If the autonomy of a linguistic competence cannot, any more than can the pragmatics of its performances, be based on universals, perhaps one might consider that it corresponds to a certain transitory crystallisation of a state of language in relation to which individual performances will have to determine themselves?
How, then, is one to account for the nature of the constraints that ensure this stabilisation, constraints that phonologists have attributed to a structure that is intrinsic to language and generativists to hereditarily encoded universals?
What is the crystallisation of a linguistic power formation? One can understand nothing of this question if one represents power as just being a social superstructure. Power is not just micropolitical power, it is also the power of the superego, the famous power over oneself, which makes one shake with fear, which engenders somatisations, neuroses, suicides, etc. The stability of a ‘state of language’ certainly always corresponds to an equilibrium between these powers; these latter of course, are not arranged in relation to each other in no matter what way – it is not a question of an amorphous matter. Thus one can only account for the stabilisation of a ‘stratum of competence’ on condition that one render homogeneous domains that are as different as those of:
•the entirety of semiotisation activities (going from internal perceptions to modes of communication arising from the mass media);
•micropolitical levels (arising from the formation of bodies without organs);
•machinic indices and abstract machines (arising from the machinic phylum and the plane of consistency);
•which result in the putting into correspondence within each stratum of diverse systems of segmentarity and deterritorialising lines of flight.
Each pragmatic sequence involves a composition of powers at all levels and of every kind; its effectiveness depends on the dominant mode of semiotisation that it puts to work, due in particular to whether or not a diagrammatic semiotics liberates the functioning of certain abstract machines (financial, scientific, artistic, etc.). Thus we are led to define a micropolitical pragmatics as an activity of the assemblage of modes of semiotisation that everywhere exceed linguistic personology – towards corporeal intensities on the infra side, and towards the socius on the supra side. From this point of view, one would have to stop considering pragmatics as on the outskirts of syntax and semantics. Semiological (linguistic) pragmatics only represents a particular case of a more general semiotic pragmatics. The crystallisation of a signifying power, that we would put on the side of generative pragmatics (linguistic semiology), corresponds to a stratification of the libido, it coiling up into a system of redundancy of expression and redundancy of contents, the articulation of which has the effect of disempowering utterances of enclosing them either in the [worldliness] of an instituted power, or in an idiosyncratic system arising from madness or creation, for example. But before being stabilised as a language or a dialect, this kind of micropolitical competence is first experienced as a collective performance: every degree of fluidity is thus possible in the passage from an individual performance, even one that is marginal or delirious, to the completely sclerotic encoding of the dictionary or academic grammar kind. Besides naturalising the foundations of language, the brutal opposition of competence and performance squeezes [in English in the original] collective assemblages of enunciation – that is to say the groups that, in linguistic matters, are genuinely creative – to the profit of an alternative between an individuated or a universal subjectivity. One can approve the position of psycholinguists such as T.G. Bever, who consider that judgements of grammaticality are ‘forms of behavior like any other’1 without for all that falling into the trap of linguistic ‘psychologisation’. That a signifying grammaticalisation might take power over semiotic ensembles relative to capitalistic social fields, thus contributing to their stratification, doesn’t in the slightest imply that such ensembles can only be based on the universals that are supposed to rule over them. In fact, one is in the presence of the same type of universalisation procedure with a retroactive effect used by all power formations that wanted to give themselves the apparent legitimacy of divine right, and in particular those that sought to ‘justify’ the expansionism of capitalist exchangism. From the fact that one can always ‘structuralise’ monetary, linguistic, musical, etc. performances, that one can always discursivise them, binarise them, one considers that they have always been there, or even that their elements carried within themselves the seeds of the generation of the form of Capital, the Signifier, Music … But the process of power and the machinic mutations that have fixed and stabilised this form, furnished and delimited its creative potentialities, the metastable equilibria of its assemblages of enunciation and its group-subjects are, for their part, absolutely undecomposable, irreducible to a range of discrete and in principle discursive elements. If – as we will try to show later on – the abstract machines that are in question here can always be complexified, they can never, by contrast, be decomposed without losing their mutational specificity. And they aren’t [acquired] in little pieces, through learning or conditioning. They latch on to a process ready-made, they co-opt themselves in an assemblage that they can transform from top to bottom.
Abstract machines thus have nothing to do with the supposed ‘stages’ that it is claimed punctuate the ‘development’ of the child. The passage from one age in life to another doesn’t depend on the developmental programmes constructed by psychologists or psychoanalysts. It is linked to original reassemblages of different modes of encoding and semiotisation the nature and linking together of which cannot be determined a priori. The ‘stages’ in question have nothing automatic about them; as an individuated organic totality, the child only constitutes an intersection between the multiple material, socio-economic, semiotic, sets that traverse it.
The intrusion of the biological components of puberty in the life of an adolescent, for example, is inseparable from the micro-social context within which they appear; they trigger a series of machinic indices that have been put together elsewhere, they liberate a new abstract machine that will be manifested in the most varied of registers: the reorganisation of perceptual codes, turning in on oneself and/or poetic, cosmic, social externalisation, opposition to paternal values, etc. But in reality this triggering is not unilateral, other ‘external’ semiotic components can likewise accelerate, inhibit or reorient the effects of the biological semiotic effects of puberty. Under these conditions, where do the interactions between the biological and the social start and finish? Certainly not with the delimitation of an individual, considered as an organic totality or a sub-set of a family group! Little by little, all the machines of the socius are called into question by such phenomena, and reciprocally, biology in its entirety, at its most molecular level, is concerned with the interactions of the social field! On the plane of the individual, one thus ought not to separate the manifestations of puberty, considered in their organic, family, educational, context, from the upheavals which, on a broader social plane, call back into question the collective economy of desire. How can one fail to recognise that society in its entirety is constantly traversed, in its most intimate fibres, by these phenomena of biological change, which tirelessly sweeps childhood and adolescence away, generation after generation? It is true that the flights of desire of which they are the bearers are kept systematically in hand by the codes of the family, the school, medicine, sport, the army, and all the regulations and laws that are supposed to order the ‘normal’ behaviour of the individual. But it nevertheless happens that they managed to make collective machines of desire crystallise at the largest of scales (from neighbourhood gangs to Woodstock or May 68, etc.). And what were only scattered machinic indices, the quickly disempowered outlines of deterritorialisation, then become abstract machines able to catalyse new semiotic assemblages of desire.
Let’s return to the relative positions and functions of machinic indices, abstract machines, and semiotic assemblages on the basis of some different examples. In the first place, let us consider the embryonic writing that is manifested in the drawings of children, until the age of three or four. One can only talk about an index of writing here. Nothing is played out, nothing is crystallised, everything is possible still. But taken in charge by the educational machine, this index undergoes a profound reorganisation. Drawing loses its polyvocity. There is a disjunction between, on the one hand, drawing, which is impoverished and imitative, and, on the other, a writing that is directed entirely towards adult expression and is tyrannised by a concern for conformity with the dominant norms. How does the assemblage of semiotics in the school thus manage to take power over the child’s intensities of desire? Previously we have evoked the insufficiency of explanations that are content to consider the repressive action of equipments of power over the machinic indices ‘of’ the child. What one must try to grasp is why, in one case, this repression achieves its goal, and in another, it misses it. Once again, it seems to us to be impossible to avoid the intermediary instance constituted by abstract machines. If the crystallisation of an abstract machine tied to repression fails, the assemblage of power will lack its effect, subjects will become maladjusted, retarded, disturbed, psychotic etc., all things that supporters of the established order will blame on a deficit, although it would be easy to see that under non-repressive conditions, these same children incessantly enrich their ‘pre-school’ semiotic creativity. The passage to the stage of ‘working normally’ in class, the acquisition of an average competence in matters of reading, writing and arithmetic, etc., thus do not depend on the mechanical triggering of sensori-motor schemas internalised in the course of various ‘stages’ in the development of language. The stages in question here are not of a psychogenetic but of a repressive-genetic order. Instead of considering a ‘latency period’ which is as if destined, with the ‘waning of the Oedipus complex’ to punctuate the child’s life, it would doubtless be more advisable to study concrete social constellations and their particular technologies of semiotic subjection, in so far as they lead to the child’s surrounding by the family and education at a decisive moment of his ‘entering the world’ (one might talk here of a ‘school-barracks’ complex, to borrow Fernand Oury’s expression).
Abstract machines, which the supposed ‘psychogenetic’ stages put into play, cannot be assimilated to general schema at the level of perception, memory, logical integration, the structure of behaviour. In fact, they crystallise heteroclite components, they mix up ‘regressive fixations’ and archaic modes of territorialisation with ultra-deterritorialised semiotic components. A child who wets the bed, for example, comes up against an abstract formula – a body without organs – in which, in the same repressive formulation, a postural semiotics directed towards a turning in on oneself is associated with an affective semiotics directed towards a dependency on one’s family, and sado-masochistic educational and therapeutic machines, running from special beds to behavioural techniques so-called for the ‘reinforcing of the right reactions’ or the tyrannical interpretations of the psychoanalytic apparatus. But the abstract machine of ‘wetting the bed’ nonetheless retains the singularity of its mute dances, which will always remain more or less irreducible to the discursive-repressive analyses of therapists of every stripe! The possible good will of the child nevertheless always risks being found to be at fault. Even if he plays the game of repression, even if he invests it explicitly, the dimension of singularity of his system of abstract machines will allow him to escape from it partially.
In any case, repression doesn’t seek to completely submerge the child as an organic totality, but to graft itself onto the elements that are constitutive of his modes of semiotisation. Thus there isn’t purely and simply an application of the repressive ensemble to the ensemble of desiring machines, but processes of mediatisation by way of the abstract machines traversing the socius and the individual. When, as a secondary symptom, the bed-wetter manifests an inability to do division at school, for example, that doesn’t signify the existence of a deficit of logical competence – on the contrary, one notices that he is very frequently capable of dealing with very difficult abstract problems – but only that he has organised for himself a repressive jouissance in the framework of the rhizome ‘school-teacher-parent-notational system-repressive faciality traits-prohibitions bearing on masturbation’, etc. His refusal of a certain kind of logical discursivity manifests his desire to globalise the assemblage in question. He thus furnishes a sort of extra-corporeal erogenous zone for himself, territorialised on a particular stopping point: the division question thus becomes a machine point, the index of a potential line of flight. Under other circumstances, the same child could just as easily become mute or start to ejaculate on reading the statement of a problem … In fact, the machines of family and educational power can only find their efficacy to the extent that they manage to cling on to such bio-psycho-social zones, which do not inevitably take the form of labelled symptoms. The adaptive and recuperative therapy that consists in enlarging, in normalising, the semiotic conjunctions called into question by, for example, a child who territorialises a zone of stuttering, seeks to convert his libido to a relatively more deterritorialised zone: an anxiety linked to competitiveness at school, without for all that completely paralysing him.
Thus, by way of abstract machines, the libido doesn’t cease circulating between instances of social repression and those of individual semiotisation. But there is nothing automatic about this circulation, nothing necessary. It must always bring together two conditions to be possible: 1) ‘individual’ desire must crystallise its indices, its machinic points, on an abstract machine; 2) it must be possible for certain elements of the repressive socius to be connectable to this abstract machine. With an abstract machine, the possibility of a different assemblage of the world unfolds emptily. On leaving childhood, for example, an adolescent will see, as if in a flash, all the richness and threat that is harboured by the new system of enunciation in which he is engaged and which he both takes part in and is taken up by. Equally, the abstract machine constitutes a fundamentally metastable instance between the intensities of desire and the dominant semiological stratifications. However, unlike the machinic indices, which only anticipate their crystallisation, abstract machines subsist in the virtual state, even when they do not consolidate their pathways to manifestation. Whilst indices can be scattered at any moment and allow the forceful returning of old stratifications to establish itself, abstract machines will, under all circumstances and everywhere, continue to threaten them with a possible revolution. This is how a capitalist abstract machine has haunted every social system from the moment that a despotic State power succeeded in taking off from the archaic territorialities of the Neolithic (Urstaat). It is by a sort of immediate semiotic contamination that the most deterritorialised abstract machinisms are transmitted from one system to another. But whilst there is a potential transmission of abstract machines from the adult world ‘to’ the world of childhood, from the civilized world ‘to’ the barbarian, on the side of childhood ‘without’ adults, savages ‘without’ the civilized, there are only indices – of writing and the capitalist economy, for example. At this level, nothing has been played out definitively; everything depends on the constitution of collective assemblages of enunciation; a new assemblage can close up around a closed system of semiologisation – a dualist signifier-signified substance, or it can set off diagrammatic chain reactions, machinic flights/leaks of desire that will cross the ‘wall of significations’ and bring about direct connections between the points of deterritorialisation of sign machines and those of material and social ensembles. One might say that the abstract machine ‘materialises’ a triple possibility:
•either its own dissolution and a return to the ‘anarchy’ of machinic indices;
•or a relatively deterritorialised stratification in the form of abstraction by the putting into play of a significative semiology;
•or an active destratification, by the diagrammatising effect and putting into circulation of a-signifying particle-signs.
An abstract machine thus does not belong to one amongst a number of stages; it can participate in several stages at once, in one modality or another: at the level of indices, where it represents the potential for a machinic integration at a ‘higher’ degree, which will or won’t be recuperated by a stratum; and at the level of the strata, where it represents the potential for a destratifying diagrammatisation. Pure quanta of potential deterritorialisation, the abstract machines are everywhere and nowhere, before and after the crystallisation of the oppositions of machine and structure, representation and referent, object and subject. Thus the abstract machines make the threat of a reifying totalisation weigh on multiplicities as much as they do the possibility of a deterritorialising multiplication of stratifications that they open up. Independently of the appearance of an autonomous semiotic machine that distributes signs, things and representations over the separated planes of content and expression, their existence prohibits us from reducing them to a logico-mathematical system or to a priori forms. Their existence after the stratification of signifying semiologies, on the other hand, prohibits us from considering them as simple structural invariants of stratification or transcendental abstractions. Although the strata are nothing, for them, but the provisional residues of processes of deterritorialisation, being nothing in themselves from the substantial point of view, in order to become manifest they are constrained permanently to stratify and destratify themselves. But for all that they are not restricted to a disempowering face-to-face of the form-matter kind. There is thus a fundamental dissymmetry between the closed formalism of the strata that are ‘established’ in existence, and the active, open formalisation that is piloted by the abstract machines at the level of machinic indices and diagrammatic effects that mark the at once both creative and irreversible character of processes of deterritorialisation. Under these conditions, a homeostatic equilibria of strata will never be guaranteed: they are threatened from the ‘outside’ by the work of interstratic deterritorialisation of the abstract machines, which can result in the reshuffling, assemblages and creation of new strata, and from the ‘inside’ by the metabolism of lines of flight criss-crossing them everywhere.
Before the manifestation of the possible in semiotic structures or social material stratifications, the possible doesn’t exist as a purely logical matter; it doesn’t start out from nothing, either. It is organised in the form of quanta of freedom, in a sort of system of valences, the differentiation and complexity of which gives nothing away to the chains of organic chemistry or genetic codes.2 It puts into play matters of expression that are differentiated as a function of their degree of deterritorialisation. The plane of consistency, which deploys the infinite set of machinic potentialities, constitutes a sort of sensory plate for the locating, selection and articulation of points of active deterritorialisation within the strata. There is no possibility in general, but only by starting from a process of deterritorialisation which must not be confused with a global and undifferentiated nihilation. Thus there exists a sort of matter of deterritorialisation, a matter of the possible, which constitutes the essence of politics, but a trans-human, trans-sexual, trans-cosmic politics. The process of deterritorialisation always leaves remains, either in the form of – spatio-temporal, energised, substantialised – stratification, or in the form of the residual possibilities of the line of flight and of the generation of new connections. Deterritorialisation never stops when under way, that is how it differs from a nothingness that one represents as closed in on itself, maintaining disempowering mirror relations with the stratified real. The system of abstract machines thus constitutes an active limit, a productive limit beyond the most deterritorialised limits, and on this side of a nothingness as the terminal point of all process. Abstract machines are thus not a scientific affair, nor an affair of culture, ideology, or education, but an affair of the politics of desire before subjects and objects have been specified. It is not a question here of a freedom linked intrinsically to the human condition, of a freedom of the ‘for itself’ in a radical opposition with an ‘in itself’ that is stratified and thereby with no connection with anything other than its own impotence. In passing from one assemblage to another, one receives or one loses a certain quantum of deterritorialising connection; deterritorialisation cannot be assimilated to a necessary causality, it can be vectored either along the lines of a stratification or along the lines of an open ‘possibilisation’.
Let’s come back once again to the supposed ‘latency’ period which, according to the Freudians, marks the ‘development’ of the child. It would be manifested between six and eight years by a ‘childhood amnesia’ that results from a repression bearing on the whole Oedipal and pre-Oedipal past of the child. But, Freud tells us, all memory is not, for all that, abolished: ‘vague, incomprehensible memories’3 remain all the same. Incomprehensible for who? For the civilized, normal, white adult! In fact it is not a matter of memories here but of the entirety of the modes of semiotisation of the child, of its sensations, its feelings, its sexual impulses, which receive a formidable snuffing out. Why would one find the existence of a mechanism for the intrinsic repression of the development of the drives of the child – which will subsequently be linked to the universal antagonism between Eros and Thanatos – if not so as to mask the entrance on stage of repressive social assemblages? Why does the semiotic politics of the child invert itself, why does it take the side of repression? Why do the factors of deterritorialisation, which unbalance the earlier territorialities instead of opening the process up to a greater semiotic creativity, vectorise the child to the abstractions of the dominant system?
As soon as one attempts to give up the schematic responses of psychogenetic determinism, the questions change completely and are enriched. In the context of the repressive powers of the family and school of a given society, as a function of what particularity does a child resist or succumb to the ‘temptation’ of an investment in repression? In the case of the ‘latency period’, on the very concrete terrain of existing systems, what sort of educational abstract machine connects to the abstract machines of the child? In what ways do the semiotics put into action by nurseries continue the [extinguishing] actions of the ‘educational’ interventions of the parents? (We know now that it is from the nursery on that the division between work and ‘play’ time is put in place.) In what way does learning a writing that is detached from any living use, at school, sterilise the ulterior possibilities of a creative diagrammatism? How do the semiotics of educational space and time (division between school days and holidays, division between the space of the class and the space of the teacher, the space of the playground, the street, etc.), how do the semiotics of discipline (sitting in rows and ranking, grades, emulation, punishments, etc.) succeed in crushing the ‘pre-school’ semiotics of the child, sometimes definitively? And how do they outline the semiotic conditionings of the factory, the office and the barracks? As we have tried to show previously, the compulsory schooling machine doesn’t have as its primary goal the transmission of information, of knowledge, a ‘culture’, but the top to bottom transformation of the semiotic coordinates of the child. In these conditions one can consider that the real function of the ‘latency period’ is a modern equivalent of the initiation camps in primitive societies, which fabricate complete ‘persons’, that is, adult males who meet the essential norms of the group.4 But here, instead of lasting fifteen days, the initiation camp lasts fifteen years, and its objective is to enslave individuals to capitalist systems of production, right down to the least useful fibre in their bodies. Childhood amnesia, correlated to the latency period, thus marks the extinction of semiotics that are not subjected to the signifying semiologies of the dominant powers. And if the neurotics, like ‘pre-Oedipal’ children, escape its net, this is precisely because, for one reason or another, the systems of encirclement by these powers have failed to get a hold over them. Consequently, childhood intensities continue to work away and to upset them, to turn them against ‘normal’ values and significations. The role of memory – either the natural memory of the adult who recalls his childhood nostalgically or the artificial memory of psychoanalytic anamnesis – consists in doubling up the first erasing of these intensities and in recognising childhood according to a set of norms.
To have a grasp of reality, the assemblages of discourse are required, in whatever way it might be, to free themselves from the constraints of language, considered as a system closed up on itself. And it is the classic break between langue and parole that a pragmatics will, at a minimum, have to call into question. But although the linguistics of enunciation has already oriented itself in this direction, no micropolitical analysis of these assemblages, at the level of their collective or individual unconscious effects will become possible if it doesn’t call into question more fundamentally the concepts that delimit the different disciplines arising from what are usually called the human sciences. To succeed in constituting itself, a pragmatics of the unconscious will thus not only have to free itself from the dominant ideologies and universals of psychology, sociology, and psychoanalysis, but equally from a certain conception of the unity and autonomy of language, considered as plane of expression as well as a social entity – that is to say, in short, from the key ‘conquests’ of linguistics since Saussure. For our part we consider that there is no language in itself. What is specific about the phenomenon of language is precisely that it never refers to itself, that it always remains open to all the other modes of semiotisation. Whenever it closes up round a language, a dialect, a patois, a specialised language, a delusion, this always results from a certain kind of political or micropolitical operation. There is nothing less logical, less mathematical, than a language. Its ‘structure’ results from the petrification of a sort of rag-bag, whose elements come from borrowings, amalgamations, agglutinations, misunderstandings – a sort of underhand humour that presides over its generalisations. It is the same with linguistic laws as it is with anthropological laws, those bearing on incest, for example: seen from the distance of the grammarian or the ethnologist, they seem to have a certain coherence, but as soon as one gets a little closer, everything gets tangled up and one notices that it is a matter of systems of arrangements that can be pulled in numerous directions or turned around in all sorts of ways.
The relativity of the relations between concrete semiotic performances and a structural linguistic competence, or between languages themselves, is thus imposed not just on the synchronic but also the diachronic plane. The unity of a language is inseparable from the constitution of a power formation. One never finds clear frontiers on the map of dialects, only borderlands or zones of transition. There is no mother tongue, but phenomena of semiotic power takeover by a group, an ethnic group or a nation. Language stabilises around a parish, a bishopric, a capital. It evolves by flows along river valleys or train tracks; it spreads like patches of oil.5
But the fluidity of the competence-performance relation makes it something on this side of dialect. One may consider that every individual passes constantly from language to language. He successively speaks as ‘father to son’, as a teacher or as a boss; to his lover, he speaks an infantilised language; while sleeping he is plunged into an oneiric discourse, then abruptly returns to a professional language when the phone rings. Each time a whole set of semantic, syntactic, phonological and prosodic dimensions are put into play – not to mention the poetic, stylistic, rhetorical and micropolitical dimensions of discourse. Studying linguistic change, Françoise Robert notes that linguistic mutations are manifested by ‘gradual modifications, not of the phenomena themselves … but of their frequency, their establishment in language’.6 And it is true that one does not observe the sudden ruptures that are implied by the clear-cut distinction between synchrony and diachrony (a point on which Chomsky did not distinguish himself from Saussure, who only intended to take into account innovations at the moment when ‘the collective welcome them’7). Thus the autonomy of a pragmatic micropolitics is unfounded if the break between the exercise of individual speech and the coding of language in the socius is maintained. For Chomsky, as Françoise Robert also remarks, the reference to an ideal locutor-auditor, who belongs to a completely homogenous linguistic community, in fact results in investing the separation between competence and performance with a normative function. And, in the last resort, this norm is reduced to that of the linguist himself.8 In our opinion then, the apparent unity of a language doesn’t depend on the constitution of a structural competence. According to Weinrich’s formula, language is an ‘essentially heterogeneous reality’.9 In the final analysis, its homogeneity can only result from phenomena of a political order, independent of the structural decompositions that can otherwise be carried out on it. And what characterises a political event is its being the bearer of a historical singularity that is undecomposable or that an analysis will necessarily decentre in other dimensions, other registers. Things happen in the same way as in the chemical analysis of a biological phenomenon, or in the economic analysis of a social phenomenon: there is no more a chemical structure of a biological fact, or a chemical competence with regard to a biological performance, than there is a capitalist or socialist structural competence with regard to economic or monetary performances. There are no biological or economic universals. And yet at each one of these levels, abstract machines are differentiated, manifested, and stratified at different crossroads-points of the machinic phylum, without depending on any transcendental formalism, any heredity, any linguistic essence, any economic fate. Our hypothesis, of a mutational phylum of abstract machines, should allow two kinds of obstacle in the domain of pragmatics to be avoided:
•a pure and simple pinning of linguistic machines onto social structures, as in the linguistic dogmatism of Marr, or as in certain contemporary psycho-linguistic currents;
•a structuralist or generative formalisation, which cuts the production of utterances off from the collective assemblages of enunciation.
The differential relations between what we will call the tracings of performance and the maps of competence do not play just at the level of diverse kinds of segmented encoding. We consider that the relative structure of ‘competence’ in one domain in relation to another in fact depends on whether or not it puts into play a segmentarity that is finer, more machinic, more molecular, more deterritorialised than the more molar segmentarity of the second, which thus finds itself taking a ‘performative’ position. A hierarchical relationship of double segmentarity is thus established, which fixes the possibilities for semiotic innovation within a strict margin. Only the appearance of a deterritorialising line of flight (the diagrammatic use of signs with a linguistic origin in aesthetic or scientific domains, etc.) can overthrow such an equilibrium. We have seen that at the level of past-ified, spatialised or semiologically substance-ified strata, equilibria, relations of force, can only manifest themselves on the basis of a relative deterritorialisation, the placing into correspondence of at least two systems of segmentarity (for example, the molecular segmentarity of the figures of expression of the second articulation), whilst at the level of machinic mutations, the strata are undone or reorganised by diagrammatic processes that put into play a deterritorialisation that is quantified by systems of abstract machines. But the lines of diagrammatic deterritorialisation do not definitively transcend segmentary stratifications. Mad vectors of possibilities, which cannot be realised in the existing context, as well as veritable machinic mutations, can result from their interactions with stratified systems.10
As we have seen, not only are abstract machines not outside history, ‘before’ my spatial, temporal and substantial coordinates – deictic performances, one might say – but they do not result in the unification of diverse modes of semiotisation.11 Abstract and singular machines, they make history by undoing dominant realities and signification; they constitute the umbilicus, the point of emergence and creationism of the machinic phylum. Thus there cannot be an abstract set of the abstract machines. No logical category can subsume machinic consistency (hence the difference that we have already signalled between logical and machinic consistency). Being undecomposable on an intensional plane, one cannot insert abstract machines in an extensional class.12 Given that there is no abstract machine hanging over history, no ‘subject’ of history, and that machinic multiplicities traverse the different strata both on a diachronic and on a synchronic plane, one cannot say that the general movement of their line of deterritorialisation manifests a universal and homogeneous tendency, because it is interrupted at every level by strata of reterritorialisation, onto which microscopic buds of deterritorialisation are grafted once again. In these conditions, a pragmatic approach to the unconscious would have to escape from two kinds of pitfall:
1An analysis that is centred exclusively on a verbal material and tends to a ‘significantisation’ of behaviours and affects by means of a systematic gridding of semantic contents and enunciative strategies (politics of transference), based on a meta-syntactic interpretative grid.
2A return to the analysis of personological strategies, as is the case with Anglo-Saxon family therapies, and a return to lived experience, to corporeal abreaction, etc.
Before its engagement in the detail of utterance production and modes of semiotisation, the abstract machine has to determine the micropolitical lines creating the ensemble of assemblages of enunciation and power formations at the most abstract level. In other words, in each case and in each situation, it has to construct a map of the unconscious – with its strata, its lines of deterritorialisation, its black holes – open to opportunities for experimentation (and that in opposition with the infinite tracing of Oedipal triangulations, which merely make all previous impasses, all modes of signifying subjection, resonate together). In effect, we consider that the pragmatic articulation of encoding strata closed in on themselves always leave open the possibility of a passage from one stratum to another, by way of the abstract machines traversing different modes of territorialisation, The different kinds of consistency – biological, ethological, semiological, sociological, etc. – therefore do not depend on structural or generative super-stratum; they are worked from the ‘inside’ by a network of machinic connections. Machinic consistency is not totalising but deterritorialising. It guarantees the always possible conjunction of the most different of systems of stratification, and it is in this respect that it is in some way the basic element out of which a pragmatic can constitute itself.
After having relied on psychoanalysis, linguistics, semiology, will the normative gridding of the human sciences shift to a new field of combat, that of pragmatics? The latter is defined by Herbert E. Brekle as the ‘condition for the production of speech acts’. And straightaway it is associated with communication: pragmatics is the communicative dimension of language. Communication being inseparable here from the bipolar speaker-listener axis, pragmatics thus finds its fate is linked to the existence of the stratum of individuated subjectivity and the individual/socius opposition. A different condition for the possible extraction of the autonomy of pragmatics will thus consist in specifying, positively this time, its specific modes of semiotisation, its particular way of freeing itself from the semiological modes of ‘structuralisation’ of the languages of power. Here the collective character of machinic enunciation is opposed to the individuation of signifying enunciation, and the politics of sense to that of signification. Such a pragmatics thus presents two faces: one that links it to the stratum of subjectivation and alienates it in communication, and one that links it to collective assemblages capable of producing utterances that mesh directly with machinic processes. The always possible bogging down of modes of semiotisation would thus form an intrinsic part of pragmatic components. And the pragmatics of enunciative linguistics, in which language is closed in on itself in a function of disempowerment, would thus be just a particular case of a more general (diagrammatic) pragmatics, open to the ensemble of non-linguistic modes of encoding and semiotisation. In sum, the autonomy of pragmatics will be founded on the essential impossibility of guaranteeing its own autonomy. And rather than seeking to give itself a pseudo-scientific status, it will define itself as an activity of micropolitical assemblage.
What might the characteristics of a generative and transformational pragmatics be? In the first place, its modes of engendering would not be trees, but rhizomes (or trellises). A priori there would be no reason for a pragmatic chain to begin at point S so as then to be derived by successive dichotomies; any point whatever of the rhizome can be connected to any other point. Besides, no trait will necessarily refer to a linguistic trait. A linguistic chain can be connected here to the chain of a non-linguistic semiology, or to an assemblage that is social, biological, etc. Segmentary stratifications will be correlated here with deterritorialising lines of flight. A rhizome cannot, therefore, be formalised on the basis of a logical or mathematical meta-language. It will not be indebted to any structuralist or generative model. As a process of machinic diagrammatisation, it cannot be reduced to a system of representation, and it implies the putting into play of a collective assemblage of enunciation. The preparation of the pragmatic rhizome arising from such and such an assemblage will not have as its goal the description of a state of fact, the re-balancing of inter-subjective relations, or the exploration of the mysteries of an unconscious hidden away in the shadowy corners of memory. On the contrary, it will be turned entirely towards an experimentation flush with the real. It will not decipher an always already constituted unconscious, closed in on itself, it will construct the unconscious. It will contribute to the connection of fields, the unblocking of stratified, empty or cancerous bodies without organs, and to their maximal opening onto the machinic plane of consistency. It will be led to put into play diverse semiotics and modes of coding, of a biological, sensory, perceptual order, on the order of a thinking with images, categorical thought, semiotics of gesture and word, political and social fields, formalised writings, arts, music, refrains … Unlike psychoanalysis, which always seeks to reduce each utterance and each libidinal production to an overcoding structure, a schizo-analytic pragmatics will have as its objective, the pinpointing of their repetitive elements in what we will call systems of tracings, which can be articulated with a map of the unconscious.
The map is opposed to structure here; the map is open, it can be connected in each of its dimensions, it can be torn up, it can be adapted to every kind of setup. A pragmatic map can be put to work by an isolated individual or by a group, one can draw it on a wall, one can conceive it as a work of art, one can conduct it like a political action or as a meditation. What matters is to determine how, given a kind of performance, a particular assemblage of enunciation, a redundant tracing, does or doesn’t modify the unconscious map of a local pragmatic competence.13 These maps of competence do not depend in an absolute fashion on a broader competence. Just as there is no universal competence, there is no universal cartography: such and such a map, which serves as a marker for one collective performance (that of an anti-psychiatric community or a groupuscule, for example), could be valid as a performance for such and such other social group (psychiatry in France as a whole, or the ensemble of political movements, for example).
One rediscovers the subject-group/subjugated group alternative here, which must never be taken as an absolute opposition. The relations of alienation between fields of competence always imply a certain margin, which it falls to pragmatics to localise and utilise. In other words, in no matter what situation, a diagrammatic politics is always possible. Pragmatics refuses any idea of fatalism, whatever name one gives it: divine, historical, economic, structural, hereditary or syntagmatic. By taking into account the entirety of his semiotic productions, studying the unconscious in the case of Little Hans would have consisted in establishing which kind of tree or rhizome his libido would have been led to invest. How, at such and such a moment, the branch of the neighbours was cut off, following what manoeuvrings the Oedipal tree contracted, what role Professor Freud’s branch and its deterritorialising activity played, why the libido was constrained to take refuge in the semiotisation of a becoming-horse, etc. In this way, phobia would no longer be considered a psychopathological result but as the libidinal pragmatics of a child who was not able to find any other micropolitical solution to escape from familial and psychoanalytic transformations. Pragmatics would thus imply, in the first place, an active refusal of every conception of the unconscious as a genetic stage, as structural destiny. For a group, it would necessitate a permanent searching for investments of desire able to thwart the reifications of bureaucracy, leadership, etc. ‘Working’ the map of the group would consist in carrying out a reshuffling and transformation of the body without organs of the group – that is to say, the locus of investment of desire ‘anterior’ to any specification, any organisation centred on an object – necessitated by a micropolitics compatible with these investments. One cannot just give such a pragmatics its part to do: it can only challenge the hegemonic vocation of linguistics, psychoanalysis, social psychology, the entirety of the human, social, legal, economic sciences.
What is the nature of the relations between the two kinds of components – generative and transformational – of pragmatics, the existence of which we have simply evoked? As we have said, pragmatics has, hitherto, been considered as a domain that can only be adjacent to linguistics. This was true for Austin and Searle, and it is still true for Ducrot, despite the fact that he calls into question communication as the essential characteristic of language, and despite the richness of his analysis of presupposition, which opens linguistics up to a veritable new micropolitical field.14 We have seen that whilst the pragmatics that we are envisaging is essentially aimed at the ensemble of non-linguistic semiotics fields, it nevertheless entertains a particular relationship with linguistic semiologies, this domain being defined as that of generative pragmatics. Pragmatics would thus be divided into two components – and not two regions, as these components will constantly recompose themselves: a generative pragmatics corresponding to the modes of ‘linguisticisation’ of semiotics and a non-linguistic, non-signifying, transformational pragmatics.
The question was already posed at the level of the independence of ‘analogical’ semiotics. Should one accept their fundamental dependence on linguistic semiology, like the majority of semioticians, accept their fundamental dependence on linguistic semiology? Or should they be considered as autonomous modes of semiotisation, able, under certain conditions, to pass into the control of a signifying transformation? Should one not, on the contrary, consider that what could be called the ‘axiom of structure’ (which has, since Saussure, consisted in separating language from acts of language and expression), is just a particular case, resulting from a contingent semiotic conjunction? Does the normal, terminal, regime of symbolic semiotics depend on the linguistic machine of expression? On the contrary, we previously indicated that we consider that there is nothing ineluctable, nothing universal, about signifying transformations, and that they are linked to a certain kind of regime of individuation, enunciation and inter-subjective communication. These signifying transformations derive their power from their reliance on a certain kind of a-signifying machine of expression (double articulation machine, which can be described in terms of syntagmatic trees or more abstract formalisations), which organises and stabilises the entirety of semiotic compositions as a plane of content and plane of expression. The strength of the machine for signifying disempowerment resides in its capacity to crush, to neutralise all contents. The function of the signifying transformation is to generate, to structuralise semiotic productions of all kinds. By means of which systems of institutional constraints is what Herbert E. Brekle designates as ‘communicative competence’ determined? These are the questions to which a generate pragmatics must respond.
Let us return now to the relations between the different semiotic components, which we presented in the table on p. 129, and let us examine in particular the fact that non-interpretative (symbolic, diagrammatic) transformational components are able to break the hegemony of interpretive (analogical and signifying) generative components.
The anthropological study of phenomena of acculturation shows us that the putting into place of a signifying transformation never goes without saying. Primitive societies can even actively oppose it. It is in this way that certain mythographic systems have long been able to resist the exclusive domination of a semiology in which the expression-content relation is structured according to syntagmatic and paradigmatic axes. In the domain of myth, kinship relations, political anthropology, etc., symbolic semiologies cannot be automatically reduced to the dichotomous relations of a signifying economy. There is a big danger here of a hasty ‘structuralisation’ of ethnographic data, consisting in interpreting kinship relations in terms of a generalised exchangism, for example. The installing of invariant15 significations doesn’t go without saying. In symbolic semiotics, the planes of content are linked to one another, slide around in relation to each other, without being organised on the structured plane of the signified. It is only with the accomplishment of the hegemony of capitalism, in the nineteenth century, that the ‘absolute stability of signifieds, under the proliferation of relations of designation […] so as to be able to found the comparison of forms’16 imposed itself definitively. A certain kind of dictatorship of the signifier thus seems linked to a certain historical context and, as a consequence, cannot be considered either as immutable or as universal. This signifying power can be neutralised, even overthrown by transformations. This is what happens in contemporary African societies, for example, where a fixation on tribal modes of solidarity, or sudden returns to animist practices, serve as a counter-weight to the expansion of semiologies of the Western kind. Equally, at an individual level, with the ‘taking power’ by an oneiric semiology, of perceptual semiotics, linguistic semiologies, etc., under the effect of sleep, drugs, amorous exaltation, etc.
Another general type of pragmatic transformation can bring about a semiotic freeing up of the disempowering signifier-signified couple: diagrammatic transformation. Two kinds of semiotic system, the distinction between which had nevertheless been outlined by C.S. Peirce, have generally been confused under the category of icon:17
aImages, in which the sign functions through analogy, by evoking the object denoted (in the case of a semiotic functioning on the basis of spatial elements, these generally put into play at least two dimensions).
bDiagrams, which function in such a way that the elements of the form of content are transferred onto the plane of the form of expression by means of what we will call a sign-particle system that simulates the process denoted, and that generally according to a linear mode of coding.18 Peirce defined diagrams as being ‘icons of relation’. The diagrammatic sign doesn’t imitate objects, but articulates properties, functions.19 Content is deterritorialised by its mode of formalisation. Symbolic semantic and semiological signifying redundancies are emptied of their substance (a polyphonic and harmonic formalisation in music, mathematics in physics, axiomatics in mathematics).20
Thus diagrammatism does not objectify a world, the representation of which it would stabilise, but assembles a new type of reality. It ruptures with the organisation of dominant significations. Diagrammatic semiotic processes in fact constitute components that are indispensable to the machinic assemblages of human societies. For example, it is impossible to conceive the assemblages of a scientific experiment without the putting to work of such a process (in the form of plans, topological, mathematical, axiomatic, informatics descriptions, etc.). That such sign machines can function directly within material and social machines, with the mediation of processes of significant subjectivation, is something that has become daily more evident; but the decisive step that it seems to us to be necessary to take, in order to found a pragmatic politics, is to see that the common essence of semiotic and material machines results from the same kind of abstract machine. Positivist realism has led to the crushing of the creative dimension of diagrammatism, reducing it to the general category of analogy; first, diagrammatism is recuperated as a sub-product of the icon, then, second, the icon is recuperated under the category of analogy, itself considered as a sub-product of signification. But – and we can’t insist on this enough – the relation of signification (signifier-signified), is only a particular case of the mechanism of semiotic machines, which function by prolonging one another. In this regard, Bettin and Casetti have pointed out how reductive the commentary on Peirce’s writings has been, because, unlike their habitual presentation, his categories are never closed in on themselves, and there is no irreversible break between the systems of signs and their object. An iconic sign can always be the sign of another system, and the systems of objects themselves already function as a sign machine in a society’s knowledge, inserting themselves into what he calls, the ‘progressive chain of interpretative definitions’. And the establishment of a stabilised system of significations in effect indeed seems to us always to be correlated with the placing under guard of symbolic semiologies in their diversity. As Lotman writes ‘the greater the distance between structures made equivalent to each other in the process of recoding, the greater the disparity in their nature, the richer will be the content of the very act of switching from one system to the other’.21
Analogy only constitutes the first level of this operation of levelling and translateabilisation of semiotic chains of all kinds (doubtless one ought to be led to consider ‘degrees of analogism’). Analogy and signifiance constitute two modes of the same politics of the reterritorialisation and subjectivation of contents. But whilst analogy organises them into relatively informal fields, articulated through relatively territorialised assemblages of enunciation, significance, with its doubly articulated chains, grids them in paradigmatic and syntagmatic coordinates that are much more strictly articulated with individuated assemblages of enunciation directly subjected to capitalist social systems. Analogic formalisation is less rigorous, less deterritorialised, than that of signifiance: it brings into view strata of expression that retain their own consistency, producing what we have called ‘fields of interpretance’. One symbol interprets another, which itself interprets a third, and so on, without the process hitting a terminating signified, the sense of which would be blocked in, for example, a dictionary, and without the chain being liable to respect a grammaticality that fixes rigorous rules of syntagmatic concatenation. The work of signifying generation on content brings an additional degree of deterritorialisation into play: it isn’t based on analogic motivations any longer, but on the ‘arbitrariness’ of a machine of a-signifying signs,22 which phonologises, graphematises, morphologises, lexicalises, syntacticises, rhetoricises them. Certainly, analogic transformations are not specific to one particular kind of assemblage of enunciation; they can equally be applied to diagrammatic semiotics. But in this case, the same signs are treated in terms of two generative and transformational semiotic politics: on the one hand, they function as symbols in an analogic mode, and on the other, as figures of expression in a diagrammatic mode. This mixed system corresponds precisely to the signifying mode of representation, which puts an a-signifying machine at the service of signifiance. Empty signs, without any semantic content, the phonic or graphic image of the word ‘table’, for example, are seen as a table.23 Thus, by territorialising artificial analogons, diagrammatisation closes up on a world of quasi-objects. But, unlike the world of symbolic representations, this world is ‘worked’ from the inside by syntax and logic, on which the formalisation of significations and dominant propositions rests. On the one hand, it invites us to insert ourselves into a reality that ‘goes without saying’, a reality of the everyday, and on the other, it draws us, as if in spite of ourselves, into the circle of its pragmatic implications, and its signifying chains alienates us in an immense social and technical machine, that of Charlie Chaplin in Modern Times. All libido is thus captured, functionalised, subjectivised as a function of the demands of the economy of capitalist flows.
The generative components of analogy and signifiance are thus not to be placed on the same plane as the transformational components of symbolism and diagrammatism, and the distinction, now traditional, between ‘analog’ and ‘digital’ semiotics24 seems to us not to need to be maintained. We find ourselves in the presence of two general kinds of component:
•Symbolic and diagrammatic transformations, which constitute semiotic domains that are distinct from one another, and the difference of which even gets accentuated to the extent that the process of deterritorialisation that marks the evolution of the second develops.
•Analogic and signifying ‘generations’, which do not constitute distinct semiotic domains, but both participate in the same reterritorialisation and subjectivation function. The constraints that these components impose on the two previous components, when they are applied to them, have as their goal making them compatible with the values and coordinates of a particular vision of the world. They generate a world by making the possibility of the appearance of different worlds degenerate; thus we could call them degenerative components, in opposition to the pragmatic (symbolic and diagrammatic) transformations which, each in their own way, overthrows the dominant system of redundancies, reordering the vision of a world.
Like the semiotic components that they put into play, pragmatic assemblages of enunciation cannot be reduced to the composition of standard elements, universal subjective positions of the kind theorised by Lacan, for example (discourses of the master, the hysteric, knowledge, the analyst). And the classification that we have set out in the table on page 129 is entirely relative! Thus, in fact, territorialised assemblages of enunciation only correspond to a dominance of analogical transformations of interpretance, and can equally put into play symbolic, diagrammatic and signifying semiotics (example: the discourse of primitive societies, in so far as it ‘refuses’ the reductive effects of signifying generation, bases itself on symbolic techniques relatively non-interpretative, but this refusal implies by contrast the existence of a threatening signifying economy). The individuation of enunciation, whilst being specific to the dominance of signifying transformations, equally puts into play deterritorialised and overcoded symbolic transformations (of the figure-ground kind) and a diagrammatic redundancy organising symbolic formations according to a plane of content (consciential transformation). This second degree formalisation thus has as consequence the production of a new kind of effect, that one could call the effect of lack. Each content is doubled by a lack, it is ‘lacking’ the formalism that overcodes it. The unity of linguistic semiology thus becomes the formal signifying unity that Hjelmslev brought to light between the form of expression and the form of content. The fundamentally metastable character of this effect of lack produced through consciential transformation has as its corollary a sort of vertigo of unbearable, maddeningly anguishing, deterritorialisation. It must be filled without delay; and it entails the intervention of a certain number of reterritorialising pragmatic components: a transformation of faciality; transformation of the double, transformation of the couple, transformation of paranoid knowledge, etc. The mad vector of consciential transformation that this absolute deterritorialisation represents is thus conjured away through artificial reterritorialisations, which it is worth differentiating from the territorialised assemblages of enunciation evoked earlier. Now, there is no methodological necessity that forces us to consider that the semiotic components, on the basis of which we started our description, have real priority. A ‘rhizomatic’ analysis could just as well be carried out on the basis of less classically semiotic components, such as those that are knotted together around the black holes of anxiety, faciality, power formations, etc.
Table summarising the formation of semiotic fields on the basis of transformational and generative components
| Transformations | Generations | Semiological fields |
| A. Symbolic (e.g. dreams) | ↗AC Interpretative semiology (e.g. magic) |
|
↘– analogical (interpretance) |
→AD Signifying semiology (e.g. psychoanalysis) |
|
↗– signifying (signifiance) |
→BC Interpretative logography (e.g. geomancy, tarot) |
|
| B. Diagrammatic (e.g. systems of graphemes) | ↘BD Doubly articulated languages |
It is the same with the machinic assemblages of enunciation that are characteristic of the domain of diagrammatic pragmatic transformations. They remain haunted by subjects of enunciation. But the representation of a locutor-auditor as fictive pole of the production of utterances becoming increasingly abstract with them, the fact that ‘it continues to speak’ through the mouths of individuals takes on an increasingly relative scope. The utterance is produced and understood through a complex assemblage of individuals, organs, material and social machines, mathematical and scientific semiotic machines, etc., which constitute the veritable nucleus of enunciation. That being the case, this kind of assemblage cannot be separated in practice from the artificial reterritorialisations of enunciation that are correlative to it and which are always manifested within mixed semantics. It is in reaction to the vertiginous deterritorialisation of the subject that is implied, either through consciential transformation, or by a desubjectifying diagrammatic transformation, that a system of collective ‘reassurance’ artificially reproduces a territorialisation of enunciation. Thus, after the collapse of systems of territorialised, familial communities, the illusion of a return to the territorialised assemblages of primitive societies could even be maintained (the illusion of a ‘return to nature’, of a return to originary significations). Thus an artificial conjugal nuclear family will be recreated or, faced with the internationalisation of production and of the market, one will witness a massive return to questions of nationality, of regional particularisms, racisms, etc.
Without losing from view the arbitrary character of the systematic classifications that we are proposing, let us now examine certain limit assemblages, such as they can be determined on the basis of the distribution of their components – this time of a different, ternary order. We will insist once again on the fact that a monographic approach – a ‘rhizomatic’ analysis – of real situations, would therefore not start from the simple to go to the complex, but, on the contrary, would begin from the complex so as to envisage the ‘elementary’ components only to the extent that such an undertaking would allow it to explore more precisely certain singular traits of these components, leading to an even greater complexification of the assemblages of enunciation, and permitting a richer, more open, creative experimentation to be envisaged. The triadic system that we are proposing here can therefore not be assimilated to a method like that of C.S. Peirce, for example. The association of five, seven or n components might, in principle, have been preferable for him. Nevertheless, it should allow us to examine the limit cases, which anthropologists, historians or economists would doubtless make into typical cases, structural archetypes.
| Assemblages of enunciation | Machinic instances | Semiotic components | |
| Composition a | Territorialised | Index | Symbolic |
| Composition b | Individuated | Abstract machine | Signifying (abstraction) |
| Composition c | Collective | Machinic assemblage | A-signifying |
Numerous symbolic semiotics – those of childhood, of the mad, of primitive societies – are inseparable from the existence of stratified territorialities. Thus they do not depend, in the first place, on a substance of expression that would traverse and unify its different modes of semiotisation. They constitute a system of articulation of modes of encoding and formalisation in the raising up of a universal substance of expression. For example, in the territorialised assemblage of certain primitive societies, one will find an activity of mythographic formation developing on the basis of traits of matters of expression that do not enter into correspondence, that are not translateabilisable with those of gestural, perceptual, economic and other semiotics. That does not signify that these diverse modes of semiotisation are without relation to each other. But what brings about this relation is precisely the kind of territorialisation of the group, its internal topology, its translations into itself and outside its territory. Here, the territorialised assemblage of the group occupies the place that will become that of signifying substance in the system of despotic individualisation of enunciation.
Primitive societies refuse, by warding off, the bringing to light of a signifying substance; their politics is that of a group enactment of semiotic conjunctions. Already it is a matter here of a sort of pragmatic rhizome, but a rhizome that seeks to contain, to dominate deterritorialising flights. The systems of indices precisely mark on this rhizome the inscription of such a threat, of such a refusal to fall into signifying abstraction or into deterritorialised machinic assemblages. One index would, for example, be the fact that the death of a cow first calls for a recourse to the practices of geomancy, then, to the extent that the right results have been obtained from this procedure, the recourse to a ritual sacrifice, then to a trial for witchcraft, a marabout etc., without a synthesis being effectuated at any moment between these different undertakings, without a paradigm stabilising their general signification being extracted.
The group assembles the semiotics, it doesn’t interpret, it experiments. This real passage operates by respecting the particular traits of each matter of expression. Besides, and here is an essential difference with the rhizomes that depend on a deterritorialised machinic phylum, these territorialised assemblages do not hierarchise planes. Machinic deterritorialisations exist (for example: an embryonic writing) but they will be treated on the same plane as the territorialised assemblages. It is as if these societies entertained an active misunderstanding of the powers of deterritorialisation contained in certain indices. This kind of assemblage thus doesn’t exclude either the signifier, or diagrammatism, it simply refuses the power takeover by an overcoding instance or a deterritorialisation machine. A religious machine can be the bearer of universalising abstractions, but it will be prevented from escaping its territory, its totem, for example. It doesn’t aspire to a general translateability of kind of capitalist religions. It equally avoids symbolism falling into the equivalent of signifying translateability that iconism constitutes for it. The differential coefficients of deterritorialisation are not extracted from their territory, their original matter.
These societies lead an active struggle against the erection of a signifying object on high, whether in the form of a capitalisation of power, at the level of the chiefs, or in the form of a concentration of systems of semiotic enslavement in technical machines or writing machines. In other words, they endeavour to ensure that all systems of deterritorialisation remain or return to the state of indices, qualitative indices that will be neither quantified nor systematised. It is only during the ‘passage’ to societies dominated by signifying semiologies or a-signifying semiotics that such a quantification, such an accumulation of effects of deterritorialisation can be put to work. Here deterritorialisations still remain directly plugged into the intensities of desire, the body, the group, the territory.
Composition b corresponds to a process of evolution of the old territories that are traversed by machinic systems that hollow them out everywhere. Indices link together, accumulate. In the societies of the Pueblos, as with the Hopi Indians (whose ‘theocratism’, according to Levi-Strauss, evokes, in an unrefined form, Aztec civilisations), one begins to interpret indices in relation to one another; it is the reign of ‘dwelling on the past’, of bad conscience, of guilt.25 Abstract machines capitalise the indices and sketch out the constitution of machinic assemblages. In such conditions, these societies become vulnerable to contamination by abstract capitalist machines. But it is with societies that autonomise a despotic State machine that this signifying power will truly acquire its autonomy. How will the escalating deterritorialisations and systems of defence against capitalist flows be effectuated, what will they cling on to? What ceased to be possible in a territory will become so again in a system of semiological substance. The characteristics of this substance are disempowerment and dualism. What is retained by this substance are no longer intensities as such, but their differential character. Precisely the ensemble of these differential relations constitutes signifying substance. This signifying disempowerment is correlated with conscientalisation, the emergence of myths of the double, the totalisation of intensive effects on the person, the dualism of phallic power and already, in an embryonic fashion, systems of enslavement by semiotics of faciality and conjugality. Once it has crystallised, this substance contaminates all the old matters of expression. It constitutes a sort of sky that looms over intensities, pinning them down like butterflies, reducing them to the state of neutralised indices.
It deploys a formal subjectivity that is substituted for deterritorialised assemblages. Unlike these latter, this subjectivity has no need of being enacted, as it haunts each intensive system as differential value; it functions as a capital of differences; it is the matrix of all the capitalisations of power, whether they concern the State, matrimonial or economic exchanges, and, in general, all the systems for the capitalisation of decoded flows that we have characterised as capitalist. The semiological substance of individuated (or individuating) assemblages of enunciation is dualist in that it deploys a surface of representation that is constantly divided into two sub-systems: a substance of expression and a substance of content. The ensemble of intensive effects is formalised, secretly kept in hand by the formalisation of expression. Inversely, the diagrammatic machines that are put into play by them are kept in hand by the organisation, the finalisation, of the significations of content. This process of the bi-univocalisation of all intensities has as its corollary a linearisation, a flattening, of the old systems of territorialised rhizomes.
All the material intensities that contribute to the formalisation of expression must be put into order. It is no longer appropriate to speak by singing and dancing. What counts now is solely the assemblage of differential characteristics of the system as a whole, in so far as it contributes to the functioning of new deterritorialised powers. Under these conditions, the prosodic components that arise from song, from mimicry, from gestures, posture, etc., from ‘primitive’ speech can do nothing other than degenerate. One passes from one element to another according to a syntactic order and no longer the apparent disorder of territorialised assemblages. One compares, one measures, the coefficients of deterritorialisation of each fulfilling of form. The strata will have to submit themselves and to be hierarchised in this passage, there will not be any more contour, just a linear passage, constituting the most economic means of effectuating such a comparison and hierarchisation. In the absence of such a neutralisation, the possibility of the irruption of a system of intensity would subsist. But signifying substance is hegemonic, it cannot take such a risk.
In fact, it remains in a metastable state, because in order to be able to semiotise the structuring and hierarchisation of power formations on which it rests, it must have recourse to a putter to work of diagrammatic machines, the effects of which also risk making themselves felt on the side of content by the triggering of new machinic assemblages. How, under these conditions, is one to keep such a sign machine in hand? At every moment, and for everything, it will be necessary only to retain from it what can be fixed in a system of abstraction and a formal syntax. For example, the appearance, in the history of music, of a polyphonic, then a harmonic, writing component, which threatened to make music explode in a kind of generalised baroque, was warded off for a long time by religious power, which endeavoured to retain only those traits of musical expression that were mathematisable.
Thus a sort of universal syntax of musical writing was established, inseparable from the power formations weighing on musicians (teaching, patronage, etc.). It was only when other, more deterritorialised components, come on the scene and call into question the musical compromise called, paradoxically, ‘baroque’, that the continuous process of fragmentation that the evolution of modern music represents, will be sketched out. But this semiotic deterritorialisation of music is inseparable from those that have worked over the representations of the world in the religious, philosophical and scientific domains. And there too one will discover systems of reterritorialisation to check the proliferation of abstract machines and translateabilise them into a general conception of the world. Abstraction functions here as a locus for rebounding, a stopping point for semiotic systems susceptible of being organised into a machinic rhizome. The abstract machine corresponds here to the Hjelmslevian intuition regarding form, according to which it is, in some way, the same abstract machine that is manifested in the substance of expression and the substance of content. One might say that it is the same dualising substance that secretes abstraction and contains intensities in reductive systems of dichotomous trees. But the transcendent formalism, which results from what we have called a paradigmatic perversion, is nonetheless under threat from a double danger: on the side of content, the explosion, the flourishing, of intensive multiplicities; on the side of expression, the implacable diagrammatism of sign machines.
The figure-ground, form-matter oppositions of territorialised assemblages, and the dualism of the signifying substance of individuated assemblages, cease to be pertinent here. In appearance, one is returning here to a polyvocal expression of the kind of territorialised assemblages. But one is not dealing here with well localised assemblages of persons, techniques, myths, etc., with the enactment of bodies, organs and territories on the basis of a system of signifying subjection, but with a machinic assemblage, a non-human machine, at the heart of which the overcodings of despotic abstraction no longer lay down the law in the same way. What now looms over this semiotic system is no longer a territorialised assemblage or a formal subjectivity, but the plane of consistency of the ensemble of possible machinic assemblages. The machinic assemblage of enunciation re-articulates machinic indices at an intensive level, and no longer solely at a differential level. In addition it vectorises systems of stratification by polarising territorialised systems towards deterritorialised systems. One has thus left the register of the autonomy of territorialised assemblages or of the comparative dualism of intensities of signifying substance in individuated assemblages. The machinic rhizome is vectored and vectorising. A general vectorisation of destratification processes is substituted for global hierarchies.
One is not for all that in the presence of an autonomised machinic substance: machinic components are not stratified: as they are enacted, they constitute a phylum that implies not just their actual state and the historical and logical links that have led there, but also their diagrammatic potentialities. The virtual, the theoretical and the experimental to come thus form a part of the machinic phylum.26 We will therefore not reintroduce a dualism between material and semiotic deterritorialisation at this level, because one is always in the presence of a multiplicity of matters of expression and semiotic systems corresponding to a diversity of particular modes of deterritorialisation. There is thus no place for grouping [together], for example, energetic, physic-chemical, biological, etc., intensities on the one hand, and aesthetic, revolutionary, scientific, etc., intensities on the other. The multiplicity of systems of intensity is conjugated, ‘rhizomatises’ over itself: the machinic assemblage brings about conjunctions between ‘scientifically formed’, ‘aesthetically formed’ matters, without giving them any privilege, in so far as they issue from an autonomised sign machine. No system has any priority over any other as of right; material components are not necessarily more territorialised than semiotic components. What is important here is not a particular differential index, nor a range of differential indices, it is the assemblage of quanta of deterritorialisation enacted. Certain intensive systems have quantum superpower in relation to others.
A mathematical sign machine can temporarily become superpowered in relation to the system of deterritorialisation in play – in physics, for example, in conjunction with theoretical and experimental components. Inversely an intensive effect27 can become superpowered in relation to an entire sector of theoretical physics. Indices and abstract machines continue to exist in machinic assemblages but instead of the indices turning round and round in a territorialised assemblage, enacted by human collectivities in a given territory, or the abstract machines remaining tightly fixed to a dualising substance, they now only function in so far as they are bearers of certain quanta of deterritorialisation. This point is primordial, because, we repeat, there is no hierarchy between indices, abstract machines and machinic assemblages. For example, the ‘feelings’, the private life, of a scientific researcher, the fact that he falls in love or goes mad, can introduce a deterritorialising charge of the greatest in the machinic assemblage that constitutes his research. An erotic index, a libidinal charge, will perhaps be able to unblock systems of abstract machines and systems of experimental assemblages, or even throw them out of gear completely. Inversely, an abstract machine might fertilise a system of indices: it is perhaps the fact that an abstract machine, of a theoretical or experimental order, has been introduced into his system that makes our researcher ‘decide’ to fall in love or go mad. Passions, all passions, not just those of artists and scientific experts, whatever they may be, whatever they put into play, should cease being separated from oeuvres so that they can be related to the recipes relative to the interpersonal strategies that obsess psychoanalysis. Machinic assemblages are bearers of indices as much as abstract machines are. One may even consider that in a sense, there are only machinic assemblages, whether virtual or manifest, and that territorialised assemblages and abstract machines are already potentially machinic assemblages.
We have only considered here limit situations that translate the fact that:
1territorialised machinic assemblages at level a mark a fear and a warding off of deterritorialisation at level b;
2those of level b mark, in another form, a refusal and repression of the diagrammatic effects at level c, by way of the systems of abstract machines;
3the machinic assemblages of level c mark, on the one hand, a return to territorialised indices and, on the other hand, a beyond of the abstract machines of level b, in that they bring a deterritorialising charge to the indices that allows them to pass through the ‘wall of the signifier’.
‘Do it’ could be the order-word for a pragmatic micropolitics. Not only can the Chomskyan axiom of grammaticality (S) no longer be accepted as going without saying, but it becomes the object of a sort of militant opposition. One refuses to consider that semiotic assemblages of all kinds have necessarily to organise themselves into phrases that are compatible with the system of dominant significations. A pragmatic order-word will therefore not seek to interpret, to reorganise significations, to compose with them; it will postulate that beyond their systems of redundancy, it is always possible to transform a semiotic assemblage. There is a primary political decision here, a primary axiom of pragmatics: the refusal to legitimate the signifying power manifested by the ‘evidence’ of dominant ‘grammaticalities’. The appreciation of a ‘degree of grammaticality’ then becomes a political matter. Rather than agreeing to remain prisoner of the redundancy of signifying tracings, one will endeavour to fabricate a new map of competence, new a-signifying diagrammatic coordinates. This is what the Leninists did during their rupture with the social-democrats, when they decided, with a certain arbitrariness, that on the basis of the constitution of a party of a new kind a split would be created between the proletarian avant-garde and the masses, the effect of which would be to radically transform their passive attitude, their tendency to spontaneity, and their ‘economist’ tendency. The fact that his ‘Leninist transformation’ later toppled over into the field of redundancy of Stalinist bureaucracy shows that in this domain, the systems of maps and tracings can always be inverted, that no structural foundation, no theoretical legitimation can definitively guarantee the maintenance of a revolutionary ‘competence’. Whatever the case may be, the Leninists made a new matter of expression rise up from the social field, a new map of the political unconscious, in relation to which all productions of utterances, including those of bourgeois movements, would be constrained to determine themselves. Another transformation of the unconscious map of the revolutionary movement had been produced by the Marxists of the First International, who literally ‘invented’ a new kind of working class, anticipating the sociological transformations that industrial societies were to experience (in effect, the class on which the communist movement of Marx’s era rested was essentially composed of artisans and journeymen: it was only at the end of the nineteenth century that it really began to be proletarianised). A micropolitical pragmatics will never accept systems of redundancy, which seem to be the most stuck in an ‘impasse’, as a fait accompli; it will endeavour to make processes of diagrammatisation emerge, ‘analysers’, collective assemblages of enunciation that will depose individuated modes of subjectivation and will form the basis on which previous micropolitical relations will be registered and reshuffled. But, once again, it cannot be a matter here simply of organisational, programmatic or theoretical instruments, but fundamentally of mutations in social pragmatics.
The task of a revolutionary pragmatics will thus consist in bringing about connections between transformational systems able to annul the effects of signifying generation. One is thus in the presence of two micropolitical orientations concerning semiotic systems as a whole. Diagrammatic transformations are thus able to carry their effects into no matter what semiotic register: whether it is a matter of symbolic semiologies (with mimetic or transitivist effects, for example), signifying semiologies (with systems of expression based on a limited range of discrete elements: phonemes, graphemes, distinctive features, etc.), or even natural modes of encoding. In each situation the pragmatic objective will consist in setting out the nature of the crystallisations of power that operate around a dominant transformational component: the map of black holes, semiotic branches and lines of flight (in Asiatic empires, the establishment of a despotic signifying writing, for example, or the emergence of a systematic signifying delirium in paranoia). Overthrow by a new diagrammatic component will reduce the effects of signifiance and individuation and lead to enunciation being nothing more than one element amongst others in machinic assemblages (the emancipation of a writing machine from its signifying function in poetic, musical, mathematical work, etc.). Pragmatic transformations will assemble their composition synchronically as a function of diverse political strategies; but they will equally organise their mutations diachronically on a machinic rhizome. Although evolution goes globally in the direction of a growing deterritorialisation, punctuated by always more brutal reterritorialisation on artificial stratifications, one really cannot set out general laws concerning them. And that is how it should be!
Pragmatic assemblages are machinic; they do not depend on universal laws properly speaking; they are subject to historical mutation. Thus one can speak of a ‘romantic complex’, of a ‘Popular Front complex’, a ‘Resistance complex’, a ‘positivist complex’, all of which have maintained their effects beyond their original historical localisation, without it being possible to give them the universal character that psychoanalysts accord to the Oedipus complex, or Maoists to the ‘revisionist’ complex. Pragmatic markers are not universals, they can always be called into question. Let us consider, for example, the fact that the most territorialised segmentarities have a ‘tendency’ to take control of more molar segmentarities. This is, in effect, a kind of law. But it only remains valid in the context of a given period, to the point when a revolutionary situation, overturning the maps of competence, reveals the existence of another machinism that was in the subterranean process of gnawing away at an earlier equilibrium. Differentiating coefficients of deterritorialisation ought nonetheless to allow political sequences to be vectorised – a ‘line’ of schizophrenisation versus a paranoid ‘line’ for example – in the struggle against bureaucratic transformations. But one will never be able to deduce from this, as some have believed they could consider so doing on the basis of Anti-Oedipus, that it is a matter here of a new Manichean alternative. It will only ever be a matter of a provisional orientation. Different kinds of entrance points must always be possible in a pragmatic system: that of performances of tracings or that of the competence of maps. In the first case, one will accept the repetitive character of deadlocked libidinal investments, one will even rely on them, so as to guarantee the minimal deterritorialisation of a body without organs on the basis of which other transformational operations will be possible (example: the positive aspects of regionalist struggles). In the other case, one will rely directly on a line of flight able to make the strata explode and bring about new semiotic branchings. Schematically speaking, and to borrow a different terminology, one can say of the generative pragmatics that it will concern itself specifically with empty and cancerous bodies without organs, whilst the transformational pragmatics will concern itself with full bodies without organs connected to the plane of consistency. But what brings these two points of entry together is that the simple fact of introducing a mode of semiotisation that concerns them in particular, the simple fact of memorising potentialities, of noting tracings and drawing up maps already sketches out diagrammatic effects: the simple fact of deciding to write down one’s dreams, for example, rather than passively interpreting them, the simple fact of sketching or miming them, could transform the map of the unconscious. One of the formidable traps of psychoanalysis is that it has managed to rely on the minimal transformation that the simple fact of having a discourse outside the habitual conditions of enunciation represents: the entire ‘mission’ of psychoanalysis having hitherto consisted in ‘extinguishing’ the diagrammatic effects of this transformation through the technique of the transference, and in pushing the discourse of the patient back into new grids of signifying redundancy.
A pragmatics of collective assemblages of enunciation will therefore oscillate constantly between these two kinds of semiotic micropolitics, elaborating from them a sort of technology for the calling into question of dominant significations. Under these conditions, discourse itself could become a war machine, with the constant risk of the re-establishing of a system of signifying redundancy.
Let us note that in effect, from the point of view of a transformational pragmatics, there is no fundamental difference between a war machine and diagrammatic linguistic machine, for the reason that at the level of the plane of consistency one cannot distinguish between the abstract machines that are manifested by a semiological substance of expression, and those that are manifested by the intensive traits of a more ‘material’ diagrammatic machine. Both are a part of the same kind of rhizome. Let us add that appreciating the effects of redundancy produced by a pragmatic transformation is not an unimportant objective; it is not, in effect, a matter of proposing a politics of novelty for novelty’s sake – a mimetic conversion to madness on the pretext of playing of a schizophrenic line against a paranoid line, for example! Pragmatic map-tracing assemblages intervene essentially at the level of the traits of matters of expression. In the last resort they are what determine the regime of coefficients of deterritorialisation, the rhythms of induction, the viscosity, the boomerang effects and so on, that are compatible with the fabrication of a body without organs (the injections of ‘caution’ so as not to bodge a body without organs). Tracking them thus doesn’t depend here on theoretical analyses but on a composition of systems of intensities. In sum, the redundancy of traits of the matter of expression relay a generative tree, a new rhizome can connect itself up and – this is perhaps the most general case – a microscopic element of a tree, a radicle, will outline the production of a new kind of local competence, whilst overcoded in a generative tree, one of the different semiotic components (perceptual, sensory, from thinking in image, speech, the socius, writing) will in any case crack. An intensive trait starts to work on its own count, a hallucinatory perception, synaesthesia, a perverse mutation, a play of images, detach themselves and in a single blow, the hegemony of the signifier is called into question.28 Generative trees, constructed according to the Chomskyan syntagmatic model, and which Jim McCawley, Jerrold Sadock, Dieter Wunderlich, etc., are trying to adapt for linguistic pragmatics,29 could thus open up and bud in all directions. A performative utterance, a promise, an order, can change the import of a situation – which is nothing to do with its signification – as a function of the appearance of a new transformation. It is obvious that a sermon does not have the same impact when it is given in the content of a transformation of conjugal, police or religious ‘power’. Saying ‘I swear’ before a judge or in a psychodramatic scene doesn’t have the same function, doesn’t involve the same kind of persona, nor the same kind of intersubjectivity.
The question, then, is not only one of knowing if a pragmatic transformation intervenes at different levels – semantic, syntactic, phonological, prosodic, etc. – but of studying how it intervenes on a micropolitical plane. And in the instance that its impact is not seen, this is because the analysis has been taken to its terminal point! This attitude is exactly the inverse of linguists who seek to minimise the role of pragmatic components and only agree to take them into account when they can no longer avoid them. Here one is no longer interrogating syntax and semantics so as to detect whether they harbour pragmatic elements: one interrogates the pragmatic semiotic compositions of assemblages of enunciation so as to detect the paralysing effects of signifying redundancies. When Bukharin takes the oath, from the point of view of the militant persona that he intended to remain faithful to until his death, this ambiguity can already be sensed in the official accounts. There is every reason to think that a syntactic, phonological, analysis of his discourse would allow the effects of the transformation ‘Moscow Trial’ on his oral expression to be brought out, and the international success that this formula has experienced. (Evidently it would be absurd to consider that such transformations of power, linked to school, to the tribunal, the party, the family, can be typified once and for all, in so far as they modify the signification of a performative, for example, or to seek to extract ‘universals’ from them.)
Generally acts of citizenship are considered to be the crowning point of a series that begins with a commitment to family values. Thus modes of mental organisation are staged, going from the most primitive of levels, like that of oral fixation, to the most ethereal of levels of sublimation. But in reality, things are not like that: every ‘stage’ can play a role at any time, and any one can come back on the system at a given point and blow it up. Let us repeat: no genetic finality, no general competence in a dominant adult language, will ever constitute a totalising reference for a particular performance. The objective of generative pragmatics is to determine in what way there is a coincidence between maps and what disjunctions might be utilised, what the scope of a power takeover by the signifier in a given system is, what the nature of the power formations that are plugged into the signifier S that organises and overcodes a corpus of utterances and propositions is. A repressive proposition, for example, doesn’t function in the same way when it is assembled in a molar military enunciation or in a molecular micro-fascist enunciation. Particular dialects, even idiolects, correspond to each situational rhizome. And in the case in which these are traversed by a language system, by a general grammaticality, it will always be a matter of a dominant overcoding instance functioning like Francophony in relation to the vernacular languages of the old French colonies, relayed today by new power formations.30