Power and Insight in Jain Discourse I-IV

In contrast to earlier Jainological emphasis on the unchanging and dogmatic nature of doctrinal Jainism, recent historical-philological and anthropological scholarship focuses predominately on historically changing, syncretic and hybrid features of Jain beliefs and practices, and on the role of agency in the construction of socioreligious identity.^[1] Contrary to culturalist self-images and academic representations, it is widely recognised that the 'differences which separate Jainism from Hinduism and Buddhism … are largely differences of emphasis for all are built from common material' (WILLIAMS (1983: xxii));^[2] and that 'even though Jainism is a distinct religion and not a sect of Hinduism, still it is a fact that in the past [and present] many Jains used to regard themselves as Hindus and were also regarded by others as Hindus' (SANGAVE (1980: 3)). This raises questions about the characteris tic features and the politics of Jain discourse, the principle medium of Jain cultural synthesis.^[3] DUMONT (1980: 210) once stated that the Jains, like any non-Hindu group in India, 'cannot be regarded as independent of the environment in which it is set, as really constituting a society by itself, however strongly its own values push it in this direction.' Yet, whether, or to what extent, the Jains 'derive their raison d'être from their distinctive ways of manoeuvring within a [hierarchical] structure that they share with the whole society' (MARRIOTT (1976: 131)) needs further research.^[4] LAIDLAW (1985), (1995: 95), in one of the few studies of Jain discourse to date,^[5] argues that 'Jain cultural distinctiveness does not rest on rituals or practices in which people are marked as different and counted in or out' but on 'a range of practices and relationships through which Jains participate in Hindu public culture in India, and do so as Jains.' Jain culture is defined as a shared ethical life-style, or 'class psychology', grounded in 'a set of processes and practices which cluster around the ownership, management, funding, and use of property' (LAIDLAW (1985: 147, cf. 349 f.)). Socioreligious group formations beyond the institutions of family, caste and religious trusts are seen as ephemeral and dependent on instrumental processes of strategic mass mobilisation by individual lay leaders. CARRITHERS (1992: 118) studied how in conventional settings Jain public speakers 'create, manipulate, and transform' connections between listeners (śrāvaka), in particular through the narration of religious stories in communal rhetoric.^[6] LAIDLAW's (1985: 55 f.) theory of the Jain 'language game' and CARRITHERS' (1991: 262) work on the 'rhetoric of samāj' both successfully move away from essentialist notions of communal identity. But they achieve this only at the price of recurring to instrumentalist definitions of community formation, disregarding the key dimensions of felt togetherness and shared belief and custom.^[7]

In this essay, I propose to avoid both a priori definitions of socio-cultural identity and instrumentalist theories of community formation by analysing the stated principles of Jain religious discourse itself. I will compare and contrast these principles with the categories of Jürgen HABERMAS' (1980–81), who in his Theory of Communicative Action offers a seemingly non-reductionist interpretation of linguistically mediated processes of socio-cultural synthesis. In contrast to the explicit normative ideals of the Jains, rooted in an ontology of karman, Habermas' theoretical investigation presents itself as a non-ontological reconstruction of regulative ideals implicitly presupposed by all actual human discourse.^[8] Some preliminary remarks on the architecture of his theory are necessary.

Habermas' model seeks to transcend the false alternative of 'community' and 'society'—which still dominates the sociology of Indian religions—by focusing on the relationship between 'lifeworld' and 'system' instead, as differentiated in contemporary modern society. 'Lifeworld', a term imported from phenomenology, is defined as the horizon or context of linguistically mediated communicative action.^[9] It is, in his view, constituted by language and has three components or actor-world relations, in which communications are simultaneously embedded: cultural symbols, social norms, and personal aims. Lifeworlds are conceived as thematic resources for the intersubjective construction of social situations though symbolic or communicative action.^[10] In case of disagreement, situations are ideally defined rationally and consensually, through co-operative processes of interpretation based on the rejection or acceptance of claims of objective truth (cognition oriented), normative rightness (action oriented), and subjective sincerity (person oriented). The limited explanatory scope of the lifeworld perspective conceded, HABERMAS (1981: 180) / (1984–1987 II: 118) defines society as a whole 'simultaneously as a system and a lifeworld': 'societies are systematically stabilised complexes of action of socially integrated groups' (1980–1981: 228) / (1984–1987 II: 152). This definition acknowledges that society is not constituted through symbolic or communicative action alone, but also—and increasingly so—through systemic processes, i.e. the unintended consequences of action and interaction mediated by 'steering media' such as money or power rather than by language. Habermas thus situates the social role of discourse within a theory of differentiation of system and lifeworld. The degree of differentiation determines the extent to which social integration can / must be achieved through symbolic or communicative action alone.^[11] Habermas argues that lifeworld and system perspectives are mutually incompatible. His proposed synthesis (chided as 'eclectic' by his critics) prescribes a systematic alternation of the two perspectives, thus addressing the problem in a similar way as Jain perspectivism. Within this framework, Habermas' contribution to discourse analysis results from a single conceptual move: the substitution of the pivotal concept of subjective 'intentionality' (championed by Weberian, Husserlian or Wittgensteinian variants of interpretive sociology)^[12] with the notion of intersubjective 'communication' (Verständigung), which in his view is the inherent telos of human language.^[13] The intersubjective alternative to conventional subject-philosophical approaches^[14] enables Habermas to criticise empirical discourses of power, based on 'instrumental action', as 'deviations' from an 'original' mode of unconstrained 'communicative action', implicitly presupposed by all interlocutors.

Habermas' characterisation of the constitutive role of implicit idealisations for linguistically mediated interaction by the term 'ideal speech situation' has been widely criticised (in similar ways as Chomsky's 'ideal speech community'), since, by definition, ideal situations are rarely, if ever, empirically encountered, and not even consciously contemplated by the majority of interlocutors.^[15] Though it is a truism that ideals can only influence behaviour if they differ from it, a society which relies entirely on explicit consensus is both a modern utopia and a nightmare, since everything can become problematic under the imperative of rational control in an ideal world of unconstrained intersubjectivity. To avoid the 'cost' of social reflexivity, social life has to rely on traditions, habits, routines and systemic processes (mediated by institutions or markets) which are taken for granted, until questioned. This is recognised by the model. Habermas insists, however, that the fundamental unspoken expectations underlying all social interaction can be analytically reconstructed. Conflict, for instance over values, can only be peacefully resolved if the normative presuppositions of communicative action, such as common interest in the avoidance of violence, are implicitly observed.^[16] Hence, rather than dismissing Habermas' 'utopianism' outright,^[17] it may be more fruitful to ask whether the principles of discourse identified by Habermas function indeed as universal presuppositions of communicative action or are merely one set of possible idealisations amongst many. The question is pertinent, since 'communicative action', according to Habermas, is predicated on the implicit recognition of the values of individual autonomy and equality, and the existence of domination-free social spaces and interactional competencies, which are rarely given in any concrete situation. Does Habermas' model, then, merely impose modern European ideals or is his theory indeed of universal relevance?^[18]

This question can be explored by comparing Habermas' dialogical model of the 'ideal speech situation' with other models of highly idealised speech situations of a similar level of abstraction, such as the Jain theory of speaking, which, at first sight, seems to be predicated on hierarchical, or subject oriented, rather than egalitarian, or intersubjective, normative presuppositions.^[19] From Habermas' perspective, the principles informing hierarchical systems, even if culturally dominant, cannot be universalised, since they themselves are predicated on the principles of communicative action which are (from the perspective of analytical reconstruction) consciously or unconsciously presupposed in all linguistically mediated interaction. Conversely, from a Jain perspective, the ontology of soul, non-soul, karman and the principle of non-violence are implicitly presupposed in all universally acceptable actions. In both cases, moments of 'insight' can be generated through the acceptance^[20] and situational projection of the respective model.^[21] It is debatable whether any comparison between contrasting philosophical or religious models implies a dialectical third perspective which will 'always be more general than the most general postulates of a religion and the most general rules of investigation itself' (PIATIGORSKY (1985: 210)) or is simply an addition without being 'higher or lower' (MURTI (1955: 127)). Frequently cited examples of overarching perspectives that are not predicated on specific comparisons are the dialectic of the categories of reflection (Hegel), the politics of cultural hegemony (Gramsci), or indeed the Jaina conception of a disjunctive synthesis of differences or alternatives (anekānta-vāda), which according to MURTI (1955: 128) is 'more a syncretism than a synthesis'. A non-relativistic scenario is plausible if one model is able to reconstruct another on its own terms in a non-reductive way, or if it improves the other model, without losing information, while the reverse is not possible. In such a case, the analytical superiority of one model over the other must be conceded in principle.

A peculiar feature of Habermas' universal pragmatic theory is that it can only be operationalised by 'reversing step by step the strong idealisations' of the concept of communicative action to approximate the complexity of natural situations. Because most methodological provisions and theoretical assumptions have to be dropped in this process, universal pragmatics becomes, for all practical purposes, indistinguishable from empirical pragmatics, except for the additional conceptual sensitivity 'needed to recognise the rational basis of linguistic communication in the confusing complexity of the everyday observed' (HABERMAS (1980: 444) / (1984–1987 I: 331).^[22] What is gained is the ability to discover different levels of the linguistically represented reality, and communicative pathologies, such as veiled power relationships or systemic distortions of rational communication through the use of generalised media of communication.^[23] For formal pragmatic investigations of South Asian discourse the fundamental empirical question is not whether, for instance, mantras or ritual language can be considered as 'speech acts', or in which sense.^[24] The question is rather, as Richard BURGHART (1996: 301) put it, 'how does non-distorted speech communication take place in hierarchical structures'? BURGHART (1983), (1985) was the first South Asianist to tentatively explore the possibilities of the theory of communicative action for an understanding of religious and political discourse in South Asia. Since his premature death, few advances have been made in rendering Habermas' highly abstract theory fruitful for South Asian studies.

In this essay, I make a fresh attempt in exploring the analytical potential of Habermas' communication theoretical approach for South Asian Studies by contrasting Habermas' discourse ethics, the reflective form of communicative action, with Jain discourse ethics, a reflective form of non-violent action. I will focus particularly on their respective theorisation of the subtle role of power in processes of indirect communication. Habermas opened up a new critical perspective by studying the constitutive role of idealisation in discourse and its exploitation. Alternative approaches, such as FOUCAULT's (1981) and BOURDIEU's (1991a), by contrast, emphasise the ubiquity and institutionalised nature of power. This view exerts a strong influence on current empirical investigations of Jain processes of social self-identification.^[25] In comparing Habermas' and Jain theories of discourse, I pursue three main arguments as far as the principles of Jain discourse is concerned:

1. Despite being differently constructed, the Jain theory of speech plays a similar role within Indian philosophy as Habermas' theory of communicative action does within Western philosophy. Both aim at the integration of a variety of perspectives, proclaim the primacy of morality over truth and logic, and predicate critical analysis of typical speech acts, especially latent strategic speech acts (perlocutions), on idealised normative presuppositions.^[26]

2. The orientation towards, and mastery of, the principles of Jainism generates interactional competencies regarding the non-violent resolution of conflicts, and cognitive distancing effects, which enable competent agents to intentionally create ambiguous symbols (utterances and gestures), and to manipulate identities through the re-interpretation of culturally normative or conventional presuppositions. The same can be said of the cognitive functions of modern theories of communication.

3. The perceived plurivocality or multifunctionality of symbols—one of the main features of syncretism and socio-cultural synthesis in general—is in the Jain case not only a feature of rule-application, or a consequence of external imposition or extrinsic borrowing etc.,^[27] but also a consequence of religious knowledge, which can generate effects of insight qua (re-) interpretation of any given content. From the conventional point of view of communicative action, the principles of Jain hermeneutics produce systematically distorted communication, albeit one that is ideally oriented to salvific rather than material ends.

My basic contention is that philosophy (and logic), whether preoccupied with questions of universal validity, scepticism or pluralism, is always embedded in socio-cultural milieus which it both reflects and influences in varying degrees. Philosophy is always not 'merely philosophy', but a form of social discourse with social functions, manifest or latent. Philosophy does not merely consist in sets of propositions and logical or argumentative procedures but has also, directly or indirectly, pragmatic and expressive dimensions, and presupposes matching lifeforms and institutions for its social recognition.^[28] The comparison between two expressly non-absolutist universalist theories, the theory of communicative action and the Jain theory of language usage, in my view, demonstrates that philosophies are intrinsically connected with a selective range of matching life-forms, while recognising that most socio-cultural milieus are culturally hybrid and contain elements which are universally acceptable. To what extent procedures and contents of 'rational inquiry'^[29] are influenced by and influence social context is a question for empirical research.

The aim of this essay is to outline a new approach for the analysis of religious discourse in South Asia, and to prepare the ground for future critical sociolinguistic studies of Jain discourse. For this purpose key theoretical issues of philosophical pluralism and cross-cultural comparison are explored in a heuristic way. The essay is in eight parts. First, I am going to review the general problematic of Jain syncretism (I) and the existing academic literature on Jain rhetoric and discourse (II), followed by an overview of Habermas' theory of communicative action (III) and aspects of the work of Grice and Brown and Levinson, which will prove useful for operationalising Habermas' theory (IV). To prepare the ground for empirical investigation, I will then propose a typology of characteristic social settings of Jain religious discourse, and discuss their normative implications (V). Thereafter, I analyse the key features of the Jain theory of speech (VI), and of Jain discourse ethics in form of the Jain tetrad of the modes of speech (VII). Finally, I draw some general conclusions by comparing and contrasting the normative presuppositions of the theory of communicative action and the Jain theory of speech, which both in their own way offer critical analytical perspectives on the role of power and violence in human communication (VIII).^[30]

— I —

One of the key arguments of this essay is that Jainism, as a meta-philosophy whose social efficacy is predicated on the systematic reinterpretation of conventional perspectives, constitutes a form of discourse which produces syncretic patterns. At present, syncretism is predominantly understood as a transitional phase within an overall dialectical process of religio-historical development— syncretisation or acculturation—which often involves parallel processes of linguistic syncretism and/or group formation.^[31] PYE (1994: 220), for instance, defines the 'syncretic situation' as 'the temporary ambiguous existence of elements from diverse religions and other contexts within a coherent religious pattern.' He locates syncretism between a 'mere mixture' and a 'coherent mixture' or 'synthesis':

'If coherent mixture, or synthesis, represents the conclusion to a process which is thereby completed, syncretism by contrast is to be understood as dynamically open and indeed patent of resolutions other than synthesis. These might be, in particular, the outright dominance of one strand of meaning by another (assimilation), or the avoidance of synthesis through the drawing a part of the distinct elements and the consequent collapse of the syncretism (dissolution).'^[32]

Earlier, BECHERT (1978: 20–3) had proposed a similar typology of syncretic phenomena, i.e. of 'the different forms in which religious traditions have influenced each other', associating them with particular ideological or cultural systems: (1) The marginal acceptance of single elements (e.g. Jains), (2) proper syncretism 'where elements from different religious traditions gain equal weight' (e.g. Nepal, Bali, Sri Lanka), (3) full integration (Neo-Hinduism), (4) perfect synthesis (e.g. Sikhs). According to Bechert, Jainism (in general) is an example of type one, because its 'essential characteristics' are not touched by the assimilation of new elements. Bechert's assessment of Jainism was probably influenced by BRUHN's (1954: 136) remarks on the lacking 'mixture of traditions' and the prevalence of a combinatorial coexistence of elements from diverse traditions in Jain literature. BRUHN (1987a: 109) later pointed to the frequent co-occurrence of various syncretic phenomena within a single tradition. He distinguished, for instance, between 'syncretism' and 'import' in Jain literature. In contrast to Bechert and Pye, in his view 'syncretism' denotes the end product of the process of syncretisation, i.e. a 'real' synthesis of elements (from the participants point of view), whereas 'import' describes a situation of 'unreal' synthesis, where new elements are incorporated but not yet properly integrated (danger of disintegration). Bruhn also suggested distinguishing more clearly between different sources and periods, in order to achieve greater realism. In this context belong analytical distinctions between (a) the 'hinduised' (WILLIAMS (1983: xx)) or 'pseudo-jainised' (JAINI (1974: 335), (1979: 291–4) ritual and literature of (post-) medieval temple-worshipping Jain traditions, (b) the 'islamicised' (JAINI (1974: 314 n. 63)) iconoclastic reform movements which emerged in the Mughal period, and (c) contemporary 'westernised' developments.^[33] Whatever the merit of such typologies, which contrary to WEBER (1988) are often based on the supposition of 'essential characteristics', it is apparent that one of the main ambitions of present research is the construction of comprehensive classifications of various forms of syncretisation and their strategic uses.^[34]

An important debate between GOMBRICH (1971: 49) and BECHERT (1978: 20–4) on the question of the relation between literary syncretism (eclecticism) and the syncretism of popular religious practice in contemporary (Theravāda) Buddhism is also relevant for the understanding of similar phenomena amongst the Jains. Gombrich describes non-monastic forms of Buddhism as 'accretive' or corrupted forms. Bechert, on the other hand, criticises his devaluation of popular beliefs and of the political role of religion as 'elitist'. Instead, he interprets Buddhist 'cultures' as 'systems' or organic totalities, encompassing both saṅgha and society. TAMBIAH (1977), too, focuses less on Buddhist doctrine and the saṅgha and more on cultural history, emphasising especially the constitutive role of local cosmologies ('pantheons') which are implicated in the cults of Buddhist kingship. This approach, which favours a typified 'common man's' view from within and privileges 'hierarchisation' ('hegemony' or 'totalisation') as the most important strategy of acculturation, was pioneered by DUMONT (1980: 427 n. 6, 433 n. 19), who argued that historically the 'worldly religion' of 'Hinduism' emerged as a product of cumulative interactive processes between the 'two ideal types' of Brāhmaṇism and Jainism / Buddhism which superimposed an 'individual religion … on to the religion of the group [caste]' (DUMONT (1980: 275)).^[35] From the perspective of an individual, GUMPERZ (1972: 230 f.) pointed out, superposed structures demand a wider socioreligious repertoire, including role compartmentalisation and perspective variation. Dumont's view that, from a lay participant's point of view, soteriological cults appear as religions of individual choice which are superimposed upon worldly religion, lends support to both accretionary and syncretistic interpretations. It also highlights the marginal historical role of Jainism in India, which, for want of political support, was nowhere able to achieve a culturally dominant position comparable to Buddhism in the countries of Theravāda Buddhism, and consequently not forced to develop its own (hegemonic) social system. Jainism always remained primarily a monastic religion which relied on the institutions of Hinduism and the state to legislate for society. Jain philosophical syncretism conceives merely of a negative totality based on the disjunctive synthesis of differences within an infinite horizon of plural perspectives. Yet, negative philosophical forms of syncretism are to be distinguished from positive linguistic or socioreligious forms of syncretism, which are less prominent in Jain discourse, but dominant in practice.^[36] The Jain case shows that it is an empirical question whether a given form of popular religion appears to be predominantly accretic or syncretic. It also underlines the crucial importance of configurations of power for competitive processes of doctrinal syncretism and socio-cultural synthesis. The religious status of 'popular Jainism'—'deviation', 'cultural bedrock' or 'modern political essentialisation'—is the subject of ongoing disputes between rivalling Jain leaders. Epistemologies and religious rituals for Jain laity were constructed intentionally by Jain monks. Yet, the extent to which social life is regulated by Jain social philosophy varies locally and from sect to sect, and from caste to caste. Iconoclastic Jain sects rely on the Hindu social system alone, whereas temple-worshipping sects accept the 'hinduised / jainised' practices of popular Jainism as an integral part of Jain religion,^[37] while communal reformers demand the 'eradication of every non-Jain element from the Jaina community' (SANGAVE (1980: 410)) in order to form entirely new social entities. Present political attempts to ethnicise the 'Jain community' by propagating intra-religious, trans-sect and trans-caste marriages are unlikely to succeed, however, because of ongoing internal sectarian rivalries, exclusive caste and class affiliations of the laity, and the continuing existence of mixed religious castes. Effectively, Jain communalism contributes to the strengthening of the cultural self-consciousness of an important faction of the new Indian business class, but does not alter the hierarchical structure of the society itself.

Jain laity usually practises 'Jain' and 'Hindu' rituals side by side, combining soteriological religion with worldly religion without mixing the two, as described by Dumont.^[38] Even the lay followers (śrāvaka) of contemporary Jain reformist groups (e.g. 'Jain communalists') cannot avoid combining 'Jain' and 'Hindu' religious practices, because of lacking Jain life-cycle rituals.^[39] Sometimes popular practices are 'jainised' by ascetics, and in this way legitimately incorporated into Jain religion. 'Jain marriages' for instance, and similar life-cycle rituals, are created simply by adding a Jain mantra to customary local procedures; and 'Jain pūjās' are rendered possible if interpreted as forms of dāna, i.e. without expectation of return etc. (WILLIAMS (1983: xx–xxv, 216)). Re-interpretation and modification through addition etc., are essential techniques for incorporating elements from other traditions and for constructing cosmologies and embryonic Jain social systems along the lines of pre-existing Hindu and Buddhist models. Yet, only few elements of 'Hindu' popular religion have been fully integrated, predominantly into the ritual corpus of temple-worshipping sects. Aniconic Jain sects do not practice socioreligious rituals to the same degree as image-worshipping sects, and thus have a less clearly defined socioreligious identity. Structure and semantics of the ritual terminology correspond to socioreligious structure. The paradigmatic case of an apparently 'non-Jain' popular ritual which was appropriated and re-interpreted by medieval Jain ascetics to build up a Jain system of lay rituals is pūjā.^[40] Its ambiguous status between soteriological and world-affirming orientations is reflected in the intentional multivocality of the religious terminology employed in this and similar lay rituals, as WILLIAMS (1983), LAIDLAW (1985) and CORT (1989), (1991) demonstrated.^[41] The socioreligious dimension constituted by a system of jainised lay rituals seems to be predicated on generalised indirectness.

— II —

Modern writers on Jainism have often noted the abundance of similes and double meanings (śleṣa) in Jain narrative and ritual literature,^[42] and their strategic use to infuse conventional language and popular stories with different meanings, derived from Jain ethics.^[43] WILLIAMS (1983: xviii–ix) was the first scholar to highlight the ways in which medieval Jain writers, such as the Digambara ācārya Jinasena (9th CE), instrumentalised śaivaite terms (amongst others) as 'vehicles' for the indirect communication of their own religious views:

'Jain writers have shown a remarkable aptitude for the subtle handling of words … The polyvalence of certain expressions even within the limits of the same text is often disconcerting: guṇa in particular is greatly overworked and so are kriyā and karman. Indeed one is led to wonder whether the double meanings given to many words and their formal identity with Hindu terms may not be voluntary. Examples of such coincidences (with the Jaina meanings noted in parentheses) are: śiva (mokṣa), liṅga (the monks symbols such as the rajo-harana), guṇatraya (the ratna-traya), paśupati (the Jina) mahā-deva (the Jina) whilst on the other hand the word Digambara itself can be an epithet of Śiva.'

WILLIAMS (1983: xix) sees the reasons for the intentionally multivocal use of terms in the political assertiveness of 'Hinduism' in medieval South Indian society which forced the Jains to conceal their 'heterodox' beliefs behind a conformist public facade as a way of social self-protection:

'It may be that such resemblances were intended to render Jaina doctrines attractive to śaivas or that śaiva persecution made it desirable to give to certain Jaina texts an innocuous aspect. Certainly the Jaina's concept of asatya^[44] would make it easy for them to adopt an attitude similar to that of those Shiite sectarians who in the early days of Islam maintained an outward conformity by concealing their real beliefs under forms of words.'

Numerous studies on diglossia, multilingualism / multifaithism and code-switching demonstrated in the meantime that the strategy of 'outward conformity and inward dissent', based on the method of differentiating hierarchical levels / media of discourse, is not limited to a certain historical period in South Asia or to the Jains in particular, but a universal feature, especially of dependent subaltern groups, minorities, or elites. DUMONT (1980: 194) analysed the method of 'extrinsic borrowing' 'from superiors of certain features as social signs and not as functional features' in terms of a theory of acculturation, which distinguishes three contemporary types of cultural interaction:^[45] 'rejection, mixture, in which traditional and modern features exist happily side by side, and combination, which unites them intimately in new forms of a hybrid nature and ambiguous orientation' (DUMONT (1980: 229)).^[46] Others emphasised the political implications of counter-hegemonic adaptive strategies. SEAL (1968), RUDOLPH–RUDOLPH (1984), HAYNES (1991), and JAFFRELOT (1993), for instance, investigated the paradoxical effects of double-strategies employed by political mediators in (colonial) South Asia, which used 'modern language' in the public sphere, i.e. institutions of the state and print media, and 'traditional language' within their own community. Sociolinguistic theories of discourse and multilingualism will prove useful for the future study of the interaction of different levels within Jain religious language and of Jain discourse being superimposed on different contexts.^[47] In Jain literature, DERRETT (1980: 144) identified 'double standards'. CARRITHERS (1991: 266 f.) showed how the multivocal 'political rhetoric' of the leaders of Jain lay communities (samāja) gains persuasive force only if indirectly tapping into a 'diffuse realm of religious sentiments'. LAIDLAW (1985: 60) and CORT (1989: 449–70) suggested, conversely, that the official renunciatory 'religious ideology' of the Jains implicitly relies on a 'diffuse Jain ideology of wellbeing', and how 'symbolically rich' multivalent concepts, such as lābha or maṅgala which can mean either 'profit' or 'power' both in the world and in the religious sphere, 'bridge the two ideologies' (CORT (1989: 465)). The merit of Carrithers', Laidlaw's and Cort's approaches lies in the attempt to interpret the implicit links between the Jain religious discourse and the socio-economic sphere in terms of a theory of symbolisation.^[48] But they suffer from an exclusive focus on the laity and from the artificial treatment of Jain communities as quasi-ethnic groups isolated from the wider context of Indian society, which makes a critical analysis of the contextual relationships between Jainism and society, Jainism and power etc.— masked by the use of multivalent symbols—virtually impossible.^[49] 'Clearly', writes CORT (2001: 11), 'most scholars of ideology view power as ubiquitous.' LAIDLAW (1995), following Foucault, shares this view. However, while Williams' remarks need to be qualified, I think he and his successors posed the crucial question for an understanding of the characteristic 'syncretic', 'hybrid' or 'parasitic' form of much Jain narrative literature and lay ritual, by pointing to the intentional multivocality of basic concepts of popular Jainism, invented by ascetics, and their political-rhetorical function within contexts of competitive religious proselytisation. Williams did not pursue this line of research further, but confined himself to the sociologically less interesting search for the precise entailments of single terms. In order to handle the potentially boundless increase of investigations of such terms, BRUHN (1983: 61) proposed to limit 'rhetorical studies' to specific Jain genres, and most of Williams' successors followed this path.

I chose a different strategy in this article, turning away from the description and analysis of literary genres and doctrinal semantics (a task for the philologist) to the investigation of the pragmatics of Jain discourse. The limited aim of this study is to explore the methodological preconditions for an investigation of contextual implications of intentional multivocal utterances in Jain religious language. To accomplish this, a prior comparative analysis of the constitutive principles of Jain discourse and its typical normative contexts is required. Particularly significant is the question of the ways in which the specific ethical principles of Jain discourse interlink both with contextual norms and with universal moral presuppositions of communication per se, upon which intentional language usage indirectly relies, if Habermas is to be believed. I will seek to demonstrate that critical reflection on language usage on a level of abstraction similar to universal pragmatics is doctrinally prescribed for Jain ascetics, who need to consider the ethical implications of their own religious rhetoric in different contexts.^[50] Because the social implications of the Jain rules of speech themselves are hardly illuminated in Jain hermeneutical literature,^[51] I will start my investigation with a theoretical analysis of characteristic Jain hermeneutical procedures and discursive strategies in typical situations of linguistically mediated interaction.

The social function of Jain theories of speech, I argue, can be understood from the perspective of a theory of interactional competence. The term interactional competence encompasses cognitive, linguistic and rhetorical ability, and hence the capacity to use language both for the pursuit of power and insight. Interestingly enough, the use of mental and linguistic violence is a necessary requirement for accomplishing both aims. Power and influence are intrinsically connected with violence. It is one of the most intriguing questions how power works through discourse. Power has been defined by Max WEBER (1972: 28), from an intentionalist perspective, as the 'opportunity existing within a social relationship which permits one to carry out one's own will even against resistance and regardless of the basis on which this opportunity rests.'^[52] Rhetoric has similarly been characterised as a competencebased social technique by means of which a minority, or an elite, may gain or exercise social influence, personal prestige or persuasive power, over a majority. Its capacity to influence is often predicated on the measured violation of a conventional structure by means of the manipulation of linguistic and non-linguistic media of communication; i.e. it presupposes a 'deviance' from what is culturally felt as being 'normal' without threatening the co-operation of the listeners. Influence is further strengthened through commitment and dogmatism on the side of the proponent.^[53] By such means, a speaker may exert a disproportionate influence on individuals and society at large and stimulate changes in a pre-meditated direction. Social influence is rooted in conflict and striving for a new consensus: 'Conflict is … at the root of influence, either because it arises from the presence of a difference or because the existence of a disagreement brings it into the open' (MOSCOVICI (1985: 353)).

Rhetoric has another, ethical-pedagogical, side: it may not only serve as an instrument of power and manipulation, but (from the perspective of the addressee) also as a critical analytical method of truth-finding and insight. Both aspects are often intrinsically related and can only be dissected analytically.^[54] The ways in which Jain ascetics play with the two aspects of rhetoric, how they use multivocal language strategically as a means of normative influence and religious conversion, are interesting questions for sociological research.^[55] If, as Williams argued, the multivocal categories of/for popular Jainism are intentionally constructed by proselytising ascetics, and if this is done to generate effects of religious insight (samyag-darśana) in the audience, for instance through the rhetorical provocation of vairāgya-shocks (aversion leading to renunciation),^[56] how then does Jain philosophy account for the element of violence which is necessarily implied in acts of persuasion? In other words, if social influence can only be achieved through strategic acts of violence, how do ascetics conceive of the moral paradox involved in the violent production of non-violent attitudes?

— III —

For the analysis of the role of violence in rhetoric, or speech acts in general, a standard of non-violent speech is required. Such a standard is offered in the Jain scriptures. It will be discussed in Chapters V–VII. Comparable paradigms in contemporary Western philosophy are the conversational maxims of GRICE (1975) and the universal pragmatic validity claims of HABERMAS (1980), both echoing Kant rather than Carnap. A comparison with the Jain model promises to elicit key differences and highlight the specific nature of the principles of Jain discourse. In the following discussion of the two models, I adopt HABERMAS' (1980–1981: 440 ff.) / (1984–1987 I: 328 ff.) suggestion to treat empirical pragmatic models as if they were conscious operationalisations of universal pragmatics.^[57] The latter claims to be a more general than the former since it does not only theorise principles of linguistic exchange but also the morally binding force of rational argument and the socially constitutive function of rationally constituted consensus. Contrasting claims to universality of the models of Habermas and the Jains, the former would claim (on debatable grounds), can be tested through comparative analysis.

In his Theory of Communicative Action, HABERMAS (1980: 384–88) / (1984–1987 I: 284–89) distinguishes between 'communicative' and 'strategic' types of linguistically mediated interaction to be able to discriminate between consensus-oriented and manipulative forms of language usage. The key variable in his model is the dominant social orientation informing language usage.^[58] Communicative action is orientated towards consensus through rational argumentation, understanding and insight, whereas strategic action (success-oriented action) is orientated towards power and the pursuit of self-interest through the manipulative use of speech.^[59] Despite the ambiguity of his key terms Verständigung ('communicative understanding') and Einsicht ('insight'),^[60] Habermas differentiates clearly between linguistic and normative aspects of communicative action, that is, the understanding of the meaning of a speech act and the acceptance of implicit validity claims. In contrast to speech act theoreticians such as Austin, Grice, and others, Habermas does not privilege the intention of the speaker as determinative of the social meaning of a speech act, which he sees as negotiable, but focuses on the (contextually varying) conditions of acceptability, that is, the implicit or explicit reasons on which the validity claims of a speech act are based, which can be accepted or rejected in processes of intersubjective communication. The question is: What does it mean to understand a speech act? Not: What does it mean to understand an intention? Habermas expands AUSTIN's (1962) distinction between 'locutionary', 'illocutionary' and 'perlocutionary' speech acts, which he re-labels, partly following SEARLE's (1979: 12–20) terminology, as 'constatives', 'regulatives' and 'expressives' in analogy to his three universal validity claims 'truth', 'rightness' and 'truthfulness', by focusing not so much on speech acts themselves but on the social function of speech for the co-ordination of action. The advantage of Habermas' distinction between 'speech acts' and 'linguistically mediated interaction', prefigured in the work of sociolinguists such as GUMPERZ (1964), (1972) and HYMES (1972b),^[61] is that it allows to analyse the contextual moral and legal implications of language usage, i.e. the socially binding force of the validity claims implied in illocutionary speech acts, which cannot be reduced to the 'power of words' themselves nor to the underlying intention of the speaker.

'Communicative action' is not the same as 'communication'. It refers to a situation where the intersubjective co-ordination of action-plans is reached by way of rational agreement, that is, where an explicit consensus is constitutive for social integration (HABERMAS (1998: 396–98)). 'Strategic action' (success− / influenceoriented action), on the other hand, refers to a situation of exploitation, because it is not able to generate new normative consensus through rationally motivated acceptance of reasons itself, but only through the force of pre-existing norms and institutional or other configurations, which are external to the communicative process.^[62]

Hence, by defining communicative action as the original mode of language usage, directly oriented towards normative agreement, Habermas is able to criticise strategic action as a success orientated mode of language usage, which is parasitic on the former, because it is only indirectly orientated towards communicative understanding. The whole purpose of Habermas' theory lies in the analysis of the conditions of the possibility of a consensual constitution of social order, without grounding the emancipatory potential of discourse either in metaphysical postulates or in institutional configurations of power as suggested by FOUCAULT (1981), BOURDIEU (1991a), (1991b), or BLOCH (1975). The critical emancipatory interest which informs the theory of communicative action is directed against (illegitimate) power and in favour of consensual (and non-violent) forms of conflict resolution and institution building. HABERMAS (1998: 449) argues that the emancipatory potential of language is not a metaphysical ideal, but manifest in the 'unconcealed idealising surpluses of an innerworldly transcendence' in form of the universal validity claims implicitly presupposed in all processes of communication, as the vanishing points (Fluchtpunkte) of infinite processes of open intersubjective interpretation.^[63]

How can we explain the power of non-institutionally bound illocutionary acts to produce feelings of normative obligation? What motivates a listener freely to submit him / herself to normative constraints if not self-interest or external force? At first sight, HABERMAS' (1980) argument that illocutionary binding effects are the product of processes of rational understanding, i.e. of insight into the validity of reasons (e.g. common values and convictions), seems to rely on the pre-existence of an implicit normative consensus and of the rhetorical ability of the speaker (HABERMAS (1980: 386)). However, like other proponents of formal pragmatics (Apel, Allwood, Grice),^[64] Habermas attempts to circumvent both normative reductionism and subjectivism, by not focusing on the local 'ethical' context of specific empirical cases, but on the general 'moral' presuppositions of communication, that is, the universal normative conditions of intersubjective recognition which must be fulfilled for a speech act to be accepted. He states that 'we understand a speech act, if we know, what makes him acceptable' (HABERMAS (1980: 400)). Crucially, he assumes that every linguistically mediated interaction presupposes a set of idealisations on the side of the interlocutors themselves. He explicates these implicit basic conditions of communicative action in terms of a model of three universal pragmatic^[65] validity claims underlying all grammatically comprehensible utterances: (1) (propositional) truth, (2) (normative) rightness, and (3) (expressive) sincerity.^[66] Habermas argues that the binding power of rea sons, that is, of rationally motivated validity claims whose acceptance implies weak normative obligations,^[67] is not rooted in content but in procedural form, i.e. in the guarantee of the speaker, if necessary, to justify his / her claims in terms of these three types of validity claims and to give reasons which can be criticised and rejected (only) with better reasons (HABERMAS (1980: 406)).^[68] Precondition is the principal autonomy of the interlocutors, their ability to say 'no'. Only under this provision does acceptance imply voluntary agreement. For Habermas, who follows Kant and Durkheim here, the socially constitutive power of discourse is predicated on the independence of interdependent interlocutors, who co-operate on the basis of negotiated agreement. This, critics object, is the logic of the modern market based on functional division of labour.^[69]

HABERMAS (1981: 62 f.) / (1984–1987 II: 38 f.), (1991: 25, 44) acknowledges, the readiness to accept the binding power of agreed normative claims, instead of traditional authority, is itself a product of processes of social conditioning and presupposes specific forms of life and socialisation. The development of moral consciousness is predicated on the internalisation of the perspective of the generalised other or threatening or factually exercised sanctions. Moreover, the social manifestation of the idealtypical 'standard form' of rational argumentation presupposes situations or institutions where unconstrained rational argumentation between equals plays a constitutive role for society, such as the public sphere, the parliament, the courts of justice, or the university.^[70] According to Habermas, modern social milieus such as these are the product of evolutionary processes of social differentiation generating an increase in both inter dependence and individual autonomy and responsibility through a progressive universalisation of law and acquisition of moral competence. He argues that in cases where communicative action is consciously used for the production of normative consensus and social co-operation, discourse takes over the social function of ritual, and becomes a second order ritual (HABERMAS (1981: 86)).^[71] With reference to the apparently increasing importance of explicit negotiation in processes of social self-identification, Habermas defines this process of rationalisation as the 'linguistification (Versprachlichung) of the sacred', and argues that it goes hand in hand with the progressive 'technicisation' or 'colonisation of the lifeworld' by systemic processes.

How to analyse Jain discourse from the point of view of the theory of communicative action? Generally, religious discourse is depicted by HABERMAS (1980–1981) as a form of communication based on structures of traditional authority in which status rather than argumentation functions as a medium of generalised acceptability.^[72] The role of power and insight in Jain discourse could be analysed in these terms. But one has to be aware of the fact that the 'rational reconstruction' of the universal normative presuppositions of linguistic exchange projects its own idealisations into human reality. The theory of communicative action itself contains a religious element in its notion of the 'unconscious innerworldly transcendence' that is implicitly presupposed in communicative action. A deeper analysis of the differences between the theory of communicative action and the Jain doctrine of speech therefore requires a comparison between the normative ideals underlying Habermas' theories of communicative action and discourse ethics^[73] with those of Jain discourse ethics (to my knowledge there are no Jain attempts to rationally reconstruct implicit rules), supplemented by comparative analyses of rules and discourses of norm application in typical speech situations (See infra section V–VI).

The daily sermon (pravacana) of Jain ascetics is a good example for a transitional ritual, where discourse serves both as an instrument for the reproduction of the traditional authority of the saṅgha, and for the eventual production of new normative consensus via the evocation of religious insight amongst yet unconverted listeners. A Jain discourse of conversion presupposes mutual independence of interlocutors and creates new consensus via the unconstrained acceptance of validity claims. The main difference compared with the ideal situation of communicative action is the institutionalised one-sidedness of a monological discourse and its coded semantic form which does not allow for a negotiated modification of the religious dogma itself, at least not in the short run. Because insight and power go hand in hand in traditionbased religion, conversion discourses always take the form of strategic action, producing insight (psychological and verbal acceptance of the dogma) as a perlocutionary effect predicated on the intentional deconstruction of the pre-existing conventional normative consensus.^[74] In this case, latent strategic action serves as the vehicle for both communicative action or even mythopoetic world disclosure, and for the unquestioned reproduction of the hierarchy of power. In practice, the two fundamental ways of interpretation are irreducible to each other. But they can be analytically differentiated. In the following, I analyse the plurivocality of Jain discourse primarily in terms of its social implications—as a form of latent strategic action—though multiple ambiguities can be distinguished in every speech act in a conventional speech situation.

It is an open question to what extent the theory of communicative action falls under this verdict as well.^[75] HABERMAS (1986: 383), who after all regards communicative action as the original or constitutive mode of language usage, insists on the empirical inevitability of indirect violence and power in the initial stages of the historical evolution of abstract norms. He distinguishes two developmental phases of communicative co-operation and social integration: Communicative action, he argues, is originally embedded in contexts of latent strategic action, because, initially, situation definitions do not sufficiently overlap. The participants therefore have to use indirect communication 'following the model of intentional semantic approaches (Grice)', in order to avoid a breakdown of co-operation. In the opening phases of all processes of co-operation indirect communication plays an important role for the creation of overlapping definitions of the situation, as an alternative of meta-communication: the speaker makes the hearer understand something qua perlocutionary effects, i.e. latent strategic speech acts, whose contextual implications are not directly expressed, and need to be inferred by the hearer.^[76] However, these strategic sequences are embedded within the general context of communicatively oriented speech.^[77]

For HABERMAS (1980: 402–8) / (1984–1987 I: 298–304), (1981: 57–63) / (1984– 1987 II: 34–9), as for SEARLE (1993: 91, 99), normative authority is originally embedded within imperative authority, and takes the form of context-specific intentions and speech acts of particular individuals. This fits the case of the Jain sermon. However, Habermas' evolutionary theory of the genesis of norms is highly speculative and certainly the weakest part of his conceptual system. Moreover, he has not elaborated his theory of latent strategic language usage. But he indicates the systematic role of latent strategic action within a typology of social actions,^[78] and shows how shared convictions may serve as sources of legitimate power. Power enters unnoticeable into the pores of everyday-life communication via two forms of pseudo-communication concealing strategic intent under the facade of consensual action: (a) conscious deception or manipulation, and (b) systematically distorted communication (HABERMAS (1984: 548)).^[79] In cases of manipulation, at least one actor intentionally deceives others by hiding the fact that s/he does not comply with the three universal pragmatic validity claims. In cases of systematically distorted communication at least one of the participants deceives him / herself about the fact that s/he acts strategically, while maintaining the external appearance of communicative action (HABERMAS (1984: 461)). In both cases a deviation from universal pragmatic presuppositions takes place, through the splitting up or doubling of communication. In the first case, a competent speaker generates intentionally multivocal expressions for social purposes. In the second case, the internal organisation of speech is distorted by way of a privatised use of language. In psychoanalysis this is regarded as a process of de-symbolisation. According to Habermas it is symptomatic for a loss of interactional competence (which may or may not be culturally normalised) (HABERMAS (1984: 251–4)). He argues that distorted forms of communication often occur in situations of social dependency, where they serve as unconscious defence mechanisms concealing conflict smouldering beneath surface consensus (HABERMAS (1984: 232, 264–9)).^[80] A manipulative, intentionally symbolic usage of language, on the other hand, creating conflict artificially for the sake of influence, is to be expected in situations of social dominance, and requires cognitive and communicative competencies. We will see that ambiguous language can be used for religious purposes as well.

It could be asked whether the structural multivocality generated by Jain discourse, superposed on varying contexts, can from different viewpoints be interpreted either as a form of latent strategic action or of systematically distorted communication. The perceived ambiguity of certain terms or utterances in Jain discourse is a function of processes of religious interpretation. Depending on circumstances, they may involve forms of symbolisation as well as the expression of latent structural conflicts, resulting for instance from the typical position of many Jains as subaltern elites within the hierarchical structure of Indian society. On the one hand, Jain ascetics use the strategy of intentional re-interpretation of dominant forms of discourse and popular rites, thus creating a private language for competent members of their religious community. On the other hand, the propagation of Jainism through proselytising renouncers eventually triggers effects of moral insight amongst some listeners, by providing a language for example for the explanation and expression of diffuse feelings or implicit conflicts, which may explain part of its appeal. It is important to note that the Jain tradition itself provides a yardstick for the critique of self-deceptions and symptoms of alienated modes of life with its ideal of non-violence. The authenticity associated with such a life-project could be understood as a higherlevel validity claim, in analogy to the claim to truthfulness in expressive speech acts.^[81] However, better analytical categories are required to investigate the empirical diversity of forms of concealed strategic action. As Habermas indicates, some of them are provided by the intentional semantic approaches of Grice and Searle.

— IV —

One of the most interesting contributions to the theory of linguistic communication and rhetoric is GRICE's (1975) analysis of the communicative function of 'conversational implicature'. In his article Logic and Conversation, Grice provides a pragmatic description of how multiple meanings are purposefully generated. He investigates what he calls 'implicature', that is, instances in which 'a speaker deliberately says something which is not, in fact, what he means' (GRICE (1975: 43 f.)). Grice's analysis is based on the fundamental pragmatic postulate that there is a assumption by all conversationalists of the rational and efficient nature of talk—a supposition, which can be formally stated in terms of a set of counterfactual principles and maxims concerning the efficient and univocal transmission of information, given that both parties co-operate and are interested to continue the conversation. Against this assumption polite, multivocal ways of speaking appear as deviations, requiring rational explanation on the part of the recipient, who finds in considerations of politeness reasons for the speaker's apparent irrationality or inefficiency. Given the speaker's rationality of expression, and the hearer's willingness to co-operate, unclear, ambiguous language forces the hearer to work out the implied suppositions of a multivocal statement. The latter must ask her / himself why a speaker, who can express her / himself unambiguously, has chosen not to say directly what s/he means, but rather speaks in a veiled language. In such cases, Grice argues, the hearer (interpreter) will assume that the speaker has deliberately 'flouted' the mutually presupposed conversational principles to 'force an implicature' onto the hearer, who then has to infer what the speaker intended without saying so openly. Particularly interesting are cases in which someone deliberately violates a maxim, not to mislead, but to influence others by getting another message across. Strategically constructed 'deviations' such as these, which are 'parasitic' to direct communication, are particularly successful in achieving social influence, because they induce the necessary element of conflict in a manifestly non-violent form, not breaching social norms openly but only indirectly via implicit 'exploitation' or 'flouting' of presupposed conversational principles. It is precisely this 'non-violent' form of social influence which I regard as central to Jain rhetoric.

The construction of implicatures is one of the foremost rhetorical and poetical devices. Yet, only under certain contextual conditions will a conversational act of flouting lead to social consequences. Without the hearer's willingness and ability to cooperate, and to invest the effort to work out or to grasp intuitively the implications of an ambiguous statement, rhetorical communication with the help of implicature would not be effective. The uses of implicatures presupposes the invocation of contextual knowledge or common experiential ground on the part of the hearer, who is forced to generate a meaningful interpretation in response to the speaker's deliberate violation of the principles of conversation, which implicitly forces on the hearer a choice between deference to the veiled imperative or discontinuation of the communication. The efficacy of this kind of indirect communication is ultimately based on ego's exploitation of alter's willingness to adhere to the general cooperative principle of language exchange under given normative conditions. Accordingly, the principal means of physically 'non-violent' resistance is non-cooperation.^[82]

Grice's theory of language usage is, in itself, of limited use in anthropological research. Theories of information processing cannot sufficiently account for the empirical role of conventional meaning and contextual knowledge, as HYMES (1972a), (1972b) convincingly argued.^[83] BROWN'–LEVINSON's (1978) theory of politeness, which builds upon Grice's principles, therefore links the theory of conversational implicature with Goffman's theory of social self-identity or 'face' (self-esteem or public self-image). It argues that talk exchange can be related to contextual, that is, institutional, variables by considering the relative social position of speaker and hearer, which decisively influences choices of politeness strategy. 'Face' is relationally defined, 'that is, normally everyone's face depends on everyone else's being maintained, and since people can be expected to defend their faces if threatened, and in defending their own to threaten others' faces, it is in general in every participant's best interest to maintain each other's face' (BROWN–LEVINSON (1978: 66)). Two kinds of 'face wants' are distinguished: 'negative face wants' (to be unimpeded by others)—often achieved by strategies of non-interaction (which Marriott regards as typical for Jains, and vaiśyas in general)—and 'positive face wants' (to be desirable to others) (BROWN–LEVINSON (1978: 66)). 'Face threatening acts' (FTA's) are illocutionary acts which might cause a loss of face by way of orders, requests, threats, offers, suggestions etc. They run contrary to particular face wants, and motivate certain strategies of politeness, employed to minimise the threat, and to maintain face in a particular situation (BROWN–LEVINSON (1978: 70–3)).

BROWN and LEVINSON distinguish four main strategies of politeness: (a) 'bald on record' (being as clear as possible), (b) 'positive politeness' (the expression of solidarity), (c) 'negative politeness' (the expression of restraint), and (d) 'off record' (the avoidance of unequivocal impositions) (BROWN–LEVINSON (1978: 73–5)).^[84] Any of the four strategies may satisfy the opponent's face wants to avoid conflict and to minimise threat. Yet, politeness is also used in order to stimulate conflict in a calculated way. Positive politeness and negative politeness—the dominant Jain strategy— I would argue, are primarily strategies of conflict avoidance. The most interesting strategy, also often used in Jain discourse, is to go off-record, that is, to construct intentionally multivocal statements: 'A communicative act is done off record if it is done in such a way that it is not possible to attribute only one clear communicative intention to the act. In other words, the actor leaves himself an "out" by providing himself with a number of defensible interpretations; he cannot be held to have committed himself to just one particular interpretation of this act. Thus if a speaker wants to do a FTA, but wants to avoid the responsibility for doing it, he can do it off record and leave it open to the addressee to decide how to interpret it' (BROWN–LEVINSON (1978: 216)). The purpose of off-record strategies is to make the addressee responsible for decoding the intentions. An FTA, like any word, rule or symbolic act,^[85] gains its meaning through context and becomes threatening only through its realisation via 'wrong or correct' interpretation.^[86] The near universal usefulness of off-record strategies for purposes of persuasion and social influence is aptly expressed by STRECKER (1988: 114): 'To go off-record is one of the most pervasive strategies in social interaction whenever actors want to avoid harsh confrontation and the possibility of conflict, and when they want to persuade others, to influence them so that they do what they cannot be openly coerced to do. All this is done by means of indirect message construction.' Important for my argument is that politeness and symbolisation is used to avoid open conflict. A great deal of politeness and symbolisation is therefore to be expected from the Jains.

BROWN–LEVINSON (1978: 254–6) conclude their discussion of politeness phenomena with a typology of characteristic patterns of strategy distribution in speaker / hearer relationships regarding the contextual variables of power (P) and social distance (D) (which could be associated with 'active' directive and 'passive' commissive orientations respectively). In this way, they try to tie Grice's concept of conversational implicature into a context-sensitive social theory of linguistically mediated interaction.^[87] They predict that in a situation where one of the interlocutors has high social power over another, and where social distance is low, the speaker with high social power will predominately use on-record strategies, and the speaker with low social power will use negative politeness and off-record strategies. STRECKER (1988: 165 f.), however, notes that individuals in an inferior position are generally not allowed to force an implicature upon the socially powerful (only vice versa). He suggests that the dominant pattern of strategy distribution in asymmetrical social contexts is the combination of off-record strategies on the part of superiors and on-record strategies and negative politeness on the part of inferiors.

The still unresolved discrepancy between Brown and Levinson's and Strecker's interpretation provokes the question whether it makes sense at all to link types of socially-purposeful language usage to certain 'social positions', given the wide range of contextually acceptable strategy choices. Off-record strategies, in fact, can be used for many purposes: for purely aesthetic reasons, for the negotiation of unclear social relationships, for the veiling of heavy FTA's or forms of control, or for the stabilisation of asymmetrical social relationships. More importantly, a statement in itself does not force a reader / listener to adopt a particular type of interpretation (and no particular action follows from it in any structured way). Everything, in fact, depends on the wider social context, particularly on the informal or institutionalised norms and social conventions of a given field of discourse, which sanction the way in which language is used and which interpretations are regarded as feasible. As Habermas shows, it is only in situations where social pressures and normative sanctions are attached to language usage, that normative claims can be enforced through extra-linguistic sanctions. It is impossible to force anybody to comply with pragmatic conditions for the fulfilment of an implication, in the Gricean sense, without implicitly presupposing an institutional context and referring to a set of known social sanctions attached to the specific norms of a particular discursive field.^[88]

This point has evaded most linguistic theories of politeness, because—as Habermas points out—they do not clearly distinguish between co-operation in conversation and the role of speech acts for social co-operation. Therefore they cannot explain, for instance, how, in certain situations, superiors may rely on contextual pressures to force inferiors to decode their veiled speech and to fulfil the indirectly communicated demands. It is not the power of the words themselves nor the speaker's intentions which inform the 'calling in' of conversational implicature, but the role of the presupposed norms and sanctions which assure the perpetuation of social systems which, for instance, force inferiors to infer unequivocally the indexical meaning of intentionally ambiguous statements of their superiors, and to practically fulfil their unspoken commands without questioning the normative basis of their social co-operation.^[89] Because they accept the unnecessary constraints of Grice's theory of information processing, even Brown and Levinson and Strecker cannot avoid reductionist interpretations of language usage in terms of pre-existing 'social positions' and differential 'power'. Habermas' distinction between communicative and success orientated language usage, however, seems to open a path for a non-reductionist understanding, for instance, of the social function of positive politeness, which, in the words of Brown and Levinson, can only 'express' but not 'generate' solidarity. Reductionism can only be avoided if not merely the reflexive but also the constitutive function of discourse for the construction of social identities is considered. From this perspective, strategies of negative politeness and off-record appear to have greater power than positive politeness or onrecord strategies, particularly in social situations where coercion cannot succeed and where co-operation is primarily engendered through indirect communication.