3 Honorifics: The cultural specificity of a universal mechanism in Japanese Barbara Pizziconi 3.1 Introduction Honorifics have been the object of linguistic enquiry for much longer than the broader notion of ‘politeness’ itself,1 but in spite of such long-standing investigations they pose an epistemological challenge. As grammaticalised linguistic devices typically interpreted as markers of deference to people of higher status, they are commonly conceived as exhibiting a ‘core’ deferential meaning, coded in the very honorific form, therefore constant across instances of use and always presupposed. Deferential forms are assumed to selectively mark literal or metaphorical distinctions of rank or horizontal distance, and therefore we are left to explain how meanings other than social ranks and roles – e.g. affective stances such as aggression or hypocrisy, intimacy, affection – can be routinely conveyed by the same forms. For example, in studies of business discourse, use of honorifics can be taken to modulate power differentials, through the signalling of deference to such power, or the exercise of power itself; in studies on gendered discourse, they can be taken to index, via manipulation of the core meaning of deference, positions of subordination or authority. Issues of power are, of course, of central importance for performing gender or professional identities, but assuming the expression of a deferential intent in any occurrence of an honorific form carries the risk of over-attribution or blatant misinterpretation. Japanese studies of honorifics have identified a number of regular uses of honorific forms that have traditionally been categorised as ‘secondary’ meanings. The Japanese linguist Hatsutaroo Ooishi (1986 [1975]: 65ff.) for instance lists the following: reverence (agame 催栽殺), respect (sonchoo 梺臍), distance I wish to thank the organisers of the 4th International Symposium of the Linguistic Politeness Research Group: ‘East meets West’ July 2008, Budapest: Francesca Bargiela-Chiappini, Dániel Z. Kádár, and Huba Bartos for their invitation to deliver the plenary talk which led to this chapter, and the editors of this volume for their comments and suggestions. A thank you goes also to Miriam Locher and Heiko Narrog for reading and making very perceptive observations on a first draft. Any remaining inaccuracies are my own. 45 46 Barbara Pizziconi (hedate 笹冴咲), formality (aratamari 催皿財撮晒), dignity (igen 懦乱), grace/good manners (hin’i 仂溺), irony (hiniku ), contempt/disdain (keibetsu ), endearment (shin’ai 粫). Others mention ‘polite hypocrisy’ or ‘nasty politeness’ (inginburei 紮纎靖 ), or a ‘speech-beautifying’ (bika 杢) function. Thus the range of so-called secondary meanings includes social as well as affective meanings. They are thought to be ‘derivative’ from an original meaning of genuine deference,2 and are explained as the effects of pragmatic inferences or historical developments, but effectively parasitic to an invariably ‘coded’ deferential meaning. However, established as it is, such a view of honorific meaning is in striking contrast with the understanding that politeness is an emergent property of interaction – notably most popular in recent scholarly research on English, a non-honorific language (Watts, 2003: 153; Locher, 2004; Arundale, 2006; and on honorific languages: Agha, 2007) – and it is also at variance with thinking in pragmatics and discourse analysis, which tends to emphasise the underspecification of linguistic meanings (cf. Clark, 1996: 50ff.; Levinson, 1983: 48ff.), and the contribution of specific contextual and interactional conditions to meaning interpretation or communication. This chapter addresses the theme of this volume as a whole by asking the question: in which ways do the lavish honorific systems such as those that can be observed in many (East) Asian languages require different theoretical analyses of the phenomenon of linguistic politeness? In the light of the developments in politeness theory mentioned above, and in view of the weakness of analyses that have to assume deferential attitudes even when they do not exist, I submit some observations in support of a re-conceptualisation of the meaning and function of honorific forms. I will be claiming that deference is indeed not directly coded in any linguistic form, not even in ‘specialised’ forms such as honorifics, and that the dynamics of utterance interpretation – including their contextual and discursive foundation – are not substantially altered by the use of honorifics or non-honorific devices. This chapter exemplifies these arguments by focusing on Japanese. I will begin by reporting briefly on the contribution of one scholar, Hatsutaroo Ooishi, and his view of honorifics as abstract representations of interactional schemata; I will then elaborate by drawing extensively on the work of Asif Agha on indexicality, and explain how Japanese honorifics can be accounted for in terms of their indexical properties. I will conclude by showing how the same, arguably universal, inferential mechanisms hinging on principles of indexicality are shared by honorifics and other devices such as speech acts (as a type of polite strategy utilised extensively in honorific-poor languages such as English). This approach underscores how indexicality enables infinite variability in use and interpretation, diachronically and synchronically, and suggests that variability is the norm rather than the exception, inherent in the nature of linguistic signs. Honorifics 47 I will then ask why, in spite of so much inherent variability, ‘typical’ interpretations of honorific meanings exist at all. Thus, rather than dismissing stereotypical representations tout court as deceiving and unhelpful, this study homes in on them and explores their role in folk as well as scientific models of politeness. The discussion of Japanese honorifics that I present here aims at addressing an issue that, I think, is relevant to all enquiries on politeness, irrespective of the language in question. Ever since Sachiko Ide’s (Hill et al., 1986; Ide, 1989) popularisation of the term wakimae (‘discernment’), and maybe even in spite of some of Ide’s comments on this subject, the field has grown used to discussing politeness systems in ‘binary’ terms, i.e. those with very developed honorific repertoires on the one hand (that force the speakers to make obligatory choices in virtually every utterance) and those without (that resort to various linguistic strategies, and leave their speakers more ‘leeway’, as it were, in the expression of deference, the marking of rank, etc.). The analysis in this chapter provides a motivated argument for maintaining that the existence of elaborate honorific repertories cannot account for qualitative differences between ‘Eastern’ and ‘Western’ politeness, and can only be investigated in terms of the reflexive models of behaviour and cultural discourses at work in specific social groups. Before I present my analysis, I wish to make explicitly clear that, by refocusing on honorifics rather than the broader notion of politeness, I do not intend to argue in favour of a formalist theoretical position that neglects the importance of the interactional context in constructing meaning, or that disregards the perspective of participant-meaning in favour of sentence-meaning – quite the contrary. I am interested in re-examining the properties of honorifics in terms of how social persons (who not only inhabit social positions but also affective stances, and bring their personal histories and their ideologies to bear on interpretation), rather than model persons (who inhabit ideal and decontextualised worlds), exploit such properties to interpret and achieve specific interpersonal goals. 3.2 A brief introduction to Japanese honorifics and their taxonomy The term commonly used in Japan to refer to the phenomenon of linguistic politeness is (among some others) keigo. This is a Sino-Japanese compound of two characters (谺 ), the first signifying ‘to respect, to honour’ and the second ‘language’ or ‘speech’, and commonly rendered in English as ‘polite language’ or ‘honorific language’. Although this term has been used also to refer broadly to what we, today, call ‘politeness’, technically speaking, it designates only those features of politeness that in Japanese are grammaticalised, i.e. honorifics. Table 3.1 shows a broadly accepted classification of honorifics in Standard Japanese3 (Tsujimura, 1992: 227), but it is in fact a hybrid of different proposals; many others have been presented, with minor variations of subcategories based 48 Barbara Pizziconi Table 3.1 Synoptic table of keigo categories and terminology Focus of deference Addressee honorifics taisha keigo 梏 谺 Referent honorifics sozai keigo ㊌閭谺 Honorification type Beautifying expressions Bika hyoogen 杢↘ " Higher-rank expressions jooi hyoogen 濯溺↘ " Lower-rank expressions kai hyoogen 琢溺↘ " Linguistic form ‘Polite forms’ teineigo 瀧柧 " ‘Deferential forms’ sonkeigo 梺谺 ‘Humble forms’ kenjoogo mostly on the syntactic behaviour of the items in questions, or with variable emphasis on semantic or syntactic criteria (cf. Wetzel, 2004; Pizziconi, 2004). First of all, it is interesting to note that the labels used for classes of keigo forms (‘Polite’, ‘Humble’, ‘Deferential’) correspond to the general metapragmatic functions of such forms. Such labels, that with the customary variation across individual scholars’ definitions we find in most treatises on politeness, are also, thanks to their relative semantic transparency and their use in school syllabi, common currency in everyday use among educated speakers. This point is important because it allows speakers (and linguists) an enhanced reflexive awareness about what we can call the ‘typical’ function of such forms, but also reinforces the belief that such typical function is their only function. I will come back to this later on. The main criterion of categorisation here is in terms of focus of deference: the language possesses some specialised forms that indicate whether the target of honorification is the addressee, i.e. the interlocutor, or the referent, the person/object/event talked about. So apart from the special case of when an addressee is also a referent these two categories have the potential to systematically distinguish deference towards referents who are present in the speech event, and those who are not. Within the category of referent honorifics we have a further distinction between forms of ‘object exaltation’ (‘Higher-rank expressions’ in Table 3.1, used to refer to prototypically respected others) or ‘humilification’ (‘Lowerrank expressions’, to refer humbly to self or in-groups). An additional category traditionally recognised in Japanese linguistics is that of bikago 杢 or ‘beautifying language’ (Tsujimura, 1992: 98ff.). Bikago is taken to represent the speaker’s concern for the quality of the expression itself (corresponding to the type of concern that in English motivates the choice of ‘ladies/gents’ over Honorifics 49 ‘toilets’, ‘to rest’ over ‘to sleep’). Because this is normally triggered by the relation to the addressee (in other words, the audience to speakers’ talk), it is considered by some a type of addressee honorific, but because it is also affected by considerations regarding the content matter it can also be considered a type of referent honorific. It has therefore been given a trans-categorial status, not formally included in either addressee or referent honorifics, but sharing some of their features (as well as forms). This ‘concern for the quality of the expression’ also suggests the speaker’s concern for his/her own image, and hence the use of bikago is commonly construed as a mark of ‘refinement’ (or at least a claim to a refined image). Although a certain user ‘image’ (or speaker ‘typification’, as I will call it later, following Agha, 2007) is a kind of by-product of the use of any honorific category (cf. Ide, 2005 on honorifics as expressions of ‘dignity’ and ‘elegance’), bikago is not strictly associated with other-honorification, and therefore more than other categories, it has mostly been discussed as an index of personhood. As for the range of honorific forms included in these various categories, it is indeed remarkable, involving suppletive forms as well as morphosyntactic manipulations, which affect virtually all grammatical categories: pronouns, nouns, adverbs, verbs, etc.4 As Table 3.2 shows, in terms of linguistic repertoire, Japanese honorifics are very rich, and indeed much richer than those available to most European languages. Other Asian languages – Javanese, Lhasa Tibetan, Korean – have similarly complex and developed systems, which certainly justifies at least the assumption of a ‘special sensitivity’ of Eastern cultures to interactional matters. Grammaticalised systems can be understood as the fossilisation of constraints on the hearers’ search for relevance (Lapolla, 1997), or a progressive ‘fixation’ of constraints on interpretation. They are therefore the result of an evolutionary process of ‘short-circuiting’ procedural information of particular salience. Languages with rich honorific systems exemplify cases in which the salience of meanings pertaining to social interaction has generated semiotic systems dedicated to the signalling of social relations. However, it has been noted that there is no obvious or necessary link between social and linguistic structures (Irvine, 1998: 51). The existence of a royal court does not predict the existence of an honorific system (although of course there will exist means to express deference), and vice versa, relatively low social stratification can be observed in cultures that make use of honorifics. Moreover, a close look at the area prototypically associated with rich politeness systems (i.e. Asia) reveals that there is huge variation in the scope of lexical repertoires, the productivity of grammatical structures, and how the various categories are realised in the language. Japanese, for example, does not possess grammaticalised bystander honorifics, as one finds in Javanese. Korean’s marking of humble honorifics is much less developed than those of 50 Barbara Pizziconi Table 3.2 List of some honorification devices Category Examples Lexical devices (suppletive forms) Pronouns for the first person include various alternants associated to various degrees of formality: I = watakushi, watashi, atakushi; boku, ore, etc. + other nouns functionally interpreted, in interaction, as first person reference: e.g. kochira (‘over here’)5 Nouns wife = kanai, tsuma, waifu; okusan, okusama; kamisan, etc. (see Section 3.3.2) Adverbials today = kyoo (unmarked), honjitsu (formal) Predicates to eat = taberu (unmarked); kuu (informal); itadaku (humble); meshiagaru (deferential), etc. copula = da (colloquial), dearu (written), desu (neutral/formal), degozaimasu (formal) Morphosyntactic devices Affixes: Nouns and adjectives Predicates honorific prefix o/go- : go-kazoku (family); o-isogashii (busy), etc. o-V-ni naru (deferential); o-V-suru (humble); V-masu (polite)6 Japanese, etc. Within macro-cultural commonalities in the marking of interactional matters, language-/culture-specific interests have developed in different functional domains. In the next section, I will look at the contribution of one particular Japanese scholar to the study of keigo, which has some interesting implications for an indexical view of honorific devices. 3.2.1 Ooishi (1986 [1975]): The interactional schema Like other politeness theorists in the Japanese school, Hatsutaroo Ooishi attempts to provide schematic characterisations of keigo in order to tease out the multilayered meanings that the honorific register can convey, but his particular take provides some important insights into the fundamental distinction between the denotational and interactional levels that are compounded in an Honorifics 51 honorific expression. His main contribution lies in the analysis of what I will call, following Agha (2007: 46), “interactional schema”. Ooishi (1986 [1975])7 starts from an analysis of single honorific forms, subdivides them in roughly the same categories I have presented in Table 3.1, and then describes each category in terms of a diagram of the relationship between various interactional roles that that category defines. Thus, what he describes are the mappings of different relational patterns indexed by each honorific category. The simplest one is the pattern for the addressee honorifics (or Speaker– Hearer axis). Figure 3.1 shows the denotational information, i.e. that the grammatical/semantic role of ‘Subject’ is occupied by the addressee (or hearer), as well as the vector of deference, showing that the Subject is the target of the speaker’s deference.8 A (+Subject) S Figure 3.1 Addressee honorifics: relational pattern between S and H 歳撮載採腰 Ikimasu ka? Are you going [+POL]? (1) 諜踰昨圄銓踰崎載腰 Kyoo wa doyoobi desu. Today is [+POL] Saturday. In the pattern for referent honorifics (i.e. the categories of honorifics related to the person, action, state or object talked about) we have three participants, as shown in Figure 3.2.9 Similar diagrams obtain for the deferential and humble forms. B Ref (−Subj) A Ref (+Subj) H S S H Figure 3.2 Referent honorifics: relational pattern between S–H–Ref: (A) deferential forms, (B) humble forms (2) 彦 栽最皿堺菜雑三腰 Sensei ga irassharu. The teacher is going [+DEF]. 栽 躬踰撮崎作 彦 作 才 ""椰犀載三 (Watashi ga) ashita made ni (sensei ni) o todokesuru. (I) will deliver [+HUM] it (to the teacher) by tomorrow. 52 Barbara Pizziconi In the case of the deferential forms in A, the interactional schema tells us that the (grammatical/semantic) role of subject is currently occupied by the referent identified as ‘teacher’, and additionally that that referent is the target of the speaker’s deference. In the case of the humble forms in B, the target of the speaker deference is again the teacher, but here the role of subject is occupied by the speaker (these are the expressions in which the deferential effect on the referent is obtained by a ‘lowering’ of the speaker, rather than a direct ‘raising’ of the referent). Referent honorification can be even more complex, when more than one referent is involved. In cases when the action of some referent affects, in some way, the state of a second referent, the pattern shown in Figure 3.3 obtains. Ref2 S H Ref1 Figure 3.3 Referent honorifics: deferential forms, relation S–A–Ref1–Ref2 (3) 痞栽 躬踰撮崎作 彦 作 才椰犀載三腰 (Kare ga) ashita made ni (sensei ni) o todokesuru. (He)1 will deliver it [+HUM] (to the teacher)2 by tomorrow. Predicate honorification triggers this complex interactional diagram even in cases when the elements overtly expressed in an utterance are only minimal, and everything but the predicate is omitted: ‘(he)’ll deliver (it) [+HUM]’ can index the relationships between three participants: the speaker, the subject (referent 1) and a target of deference (referent 2). Ooishi does not particularly elaborate on the theoretical significance of this analysis – he does not attempt a generalisation about what such scenarios mean in terms of a qualification of keigo properties, and his discussion is merely descriptive. He only claims that the patterns he described constitute a characterisation of the features of treatment in terms of “attitude and viewpoint” (1986: 93). But these diagrams are important because they highlight honorifics’ crucial indexical properties: interactional and deferential indexicality. 3.2.2 The indexical view of honorifics Honorifics are a type of deictic sign (Fillmore, 1971; Levinson, 1983), and as such they anchor various dimensions of the speech event to the current utterance. They provide an indexing of (the speaker’s evaluation of) social relations, and map interactional roles at the time of utterance onto social locations, i.e. statuses (Shibatani, 1990). As we saw above, such deictic mapping involves Honorifics 53 two dimensions, one that links referents to semantico-grammatical categories (subject, object, etc.), and one that links them to social locations. The first mapping is one of interactional roles. Honorifics do not identify specific persons, but identify the roles these persons occupy in the course of the current utterance (for example, the referent is/is not the speaker, etc.; cf. Agha, 2007: 315). The second kind of mapping is that of deference: this specifies a source of deference (which Agha calls origo), and a focus of deference. These mappings therefore provide templates of interactional roles, or “interactional schema” (Agha, 2007: 46), which are independent from, but ‘superimpose’ themselves on, and link, the denotational meaning of the utterance and the situation of utterance. The quality of this mapping is abstract (as in the figures above) and a-contextual, but it imposes structure on the participation framework, from which event-specific meanings can be derived. The Table 3.3 illustrates this schematically with the verb ‘to do’, and its honorific and non-honorific correlates. The use of honorific forms therefore includes information about: • denotational reference, i.e. semantic reference to the action type, plus lexico-grammatical information (verb, activity/stativity, animate subject, etc.…); • the interactional schema, i.e. information about the position of the grammatical subject within the interactional roles at the time of utterance: speaker, addressee, referent; • deferential indexing, i.e. honorific forms identify the focus of deference, as well as the origo, or source of deference. For example, in the humble verb itasu the origo of deference is (by default) the speaker.10 Agha calls tables like Table 3.3 “tables of categorial default”. This is intended to indicate that while the meaning of ‘deference’ often represents the ‘default’ reading, this can be altered by co-occurring signs that are either Table 3.3 Categorial defaults for verb ‘to do’ (suppletive forms) Denotation Lexeme: ‘to do’ Interactional indexing Interactional role (grammatical subject is …) Deferential role (grammatical subject is …) nasaru [+HON] +verb, +active –speaker Focus suru +verb, +active ±speaker Undetermined itasu [+HUM] +verb, +active +speaker Origo 54 Barbara Pizziconi congruent or incongruent with that reading. But which co-occurring signs provide such specifications? These need not be co-occurring linguistic forms. Any kind of evidence available to a speaker will be read in conjunction with the honorifics: the assumed identity of the participants, the nature of their relationship at the point of utterance, physical indicators of the speaker’s attitude, as well as, of course, other co-textual elements. In the following section, I will discuss how the available evidence regarding the relationship between participants becomes a discriminating factor in establishing the actual (contextually derived) meaning of honorific forms. 3.3 Interpreting honorifics: Constructing multiple scenarios Imagine we overhear only one utterance in a conversation between two parties that we do not know anything about, and that we cannot even see. 躬踰撮崎作椰犀咲最財冴犀撮載採 ashita madeni todokete itadakemasu ka deliver BEN tomorrow by [+HUM] [+POL] non-past INT ‘Could you kindly deliver (it) by tomorrow?’ denotational level interactional level What we can infer based on what we hear and on a ‘default’ reading of the honorific elements is, arguably, something along these lines. At the denotational level, that some form of delivering (an act of object transfer by a human subject) is being discussed. At the interactional level, that the delivery has not happened at the time of utterance (non-past form of the verb, and a tense deictic – tomorrow – that can only be interpreted in relation to the present moment of utterance); and that it is being asked of someone (a deictic of (non-) factuality: note that only in context is this assigned the meaning of request) who is not the speaker (at this point still underdetermined as to whether second or third person). But we also get a deferential sketch: typically, that someone (typically, the speaker) is paying deference to someone else (another indexical, whose reference can only be worked out depending on the known or presupposed participation framework and identity of participants, and is therefore, at this point, underdetermined). Based on the denotational information and interactional sketches indexed by all the above elements, the overhearer can construe a possible scenario of usage and conclude that the one that most typically matches this utterance’s meaning is, for example, that of two colleagues negotiating the delivery of some document or other object, in which the one doing the request is in a subordinate position, and therefore ‘packages’ this request in the appropriate amount of deference. Alternatively, one could conclude that there is no vertical relation Honorifics 55 involved, the two colleagues are more or less of the same rank but not very intimate with each other, and the honorific is motivated by horizontal distance. In this ‘typical’ scenario, the indexing of deference projected by the deferential form itadaku and the polite form -masu are also arguably congruent with other signals of deference such as the question form (or indirect request), or the use of a benefactive auxiliary. The readings of ‘deference’ or ‘horizontal distance’ are provisional: other elements (other ‘signs’) are needed to disambiguate the intended ‘meaning’. Note the use of the key word ‘typically’ in the analysis I have just proposed. The meaning of deference is not an absolute and invariable meaning, invariably coded as such independently from contexts, but happens to be, for reasons that have to do with socio-cultural discourses underlying the use of those honorific forms, the default reading. I will come back to this below. Let us now imagine that we then turn the corner, and are able to observe that the speaker is a grown-up woman who is addressing a child, conceivably her daughter. The metapragmatic incongruence of the fact that the subject of the act of delivering and therefore the target of deference is occupied by a child, moreover arguably the speaker’s own child, forces a reanalysis of the ‘meaning’ of the honorific: by social convention, children do not qualify as recipients of social deference11 – they generally do not enjoy higher status, and one’s children are not socially distant referents. So we are forced to construe a different scenario: perhaps a kind of ‘role-playing’ (the daughter playing the company’s department head, the mother playing section head). In this case, the use of a humble verb, rather than conveying deference as such, effectively construes the change of ‘footing’ (i.e. it construes a different capacity, or ‘principal’, Goffman, 1981). This non-default reading, in which the honorific forms index a virtual professional persona, emerges from this particular scene, and will disappear as the scene changes. Our knowledge of the ‘typical’ deferential reading, the ‘typical’ interactional schema and the incongruence of the child in the role of deference target prompt a remodelling of the scene we are observing, and a different reading of the honorifics’ meaning, which is not ‘deference’. Let us consider one final possibility. As we turn the corner, we discover that the two speakers are a married couple we know. The interactional schema instantiated by the honorific forms is again partially incongruent with the evidence we have now acquired: the target of deference is a husband, whom we know is habitually, or typically, addressed by the wife with more intimate forms (other couples could have different linguistic habits). We therefore must reassign a suitable meaning to the honorifics: perhaps the meaning of affective distance. This suggests a rather different possible scenario, one in which the documents that need delivering are perhaps divorce papers, and in which the wife is displaying a frosty, or even hostile stance towards her soon to be exhusband. Notably, this effect is partially a function of one’s ideology (another 56 Barbara Pizziconi contextual variable) vis-à-vis the social status of wives in relation to their husbands and what is consequently the appropriate register for them to use, and it is subject to variable interpretations by those who subscribe to different ideologies (more on this in section 3.3.2 below). The point is that since honorifics are deictic forms, their interpretation is anchored to other variables, or co-occurring signs of the event. The only inherent property of the deictic is a constraint on some interactional schema, but the resulting meaning is an emergent property of that specific utterance, uttered by those specific participants, in conjunction with other signals relating to these participants’ identities (gender, age, attire, facial expressions, tone, preceding and subsequent portions of discourse, etc.) that are either congruent with the default deferential schema (in which case the default interpretation is ratified) or are not (in which case alternative non-default interpretations are sought). These emergent meanings are volatile: they disappear once the context is modified. 3.3.1 ‘Default’ interpretations As we have seen, even for one and the same overhearer, the stereotypical interpretation of this utterance in the absence of other cues is just one of many possible interpretations. Additional cues add further constraints to its interpretation. This suggests that we should abandon a view of honorifics that sees deference (a categorial default) as a constant inherently ‘coded’ meaning and an obligatory interpretation, re-examine the implications of their indexical and deictic properties, and consider how ‘defaults’ are troped on to generate different effects (Agha, 2007: 324). Let me summarise here with Agha’s definition of categorial defaults. Honorific categories, for example addressee or referent honorifics, represent different “categorial text-defaults, i.e. differences in the baseline sketch of deference implemented by a form in the absence of co-textual effects to the contrary” (Agha, 2007: 316; my emphasis). The diagrams by Ooishi that we examined above specify such different ‘baseline sketches’ of deference for some Japanese categories of honorifics, that is, that someone (focus of deference) is deferred to by someone (origo of reference). Moreover, “to say that deference indexicals have categorial deference effects is to say that honorific expressions regularly specify particular contextual variables as default foci of deference relative to context. Categorial effects are gradiently defeasible by co-occurring signs in a variety of ways” (Agha, 2007: 322). Therefore, the emergent effects can be partially different from the categorial default effects of the lexeme in question. Although native users of Japanese may easily typify categories of honorifics in terms of metapragmatic judgements such as ‘sonkeigo is used to show respect to someone’, or ‘teineigo is used to speak politely to someone’, the terms ‘respect’ or ‘polite’ must be taken as stereotypical formulations of default interpretations, Honorifics 57 always subject to reinterpretations in contextualised instances of use, based on computations of all co-occurring signs (Agha, 2007: 322). The deference reading may be the default reported norm of usage, but not necessarily a constant norm of actual use, nor an inherent or invariant property of honorifics. A stereotypical definition is one of many possible constructions, possibly the most frequent, conventional, or standard. While frequency is a statistical notion that can be measured by some sort of large sampling of actual data, the criteria of conventionality and standard are a function of the particular register(s) that the overhearer of this utterance is familiar with, as a language user with a specific social background (say, educated, upper-class, professional, living in the twentieth century, etc.). A hundred years ago the sentence in (3) above exchanged between husband and wife would possibly not have been construed as antagonistic, because the subordinate position of women in society may have justified the ‘deferential’ reading. Alternatively, in upper-class environments in which a more formal register is the norm, the sentence may be perfectly appropriate as a marker of refined demeanour, and therefore would not trigger any particular ‘affective’ effect. Other overhearers may in principle interpret the utterance rather differently (they may not make sense of it at all, they may think it is an extremely pompous or old-fashioned phrasing, etc.). So, when saying that certain metapragmatics functions are stereotypical readings, we imply not only that they will be construed differently under different contextual circumstances, but also that such readings are not necessarily uniform or invariant across social groups. Thus, while different social groups can recognise patterns of usage of this or that form (and partly because some normalised patterns can be recognised, which ensures mutual coordination), they may have contrasting, even conflicting ideas about the stereotypical meanings that these forms realise. Since speaker ideologies participate in the ‘reading’ of a recognisable sign (through evaluations of what is conventional or standard, which depend on community-specific practices), we can expect such ideologies to affect its usage, and therefore multiple normative patterns to exist simultaneously. In the next section I will briefly elaborate on this question, and then return to discussing whether the strength of these normative patterns can be said to be more significant in languages with grammaticalised honorifics. 3.3.2 Ideologies of use and registers Let us examine the question of normativity first through a set of lexemes for /wife/. These are relevant to issues of honorification in the sense that their distribution is in part constrained by status and group-consciousness. The same reasoning applies to grammaticalised honorifics, but I start from the analysis of this term for its relative lexical transparency. I will not describe their individual usage in great detail for reasons of space, but will simply note general points.12 58 Barbara Pizziconi Table 3.4 Categorial defaults for ‘wife’ terms Denotation: Syntactic/ semantic schema +noun +human +female +married Interactional indexing Deictic term: /wife/ Translation and gloss Interactional role Deferential focus –in-group + okusan/okusama 恟斎燦疾恟┞ Lit. ‘inner area of the home’; as the honorific suffix -san/-sama indicates, in Standard Japanese normally used for wives of other men +in-group – kanai 朧敏 Lit. ‘inside the home’; in Standard Japanese, the normative term to refer to one’s wife +in-group – waifu 軸伺侍 Loanword from English: ‘wife’; used by younger men (also attested: furao 侍汐賖 [Frau] from German) +in-group – tsuma 愼 Arguably the most common term for ‘wife’13 +in-group – nyooboo 悁冐 Lit. ‘woman of the palace’: high-ranking female servant of a nobleman +in-group – gusai 粢愼 Lit. ‘stupid wife’, archaic: humble form used to display politeness to addressee First of all, we should note that there are no categorial defaults for number (singular/plural), or grammatical/semantic category (subject/non-subject). These terms all share the same semantic/syntactic information. In terms of interactional roles, they can be described on the basis of the general pragmalinguistic rule that distinguishes between terms typically used for out-groups (marked by the honorific suffix -san) and those used for in-groups only. This provides a first indication of how speakers position themselves vis-à-vis some object of reference and their interlocutor. Okusan/okusama makes the referent immediately identifiable, by marking it honorifically and therefore excluding reference to the speaker’s own wife. Kanai obtains reference by the inverse Honorifics 59 reasoning. We could argue that this general principle is widely recognised across Japanese speakers, and constitutes a relatively stable pattern of usage. However, other components of these indexical signs are subject to contrasting social evaluations, which introduce some divergence across more localised (group-specific) patterns of use. The etymological glosses transparently show a role differentiation which today can be read as sexism: women belong ‘inside the house’, are associated with the inner quarters of a home. So speakers who believe the place of a woman to be the home find the normative term kanai fully appropriate, and speakers who resent that association, believe in gender equality and consider themselves progressive and cosmopolitan, would resort to, for example, waifu. ‘Old-fashioned’ (and older) users would not object to a very humbling term (which indirectly ‘raises’ the addressee) such as gusai, especially in some stylised register such as letter-writing.14 So while the models of use may be partially congruent (reference to in-group or out-group can be relatively straightforward via the marker -san) it is easy to see that these terms are not socially neutral, and that different social groups (for example feminists) will evaluate them differently, challenge them and attempt to transform their normative status. These terms constitute part of the linguistic weaponry in the social contest over the role of women in the family and society, and come to be associated with different ‘types of speakers’ and different types of ideologies (about the family, gender roles, political orientation, etc.).15 The relationship of specific lexical items with underlying ideologies of use is not at all static or straightforward, and cannot be evinced directly from a list of repertoires. The social meaning of a linguistic form at a certain point in time – whether used ‘conservatively’ or ‘innovatively’ – hinges on reference to its prior history and recognisable patterns of use, but it can show unpredictable developmental trajectories, as the history of just one of the terms, okusan/ okusama lit. ‘inside the house’, demonstrates. This term, presently considered by many to be the standard term of reference for others’ wives – referred initially to the (legitimate) wives of daimyoo 彖琳 (high-rank military leaders) and court nobles (Inoue, 1999: 66), and it was correspondingly associated with high status. From the late Meiji period (1868–1912), a period of intense social reformation and modernisation, it progressively spread to married women in urban households, i.e. wives of the rising class of officials, businessmen and teachers. Such referential shift was possibly sustained by its deferential meaning (a working woman would use it in reference to a non-working, wealthy woman, whose role was effectively less powerful but socially more prestigious; Ueno, 1987: 79). While the use of the term in a deferential sense hinged precisely on its pre-modern usage, this progressive shift in reference, denoting an increasing number of ‘wife types’, led to an effective weakening of its deferential meaning. This trend continued after the war, when in spite of some resistance, even female heads of business households (who would have previously 60 Barbara Pizziconi been referred to with the term okamisan, lit. the ‘head/boss’) also came to be routinely called okusan/okusama (Inoue, 1999: 66). In more recent times, the use of this term (together with the term kanai, lit. also ‘inside the house’, used for one’s own wife) has been openly protested against by women’s groups (Gottlieb, 2006: 131), who arguably object to its etymology (its conceptual meaning) rather than the original deferential denotation, which is now rather less prominent. Meanwhile, there is another use of the term okusan: the term for wives of other men is being reanalysed as a more respectful and preferable choice for reference to one’s own wife by speakers who profess an egalitarian credo and who are reluctant to use the humbling kanai. Paradoxically, we have two groups with a comparable political agenda (women’s emancipation) who reanalyse the available linguistic forms with opposite results: while one group rejects the form okusan as sexist and demeaning altogether, the other revalues it as a preferable alternative to the normative one. Incidentally, the latter strategy represents a breach of the overarching principle of in-group/out-group distinction (as the suffix -san should not be used for one’s own wife), often quoted as a characteristic Confucian legacy and a defining feature of the Japanese polite ethos. Although through different interpretative paths and with different outcomes, both solutions represent a challenge to the ideology underlying the use of normative terms, rest precisely on users’ familiarity with some established registers of use, and intend to be emblematic of a new (feminist, egalitarian) speaker ‘type’. A novel use of the form okusan when kanai is expected can lead to variable interpretations on the part of an interlocutor, ranging from offence (the hearer may feel that the speaker’s wife is being elevated above him/herself in the evoked interactional sketch), to appreciation of the speaker’s consideration of his wife, and his egalitarian intent. Other co-occurring signals (both conceptual and formal) would further constrain such inferences. Such novel interactional effects, as we said above, would not be ‘coded’ in the term used, but last only until some of the relevant contextual variables are modified, and then disappear. Some of these novel formations will not last long, will be stigmatised, and eventually disappear. However, some effects can ‘catch up’ and become defaults for entire categories of users (say, young people, feminists, etc.) through processes of enregisterment (Agha, 2007: 145, 278). Larger-scale processes can result in permanent diachronic changes incorporated in standard varieties (like the addressee honorific suffix -masu, said to derive from the humble referent honorific mairasu). But the potential to become “emblems of social difference” (Agha, 2007: 293) is not only available to lexically transparent terms like these. Any sign can participate in speaker indexicality, including non-lexical honorific morphemes, as we can see in a case of ‘ambiguous’ reading reported by Ooishi (1986 [1975]: 108). Honorifics (5) 61 Scene: two girls at work: 輜踰昨悌 桟才定擦菜撮菜財腰 Kinoo wa kaisha o o-yasumi shimashita. Yesterday I took [+HUM] a day off work. I explained above how honorific forms specify some contextual variables as default foci of deference. In this case (a humble form) the focus of deference is undetermined: is it the company? Is it the addressee? And why would deference need to be paid to either to convey the rather ‘neutral’ propositional meaning above? In the absence of any additional cues that justify these readings (or a known scenario), the only element that can aid interpretation is the sketch provided by the honorific (humble) form o-V-suru, which not only points to a target of deference (unhelpful in this case) but also the source of deference, the speaker. This is why cases like these (an instance of bikago, cf. Table 3.1 and explanation in Section 3.2) are interpreted primarily as indices of speakerhood. The very use of keigo denotes a speaker who is conversant with its grammar, lexicon and complex norms of usage, one who has mastered a socially valuable tool, a precious commodity not available to all (this belief is indeed confirmed by the plethora of manuals on ‘correct’ use). Use of keigo is interpreted as a ‘barometer of the speaker’s level of education’ (Inoue, 1999: 19) and as emblematic of the speaker’s refinement (Ide, 2005: 60), and users of bikago could conceivably use these forms to portray themselves as knowledgeable. However, this use of keigo as a form of ‘designer accessory’ is socially controversial (in the same way as mini-skirts or hoodies), and Ooishi (who attributes this ‘anomalous’ use mostly to ‘young women’) notes it for its social markedness. Different users exploit the properties of honorific forms in ways that are not universally validated, based on ideologies of language use not uniformly distributed in society, and reflecting variable orientations to the social value of such forms. This generates variable metapragmatic judgements, such as attributions of sophistication, affectedness, ignorance, etc. Variable evaluations are not only observed in cases like (5) above, where the deferential sketch is unclear. Much anecdotal evidence exists of social friction linked to rather routine uses of honorifics; Inoue (1999: 14) even recounts a number of murder incidents whose triggers, if not causes, were incompatible norms of linguistic behaviour between speech participants. A recent survey of 1,975 people by the Agency for Cultural Affairs (Bunkachoo, 2007) asking Japanese speakers whether they found other people’s manner of speech questionable reported that 71 per cent did indeed; 80 per cent replied that they thought the language was in a state of disarray (midarete iru 遅傘咲最三), and of these, 67 per cent declared that keigo was the major cause of such a deplorable state, an increase of 11 per cent from the previous survey in 2002. 62 Barbara Pizziconi The unequal distribution of keigo knowledge and the coexistence of multiple ideological practices (cf. Mills, 2003 for class- and gender-based ideology; Okamoto, 1997: 799 for age-based ideologies; and Miller, 1971: 602ff. for more general political ideologies) entail that different types of speakers, while partially recognising the indexical value of honorific forms, find enough ‘leeway’ in the interactional sketches that honorifics diagram to produce divergent and innovative patterns of usage (which does away with a requirement of full ‘sharedness’ to appropriately understand and use honorifics). 3.3.3 A first summary We can conclude from the discussion so far that honorifics are neither necessary nor sufficient to signal ‘respect’ or ‘deference’ (Agha, 2007: 302). They are not sufficient because unless some other variables presupposed or made available in the interaction permit the assumption of a deferential relationship, contextually they are not necessarily interpreted as indices of respect (cf. the scenarios discussed above, or the case of bikago). They are also not necessary, because deference can be expressed without them. For example, we can show deference by avoiding speaking altogether, by lowering our gaze, by particular attires (cf. wearing black at funerals), by using ‘vague’ language (for example omitting altogether a predicate which would require explicit honorific marking), by using ‘deflected communication’ (i.e. talking to someone by addressing a third party instead; Pizziconi, 2009), by praising or complimenting someone, etc. However, that honorifics are not strictly necessary to express polite attitudes or deference may be interpreted in at least two ways, one of which may be problematic in view of some arguments made in past research on Japanese. In a general sense, it is, I think, uncontroversial that in Japanese it is possible to show a deferent attitude even without explicit honorific marking: the examples mentioned in the preceding paragraph would all apply to Japanese as well. However, some, most notably Sachiko Ide, have argued that there are contexts in which speakers do not have much choice but to submit to the requirements of the situation (or ba 寉), and must use honorifics as prescribed by social norms and expectations. The notion that Japanese speakers must abide by the rules of the situation and do not have the range of expressive choices that is available, for example, to speakers of English, is for Ide (2006: 113) what qualitatively distinguishes a language like Japanese, and is captured by the principle of wakimae 獸宰 (or ‘discernment’). If this is true, then honorifics would be absolutely necessary, at least in some cases. And if so, it can be argued that they would indeed constitute a powerful socio-cognitive constraint on the linguistic (and cognitive) freedom of users of Japanese. The following section will provide a critical discussion of this argument. Honorifics 3.4 63 Politeness in languages with and without rich honorific systems Ide (2006), revisiting the results of the empirical study on Japanese and American polite behaviour presented in Ide et al. (1986), maintains that the responses of speakers of Japanese and English who are asked to match a list of some twenty forms of request (for borrowing a pen) of varying degrees of politeness/ formality, to a range of about twenty ideal addressees in varying degrees of power over and distance to the speaker, show rather different patterns. She calls the Japanese pattern ‘clear-cut’ (my translation, kukkiri gata in the original) and the American one ‘fuzzy’ (bonyari gata), referring to the fact that while Japanese responses cluster in two mutually exclusive areas – forms used only with non-intimate or out-groups (among other things, the speech level markers desu/-masu)16 and forms used only with intimate or in-groups (among other things, plain forms) – American responses are more evenly distributed, i.e. Americans use the same forms with a wider range of addressees. This is indeed a noteworthy result, suggesting that while American speakers enjoy a higher degree of linguistic freedom (within some rather broad constraints they can select a form mostly on the basis of individual intentions, or ‘volition’), Japanese speakers are much more constrained by very particular contextual factors (primarily the in-group/out-group dimension, but also social position, social role, age, gender, degree of formality of the situation, affective and psychological distance, etc.). Once such contextual parameters set the scene and define the corresponding rule, Japanese speakers are said to be forced to abide by it, in order not to sound inappropriate, or in order not to generate unwanted implicatures. If so, this would suggest that social factors constrain Japanese speakers’ awareness of social institutions, social power and social ranks to a greater extent than they do for speakers of English. 3.4.1 The principle of wakimae The term wakimae has been introduced, and widely adopted, in politeness literature in reaction to a limitation of the Brown and Levinsonian model, i.e. its emphasis on rationality and intentionality. But to use this term to characterise qualitative differences between Japanese and other languages, or between languages with and without highly developed honorific systems is, in my view, untenable.17 This is because ‘discernment’ of situations, (including social positions and social norms) is crucial not just in honorific usage but in language use as a whole. Much research on language has explored the rather staggering implications of the truism that ‘context affects language use’ (see Hanks, 1996: 232; Gumperz and Levinson, 1996: 225; Duranti and Goodwin, 1992; as well as most work in cognitive linguistics), and further work on the indexical properties of language is adding depth to this notion (Agha, 2007). In these 64 Barbara Pizziconi approaches, contextual parameters are seen not as mere accessories of utterance interpretation, but as constitutive of it.18 In interpreting utterances, and in particular the social value of utterances, interactional variables (of which presupposed social relations are just one type), are simply indispensable. This applies to the close set of deictics called honorifics, as we have seen above, as well as the verbal strategies of an honorificpoor language like English, as I will show below.19 Speech acts too involve a deictic formulation of a sketch of referents (Agha, 2007: 45). Let us imagine being a junior lecturer in a university, being asked by one’s department head to carry out a job that falls within one’s duties: (6) a. Send me the timetable for the Japan section. b. Could you send me the timetable for the Japan section? c. I don’t suppose you’d be so kind as to send me the timetable for the Japan section? The preparatory conditions for orders include that the receiver has the obligation to do the action, and that the speaker has the right to tell the receiver to do the action. So while ‘Send me …’ presupposes that the speaker has rights/ power over the hearer, ‘Could you …’ and ‘I don’t suppose you’d …’ do not. Each utterance therefore makes apparent to each the speakers’ and hearers’ mutual power relations, and their entitlements to demand or refuse action (see Table 3.5). The social roles of department head and junior lecturer (when indeed these roles are believed to instantiate a hierarchical relationship), together with other circumstantial variables, constrain the choice of the possible strategy in much the same way as, in Japanese, a notion of in-group/out-group constrains the choice of the honorific linguistic form (the speech level markers desu/-masu). Moreover, such presupposed social roles also ‘frame’ the interpretation of the forms’ meaning, much as we have seen with honorifics. So – based on my knowledge and my understanding of academic hierarchies, by all means subjective and disputable – I could judge (6a) as not being particularly considerate to me as a colleague (it construes me as having no rights), (6b) as being Table 3.5 Sketches of power relation between speaker and hearer Linguistic form Strategy S’s right to order action H’s right to refuse Imperative Indirect request Pessimistic statement order enquire about ability question willingness + – – – + + Honorifics 65 acceptable (gives me rights of choice), and (6c) as being ironical rather than polite (it overstates my right of choice). But a change of any of the presupposed social roles, or even other interactional variables (for example, the history of the exchange: if the department head had already requested the timetable and this is just a reminder of previous conversations) naturally changes the resulting interpretation: (6a) could sound no longer bossy, and instead acceptable. The resulting effects (or judgements), such as ‘authoritarian’, ‘moderately polite’ or ‘ironical’, are not coded in the modal verbs, in the proposition’s mood, or a particular formulaic phrasing. These are ‘emergent’ meanings that I can only compute based on my knowledge of the indexical properties of those linguistic forms and my discernment of the type of relationships involved, as well as the history of communication with the head and the circumstances of the request, other co-occurring signs such as his/her tone, facial expression, cotextual elements, etc.20 This demonstrates, in my view, that wakimae, far from characterising one language in contrast to others, is a principle of interpretation of any linguistic sign associated with registers of social indexicality. This suggests that whether the discernment of a social relationship or social status is achieved through an honorific or a non-honorific form is relatively immaterial. Honorifics appear to be merely fossilised (i.e. grammaticalised) forms of the same inferential processes involved in speech act interpretation. Both honorific and non-honorific devices equally evoke interactional sketches and can be associated with ‘types’ of speaker. Hence there is no dichotomy between ‘discernment’ and ‘volition’; all linguistic signs are chosen in accordance with their situational and interpersonal appropriateness (based on a subjective understanding of the social norm that defines what is appropriate), and at the same time are manipulated strategically (Pizziconi, 2003; Ide, 2006: 108ff.; Agha, 2007). The difference between honorifics and non-honorifics is a matter of gradation. In so far as honorific markers are the result of extreme specialisation, i.e. extensively routinised and grammaticalised constraints on (among other things) the indexing of interpersonal relationships, they are more ‘focused’ than speech acts, conventional formulae, or other devices indirectly pointing to such relationships. Although languages differ in the extent to which such processes of interpretation are grammaticalised and in the type of specific dimensions that are grammaticalised, the inferential and indexical processes involved are the same. It is clear that Brown and Levinson’s model overestimated individual intentionality. However, a conceptualisation of wakimae as the ‘automatic’ or ‘obligatory’ selection of a form based on certain situational settings (and its particular importance in Japanese society, Ide, 2006: 72), overestimates the role of normative defaults. Such stereotypical defaults are in fact more stable in situations of ‘extreme’ social and psychological distance or in institutionalised discourse (e.g. discourse in academic or business contexts, or formulaic forms 66 Barbara Pizziconi of greeting; see Dunn, 2005 for an example of how such defaults are troped on in highly formalised and ceremonial wedding speeches to convey different personae), and tend to be very unstable in all other cases, where available scripts of normative behaviour are fewer or fuzzier (e.g. language between schoolmates, Inoue, 1999: 19). The degree of freedom over the choice of a linguistic expression is proportional to the extent to which that expression evokes discernible mappings of interpersonal relationships (and an interpersonal map of power), and to the strength of normative canons. When what we say has interactional consequences (which is effectively all the time), we need to tread carefully, whether choosing the right honorifics, verbal strategies, tone of voice, or even posture and attire, or the very timing of the utterance. However, immaterial as the distinction between honorific and non-honorific systems may be regarding the technicalities of deference expression, from a social (and cultural) point of view the availability of an honorific system is certainly not inconsequential. 3.4.2 Sociocultural relevance of elaborated honorific systems The fact that Japanese possesses easily recognisable grammaticalised devices for honorifics indeed constitutes a powerful constraint on utterance interpretation. I said before that grammaticalisation can be seen as a kind of stabilisation, into specialised morphemes, of some well-trodden inferential paths. All forms (lexical and non-lexical) constrain interpretation: grammatical, obligatory forms develop out of lexical, non-obligatory forms: a kind of inferential short-cut to recursive and salient constraints on interpretation.21 The specific paths that get ‘grammaticalised’ are therefore language and culture-specific meanings, which proficient speakers of that language must express, and hearers must recognise (Lapolla, 1997). In the case of honorific devices, such inferential paths have to do with some salient interactional meanings. Thus, for example, the formal and non-formal markers of the predicate in Japanese (desu/-masu suffixes vs. zero morphemes on plain predicate forms) constrain interpretation regarding the speaker’s relation to the addressee, respectively non-intimate and intimate. These registers come to be stereotypically associated with ‘types’ of participant roles; the formal register is, for example, associated with sets of ‘distant’ addressees like those listed in Ide’s survey of reported usage: “professor”, “middle-aged stranger”, “department store clerk”, etc., even though, in actual usage, other constraints could override this ‘rule of thumb’. Such constraints are co-textual and contextual variables, that can include anything from a shift to more transactional concerns due to the nature of the task (as illustrated by Ide, 2006: 110, similar to what Maynard, 1993: 197 calls a “Low Hearer Awareness Situation”), to the observable attributes of a speaker (as we saw in the scenarios discussed above); these generate a Honorifics 67 variety of non-default effects such as shifts of social persona (Dunn, 2005) or special rhetorical effects (see Okamoto, 2002 on irony), etc. The point is that in spite of much documented variability in meanings and expressive effects, in spite of often apparently incongruous interpretations of the same form (e.g. as an index of genuine deference but also ‘hypocritical politeness’, or inginburei 紮纎靖 ), these rather conspicuous markers lend themselves quite easily to typifications. Once certain forms have become part of a recognisable register (a type of normalised discourse under certain contextual variables), they can be isolated and objectified in metapragmatic comments, and take on a life of their own. They can be talked about, elaborated on argumentatively, support discourses about appropriateness22 and ideological orientations (cf. Miller, 1971). They can be extracted from interactional contexts and made the object of language policies. The first post-war policy statement of 1952, ‘Keigo from now on’ (Korekara no keigo) did just that, and made explicit recommendations about the use of specific forms. It established desu/-masu as the predicate forms to be used to set the ‘basic tone’ of a conversation (taiwa no kichoo 梏 搾嫗 ). While acknowledging that other specific settings may require other forms (such as the dearimasu in lectures, degozaimasu in formal occasions, da in intimate conversation), it explicitly recommended desu/-masu as the standard form for ‘general conversation among members of society’ (cf. Wetzel, 2004: 120, 126), vague as such an a-contextual prescription may be. Fifty years or so later, a review of keigo policy recognised that keigo usage in contemporary Japanese society is far from being homogeneous. So while desu/-masu forms are typically referred to as the ‘default’ polite forms for non-intimate interaction in the standard variety of Japanese, anecdotal and empirical reports of actual usage show a considerable degree of intracultural variability, and indicate the users’ perception of some difficulty in the appropriate choice (Ooishi, 1986 [1975]: 178; Inoue, 1999: 19; Takiura, 2001b: 57; Ide, 2006: 141). The 1994 review noted that social processes of urbanisation, internationalisation, information technology, an ageing population, different perceptions about gender, etc., generate a variety of values and ideologies that make “smooth modern communication” potentially problematic (Wetzel, 2004: 129, 144). Thus, while refraining from specific prescriptivism, it positively recommended the use of ‘appropriate language’ to “foster warm human connectedness”. For Wetzel (2004: 65), such blurry guidelines are a symptom of the academic and political establishment’s troubled response to the conflicting agendas of increasingly diversified political subjects. Normative prescriptivism is not the prerogative of public institutions. Wetzel describes a whole industry of ‘how-to’ advice involving the media, expert academics in varying degrees of popularising roles, educational enterprises, business organisations, etc., all intent on some form of ‘policing’ of keigo (2004: 68 Barbara Pizziconi 66). The 2007 survey of the Agency for Cultural Affairs (Bunkachoo) indicates that one in four respondents does not feel sufficiently confident in their knowledge of keigo,23 which suggests a potential market for keigo ‘business’. The more keigo becomes objectified, commodified and marketed, exploited in the creation of desirable (self-)images, the more the value of this form of symbolic capital is bound to increase, and Ide is certainly right in highlighting its importance in contemporary Japanese society. But to what extent is this value shared by all Japanese speakers? Wetzel and Inoue, referring to previous work on ‘naturalized ideologies’, note that ‘ideological-discursive formations’ … have entered the realm of ‘common sense’ to the extent that no one involved in the interaction recognises or questions their ideological nature. Few Japanese question the norms that underlie the skills and techniques they learn in telephone etiquette or other ‘social manners’ classes; in fact, open concern for prescribed behavior is a hallmark of Japanese society. (1999: 75) Whether this concern for prescribed behaviour is universal is an empirical question, but there is no doubt that this phenomenon is not in itself a product of a traditional or historical predisposition to matters of etiquette. Rather, it is a product of modernity, an anxiety generated by the coexistence of multiple, and at times conflicting cultures, subcultures, countercultures, and the distinct role played by keigo in their demarcation. Commentators’ perception that the conflict between such cultures is hardly ever brought into the open and problematised in public arenas is in itself an index of the strength of the hegemonic power of such naturalised discourses on etiquette. Strategic choices such as those which I reported regarding terms of reference for ‘wife’, or the use of humble honorifics in the absence of a target of deference, as well as any ‘creative’ use of an honorific form regardless of its direction, are, needless to say, never entirely ‘free’. One can choose to refer to one’s own wife with a term previously restricted to other people’s wives in order to reject the sexist implications of the ‘normative’ term, but this has implications for one’s interlocutor (who may feel offended), and is subject to a judgement of the political view it represents. Constraints over such choices derive from the speaker’s political power, e.g. his/her degree of authority in making these views manifest. The choice is even more socially sensitive when it comes to terms of reference for other people’s wives. A misuse in this case not only exposes the speaker’s ideology (which may be subject to negative evaluation), but it may result in potential offence (e.g. one may be happy to replace kanai with the less common tsureai 傘倫最, ‘one’s partner’ – but would this term be sufficiently respectful for referring to others’ wives?). Similarly, the deviant use of honorific markers is always a sensitive matter as it can be indicative of informed defiance as well as ignorance, may or may not be recognised as intended, may be a deceiving indication of this or that ideological stance. Indeed the indexical meanings evoked Honorifics 69 by honorifics and other markers pose a constant challenge to ‘discerning’ speakers, because of their inherent iconism and the complexity of their layering. The more detailed, complex and sophisticated the lexical repertoire of social deixis, the greater the risk of misalignments or mismatches in speakers’ orientation. In this sense, speakers of languages with elaborated social registers are indeed more ‘constrained’ – but more by the scope of linguistic choices (a systemic feature) rather than the social pressure to conform (a sociocultural fact). 3.5 Conclusions The analysis in this chapter suggests that the range of meanings that honorifics can convey in actual instances of use is broader than the meanings stereotypically attributed to them in metalinguistic commentaries by language users (Agha, 2007: 307). Typically, honorifics are understood to be indices of deferential, humble or polite stance; typically they are said to mark vertical distance, but other typical reports involve horizontal distance; however, we have examined one case of honorifics used to convey the affective stance of anger, and by the same mechanisms they can be used to index irony, flattery, annoyance, formality (concern for the situation), ‘hypocritical politeness’ (inginburei), etc. These are only a few of other meanings systematically observed by Japanese linguists, but the range of possible meanings is a function of the specifics of the co-text and context, and therefore, always potentially extendable to new effects. An indexical view of honorifics allows us to have a unified account of a range of different effects that are achieved through them: from textual effects (reference maintenance, manipulation of footing, voicing, etc.) to affective effects (anger, aggression, offence, disdain, endearment), as well of course as default social effects: these turn out to be just particular, stereotypical, effects. Also, we have seen that honorifics form part of the repertoires of social registers. Recognisable patterns in the way types of user orient themselves towards the honorific repertoires (e.g. men or women use such and such a form, liberals or conservatives use such and such a form) transform them into emblems of speakers’ values, demeanour, ideology, which can be used as justifications for social judgements. Thus the indexical approach outlined by Agha and illustrated here in relation to Japanese provides further evidence for rejecting a ‘coding’ view of language and is perfectly compatible with interactionist approaches (that is, that politeness is an emergent feature of interaction, cf. Arundale, 2006; and Watts, 2003) or notions of argumentation (the idea that we speak from disparate ideological positions), and of social differential in the knowledge and control over linguistic resources. I have maintained that indexicality (here, in particular, of interpersonal relationships) governs the interpretation of linguistic forms in the same way 70 Barbara Pizziconi in honorific-rich and honorific-poor languages. Hence, even though from a diachronic point of view honorifics can be seen as sediments of inferential processes (as noted also by Brown and Levinson, 1987: 260), which therefore short-circuit such processes and result in more constrained, conventionalised interpretations, their indexical nature in no way prevents contextually variable interpretations and limitless innovations. I conclude that wakimae, or a principle of discernment of social relations said to govern the appropriate use of Japanese honorifics, is not a sufficient principle for defining any specific feature of Japanese politeness. Nevertheless, honorifics do demarcate social relations in finer detail (and with higher frequency due to their distribution over several grammatical categories) than non-grammaticalised, strategic or inferential devices. Even minor manipulations therefore result in more conspicuous effects. Users of Japanese or other honorific-rich languages are not necessarily more ‘socially sensitive’, but they are likely to be more sensitive to the way in which social relations and social reality are created and transformed by linguistic signs. They are, however, also not necessarily more critically aware of their significance from the viewpoint of the hegemonic contests that are played through them. The sociocultural relevance of the availability of honorific systems should not be underestimated. As isolatable linguistic forms, stereotypically associated with notions of deference, social ranking and demeanour, they facilitate explicit metapragmatic reasoning, the creation of reflexive models of social behaviour, discourses of appropriateness, and even language policies that target issues of morality and civic education.