Self-mockery in Japanese (1).

Suzuki, Satoko

Abstract

Self-mockery is defined as a linguistic act in which the speaker makes a statement and then denies, invalidates, or expresses his/her nonserious attitude toward the content of the utterance. It is a metalinguistic commentary that is distinct from sarcasm/irony in that (i) it does not involve contempt for others, (ii) it does not echo somebody else's utterance, (iii) double-voicedness associated with sarcasm/irony is expressed in two phrases in self-mockery, and (iv) the speaker may actually believe in the content of the utterance s/he invalidates. It is also distinct from joking in that the invalidating signal is explicitly expressed and physically separated from the invalidated utterance.

In Japanese expressions several factors contribute to the meaning of self-mockery. First, by activating the image of a double voice associated with quotation, the speaker effectively dissociates him/herself from the content of the utterance and evokes the idea of pretense. Second, the lexical meaning that refers to lack of specification implies lack of the speaker's willingness for commitment. Third, the use of a pause or other framing device signals that the speaker is dissociated from the content of the utterance. Fourth, the nonfinite verbal form makes the utterance less imposing.

1. Introduction

This paper deals with the linguistic act of self-mockery. (2) The term is taken from Maynard (1996), which provides a brief but insightful section on the phenomenon. (3) In self-mockery the speaker makes a statement and then denies, invalidates, or expresses his/her nonserious attitude toward the content of the utterance by adding a certain phrase. An example is given below. (4)

(1) A1: E, ja moo ima wa shuushoku katsudoo o oete.
 oh then already now TP employment activity OB finishing
 `Oh, then you are now done with the job search.'
 B1: Un.
 yeah
 `Yeah.'
 A2: Soredee?
 and
 `And?'
 B2: Doo shiyoo?
 how shall-do
 `What should I do?'
 A3: Doo shite-ru?
 how are-doing
 `What are you doing?'
 B3: Benkyoo, toka itte [laughter] ima-sara. Iya
 study something-like saying this-late well
 ryokoo shiyoo kanaa to mo omou kedo.
 travel shall-do I-wonder QP also think but
 `Studying, just kidding [laughter], at this belated time.
 Well, I'm also thinking of traveling.'

(1) is a conversation between two senior university students. Asked about what she is doing now that she has completed a job search, B says benkyoo `studying' and then quickly tries to buffer the seriousness of the answer by adding the phrase toka itte, which is a combination of toka (`something like') and itte (the nonfinite form of the verb iu `say'). The feeling this phrase conveys is similar to the English phrase just kidding. By attaching phrases such as just kidding or toka itte, the speaker mocks his/her own utterance. Other phrases in Japanese that produce similar effects (nanchatte, nante, and tte) will be discussed later.

Why should the phenomenon of self-mockery be studied? In his recent book Haiman (1998) discusses the notion of the speaker as a divided self. He observes that this notion, or more specifically "the speaker's self-conscious alienation from the actual referential content of his or her message" (Haiman 1998: 10), is the key factor in understanding a wide range of linguistic phenomena from sarcasm and politeness to ritual language. As will be argued later in this paper, the notion of the speaker as a divided self also plays a central role in self-mockery.

Among the phenomena that involve the notion of the speaker as a divided self, sarcasm/irony (5) has received considerable attention (e.g. Grice 1975, 1978; Sperber and Wilson 1981; Clark and Gerrig 1984; Wilson and Sperber 1992; Haiman 1989, 1990, 1998). Self-mockery is very similar to sarcasm/irony. They both involve the speaker's mockery of the message that is expressed in an utterance. In both linguistic acts the speaker is psychologically detached from the content of the message and suggests that s/he does not really mean what s/he says. As the study of sarcasm/irony helps us understand various aspects of human communication such as the notion of multivoicedness, the study of self-mockery should also be enlightening. Yet, there has been very little written about self-mockery in the linguistic literature. (6) This paper is an attempt to make a contribution to the study of this topic.

The data in Japanese come from various sources such as taped conversations, commercial video programs, television programs, web pages, books, and magazines. The sources are mentioned at the end of each example and in the Data references, following the References. No mention of the source at the end of an example means that it is taken from a collection of taped conversations. The collection includes conversations of twelve Japanese undergraduate students who were enrolled in a Japanese university at the time of the recording. (7) English examples are taken from the Santa Barbara Corpus of Spoken American English, Part 1 (Santa Barbara 2000), as well as a collection of personal correspondence.

The outline of this paper is as follows. Section 2 introduces other phrases that express self-mockery and discusses a basic form of self-mockery. Section 3 characterizes self-mockery in contrast to sarcasm/irony and joking, two linguistic acts that are similar to self-mockery in some respects. The section also defines self-mockery as a metalinguistic phenomenon. Section 4 discusses factors that contribute to the function of expressing self-mockery in Japanese. At the end of the section the special status of the expression nanchatte is discussed. Section 5 discusses motivations for self-mockery and implications for future studies.

2. Expressions of self-mockery

In addition to the phrase toka itte discussed in the Introduction, nanchatte, the contracted form of nante (`something like') itte (the nonfinite form of the verb iu `say') shimatte (the nonfinite form of the auxiliary verb shimau `do -- completely, end up --ing') may also be used for expressing self-mockery. The first vowel is often lengthened (naanchatte) perhaps to emphasize the nonserious nature of the utterance. I even found in the spoken data nahahanchatte, with the constructed sound of laughter (haha) inserted in place of the lengthened vowel. The last vowel may also be lengthened (nanchattee) (8) as in the following example, which is taken from a web page devoted to personal writings about a singer named Miyuki Nakajima.

(2) Takusan kaita kedo jitsu wa shinpuru. Tada "nakajima
 much wrote but actuality TP simple just Nakajima
 miyuki" to iu sonzai ga ureshii dake kamo nee. Tooi
 Miyuki QP say presence SB happy only may FP far
 mukashi kara, watashi wa zutto konna hito o
 past from I TP throughout this-kind person OB
 sagashite-iru no kamoshirenai. Nanchattee.
 is-looking NM may something-like-saying-ending-up
 `I wrote a lot, but actually it is simple. I may be just happy about
 the presence of "Miyuki Nakajima." I may have been looking for
 a person like this from a long long time ago. Just kidding (Miyuki
 joo, October 20, 2000).

Perhaps embarrassed about the earnestness of the statement "I may have been looking for a person like this from a long long time ago," the writer adds nanchattee as if trying to erase the awkwardness.

Nanchatte also has a shortened form, na(a)nte, which is often followed by an interactional particle ne or na.

(3) Iroiro osewa ni narimashita ... nante ne.
 various assistance AV became something-like FP
 `I am indebted to you for your kind assistance on various matters
 ... just kidding' (Saigo no koi, Part 3).

(3) is taken from a TV drama in which a young woman thanks a medical student who took care of her brother as she is leaving the hospital. Since she has become good friends with the medical student, she is embarrassed about the formality of her statement. The addition of nante after the statement is her attempt to mock it. A little later, she restates her gratitude.

(4) Iroiro arigatoo ... honto hi.
 various thank-you real AV
 `Thank you for various things ... seriously' (Saigo no koi, Part 3).

By attaching the adverbial phrase honto ni `seriously', the speaker makes sure that this statement is interpreted as a sincere expression of her gratitude. In this way, the use of nante is contrasted with the phrase that represents sincerity.

Another expression that may be used to express self-mockery is tte. Tte is related to a quotative particle to. It is sometimes considered to be an informal variant of to and other times a contracted form of the combination of to and itte, the nonfinite form of the verb iu `saying.' Much has been written about various functions of the utterance-final tte (e.g. Morishige 1954; Miura 1974; Itani 1994; Horiguchi 1995; Okamoto 1995; Maynard 1996, 1997; Suzuki 1996, 1998a), but a self-mocking function has been overlooked. This is probably because the expression of self-mockery is very subtle in the use of tte, as shown in (5).

(5) A: Kubininatta hito ga iru yoo na nyuansu de ...
 got:fired person SB exit appear CP nuance with
 `(He was talking) with the nuance that there was somebody
 who got fired.'
 B: Sore wa anata desu tte.
 that TP you CP QP
 `That's you' [said in a playful tone].

If speaker B of (5) had said sore wa anata desu `that's you (who got fired)' instead of sore wa anata desu tte, he would have come across as being serious and potentially hurt the addressee's feelings. The addition of tte at the end of the sentence provides a playful, kidding tone and distances the speaker from the previous statement.

In all of the expressions of self-mockery discussed above, an utterance is made and then a phrase is added to deny, invalidate, or express the speaker's nonserious attitude toward the content of the utterance. The basic common structure is as follows:

(6) UTTERANCE + self-mocking expression

I will use the term UTTERANCE to refer to the part of the discourse that is mocked/invalidated. It is often the case (although not always) that the UTTERANCE and the self-mocking expression are separated physically. In spoken data I found a pause or the conversational partner's utterance separating the two. In written data a comma, a period, or the symbol that denotes a considerable pause (...) is often found between the two. A new, separate line may also be used for the self-mocking phrase. Further, because of the nature of self-mockery, sounds of laughter (or the Chinese character for laughter in parenthesis in the case of written discourse) often accompany the expression.

3. Characteristics of self-mockery

3.1. Self-mockery and sarcasm/irony

Although there are similarities between self-mockery and sarcasm/irony, as mentioned in the Introduction, the two are distinct from each other in several respects. First, self-mockery does not involve hostility or contempt for others. Haiman (1998: 25) repeats the conventional view that in using sarcasm/irony "the speaker expresses hostility or ridicule of another speaker." The "other speaker" could be the present interlocutor, the third person who is absent, or the opinions or conventions of the society. Sperber and Wilson (1981: 314) observe that the target of this verbal aggression is the originator, real or imagined, of the utterance that is echoed in the sarcastic comment. Such hostility for others is not present in the expression of self-mockery. If there is a "target," it is not others, but the speaker him/herself that is ridiculed.

The above-mentioned difference between self-mockery and sarcasm/irony is related to the more fundamental difference between the two. As argued by Sperber and Wilson (1981) and Wilson and Sperber (1992), many of the cases of sarcasm/irony involve echoic mention/ interpretation. (9,10) In being sarcastic, the speaker "echoes a thought she attributes to someone else" (Wilson and Sperber 1992: 60; emphasis added). In self-mockery, nobody's thought or utterance is echoed. Instead, the speaker's own immediately preceding utterance is deflated.

Another difference between self-mockery and sarcasm/irony is its form. A classic example of sarcasm/irony is (7), uttered on a dark rainy day:

(7) What lovely weather!

This is a double-voice discourse. The speaker's self is divided. One delivers its ostensible message and another its metamessage, which invalidates the ostensible message. Both voices are usually delivered in a single utterance. In contrast, the two voices in self-mockery are always physically separated. The UTTERANCE is given and then another voice invalidates its effect by appending a self-mocking expression such as nanchatte and just kidding. More will be said about this separation in section 3.2.

The fourth difference between self-mockery and sarcasm/irony concerns the speaker's commitment. In uttering a sarcastic/ironic utterance such as (7), the speaker believes something different from what is ostensibly expressed in (7), presumably something like "What nasty weather!" The same could be said in some cases of self-mockery. For example, in (8) what is followed by the self-mocking phrase naanchatte is presented as something in which the writer of the web page does not really believe as opposed to the immediately following discourse in which the writer's sincere opinion is stated.

(8) (In responding to a comment on why people who previously were
 frequent participants in the electronic chat are not participating
 these days)
 Akita to iu yori wa, "konomama tsuzukete-iru to
 got-bored QP say than TP as-is is-continuing if
 seikatsu ga hatansuru" to osore o idaite, sessei ni
 life SB get-ruined QP fear OB hold temperance AV
 tsutome-hajimeta to iu koto de wa?
 began-effort QP say thing CP TP
 Naanchatte.
 something-like-saying-ending-up
 (SPACE)
 Joohooryoo toshite kanari tarite-shimatta node,
 information as considerably ended-up-enough so
 mohaya atarashii joohoo o eru hitsuyoo ga nai,
 any-more new information OB obtain need SB not-exist
 to iu no ga hontoo no tokoro da to omoimasu.
 QP say NM SB actual GN case CP QP think
 `Rather than getting bored, I wonder if they fear that their lives
 will be ruined if they continue [to participate in the chat] in this
 way and started to make efforts to be temperate?
 Just kidding.
 (SPACE)
 Actually I think it is the case that they have received enough
 information, so they do not need to acquire new information any
 more' (Archives, October 20, 2000).

However, self-mockery could be used when the speaker is truly committed to what the UTTERANCE expresses. In such cases self-mockery is used as a kind of face-saving device. An example of such use from Maynard (1996: 221), which was originally taken from a comic book, is given below. The transcription is slightly modified.

(9) Hoshina: Kedo, omee no me de wakaru.
 but you GN eye with understand
 `But I understand you by (looking into) your eyes.
 Sari: Kaa (blushes in embarrassment)
 `Ohh ...' (embarrassed).
 Hoshina: Naanchatte na.
 something-like-saying-ending-up FP
 `Umm, just kidding' (Maynard 1996: 221).

Maynard (1996: 122) explains the background of (9). Hoshina and Sari are beginning to fall in love. In this scene Hoshina catches himself expressing his love. He feels embarrassed and vulnerable. In order "to avoid being taken too seriously and to circumvent possible rejection," he qualifies his own statement by adding nanchatte. In doing so, "he retains the possibility of backing out from his confession of love without losing face." In this case it is likely that the speaker is committed to what he says in the UTTERANCE, his confession of love, but for various reasons (e.g. fear of rejection) he decides to make a mockery of it. His intention is not necessarily to communicate something different from what the UTTERANCE says, but to mitigate its effect.

When self-mockery is used in this manner (i.e. as a self-protective strategy), the addressee is not likely to interpret the UTTERANCE as invalidated despite the invalidating force of self-mocking expressions. This can be seen in (10).

(10) A1: Ima hikooki tonnai-to moo dame na n da yo
 now air-plane if-not-take already no-good CP NM CP FP
 nee.
 FP
 `If I don't reserve a plane ticket now, it'd be too late, you
 know.'
 B1: Un un.
 uh-huh uh-huh
 `Uh-huh.'
 A2: Hitori de iku-na tte yappa mata oya ga
 one-person by not-go QP as-expected also parents SB
 urusai koto iidashichattee.
 annoying thing start-saying-ending-up
 `As expected, my parents also started saying I shouldn't
 go alone.'
 B2: Uu uun.
 uh-huh uh-huh
 `Uh-huh.'
 A3: Ikanai, mitchan? toka itte.
 not-go [name] something-like saying
 `Won't you go with me, Mitchan? Just kidding.'
 B3: Sasuga ni kono natsu wa benkyooshinai-to
 as-expected AV this summer TP if-not-study
 dame deshoo.
 no-good probably
 `This summer, as may be expected, I will have to study.'

Speaker A is thinking about going to Russia in the summer and is telling speaker B her plan. In line A3, A asks B to join her in her travel. Perhaps in anticipation of R's rejection, A adds toka itte to show she is not really serious. She may actually be completely serious in her invitation to B, much as Hoshina is likely serious in his confession of love in (9), but she chooses to present the invitation in a nonserious manner. Despite this nullification of the effect of invitation in A's utterance, B interprets her utterance as an invitation and answers accordingly (i.e. declines it).

3.2. Self-mockery and joking

Another linguistic phenomenon that is similar to self-mockery is joking. (11) Fonagy (1982: 33) characterizes joking as "a verbal act immediately followed by its invalidation: `I didn't mean it, I'm only joking.'" From this description, it seems that self-mockery and joking are identical verbal acts. However, Fonagy goes on to say that in its basic form joking contains both phrases (i.e. the message and its invalidation) in a single speech act. (12) In this basic form "the assertion invalidates itself through exaggeration. Absurdity (13) functions as an implicit joke mark, a signal that implies the assertion is not to be treated seriously." The following example illustrates this point.

(11) (Said to a child who has been eating spinach) It's a good thing if one first puts one's hands into the spinach and then wipes them on Daddy's book. Like this (making the appropriate gesture). It would make a nice illustration (Fonagy 1982: 32).

To an adult, the above utterance is apparently a joke since the idea of wiping spinach on a book making a nice illustration is absurd. Absurdity of the utterance itself implicitly invalidates the ostensible message.

In contrast, as already mentioned in the comparison with sarcasm/irony, the two phrases (i.e. the message and its invalidation) are physically separate in self-mockery. In addition, what Fonagy calls the "joke mark" (i.e. the invalidating signal) is explicit rather than implicit. Expressions such as toka itte and just kidding are uttered following the UTTERANCE rather than being implied. (14)

As mentioned earlier, the separation between the message and its invalidation is often overtly marked with a pause or other types of device between the UTTERANCE and a self-mocking expression such as nanchatte. In some cases the content in the UTTERANCE may be given in earnest, but then the invalidation is provided as an afterthought. A case in point is (9), which is repeated here.

(9) Hoshina: Kedo, omee no me de wakaru.
 but you GN eye with understand
 `But I understand you by (looking into) your eyes.
 Sari: Kaa (blushes in embarrassment)
 `Ohh ...' (embarrassed).
 Hoshina: Naanchatte na.
 something-like-saying-ending-up FP
 `Umm, just kidding' (Maynard 1996: 221).

In (9) Hoshina may have given his first statement, his confession of love, in earnest. His intention may have been to communicate seriously what he said. However, after seeing Sari's reaction, which is the expression of embarrassment, he may have added his second utterance Nanchatte na `Umm, just kidding' to mitigate the effect of his first utterance.

3.3. Self-mockery as a metalinguistic phenomenon

From the discussion in the last two subsections, it is clear that the self-mocking phrase (toka itte, nanchatte, just kidding, etc.) is a metalinguistic remark (i.e. language about language). Sometimes the nonserious nature of the UTTERANCE is felt by the speaker to be rather obvious, given the encoded meaning and the contextual factors available to the addressee. Yet to be on the safe side, the speaker chooses to provide the addressee with a metalinguistic commentary on the nature of the message. In other contexts the speaker fears that the addressee will miss the nonserious nature of the UTTERANCE unless a specific instruction by means of a self-mocking phrase is given. In yet other situations the explicit invalidation is felt necessary even though the UTTERANCE was in fact originally meant to be taken seriously.

4. Self-mockery in Japanese

As already mentioned, the English phrase that may be equivalent to the Japanese phrases is (I'm) just kidding. (15) Similar to the Japanese expressions, just kidding is appended to an immediately preceding utterance. The following example illustrates this.

(12) And we turned around and our car was gone ... Just kidding (Santa Barbara 2000: Part 1, 0013).

Just kidding is sometimes combined with no as in (13), which is taken from a personal electronic-mail correspondence.

(13) (In responding to a message "I was wondering if I could ask you to read my manuscripts in the future and give me comments and/or suggestions.")

Depends how productive you are! No I'm just kidding, [...]

Here the effect of just kidding is reinforced with the presence of the invalidating no.

Similar to various Japanese examples shown in section 2, just kidding may be used in situations that the speaker finds embarrassing. I had a student in my Japanese linguistics class who used just kidding in almost all of her statements made in class. She would state her opinion, and then say just kidding after it. Her facial expressions and the tone of her voice indicated that she was not very confident of her opinions and that she was embarrassed. Appending just kidding to her remark seemed to be her attempt to communicate to her classmates, "Do not take me seriously."

In this way the English expression just kidding and the Japanese self-mocking phrases are very similar. They both deny, invalidate, or express a nonserious attitude toward the content of the utterance to which they are attached. The difference between the English and Japanese expressions is that in English the functions are indicated explicitly. With just kidding, the speaker's nonserious attitude is explicitly represented in the lexical item kid. When it is used in combination with no as in (13), the function of denial and invalidation is spelled out.

In contrast, the equivalent Japanese expressions do not have elements in them that explicitly spell out their functions. As mentioned earlier, toka itte is the combination of `something like' and the nonfinite form of the verb iu `say'. Nanchatte is the contracted form of nante (`something like') itte (`say') shimatte (the nonfinite form of the auxiliary verb shimau `do -- completely, end up --ing). Tte is either a variant of the quotative particle to or the reduced form of the combination of to and itte. None of these lexical items specifically expresses a nonserious attitude or invalidation. Why do these signify self-mockery, then? I would like to argue that the various factors discussed below (self-quotation, the lexical meaning of toka and nante, and the nonfinite form of the verb) contribute to the function. When only one or two factors are involved, the meaning of self-mockery may be rather weak, as in the case of tte. When more or all of the factors are present, the speaker's intent of self-mockery is more strongly expressed. (16)

4.1. Self-quotation

All of the self-mocking phrases in Japanese contain either itte (the nonfinite form of the verb iu `say'), a quotative particle, or both the verb and a particle. Toka itte is composed of toka, which may be used to mark a quote, and itte. Nanchatte contains both nante, which may be used as a quotative particle, and itte. Tte is either a variant of the quotative particle to or a combination of to and itte. Quotative particles are known to express that the act of saying is conducted even when the verb of saying is not present. This means that all of the self-mocking expressions in Japanese make reference to the act of saying. The subject of saying is understood to be the speaker although it is not explicitly expressed. Therefore, the basic structure of self-mockery in Japanese expressed in (6) can be roughly translated as follows:

(14) UTTERANCE, (something like that), I said/am saying. (17)

Why is the speaker compelled to express "I said/am saying"? The fact that the speaker is saying what s/he is saying is clear from the act of saying itself. There clearly has to be significance in the speaker's explicitly referring to this act of saying.

Maynard (1996) treats this as a case of self-quotation. Quotation is most typically used when the speaker repeats what somebody else said. In self-quotation, in contrast, an utterance is made, and the speaker specifically expresses that s/he said/is saying it.

Quotation is a place where the idea of the speaker as a divided self is most vividly expressed. As Volosinov (1986: 116) notes, quotation is "regarded by the speaker as an utterance belonging to someone else, an utterance that was originally totally independent, complete in its construction, and lying outside the given context." Haiman (1998: 26) observes that the metamessage in quotation is paraphrasable as "this is not really me: I'm just playing a role, mouthing someone else's words."

Maynard's work on self-quotation (1996) is significant because she points out that this idea of multivoicedness in quotation is applicable to self-quotation as well. (18) In saying "I am saying [X]," Maynard (1996: 212) notes that "Although in physical terms only one person stands behind the statement [...], one can say that there exists a character who utters [X] and a speaker who frames and controls the character's voice."

Why is self-quotation used in the expression of self-mockery in Japanese? In self-mockery the speaker is divided into two. One part of the speaker delivers the UTTERANCE. Another part of the speaker invalidates and expresses his/her nonserious attitude toward the UTTERANCE. Multivoicedness associated with (self-) quotation is well-suited for this phenomenon. By saying "I am saying (something like) the UTTERANCE" the speaker activates the image of a double voice. Put differently, the speaker presents the UTTERANCE "as if it represents someone else's voice" (Maynard 1996: 224). The speaker effectively dissociates him/herself from the UTTERANCE and evokes the idea of pretense/fakery ("I'm just playing a role"), which presents the speaker's denial of and nonserious attitude toward the UTTERANCE.

4.2. Lack of specification

In the last section it was argued that by explicitly expressing that the speaker said/is saying the UTTERANCE, s/he communicates that there is a double voice involved; one delivers the message in the UTTERANCE; another is dissociated from it. This idea of a double voice is extremely important in the discussion of self-mockery. However, we should note that the expression tte, which involves self-quotation, is not always associated with self-mockery, whereas toka itte and nanchatte (and its shortened form nante) are more readily identified with the self-mocking function. In addition to the reference to self-quotation, the latter expressions contain a particle with a similar lexical meaning. Toka in toka itte is considered to be a combination of either to `and' and ka (the question marker) or to and ka `or' (Shinmura 1991 [1955]: 1829) and may be translated as `something like'. (19) Nante in nanchatte is considered to have developed from the combination of nado (which is described to have derived from the combination of nani `what' [Kindaichi et al. 1993: 274]) and to `and' (20) and may also be translated as `something like'. (21) In this section I would like to argue that this lexical meaning (`something like') of the particles toka and nante contributes to the function of self-mockery along with the image of a double voice associated with self-quotation.

In self-mockery the speaker invalidates the effect of the UTTERANCE. This indicates that the speaker has a low opinion of the message expressed in the UTTERANCE. In cases in which the speaker uses self-mockery as a self-protective strategy, s/he may not actually have a negative attitude toward the message in the UTTERANCE but presents the discourse as if s/he did. Whether it is serious or fake, the sense of negative evaluation is present.

This sense often accompanies toka and nante. In addition to forming part of the self-mocking phrases, these particles have diverse functions such as marking of a topic and marking of a quote. In various uses the speaker's negative attitude can be detected, as illustrated in the following examples.

(15) Josei-tachi mo katsute wa "san koo" toka itte,
 women also before TP "three highs" something-like saying
 jibun de jibun no kubi o shimete-ita kedo ...
 self by self GN neck OB were-strangling but
 `Women also used to say (as criteria for a husband) something
 like "Three Highs" and were putting themselves into difficult
 positions, but ...' (More 1996:121).

(16) Zuibun fukoohei na hanashi da to omoimasen ka.
 considerably unfair CP story CP QP not-think FP
 Juunana, hachi-sai no koro no, tatta isshu no nooryoku
 seventeen eighteen GN age GN only one-kind GN ability
 ga sonohito no jinsei o sayuusuru nante.
 SB that-person GN life OB determine something-like
 `Don't you think it is a considerably unfair state of affairs
 (which is something like) that only one kind of ability (that
 one has) when one is seventeen or eighteen determines one's
 life?' (Chiba 1985: 30).

(17) Kimi ga sonna ni shinkeishitsu ni, ofukuro mitai na
 you SB that-kind AV high-strung AV mother like CP
 taido de, sore wa nanbaime no kakuteru na no?
 attitude with that TP how-many-glass GN cocktail CP NM
 nante iu koto wa nai.
 something-like say thing TP not-exist
 `You do not have to be so high-strung and say with the mom-like
 attitude something like "How many cocktails have you had?"'
 (Kanai 1995: 99).

In (15) toka is used as a quotation marker. The toka-marked `Three Highs' refers to a phrase that describes three criteria for selecting a husband: high income, high education, and physical height. The phrase and the idea behind it were popular among young women in Japan at one time. The speaker of (15) is critical of such a trend that places importance on money, prestige, and appearance. From this background it is clear that the speaker feels contempt toward the phrase san koo `Three Highs'. The clause in (16), juunana, hachi-sai no toki no, tatta isshu no nooryoku ga sonohito no jinsei o sayuusuru `only one kind of ability (that one has) when one is seventeen or eighteen determines one's life' is marked with nante. From the comment regarding this, zuibun fukoohei na hanashi da to omoimasen ka `Don't you think it is a considerably unfair state of affairs?', we can tell that the speaker views the content of the clause negatively. (17) is a husband's remark to his wife. Other expressions in the same sentence such as shinkeishitsu ni `in a high-strung manner' and ofukuro mitai na taido de `with the mom-like attitude' make it clear that he does not appreciate his wife's question, which he marks with nante. In this way toka and nante often mark phrases toward which the speaker has a negative attitude.

The reader may wonder if it is really these particles that are responsible for communicating the speaker's negative attitude. After all, these particles have the lexical meaning, `something like', which contributes to the propositional meaning of the sentences. Actually, the contexts of these sentences are such that the lexical meaning, `something like', is not required. In (15) san koo `Three Highs' is the exact phrase that women used to use. Therefore, the speaker does not have to use toka `something like', which signifies lack of specification. The speaker could have used a regular quotative particle to. The writer of (16) is criticizing the Japanese social system, in which the result of a college entrance examination plays a central role in one's life. Therefore, the content of the nante-marked clause (that one's life is determined by the academic test-taking skills one has at the age of seventeen or eighteen) is exactly the target of her criticism. She does not have to use nante `something like', which indicates lack of specification. The nante-marked quote in (17) is presumably what the wife just said. The husband does not need to obscure the wording of the quote with nante.

Why do the speakers of (15), (16), and (17) use phrases with the lexical meaning `something like', then? Because the phrases are not needed propositionally, I would like to argue that they are present to convey the speaker's attitude, which in this case is negative. Why are these phrases associated with the implication of the speaker's negative attitude? I would like to propose that this is because they represent lack of specification.

Both X toka and X nante `something like X' indicate lack of specification. When the speaker is not able to provide specification, such as in recalling distant past incidents or talking about hypothetical situations, these phrases are propositionally meaningful. What is in question may be X, or may not be exactly X; the speaker cannot fully commit to the accuracy of X. When the circumstance is such that the speaker could say `X', as in the cases of (15), (16), and (17), but nonetheless elects to say `something like X', s/he is expressing his/her attitude toward X. By being nonspecific, s/he is signaling that s/he is not willing to commit herself/himself to X. It is natural to interpret this speaker's unwillingness to commit to X as the speaker's negative evaluation of X. (22)

Because of this aspect, nante and toka may be used to express the speaker's negative attitude. The association of nante and toka with the speaker's negative evaluation contributes to the meaning of self-mockery in the phrases that contain toka and nante, toka itte, and nanchatte, respectively. As discussed earlier, because these phrases include the verb of saying itte and thus allude to self-quotation, they suggest a double-voice discourse. In addition, because the particles within these phrases that mark the UTTERANCE imply the speaker's negative attitude, the addressee will know that one of the voices involved in the double-voiced discourse (i.e. the one that does not deliver the ostensible message expressed in the UTTERANCE) looks down upon the message.

4.3. Other factors

There are two other factors that are less prominent than the two factors already mentioned but are nonetheless relevant to the discussion of self-mockery. One is the pause or other type of device that physically separates the UTTERANCE and the self-mocking phrase. Adachi (1996) analyzes a quotative particle datte as a sarcasm marker in Japanese. He argues that the sarcastic meaning is emphasized if the speaker pauses between the quote and datte because a pause functions as a framing device, which signifies that "the quote is treated as an independent entity which is not incorporated in the speaker's `own' utterance" (Adachi 1996: 12). The quoter stands outside the frame, dissociated from the quote. His analysis of the pause as the framing device is useful in the analysis of self-mockery. The pause or another type of device that physically separates the UTTERANCE from the self-mocking phrase reinforces the speaker's (presentation of) detachment from the UTTERANCE that is already inherent in self-quotation. The speaker treats the UTTERANCE as if it were a separate, independent entity. This follows the general iconicity principle that linguistic distance reflects conceptual distance (Haiman 1985).

Another factor that may be relevant is the nonfinite form of the verb, or the te form, as it is often called in Japanese linguistics, (23) which is used at the end of these phrases. This nonfinite form is usually used to connect a string of utterances, as shown in (18).

(18) A: Watashi kizuiteta kedo sa, sonna koto iwaretara kaita
 I noticed but FP that-kind thing if-be-told wrote
 hito no tachiba ga nai jan toka
 person GN position SB not-exist TG something-like
 omotte,
 thinking
 `I noticed (the mistake on the blackboard), but I thought that
 the person who wrote it would lose face if somebody pointed
 that out.'
 B: Un un
 uh huh
 `Uh huh'
 [part of discourse omitted]
 A: Dakara / nanka ii ya to omotte, nani mo
 so somehow okay FP QP thinking what even
 iwanaideita no.
 did-not-say NM
 `So, I thought, well, it's okay and I didn't say anything.'

In (18) the te form of omou `think', omotte, is used twice. The speaker (a college student) is talking about an incident in which she found a mistake in what a fellow student wrote on the blackboard but did not point the mistake out. The clauses that end with omotte provide reasons or background information on why she did not point out the mistake. Not all te-form-ending clauses have the same function. Some provide background information for another part of the discourse, as in (18). Some supply elaboration of something that is discussed elsewhere in the discourse, while others express contrastive relations. (24) A common feature among them is that they are linked to some other part of the discourse. In this usage the te form is considered to imply that the talk is to continue (Maynard 1989: 38).

When an utterance ends, the norm is to finish an utterance with the tensed form of a verb, which may or may not be followed by an interactional particle. However, in casual conversation it is quite common to end an utterance with the te form of a verb, as is the case in self-mocking remarks. In her analysis of Japanese conversation, Maynard (1989: 38) observes that the nonfinite form endings are used "primarily to soften the statement by leaving the propositional content with a feeling of incompleteness." This sense of incompleteness obviously comes from the fact that the nonfinite form is usually used to connect strings of discourse. The addressee expects the finite verb form to appear to conclude an utterance. When the utterance ends with the nonfinite form, the addressee feels that it is not complete.

Maynard (1989: 38) also says that this sense of incompleteness gives a less imposing feeling toward the recipient." (25) This association between the nonfinite form and nonimposition may be related to the discussion of self-mockery. Recall (9), in which the speaker confesses his love to the addressee. After the confession, he uses the self-mocking expression nanchatte as if to "lighten up" the atmosphere. The nonfinite form used may be interpreted as his attempt to be less imposing. The same interpretation could be given to (10), in which the speaker invites the addressee to take a trip with her. The inconclusive character of the nonfinite form is effective in such situations.

4.4. Special status of nanchatte

Before concluding, I would like to mention that one of the expressions discussed here, nanchatte, has a special status. It seems that nanchatte has been grammaticalized to be the marker of self-mockery, whereas other expressions have not.

While nanchatte seems always to be used in contexts in which the speaker's self-mockery is expressed, others may be used in contexts that do not involve self-mockery. For example, toka itte is used with its literal meaning `saying something like' in the following example.

(19) (After seeing Princess Masako's engagement ceremony on TV)
 A, konna hito na n da, toka itte.
 oh this-kind person CP NM CP something-like saying
 `I said something like, "Oh, she is this kind of person."'

In (19) the speaker is recounting the time when she saw the Japanese princess on television for the first time while she was living in England. She uses toka itte `(I) said something like' because she cannot remember exactly what she said at the time. In this example the connotation of self-mockery is absent. The lexical meaning of toka itte `saying something like' is used sincerely.

Toka itte can also be used to mark somebody else's utterance rather than one's own, as shown in (20). So can tte, as in (21).

(20) De, dakara boku wa oshie-hajimete kara shikuhakkushite, koko
 and so I TP start-teach since struggle here
 made kita n desu, toka itte
 till came NM CP something-like saying
 `And (he said something like), "So, since I started teaching,
 I have had great struggles before coming this far."'

(21) Sasaki no senyoosha o moo ichidai kattemorau tte
 Sasaki GN private-use-car OB more one receive-buy QP
 `(Sasaki's friend was saying) Sasaki will have them buy one more
 car for his private use.'

Even nante can mark a quote of another speaker rather than that of the speaker him/herself.

(22) (Quoting an utterance of a fellow student who is taking classes at
 a school specially designed for the certification program to become
 government workers)
 De, keizaigaku to nani, minpoo to keizal mikuro to
 and economics and what civil-law and economics micro and
 makuro to ryoohoo de, ato zaiseigaku mo aru naa,
 macro and both CP rest finance also exist FP
 nante.
 something-like
 `And, I have classes of economics, and what, the civil law,
 economics, both micro and macro, and also finance (something like
 that, he was saying).'

In contrast, I did not find any instances of nanchatte that do not imply self-mockery. (26) This suggests that the semantic meaning of nanchatte has been bleached and that nanchatte has become grammaticalized to signal self-mockery. (27) The exaggerated lengthening of the first or last vowel as well as the highly marked intonation that often accompanies nanchatte confirm this. Further, the fact that nanchatte has undergone phonological reduction (from nante itte shimatte) also suggests that its grammaticalization process is more advanced than that of other expressions.

In fact, an even more advanced stage of semantic change may be taking place. Gendai Yoogo no Kiso Chishiki `Basic Knowledge of Contemporary Terms' is a book series annually published to record new words in Japan. In the section entitled Wakamono Yoogo no Kaisetsu `Explanation of Youth Terms' in its 1996 version, in which various slang words and expressions are described, I found the following entry:

Nanchatte --: Fake. Famous brand names follow it as in Sono baggu, nanchatte shaneru janai no? `That bag, isn't it nanchatte (fake) Chanel?' (Horiuchi 1996: 1036).

This nanchatte as an adjective that means `fake' is obviously derived from the use of nanchatte discussed in this paper. It captures the intuition of native speakers regarding the suggestion of fakery associated with nanchatte. This innovative use of nanchatte as an adjective indicates that nanchatte's original lexical meanings have been completely bleached.

While this newer usage is still not well known among general speakers, the self-mocking function is readily recognized. It even appears in the popular discourse. For example, the novelist Eimi Yamada discusses the expression in her anthology called Kairaku no Dooshi `Verbs of Pleasure' (1997). (28)

Why has nanchatte come to have such a special status? More specifically, from among other similar expressions, why has it been chosen to be grammaticalized to be the marker of self-mockery? One reason is that nanchatte is equipped with all of the factors that contribute to the expression of self-mockery: reference to self-quotation, the lexical item that indicates lack of specification, the pause or other type of separation device, and the nonfinite verbal form. These contributing factors reinforce each other and make the association of nanchatte with the self-mocking function stronger. In addition, there may be two other reasons. First, of the two lexical items that indicate lack of specification, toka and nante, the pejorative meaning of nante is much more readily recognized than that of toka. Dictionaries discuss this function of nante. For example, Shinmura (1991 [1955]: 1938) lists nante as a particle that "in many cases expresses meanings of contempt or unexpectedness." (29) Because of this strong association of nante with pejorative meaning, nanchatte may have been chosen to be developed into the special marker of self-mockery.

Another factor may be attributed to the role of the auxiliary verb shimau. As mentioned earlier, nanchatte is a contracted form of nante itte shimatte. Shimatte is the nonfinite form of shimau. Shimau as a verb means `put away' or `finish'. Combined with the nonfinite form of another verb, shimau functions as an auxiliary that means `end up --ing' or `do -- completely'. Much has been written about the auxiliary (e.g. Alfonso 1966; Martin 1975; Teramura 1982; Soga 1983; Makino and Tsutsui 1986; Aoki and Okamoto 1988; Ono 1992; Ono and Suzuki 1992; Yoshida 1994; Strauss 1994; Strauss and Sohn 1998). It is primarily used as an aspectual marker that indicates that an event or action is completed. As an extension of this function, events marked by the auxiliary may also imply "a sense of irreversibility, the lack of control over some situation, the automatic or spontaneous occurrence of an action or event, and/or the speaker's reaction to some unexpected happenstance" (Strauss 1994: 259). These implications of the auxiliary may have reinforced the meaning of self-mockery in nanchatte. (30) In using shimau, Aoki and Okamoto (1988: 248) observe, the speaker is saying that something has been irreversibly done, nothing can be done to undo it, and therefore, s/he is helpless. The implication is that "the speaker wants to disclaim that it was through any planned activity" that something was done. This sense of lack of volition or control is the key notion here. The speaker is abandoning his/her responsibility for what has happened. In the case of nanchatte (nante itte shimatte), the auxiliary follows the verb of saying. The speaker is claiming lack of responsibility for saying what was said ("I unwittingly said ..."). Put differently, the speaker is distancing him/herself from the self who did the act of saying something. The speaker is presented as a divided self. As was discussed earlier, this presentation is closely connected to the idea of self-mockery since one can only mock by detaching oneself from the content one is mocking.

5. Concluding remarks

In an interview, an essayist named Sawako Agawa says the following in referring to the fact that her essays are often of the humorous kind:

It is the other side of my sense of embarrassment. While writing, essays tend to become preachy, saying things like "Love thy neighbor," "Be kind to others." I am too embarrassed to end essays like that, so I want to have a humorous surprising ending, saying nanchatte (Croissant 2000: 105).

As Agawa points out regarding nanchatte, it seems that the sense of embarrassment is often the underlying sentiment behind the use of self-mockery. Many of the examples cited in this paper involve the speaker's feeling of embarrassment. For example, the writer of (2) seems to be embarrassed by the earnestness of her UTTERANCE ("I may have been looking for a person like this from a long long time ago"). The speaker of (3) is embarrassed by the formality of the UTTERANCE ("I am indebted to you for your kind assistance on various matters"). The speaker of (9) feels awkward and vulnerable in his confession of love.

This sense of embarrassment causes the speaker to take a step back and dissociate him/herself from the statement s/he just made. By evoking the image of the speaker as a divided self, the speaker tries to disclaim responsibility and thus to escape from the awkwardness. As was discussed earlier, it is a self-protective strategy.

Somewhat paradoxically, this means of self-protection may be used to respect the wish of the addressee at the same time because self-mockery weakens the force of speech acts. Take (10), for example, in which the speaker asks the addressee to go on a trip with her. The speaker is able to say what she wants to say (i.e. the invitation for the trip) by uttering the UTTERANCE. By appending to it toka itte she places distance between herself and the speech act so that she can avoid the possible awkwardness that results from rejection. At the same time, by using the self-mocking expression and hence making the invitation less imposing, she honors the addressee's desire of not wanting to be put on the spot.

Although I have treated the English phrase just kidding as an equivalent of the self-mocking phrases in Japanese in this paper, I am not sure if the kinds of context in which it is used and thus the underlying motivations completely overlap with those of the Japanese counterparts. That is, I am not sure if the sense of embarrassment that is closely connected to the expression of self-mockery in Japanese is as prevalent in English-speaking cultures. My intuition tells me that it is not and that just kidding is probably used in more diverse contexts. Cross-linguistic and cross-cultural comparisons may be interesting topics for future studies.

Further, there are diverse kinds of speaker within the same language community. For example, the student who appended just kidding after almost all of her statements in classroom discussions in my course was a young Asian-American woman. Her age, gender, and cultural background may have played a role in her use of self-mockery.

With regard to his culture (i.e. Anglo-American male culture) Haiman (1998: 94) has the following to say:

Jocularity is so much a given in popular discourse that we sometimes have to signal its absence with the special metamessage "I am sincere" [...]. (31) So pervasive is the jocular style that we become aware of it only in its absence.

This comment suggests that among some English-speaking speakers the invalidating signal such as just kidding is almost superfluous. The speaker's insincerity in uttering the UTTERANCE is taken for granted and thus does not require special marking. Will this be the advanced stage of the expression of self-mockery in Japanese? If so, has there been any indication? If not, are there any correlations between this type of linguistic behavior and cultural factors? These are some of the questions further studies may explore.

Macalester College

Received 11 October 1999
Revised version received
23 January 2001

Notes

(1.) I would like to express my gratitude to John Haiman and Thorstein Fretheim for looking at an earlier version of this paper and giving me their suggestions and encouragement. I would also like to thank the two anonymous reviewers for their criticism and advice. All errors are my own. Correspondence address: Macalester College, 1600 Grand Avenue, St. Paul, MN 55105, USA. E-mail: suzuki@macalester.edu.

(2.) An earlier and shorter version of this paper was presented at the annual meeting of the High Desert Linguistics Society at the University of New Mexico in 2000.

(3.) Maynard (1996) also uses the term (self-) parody to refer to the same phenomenon. I elected to use the term self-mockery because the traditional definition of parody is something like "the imitative reference of one literary text to another, often with an implied critique of the object text" (Dane 1988: 4) and thus refers to a linguistic act that is related to but different from what this paper is concerned with.

(4.) Note the abbreviations used in the literal glosses:

AV adverbial marker
CP copua
FP final particle
GN genitive marker
NM nominalizer
OB direct object marker
QP quotative particle
SB subject marker
TG tag-like expression
TP topic marker

The conversational data, as is typical, contained instances of false start, repetition, overlaps, etc. To facilitate analysis and smooth presentation, the transcripts have been "cleaned up" to some extent. The following conventions are used in the transcription.

. falling intonation followed by noticeable pause
? rising intonation followed by noticeable pause
' continuing intonation, which may contain slight rise or
 fall in contour and may be followed by a short pause
... noticeable pause without falling intonation

(5.) These two terms are sometimes used interchangeably, so I will use both of them in my discussion henceforth. Based on the common usage, Haiman (1998: 20) makes two distinctions between the two. First, while situations may be ironic, only people can be sarcastic. Second, people may be unintentionally ironic, but intention is required in expressing sarcasm.

(6.) Fonagy (1982: 33, 63) briefly discusses the invalidating force of utterances such as I was just kidding.

(7.) I would like to thank Professor Yoshikazu Kawaguchi of Waseda University for his generous support while I was collecting this data.

(8.) I will use nanchatte to represent all of its variations in this paper, although this pronunciation and spelling may have become less common than the variations. For example, among the data found on the Internet, four were spelled as nanchatte, twelve as naanchatte, and nineteen as nanchattee.

(9.) Clark and Gerrig (1984) argue against Sperber and Wilson's (1981) mention theory of sarcasm/irony and theorize that the key notion in understanding the phenomenon is pretense. Haiman (1998: 25-26) observes that there is no real conflict between the ideas of pretense and mention and that the former simply subsumes the latter. He also states "At the deepest level, [...] the difference between mention and pretense dissolves completely, in that both derive from the more fundamental notion of repetition" (Haiman 1998: 26).

(10.) Wilson and Sperber (1992) modify their original theory (Sperber and Wilson 1981) and state that sarcasm/irony is a variety of echoic interpretation rather than echoic mention, where mention is equated with literal interpretation. Even after this modification, I believe Haiman's (1998) observation, discussed in note 9, is still relevant since it is concerned with the notion of echo.

(11.) For a review of classic and modern literature on the topic, see Johnson (1978).

(12.) In this sense joking is similar to sarcasm/irony.

(13.) Absurdity is produced because "the joker artificially creates a conflict between the deictic field and the verbal field [...]. His statement is purposely contradicted by the situation (which in this case may be broadened to include general beliefs and moral principles)" (Fonagy 1982: 33).

(14.) Put differently, I would like to define self-mockery as a linguistic act in which the invalidating signal is explicitly expressed and physically separated from the invalidated utterance, to distinguish it from joking.

(15.) This is not to say that all of the functions of just kidding and those of the Japanese expressions discussed in this paper are identical.

(16.) The meaning of self-mockery is especially strong in nanchatte, which will be discussed in detail in section 4.4.

(17.) Since the verb is in the nonfinite form, it is tenseless. It is common in Japanese conversation to use the nonfinite form at the end of an utterance (Maynard 1989: 38).

(18.) In fact, Maynard (1996:211) argues that the notion of multivoicedness associated with quotation appears in its most crystallized form in self-quotation precisely because self-quotation is a seemingly unexpected place for the phenomenon of many voices to appear.

(19.) Martin (1975:1021) translates toka as `something to the effect that' and `or something'.

(20.) Martin (1975: 163) considers nante to have developed from the combination of nani and to, the same source as for nado.

(21.) Alfonso (1966: 1138) translates nante as `something like that' and `something of the sort'. Martin (1975:160-161) says that nante is a synonym of nado in the meaning `the likes of; such a thing/person/place', etc.

(22.) The implication of the speaker's noncommittal attitude is not always interpreted as his/her negative evaluation. In contexts that call for consideration of politeness, the implication may be considered as the speaker's strategy of avoiding imposition. In fact, the particles toka and nante may be used as hedging expressions in certain contexts. See Suzuki (1998b) for a discussion of the relationship between the function of expressing the speaker's pejorative attitude and the hedging function.

(23.) Some linguists use the term "the gerund" to refer to this form. Others avoid the term and use instead "the te form" or simply te. Hasegawa (1996: 765) says, "Although `verbal + TE' exhibits some similarities with the gerund of Indo-European and other languages, it cannot in principle function as a nominal, and indeed in some uses TE functions more like the English conjunction and." Bisang (1998) analyzes the te form as a converb.

(24.) Hasegawa (1996) lists the following categories in the re-form linkage: circumstance, additive, cause-effect, means-end, contrastive, concession, and conditional.

(25.) John Haiman (personal communication) notes that utterances that are marked by nonfinal syntax in general often signal obsequiousness. He also observes that in English conversation the word "period" (an utterance terminator) is often used to signal commitment and vehement force.

(26.) I found 54 instances of nanchatte and its variations in the data.

(27.) Because nanchatte does not have a dual meaning to be resolved in context, Thorstein Fretheim (personal communication) suggests that it has developed a procedural, as opposed to a conceptual, lexical meaning: the addressee is being instructed to process UTTERANCE as a self-mocking contribution, A linguistic expression is considered to have a procedural meaning when it functions to "constrain the inferential phase of comprehension by reducing the hypothesis space that has to be searched in arriving at the intended interpretation" (Wilson and Sperber 1993: 21). See Blakemore (1987) and Wilson and Sperber (1993) for discussions on conceptual and procedural meanings.

(28.) Another appearance of nanchatte in the popular discourse is found in a web page entitled Tookoo! Nau na Shigo Jiten `Reader's Contribution! Dictionary of Hip Obsolete Words'. It defines na(a)nchatte as the expression that "is used when the speaker's utterance has spoiled the conversation or when the atmosphere becomes chilled because the speaker's joke has not been well received" (jibun no hatsugen de ba ga shiraketa toki ya majime na gyagu ga ukenakute osamui toki ni hassuru).

As the title of the web page suggests, the popular perception is that nanchatte is a dated expression. This is at least partly due to the fact that the expression was made popular in the mass media. Because of this exposure in the popular culture, the expression is considered to be slang. As is well known, slang comes to be viewed as outdated as time goes by. However, nanchatte is not obsolete. A recent (October, 2000) survey on an Internet search engine produced 35 instances of nanchatte and its variants used on various web pages.

Another popular perception is that the expression is used only after what is considered dajare, a poor joke that involves some kind of pun. Yamada's short story specifically deals with this. However, the actual use of nanchatte is not limited to the cooccurrence with dajare, as can be seen in the authentic examples in this paper.

(29.) The original Japanese is ooku keishi ya igai no i o arawasu.

(30.) Strauss and Sohn (1998: 222) observe that in the latest stage of grammaticalization of shimau these affective meanings have been bleached and that the auxiliary is used "as a colloquial marker of style or an informal in-group marker." This current function/ meaning coexists with the auxiliary's primary function as an aspectual marker as well as its affective implications. I would argue that the auxiliary's affective meanings influenced the process of nanchatte's grammaticalization even though the auxiliary's most recent function does not include these meanings.

(31.) I found several examples of this behavior in the English data I have. For example, the speaker in the following example says I am not kidding to emphasize his seriousness.

(i) I mean I'm gonna start dancing with those Brazilian women. ... So I can learn how to beat my hips. I mean, cause their hips are <<SLAPPING beating up against you ... you know, like that fast SLAPPING>> Hundred cycles per second, or something? [...] I mean, I am not kidding. It ... it is a different sensation entirely when you are dancing with them (Santa Barbara 2000: Part 1, 0002).

References

Adachi, Takanori (1996). Sarcasm in Japanese. Studies in Language 20(1), 1-36.

Alfonso, Anthony (1966). Japanese Language Patterns: A Structural Approach. Tokyo: Sophia University.

Aoki, Haruo; and Okamoto, Shigeko (1988). Rules for Conversational Rituals in Japanese. Tokyo: Taishuukan.

Bisang, Walter (1998). Adverbiality: the view from the Far East. In Adverbial Constructions in the Languages of Europe, Johan van der Auwera (ed.), 641-812. Berlin and New York: Mouton de Gruyter.

Blakemore, Diane (1987). Semantic Constraints on Relevance. Oxford: Blackwell.

Clark, Herbert H.; and Gerrig, Richard J. (1984). On the pretense theory of irony. Journal of Experimental Psychology: General 113(1), 121-126.

Dane, Joseph A. (1988). Parody: Critical Concepts Versus Literary Practices. Aristophanes to Sterne. Norman: University of Oklahoma Press.

Fonagy, Ivan (1982). He is only joking: joke, metaphor and language development. In Hungarian Linguistics, Ferenc Kiefer (ed.), 31-108. Amsterdam: Benjamins.

Grice, Paul H. (1975). Logic and conversation. In Syntax and Semantics 3: Speech Acts, Peter Cole and Jerry L. Morgan (eds.), 41-58. New York: Academic Press.

--(1978). Further notes on logic and conversation. In Syntax and Semantics 9: Pragmatics, Peter Cole (ed.), 113-127. New York: Academic Press.

Haiman, John (1985). Natural Syntax: Iconicity and Erosion. Cambridge: Cambridge University Press.

--(1989). Alienation in grammar. Studies in Language 13(1), 129-170.

--(1990). Sarcasm as theater. Cognitive Linguistics 1(2), 181-205.

--(1998). Talk Is Cheap. New York: Oxford University Press.

Hasegawa, Yoko (1996). The (nonvacuous) semantics of TE-linkage in Japanese. Journal of Pragmatics 25(6), 763-790.

Horiguchi, Junko (1995). Kaiwa ni okeru inyoo no "-tte" ni yoru shuuketsu ni tsuite. Nihongo Kyooiku 85(3), 12-24.

Horiuchi, Katsuaki (1996). Wakamono yoogo no kaisetsu. In Gendai Yoogo no Kiso Chishiki, 1033-1036. Tokyo: Jiyuukokuminsha.

Itani, Reiko (1994). A relevance-based analysis of hearsay particles: Japanese utterance-final tte. University College London Working Papers in Linguistics 6, 379-400.

Johnson, Ragnar (1978). Jokes, theories, anthropology. Semiotica 22 (3/4), 309-334.

Kindaichi, Haruhiko; et al. (1993). Nihon Kokugo Daijiten, vol. 7. Tokyo: Shogakkan.

Makino, Seiichi; and Tsutsui, Michio (1986). A Dictionary of Basic Japanese Grammar. Tokyo: Japan Times.

Martin, Samuel E. (1975). A Reference Grammar of Japanese. New Haven: Yale University Press.

Maynard, Senko K. (1989). Japanese Conversation: Self-Contextualization Through Structure and Interactional Management. Norwood, NJ: Ablex.

--(1990). An Introduction to Japanese Grammar and Communication Strategies. Tokyo: Japan Times.

--(1996). Multivoicedness in speech and thought representation: the case of self-quotation in Japanese. Journal of Pragmatics 25, 207 226.

--(1997). Meta-quotation: the thematic and interactional significance of tte in Japanese comics. Functions of Language 4(1), 23-46.

Miura, Akira (1974). "To" to "tte." Nihongo Kyooiku 24, 23-28.

Morishige, Satoshi (1954). "Te, tte" "teba, tteba" "tara, ttara" ni tsuite. Kokugo Kokubun 23(11), 51-62.

Okamoto, Shigeko (1995). Pragmaticization of meaning in some sentence-final particles in Japanese. In Essays in Semantics and Pragmatics, Masayoshi Shibatani and Sandra Thompson (eds.), 219-246. Amsterdam: Benjamins.

Ono, Tsuyoshi (1992). The grammaticalization of the Japanese verbs oku and shimau. Cognitive Linguistics 3(4), 367-390.

--; and Suzuki, Ryoko (1992). The development of a marker of speaker's attitude: the pragmatic use of the Japanese grammaticized verb shimau in conversation. In Proceedings of the 18th Annual Meeting of the Berkeley Linguistics Society, Laura A. Buszard-Welcher, Lionel Wee, and William Weigel (eds.), 204-213. Berkeley: Berkeley Linguistics Society.

Shinmura, Izuru (1991 [1955]). Koojien. Tokyo: Iwanami Shoten.

Soga, Matsuo (1983). Tense and Aspect in Modern Colloquial Japanese. Vancouver: University of British Columbia Press.

Sperber, Dan; and Wilson, Deirdre (1981). Irony and the use-mention distinction. In Radical Pragmatics, Peter Cole (ed.), 295-318. New York: Academic Press.

Strauss, Susan (1994). A cross-linguistic analysis of Japanese, Korean and Spanish: -te shimau, -a/e pelita, and the "Romance reflexive" se. In Japanese/Korean Linguistics, vol. 4, Noriko Akatsuka (ed.), 257-273. Stanford: CSLI.

--; and Sohn, Sung-Ock (1998). Grammaticalization, aspect, and emotion: the case of Japanese -te shimau and Korean -a/e pelita. In Japanese/Korean Linguistics, vol. 8, David J. Silva (ed.), 217-230. Stanford: CSLI.

Suzuki, Satoko (1996). The discourse function of the quotation marker tte in conversational Japanese. In Proceedings of the 22nd Annual Meeting of the Berkeley Linguistics Society, Jan Johnson, Matthew L. Juge, and Jeff L. Moxley (eds.), 387-393. Berkeley: Berkeley Linguistics Society.

--(1998a). Tte and nante: markers of psychological distance in Japanese conversation. Journal of Pragmatics 29,429 -462.

--(1998b). Pejorative connotation: a case of Japanese. In Discourse Markers: Descriptions and Theory, Andreas H. Jucker and Yael Ziv (eds.), 261-276. Amsterdam: Benjamins.

Teramura, Hideo (1982). Nihongo no Shintakusu to Imi, Part 2. Tokyo: Kuroshio Shuppan.

Volosinov, V. N. (1973). Marxism and the Philosophy of Language, Ladislav Matejka and I. R. Titunik (trans.). New York: Seminar.

Wilson, Deirdre; and Sperber, Dan (1992). On verbal irony. Lingua 87, 53-76.

-- and Sperber, Dan (1993). Linguistic form and relevance. Lingua 90, 1-25.

Yamada, Eimi (1997). Kairaku no Dooshi. Tokyo: Bungei Shunjuu.

Yoshida, Eri (1994). Speaker's subjectivity and the use of shimau, in Japanese spoken narratives. In Japanese Korean Linguistics, vol. 4, Noriko Akatsuka (ed.), 183-199. Stanford: CSLI.

Data references

Archives. http://www.and.or.jp/diable/logs9707.html

Chiba, Atsuko (1985). Chotto Okashiizo, Nihonjin. Tokyo: Shinchoosha.

Croissant (2000). August 25. Tokyo: Magazine House.

Kanai, Mieko (1995). Ren'ai Taiheiki, vol. 1. Tokyo: Shuueisha.

Miyuki Joo. http://www.est.hi-ho.ne.jp/rediriver/missm.htm

More (1996). April. Tokyo: Shuueisha.

Saigo no Koi (1997). TV drama broadcast on TBS, Japan, from July 11, 1997 to September 19, 1997.

Santa Barbara (2000). Santa Barbara Corpus of Spoken American English, Part One. Santa Barbara: University of California, Santa Barbara Center for the Study of Discourse.

Tookoo! Nau na Shigo Jiten. http://www.orange.ne.jp/-kibita/dwd/dwddic11.html#n