Итог полностью

The Economy of Linguistic Exchanges

Sociology can free itself from all the forms of domination which linguistics and its concepts still exercise today over the social sciences only by bringing to light the operations of object construction through which this science was established, and the social conditions of the production and circulation of its fundamental concepts. The linguistic model was transposed with such ease into the domain of anthropology and sociology because one accepted the core intention of linguistics, namely, the intellecrualist philosophy which treats language as an object of contemplation rather than as an instrument of action and power. To accept the Saussurian model and its presuppositions is to treat the social world as a universe of symbolic exchanges and to reduce action to an act of communication which, like Saussure's parole, is destined to be deciphered by means of a cipher or a code, language or culture.¹

In order to break with this social philosophy one must show that, although it is legitimate to treat social relations - even relations of domination - as symbolic interactions, that is, as relations of communication implying cognition and recognition, one must not forget that the relations of communication par excellence - linguistic exchanges - are also relations of symbolic power in which the power relations between speakers or their respective groups are actualized. In short, one must move beyond the usual opposition between economism and culturalism, in order to develop an economy of symbolic exchanges.

Every speech act and, more generally, every action, is a conjuncture, an encounter between independent causal series. On the one hand, there are the socially constructed dispositions of the linguistic habitus, which imply a certain propensity to speak and to say determinate things (the expressive interest) and a certain capacity to speak, which involves both the linguistic capacity to generate an infinite number of grammatically correct discourses, and the social capacity to use this competence adequately in a determinate situation. On the other hand, there are the structures of the linguistic market, which impose themselves as a system of specific sanctions and censorships.

This simple model of linguistic production and circulation, as the

relation between linguistic habitus and the markets on which they offer their products, does not seek either to challenge or to replace a strictly linguistic analysis of the code. But it does enable us to understand the errors and failures to which linguistics succumbs when, relying on only one of the factors involved - a strictly linguistic competence, abstractly defined, ignoring everything that it owes to the social conditions of its production - it tries to give an adequate account of discourse in all its conjunctural singularity. In fact, as long as they are unaware of the limits that constitute their science, linguists have no choice but to search desperately in language for something that is actually inscribed in the social relations within which it functions, or to engage in a sociology without knowing it, that is, with the risk of discovering, in grammar itself, something that their spontaneous sociology has unwittingly imported into it.

Grammar defines meaning only very partially: it is in relation to a market that the complete determination of the signification of discourse occurs. Part (and not the least) of the determinations that constitute the practical definition of sense comes to discourse automatically and from outside. The objective meaning engendered in linguistic circulation is based, first of all, on the distinctive value which results from the relationship that the speakers establish, consciously or unconsciously, between the linguistic product offered by a socially characterized speaker, and the other products offered simultaneously in a determinate social space. It is also based on the fact that the linguistic product is only completely realized as a message if it is treated as such, that is to say, if it is decoded, and the associated fact that the schemes of interpretation used by those receiving the message in their creative appropriation of the product offered may diverge, to a greater or lesser extent, from those which guided its production. Through these unavoidable effects, the market plays a part in shaping not only the symbolic value but also the meaning of discourse.

One could re-examine from this standpoint the question of style: this ‘individual deviation from the linguistic norm’, this particular elaboration which tends to give discourse its distinctive properties, is a being-perceived which exists only in relation to perceiving subjects, endowed with the diacritical dispositions which enable them to make distinctions between different ways of saying, distinctive manners of speaking. It follows that style, whether it be a matter of poetry as compared with prose or of the diction of a particular (social, sexual or generational) class compared with that of another class, exists

only in relation to agents endowed with schemes of perception and appreciation that enable them to constitute it as a set of systematic differences, apprehended syncretically. What circulates on the linguistic market is not ‘language' as such, but rather discourses that are stylistically marked both in their production, in so far as each speaker fashions an idiolect from the common language, and in their reception, in so far as each recipient helps to produce the message which he perceives and appreciates by bringing to it everything that makes up his singular and collective experience.

One can extend to all discourse what has been said of poetic discourse alone, because it manifests to the highest degree, when it is successful, the effect which consists in awakening experiences which vary from one individual lo another. If. in contrast to denotation, which represents ‘the stable part, common to all speakers'.² connotation refers to the singularity of individual experiences, this is because it is constituted in a socially characterized relation to which the recipients bring the diversity of their instruments of symbolic appropriation. The paradox of communication is that it presupposes a common medium, but one which works - as is clearly seen in the limiting case in which, as often in poetry, the aim is to transmit emotions - only by eliciting and reviving singular, and therefore socially marked, experiences. The all-purpose word in the dictionary. a product of the neutralization of the practical relations within which it functions, has no social existence: in practice, it is always immersed in situations, to such an extent that the core meaning which remains relatively invariant through (he diversity of markets may pass unnoticed.³ As Vendryes pointed out. if words always assumed all their meanings at once, discourse would be an endless play on words: but if. as in the case of the French verb loiter (to rent, from locare) and louer (to praise, from laudare). all the meanings it can take on were totally independent, all plays on words (especially of the ideological sort) would become impossible.⁴ The different meanings of a word are defined in the relation between the invariant core and the specific logic of the different markets, themselves objectively situated with respect to the market in which the most common meaning is defined. They exist simultaneously only for the academic mind which elucidates them by breaking the organic solidarity between competence and market.

Religion and politics achieve their most successful ideological effects by exploiting the possibilities contained in the polysemy inherent in the social ubiquity of I he legitimate language. In a differentiated society, what are called ‘common* nouns - work.

family, mother, love, etc. - assume in reality different and even antagonistic meanings, because the members of the same ‘linguistic community' use more or less the same language and not several different languages. The unification of the linguistic market means that there are no doubt more and more meanings for each sign.⁵ Mikhail Bakhtin reminds us that, in revolutionary situations, common words take on opposite meanings. In fact, there are no neutral words: surveys show, for example, that the words most commonly used to express tastes often receive different, sometimes opposite, meanings from one social class to another. The word soigne (neat, clean, conscientious), for example, used approvingly by the petits bourgeois, is rejected by intellectuals for whom, precisely, it evokes everything that is petit-bourgeois, petty and mean-spirited. The polysemy of religious language, and the ideological effect of the unification of opposites or denial of divisions which it produces, derive from the fact that, at the cost of the re-interpretations implied in the production and reception of the common language by speakers occupying different positions in the social space, and therefore endowed with different intentions and interests, it manages to speak to all groups and all groups speak it - unlike, for example, mathematical language, which can secure the univocal meaning of the word ’group' only by strictly controlling the homogeneity of the group of mathematicians. Religions which are called universal are not universal in the same sense and on the same conditions as science.

Recourse to a neutralized language is obligatory whenever it is a matter of establishing a practical consensus between agents or groups of agents having partially or totally different interests. This is the case, of course, first and foremost in the field of legitimate political struggle, but also in the transactions and interactions of everyday life. Communication between classes (or. in colonial or semi-colonial societies, between ethnic groups) always represents a critical situation for the language that is used, whichever it may be. Il tends to provoke a return to the sense that is most overtly charged with social connotations: ‘When you use the word paysan (peasant) in the presence of someone who has just left the countryside, you never know how he is going to take it.' Hence there are no longer any innocent words. This objective effect of unveiling destroys the apparent unity of ordinary language. Each word, each expression, threatens to take on two antagonistic senses, reflecting the way in which it is understood by the sender and the receiver. The logic of the verbal automatisms which insidiously lead back to ordinary

usage, with all its associated values and prejudices, harbours the permanent danger of the “gaff which can instantly destroy a consensus carefully maintained by means of strategies of mutual accommodation.

But one cannot fully understand the symbolic efficacy of political and religious languages if one reduces it to the effect of the misunderstandings which lead individuals who are opposed in all respects to recognize themselves in the same message. Specialized discourses can derive their efficacy from the hidden correspondence between the structure of the social space within which they are produced - the political field, the religious field, the artistic field, the philosophical field, etc. - and the structure of the field of social classes within which the recipients are situated and in relation to which they interpret the message. The homology between the oppositions constitutive of the specialized fields and the field of social classes is the source of an essential ambiguity which is particularly apparent when esoteric discourses are diffused outside the restricted field and undergo a kind of automatic universalization, ceasing to be merely the utterances of dominant or dominated agents within a specific field and becoming statements valid for all dominant or all dominated individuals.

The fact remains that social science has to take account of the autonomy of language, its specific logic, and its particular rules of operation. In particular, one cannot understand the symbolic effects of language without making allowance for the fact, frequently attested, that language is the exemplary formal mechanism whose generative capacities are without limits. There is nothing that cannot be said and it is possible to say nothing. One can say everything in language, that is, within the limits of grammaticality. We have known since Frege that words can have meaning without referring to anything. In other words, formal rigour can mask semantic freewheeling. All religious theologies and all political theodicies have taken advantage of the fact that the generative capacities of language can surpass the limits of intuition or empirical verification and produce statements that are formally impeccable but semantically empty. Rituals are the limiting case of situations of imposition in which, through the exercise of a technical competence which may be very imperfect, a social competence is exercised - namely, that of the legitimate speaker, authorized to speak and to speak with authority. Benveniste pointed out that in Indo-European languages the words which are used to utter the law are related to the verb 'to speak'. The right utterance, the one which is formally correct, thereby claims,

and with a good chance of success, to utter what is right, i.e. what ought to be. Those who, like Max Weber, have set the magical or charismatic law of the collective oath or the ordeal in opposition to a rational law based on calculability and predictability, forget that the most rigorously rationalized law is never anything more than an act of social magic which works.

Legal discourse is a creative speech which brings into existence that which it utters. It is the limit aimed at by all performative utterances - blessings, curses, orders, wishes or insults. In other words, it is the divine word, the word of divine right, which, like the intuitus originarius which Kant ascribed to God, creates what it states, in contrast to all derived, observational statements, which simply record a pre-existent given. One should never forget that language, by virtue of the infinite generative but also originative capacity - in the Kantian sense - which it derives from its power to produce existence by producing the collectively recognized, and thus realized, representation of existence, is no doubt the principal support of the dream of absolute power.

1

'As you say, my good knight! There ought to be laws to protect the body of acquired knowledge.

Take one of our good pupils, for example: modest and diligent, from his earliest grammar classes he's kept a little notebook full of phrases.

After hanging on the lips of his teachers for twenty years, he’s managed to build up an intellectual stock in trade; doesn’t it belong to him as if it were a house, or money?'

P. Claudel, Le Soulier de Satin

‘Language forms a kind of wealth, which all can make use of at once without causing any diminution of the store, and which thus admits a complete community of enjoyment; for all, freely participating in the general treasure, unconsciously aid in its preservation'.¹ In describing symbolic appropriation as a sort of mystical participation, universally and uniformly accessible and therefore excluding any form of dispossession, Auguste Comte offers an exemplary expression of the illusion of linguistic communism which haunts all linguistic theory. Thus, Saussure resolves the question of the social and economic conditions of the appropriation of language without ever needing to raise it. He does this by resorting, like Comte, to the metaphor of treasure, which he applies indiscriminately to the ‘community’ and the individual: he speaks of ‘inner treasure’, of a ‘treasure deposited by the practice of speech in subjects belonging to the same community’, of ‘the sum of individual treasures of language', and of the ‘sum of imprints deposited in each brain’.

Chomsky has the merit of explicitly crediting the speaking subject in his universality with the perfect competence which the Saussurian

tradition granted him tacitly: ‘Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention or interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance. This seems to me to have been the position of the founders of modern general linguistics, and no cogent reason for modifying it has been offered.'² In short, from this standpoint, Chomskyan ’competence’ is simply another name for Saussure’s langue.^ Corresponding to language as a ‘universal treasure’, as the collective property of the whole group, there is linguistic competence as the 'deposit' of this ’treasure’ in each individual or as the participation of each member of the ’linguistic community’ in this public good. The shift in vocabulary conceals the fictio juris through which Chomsky, converting the immanent laws of legitimate discourse into universal norms of correct linguistic practice, sidesteps the question of the economic and social conditions of the acquisition of the legitimate competence and of the constitution of the market in which this definition of the legitimate and the illegitimate is established and imposed.⁴

Official Language and Political Unity

As a demonstration of how linguists merely incorporate into their theory a pre-constructed object, ignoring its social laws of construction and masking its social genesis, there is no better example than the passage in his Course in General Linguistics in which Saussure discusses the relation between language and space.⁵ Seeking to prove that it is not space which defines language but language which defines its space, Saussure observes that neither dialects nor languages have natural limits, a phonetic innovation (substitution of ‘s’ for Latin ‘c’, for example) determining its own area of diffusion by the intrinsic force of its autonomous logic, through the set of speaking subjects who are willing to make themselves its bearers. This philosophy of history, which makes the internal dynamics of a language the sole principle of the limits of its diffusion, conceals the properly political process of unification whereby a determinate set of ‘speaking subjects’ is led in practice to accept the official language.

Saussure’s langue, a code both legislative and communicative which exists and subsists independently of its users (‘speaking

subjects’) and its uses (parole), has in fact all the properties commonly attributed to official language. As opposed to dialect, it has benefited from the institutional conditions necessary for its generalized codification and imposition. Thus known and recognized (more or less completely) throughout the whole jurisdiction of a certain political authority, it helps in turn to reinforce the authority which is the source of its dominance. It does this by ensuring among all members of the ‘linguistic community’, traditionally defined, since Bloomfield, as a ‘group of people who use the same system of linguistic signs’,⁶ the minimum of communication which is the precondition for economic production and even for symbolic domination.

To speak of the language, without further specification, as linguists do, is tacitly to accept the official definition of the official language of a political unit. This language is the one which, within the territorial limits of that unit, imposes itself on the whole population as the only legitimate language, especially in situations that are characterized in French as more officielle (a very exact translation of the word ‘formal’ used by English-speaking linguists).⁷ Produced by authors who have the authority to write, fixed and codified by grammarians and teachers who are also charged with the task of inculcating its mastery, the language is a code, in the sense of a cipher enabling equivalences to be established between sounds and meanings, but also in the sense of a system of norms regulating linguistic practices.

The official language is bound up with the state, both in its genesis and in its social uses. It is in the process of state formation that the conditions are created for the constitution of a unified linguistic market, dominated by the official language. Obligatory on official occasions and in official places (schools, public administrations, political institutions, etc.), this state language becomes the theoretical norm against which all linguistic practices are objectively measured. Ignorance is no excuse; this linguistic law has its body of jurists - the grammarians — and its agents of regulation and imposition - the teachers - who are empowered universally to subject the linguistic performance of speaking subjects to examination and to the legal sanction of academic qualification.

In order for one mode of expression among others (a particular language in the case of bilingualism, a particular use of language in the case of a society divided into classes) to impose itself as the only legitimate one. the linguistic market has to be unified and the different dialects (of class, region or ethnic group) have to be measured practically against the legitimate language or usage.

Integration into a single ‘linguistic community’, which is a product of the political domination that is endlessly reproduced by institutions capable of imposing universal recognition of the dominant language, is the condition for the establishment of relations of linguistic domination.

The ‘Standard’ Language: A ‘Normalized’ Product

Like the different crafts and trades which, before the advent of large-scale industry, constituted, in Marx’s phrase, so many separate ‘enclosures’, local variants of the langue d’oil differed from one parish to another until the eighteenth century. This is still true today of the regional dialects and, as the dialecticians’ maps show, the phonological, morphological and lexicological features are distributed in patterns which are never entirely superimposable and which only ever correspond to religious or administrative boundaries through rare coincidence.⁸ In fact, in the absence of objectification in writing and especially of the quasi-legal codification which is inseparable from the constitution of an official language, ‘languages’ exist only in the practical state, i.e. in the form of so many linguistic habitus which are at least partially orchestrated, and of the oral productions of these habitus.⁹ So long as a language is only expected to ensure a minimum of mutual understanding in the (very rare) encounters between people from neighbouring villages or different regions, there is no question of making one usage the norm for another (despite the fact that the differences perceived may well serve as pretexts for declaring one superior to the other).

Until the French Revolution, the process of linguistic unification went hand in hand with the process of constructing the monarchical state. The ‘dialects’, which often possessed some of the properties attributed to ‘languages’ (since most of them were used in written form to record contracts, the minutes of local assemblies, etc.), and literary languages (such as the poetic language of the pays d'oc), like artificial languages distinct from each of the dialects used over the whole territory in which they were current, gave way progressively, from the fourteenth century on, at least in the central provinces of the pays d’oil, to the common language which was developed in Paris in cultivated circles and which, having been promoted to the status of official language, was used in the form given to it by scholarly, i.e. written, uses. Correiatively, the popular and purely oral uses of all the regional dialects which had thus been supplanted degenerated into patois, as a result of the compartmentaliza-

lion (linked to the abandonment of the written form) and internal disintegration (through lexical and syntactic borrowing) produced by the social devaluation which they suffered. Having been abandoned to the peasants, they were negatively and pejoratively defined in opposition to distinguished or literate usages. One indication of this, among many others, is the shift in the meaning assigned to the word patois,, which ceased to mean ‘incomprehensible speech* and began to refer to ‘corrupted and coarse speech, such as (hat of the common people’ (Furc-tiere's Dictionary’, 1690),

The linguistic situation was very different in the langue d^roc regions. Not until the sixteenth century, with the progressive constitution of an administrative organization linked to royal power (involving the appearance of a multitude of subordinate administrative agents, lieutenants, provosts, magistrates, etc,), did the Parisian dialect begin to take over from the various langue d oc dialects in legal documents. The imposition of French as the official language did not result in the total abolition of the written use of dialects, whether in administrative, political or even literary texts (dialect literature continued to exist during the ancien regime), and their oral uses remained predominant, A situation of bilingualism tended to arise. Whereas the lower classes, particularly the peasantry, were limited to the local dialect, the aristocracy, the commercial and business bourgeoisie and particularly the literate petite bourgeoisie (precisely those who responded to Abbe Gregoire’s survey and who had, to varying degrees, attended the Jesuit colleges, which were institutions of linguistic unification) had access much more frequently to the use of the official language, written or spoken, while at the same time possessing the dialect (which was still used in most private and even public situations), a situation in which they were destined to fulfil the function of intermediaries

The members of these local bourgeoisies of priests, doctors or teachers, who owed their position to their mastery of the instruments of expression, had everything to gain from the Revolutionary policy of linguistic unification. Promotion of the official language to the status of national language gave them that de facto monopoly of politics, and more generally of communication with the central government and its representatives. that has defined local notables under all the French republics.

The imposition of the legitimate language in opposition to the dialects and patois was an integral part of (he political strategies aimed at perpetuating the gains of the Revolution through the production and the reproduction of the ‘new man’. Condillac's theory, which saw language as a method, made it possible to identify revolutionary language with revolutionary thought. To reform language, to purge it of the usages linked to the old society and impose it in its purified form, was to impose a thought that would itself be purged and purified. It would be naive to attribute the policy of linguistic unification solely to the technical needs of communication between the different parts of the territory, particular-

ly between Paris and the provinces, or to see it as the direct product of a slate centralism determined to crush ‘local characteristics'. The conflict between the French of the revolutionary intelligentsia and the dialects or patois was a struggle for symbolic power in which what was al stake was the formation and re-formation of mental structures. In short, it was not only a question of communicating but of gaining recognition for a new language of authority, with its new political vocabulary, its terms of address and reference, its metaphors, its euphemisms and the representation of the social world which it conveys, and which, because it is linked to the new interests of new groups, is inexpressible in the local idioms shaped by usages linked io the specific interests of peasant groups.

Thus, only when the making of the ‘nation', an entirely abstract group based on law, creates new usages and functions does it become indispensable to forge a standard language, impersonal and anonymous like the official uses it has to serve, and by the same token to undertake the work of normalizing the products of the linguistic habitus. The dictionary is the exemplary result of this labour of codification and normalization. It assembles, by scholarly recording, the totality of the linguistic resources accumulated in the course of time and, in particular, all the possible uses of the same word (or all the possible expressions of the same sense), juxtaposing uses that are socially at odds, and even mutually exclusive (to the point of marking those which exceed the bounds of acceptability with a sign of exclusion such as Obs., Coll, or SI.}. It thereby gives a fairly exact image of language as Saussure understands it, ‘the sum of individual treasuries of language', which is predisposed to fulfil the functions of a ‘universal’ code. The normalized language is capable of functioning outside the constraints and without the assistance of the situation, and is suitable for transmitting and decoding by any sender and receiver, who may know nothing of one another. Hence it concurs with the demands of bureaucratic predictability and calculability, which presuppose universal functionaries and clients, having no other qualities than those assigned to them by the administrative definition of their condition.

In the process which leads to the construction, legitimation and imposition of an official language, the educational system plays a decisive role: ‘fashioning the similarities from which that community of consciousness which is the cement of the nation stems.’ And Georges Davy goes on to slate the function of the schoolmaster, a maitre a parler (teacher of speaking) who is thereby also a maitre a

penser (teacher of thinking): 'He [the primary school teacher], by virtue of his function, works daily on the faculty of expression of every idea and every emotion: on language. In teaching the same clear, fixed language to children who know it only very vaguely or who even speak various dialects or patois, he is already inclining them quite naturally to see and feel things in the same way: and he works to build the common consciousness of the nation’.¹⁰ The Whorfian-or, if you like. Humboldtian¹¹ - theory of language which underlies this view of education as an instrument of ‘intellectual and moral integration', in Durkheim’s sense, has an affinity with the Durkheimian theory of consensus, an affinity which is also indicated by the shift of the word ‘code’ from law to linguistics. The code, in the sense of cipher, that governs written language, which is identified with correct language, as opposed to the implicitly inferior conversational language, acquires the force of law in and through the educational system.¹"

The educational system, whose scale of operations grew in extent and intensity throughout the nineteenth century.¹³ no doubt directly helped to devalue popular modes of expression, dismissing them as ‘slang' and ‘gibberish' (as can be seen from teachers’ marginal comments on essays) and to impose recognition of the legitimate language. But it was doubtless the dialectical relation between the school system and the labour market - or, more precisely, between the unification of the educational (and linguistic) market, linked to the introduction of educational qualifications valid nation-wide, independent (at least officially) of the social or regional characteristics of their bearers, and the unification of the labour market (including the development of the state administration and the civil service) — which played the most decisive role in devaluing dialects and establishing the new hierarchy of linguistic practices.¹⁴ To induce the holders of dominated linguistic competences to collaborate in the destruction of their instruments of expression, by endeavouring for example to speak "French’ to their children or requiring them to speak ‘French’ at home, with the more or less explicit intention of increasing their value on the educational market. it was necessary for the school system to be perceived as the principal (indeed, the only) means of access to administrative positions which were all the more attractive in areas where industrialization was least developed. This conjunction of circumstances was found in the regions of ‘dialect’ (except the east of France) rather than in the patois regions of northern France.

Unification of the Market and Symbolic Domination

In fact, while one must not forget the contribution which the political will to unification (also evident in other areas, such as law) makes to the construction of the language which linguists accept as a natural datum, one should not regard it as the sole factor responsible for the generalization of the use of the dominant language. This generalization is a dimension of the unification of the market in symbolic goods which accompanies the unification of the economy and also of cultural production and circulation. This is seen clearly in the case of the market in matrimonial exchanges, in which ‘products’ which would previously have circulated in the protected enclosure of local markets, with their own laws of price formation, are suddenly devalued by the generalization of the dominant criteria of evaluation and the discrediting of ‘peasant values’, which leads to the collapse of the value of the peasants, who are often condemned to celibacy. Visible in all areas of practice (sport, song, clothing, housing, etc.), the process of unification of both the production and the circulation of economic and cultural goods entails the progressive obsolescence of the earlier mode of production of the habitus and its products. And it is clear why, as sociolinguists have often observed, women are more disposed to adopt the legitimate language (or the legitimate pronunciation): since they are inclined towards docility with regard to the dominant usages both by the sexual division of labour, which makes them specialize in the sphere of consumption, and by the logic of marriage, which is their main if not their only avenue of social advancement and through which they circulate upwards, women are predisposed to accept, from school onwards, the new demands of the market in symbolic goods.

Thus the effects of domination which accompany the unification of the market are always exerted through a whole set of specific institutions and mechanisms, of which the specifically linguistic policy of the state and even the overt interventions of pressure groups form only the most superficial aspect. The fact that these mechanisms presuppose the political or economic unification which they help in turn to reinforce in no way implies that the progress of the official language is to be attributed to the direct effectiveness of legal or quasi-legal constraints. (These can at best impose the acquisition, but not the generalized use and therefore the autonomous reproduction, of the legitimate language.) All symbolic domination presupposes, on the part of those who submit to it, a

form of complicity which is neither passive submission to external constraint nor a free adherence to values. The recognition of the legitimacy of the official language has nothing in common with an explicitly professed, deliberate and revocable belief, or with an intentional act of accepting a 'norm’. It is inscribed, in a practical state, in dispositions which are impalpably inculcated, through a long and slow process of acquisition, by the sanctions of the linguistic market, and which are therefore adjusted, without any cynical calculation or consciously experienced constraint, to the chances of material and symbolic profit which the laws of price formation characteristic of a given market objectively offer to the holders of a given linguistic capital,¹⁵

The distinctiveness of symbolic domination lies precisely in the fact that it assumes, of those who submit to it, an attitude which challenges the usual dichotomy of freedom and constraint. The ‘choices’ of the habitus (for example, using the 'received' uvular T' instead of the rolled ‘r’ in the presence of legitimate speakers) are accomplished without consciousness or constraint, by virtue of the dispositions which, although they are unquestionably the product of social determinisms, are also constituted outside the spheres of consciousness and constraint. The propensity to reduce the search for causes to a search for responsibilities makes it impossible to see that intimidation, a symbolic violence which is not aware of what it is (to the extent that it implies no act of intimidation) can only be exerted on a person predisposed (in his habitus) to feel it, whereas others will ignore it. It is already partly true to say that the cause of the timidity lies in the relation between the situation or the intimidating person (who may deny any intimidating intention) and the person intimidated, or rather, between the social conditions of production of each of them. And little by little, one has to take account thereby of the whole social structure.

There is every reason to think that the factors which are most influential in the formation of the habitus are transmitted without passing through language and consciousness, but through suggestions inscribed in the most apparently insignificant aspects of the things, situations and practices of everyday life. Thus the modalities of practices, the ways of looking, sitting, standing, keeping silent, or even of speaking (‘reproachful looks’ or ‘tones’, 'disapproving glances' and so on) are full of injunctions that are powerful and hard to resist precisely because they are silent and insidious, insistent and insinuating. (It is this secret code which is explicitly denounced in the crises characteristic of the domestic unit, such as marital or teenage

crises: the apparent disproportion between the violence of the revolt and the causes which provoke it stems from the fact that the most anodyne actions or words are now seen for what they are — as injunctions, intimidations, warnings, threats - and denounced as such, all the more violently because they continue to act below the level of consciousness and beneath the very revolt which they provoke.) The power of suggestion which is exerted through things and persons and which, instead of telling the child what he must do, tells him what he is. and thus leads him to become durably what he has to be, is the condition for the effectiveness of all kinds of symbolic power that will subsequently he able to operate on a habitus predisposed to respond to them. The relation between two people may be such that one of them has only to appear in order to impose on the other, without even having to want to. let alone formulate any command, a definition of the situation and of himself (as intimidated, for example), which is all the more absolute and undisputed for not having to be stated.

The recognition extorted by this invisible, silent violence is expressed in explicit statements, such as those which enable Labov to establish that one finds the same evaluation of the phoneme ‘r’ among speakers who come from different classes and who therefore differ in their actual production of T’, But it is never more manifest than tn all the corrections, whether ad hoc or permanent, to which dominated speakers, as they strive desperately for correctness, consciously or unconsciously subject the stigmatized aspects of their pronunciation, their diction (involving various forms of euphemism) and their syntax, or in the disarray which leaves them ’speechless’, ‘tongue-tied’, ‘at a loss for words’, as if they were suddenly dispossessed of their own language.¹⁶

Distinctive Deviations and Social Value

Thus, if one fails to perceive both the special value objectively accorded to the legitimate use of language and the social foundations of this privilege, one inevitably falls into one or other of two opposing errors. Either one unconsciously absolutizes that which is objectively relative and in that sense arbitrary, namely the dominant usage, failing to look beyond the properties of language itself, such as the complexity of its syntactic structure, in order to identify the basis of the value that is accorded to it. particularly in the educational market; or one escapes this form of fetishism only to fall into the

naivety par excellence of the scholarly relativism which forgets that the naive gaze is not relativist, and ignores the fact of legitimacy, through an arbitrary relativization of the dominant usage, which is socially recognized as legitimate, and not only by those who are dominant.

To reproduce in scholarly discourse the fetishizing of the legitimate language which actually rakes place in society, one only has to follow the example of Basil Bernstein, who describes the properties of the ‘elaborated code' without relating this social product to the social conditions of its production and reproduction, or even, as one might expect from the sociology of education, to its academic conditions. The ‘elaborated code' is thus constituted as the absolute norm of all linguistic practices which then can only be conceived in terms of the logic of deprivation. Conversely, ignorance of what popular and educated usage owe to their objective relations and to the structure of the relation of domination between classes, which they reproduce in their own logic, leads to the canonization as such of the ‘language* of the dominated classes. Lahov leans in this direction when his concern to rehabilitate ‘popular speech' against the theorists of deprivation leads him to contrast the verbosity and pompous verbiage of middle-class adolescents with the precision and conciseness of black children from the ghettos. This overlooks the fact that, as he himself has shown (with the example of recent immigrants who judge deviant accents, including their own, with particular severity), the linguistic ‘norm* is imposed on all members of the same linguistic community', most especially in the educational market and in all formal situations in which verbosity is often de rigueur.

Political unification and the accompanying imposition of an official language establish relations between the different uses of the same language which differ fundamentally from the theoretical relations (such as that between mouton and "sheep’ which Saussure cites as the basis for the arbitrariness of the sign) between different languages, spoken by politically and economically independent groups. All linguistic practices are measured against the legitimate practices, i.e. the practices of those who are dominant. The probable value objectively assigned to the linguistic productions of different speak-ers and therefore the relation which each of them can have to the language, and hence to his own production, is defined within the system of practically competing variants which is actually established whenever the extra-linguistic conditions for the constitution of a linguistic market are fulfilled.

Thus, for example, the linguistic differences between people from

different regions cease to be incommensurable particularisms. Measured de facto against the single standard of the ‘common’ language, they are found wanting and cast into the outer darkness of regional-isms, the ‘corrupt expressions and mispronunciations’ which schoolmasters decry.¹' Reduced to the status of quaint or vulgar jargons, in either case unsuitable for formal occasions, popular uses of the official language undergo a systematic devaluation. A system of sociologically pertinent linguistic oppositions tends to be constituted, which has nothing in common with the system of linguistically pertinent linguistic oppositions. In other words, the differences which emerge from the confrontation of speech varieties are not reducible to those the linguist constructs in terms of his own criterion of pertinence. However great the proportion of the functioning of a language that is not subject to variation, there exists, in the area of pronunciation, diction and even grammar, a whole set of differences significantly associated with social differences which, though negligible in the eyes of the linguist, are pertinent from the sociologist’s standpoint because they belong to a system of linguistic oppositions which is the re-translation of a system of social differences. A structural sociology of language, inspired by Saussure but constructed in opposition to the abstraction he imposes, must take as its object the relationship between the structured systems of sociologically pertinent linguistic differences and the equally structured systems of social differences.

The social uses of language owe their specifically social value to the fact that they tend to be organized in systems of differences (between prosodic and articulatory or lexical and syntactic variants) which reproduce, in the symbolic order of differential deviations, the system of social differences. To speak is to appropriate one or other of the expressive styles already constituted in and through usage and objectively marked by their position in a hierarchy of styles which expresses the hierarchy of corresponding social groups. These styles, systems of differences which are both classified and classifying, ranked and ranking, mark those who appropriate them. And a spontaneous stylistics, armed with a practical sense of the equivalences between the two orders of differences, apprehends social classes through classes of stylistic indices.

In emphasizing the linguistically pertinent constants at the expense of the sociologically significant variations in order to construct that artefact which is the ‘common’ language, the linguist proceeds as if the capacity to speak, which is virtually universal, could be identified with the socially conditioned way of realizing this natural capacity,

which presents as many variants as there are social conditions of acquisition. The competence adequate to produce sentences that are likely to be understood may be quite inadequate to produce sentences that are likely to be listened to. likely to be recognized as acceptable in all the situations in which there is occasion to speak. Here again, social acceptability is not reducible to mere grammaticality. Speakers lacking the legitimate competence are de facto excluded from the social domains in which this competence is required, or are condemned to silence. What is rare, then, is not the capacity to speak, which, being part of our biological heritage, is universal and therefore essentially non-distinctive, but rather the competence necessary in order to speak the legitimate language which, depending on social inheritance, re-translates social distinctions into the specifically symbolic logic of differential deviations, or. in short, distinction.¹⁹

The constitution of a linguistic market creates the conditions for an objective competition in and through which the legitimate competence can function as linguistic capital, producing a profit of distinction on the occasion of each social exchange. Because it derives in part from the scarcity of the products (and of I he corresponding competences), this profit does not correspond solely to the cost of training.

The cost of training is not a simple, socially neutral notion. To an extent which varies depending on national traditions in education, the historical period and the academic discipline in question, it includes expenditure which may far exceed the minimum ’technically’ required in order to ensure the transmission of the strictly defined competence (if indeed it is possible to give a purely technical definition of the training necessary and sufficient to fulfil a function and of the function itself, bearing in mind that 'role distance’- distance from the (unction - enters increasingly into the definition of the function as one moves up the hierarchy of functions). In some cases, for example, the duration of study (which provides a good measure of the economic cost of training) tends to be valued for its own sake, independently of the result it produces (encouraging, among the ’elite schools', a kind of competition in the sheer length of courses). In other cases - not that the two options are mutually exclusive - the social quality of the competence acquired, which is reflected in the symbolic modality of practices, i.e. in the manner of performing technical acts and implementing the competence, appears as inseparable from the slowness of the acquisition, short or 'crash' courses always being suspected of leaving on their products the marks of 'cramming' or the stigmata of ‘catching up'. This conspicuous consumption of training (i.e.

of time), an apparent technical wastage which fulfils social functions of legitimation, enters into the value socially attributed to a socially guaranteed competence (which means, nowadays, one 'certified' by the educational system).

Since the profit of distinction results from the fact that the supply of products (or speakers) corresponding to a given level of linguistic (or, more generally, cultural) qualification is lower than it would be if all speakers had benefited from the conditions of acquisition of the legitimate competence to the same extent as the holders of the rarest competence,²*¹ it is logically distributed as a function of the chances of access to these conditions, that is. as a function of the position occupied in the social structure.

Despite certain appearances, we could not be further from the Saussu-rian model of homo Unguisticus who, like the economic subject in the Walrasian tradition, is formally free to do as he likes in his verbal productions (free, for example, to say 'tat' for ‘hat’, as children do) but can be understood, can exchange and communicate only on condition that he conforms to the rules of the common code. This market, which knows only pure, perfect competition among agents who are as interchangeable as the products they exchange and the ‘situations' in which they exchange, and who are all identically subject to the principle of the maximization of informative efficiency (analogous to the principle of the maximization of utilities), is, as will shortly become clearer, as remote from the real linguistic market as the ‘pure’ market of the economists is from the real economic market, with its monopolies and oligopolies.

Added to the specific effect of distinctive rarity is the fact that, by virtue of the relationship between the system of linguistic differences and the system of economic and social differences, one is dealing not with a relativistic universe of differences capable of relativizing one another, but with a hierarchical universe of deviations with respect to a form of speech that is (virtually) universally recognized as legitimate, i.e. as the standard measure of the value of linguistic products. The dominant competence functions as linguistic capital, securing a profit of distinction in its relation to other competences only in so far as certain conditions (the unification of the market and the unequal distribution of the chances of access to the means of production of the legitimate competence, and to the legitimate places of expression) are continuously fulfilled, so that the groups which possess that competence are able to impose it as the only legitimate one in the formal markets (the fashionable, educational,

political and administrative markets) and in most of the linguistic interactions in which they are involved.²¹

It is for this reason that those who seek to defend a threatened linguistic capital, such as knowledge of the classical languages in present-day France, are obliged to wage a total struggle. One cannot save the value of a competence unless one saves the market, in other words, the whole set of political and social conditions of production of the producers/consumers. The defenders of Latin or, in other contexts, of French or Arabic, often talk as if the language they favour could have some value outside the market, by intrinsic virtues such as its ‘logical’ qualities; but, in practice, they are defending the market. The position which the educational system gives to the different languages (or the different cultural contents) is such an important issue only because this institution has the monopoly in the large-scale production of producers/consumers, and therefore in the reproduction of the market without which the social value of the linguistic competence, its capacity to function as linguistic capital, would cease to exist.

The Literary Field and the Struggle for Linguistic Authority

Thus, through the medium of the structure of the linguistic field, conceived as a system of specifically linguistic relations of power based on the unequal distribution of linguistic capital (or, to put it another way, of the chances of assimilating the objectified linguistic resources), the structure of the space of expressive styles reproduces in its own terms the structure of the differences which objectively separate conditions of existence- In order fully to understand the structure of this field and, in particular, the existence, within the field of linguistic production, of a sub-field of restricted production which derives its fundamental properties from the fact that the producers within it produce first and foremost for other producers, it is necessary to distinguish between the capital necessary for the simple production of more or less legitimate ordinary speech, on the one hand, and the capital of instruments of expression (presupposing appropriation of the resources deposited in objectified form in libraries - books, and in particular in the ‘classics', grammars and dictionaries) which is needed to produce a written discourse worthy of being published, that is to say, made official, on the other. This production of instruments of production, such as rhetorical devices,

genres, legitimate styles and manners and, more generally, all the formulations destined to be ‘authoritative’ and to be cited as examples of ‘good usage’, confers on those who engage in it a power over language and thereby over the ordinary users of language, as well as over their capital.

The legitimate language no more contains within itself the power to ensure its own perpetuation in time than it has the power to define its extension in space. Only the process of continuous creation, which occurs through the unceasing struggles between the different authorities who compete within the field of specialized production for the monopolistic power to impose the legitimate mode of expression, can ensure the permanence of the legitimate language and of its value, that is, of the recognition accorded to it. It is one of the generic properties of fields that the struggle for specific stakes masks the objective collusion concerning the principles underlying the game. More precisely, the struggle tends constantly to produce and reproduce the game and its stakes by reproducing, primarily in those who are directly involved, but not in them alone, the practical commitment to the value of the game and its stakes which defines the recognition of legitimacy. What would become of the literary world if one began to argue, not about the value of this or that author’s style, but about the value of arguments about style? The game is over when people start wondering if the cake is worth the candle. The struggles among writers over the legitimate art of writing contribute, through their very existence, to producing both the legitimate language, defined by its distance from the ‘common’ language, and belief in its legitimacy.

It is not a question of the symbolic power which writers, grammarians or teachers may exert over the language in their personal capacity, and which is no doubt much more limited than the power they can exert over culture (for example, by imposing a new definition of legitimate literature which may transform the ‘market situation’). Rather, it is a question of the contribution they make, independently of any intentional pursuit of distinction, to the production, consecration and imposition of a distinct and distinctive language. In the collective labour which is pursued through the struggles for what Horace called arbitrium et jus et norma loquendi, writers - more or less authorized authors - have to reckon with the grammarians, who hold the monopoly of the consecration and canonization of legitimate writers and writing. They play their part in constructing the legitimate language by selecting, from among the products on offer, those which seem to them worthy of being consecrated and incorporated into the legitimate competence through educational

inculcation, subjecting them, for this purpose, to a process of normalization and codification intended to render them consciously assimilable and therefore easily reproducible. The grammarians, who, for their part, may find allies among establishment writers and in the academies, and who take upon themselves the power to set up and impose norms, tend to consecrate and codify a particular use of language by rationalizing it and ‘giving reason' to it. in so doing they help to determine the value which the linguistic products of the different users of the language will receive in the different markets - particularly those most directly subject to their control, such as the educational market - by delimiting the universe of acceptable pronunciations, words or expressions, and fixing a language censored and purged of all popular usages, particularly the most recent ones.

The variations corresponding to the different configurations of the relation of power between the authorities, who constantly clash in the field of literary production by appealing to very different principles of legitimation, cannot disguise the structural invariants which, in the most diverse historical situations, impel the protagonists to resort to the same strategies and the same arguments in order to assert and legitimate their right to legislate on language and in order to denounce the claims of their rivals, Thus, against the ‘fine style' of high society and the writers’ claim to possess an instinctive art of good usage, the grammarians always invoke ‘reasoned usage', the ‘feel for the language' which comes from knowledge of the principles of ‘reason’ and 'taste' which constitute grammar. Conversely, the writers, whose pretensions were most confidently expressed during the Romantic period, invoke genius against the rule, flouting the injunctions of those whom Hugo disdainfully called 'grammatists'.“

The objective dispossession of the dominated classes may never be intended as such by any of the actors engaged in literary struggles (and there have, of course, always been writers who, like Hugo, claimed to ‘revolutionize dictionaries' or who sought to mimic popular speech). The fact remains that this dispossession is inseparable from the existence of a body of professionals, objectively invested with the monopoly of the legitimate use of the legitimate language, who produce for their own use a special language predisposed to fulfil, as a by-product, a social function of distinction in the relations between classes and in the struggles they wage on the terrain of language. It is not unconnected, moreover, with the existence of the educational system which, charged with the task of sanctioning heretical products in the name of grammar and inculcating the specific norms which block the effects of the laws of evolution, contributes significantly to constituting the dominated

uses of language as such by consecrating the dominant use as the only legitimate one, by the mere fact of inculcating it. Bui one would obviously be missing the essential point if one related the activity of artists or teachers directly to the effect to which it objectively contributes, namely, the devaluation of the common language which results from the very existence of a literary language. Those who operate in the literary field contribute to symbolic domination only because the effects that their position in the field and its associated interests lead them to pursue always conceal from themselves and from others the external effects which are a by-product of this very misrecognition.

The properties which characterize linguistic excellence may be summed up in two words: distinction and correctness. The work performed in the literary field produces the appearances of an original language by resorting to a set of derivations whose common principle is that of a deviation from the most frequent, i.e. 'common', ‘ordinary’, ‘vulgar’, usages. Value always arises from deviation, deliberate or not, with respect to the most widespread usage, ‘commonplaces’, ‘ordinary sentiments’, 'trivial' phrases, ‘vulgar’ expressions, 'facile' style. In the uses of language as in life-styles, all definition is relational. Language that is 'recherche, ‘well chosen', ‘elevated’, ‘lofty’, 'dignified' or ‘distinguished’ contains a negative reference (the very words used to name it show this) to 'common' ‘everyday’, ‘ordinary’, ‘spoken’, ‘colloquial’, ‘familiar’ language and, beyond this, to ‘popular’, ‘crude’, ‘coarse’, ‘vulgar’, 'sloppy', ‘loose’, ‘trivial’, 'uncouth' language (not to mention the unspeakable, ‘gibberish’, ‘pidgin’ or ‘slang’). The oppositions from which this series is generated, and which, being derived from the legitimate language, is organized from the standpoint of the dominant users, can be reduced to two: the opposition between 'distinguished' and ‘vulgar* (or ‘rare’ and ‘common’) and the opposition between ‘tense’ (or ‘sustained’) and ‘relaxed’ (or ‘loose’), which no doubt represents the specifically linguistic version of the first, very general, opposition. It is as if the principle behind the ranking of class languages were nothing other than the degree of control they manifested and the intensity of the correctness they presupposed.

It follows that the legitimate language is a semi-artificial language which has to be sustained by a permanent effort of correction, a task which falls both to institutions specially designed for this purpose and to individual speakers. Through its grammarians, who fix and codify legitimate usage, and its teachers who impose and inculcate it through innumerable acts of correction, the educational system

tends, in this area as elsewhere, to produce the need (or its own services and its own products, i.e. the labour and instruments of correction.²⁴ The legitimate language owes its (relative) constancy in time (as in space) to the fact that it is continuously protected by a prolonged labour of inculcation against the inclination towards the economy of effort and tension which leads, for example, to analogical simplification (c.g. of irregular verbs in French - vous faisez and volts disez for vousfaites and vous dites). Moreover, the correct, i.e. corrected, expression owes the essential part of its social properties to the fact that it can be produced only by speakers possessing practical mastery of scholarly rales, explicitly constituted by a process of codification and expressly inculcated through pedagogic work. Indeed, the paradox of all institutionalized pedagogy is that it aims to implant, as schemes that function in a practical state, rules which grammarians have laboured to extract from the practice of the professionals of written expression (from the past), by a process of retrospective formulation and codification. 'Correct usage' is the product of a competence which is an incorporated grammar, the word grammar being used explicitly (and not tacitly, as it is by the linguists) in its true sense of a system of scholarly rales, derived ex post facto from expressed discourse and set up as imperative norms for discourse yet to be expressed. It follows that one cannot fully account for the properties and social effects of the legitimate language unless one takes account, not only of the social conditions of the production of literary language and its grammar, but also of the social conditions in which this scholarly code is imposed and inculcated as the principle of the production and evaluation of speech.²⁵

The Dynamics of the Linguistic Field

The laws of the transmission of linguistic capital are a particular case of the laws of the legitimate transmission of cultural capital between the generations, and it may therefore be posited that the linguistic competence measured by academic criteria depends, like the other dimensions of cultural capital, on the level of education (measured in terms of qualifications obtained) and on the social trajectory. Since mastery of the legitimate language may be acquired through familiarization, that is, by more or less prolonged exposure to the legitimate language, or through the deliberate inculcation of explicit rules, the major classes of modes of expression correspond to classes

of modes of acquisition, that is, to different forms of the combination between the two principal factors of production of the legitimate competence, namely, the family and the educational system.

In this sense, like the sociology of culture, the sociology of language is logically inseparable from a sociology of education. As a linguistic market strictly subject to the verdicts of the guardians of legitimate culture, the educational market is strictly dominated by the linguistic products of the dominant class and tends to sanction the pre-existing differences in capital. The combined effect of low cultural capital and the associated low propensity to increase it through educational investment condemns the least favoured classes to the negative sanctions of the scholastic market, i.e. exclusion or early self-exclusion induced by lack of success. The initial disparities therefore tend to be reproduced since the length of inculcation tends to vary with its efficiency: those least inclined and least able to accept and adopt the language of the school are also those exposed for the shortest time to this language and to educational monitoring, correction and sanction.

Given that the educational system possesses the delegated authority necessary to engage in a universal process of durable inculcation in matters of language, and given that it tends to vary the duration and intensity of this inculcation in proportion to inherited cultural capital, it follows that the social mechanisms of cultural transmission tend to reproduce the structural disparity between the very unequal knowledge of the legitimate language and the much more uniform recognition of this language. This disparity is one of the determinant factors in the dynamics of the linguistic field and therefore in changes in the language. For the linguistic struggles which are the ultimate source of these changes presuppose that speakers have virtually the same recognition of authorized usage, but very unequal knowledge of this usage. Thus, if the linguistic strategies of the petite bourgeoisie, and in particular its tendency to hypercorrection - a very typical expression of ‘cultural goodwill’ which is manifested in all areas of practice - have sometimes been seen as the main factor in linguistic change, this is because the disparity between knowledge and recognition, between aspirations and the means of satisfying them - a disparity that generates tension and pretension - is greatest in the intermediate regions of the social space. This pretension, a recognition of distinction which is revealed in the very effort to deny it by appropriating it, introduces a permanent pressure into the field of competition which inevitably induces new strategies of distinction on the part of the holders of distinctive marks that are socially recognized as distinguished.

The petit-bourgeois hypercorrection which seeks its models and instruments of correction from the most consecrated arbiters of legitimate usage - Academicians, grammarians, teachers - is defined in the subjective and objective relationship to popular ‘vulgarity’ and bourgeois 'distinction'. Consequently, the contribution which this striving for assimilation (to the bourgeois classes) and. at the same time, dissimilation (with respect to the lower classes) makes to linguistic change is simply more visible than the dissimilation strategies which, in turn, it provokes from the holders of a rarer competence. Conscious or unconscious avoidance of the most visible marks of the linguistic tension and exertion of petit-bourgeois speakers (for example, in French, spoken use of the past historic, associated with old-fashioned schoolmasters) can lead the bourgeois and the intellectuals towards the controlled hypocorrection which combines confident relaxation and lofty ignorance of pedantic rules with the exhibition of ease on the most dangerous ground,²⁶ Showing tension where the ordinary speaker succumbs to relaxation, facility where he betrays effort, and the ease in tension which differs utterly from petit-bourgeois or popular tension and ease: these are all strategies of distinction (for the most part unconscious) giving rise to endless refinements, with constant reversals of value which tend to discourage the search for non-relational properties of linguistic styles.

Thus, in order to account for (he new style of speaking adopted by intellectuals, which can be observed in America as well as in France - a somewhat hesitant, even faltering, interrogative manner {‘nonT, ‘right?¹, "OK?' etc.) - one would have to lake into account the whole structure of usages in relation to which it is differentially defined. On the one hand, there is the old academic manner (with - tn French - its long periods, imperfect subjunctives, etc.), associated with a devalued image of the professorial role; on the other, the new petit-bourgeois usages resulting from wider diffusion of scholarly usage and ranging from ‘liberated¹ usage, a blend of tension and relaxation which tends to characterize the new petite bourgeoisie, to the hypercorrection of an over-refined speech, immediately devalued by an all-too-visible ambition, which is the mark of the upwardly mobile petite bourgeoisie.

The fact that these distinctive practices can be understood only in relation to the universe of possible practices does not mean that they have to be traced back to a conscious concern to distinguish oneself from them. There is every reason to believe that they are rooted in a practical sense of the rarity of distinctive marks (linguistic or otherwise) and of its evolution over time. Words which become

popularized lose their discriminatory power and thereby tend to be perceived as intrinsically banal, common, facile - or (since diffusion is linked to time) as worn out. It is no doubt the weariness deriving from repeated exposure which, combined with the sense of rarity, gives rise to the unconscious drift towards more ‘distinguished* stylistic features or towards rarer usages of common features.

Thus distinctive deviations are the driving force of the unceasing movement which, though intended to annul them, tends in fact to reproduce them (a paradox which is in no way surprising once one realizes that constancy may presuppose change). Not only do the strategies of assimilation and dissimilation which underlie the changes in the different uses of language not affect the structure of the distribution of different uses of language, and consequently the system of the systems of distinctive deviations (expressive styles) in which those uses are manifested, but they tend to reproduce it (albeit in a superficially different form). Since the very motor of change is nothing less than the whole linguistic field or. more precisely, the whole set of actions and reactions which are continuously generated in the universe of competitive relations constituting the field, the centre of this perpetual movement is everywhere and nowhere. Those who remain trapped in a philosophy of cultural diffusion based on a hydraulic imagery of ‘two-step flow’ or ‘trickle-down’, and who persist in locating the principle of change in a determinate site in the linguistic field, will always be greatly disappointed. What is described as a phenomenon of diffusion is nothing other than the process resulting from the competitive struggle which leads each agent, through countless strategies of assimilation and dissimilation {vis-a-vis those who are ahead of and behind him in the social space and in time) constantly to change his substantial properties (here, pronunciation, diction, syntactic devices, etc.), while maintaining, precisely by running in the race, the disparity which underlies the race. This structural constancy of the social values of the uses of the legitimate language becomes intelligible when one knows that the logic and the aims of the strategies seeking to modify it are governed by the structure itself, through the position occupied in the structure by the agent who performs them. The ‘interactionist’ approach, which fails to go beyond the actions and reactions apprehended in their directly visible immediacy, is unable to discover that the different agents’ linguistic strategies are strictly dependent on their positions in the structure of the distribution of linguistic capital, which can in turn be shown to depend, via the structure of chances of access to the educational system, on the structure of class relations.

Hence, interactionism can know nothing of the deep mechanisms which, through surface changes, tend to reproduce the structure of distinctive deviations and to maintain the profits accruing to those who possess a rare and therefore distinctive competence.