Generic and vague uses of a second-person singular pronoun in an open-class person-reference system and speaker creativity in reported speech: the case of anata in Japanese

1 Introduction

It has been observed across languages that 2sg pronouns can be used not only as terms of address or second-person reference but also to refer to people in general or to human referents who are very low in specificity. This kind of usage has been described variously as non-canonical, universal non-specific, generic, generalized human, indefinite, referentially arbitrary and impersonal, amongst other terms (see Deringer et al. 2015; Gast et al. 2015; Henkin 2020, 2022; Kitagawa and Lehrer 1990; Kluge 2016; Laberge and Sankoff 1979; Siewierska 2004, among others).

Laberge and Sankoff (1979) used the term ‘indefinite’ in their study of Montreal French and showed that, in contexts where the indefinite pronoun on is used, the 2sg tu (as well as its polite equivalent vous) could alternatively be used as an indefinite, as in Example (1).

In (1), tu cannot be interpreted as referring to the addressee in the immediate conversational situation. Instead, it is understood as referring to a generalized human referent, ‘anyone’.^[1] Laberge and Sankoff (1979: 423) explain that by using tu with the expression le monde où … ‘the world where …’, the speaker evokes a general situation in which ‘anyone’ could participate. These authors describe the most crucial element unifying various contexts in which indefinites can be used as “the theme of generality or generalization” (1979: 423).

In other studies, this generic sense is expressed by the terms ‘generic seconds’ or ‘impersonal’. Kluge (2016: 503) uses the former term and states, “[g]eneric seconds are cross-linguistically attested in many languages but they are far from universal”. On this point, using the latter term, ‘impersonal’, Kitagawa and Lehrer (1990: 753) claim a typological pattern as follows:

The extension of the 2nd person pronoun to an impersonal is possible only in languages with small, closed pronoun sets.

According to Kitagawa and Lehrer (1990: 752), the above pattern is applicable to such languages as French (as seen in Example [1]), German, Italian, English, Gulf Arabic, Modern Hebrew, Hindi-Urdu, Persian (Farsi), and Mandarin, all of which meet their criterion of having a closed set of personal pronouns. In these languages, personal pronouns primarily function as person-deixis (Kitagawa and Lehrer 1990: 753). This ‘person-deixis’ framework allows the speaker to abstract the 2sg “away from its referential property associated with the immediate speech act domain” to a different sphere, a more generalized situation (Kitagawa and Lehrer 1990: 753).

On the other hand, regarding languages without a closed set of personal pronouns, Kitagawa and Lehrer (1990: 753) claim that:

[this linguistic feature] would place, on the other hand, such languages as Japanese and Korean among those having no recourse to impersonal use of 2nd person pronoun; neither Japanese nor Korean possesses a clearly defined closed set of personal pronouns.

The above statement is based on the widely held assumption that all personal pronouns in these languages are semantically and pragmatically “too loaded” to be used as impersonals (Kitagawa and Lehrer 1990: 756). Indeed, many languages in Asia, including Japanese and Korean, have elaborate pronominal paradigms and open-class alternative terms (Enfield and Comrie 2015: 8).^[2] The use of these terms is primarily determined by the social characteristics of the interlocutors, such as age, gender, social status, kin relations, and the social distance between them, as well as the level of formality in the conversational setting. In these languages, personal pronouns are regarded as inevitably involving social-indexicality. It is presumably for this reason that they have been assumed to be incapable of functioning as ‘mere person-deixis’. In other words, it has been considered impossible for these languages to have generic uses of any 2sg pronouns.

However, among various terms of address and reference in present-day Japanese, one 2sg pronoun, anata ‘you’, is found to be unique, as I argued in a recent study (Yonezawa 2021). I will summarize the study in Section 3.2, but briefly state the central thesis here: unlike all other 2sg pronouns in the Japanese language, anata is semantically simple and socially inert and thus, in fact, primarily functions as mere person-deixis.^[3] The discovery of the semantic simplicity of the 2sg pronoun anata in Japanese suggests the potential for it to function as a generic second. The current study shows that generic and vague uses of 2sg anata do, indeed, occur in Japanese, in reported speech environments. These uses have been almost entirely neglected in studies of the Japanese person-reference system. To date, only a few published studies have attempted to discuss these uses (e.g., Shimotani 2012), partly due to their low frequency in ordinary conversation corpora. This study fills this gap by closely examining generic and vague uses of the 2sg anata in reported speech.

In addition to the discovery of generic and vague uses of 2sg in an open-class person-reference system, another important contribution of this study to typological research is in relation to the study of pronouns in reported speech. Spronck and Nikitina (2019: 131) state that the interpretation of pronouns and other deictics in reported speech demands specific knowledge of the grammar of the language. This study further shows that, in the case of languages with an open-class pronominal system, the interpretation of pronouns in reported speech also requires a thorough understanding of their core semantics, as well as their pragmatic usage.

It should be noted that, while I use the term ‘reported speech’ in this article, this term does not mean that the speaker in the immediate conversational situation ‘reports’ verbatim utterances of the original speaker. The term ‘reported speech’ here refers to the same phenomenon that is described as “constructed dialogue” (Tannen 1989) and “represented speech” (McGregor 1997; Vandelanotte 2004; Verstraete 2011), in the sense that the current speaker, though using the form of a ‘report’, actually constructs the dialogue within the speech. The intention of the quoting speaker is to give a ‘representation’ of the speech used, not an accurate verbatim ‘report’. Tannen rejects the term ‘reported speech’ altogether, stating:

when a speaker represents an utterance as the words of another, what results is by no means describable as ‘reported speech’. Rather it is constructed dialogue. And the construction of the dialogue represents an active, creative, transforming move which expresses the relationship not between the quoted party and the topic of talk but rather the quoting party and the audience to whom the quotation is delivered. (Tannen 1989: 109)

While I continue to use the widely accepted term ‘reported speech’, I note that generic and vague uses of anata are understood most often as the quoting speaker’s ‘creation’ rather than as a verbatim quote of the original words used.

The investigation of these uses of anata in this study is conducted through data drawn from Japanese parliamentary debates. It will be shown that generic and vague uses of anata are part of speakers’ speech strategies in politicians’ effort to evoke “the theme of generality or generalization” (Laberge and Sankoff 1979: 423) in these debates, with a view to making their argument persuasive or relatable.

The article is organized as follows. Section 2 takes a brief look at some examples of generic and vague uses of you in English and ni in Mandarin and introduces the notion of a ‘continuum of reference of the 2sg’ proposed by Kluge (2016: 504), which is based on data from European languages. Sections 3 and 4 give background knowledge necessary for the later analysis: the Japanese person-reference system and the peculiarity of the 2sg anata (Section 3) as well as features of reported speech in Japanese (Section 4). Section 5 outlines the data and methodology. Section 6 provides findings and discussion, by proposing a ‘continuum of reference of the 2sg in reported speech in Japanese’, inspired by Kluge’s (2016) continuum but different from it. Section 7 gives concluding remarks with some notes on the study’s implications.

2 Generic seconds and the notion of a continuum of reference of the 2sg

In this section, I briefly touch on examples, from languages with closed pronoun sets, of the uses of 2sg pronouns that differ from the more common, fully referential uses. I present examples from English, drawn from Kitagawa and Lehrer (1990), and Mandarin, drawn from Biq (1991). I then introduce the notion of a ‘continuum of reference of the 2sg’ proposed by Kluge (2016: 504) for European languages. This notion is insightful, and it is also a useful starting point for my later discussion of the generic and vague uses of anata in Japanese.

Let us begin by looking at examples in English. Kitagawa and Lehrer (1990: 742) point to two uses of non-referential you. One they refer to as an ‘impersonal’ use, which “applies to anyone and/or everyone” (referred to in the current study as the ‘generic’ use).^[4] The other is a ‘vague’ use, which “applies to specific individuals, but they are not identified, or identifiable, by the speaker”. Consider the following examples.

Example (2) is from an interview with a man who teaches fiction writing. In this example, you and your do not refer to the addressee in the immediate conversational situation. It is not the interviewee who has to ‘examine manuscripts’ and to ‘use [his/her] mind’. The speaker is expressing the notion that anyone at all who teaches fiction has to do so (Kitagawa and Lehrer 1990). These authors state that (2) is a typical example of what they call the ‘impersonal’ you in English.

Example (3) is the ‘vague’ you. A European woman is talking to one of the authors of the study from which the example was taken (Lehrer), who is a U.S. citizen, about American political and military policy in Europe. As evidenced by the speaker’s insertion of the phrase “I don’t mean you personally”, this you is clearly not referring to her addressee. Instead, it “evokes a membership categorization device in terms of the interlocutor’s nationality” (Kluge 2016: 515). Kitagawa and Lehrer differentiate impersonal you and vague you in English, showing that impersonal you can be replaced with one or we without changing the informational content of the text, while vague you cannot. Impersonal use is unlike vague you in that it always includes the addressee; a phrase such as “I don’t mean you personally” would not make sense in this case.

In relation to this type of usage, Kluge (2016: 504) has proposed the notion of a ‘continuum of reference of the 2sg’. In her corpus of French, Spanish, German, Portuguese, and English, she identified five focal points within the continuum (Figure 1):

Figure 1:

Continuum of reference of the 2sg (Kluge 2016: 504).

Using this continuum, the referents of you in Examples (2) and (3) can be more precisely described. While Kitagawa and Lehrer (1990) interpreted Example (2) as a typical impersonal use of you stating that ‘anyone’ who is in the described situation would behave in the described way, a more precise interpretation of this example could be Kluge’s number 1 in the continuum, “I the speaker (hiding behind ‘you’)”.^[5] You in this example can be seen as subsuming self-referential value, especially with the preceding utterance “but I have a gift for teaching”. Being hidden by a 2sg, the utterance is presented in a generalizable manner inviting the addressee to view the situation ‘as if’ it is also their perspective. Regarding such cases where the addressee is invited to imagine their participation in the described situation by the use of impersonal you, Deringer et al. (2015) and Gast et al. (2015) state that, by using Moltmann’s term ‘simulation’ (2006), the addressee is “invited to simulate participating in a situation in which she does not actually participate” (Deringer et al. 2015: 319). Example (3) can be understood as number 4 “you, the person in front of me, as a representative of a larger entity”.

Moving on to Mandarin examples, (4) and (5) are uses of the Mandarin ni ‘you’ from Biq (1991: 309).^[6]

As seen in Biq’s translation, like Kitagawa and Lehrer in their interpretation of the English ‘impersonal’ (generic) you, Biq considers the referent of ni in (4) to be you, me, and even anyone. However, again, the application of Kluge’s continuum helps interpret this ni more precisely. This ni can also be understood as Kluge’s number 1 “I, the speaker (hiding behind ‘you’)”. Here again, by the use of a second-person singular pronoun, the utterance is presented in a manner that can be generalized, inviting the addressee to perceive the situation as if it were their own perspective.

A particularly important example from Biq’s study, pivotal for our later discussion of generic and vague uses of the 2sg anata in Japanese, is Example (5) below. Biq (1991: 311) notes that it shows the speaker’s strategy to create a deictic shift by using ni in reported speech, more specifically, in the form of ‘direct quotation’. Biq (1991: 310–314) calls this the “dramatic ni”.

In the quoted part, the speaker is an imaginary ‘person X’. The current speaker has substituted the voice of the imaginary character in the hypothetical situation for his/her own viewpoint. These uses of ni are cases of Kluge’s number^{[7] [8]} 5 “you, the person in front of me”, not in the speech event but in the hypothetical situation, in the sense that ‘person X’ is represented as talking to ‘person Y’. Indeed, reported speech is described as Speech and Thought Representation (Maynard 1996; Vandelanotte 2009), serving to prompt a deictic shift (Duchan et al. 1995), and providing an alternative ‘mental space’ (Kluge 2016: 508). Regarding mental space, Sweetser and Fauconnier (1996: 11) explain, “as we think and talk, mental spaces are set up, structured, and linked under pressure from grammar, context, and culture. The effect is to create a network of spaces through which we move as discourse unfolds”. Rubba (1996: 234) states, “mental spaces provide grounds for interpreting deictics in quoted speech”. In Japanese, it is in reported speech, as a creative form of direct quotation, that the continuum of generic and vague uses of anata is observed (as briefly noted in Section 1 and discussed in detail in Section 6). In this sense, these uses share features of what Biq (1991) calls “dramatic ni”, in that the quoted speaker is viewed as one of the “dramatis personae” (Kitagawa and Lehrer 1990: 752) or “the animated role of the mental space’s protagonist” (Kluge 2016: 507) in the world of quoted discourse, that is, a different mental space away from the current conversational situation.

However, a crucial difference between cases of Mandarin and Japanese is that in Mandarin, as a language with a closed pronominal system, the current speaker has little choice but to use ni in this context. On the contrary, in Japanese, as a language with an open-class person-reference system, there are numerous possibilities in the choice of representing the quoted speaker’s person-reference terms. In Japanese, current speakers have more freedom to put different person-reference terms in the mouths of those whose dialogue they are representing. To expand on this point, I briefly introduce the Japanese person-reference system in the next section.

3 Person-reference system in Japanese

4 Reported speech in Japanese

Before moving on to the main discussion, features of reported speech in Japanese should be noted. Generally, issues concerning the distinction between so-called direct and indirect speech have been known for some time (Coulmas 1986b; Evans 2013). Here, by ‘direct speech’, I mean that the reported/quoted part is presented as if it reflects an authentic statement made by the reported speaker, even though reported/quoted components are ultimately a creation of the current speaker, as discussed in Section 1. On the other hand, ‘indirect speech’ explicitly indicates a formulation by the current speaker, described by Mushin (2001: 117) as “a representation of the ‘gist’ of the original utterance”.

In Japanese, while the direct and indirect distinction is relevant, both grammatical and pragmatic elements contribute to obscure this distinction (Coulmas 1985, 1986a; Kamada 2000). Example (15) is such a case where both direct and indirect readings are possible. As Japanese reported speech does not involve the deictic shift in tense that occurs in the two English translations (a and b) (will in direct speech is converted to would in indirect speech), Example (15) can be interpreted in both of the ways shown in the translations.

On the one hand, the tentative form of copula daroo in (15) can be understood as a casual sentence ending, unmarked for interaction or politeness, which is perfectly appropriate in conversation between friends. In this reading, the quoted part, kyoo wa daremo boku no uchi e konai daroo, can be understood as a direct quote (translation a) and boku no uchi is interpreted as referring to Taro’s house. On the other, daroo can also be read as the plain form grammatically appropriate within an indirect quote, from which all markers of deixis and interactional aspects have been stripped. In this case, the sentence is interpreted as indirect speech (translation b), with boku no uchi understood as the quoting speaker’s house.

The distinction between direct and indirect speech in Japanese can, however, often be identified based on other deictic switches (not including tense, as is common in English) (Coulmas 1985, 1986a; Kamada 2000). Examples (16) and (17) are the cases where the direct or indirect reading is clearly marked by deictic expressions.

In (16), the verb kudasaru is an honorific form of kureru ‘give’, which requires the addressee to be the subject. Thus, the addressee is the giver and the speaker is the recipient. This indicates the presence of the addressee and demands direct reading. On the other hand, the plain form of ‘do’, yaru, and yôni ‘in a way’, which is “chiefly used as an idiomatized marker of the imperative and hortative moods in indirect speech”, ensures that (17) is understood as indirect speech (Coulmas 1986a: 166).

Social deixis plays a particularly important role in distinguishing a direct from an indirect speech quotation. These social-deictic elements include interjections, person-reference terms, attitudinal expressions such as honorifics and politeness markers, and interactional particles. For example, in (15), if daroo occurs with the interactional particle ne, which is typically used in spoken conversation (e.g., McGloin 1990; Ogi 2017), the ending of the quoted part daroo ne signals the quoted sentence as the utterance of the original speaker to the original addressee. In other words, the sentence is interpreted as direct speech. If the polite form of daroo, which is deshoo, is used, this also indicates the presence of an addressee, to whom the original speaker showed deference in the quoted conversational situation. Thus, it is also clearly signaled as direct.

The use of 2sg pronouns in quotation in Japanese can also contribute to shifting the deictic perspective, which leads the quoted part to be interpreted as a direct quote. Example (18) is from a TV interview show, cited in Kamada (2000: 70). The speaker is an actor, who is a guest on the show. He is talking about a memory of something his middle-school teacher said to him.

In (18), the use of 2sg kimi indicates the start of the direct quotation. As shown in Table 1, kimi is an informal 2sg pronoun used toward an equal or an inferior addressee. The referent of this use of kimi cannot be interpreted as the addressee of the current conversational situation since the speaker is talking to the TV show host, who is much older than the speaker and is a respected figure. With the use of kimi, which a teacher can typically use to address/refer to a student, the speaker triggers the shift of the interlocutors’ perspective to the teacher-student conversation in his recollected past.

As such, although there are cases, such as the earlier Example (15), where both direct and indirect readings are possible, the distinction of direct and indirect speech in Japanese is relevant and is signaled by a range of grammatical or pragmatic features, including the use of personal pronouns.

It is important to reiterate here again that, as noted earlier, reported speech is most often the speaker’s ‘creation’ rather than the quoting speaker’s exact ‘report’ of the original utterances (e.g. Coulmas 1986b; Tannen 1989). In Example (18), the current speaker’s use of kimi does not necessarily mean that the term kimi is precisely the word the original speaker used. Recall also that Japanese, as a language with an open-class person-reference system, provides the speaker numerous choices of person-reference terms (constrained, of course, by the social relationship between the interlocutors). In (18), it is the current speaker who employs kimi, a typical 2sg pronoun used by a teacher toward a student. In so doing, the current speaker creates the dialogue which is perceived to be close to the original conversation that he is recollecting. In the reported speech, kimi in (18) is interpreted as, or close to, Kluge’s number 5 “you, the person in front of me” within the represented dialogue (see Section 2).

The creativity of reported speech is particularly relevant to my central discussion of the use of the 2sg anata. As noted earlier, because anata is rarely observed in daily conversations, its occurrence in reported speech is most likely the current speaker’s creation. As will be shown, this context provides rich soil for generic and vague uses of anata.

5 The data and methodology

As noted earlier, the frequency of the occurrence of anata has been found to be extremely low in ordinary conversation corpora (Usami 2007). It is understandable that previous studies have treated the occurrence of anata in reported speech as an ‘exception’ and have thus overlooked the existence of generic and vague uses of the term. On the other hand, in parliamentary debates, the use of anata is abundantly observed (I will discuss this further in Section 6.1). Thus, the data for this study are drawn from Japanese parliamentary debates, extracted from the online database, Kokkai Kaigiroku Kensaku Shisutemu ‘Minutes of the Japanese Diet Retrieval System’.^[13] This study analyses the occurrences of 2sg anata in the various Diet meetings held throughout a single year (February 25, 2012 to February 25, 2013).

It should be noted that there might be a concern about the use of parliamentary debate as spontaneous spoken data for two reasons. First, in parliament, the speakers often have prepared scripts or notes for their speech. They refer to them during the speech, or occasionally may simply read certain documents aloud. However, they also rephrase, insert, and add interactive features. Further, in many cases during the dynamics of debating, they also speak freely without looking at notes and scripts. Thus, there are abundant elements which can be analyzed as spoken discourse. In terms of the focus of this study, the use of anata in reported speech, it is particularly hard to imagine that the excerpts we will investigate in the following subsections are scripted scenarios, because they take the form of inserted conversations which are ‘acted out’ by the speaker. That said, by cross-checking each meeting transcript with its recording (available online), I ensured that the uses of anata did occur in unscripted parts and that the transcription reflected exactly what was uttered. To the best of my observations through this cross-checking, the uses of anata discussed in this study are all, indeed, part of spontaneous utterances.

The second reason for potential concern about the use of parliamentary debate as spontaneous spoken data is its obvious differences from everyday conversation. During parliamentary talk, each speaker is officially given the floor for a certain length of time and is generally uninterrupted. In this sense, parliamentary talk has features of a range of different styles including lecture style, prepared debate, persuasive rhetorical style, and narrative, all of which are very different from spontaneous conversation, where frequent interruption and co-construction between conversation participants are the norm in Japanese. Iwasaki (2015) states that parliamentary debate is a genre which exhibits a mixture of spoken features and aspects of written grammar. However, this does not mean that parliamentary talk is unsuitable for spoken discourse analysis. Hopper and Thompson (1980: 282) used narrative for their analysis of transitivity, stating “the phenomena exemplified in narrative are also present in other genres”, but simply to a lesser degree in genres such as spontaneous conversation. In the same way, the analysis of the examples of this study will certainly serve to inform an understanding of the small number of examples that have been found in spoken corpora and that have been excluded by previous studies as exceptional.

6 Findings and discussion

6.1 Quantitative results

The total number of occurrences of 2sg anata in the data reached 1610. Figure 2 shows that among all uses of anata in the data, 1418 occurrences (88 %) were terms of address or interlocutor reference in the immediate speech situation. The remaining 192 occurrences (12 %) comprise the use of anata in reported speech and are the focus of the current study.

Figure 2:

The occurrences of anata in the data: immediate speech situation and reported speech.¹⁴

Although the uses of anata as terms of address or interlocutor reference in the immediate discourse situation are not the central issue in this article, the above quantitative result may trigger a question as to why these uses of anata are observed far more often in parliamentary debate compared to ordinary conversations, in which anata is rarely observed. There are two reasons for this.

First, recall the classification of relationships by Yonezawa (2021), noted in Section 3.2, into ‘socially definable’ and ‘socially undefinable’. When the interlocutors’ relationship is socially definable, there are always socially^[14] expected terms of address and reference. If the speaker were to use anata, which absolutely specifies the second person entity, it would violate the important sociocultural norm of ‘relationship acknowledgment’ in Japanese communication. Thus, in daily conversations, anata is largely avoided. On the flip side, however, anata can be used to deliberately violate the expected social norm. For example, when the speaker argues with or criticizes the addressee, or wishes to ignore the conventional expectations of their relationship, the use of anata functions to express the speaker’s attitude of rejecting the given social relationship (Yonezawa 2021: 118–136). These situations are ‘unusual’ in ordinary conversations. In parliamentary debates, on the other hand, such situations are very much on parade; members of parliament (MPs) constantly attack, criticize, denigrate, and scorn their addressee, in many cases even the Prime Minister or a minister who is senior to the speaker. Instead of using conventional address terms to address/refer to superior addressees, such as soori ‘Prime Minister’ and daijin ‘Minister’, MPs often use anata as a strong expression of impoliteness to challenge those with greater power. For this reason, this function of the use of anata (i.e., rejection of a given social relationship) is particularly frequently observed in parliamentary debates. In terms of the categorization method, the author categorized inferior speakers’ use of anata towards superior addressees, such as regular MPs addressing/referring to a minister or the Prime Minister, observed in contexts such as attacking, criticizing, and denigrating the addressee, into the category of ‘rejection of a given relationship’, as the use of anata in these utterances clearly contributes to an attempt to deny the addressee as deserving their elevated position. Of the 88 % of uses of anata in the immediate discourse situation, this function of rejecting a given social relationship occupies 67 %, as shown in Figure 3.

Figure 3:

The occurrences of anata in the data: breakdown of types in the immediate speech situation.

Second, when the interlocutors’ social relationship is undefinable (also discussed in Section 3.2), the use of anata, as a pure marker of person-deixis, is suitable without creating any strong pragmatic effect. This includes when the speaker refers to a ‘depersonalized’ addressee. A common example of this in the Japanese Diet is on occasions when the House of Councilors and the House of Representatives call for shoonin kanmon ‘summoning of a witness’ to investigate a particular case, which may involve an assessment of criminal liability of a certain MP. In shoonin kanmon, regardless of his/her social position, the witness is treated as a depersonalized, mere addressee. The questioner often goes on to refer to the witness as addressee using anata without indicating any social elements. This is similar to a courtroom setting, where all individuals in a court are placed equally before judges and treated as depersonalized addressees. Of the 88 % of the use of anata in the immediate speech situation in the data, reference to a depersonalized addressee in witness summons occupies 21 % (Figure 3).

Our main discussion here is the use of anata in reported speech. As Figures 2 and 3 show, 12 % of the uses of anata in the data are observed in reported speech. As noted in Section 4, the use of personal pronouns contributes to shifting the deictic perspective, providing a direct-speech reading. The author manually coded all the occurrences of anata used in reported speech, all of which were interpreted as direct speech. Interestingly, these are almost entirely generic and vague uses. (Details will be presented in Section 6.3.). As will be shown, these uses are part of the political art of speech in politicians’ efforts to create ‘generalized’ narratives (Kitagawa and Lehrer 1990).

6.2 A continuum of reference of the 2sg in Japanese in reported speech

The analysis of the data reveals that anata can refer to human referents with varying degrees of specificity in reported speech in Japanese. Minagawa (2009: 43) states that ‘specificity’ is “the speaker’s commitment to refer to a particular entity which she has in mind”. This is in line with Givón’s (1993: 224) notion of “referential intent”. Givón (1993: 226) points out that specificity involves a continuum rather than a clear division and the interpretation of the degree of specificity of reference cannot always be determined in a “strict logical-semantic sense” but is rather determined in a pragmatic sense. For example, the singular indefinite noun in He sold a house is interpreted as highly specific, but the plural in He sold houses downgrades referential intent. The interpretation of low specificity in the latter sentence is generated in the pragmatic sense: the specific identity of each house sold simply “doesn’t matter” (Givón 1993: 226).

Considering the notions of degree of specificity (Minagawa 2009), referential intent (Givón 1993), and Kluge’s (2016) ‘continuum of reference of the 2sg’ (see Section 2), and on the basis of the data analyzed for this study, I propose the following continuum of reference of Japanese 2sg in reported speech (Figure 4). The concept was initially discussed briefly in Yonezawa (2021: 101) but due to the different focus in that work, it was not explored fully there. I advance the concept further here, with examples from the current data. This continuum differs from Kluge’s, presented in Section 2, which is based on data from European languages; it is organized according to the degree of specificity of the referent, from lowest (1) to highest (5),^[15] and never involves any hidden reference to the current speaker (‘I’), nor any reference to the current addressee (‘you’) within the current speech context. This relates to the fact that, unlike Kluge’s continuum, it is here applied only to the environment of reported speech in Japanese. Figure 4 presents each focal point together with a picture that helps to visualize the concept.

Figure 4:

Continuum of reference of Japanese 2sg in reported speech (modified from Yonezawa [2021: 101]).

Number 1 ‘anyone’ is what this study calls ‘generic use’. In this case, anata refers to people in general, and in indirect speech style, it can be replaced by generic terms such as hitobito ‘people’, minasan ‘everyone’, and daredemo ‘anyone’.

Number 2, ‘you, as a representative of a type of person in a hypothetical situation’, is ‘vague’ use. In indirect speech, anata can be replaced by terms such as sooiu hito(tachi) ‘that kind of person/people’, aru hito ‘a certain person’, and hitori … moo hitori ‘one … another’. As the constructed dialogue is a hypothetical situation, this leads to the referent being abstracted and interpreted as a representative of a type of person who is in the described situation or has the described attributes.

Number 3, ‘you, a non-identifiable individual’, is also vague use. The use of anata in this case refers to an individual who exists and is talked about in the reported situation. However, the current speaker does not commit to specifying the identity of the individual. In direct speech, the individual would be referred to with terms such as sono hito ‘that person’ and kono hito ‘this person’ but remain unidentifiable.

Number 4, ‘you, as a representative of a collective entity’, is also vague use. In this case, the use of anata refers to a collective entity.^[16] These collective entities are actual existing entities, such as nations, institutions, places, and associations; thus, this type can be said to refer to a concrete, non-hypothetical group of people. In indirect speech, anata can be replaced by an actual entity, such as ‘the United States’ and ‘the University of Tokyo’.

Finally, number 5 ‘you, an identifiable, specific individual’, is the only referential use in the continuum. Anata uniquely refers to a specific and identifiable person within the reported situation. In indirect speech, anata can be replaced by the individual’s actual name or position, such as ‘Mr Suzuki’ and ‘Prime Minister (Noda)’. If anata is referring to the speaker him/herself, it can be replaced by a 1sg pronoun in indirect speech.

The method of coding used to categorize anata into each focal point converted the direct speech style of the example into an indirect one and checked what kinds of terms anata could be replaced with in that context. It is important to note that numbers 1 to 5 in Figure 4 are prototypical focal points and not discrete categories, meaning that the boundary of each focal point is not absolute (see Section 6.3.3, for an example of a borderline case). Thus, in addition to the author herself, another native speaker also coded the data. The interrater reliability was measured using Cohen’s Kappa statistic. The Kappa value was 0.935, which indicates near perfect agreement. That said, both coders identified borderline cases before reaching their choice of categories. These borderline cases were also discussed with other native speaker colleagues.^[17] Two of these cases are presented as Examples (25) and (26) in Section 6.3.3. In the next section, I will discuss examples from the data for each focal point in the continuum. The majority of examples given in the next section represent typical members of each category.

6.3 The continuum of reference in 2sg anata in reported speech in the data

In Section 6.1, I have shown that 12 % of the uses of anata in the data (192 occurrences) are observed in reported speech. The main discussion in this article is the use of anata in this environment. The chart in Figure 5 shows the distribution of these 192 occurrences of anata across each focal point within the continuum proposed in Section 6.2.

Figure 5:

Distribution across the continuum of 2sg anata in reported speech.

Among the five focal points, number 1 ‘anyone’ is what this study calls ‘generic’ use. In the data, 14 % of the occurrences belong to this category. Number 2 ‘you, as a representative of a type of person in a hypothetical situation’, number 3 ‘you, a non-identifiable individual’, and number 4 ‘you, as a representative of a collective entity’, are what this study refers to as ‘vague’ uses. In the data, 45 % of the occurrences of anata in reported speech are interpreted as number 2, 16 % as number 3, and 17 % as number 4. Given this, the uses of anata in reported speech in the data are almost entirely generic and vague uses (92 %). Only 8 % belong to the case where the speaker cited in reported speech is referring to an identifiable, specific addressee in the reported speech situation, i.e., number 5 ‘you, an identifiable, specific individual’. Note that this is a referential use and thus does not belong to generic and vague uses. That said, I will still discuss number 5 as part of the analysis of the continuum.

As I have repeatedly mentioned, the continuum of reference in 2sg anata in Japanese is here only applied to the reported speech environment. Thus, the discussion in the following subsections not only elucidates generic and vague uses of 2sg anata but also increases our understanding of the different ways in which current speaker attitudes in reported speech may be encoded in the languages of the world (Spronck 2012: 87). The phenomenon is also in line with the claim made by Spronck and Nikitina (2019) that the reported clause often constitutes a grammatical environment in which indexical elements can have functions that they do not have outside this environment.

6.3.1 ‘Anyone’

Starting the discussion with focal point 1 ‘anyone’ in the continuum, Example (19) from the data examined demonstrates the case where anata refers to people in general, a use which this study refers to as a generic use of anata. In Example (19), the speaker is talking about a guaranteed minimum pension and insisting that the government should at least ensure that its citizens receive a pension to meet certain necessities.

(19)

Yahari	kokumin=ni	nattokushi-te-mora-u	niwa,	watashi-tachi=wa
still	citizen=dat	satisfy-ger-give.me-npst	in.order.to	1-pl=top
kore	to	kore	to	kore	dake=wa	anata =no	kurashi=o
this	and	this	and	this	only=top	2sg=gen	life=acc
sasae-mas-u	yo	to	iu	no	de-nak-ereba…
support-pol-npst	ip	quot	say	nmlz	cop-neg-cond
‘Still, in order to obtain citizens’ understanding, (we) must say “we will support everyone’s (lit. your [ anata gen]) life at least for this, this, and this …, for sure”.’
(Tomoko Abe at the Welfare and Labor Committee meeting, November 14, 2012)

In Example (19), anata is referring to people in general, namely, all Japanese citizens. The part watashitachi wa kore to kore to kore dake wa anata no kurashi o sasaemasu yo ‘we will support everyone’s (lit. your [ anata gen]) life in terms of at least this, this, and this, for sure’ is a style of direct speech in an imaginary situation where a government spokesperson is addressing a citizen/citizens using anata ‘you’. If it were an indirect speech version, anata could be replaced with hitobito ‘people’, kokumin ‘citizen(s)’, or minasan ‘everyone’, and without any social-deictic markers such as the polite form of the verb and the interactional particle, the sentence would become hitobito no kurashi o sasaeru ‘(we) will support people’s lives’ or kokumin no kurashi o sasaeru ‘(we) will support citizens’ lives’. This use of anata is read as number 1, ‘anyone’, in the continuum of 2sg anata in reported speech.

In Example (20), the speaker is criticizing the Prime Minister’s activity on a particular day, referring to the Prime Minister’s Daily Update. The speaker asserts that the Prime Minister’s private dinner with the Chief Cabinet Secretary in a hotel restaurant clearly reveals corruption. The speaker accuses the Prime Minister, saying that the Prime Minister cannot get away with this behavior just because he thinks no one is watching. Then, the speaker cites a part of the Japanese version of an old Chinese proverb Ten shiru, chi shiru, ware shiru, ko shiru ‘Heaven knows, earth knows, I know, (and) children know’, meaning that ‘An evil deed is certainly discovered’.

(20)

Shiryoo=o	shita	made	yon-de-itadak-u	to	wakar-u
document=acc	bottom	up.to	read-ger-receive.hum-npst	cond	understand-npst

n	des-u	ga,	ko	shiru	ware	shiru…	tsumari,	anata	mo
nmlz	cop.pol-npst	but	children	know	1sg	know	in.other.words	2sg	also

shit-te		i-mas-u		yo	ne,	watashi	mo	shit-te	i-mas-u	to
know-ger		be-pol-npst		ip	ip	1sg	also	know-ger	be-pol-npst	quot

iu	koto	na	n	des-u.
say	thing	cop	nmlz	cop.pol-npst
‘If (you) look at the whole document, (I) think (you) can see (it). “Children know, I know …”. In other words, it means “Anyone (lit. you [ anata ]) knows, right? I know, too”.’
(Jiro Ono at the Financial Committee meeting, March 13, 2012)

In Example (20), to explain the old Chinese saying, the current speaker constructs a dialogue as if an imaginary speaker is talking to an imaginary addressee, using anata. The use of anata here again leads to a generic reading, namely, number 1, ‘anyone’. In the equivalent indirect speech version, anata would be replaced by minna ‘everyone’ or daredemo ‘anyone’, and the quoted part would be minna mo shitteiru/daredemo shitteiru, watashi mo shitteiru ‘everyone knows/anyone knows; I know, too’.

Sidnell (2021: 539), in his study of Vietnamese conversation, describes part of a conversation where the speaker integrates proverbial wisdom with generic references. He states that, as “these references are embedded within a generic proposition”, they purport to “articulate a timeless truth” about the topic under discussion. Creissels (2013: 59–60) also states that in Mandinka, which belongs to the Mande language family, the non-specific use of the second-person singular pronoun is commonly observed in generalizations about humans in certain situations, often in conditional sentences, and this use of the second person pronoun is widely found in proverbs. In Example (20), the speaker links the Prime Minister’s activity with a generic proposition by using the Chinese proverb. In so doing, the speaker attempts to generalize what he thinks is wrong as a universally obvious fact, “a timeless truth” (Sidnell 2021: 539). The generic use of anata, which is only possible in reported speech, is perfectly suited to this context, and it is the politician’s technique to make their argument convincing.

6.3.2 ‘You, as a representative of a type of person in a hypothetical situation’

This section and the next two sections (6.3.3 and 6.3.4) discuss the vague use of anata. This section analyzes the case where the use of anata is classified into number 2, ‘you, as a representative of a type of person in a hypothetical situation’.

Observe Example (21). In this example, Prime Minister Noda is talking about the effort of the Ministry of Education, Sports, Culture, and Science (the Ministry of Education hereafter) to establish more systematic support for victims of bullying in schools.

(21)

Moshi	ijimer-are-te-ir-u	okosan=ga	ir-u-to-sur-u-naraba
if	bully-pass-ger-be-npst	hon.child=nom	exist-npst-cond-suppose-npst-cond
sore=wa	moo	anata =wa	hitori	de=wa-na-i,	anata =o
it=top	anymore	2sg=top	alone	cop.ger=top-neg-npst	2sg=acc
mamor-u	hito=ga	ir-u	tasuker-u	hito=ga	ir-u
protect-npst	person=nom	exist-npst	help-npst	person=nom	exist-npst

soo	shinji-te	kanarazu […]	kooiu	soodan=no	madoguchi=ni
so	believe-ger	certainly	this.kind.of	consultation=dat	service=dat
zehi	soodan=o	shi-te-itadak-i-ta-i	to	iu
by.all.means	consultation=acc	do-ger-receive.humb-inf-want-npst	quot	say
fuu-ni	omo-i-mas-u.
like-adv	think-inf-pol-npst
‘If there are children who are bullied, then (they should) believe, “You ( anata ) are not alone anymore. There are people who will protect you ( anata ) and help (you)”, and, by all means, I would like (them) to come to this kind of consultation service counter.’
(Yoshihiko Noda at the Special Committee on Comprehensive Reform of Social Security and Tax, July 18, 2012)

In this example, anata is not referring to an identifiable specific child, but its referent is ‘a certain child/children’ in a hypothetical situation described as ijimerareteiru okosan ‘a child/children who are bullied’. The utterance in a reported imaginary scene constructed by the current speaker in the form of a direct speech, anata wa hitori dewa nai, anata o mamoru hito ga iru … ‘You are not alone, there are people who will protect you …’, is as if person X is talking to child/children Y. However, if one were to use indirect speech, this anata could be replaced by non-specific terms, such as sooiu ko ‘that kind of child’ or sooiu kotachi ‘those kinds of children’ rather than specific sono ko ‘the child’. The low specificity of the referent of anata provides the interpretation that the referent is understood as a representation of children who are bullied. Thus, this anata is understood as number 2, ‘you, as a representation of a type of person in a hypothetical situation’.

In Example (22) below, the speaker is describing a general tendency in Japan’s aging society, where problems arise due to the combination of a decrease in the number of children and an increase in the percentage of unmarried people.

(22)

Kyoodai-suu=no	genshoo,	mikon-ritsu=no	jooshoo	niyotte,
sibling-number=gen	decrease	unmarried-rate=gen	increase	due.to
kazoku=no	naka=de,	anata =wa	hatarai-te,	anata =wa	kaigo=o
family=gen	inside=loc	2sg=top	work-ger	2sg=top	care=acc
shi-te,	to	i-u	yoo-na	yakuwari	buntan=ga
do-ger	quot	say-npst	like-lnk	role	sharing=nom
muzukashi-i,	sooshita	jookyoo-ni	gozai-mas-u.
difficult-npst	that.sort.of	circumstance-loc	exist.pol-pol-npst
‘Due to the decrease in the number of siblings and increase in the percentage of unmarried people, (we are) in a circumstance where it is difficult to distribute roles in families, like “You ( anata ) work” and “You ( anata ) care for the elderly”.’
(Yasuko Oshima at the Financial Committee meeting, March 22, 2012)

The first anata and the second anata are referring to different individuals. However, in each case, the use of anata is an absolute specification of an imaginary interlocutor in a hypothetical situation. Again, if one were to use indirect speech, these instances of anata in (22) could be replaced by more generic phrases such as hitori ‘one’ and moo hitori ‘another’, as in hitori wa hataraite, moo hitori wa kaigo o shite ‘one works and another takes care of the elderly’. This use of anata in (22) is indicating the imaginary interlocutors in many different families who would take on the roles described. Thus, this anata is also interpreted as number 2 in the continuum, namely, ‘you, as a representative of a type of person in a hypothetical situation’.

As noted earlier, this type of anata, such as in Examples (21) and (22), is used in hypothetical situations; thus, anata is abstracted and plays a role in representing a hypothetically conceptualized type of person who is in the described imaginary situation. If anata refers to an unidentifiable individual in a non-hypothetical situation, the specificity becomes higher, which I discuss in the following section.

6.3.3 ‘You, a non-identifiable individual’

In Example (23), the use of anata is classified as number 3 ‘you, a non-identifiable individual’. The speaker is talking about his visit to a debris-processing site after the 2011 tsunami disaster in Japan. He is impressed by the inner strength of the people there because they are working rather talkatively even though they have lost their families. He relates the following scene involving constructed dialogue at the site:

(23)

Aa,	anata=mo	sooiu	tsunami=o	keiken.shi-mashi-ta	ka,
oh	2sg=also	that.sort.of	tsunami=acc	experience-pol-pst	q
anata =mo	des-u	ka,	aa,	anata =wa	obasan=o	nakushi-te,
2sg=also	cop.pol-npst	q	oh	2sg=top	aunt=acc	lose-ger
watashi=wa	miuchi=o	kooiu	fuu-ni	nakushi-te	to,
1sg=top	relatives=acc	this.sort.of	way-adv	lose-ger	quot
soko=de	iroiro-na	hanashi=ga	deki-ru	n	des-u	ne.
there=loc	various-lnk	talk=nom	can.do-npst	nmlz	cop.pol-npst	ip
‘“Oh, did you ( anata ) also experience that sort of tsunami?”, “You ( anata ) too?”, “Oh, you ( anata ) lost your aunt and I lost my family in this way”, in that situation, (they) are able to talk about a range of things.’
(Yutaka Kumagai at the Accounts Committee meeting, August 27, 2012)

In Example (23), each of these three uses of anata refers to different individuals although none of them are identifiable. The quotation marker to has scope over all the reported utterances in which anata appears. The current speaker is not attempting to ‘report’ the exact conversations held among the victims at the site. Instead, he is substituting anata, that is absolute specification of the interlocutor, for the term the original speaker would really have used. In an original situation of this kind, it is reasonable to assume that the original speaker would probably have used the addressee’s name with an appropriate title, which is the normative way for adult acquaintances to address each other. However, the current speaker’s main intention is not to quote the exact conversation he heard, but to describe a representative scene he encountered at the tsunami debris site. The current speaker makes the referent ‘vague’ by lowering his commitment to the specificity of each individual. This is because the specific identity of each individual in the described scene “doesn’t matter” (Givón 1993: 226) in this context. Anata in this example is interpreted as number 3, ‘you, a non-identifiable individual’. It is part of this politician’s narrative style in his attempt to describe the scene vividly, and yet in a generalized way. He achieves vivid description by creating a deictic shift and hence a different mental space, as if the dramatis personae are talking to each other. At the same time, he succeeds in making the scene relatable and generalizable by lowering the specificity of each referent with the use of anata.

In another example, the speaker, MP Mori, is criticizing the Reconstruction Agency regarding its bureaucrats’ irresponsible and unkind treatment of business owners in Fukushima, who have been struggling with the process of applying for financial support. Mori had been told that the bureaucrats had explained everything to the business owners in Fukushima, to ensure there were no misunderstandings. However, on pressing the bureaucrats to give her concrete and specific examples, one of them revealed that they actually did not know the details. They excused themselves, claiming that it was the job of some other section, not theirs. Mori goes on to say:

(24)

Anata	sakki=wa	madoguchi=de	chanto	setsumei.sur-u	yoo-ni
2sg	just.then=top	counter=loc	properly	explain-npst	like-adv

shidoo=o	shi-te-ir-u	to	it-ta	ja-na-i	des-u	ka,
instruction=acc	do-ger-be-npst	quot	say-pst	cop-neg-npst	cop-npst	q

demo	jissai	genjitsu-ni	dooiu	serifu=de	it-te-ir-u	n
but	actually	reality-adv	how	wording=inst	say-ger-be-npst	nmlz

da,	memo	shi-te	mot-te-ki-te-kur-e	to	it-tara
cop.npst	memo	do-ger	bring-ger-come-ger-give.me-imp	quot	say-cond

madoguchi	betsu=no	shokan	da	to	i-u	n	des-u.
reception	different=gen	jurisdiction	cop	quot	say-npst	nmlz	cop-npst
‘(I said) “Didn’t you ( anata ) say just now that (you) gave (them) proper, explanatory instructions at the counter?” But when I said “exactly what wording did (you) actually use to instruct them? Write it down and bring the memos to (me)”, (they) said, “the service counter is under a different jurisdiction”.’
(Masamo Mori at the Accounts Committee meeting, April 13, 2012)

The current speaker is reporting what happened during her interaction with the bureaucrats. She is representing her conversation with one of them in (24) in a direct speech style. However, the specific identity of this reported interlocutor is not her main concern. The current speaker, again, lowers her commitment to the specificity of the referent, making it ‘vague’. If it were indirect speech, anata would be replaced by sono hito ‘that person’ without a specific antecedent.

At the beginning of Section 6.3, I noted that Figure 4 presented in Section 6.2 indicated prototypical focal points and not discrete categories, and that the boundary of each focal point was not absolute. Example (25) below is categorized as number 3, but Example (26) below can be seen as a borderline case. In (25), the speaker initially explains what he often observes on the street. In this sense, he is recollecting one of the scenes he has witnessed in the past. However, as the narrative continues, the speaker merges what he really witnessed with what he constructs: a hypothetical scene which could commonly happen, shown in (26). First, observe Example (25).

(25)

Yoku	yar-are-te-ir-u	tokoro=o	mi-te-i-mas-u-to,	usetsu
often	do-pass-ger-be-npst	scene=acc	see-ger-be-pol-npst-cond	right.turn

kinshi.	[…]	Ittan	teishi=o	kuruma=ga	oodan.shi-ta	tokoro	de
prohibition		temporal	stop=acc	car=nom	cross-past	place	at
hai	to	bakari-ni	kamo=ga	ki-ta	yoo-na	katachi=de
yes	quot	as.if-adv	easy.game=nom	come-pst	like-lnk	form=ins
keisatsukan=ga	hyokkori	de-te-ki-te	kore=de	anata =wa
police.officer=nom	unexpectedly	come.out-ger-come-ger	this=ins	2sg=top

ihan	des-u	yo	to.
violation	cop-npst	IP	quot
‘A time (I) often see cases of (people) getting caught by (police), (is at) a no right turn. Just after a car crosses a stop intersection, as if saying “Yes, easy game!”, a police officer suddenly appears (and says), “you ( anata ) just violated (traffic regulations), for sure”.’
(Satoshi Ninoyu at the Audit Committee meeting, September 3, 2012)

At this stage in the narrative, the scene can be interpreted as one which the speaker has often witnessed. The speaker’s initial utterance, Yoku yarareteiru tokoro o miteimasu to ‘A time (I) often see cases of (people) getting caught by (police)’, adds factuality to the subsequent utterance and enables the constructed dialogue to be interpreted as something that the speaker actually witnessed, with the referent of anata understood as someone he really saw. In this sense, this anata is interpreted as number 3, ‘you, a non-identifiable individual’, being used in a non-hypothetical situation. However, when the speaker continues the narrative and criticizes the general issue of ambiguity in police officers’ judgments in dealing with violations of traffic rules, what the speaker actually witnessed in the past and what could commonly happen are conflated in his speech. Observe Example (26).

(26)

Keisatsukan=wa	anata=wa	tomat-te-i-nakat-ta	to.	Iya	watashi=wa
police.officer=top	2sg=top	stop-ger-be-neg-pst	quot	no	1sg=top
kanarazu	tomar-i-mashi-ta	to.	Sayuu	mi-mashi-ta	to.
definitely	stop-inf-pol-pst	quot	left.and.right	look-pol-pst	quot
[…] Soo.suru.to	chotto	kiyowa-na	hito=wa	keisatsu=ni
then	a.little	timid-lnk	person=top	police.officer=dat
make-te-shima-u. […]	Ki=no	tsuyo-i	hito=wa	keisatsukan=ga
lose-ger-end-npst	mind=gen	strong-npst	person=top	police.officer=nom

mata	sore.ja	doozo	it-te-kudasa-i	to	koo
again	then	please	go-ger-give.me.resp-imp	quot	this.way

nar-u	n	des-u	ga,	ano	handan=wa	hijoo-ni
become-npst	npst	cop-npst	but	that	judgment=top	very-adv
muzukashi-i.
difficult-npst
‘A police officer (would say), “You ( anata ) didn’t stop”. “No, I definitely stopped”. “(I) looked left and right”. […] Then, a person who is a little timid would end up just giving in to the police officer. In the case of a person who is strong-minded (though), the police officer will be like, “Okay then, off you go”, but that kind of judgment is very difficult.’
(Satoshi Ninoyu at the Audit Committee meeting, September 3, 2012)

In Example (26), the lack of an additional phrase, such as seen in (25) (Yoku yarareteiru tokoro o miteimasu to ‘A time [I] often see cases of [people] getting caught by [police]’), together with the use of the present tense, as in maketeshimau ‘ends up just giving in’ and koo naru ‘will (finally) be like’, provides the possibility that the constructed dialogue is interpreted as a hypothetical situation. In this reading, the referent of anata is abstracted and can be interpreted as number 2, ‘you, a representative of a type of person in a hypothetical situation’. However, this could well be a scene that the speaker has witnessed in his actual experience, as in (25). In this sense, Example (26) is a case which could be understood in either way.

Examples of focal points 2 and 3, as well as ambiguous cases, such as Example (26), again show that lowering the specificity of the referent by the use of anata is part of politicians’ speech techniques designed to invite the hearers to enter a different mental space and align with the perspectives of protagonists in a constructed dialogue, thus making a specific example generalizable and relatable.

6.3.4 ‘You, as a representative of a collective entity’

This section discusses number 4 in the continuum, ‘you, as a representative of a collective entity’, which is the case where the referent of anata is a collective entity (rather than a type of person in a hypothetical situation, as in number 2 on the continuum), such as a nation, a city, or an institution.

In Example (27), the speaker is talking about Hoppoo ryoodo mondai ‘the Northern Territories dispute’ between Japan and Russia and asking the Foreign Minister to seek support from the United States.

(27)

Zehi,	kono	kooshoo=no	kekka=o	teinei-ni	seijitsu-ni	soshite
by.all.means	this	negotiation=gen	result=acc	polite-adv	sincere-adv	and

anata =no	tasuke=ga	hitsuyoo-na	n	da	to	i-u	koto=o
2sg=gen	help=nom	necessary-lnk	nmlz	cop.npst	quot	say-npst	thing=acc
Amerika=ni	i-u	koto	da	to	omo-i-mas-u.
U.S.=dat	say-npst	thing	cop	quot	think-inf-pol-npst
‘By all means, (explain) the results of this negotiation properly and sincerely, and then I think (you) should say “(we) need your ( anata =gen) help” to the U.S.’
(Kuniko Inoguchi at Diplomatic Defense Committee meeting, July 26, 2012)

The speaker states that Japan should say anata no tasuke ga hitsuyoo na n da ‘(we) need your help’. Although in the English translation ‘your’ would probably be thought of as having plural reference to the people or government of the United States of America, in Japanese the use of the singular anata makes it clear that this is to be addressed to an individual, such as the American ambassador or the Secretary of Defense. In indirect speech, anata is replaced by Amerika ‘the U.S.’, a collective entity, nation.

There are cases where anata is used in a possessive construction, as in anata no X (your X) or anata no tokoro no X ‘your place’s X’ and the whole phrase refers to a collective entity, as Example (28) below. In Example (28), the current speaker is talking about what happened to the Kawasaki Social Insurance Hospital, when it was suddenly announced through the news that it would be privatized. Because of this, its employees have been confused and anxious.

(28)

Tokoroga	totsuzen	nyuusu=de	anata=no	tokoro=no	byooin=wa
however	suddenly	news=inst	2sg=gen	place=top	hospital=top
moo	minkan	baikyaku	da	to	shira-s-are-ru	wake
already	private	sale	cop	quot	know-caus-pass-npst	reason

des-u	yo.	[…] Sono	fuan	kara	taishoku=o	kiboo.sur-u
cop.pol-npst	ip	that	anxiety	from	resignation-acc	wish-npst
hito=ga	fue-te-ir-u.
people=nom	increase-ger-be-npst.
‘However, suddenly through the news, (the employees) have been notified, “your ( anata =gen) hospital will be sold and privatized”. Due to that anxiety, people who wish to resign have increased.’
(Tomoko Tamura at the Health, Labor, and Welfare Committee, March 22, 2012)

The current speaker’s use of anata or the entire phrase anata no tokoro no byooin ‘your hospital (lit. your place’s hospital)’ is, of course, not referring the current addressee’s hospital but to that of the addressee in the reported situation, the Kawasaki Social Insurance Hospital, as a collective entity. This anata can also be interpreted as number 4 in the continuum, namely, ‘you, as a representative of a collective entity’.

6.3.5 ‘You, an identifiable, specific individual’ (referential use)

Number 5, ‘you, an identifiable, specific individual’, in the continuum of reference of the 2sg in Japanese is the only referential use in the continuum. It will be helpful to compare this use of anata with the use of kimi in the earlier Example (18), which also refers to a specific, identifiable individual. As discussed with regard to that example, the current speaker’s intent is clearly to refer to a specific person, in that case, he himself, addressed by his teacher, due to the use of a typical and socially expected term which he could well be quoting verbatim. The use of a socially expected 2sg item, like kimi in Example (18), would be interpreted as number 5 ‘you, an identifiable, specific individual’. It will often be interpreted this way although it will often be unclear whether the reporting speaker is quoting verbatim or simply constructing a socially expected term.

When the reporting speaker uses anata, on the other hand, it is also possible (albeit unlikely in this case) that s/he is quoting verbatim with reference to an identifiable individual. This pronoun is far from a socially expected one. As noted earlier, generic and vague uses, discussed so far in 6.3.1–6.3.4, dominate in the data, constituting 92 % of the uses of anata in reported speech in this data set, while reading number 5 is rare (8 %). The small number of examples that do occur with this reading includes the use of anata which functions to reject the given relationship or refer to a depersonalized addressee (reflects the use of anata outside the reported speech context, discussed in Section 6.1). It is, of course, not unexpected to have some examples of the use of anata which could possibly constitute a verbatim quote of the reported speaker. Even though it is infrequent outside reported speech in ordinary conversation, speakers do often quote the words of others in parliament, where its use is considerably more frequent (see Section 6.1).

Apart from that, however, a peculiar usage only possible in reported speech is also observed. It shows an interesting modification of the original speaker’s utterance, which is strongly related to the current speaker’s self-presentation in the current speech situation. Although examples of this reading are very small in number, rather than excluding them from my analysis, I briefly discuss two examples because they provide us with interesting implications for further research on the creative use of person-reference terms in reported speech. Example (29) is an utterance of Prime Minister Yoshihiko Noda.^[18] The topic is the pension system.

(29)

Tashika	shigatsu	dat-ta	desh-oo	ka.	Yosan	iinkai	de
wonder	April	cop-pst	cop-tent	q	budget	committee	loc

go-shitsumon	itadai-te,	anata	wa	ima	nani	ni	hait-te-ir-u
hon-question	receive.humb-ger	2sg	top	now	what	dat	join-ger-be-npst

ka	to	iu	go-shitsumon	de	ano	toki	naikaku	kyoosai	to
q	quot	say	hon-question	cop.ger	that	time	cabinet	mutual.aid	quot
o-kotae-shi-mashi-ta	keredomo …
hon-answer-hum-pol-pst	but
‘(I) wonder if it was April. At the budget committee meeting, I humbly received an honorable question “Which one have you ( anata ) joined?” and, at that time, I humbly answered “(It is) the Cabinet Mutual-aid Pension”, but …’
(Yoshihiko Noda at the Integrated Reform of Social Security and Tax Committee meeting, July 18, 2012)

In this example, the addressee in the quoted part is the Prime Minister himself, who was asked about his membership of a mutual-aid pension fund. In a real situation, there is little doubt that the original speaker would have used one of the conventional second person reference terms (Kim 2012). The original utterance at the Budget Committee in April is recorded in the Diet Minutes and the original speaker indeed used Soori ‘Prime Minister’ as in (30).

(30)

Ima	Soori=wa	iryoohoken	nani=ni	hait-te-ir-u
now	Prime.Minister=top	medical.insurance	what=dat	enter-ger-be-npst
ka	gozonji-des-u	ka,	go-jishin=de.
q	know.hon-cop.pol-npst	q	hon-self=ins
‘Do you (lit. Prime Minister) (honourably) know which medical insurance (you) have your (honourable) self?’
(Mitsuru Sakurai at Budget Committee meeting, April 4, 2012)

In the original situation, in addition to using the socially expected address term Soori ‘Prime minister’, the speaker also uses the suppletive respect form of ‘know’, gozonji, with the polite form of the copula, desu, and attaches the honorific prefix go- to jishin ‘self’. However, in the dialogue that Noda constructs in (29), all honorific elements are stripped away, and the term he chooses to put in the mouth of the speaker in the situation he is reporting is 2sg anata, which does not index any social hierarchy between the speaker and the addressee but merely specifies the second person entity. In so doing, Prime Minister Noda represents himself as a ‘mere addressee’ in the recollected scene. The parts of his speech outside this quoted part are full of humble expressions, even though he is speaking as the Prime Minister, such as go-shitsumon itadaite ‘humbly received an honorable question’ and okotae shimashita ‘humbly answered’. In fact, Noda’s speech style has been described as an “epitome of modesty” (Yonezawa and Jarkey 2019), integrating a wide range of humble honorific language. Portraying the original speaker as addressing him as anata in this construction can be regarded as part of Noda’s humbling speech strategy, by which he presents himself as someone who does not wish to stand above others. As Tannen noted, “construction of the dialogue represents an active, creative, transforming move which expresses the relationship not between the quoted party and the topic of talk but rather the quoting party and the audience to whom the quotation is delivered” (Tannen 1989: 109). Noda’s creative use of anata to refer to himself within reported speech certainly takes into account the presence of the audience in parliament and the wider public as part of his politeness strategy, which is an important aspect of his self-presentation.

Another example is an utterance by MP Eto, who is talking about his recent U.S. visit and meeting with an assistant to the President and Chief of Staff. It is assumed that the original conversation was in English or through an interpreter. In Example (31), a question in English, presumably something like ‘Which party are you in?’, is reproduced as follows.

(31)

Tokorode,	anata=wa	nani-too	da	to.	Jimintoo	da	to
by.they.way	2sg=top	what-party	cop	q	Liberal.Democratic.Party	cop	q
iw-are-tara	gakkaris-are-mashi-ta	kedo.
say-pass-cond	be.disappointed-hon-pol-pst	though
‘“By the way, which party are you ( anata ) in?” (they asked me). When (they) were told that it was Liberal Democratic Party, (they) were disappointed, though.’
(Taku Eto at the Agriculture, Forestry and Fisheries Committee meeting, March 21, 2012)

The current speaker, MP Eto, uses anata to refer to himself. Here, rather than constructing the reported dialogue by adjusting to Japanese social rules in a direct speech style (such as Eto-san wa nanitoo desu ka ‘Which party are you [lit. Mr. Eto] in?) the use of anata conveys a sense of a direct translation of English you. In other words, the quoted part entails a flavor of English conversation, with which the current speaker succeeds in expressing that the dialogue was achieved in an ‘egalitarian’ manner. As in Prime Minister Noda’s use of anata to refer to himself as addressee in reported speech, this use seems to function as part of this politician’s self-presentation.

7 Conclusions

This study has provided clear evidence for generic and vague uses of the 2sg pronoun anata in reported speech in Japanese. In most cases in the data, anata in the reported speech environment does not refer to an identifiable individual but rather to people in general or to a vague referent who is low in specificity. I have provided five focal points related to the degree of specificity of these uses within the continuum of reference of Japanese 2sg in reported speech (Figure 4).

Previous studies (e.g., Kitagawa and Lehrer 1990) have surmised that generic and vague uses of 2sg do not exist in languages with an open-class pronominal paradigm because personal pronouns in these languages are assumed to be inevitably social-indexical and “too loaded with semantic and pragmatic information […] to be generalized or used impersonally” (Kitagawa and Lehrer 1990: 756). However, anata has the unique property of absolutely specifying the second person without indexing any social elements of the interlocutor. It functions merely to express person-deixis. This analysis thus provides a straightforward answer to the question of why anata allows generic and vague uses in a similar way to 2sg pronouns in languages with closed pronoun classes. In this sense, while Kitagawa and Lehrer’s (1990) statement that there is no 2sg pronoun which could be used as a generic second in languages with an open-class person-reference system, such as Korean and Japanese, is found to be incorrect, their general claim that generic seconds are possible only if the relevant 2sg pronouns primarily function to mark person-deixis is consistent with the findings of this study.^[19]

That said, the use of anata discussed in this study differs in an important way from the generic and vague uses of 2sg that have been discussed to date in other languages in that its occurrence in reported speech can very rarely be taken as a verbatim report. This is because, as a term of address and interlocutor reference, mere specification of the second person entity without indicating any social elements of the interlocutors is highly restricted due to the sociocultural norm of relationship acknowledgment in Japanese communication. On the other hand, the reported speech environment, which serves to prompt a deictic shift and provides an alternative mental space, allows the reporting speaker to exploit the utility of the semantic emptiness of anata by putting it into the reported speaker’s mouth in a creative manner. As the data set for this study constitutes parliamentary debates, the analysis has rendered visible generic and vague uses of the 2sg anata as part of politicians’ art of speech, particularly in attempting to frame their narratives as generalizable in an attempt to create persuasive argument in their parliamentary talk.

The study also found that a very small number of occurrences of anata in reported speech refer to specific and identifiable individuals. These included a reflection of actual uses outside the reported speech situation but also included creative uses which show an interesting politeness strategy employed by the quoting speaker. In these creative uses, the speaker took into consideration not only the relationship between the quoted conversation parties but also that between him/herself as the quoting speaker and the addressees to whom the speech was delivered, as a strategy for managing self-presentation.

On the basis of this analysis, I would like to note two points as having implications for further research. First, among languages with an open-class pronominal paradigm, if a language also has a 2sg which is semantically simple and neutral, this 2sg may have generic and vague uses in certain environments such as reported speech. Second, languages with an open-class person-reference system may have understudied areas of the speaker’s innovative use of person-reference terms in reported speech environments which reflect the reporting speaker’s attitude rather than simply a typified social relationship. In a language with an open-class person-reference system, the abundance of terms to address/refer to a person means that the speaker has a wide range of choices to exploit as they construct or manipulate reported speech to achieve their interactional goals.

The investigation of these issues will shed further light on some fascinating aspects of the nexus of semantics, pragmatics, and the structures of a language.