LOL P: LOL M: I see H: (it’s) tasty K: (she) said, well then, I (will have) or something R: LOL P: LOL K: LOL R: (it)’s embarrassing
The segment is filled with laughter and overlaps. A nai-expression appears in K’s utterance in line 7, which is overlapped with P’s laughter. As in the case of wakannai in (10), shinjirannai ‘can’t believe; unbelievable’ (the negative potential form of Cognition shinjiru ‘to believe’) is the phonologically reduced form of shinjirarenai and occurs without overtly expressed nouns phrases. Note that K’s shinjirannai appears at the time when they seem to be wrapping up, and is produced in lower amplitude and intensity than the preceding utterances; after this, a topic shift occurs, and a speaker starts to talk about a different ramen noodle shop near a university. According to Yamada (2003), a negative in a narrative final position, like example (11), serves as ‘a summarizing comment on a narrative, so that no follow-up is required’ (p. 338). In the same line, K’s shinjirannai is considered to be what Yamada referred to as ‘reflective evaluation’ produced as ‘the narrator’s later reflection on the past events (the narrated events)’ (p. 227). It seems to suggest that shinjirannai does not just express K’s inability to believe this story, but also encodes her emotion and attitude, specifically her evaluative stance – surprise or disbelief – toward this laughable, yet embarrassing experience her friend had at the ramen noodle shop.
It should be mentioned that the affirmative counterpart of shinjirarenai may be used in a question, as in Shinjirareru? ‘Can (you) believe (it)?’ But an informal examination of Internet data suggests that the affirmative counterpart more likely occurs as attributive as in shinjiarareru hito ‘a person (I) can trust’ or shinjirareru koto ‘things that (I) can believe’. It seems to show that shinjirannai is not simply a negated version of shinjirareru ‘can believe’ but can also be viewed as a lexicalized expression that delivers a speaker’s evaluative stance.
The intersubjective nature of nai-expressions is also demonstrated in the following example. Excerpt (12) instantiates how the same expression can be used in a different way and conveys a different pragmatic meaning over the course of a conversation. Two ESL (English as a Second Language) students, T and K, are talking about various topics, such as their ESL classes and classmates in a university in the United States. Here they discuss how horrible English education in Japan is.
(12) 1 T: . . . datte sa=, because FP 2 . . . chuugakkoon-toki-no-eigo-no-sensei toka ne, junior high school-time-of-English-of-teacher TOKA FP 3 K: . . . shaberenai yo. speak:POT:NEG FP 4 T: . . . shabe- -- speak- -- 5 zettai shaberenai ko[to] nai? absolutely speak:POT:NEG thing exist:NEG 6 K: [un] . . . shaberenai. yeah speak:POT:NEG 7 T: . . . uchira-no-hoo-ga, us-of-way-GA 8 mada ima de[mo, still now even 9 .. sha]bereru yone. speak:POT FP 10 K: [un]. yeah 11 .. zettai shaberenai. absolutely speak:POT:NEG 12 T: un. yeah ---------------------------------------------------------------------------------------- T: ‘cuz people like (our) junior high English teachers K: T: speak- -- K: yeah, (they) T: Even now we can speak better (than they can), can’t they? K: yeah.
Following speaker T’s attempt to give a reason why English education is hopeless in Japan, the first instance of a nai-expression shaberenai ‘(our junior high teachers) can’t speak (English)’ occurs in K’s utterance in line 3; with a final particle of assertion, yo, it sounds rather bitter and disgruntled in tone and is not just stating her junior high school English teachers’ inability to speak English but also encoding a feeling of disapproval or possibly disgust toward her previous English teachers. K’s shaberenai facilitates the successive uses of the same expression. In fact, after its initial occurrence, shaberenai is repeated by T in line 5 and K in lines 6 and 11.
Immediately after K’s shaberenai in line 3, in line 4, T appears to start repeating the same expression but breaks off as in shabe- -- (‘speak- --’); she reformulates quickly and says, zettai shaberenai koto nai? (‘no way (they) can speak (English), don’t you think?’ lit. ‘is it not the case that (they) can’t speak (English) definitely?’) in line 5. This was produced with a slightly rising pitch to seek confirmation. T’s repetition with variation (i.e. the co-occurrence of zettai ‘absolutely’) demonstrates her rapport and reinforcement of K’s evaluative stance. K acknowledges T’s contribution and repeats shaberenai in line 6. T, then, seeks an agreement from K that their current English abilities are much better than their junior high teachers in lines 7 through 9. 3 Even before T’s turn ends, K gives her acknowledgment in line 10 and repeats the same expression with variation, zettai shaberenai ‘no way (they) can speak’ with no pauses or hesitations in line 11.
According to Tannen (2007 [1989]) ‘each time the utterance is repeated, the theme of the story or interchange is developed, slightly changed in meaning as well as form’ (p. 68). Furthermore, Svennevig (2003) argues that repletion ‘marks a strengthened affective commitment by the speaker and may thus be considered an upgrader or intensifier’ (p. 286). All those repeated utterances in (12) demonstrate that a speaker doesn’t simply repeat another conversation participant’s utterance because he or she meant the same thing; rather he or she uses repetition to elaborate on his or her opinion, create rapport, and facilitate his or her contribution to the ongoing discussion. It also bonds ‘participants to the discourse and to each other, linking individual speakers in a conversation and in relationships’ (Tannen, 2007[1989]: 61). In (12), by repeating K’s utterance, T shows an agreement and ratification of K’s opinion; by repeating her own word, K reinforces her own evaluative stance; this collaboration facilitates to build rapport and keep the conversation coherent. Viewed from an interactional perspective, the two speakers’ successive uses of shaberenai can be understood to serve as both an involvement strategy (e.g. Tannen, 2007 [1989]) as well as a cohesive device (e.g. Halliday and Hassan, 1976; Johnstone, 1987) to link new utterances/ideas to earlier discourse. That is, a sequence of shaberenai enables the speakers to build rapport, reinforce their evaluative stance and connect their feelings/attitudes with each other.
This section highlighted the pragmatically motivated uses of nai-expressions. The examples presented here indicate that nai-expressions are not just the reversals of the corresponding affirmatives but expressions that encode a speaker’s intersubjectivity.
Conclusion
An examination of nai-expressions in naturally occurring conversation has revealed fascinating tendencies regarding the negative suffix nai. Based on overall frequencies and distributional patterns, we found that the majority of them are non-activity predicate types such as Potential and Cognition. The results also support a well-established claim with respect to the rarity of overtly expressed arguments or core arguments in Japanese conversation (e.g. Clancy, 1980; Hinds, 1982; Maynard, 1989; Ono and Sadler, 2001; Ono and Thompson, 1997, 2017; Sadler, 2007); there were only 16 tokens of nai-expressions occurring with two overtly expressed core arguments in the data, as opposed to 57.4% of nai-expressions with no lexically expressed arguments. A closer examination tells us that Potential nai is more likely to occur without arguments (63.2%) than any other predicate types.
The patterns which emerged from the data also indicate that nai-expressions exhibit formulaic tendencies (Calude, 2009). We have observed that one-third of nai-expressions in the data co-occur with negative-anticipating forms such as zenzen ‘at all’ and amari ‘not very’ (Ono and Thompson, 2017). Yet, structural fixedness we have found are more apt to some specific predicate types. For example, with regard to ‘phonological cohesiveness’ (Hickey, 1993) or phonological fluency (Calude, 2009), 37.5% of nai-expressions are phonologically reduced, and all instances of wakaranai ‘don’t know/understand’ and Stative nai (including both ongoing and resultant actions consisting of a verb gerund and ~te(i)nai, as in kiiteinai ‘am not listening/asking or have not heard/asked’) in the data appear in their contracted forms (wakannai, kiitenai, respectively). We have also observed speakers’ preference to use Cognition nai occurring with the clause-ending kedo, with kedo functioning as a mitigation marker (e.g. Geyer, 2007a, 2007b; Maynard, 1989; Mori, 1999; cf. Jorden and Noda, 1987) as in wakannai kedo ‘(I) don’t know but . . .’ All in all, the data demonstrate varying degrees of formulaicity.
The current study has also demonstrated that what has been considered to be ‘grammatical’ actually turns out to be more pragmatically oriented. On the surface, nai-expressions are simply ‘a reversal of the truth value of the corresponding affirmatives’ (McGloin, 1986: 99). As has been claimed (Hudson, 2008; McGloin, 1986; Takano, 2008; Uehara and Fukushima, 2008; Yamada, 2000, 2003), however, speakers use Japanese negatives not just to express their emotional and personal stance on a particular story or event but also to help create an interpersonal space with other conversation participants and to help them be involved with the story or event (Yamada, 2003). In that regard, nai-expressions are examples of what Maynard (1993) refers to as ‘information-sharing’ and ‘subtextual emotion-sharing’ (p. 4) as well as emotive communication (Suzuki, 2006).
The patterns emerged from the data are quite similar to those in Ono and Thompson’s (2009) study on Japanese adjectives. As they pointed out that ‘we can align these nai-forms on a continuum to illustrate that negated predicates with –nai are becoming re-analyzed as i-adjectives’ (p. 138), some of the nai-expressions found in the current data may also be re-analyzed as adjectives. Particularly those that are phonologically reduced and occur with no lexically expressed arguments (e.g. wakannai ‘don’t know’ and shinjirannai ‘can’t believe; unbelievable’) appear to be quite similar to a formulaic type tamannai ‘unbearable’, which is a phonologically reduced form of tamaranai – consisting of the verb root tamar- of tamaru (‘to endure’) and the negative suffix -anai – and is found in most standard dictionaries. It seems to suggest that the boundary between adjectives and verbs may not be fixed but very fluid at least in terms of nai-expressions in conversation. Future studies can explore this further from a perspective of synchronic language change and variation.
The method used here to examine nai-expressions is a manual analysis of conversations. As argued in Kerz and Haas (2009), it serves the purpose of being more inclusive so that we would not overlook less formulaic types, such as partially fixed patterns and unusual patterns (p. 103). Having completed a preliminary examination, a next step would be to conduct a more rigorous quantitative and statistical analysis of large corpora of everyday conversation. Given how social media has become an integral part of our daily life, it would also be desirable to investigate how these nai-expressions along with other types of emotive communication are used.
Although the current study focused on one particular type of expression, it is the author’s hope that it has demonstrated the formulaic and intersubjective nature of language as well as the importance of looking at how speakers actually use a certain type of expression in everyday conversation.
Footnotes
Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.