If metonymy were just a metonymy, its ground should be a contiguity that is not contaminated by analogy, and in that case the literal
separations
within a given discursive space would be fully in control of the limits of the metonymic movement.
Paul-de-Man-Material-Events
Elle est, par con- se?
quent, tout Trope d'un usage force?
et ne?
cessaire, tout Trope d'ou` re?
- sulte un sens purement extensif; ce sens propre de second origine, inter- me?
diaire entre le sens propre primitif et le sens figure?
, mais qui par sa propre nature se rapproche plus du premier qui du second, bien qu'il ait pu e^tre lui me^me figure?
dans le principe.
11
For instance, if I speak of the "wings of an airplane" or the "wings of the building," the expression was metaphoric at the beginning, but the difference with a proper metaphor that fully operates as a figure is that there is no proper designation of the referent. I am not free to call the "wing" in any other way.
So if the only defining feature of a catachresis is its being based in a figural name that has no counterpart in a proper one, it is clear that there is no specificity in the kind of figuration introduced by catachre- sis, and that it will repeat the figures of language sensu stricto with the only differentia specifica of there being no tropological movement from the proper to the figural. Thus Fontanier can speak of a cata- chresis of metonymy, of synecdoche and of metaphor. The difficulty is that the distinction between a catachresis of metonymy and a proper metonymy depends on the possibility of establishing an uncontaminat- ed frontier between the proper and the figural. As soon as some sou- plesse is introduced, the exchanges between these polar extremes be- come more complicated: the proper becomes the extreme, the reductio ad absurdum of a continuum that is figural through and through. With this, the possibility of a radical heterogeneity on which the sharp dis- tinction between catachresis and metonymy is grounded is consider- ably eroded. The only thing we can say is that the very possibility of a hegemonic relationship depends on this erosion, on keeping an unstable equilibrium between heterogeneity and contiguity, between catachresis and metonymy--an equilibrium whose conditions of ex- tinction would be either a heterogeneity without common measure be- tween the elements, or a contiguity that becomes total and, thus, ab- sorbs within an implicitly assumed space the contiguous positions as internal differences. 12 (These two conditions of extinction of the hege-
monic link are, in fact, only one and the same: in order to be radically heterogeneous, two elements require a common ground out of which their heterogeneity can be thought. )
On the other hand, however, all hegemony tries to retotalize and to make as necessary as possible the contingent links on which its articulating power is based. In this sense, it tends to metaphorical to- talization. This is what gives it its dimension of power. It is a power, however, that maintains the traces of its contingency, and is, in that sense, essentially metonymic. Hegemony is always suspended be- tween two impossible poles: the first, that there would be no dis- placement, that contiguity becomes mere contiguity, and that all tropological movement ceases--this would be the case of what Gramsci called the "corporative class"; the second, that the meta- phorical totalization becomes complete and that purely analogical re- lations fully saturate the social space--in which case we would have the "universal class" of the classical emancipatory discourse. Both poles are excluded by the hegemonic relation. It is only on the traces of (contingent) contiguity contaminating all analogy that a hegemon- ic relation can emerge.
III
I will now attempt to illustrate these propositions with a historical example showing an extreme attempt at metaphorical totalization, whose very failure shows the space in which the undecidable logic of hegemony operates. I am referring to the work of Georges Sorel.
Sorel's work is a product of that period of socialist thought that has been called, following the characterization of Thomas Masaryk, "the crisis of Marxism. " The increasing gap between the classical Marxist dogma, as codified in the Anti-Du? hring, and the actual turn of events opened a theoretical vacuum that various intellectual projects attempt- ed to fill.
Two themes quite frequent at that time--the impossibility of unify- ing historical events by conceptual means, the grounding of historical action in conviction and will--are certainly present in Sorel, but he is going to give them a new twist and to invest them with a new meaning, as he is going to see them from the viewpoint of a far more radical his- torical possibility. Let us consider the following passage from the Re? flexions sur la violence, which shows the deep gap separating Sorel's socialism from that of most of his contemporaries:
The Politics of Rhetoric 239
240 Ernesto Laclau
Dans une socie? te? aussi enfie? vre? e par la passion du succe`s a` obtenir dans la concurrence, tous les acteurs marchent droit devant eux comme de ve? ritables automates, sans se pre? occuper des grandes ide? es des socio- logues; ils sont soumis a` des forces tre`s simples et nul d'entre eux ne songe a` se soustraire aux conditions de son e? tat. C'est alors seulement que le de? veloppement du capitalisme se poursuit avec cette rigueur qui avait tant frappe? Marx et qui lui semblait comparable a` celle d'une loi naturelle. Si, au contraire, les bourgeois, e? gare? s par les blagues des pre? di- cateurs de morale ou de sociologie, reviennent a` un ide? al de me? diocrite? conservatrice, cherchent a` corriger les abus de l'e? conomie et veulent rompre avec la barbarie de leurs anciens, alors une partie des forces qui devaient produire la tendance du capitalisme est employe? e a` l'enrayer, du hasard s'introduit et l'avenir du monde est comple`tement inde? termine? .
Cette inde? termination augmente encore si le prole? tariat se convertit a` la paix sociale en me^me temps que ses mai^tres;--ou me^me simplement s'il conside`re toutes choses sous un aspect corporatif;--tandis que le socialisme donne a` toutes les contestations e? conomiques une couleur ge? ne? rale et re? volutionnaire. 13
Let us follow this argument closely. If the objective logic of historical change that Marx had presented depends for its full development on the bourgeoisie not being dominated by the ideal of a "me? diocrite? conservatrice"--because in the latter case "du hasard s'introduit et l'avenir du monde est comple`tement inde? termine? "--everything turns on whether that ideal will prevail or not. That prevalence, however, cannot be the result of identifiable objective economic processes, given that the very possibility of those processes taking place depends, for Sorel, on the absence of the "me? diocrite? conservatrice. " Here we find the cornerstone of Sorel's thought in its mature stage: social processes do not involve only displacements in the relationship of forces between classes, because a more radical and constitutive possibility is always haunting society: the dissolution of the social fabric and the implosion of society as a totality. So, society is not only suffering domination and exploitation: it is also threatened with decadence, with the only too real possibility of its radical not being. Let us see how in Sorel this dis- tinctive possibility opens the way to a new and peculiar logic in the re- lation between groups. There are two capital moments in this logic.
The first, is that the opposition that dominates Sorel's vision of the social is not primarily the one between bourgeoisie and proletariat but, rather, that between decadence and full realization of society. If the
proletariat as a social force receives historical priority, it is because it is seen as the main instrument in confronting decadence. But--and this is a crucial point--it is not the actual victory of the proletariat against the bourgeoisie that will bring about "grandeur" and will arrest deca- dence, but the very fact of the open confrontation between the two groups. Without confrontation there is no identity; social identities re- quire conflict for their constitution. This explains Sorel's critique of democracy. If social identities require open conflict for their full consti- tution, any attempt to dilute, reduce, or even regulate that conflict can only be an instrument of decadence and corruption of those identities. Democratic participation in public institutions is the sure road leading to such corruption. So, Sorel had to see in Marxism not a scientific doctrine explaining the objective laws of capitalism, but a finalistic ide- ology of the proletariat, grounded in class struggle. Social relations, left to themselves, are just "me? lange. " Only the will of determined so- cial forces gives a consistent shape to social relations, and the determi- nation of that will depends on the violent confrontation with other wills.
But, second, if the historical justification of the action of the prole- tariat is given by its being the only force capable of opposing the deca- dence of civilization, that justification is indifferent to the contents of the proletarian program and depends entirely on the contingent ability of those contents to bring about an effect that is external to them- selves. There is no ethical justification that is intrinsic to socialism. This has two capital consequences. The first, that all social identity or social demand is constitutively split. It is, on the one hand, a concrete demand; but, on the other hand, it can also be the bearer, the incarna- tion of social "grandeur" as opposed to decadence. "Grandeur" and "decadence" do not have intrinsic contents of their own, but are the empty signifiers of a fullness of society (or its opposite, its corruption or nonbeing) that can be actualized by the most different social forces. So--and this is the second consequence--it is enough that the working class shows itself as a limited historical actor, closed in its corporative demands and incapable of incarnating the will to fullness of society, for its claims to lose all legitimacy. And the political trajectory of Sorel is a living example of the contingency of this relation between working- class demands and "grandeur": he passed from being a theoretician of revolutionary syndicalism to ally himself with a fraction of the mon- archist movement, and ended his career by being close to the Third International. The diffusion of Sorelian themes in antagonistic social
The Politics of Rhetoric 241
242 Ernesto Laclau
movements, from bolshevism to fascism, is an even more telling ex- ample of the ambiguous possibilities that his de? marche was opening. For Sorel, the action (violence) is increasingly separated from its own contents and exclusively judged by the effect the latter have on the identity of the actors.
Let us now translate these reflections on Sorel into our tropological argument. Any attempt by the proletariat to constitute its subjectivity through a variety of loosely related subject positions can only lead, ac- cording to Sorel, to corporative integration and to decadence; so all metonymic variation had to be eliminated. In that case, how to aggre- gate working-class struggles in such a way that the proletarian identity is maintained and reinforced? Through an education of the will ground- ed in the myth of the general strike. Each action of the workers-- whether a strike, a demonstration, or a factory occupation--should be seen, not in its own specificity and particular objectives, but as one more event in the formation of the revolutionary will. That is, these ac- tions are all analogous from the point of view of their ultimate deep aims and are, as a result, in a relation of metaphorical substitution with each other. Their mutual relations--as that between flies and summer in Proust's text--are necessary. The drawback of this vision is that, in that case, the myth that unifies the struggles beyond all speci- ficity cannot be specific either. The reduction of all specificity to the repetition of something analogous can only be the metaphor of meta- phoricity as such. We know what this involves: the interruption of any hegemonic operation. The metaphor of metaphoricity can only be a zero that is in no tropological relation with a one, or--at most--a zero that is in a catechretical relation with only one position. Only at that price can revolutionary closure be achieved. And this is precisely what Sorel attempts to achieve by making the general strike totally heteroge- neous with the empirical world of limited and partial struggles. The general strike is presented as a myth and not a utopia--it has lost all the detailed descriptive features of the latter; it has no particular objec- tives; it is merely an empty image that galvanizes the consciousness of the masses; it is exhausted in this last function without possibly cor- responding to any actual historical event. It is a radical nonevent that is, paradoxically, the condition of all events if there is going to be a "grandeur" in society.
In that case, why revolutionary general strike rather than anything else? Is there any ground to think that the general strike is the (necessary) catachresis of that radical nonevent which brings about "grandeur"?
Sorel cannot answer this question, and the oscillations of his political career are a clear indication that the question is unanswerable. The relationship between "grandeur" and general strike is a hegemonic in- carnation, which involves all metaphoric aggregation being ultimately grounded in (reversible) metonymic displacements. The attempt to ground the revolutionary will in a metaphoric totalization that would avoid the particularism of hegemonic variations ends in failure. As Plato knew, perhaps better than Sorel--only protracted metonymic displace- ments between Athens and Siracusa can give some hope that the king agrees to become a philosopher.
Perhaps we could still make this point in a slightly different way: it is only through the pure, irreducible event that consists in a contingent displacement not retrievable by any metaphoric reaggregation that we can have a history, in the sense of both Geschichte and Historie. It is because there is hegemony (and metonymy) that there is history. Couldn't some deconstructive strategies, such as iteration, be seen as attempts at introducing metonymy at the heart of metaphor, displace- ment at the heart of analogy? Genette (following Blanchot) tries to show, in his analysis of Proust, how the latter moved from a structure of his novel conceived as a succession of poetic instants, of punctual moments, to a conception of his whole narrative in which anamnesis is inseparable from--it is actually governed by--the whole process of narration. As he points out:
Sans me? taphore, dit (a` peu pre`s) Proust, pas de ve? ritables souvenirs; nous ajoutons pour lui (et pour tous): Sans me? tonymie, pas d'enchai^ne- ment de souvenirs, pas d'histoire, pas de roman. Car c'est la me? taphore qui retrouve le Temps perdu, mais c'est la me? tonymie qui le ranime, et le remets en marche: qui le rend a` lui-me^me et a` sa ve? ritable "essence", qui est sa propre fuite et sa propre Recherche. Ici donc, ici seulement--par la me? taphore mais dans le me? tonymie--, ici commence le Re? cit. 14
Perhaps this is, exactly, the intellectual displacement leading from Sorel to Gramsci. Whereas in the first the analogizing movement of the metaphor of metaphoricity led to a repetition that tried to eliminate the possibility of any proper event, Gramsci's notion of war of posi- tion, of a narrative-political displacement governed by a logic of pure events which always transcend any preconstituted identity, announces the beginning of a new vision of historicity--one governed by the in- eradicable tension between metonymy (or synecdoche) and metaphor.
The Politics of Rhetoric 243
244 Ernesto Laclau
IV
We now arrive at a decisive point in our argument on hegemony. If hegemony means the representation, by a particular social sector, of an impossible totality with which it is incommensurable, then it is enough that we make the space of tropological substitutions fully visible, to enable the hegemonic logic to operate freely. If the fullness of society is unachievable, the attempts at reaching it will necessarily fail, although they will be able, in the search for that impossible object, to solve a variety of partial problems; that is, the particularism of the struggles, which was systematically demoted in Sorel's analysis, now becomes central. With this, the metonymic game occupies center stage, and politics--which was for Sorel the nemesis of proletarian action--takes the upper hand.
All this becomes more visible if we compare Sorel's discursive de? - marche with other socialist discourses of the time, which oriented themselves in the opposite direction. Let us clarify, however, an impor- tant point before engaging in that comparison. Both metaphor and metonymy are tropological movements, that is, forms of condensation and displacement whose effects are achieved on the basis of going be- yond literal meaning. Now, from this point of view, classical Marxian discourse presented itself as the zero degree of tropology, as a scientific discourse describing the necessary laws of history, which did not need to go beyond the literality of their formulation in order to achieve the totalizing effects that they postulated. That this ideal of scientificity points to an impossible task, and that whatever totalizing effects Marxian discourse could have can only be achieved through putting into operation a whole arsenal of tropological movements, is well known; but the important point is that as an ideal that governed its own discursivity, literality is fully present in it and produces a whole set of concealing effects. Sorel had ceased to believe in objective, neces- sary laws of history and wanted to substitute them with an artificial necessity grounded in the power of the will; so he had, as we have seen, to have full recourse to the principle of analogy--which in a literal dis- course of necessary, objective laws would have no incidence--and to in- stall himself, fully conscious of the fact, in the terrain of metaphor. But, as we have also seen, metaphorical necessity was decisively contami- nated by metonymic contingency. What were, in that case, the politico- discursive and strategic effects of accepting as inevitable the metonymic terrain?
Let us consider the discussions in Russian social democracy at the turn of the last century. The general view was that Russia was ripe for a bourgeois-democratic revolution in which the bourgeoisie, as in all major revolutions of the West, would carry on the tasks of sweeping away the remainders of feudalism and creating a new state along liberal- democratic lines. The drawback was that the Russian bourgeoisie ar- rived in the historical arena too late and was weak and incapable of carrying out its political tasks. The need for a democratic revolution was, however, still there. This led to the conclusion--drawn at least by some sections of the social democrats--that in that case those tasks had to be taken up by some other social sector that was not its natural bearer:--in this case, the working class. This relationship by which a sector takes up tasks that are not its own is what the Russian social democrats called hegemony. So we see how the political steps that this analysis took led in the opposite direction to Sorel's. Whereas the latter tried to close the working class around its own natural tasks through metaphoric totalizations, here we find the opening of a field of meto- nymic displacements in the relations between tasks and agents, an un- decided terrain of contingent articulations in which the principle of contiguity prevails over that of analogy. It was only the contingent pe- culiarity of the Russian situation--the presence of a weak bourgeoisie and a strong working class--that was at the root of the working-class leadership in the democratic revolution.
This complicated dialectics between analogy and contiguity can be seen to expand in a plurality of directions. First, as the nontropological succession of programmed stages is interrupted, a space of logical inde- terminacy arises: "Tsarism, having entered into complete contradiction with the demands of Russia's social development, continued to exist thanks to the power of its organisation, the political nullity of the Russian bourgeoisie and its growing fear of the proletariat. "15
Second, this indeterminacy is the source of pure relations of con- tiguity that break the possibility of totalizations in terms of either syntagmatically retrievable differences or metaphorically "necessary" aggregations:
Russian capitalism did not develop from artisanal trade via the manu- facturing workshop to the factory for the reason that European capital, first in the form of trade capital and later in the form of financial and industrial capital, flooded the country at a time when most Russian artisanal trade had not yet separated itself from agriculture. Hence the
The Politics of Rhetoric 245
246 Ernesto Laclau
appearance in Russia of modern capitalist industry in a completely primitive economic environment: for instance, a huge Belgian or American industrial plant surrounded by dirty roads and villages built of straw and wood, which burn down every year, etc. The most primitive begin- nings and the most modern European endings. (339)
This gap, which interrupts any nontropological succession of necessary stages, but also any metaphoric aggregation of events around a pre- given necessary point, gives proletarian identity in Russia an open character in which contingent displacements, pure events, assume a constitutive role that no a prioristic logic can govern:
I remember an old friend, Korotkov, a cabinetmaker from Nikolayev, who wrote a song back in 1897. It was called The Proletarians' March and it began with the words: "We are the alpha and the omega, we are the beginning and the end . . . " And that's the plain truth. The first letter is there and so is the last, but all the middle of the alphabet is missing. Hence the absence of conservative traditions, the absence of castes within the proletariat, hence its revolutionary freshness, hence, as well as for other reasons, the October Revolution and the first workers' government in the world. But hence also the illiteracy, the absence of organisational know- how, the lack of system, of cultural and technical education . . . (340)
And then the inevitable consequence:
From the viewpoint of that spurious Marxism which nourishes itself on historical cliche? s and formal analogies . . . the slogan of the seizure of power by the Russian working class was bound to appear as a mon- strous denial of Marxism. . . . What then is the real substance of the problem? Russia's incontestably and incontrovertibly backward devel- opment, under the pressure of the higher culture of the West, leads not to a simple repetition of the Western European historical process but to a set of fundamentally new features which requires independent study. . . . Where there are no "special features", there is no history, but only a sort of pseudo-materialistic geometry. Instead of studying the living and changing matter of economic development it is enough to notice a few outward symptoms and adapt them to a few ready-made cliche? s. (339)
Could it be clearer? Historicity is identified with "special features" un- assimilable to any form of repetition. History is a field of contingent dis- placements that are not retrievable by any of the (analogical) figures of the same.
Of course, this field of contingent variations can be more or less ex- tended, depending on the width of the area in which the literal still pre- vails and arrests the tropological movement. Now what happened in socialist discourses, like those we are considering, was that what we have described as a tropological movement expanded ever more and covered wider and wider sections of political life. Let us consider a concept such as "combined and uneven development. " It was intro- duced to refer to the experience of social struggles in Third World countries, in which--even more than in the case of Russia--a non- orthodox combination of developments that should have correspond- ed to successive stages makes more contingent and risky hegemonic interventions possible. In the 1930s, Trotsky drew the inevitable con- clusion: an uneven development is the terrain of all social and political struggles in our time. The only thing we have is an unlimited tropologi- cal movement that is the very terrain in which society constitutes itself. And we see why metonymy is, in some sense, more "primordial" than metaphor (or, as in other of de Man's analyses, why allegory takes precedence over the symbol): because in a situation of radical contin- gency no criterion of analogy is stable; it is always governed by chang- ing relations of contiguity that no metaphorical totalization can con- trol. Metaphor--and analogy--is at most a "superstructural" effect of a partial stabilization in relations of contiguity that are not submitted to any literal principle of a priori determination.
This process of general rhetorization only takes place as far as none of the conditions in which each of the tropoi would become what it lit- erally claims to be can be met.
If metonymy were just a metonymy, its ground should be a contiguity that is not contaminated by analogy, and in that case the literal separations within a given discursive space would be fully in control of the limits of the metonymic movement. If analogy dominated unchallenged, a full totalization would have taken place that would make analogy collapse into identity--and the tropo- logical movement would cease. If synecdoche was actually able to sub- stitute the whole for the part, this would mean that the whole could have been apprehended independently of the part. If catachresis could be grounded in a tropological movement that started from total het- erogeneity, it could only take place insofar as the distinction between the homogeneous and the heterogeneous would be established with entire precision. It is as if in some way the conditions for a rhetoric whose tropological movements are going to occupy the terrain of a ground that is not itself grounded are to be found in the impossibility
The Politics of Rhetoric 247
248 Ernesto Laclau
of taking the definitions of each of the tropoi at face value, and in the need to stress the logics by which each tends to fade into the other. The same for hegemony: the conditions of its full success are the same as the conditions of its extinction.
This can be shown by a couple of historical examples. The first con- cerns Italy. At the end of World War II there was a confrontation of tendencies, within the Italian Communist Party, about the right strate- gy to be followed in the new democratic environment. There were two positions: one that asserted that the Communist Party, being the party of the working class and the latter being an enclave in the industrial north, had to limit its efforts mainly to creating forms of representa- tion for that enclave; the second, more Gramscian, maintained that the party had to build up its hegemony by spreading its activities to a va- riety of areas, the agrarian Mezzoggiorno included. How was this pos- sible, given the particularistic social and geographical location of the working class? Simply, by making the party and the unions the rallying point of a variety of democratic initiatives in a country moving away from fascist dictatorship. The democratic initiatives postulated by this approach were entirely contingent--their success was not guaranteed by any logic of history--and depending, thus, on the construction of a collective will; but, in a way different from the Sorelian will, they were not aiming for the reinforcement of a purely proletarian identity. They tended, instead, to the creation of a multifarious democratic identity, always spreading beyond itself in directions only graspable through a contingent narration. Togliatti wrote in 1957:
A class may lead society insofar as it imposes its own rule, and to this end the force of arms can also be used. It becomes a national class, how- ever, only insofar as it solves the problems of the whole of society. . . . The proletariat becomes a national class insofar as it takes on these problems as its own and thence comes to know, by the process of chang- ing it, the whole reality of national life. In this way it produces the con- ditions of its own political rule, and the road to becoming an effective ruling class is opened. . . .
We have to spread the activity of an organised vanguard over the whole area of society, into all aspects of national life. This activity must not be reduced to preaching propaganda, to phrase-making or clever tactics, but must stick closely to the conditions of collective life and give, therefore, a foundation, real possibilities and prospects to the movement of the popular masses. . . . Our struggle for the unity of popu-
lar, democratic forces is, therefore, not imposed by tactical skills, but is an historical requirement, both to maintain conquests already achieved, to defend and safeguard democracy, and to develop it. 16
Here we have a tropological space in which each of the figures tends to fade into the other. The different struggles and democratic initiatives are not united to each other by necessary links--so we have metonymi- cal relations of contiguity. The hegemonic operation tries, however, to make the condensation of these struggles as strong as possible--so the metonymics fade into metaphoric totalization. The hegemonic relation is synecdochal, as far as a particular sector in society--the working- class party, in this case--tends to represent a whole exceeding it. As, however, this whole lacks any precisely defined limits, we are dealing with an impure synecdoche: it consists in the undecidable movement between a part attempting to incarnate an undefinable whole, and a whole that can only be named through its alienation in one of its parts. Finally, the heterogeneity can only be a relative one--with the result that the line separating catachresis from metonymy becomes also un- decidable. I think that all the main categories of Gramscian theory-- war of position, collective will, organic intellectuals, integral state, his- torical bloc, hegemony--could be read rhetorically: as circumscribing the space of tropological movement that brings about new strategic flexibility in political analysis.
A comparable discussion took place in the South African context, in the years before the end of apartheid, between the so-called workers and populist tendencies within the liberation movement. The first ten- dency, based largely in the trade-union movement, asserted the need for an immediate socialist transformation constituted around a work- ing class whose protagonist role resulted from its structural centrality in capitalist society. Contingent and particularistic displacements in the search for alliances were reduced to a minimum. The populist camp, on the contrary, based on the principles of the Freedom Charter, made contingent hegemonic articulations the cornerstone of its strategy. As David Howarth and Aletta Norval have asserted,
the Charterists have suggested that the formation of political conscious- ness cannot be exclusively attributed to factory floor experiences, or de- rived simply from the agent's location in the relations of production, but occurs in a much wider discursive context. Instead of positing an ab- stract working class persona analytically separable from the complex set of discourses in which the worker is situated, the working class is
The Politics of Rhetoric 249
250 Ernesto Laclau
regarded as a real social force forming an essential component of the na- tionally oppressed people. . . . In this sense, the working class as a real social force engaged in struggle is always to some extent marked by struggles, identities and discourses which cannot simply be reduced to its position in the relations of production; their leadership will only be attained by means of actively introducing socialist discourses into the struggle for national liberation. 17
Although Howarth and Norval have pointed out the limited character of the hegemonic opening that the populist camp was postulating, it is clear that the strategic confrontation was taking place along lines simi- lar to those that we have so far discussed: in one case a principle of analogy by which the South African working class repeats a worker's identity dictated by the abstract analysis of the capitalist relations of production; in the other case, a succession of merely contiguous articu- lations that governs a predominantly contingent and contextualized narrative.
This political argument could, obviously, be extended in a variety of directions beyond the socialist discourses that we have considered here. There is one, however, that I would like to stress. I see the history of democracy as divided by one fundamental cleavage. On the one hand, we have democracy as the attempt to construct the people as one, a homogeneous social actor opposed either to "power" or to an external enemy--or to a combination of both. This is the Jacobin con- ception of democracy, with its concomitant ideal of a transparent com- munity unified--if necessary--by terror. This is the tradition that runs, with very analogous structural features, from Robespierre to Pol Pot. The discourses around which this democratic ideal is constructed are, obviously, predominantly metaphoric--although, for reasons previ- ously mentioned, they cannot conceal their metonymic foundations. On the other hand, we have democracy as respect of difference, as shown, for instance, in multiculturalism or in the new pluralism asso- ciated with contemporary social movements. Here we have discourses that are predominantly metonymic, for although--given the impos- sibility of a pure differential, nontropological closure--some effect of metaphoric aggregation is inevitable, it will be an aggregation that always keeps the traces of its own contingency and incompleteness visible. Within this basic polarity there are, obviously, all kinds of pos- sible intermediate combinations that we can start exploring through the variety of tropoi to be found in classical rhetoric.
Paul de Man's main contribution to this task lies not in anything that he had to say about politics--something his untimely death pre- vented him from doing--but rather, in two main accomplishments. The first is to have extended the field of rhetoric--or, rather, of rhetoricity-- to the ensemble of language, to have made it a constitutive dimension of language as such. The second is to have deconstructed the dominant tropoi of the Romantic tradition--such as symbol and metaphor-- showing the contingent infrastructure of more humble tropoi on which any totalizing effect is grounded. I have attempted in this essay to show the potential importance of both accomplishments for the elaboration of a theory of hegemony.
NOTES
A preliminary version of this essay was presented at the "Culture and Materiality" conference that took place at the University of California, Davis, on April 23-25, 1998. It was also discussed in the seminar on Ideology and Discourse Analysis, University of Essex, a month later. I want to thank those whose commentaries led me to introduce precisions in my text and, in some cases, partial reformulations of the argument: at Davis, Jacques Derrida, Fredrick Jameson, J. Hillis Miller, and Andrzej Warminski; at Essex, David Howarth and Aletta Norval.
1. Stefano Rosso, "An Interview with Paul de Man," in Paul de Man, The Resistance to Theory (Minneapolis: University of Minnesota Press, 1993), 121.
2. Paul de Man, "Pascal's Allegory of Persuasion," in Aesthetic Ideology (Minneapolis: University of Minnesota Press, 1996), 51-69. Subsequent page refer- ences are given in the text.
3. For the full development of this argument, see my Emancipations(s) (London: Verso, 1996), especially the essay "Why Do Empty Signifiers Matter to Politics? " 36-46.
4. Cf. the disagreement between de Man and Louis Maris, as presented by de Man in "Pascal's Allegory of Persuasion," 60.
5. What is important is to realize that these retotalizations do not operate through a simple and retrievable negation. As de Man asserts: "What is here called, for lack of a better term, a rupture or a disjunction is not to be thought of as a negation, however tragic it may be. Negation, in a mind as resilient as Pascal's, is always susceptible of being reinscribed in a system of intelligibility. . . . It is possible to find, in the terminology of rhetoric, terms that come close to designating such disruptions (e. g. , parabasis or anacoluthon), which designate the interruption of a semantic continuum in a manner that lies beyond the power of reintegration" (ibid. , p. 61). But the very fact that there are tropoi which make describable that which is beyond the power of reintegration clearly shows that we are not dealing with a simple collapse of the system but, rather, with an orderly drifting away from what would have otherwise been the conditions of its full closure. It is in the field of this drifting away that the hegemonic logics operate.
6. Paul de Man, "Reading (Proust)," in Allegories of Reading Figural Language
The Politics of Rhetoric 251
252 Ernesto Laclau
in Rousseau, Nietzsche, Rilke and Proust (New Haven: Yale University Press, 1979), 57-58.
7. Paul de Man, Blindness and Insight: Essays in the Rhetoric of Contempor- ary Criticism (Minneapolis: University of Minnesota Press, 1983), 284.
8. Ge? rard Genette, "Me? tonymie chez Proust," in Figures III (Paris, E? ditions du Seuil, 1972), 41-43. Paul de Man finds the use by Genette of the category of diegetic metaphor limited as far as Proust is concerned. Both, however, agree in privileging the metonymic transitions in Proust's text.
9. Ernesto Laclau and Chantal Mouffe, Hegemony and Socialist Strategy: Towards a Radical Democratic Politics (London: Verso, 1985).
10. I thank J. Hillis Miller for having called my attention to the need for elabo- rating more the distinction between catachresis and metonymy--a distinction that, as will be seen, is crucial for my analysis.
11. Pierre Fontanier, Les figures du discours (Paris: Flammarion, 1968), 213. "Generally speaking there is catachresis when a sign already linked to a first idea is also linked to a new idea which itself does not have or no longer has a literal mean- ing in language. It is, as a result, any trope whose use is forced and necessary, any trope from which a purely extended meaning is derived; this literal meaning of sec- ond origin, mediating between the original literal meaning and the figurative mean- ing, but which by its own nature is closer to the first than to the second, even though it might have been figurative in the beginning. "
12. With this, of course, the tropological movement would come to an end.
13. Georges Sorel, Re? flexions sur la violence (Paris: E? ditions du Seuil, 1990), 77. "In a society so enfevered by the passion of competitive success, all the actors march straight ahead like true automatons, unconcerned with the great ideas of the sociologists, and none of them considers extricating himself from that condi- tion. It is only then that the development of capitalism is pursued with that rigor which so struck Marx and which seemed to him comparable to a natural law. If, on the contrary, the bourgeois, diverted by the jokes of moral or sociological preachers, return to the ideal of conservative mediocrity, seek to correct the abuses of the economy, and wish to break with the barbarity of their ancestors, then a part of the forces which should produce the dynamic of capitalism are employed in checking it, chance is introduced, and the future of the world becomes completely indeterminate.
"This indeterminacy further increases if the proletariat is converted to social peace at the same time as their masters--or even simply that it considers everything from a corporative perspective--while socialism gives to all economic struggles a general and revolutionary character. "
14. Genette, "Me? tonymie chez Proust," 63. "Without metaphor, says Proust (more or less), there are no true memories; we can add for him (and for everybody): without metonymy, no linkage of memories, no history, no novel. Because it is metaphor that recovers lost Time, which reanimates and sets it in motion: which re- turns it to itself and its true 'essence,' which is in its own flight and its own Search. So here, here alone--through metaphor but within metonymy--Narration begins. "
15. Leon Trotsky, 1905 (London: Allen Lane, 1971), 328. Subsequent page ref- erences are given in the text.
16. Palmiro Togliatti, On Gramsci and Other Writings (London: Lawrence and Wishart, 1979), 157-59.
17. David Howarth and Aletta Norval, "Subjectivity and Strategy in South African Resistance Politics: Prospects for a New Imaginary," Essex Papers in Politics and Government, Department of Government, University of Essex (May 1992): 9.
The Politics of Rhetoric 253
254
How Can I Deny That These Hands and This Body Are Mine?
Judith Butler
I remember a sleepless night last year when I came into my living room and turned on the television set to discover that C-Span was offering a special session on feminist topics, and that the historian Elizabeth Fox- Genovese was making clear why she thought women's studies had con- tinuing relevance, and why she opposed certain radical strains in femi- nist thinking. Among those positions she most disliked she included the feminist view that no stable distinction between the sexes could be drawn or known, a view that suggests that the difference between the sexes is itself culturally variable or, worse, discursively fabricated, as if it is all a matter of language. Of course, this did not help my project of falling asleep, and I became aware of being, as it were, a sleepless body in the world accused, at least obliquely, with having made the body less rather than more relevant. Indeed, I was not altogether sure that the bad dream from which I had awoken some hours earlier was not in some sense being further played out on the screen. Was I waking or was I dreaming? After all, it was no doubt the persecutory dimension of paranoia that hounded me from the bed. Was it still paranoia to think that she was talking about me, and was there really any way to know? If it was me, then how would I know that I am the one to whom she refers?
I relate this incident to you here today not only because it fore- shadows the Cartesian dilemmas with which I will be preoccupied in the following paper, and not because I propose to answer the question of whether sexual difference is only produced in language. I will, for the moment, leave the question of sexual difference to be returned to another time. 1 The problem I do propose to address emerges every
time we try to describe the kind of action that language exercises on the body or, indeed, in the production or maintenance of bodies. And we do tend to describe language as actively producing or crafting a body every time we use, implicitly or explicitly, the language of discur- sive construction.
In the consideration of Descartes's Meditations that follows, I pro- pose to ask whether the way in which Descartes posits the irreality of his own body does not allegorize a more general problem of positing that is to be found in various forms of constructivism and various criti- cal rejoinders to a constructivism that is sometimes less well under- stood than it ought to be. The name of this paper that I have already begun, but not yet begun, is: "How Can I Deny That These Hands and This Body Are Mine? " These are, of course, Descartes's words, but they could be ours or, indeed, mine, given the dilemmas posed by contem- porary constructivism.
The language of discursive construction takes various forms in con- temporary scholarship and sometimes it does seem as if the body is cre- ated ex nihilo from the resources of discourse. To claim, for instance, that the body is fabricated in discourse is not only to figure discourse as a fabricating kind of activity, but to sidestep the important questions of "in what way" and "to what extent. " To say that the line between the sexes, for instance, must be drawn, and must be drawable, is to concede that at some level the stability of the distinction depends on a line being drawn. But to say that we must be able to draw a line in order to stabilize the distinction between the sexes may simply mean that we must first grasp this distinction in a way that allows us then to draw the line, and the drawing of the line confirms a distinction that is somehow already at hand. But it may mean, conversely, that there are certain conventions that govern how and where the line ought or ought not to be drawn, and that these conventions, as conventions, change through time, and produce a sense of anxiety and of unknow- ingness precisely at the moment in which we are compelled to draw a line in reference to the sexes. The line then lets us know what will and will not qualify as "sex"; the line works as a regulatory ideal, in Foucault's sense, or a normative criterion that permits and controls the appearance and knowability of sex. Then the question, which is not easily settled, becomes: do the conventions that demarcate sexual dif- ference determine in part what we "see" and "comprehend" as sexual difference? It is, you might surmise, not a large leap from this claim to the notion that sexual difference is fabricated in language. But I think
How Can I Deny . . . ? 255
256 Judith Butler
that we may need to look more carefully before either championing or reviling this conclusion.
The language of construction risks a certain form of linguisticism, the assumption that what is constructed by language is therefore also language, that the object of linguistic construction is nothing other than language itself. Moreover, the action of this construction is con- veyed through verbal expressions that sometimes imply a simple and unilateral creation at work.
Language is said to fabricate or to figure the body, to produce or construct it, to constitute or to make it. Thus, language is said to act, which involves a tropological understanding of language as perform- ing and performative. There is, of course, something quite scandalous involved in the strong version of construction that is sometimes at work when, for instance, the doctrine of construction implies that the body is not only made by language, but made of language, or that the body is somehow reducible to the linguistic coordinates by which it is identified and identifiable, as if there is no nonlinguistic stuff at issue. The result is not only an ontological realm understood as so many ef- fects of linguistic monism, but the tropological functioning of language as action becomes strangely literalized in the description of what it does, and how it does what it does. And though de Man often argued that the tropological dimension of discourse works against the perfor- mative, it seems here that we see, as I believe we do in de Man's discus- sion of Nietzsche, the literalization of the trope of performativity.
I want to suggest another way of approaching this question that re- fuses the reduction of linguistic construction to linguistic monism, and which calls into question the figure of language acting unilaterally and unequivocally on the object of construction. It may be that the very term construction no longer makes sense in this context, that the term deconstruction is better suited to what I propose to describe, but I con- fess to not really caring about how or whether these terms are stabilized in relation to each other or, indeed, in relation to me. My concerns are of another order, perhaps in the very tension that emerges as the prob- lem of discursive construction comes into tension with deconstruction.
For my purposes, I think it must be possible to claim that the body is not known or identifiable apart from the linguistic coordinates that establish the boundaries of the body without thereby claiming that the body is nothing other than the language by which it is known. This last claim seeks to make the body an ontological effect of the language that governs its knowability. But this view fails to note the incommensura-
bility between the two domains, an incommensurability that is not pre- cisely an opposition. Although one might accept the proposition that the body is only knowable through language, that the body is given through language, it is never fully given in that way, and to say that it is given partially can only be understood if we also acknowledge that it is given, when it is given, in parts; it is, as it were, given and withheld at the same time, and language might be said to perform both of these operations. Although the body depends on language to be known, the body also exceeds every possible linguistic effort of capture. It would be tempting to conclude that this means that the body exists outside of language, that it has an ontology separable from any linguistic one, and that we might be able to describe this separable ontology. But this is where I would hesitate, perhaps permanently, for as we begin that description of what is outside of language, the chiasm reappears: we have already contaminated, though not contained, the very body we seek to establish in its ontological purity. The body escapes its linguistic grasp, but so too does it escape the subsequent effort to determine on- tologically that very escape. The very description of the extralinguistic body allegorizes the problem of the chiasmic relation between language and body and so fails to supply the distinction it seeks to articulate.
To say that the body is figured chiasmically is to say that the follow- ing logical relations hold simultaneously: the body is given through language but is not, for that reason, reducible to language. The lan- guage through which the body emerges helps to form and establish that body in its knowability, but the language that forms the body does not fully or exclusively form it. Indeed, the movement of language that appears to create what it names, its operation as a seamless perfor- mative of the illocutionary persuasion, covers over or dissimulates the substitution, the trope, by which language appears as transitive act, that is, by which language is mobilized as a performative that simulta- neously does what it says. If language acts on the body in some way, if we want to speak, for instance, of a bodily inscription, as so much cul- tural theory does, it might be worth considering whether language lit- erally acts on a body, and whether that body is an exterior surface for such action, or whether these are figures that we mobilize when we seek to establish the efficacy of language.
This leads to a converse problem, namely, the case in which lan- guage attempts to deny its own implication in the body, in which the case for the radical disembodiment of the soul is made within language. Here it is a question of the way in which the body emerges in the very
How Can I Deny . . . ? 257
258 Judith Butler
language that seeks to deny it, which suggests that no operation of lan- guage can fully separate itself from the operation of the body.
Language itself cannot proceed without positing the body, and when it tries to proceed as if the body were not essential to its own op- eration, figures of the body reappear in spectral and partial form with- in the very language that seeks to perform their denial. Thus, language cannot escape the way in which it is implicated in bodily life, and when it attempts such an escape, the body returns in the form of spectral figures whose semantic implications undermine the explicit claims of disembodiment made within language itself. Thus, just as the effort to determine the body linguistically fails to grasp what it names, so the ef- fort to establish that failure as definitive is undermined by the figural persistence of the body.
The way in which I propose to show this chiasmic relation is through a reconsideration of the opening Meditations of Descartes, the ones in which he seeks to bring the reality of the body into question. Descartes seeks to know whether he can deny the reality of his own body and, in particular, the reality of his limbs. Suspended and inscrutable limbs reemerge in de Man's essay "Phenomenality and Materiality in Kant" in ways that suggest a metonymic relation to the problem that Descartes poses. For de Man, the body within the third Critique is understood, if we can use that word, as prior to figuration and cognition. In Descartes, it emerges as a particular kind of figure, one that suspends the ontologi- cal status of the term, and this raises the question of any absolute sepa- rability between materiality and figuration, a distinction that de Man on some occasions tries to make as absolutely as possible. I hope to make this chiasmic relation between those two texts clear toward the end of my remarks.
For instance, if I speak of the "wings of an airplane" or the "wings of the building," the expression was metaphoric at the beginning, but the difference with a proper metaphor that fully operates as a figure is that there is no proper designation of the referent. I am not free to call the "wing" in any other way.
So if the only defining feature of a catachresis is its being based in a figural name that has no counterpart in a proper one, it is clear that there is no specificity in the kind of figuration introduced by catachre- sis, and that it will repeat the figures of language sensu stricto with the only differentia specifica of there being no tropological movement from the proper to the figural. Thus Fontanier can speak of a cata- chresis of metonymy, of synecdoche and of metaphor. The difficulty is that the distinction between a catachresis of metonymy and a proper metonymy depends on the possibility of establishing an uncontaminat- ed frontier between the proper and the figural. As soon as some sou- plesse is introduced, the exchanges between these polar extremes be- come more complicated: the proper becomes the extreme, the reductio ad absurdum of a continuum that is figural through and through. With this, the possibility of a radical heterogeneity on which the sharp dis- tinction between catachresis and metonymy is grounded is consider- ably eroded. The only thing we can say is that the very possibility of a hegemonic relationship depends on this erosion, on keeping an unstable equilibrium between heterogeneity and contiguity, between catachresis and metonymy--an equilibrium whose conditions of ex- tinction would be either a heterogeneity without common measure be- tween the elements, or a contiguity that becomes total and, thus, ab- sorbs within an implicitly assumed space the contiguous positions as internal differences. 12 (These two conditions of extinction of the hege-
monic link are, in fact, only one and the same: in order to be radically heterogeneous, two elements require a common ground out of which their heterogeneity can be thought. )
On the other hand, however, all hegemony tries to retotalize and to make as necessary as possible the contingent links on which its articulating power is based. In this sense, it tends to metaphorical to- talization. This is what gives it its dimension of power. It is a power, however, that maintains the traces of its contingency, and is, in that sense, essentially metonymic. Hegemony is always suspended be- tween two impossible poles: the first, that there would be no dis- placement, that contiguity becomes mere contiguity, and that all tropological movement ceases--this would be the case of what Gramsci called the "corporative class"; the second, that the meta- phorical totalization becomes complete and that purely analogical re- lations fully saturate the social space--in which case we would have the "universal class" of the classical emancipatory discourse. Both poles are excluded by the hegemonic relation. It is only on the traces of (contingent) contiguity contaminating all analogy that a hegemon- ic relation can emerge.
III
I will now attempt to illustrate these propositions with a historical example showing an extreme attempt at metaphorical totalization, whose very failure shows the space in which the undecidable logic of hegemony operates. I am referring to the work of Georges Sorel.
Sorel's work is a product of that period of socialist thought that has been called, following the characterization of Thomas Masaryk, "the crisis of Marxism. " The increasing gap between the classical Marxist dogma, as codified in the Anti-Du? hring, and the actual turn of events opened a theoretical vacuum that various intellectual projects attempt- ed to fill.
Two themes quite frequent at that time--the impossibility of unify- ing historical events by conceptual means, the grounding of historical action in conviction and will--are certainly present in Sorel, but he is going to give them a new twist and to invest them with a new meaning, as he is going to see them from the viewpoint of a far more radical his- torical possibility. Let us consider the following passage from the Re? flexions sur la violence, which shows the deep gap separating Sorel's socialism from that of most of his contemporaries:
The Politics of Rhetoric 239
240 Ernesto Laclau
Dans une socie? te? aussi enfie? vre? e par la passion du succe`s a` obtenir dans la concurrence, tous les acteurs marchent droit devant eux comme de ve? ritables automates, sans se pre? occuper des grandes ide? es des socio- logues; ils sont soumis a` des forces tre`s simples et nul d'entre eux ne songe a` se soustraire aux conditions de son e? tat. C'est alors seulement que le de? veloppement du capitalisme se poursuit avec cette rigueur qui avait tant frappe? Marx et qui lui semblait comparable a` celle d'une loi naturelle. Si, au contraire, les bourgeois, e? gare? s par les blagues des pre? di- cateurs de morale ou de sociologie, reviennent a` un ide? al de me? diocrite? conservatrice, cherchent a` corriger les abus de l'e? conomie et veulent rompre avec la barbarie de leurs anciens, alors une partie des forces qui devaient produire la tendance du capitalisme est employe? e a` l'enrayer, du hasard s'introduit et l'avenir du monde est comple`tement inde? termine? .
Cette inde? termination augmente encore si le prole? tariat se convertit a` la paix sociale en me^me temps que ses mai^tres;--ou me^me simplement s'il conside`re toutes choses sous un aspect corporatif;--tandis que le socialisme donne a` toutes les contestations e? conomiques une couleur ge? ne? rale et re? volutionnaire. 13
Let us follow this argument closely. If the objective logic of historical change that Marx had presented depends for its full development on the bourgeoisie not being dominated by the ideal of a "me? diocrite? conservatrice"--because in the latter case "du hasard s'introduit et l'avenir du monde est comple`tement inde? termine? "--everything turns on whether that ideal will prevail or not. That prevalence, however, cannot be the result of identifiable objective economic processes, given that the very possibility of those processes taking place depends, for Sorel, on the absence of the "me? diocrite? conservatrice. " Here we find the cornerstone of Sorel's thought in its mature stage: social processes do not involve only displacements in the relationship of forces between classes, because a more radical and constitutive possibility is always haunting society: the dissolution of the social fabric and the implosion of society as a totality. So, society is not only suffering domination and exploitation: it is also threatened with decadence, with the only too real possibility of its radical not being. Let us see how in Sorel this dis- tinctive possibility opens the way to a new and peculiar logic in the re- lation between groups. There are two capital moments in this logic.
The first, is that the opposition that dominates Sorel's vision of the social is not primarily the one between bourgeoisie and proletariat but, rather, that between decadence and full realization of society. If the
proletariat as a social force receives historical priority, it is because it is seen as the main instrument in confronting decadence. But--and this is a crucial point--it is not the actual victory of the proletariat against the bourgeoisie that will bring about "grandeur" and will arrest deca- dence, but the very fact of the open confrontation between the two groups. Without confrontation there is no identity; social identities re- quire conflict for their constitution. This explains Sorel's critique of democracy. If social identities require open conflict for their full consti- tution, any attempt to dilute, reduce, or even regulate that conflict can only be an instrument of decadence and corruption of those identities. Democratic participation in public institutions is the sure road leading to such corruption. So, Sorel had to see in Marxism not a scientific doctrine explaining the objective laws of capitalism, but a finalistic ide- ology of the proletariat, grounded in class struggle. Social relations, left to themselves, are just "me? lange. " Only the will of determined so- cial forces gives a consistent shape to social relations, and the determi- nation of that will depends on the violent confrontation with other wills.
But, second, if the historical justification of the action of the prole- tariat is given by its being the only force capable of opposing the deca- dence of civilization, that justification is indifferent to the contents of the proletarian program and depends entirely on the contingent ability of those contents to bring about an effect that is external to them- selves. There is no ethical justification that is intrinsic to socialism. This has two capital consequences. The first, that all social identity or social demand is constitutively split. It is, on the one hand, a concrete demand; but, on the other hand, it can also be the bearer, the incarna- tion of social "grandeur" as opposed to decadence. "Grandeur" and "decadence" do not have intrinsic contents of their own, but are the empty signifiers of a fullness of society (or its opposite, its corruption or nonbeing) that can be actualized by the most different social forces. So--and this is the second consequence--it is enough that the working class shows itself as a limited historical actor, closed in its corporative demands and incapable of incarnating the will to fullness of society, for its claims to lose all legitimacy. And the political trajectory of Sorel is a living example of the contingency of this relation between working- class demands and "grandeur": he passed from being a theoretician of revolutionary syndicalism to ally himself with a fraction of the mon- archist movement, and ended his career by being close to the Third International. The diffusion of Sorelian themes in antagonistic social
The Politics of Rhetoric 241
242 Ernesto Laclau
movements, from bolshevism to fascism, is an even more telling ex- ample of the ambiguous possibilities that his de? marche was opening. For Sorel, the action (violence) is increasingly separated from its own contents and exclusively judged by the effect the latter have on the identity of the actors.
Let us now translate these reflections on Sorel into our tropological argument. Any attempt by the proletariat to constitute its subjectivity through a variety of loosely related subject positions can only lead, ac- cording to Sorel, to corporative integration and to decadence; so all metonymic variation had to be eliminated. In that case, how to aggre- gate working-class struggles in such a way that the proletarian identity is maintained and reinforced? Through an education of the will ground- ed in the myth of the general strike. Each action of the workers-- whether a strike, a demonstration, or a factory occupation--should be seen, not in its own specificity and particular objectives, but as one more event in the formation of the revolutionary will. That is, these ac- tions are all analogous from the point of view of their ultimate deep aims and are, as a result, in a relation of metaphorical substitution with each other. Their mutual relations--as that between flies and summer in Proust's text--are necessary. The drawback of this vision is that, in that case, the myth that unifies the struggles beyond all speci- ficity cannot be specific either. The reduction of all specificity to the repetition of something analogous can only be the metaphor of meta- phoricity as such. We know what this involves: the interruption of any hegemonic operation. The metaphor of metaphoricity can only be a zero that is in no tropological relation with a one, or--at most--a zero that is in a catechretical relation with only one position. Only at that price can revolutionary closure be achieved. And this is precisely what Sorel attempts to achieve by making the general strike totally heteroge- neous with the empirical world of limited and partial struggles. The general strike is presented as a myth and not a utopia--it has lost all the detailed descriptive features of the latter; it has no particular objec- tives; it is merely an empty image that galvanizes the consciousness of the masses; it is exhausted in this last function without possibly cor- responding to any actual historical event. It is a radical nonevent that is, paradoxically, the condition of all events if there is going to be a "grandeur" in society.
In that case, why revolutionary general strike rather than anything else? Is there any ground to think that the general strike is the (necessary) catachresis of that radical nonevent which brings about "grandeur"?
Sorel cannot answer this question, and the oscillations of his political career are a clear indication that the question is unanswerable. The relationship between "grandeur" and general strike is a hegemonic in- carnation, which involves all metaphoric aggregation being ultimately grounded in (reversible) metonymic displacements. The attempt to ground the revolutionary will in a metaphoric totalization that would avoid the particularism of hegemonic variations ends in failure. As Plato knew, perhaps better than Sorel--only protracted metonymic displace- ments between Athens and Siracusa can give some hope that the king agrees to become a philosopher.
Perhaps we could still make this point in a slightly different way: it is only through the pure, irreducible event that consists in a contingent displacement not retrievable by any metaphoric reaggregation that we can have a history, in the sense of both Geschichte and Historie. It is because there is hegemony (and metonymy) that there is history. Couldn't some deconstructive strategies, such as iteration, be seen as attempts at introducing metonymy at the heart of metaphor, displace- ment at the heart of analogy? Genette (following Blanchot) tries to show, in his analysis of Proust, how the latter moved from a structure of his novel conceived as a succession of poetic instants, of punctual moments, to a conception of his whole narrative in which anamnesis is inseparable from--it is actually governed by--the whole process of narration. As he points out:
Sans me? taphore, dit (a` peu pre`s) Proust, pas de ve? ritables souvenirs; nous ajoutons pour lui (et pour tous): Sans me? tonymie, pas d'enchai^ne- ment de souvenirs, pas d'histoire, pas de roman. Car c'est la me? taphore qui retrouve le Temps perdu, mais c'est la me? tonymie qui le ranime, et le remets en marche: qui le rend a` lui-me^me et a` sa ve? ritable "essence", qui est sa propre fuite et sa propre Recherche. Ici donc, ici seulement--par la me? taphore mais dans le me? tonymie--, ici commence le Re? cit. 14
Perhaps this is, exactly, the intellectual displacement leading from Sorel to Gramsci. Whereas in the first the analogizing movement of the metaphor of metaphoricity led to a repetition that tried to eliminate the possibility of any proper event, Gramsci's notion of war of posi- tion, of a narrative-political displacement governed by a logic of pure events which always transcend any preconstituted identity, announces the beginning of a new vision of historicity--one governed by the in- eradicable tension between metonymy (or synecdoche) and metaphor.
The Politics of Rhetoric 243
244 Ernesto Laclau
IV
We now arrive at a decisive point in our argument on hegemony. If hegemony means the representation, by a particular social sector, of an impossible totality with which it is incommensurable, then it is enough that we make the space of tropological substitutions fully visible, to enable the hegemonic logic to operate freely. If the fullness of society is unachievable, the attempts at reaching it will necessarily fail, although they will be able, in the search for that impossible object, to solve a variety of partial problems; that is, the particularism of the struggles, which was systematically demoted in Sorel's analysis, now becomes central. With this, the metonymic game occupies center stage, and politics--which was for Sorel the nemesis of proletarian action--takes the upper hand.
All this becomes more visible if we compare Sorel's discursive de? - marche with other socialist discourses of the time, which oriented themselves in the opposite direction. Let us clarify, however, an impor- tant point before engaging in that comparison. Both metaphor and metonymy are tropological movements, that is, forms of condensation and displacement whose effects are achieved on the basis of going be- yond literal meaning. Now, from this point of view, classical Marxian discourse presented itself as the zero degree of tropology, as a scientific discourse describing the necessary laws of history, which did not need to go beyond the literality of their formulation in order to achieve the totalizing effects that they postulated. That this ideal of scientificity points to an impossible task, and that whatever totalizing effects Marxian discourse could have can only be achieved through putting into operation a whole arsenal of tropological movements, is well known; but the important point is that as an ideal that governed its own discursivity, literality is fully present in it and produces a whole set of concealing effects. Sorel had ceased to believe in objective, neces- sary laws of history and wanted to substitute them with an artificial necessity grounded in the power of the will; so he had, as we have seen, to have full recourse to the principle of analogy--which in a literal dis- course of necessary, objective laws would have no incidence--and to in- stall himself, fully conscious of the fact, in the terrain of metaphor. But, as we have also seen, metaphorical necessity was decisively contami- nated by metonymic contingency. What were, in that case, the politico- discursive and strategic effects of accepting as inevitable the metonymic terrain?
Let us consider the discussions in Russian social democracy at the turn of the last century. The general view was that Russia was ripe for a bourgeois-democratic revolution in which the bourgeoisie, as in all major revolutions of the West, would carry on the tasks of sweeping away the remainders of feudalism and creating a new state along liberal- democratic lines. The drawback was that the Russian bourgeoisie ar- rived in the historical arena too late and was weak and incapable of carrying out its political tasks. The need for a democratic revolution was, however, still there. This led to the conclusion--drawn at least by some sections of the social democrats--that in that case those tasks had to be taken up by some other social sector that was not its natural bearer:--in this case, the working class. This relationship by which a sector takes up tasks that are not its own is what the Russian social democrats called hegemony. So we see how the political steps that this analysis took led in the opposite direction to Sorel's. Whereas the latter tried to close the working class around its own natural tasks through metaphoric totalizations, here we find the opening of a field of meto- nymic displacements in the relations between tasks and agents, an un- decided terrain of contingent articulations in which the principle of contiguity prevails over that of analogy. It was only the contingent pe- culiarity of the Russian situation--the presence of a weak bourgeoisie and a strong working class--that was at the root of the working-class leadership in the democratic revolution.
This complicated dialectics between analogy and contiguity can be seen to expand in a plurality of directions. First, as the nontropological succession of programmed stages is interrupted, a space of logical inde- terminacy arises: "Tsarism, having entered into complete contradiction with the demands of Russia's social development, continued to exist thanks to the power of its organisation, the political nullity of the Russian bourgeoisie and its growing fear of the proletariat. "15
Second, this indeterminacy is the source of pure relations of con- tiguity that break the possibility of totalizations in terms of either syntagmatically retrievable differences or metaphorically "necessary" aggregations:
Russian capitalism did not develop from artisanal trade via the manu- facturing workshop to the factory for the reason that European capital, first in the form of trade capital and later in the form of financial and industrial capital, flooded the country at a time when most Russian artisanal trade had not yet separated itself from agriculture. Hence the
The Politics of Rhetoric 245
246 Ernesto Laclau
appearance in Russia of modern capitalist industry in a completely primitive economic environment: for instance, a huge Belgian or American industrial plant surrounded by dirty roads and villages built of straw and wood, which burn down every year, etc. The most primitive begin- nings and the most modern European endings. (339)
This gap, which interrupts any nontropological succession of necessary stages, but also any metaphoric aggregation of events around a pre- given necessary point, gives proletarian identity in Russia an open character in which contingent displacements, pure events, assume a constitutive role that no a prioristic logic can govern:
I remember an old friend, Korotkov, a cabinetmaker from Nikolayev, who wrote a song back in 1897. It was called The Proletarians' March and it began with the words: "We are the alpha and the omega, we are the beginning and the end . . . " And that's the plain truth. The first letter is there and so is the last, but all the middle of the alphabet is missing. Hence the absence of conservative traditions, the absence of castes within the proletariat, hence its revolutionary freshness, hence, as well as for other reasons, the October Revolution and the first workers' government in the world. But hence also the illiteracy, the absence of organisational know- how, the lack of system, of cultural and technical education . . . (340)
And then the inevitable consequence:
From the viewpoint of that spurious Marxism which nourishes itself on historical cliche? s and formal analogies . . . the slogan of the seizure of power by the Russian working class was bound to appear as a mon- strous denial of Marxism. . . . What then is the real substance of the problem? Russia's incontestably and incontrovertibly backward devel- opment, under the pressure of the higher culture of the West, leads not to a simple repetition of the Western European historical process but to a set of fundamentally new features which requires independent study. . . . Where there are no "special features", there is no history, but only a sort of pseudo-materialistic geometry. Instead of studying the living and changing matter of economic development it is enough to notice a few outward symptoms and adapt them to a few ready-made cliche? s. (339)
Could it be clearer? Historicity is identified with "special features" un- assimilable to any form of repetition. History is a field of contingent dis- placements that are not retrievable by any of the (analogical) figures of the same.
Of course, this field of contingent variations can be more or less ex- tended, depending on the width of the area in which the literal still pre- vails and arrests the tropological movement. Now what happened in socialist discourses, like those we are considering, was that what we have described as a tropological movement expanded ever more and covered wider and wider sections of political life. Let us consider a concept such as "combined and uneven development. " It was intro- duced to refer to the experience of social struggles in Third World countries, in which--even more than in the case of Russia--a non- orthodox combination of developments that should have correspond- ed to successive stages makes more contingent and risky hegemonic interventions possible. In the 1930s, Trotsky drew the inevitable con- clusion: an uneven development is the terrain of all social and political struggles in our time. The only thing we have is an unlimited tropologi- cal movement that is the very terrain in which society constitutes itself. And we see why metonymy is, in some sense, more "primordial" than metaphor (or, as in other of de Man's analyses, why allegory takes precedence over the symbol): because in a situation of radical contin- gency no criterion of analogy is stable; it is always governed by chang- ing relations of contiguity that no metaphorical totalization can con- trol. Metaphor--and analogy--is at most a "superstructural" effect of a partial stabilization in relations of contiguity that are not submitted to any literal principle of a priori determination.
This process of general rhetorization only takes place as far as none of the conditions in which each of the tropoi would become what it lit- erally claims to be can be met.
If metonymy were just a metonymy, its ground should be a contiguity that is not contaminated by analogy, and in that case the literal separations within a given discursive space would be fully in control of the limits of the metonymic movement. If analogy dominated unchallenged, a full totalization would have taken place that would make analogy collapse into identity--and the tropo- logical movement would cease. If synecdoche was actually able to sub- stitute the whole for the part, this would mean that the whole could have been apprehended independently of the part. If catachresis could be grounded in a tropological movement that started from total het- erogeneity, it could only take place insofar as the distinction between the homogeneous and the heterogeneous would be established with entire precision. It is as if in some way the conditions for a rhetoric whose tropological movements are going to occupy the terrain of a ground that is not itself grounded are to be found in the impossibility
The Politics of Rhetoric 247
248 Ernesto Laclau
of taking the definitions of each of the tropoi at face value, and in the need to stress the logics by which each tends to fade into the other. The same for hegemony: the conditions of its full success are the same as the conditions of its extinction.
This can be shown by a couple of historical examples. The first con- cerns Italy. At the end of World War II there was a confrontation of tendencies, within the Italian Communist Party, about the right strate- gy to be followed in the new democratic environment. There were two positions: one that asserted that the Communist Party, being the party of the working class and the latter being an enclave in the industrial north, had to limit its efforts mainly to creating forms of representa- tion for that enclave; the second, more Gramscian, maintained that the party had to build up its hegemony by spreading its activities to a va- riety of areas, the agrarian Mezzoggiorno included. How was this pos- sible, given the particularistic social and geographical location of the working class? Simply, by making the party and the unions the rallying point of a variety of democratic initiatives in a country moving away from fascist dictatorship. The democratic initiatives postulated by this approach were entirely contingent--their success was not guaranteed by any logic of history--and depending, thus, on the construction of a collective will; but, in a way different from the Sorelian will, they were not aiming for the reinforcement of a purely proletarian identity. They tended, instead, to the creation of a multifarious democratic identity, always spreading beyond itself in directions only graspable through a contingent narration. Togliatti wrote in 1957:
A class may lead society insofar as it imposes its own rule, and to this end the force of arms can also be used. It becomes a national class, how- ever, only insofar as it solves the problems of the whole of society. . . . The proletariat becomes a national class insofar as it takes on these problems as its own and thence comes to know, by the process of chang- ing it, the whole reality of national life. In this way it produces the con- ditions of its own political rule, and the road to becoming an effective ruling class is opened. . . .
We have to spread the activity of an organised vanguard over the whole area of society, into all aspects of national life. This activity must not be reduced to preaching propaganda, to phrase-making or clever tactics, but must stick closely to the conditions of collective life and give, therefore, a foundation, real possibilities and prospects to the movement of the popular masses. . . . Our struggle for the unity of popu-
lar, democratic forces is, therefore, not imposed by tactical skills, but is an historical requirement, both to maintain conquests already achieved, to defend and safeguard democracy, and to develop it. 16
Here we have a tropological space in which each of the figures tends to fade into the other. The different struggles and democratic initiatives are not united to each other by necessary links--so we have metonymi- cal relations of contiguity. The hegemonic operation tries, however, to make the condensation of these struggles as strong as possible--so the metonymics fade into metaphoric totalization. The hegemonic relation is synecdochal, as far as a particular sector in society--the working- class party, in this case--tends to represent a whole exceeding it. As, however, this whole lacks any precisely defined limits, we are dealing with an impure synecdoche: it consists in the undecidable movement between a part attempting to incarnate an undefinable whole, and a whole that can only be named through its alienation in one of its parts. Finally, the heterogeneity can only be a relative one--with the result that the line separating catachresis from metonymy becomes also un- decidable. I think that all the main categories of Gramscian theory-- war of position, collective will, organic intellectuals, integral state, his- torical bloc, hegemony--could be read rhetorically: as circumscribing the space of tropological movement that brings about new strategic flexibility in political analysis.
A comparable discussion took place in the South African context, in the years before the end of apartheid, between the so-called workers and populist tendencies within the liberation movement. The first ten- dency, based largely in the trade-union movement, asserted the need for an immediate socialist transformation constituted around a work- ing class whose protagonist role resulted from its structural centrality in capitalist society. Contingent and particularistic displacements in the search for alliances were reduced to a minimum. The populist camp, on the contrary, based on the principles of the Freedom Charter, made contingent hegemonic articulations the cornerstone of its strategy. As David Howarth and Aletta Norval have asserted,
the Charterists have suggested that the formation of political conscious- ness cannot be exclusively attributed to factory floor experiences, or de- rived simply from the agent's location in the relations of production, but occurs in a much wider discursive context. Instead of positing an ab- stract working class persona analytically separable from the complex set of discourses in which the worker is situated, the working class is
The Politics of Rhetoric 249
250 Ernesto Laclau
regarded as a real social force forming an essential component of the na- tionally oppressed people. . . . In this sense, the working class as a real social force engaged in struggle is always to some extent marked by struggles, identities and discourses which cannot simply be reduced to its position in the relations of production; their leadership will only be attained by means of actively introducing socialist discourses into the struggle for national liberation. 17
Although Howarth and Norval have pointed out the limited character of the hegemonic opening that the populist camp was postulating, it is clear that the strategic confrontation was taking place along lines simi- lar to those that we have so far discussed: in one case a principle of analogy by which the South African working class repeats a worker's identity dictated by the abstract analysis of the capitalist relations of production; in the other case, a succession of merely contiguous articu- lations that governs a predominantly contingent and contextualized narrative.
This political argument could, obviously, be extended in a variety of directions beyond the socialist discourses that we have considered here. There is one, however, that I would like to stress. I see the history of democracy as divided by one fundamental cleavage. On the one hand, we have democracy as the attempt to construct the people as one, a homogeneous social actor opposed either to "power" or to an external enemy--or to a combination of both. This is the Jacobin con- ception of democracy, with its concomitant ideal of a transparent com- munity unified--if necessary--by terror. This is the tradition that runs, with very analogous structural features, from Robespierre to Pol Pot. The discourses around which this democratic ideal is constructed are, obviously, predominantly metaphoric--although, for reasons previ- ously mentioned, they cannot conceal their metonymic foundations. On the other hand, we have democracy as respect of difference, as shown, for instance, in multiculturalism or in the new pluralism asso- ciated with contemporary social movements. Here we have discourses that are predominantly metonymic, for although--given the impos- sibility of a pure differential, nontropological closure--some effect of metaphoric aggregation is inevitable, it will be an aggregation that always keeps the traces of its own contingency and incompleteness visible. Within this basic polarity there are, obviously, all kinds of pos- sible intermediate combinations that we can start exploring through the variety of tropoi to be found in classical rhetoric.
Paul de Man's main contribution to this task lies not in anything that he had to say about politics--something his untimely death pre- vented him from doing--but rather, in two main accomplishments. The first is to have extended the field of rhetoric--or, rather, of rhetoricity-- to the ensemble of language, to have made it a constitutive dimension of language as such. The second is to have deconstructed the dominant tropoi of the Romantic tradition--such as symbol and metaphor-- showing the contingent infrastructure of more humble tropoi on which any totalizing effect is grounded. I have attempted in this essay to show the potential importance of both accomplishments for the elaboration of a theory of hegemony.
NOTES
A preliminary version of this essay was presented at the "Culture and Materiality" conference that took place at the University of California, Davis, on April 23-25, 1998. It was also discussed in the seminar on Ideology and Discourse Analysis, University of Essex, a month later. I want to thank those whose commentaries led me to introduce precisions in my text and, in some cases, partial reformulations of the argument: at Davis, Jacques Derrida, Fredrick Jameson, J. Hillis Miller, and Andrzej Warminski; at Essex, David Howarth and Aletta Norval.
1. Stefano Rosso, "An Interview with Paul de Man," in Paul de Man, The Resistance to Theory (Minneapolis: University of Minnesota Press, 1993), 121.
2. Paul de Man, "Pascal's Allegory of Persuasion," in Aesthetic Ideology (Minneapolis: University of Minnesota Press, 1996), 51-69. Subsequent page refer- ences are given in the text.
3. For the full development of this argument, see my Emancipations(s) (London: Verso, 1996), especially the essay "Why Do Empty Signifiers Matter to Politics? " 36-46.
4. Cf. the disagreement between de Man and Louis Maris, as presented by de Man in "Pascal's Allegory of Persuasion," 60.
5. What is important is to realize that these retotalizations do not operate through a simple and retrievable negation. As de Man asserts: "What is here called, for lack of a better term, a rupture or a disjunction is not to be thought of as a negation, however tragic it may be. Negation, in a mind as resilient as Pascal's, is always susceptible of being reinscribed in a system of intelligibility. . . . It is possible to find, in the terminology of rhetoric, terms that come close to designating such disruptions (e. g. , parabasis or anacoluthon), which designate the interruption of a semantic continuum in a manner that lies beyond the power of reintegration" (ibid. , p. 61). But the very fact that there are tropoi which make describable that which is beyond the power of reintegration clearly shows that we are not dealing with a simple collapse of the system but, rather, with an orderly drifting away from what would have otherwise been the conditions of its full closure. It is in the field of this drifting away that the hegemonic logics operate.
6. Paul de Man, "Reading (Proust)," in Allegories of Reading Figural Language
The Politics of Rhetoric 251
252 Ernesto Laclau
in Rousseau, Nietzsche, Rilke and Proust (New Haven: Yale University Press, 1979), 57-58.
7. Paul de Man, Blindness and Insight: Essays in the Rhetoric of Contempor- ary Criticism (Minneapolis: University of Minnesota Press, 1983), 284.
8. Ge? rard Genette, "Me? tonymie chez Proust," in Figures III (Paris, E? ditions du Seuil, 1972), 41-43. Paul de Man finds the use by Genette of the category of diegetic metaphor limited as far as Proust is concerned. Both, however, agree in privileging the metonymic transitions in Proust's text.
9. Ernesto Laclau and Chantal Mouffe, Hegemony and Socialist Strategy: Towards a Radical Democratic Politics (London: Verso, 1985).
10. I thank J. Hillis Miller for having called my attention to the need for elabo- rating more the distinction between catachresis and metonymy--a distinction that, as will be seen, is crucial for my analysis.
11. Pierre Fontanier, Les figures du discours (Paris: Flammarion, 1968), 213. "Generally speaking there is catachresis when a sign already linked to a first idea is also linked to a new idea which itself does not have or no longer has a literal mean- ing in language. It is, as a result, any trope whose use is forced and necessary, any trope from which a purely extended meaning is derived; this literal meaning of sec- ond origin, mediating between the original literal meaning and the figurative mean- ing, but which by its own nature is closer to the first than to the second, even though it might have been figurative in the beginning. "
12. With this, of course, the tropological movement would come to an end.
13. Georges Sorel, Re? flexions sur la violence (Paris: E? ditions du Seuil, 1990), 77. "In a society so enfevered by the passion of competitive success, all the actors march straight ahead like true automatons, unconcerned with the great ideas of the sociologists, and none of them considers extricating himself from that condi- tion. It is only then that the development of capitalism is pursued with that rigor which so struck Marx and which seemed to him comparable to a natural law. If, on the contrary, the bourgeois, diverted by the jokes of moral or sociological preachers, return to the ideal of conservative mediocrity, seek to correct the abuses of the economy, and wish to break with the barbarity of their ancestors, then a part of the forces which should produce the dynamic of capitalism are employed in checking it, chance is introduced, and the future of the world becomes completely indeterminate.
"This indeterminacy further increases if the proletariat is converted to social peace at the same time as their masters--or even simply that it considers everything from a corporative perspective--while socialism gives to all economic struggles a general and revolutionary character. "
14. Genette, "Me? tonymie chez Proust," 63. "Without metaphor, says Proust (more or less), there are no true memories; we can add for him (and for everybody): without metonymy, no linkage of memories, no history, no novel. Because it is metaphor that recovers lost Time, which reanimates and sets it in motion: which re- turns it to itself and its true 'essence,' which is in its own flight and its own Search. So here, here alone--through metaphor but within metonymy--Narration begins. "
15. Leon Trotsky, 1905 (London: Allen Lane, 1971), 328. Subsequent page ref- erences are given in the text.
16. Palmiro Togliatti, On Gramsci and Other Writings (London: Lawrence and Wishart, 1979), 157-59.
17. David Howarth and Aletta Norval, "Subjectivity and Strategy in South African Resistance Politics: Prospects for a New Imaginary," Essex Papers in Politics and Government, Department of Government, University of Essex (May 1992): 9.
The Politics of Rhetoric 253
254
How Can I Deny That These Hands and This Body Are Mine?
Judith Butler
I remember a sleepless night last year when I came into my living room and turned on the television set to discover that C-Span was offering a special session on feminist topics, and that the historian Elizabeth Fox- Genovese was making clear why she thought women's studies had con- tinuing relevance, and why she opposed certain radical strains in femi- nist thinking. Among those positions she most disliked she included the feminist view that no stable distinction between the sexes could be drawn or known, a view that suggests that the difference between the sexes is itself culturally variable or, worse, discursively fabricated, as if it is all a matter of language. Of course, this did not help my project of falling asleep, and I became aware of being, as it were, a sleepless body in the world accused, at least obliquely, with having made the body less rather than more relevant. Indeed, I was not altogether sure that the bad dream from which I had awoken some hours earlier was not in some sense being further played out on the screen. Was I waking or was I dreaming? After all, it was no doubt the persecutory dimension of paranoia that hounded me from the bed. Was it still paranoia to think that she was talking about me, and was there really any way to know? If it was me, then how would I know that I am the one to whom she refers?
I relate this incident to you here today not only because it fore- shadows the Cartesian dilemmas with which I will be preoccupied in the following paper, and not because I propose to answer the question of whether sexual difference is only produced in language. I will, for the moment, leave the question of sexual difference to be returned to another time. 1 The problem I do propose to address emerges every
time we try to describe the kind of action that language exercises on the body or, indeed, in the production or maintenance of bodies. And we do tend to describe language as actively producing or crafting a body every time we use, implicitly or explicitly, the language of discur- sive construction.
In the consideration of Descartes's Meditations that follows, I pro- pose to ask whether the way in which Descartes posits the irreality of his own body does not allegorize a more general problem of positing that is to be found in various forms of constructivism and various criti- cal rejoinders to a constructivism that is sometimes less well under- stood than it ought to be. The name of this paper that I have already begun, but not yet begun, is: "How Can I Deny That These Hands and This Body Are Mine? " These are, of course, Descartes's words, but they could be ours or, indeed, mine, given the dilemmas posed by contem- porary constructivism.
The language of discursive construction takes various forms in con- temporary scholarship and sometimes it does seem as if the body is cre- ated ex nihilo from the resources of discourse. To claim, for instance, that the body is fabricated in discourse is not only to figure discourse as a fabricating kind of activity, but to sidestep the important questions of "in what way" and "to what extent. " To say that the line between the sexes, for instance, must be drawn, and must be drawable, is to concede that at some level the stability of the distinction depends on a line being drawn. But to say that we must be able to draw a line in order to stabilize the distinction between the sexes may simply mean that we must first grasp this distinction in a way that allows us then to draw the line, and the drawing of the line confirms a distinction that is somehow already at hand. But it may mean, conversely, that there are certain conventions that govern how and where the line ought or ought not to be drawn, and that these conventions, as conventions, change through time, and produce a sense of anxiety and of unknow- ingness precisely at the moment in which we are compelled to draw a line in reference to the sexes. The line then lets us know what will and will not qualify as "sex"; the line works as a regulatory ideal, in Foucault's sense, or a normative criterion that permits and controls the appearance and knowability of sex. Then the question, which is not easily settled, becomes: do the conventions that demarcate sexual dif- ference determine in part what we "see" and "comprehend" as sexual difference? It is, you might surmise, not a large leap from this claim to the notion that sexual difference is fabricated in language. But I think
How Can I Deny . . . ? 255
256 Judith Butler
that we may need to look more carefully before either championing or reviling this conclusion.
The language of construction risks a certain form of linguisticism, the assumption that what is constructed by language is therefore also language, that the object of linguistic construction is nothing other than language itself. Moreover, the action of this construction is con- veyed through verbal expressions that sometimes imply a simple and unilateral creation at work.
Language is said to fabricate or to figure the body, to produce or construct it, to constitute or to make it. Thus, language is said to act, which involves a tropological understanding of language as perform- ing and performative. There is, of course, something quite scandalous involved in the strong version of construction that is sometimes at work when, for instance, the doctrine of construction implies that the body is not only made by language, but made of language, or that the body is somehow reducible to the linguistic coordinates by which it is identified and identifiable, as if there is no nonlinguistic stuff at issue. The result is not only an ontological realm understood as so many ef- fects of linguistic monism, but the tropological functioning of language as action becomes strangely literalized in the description of what it does, and how it does what it does. And though de Man often argued that the tropological dimension of discourse works against the perfor- mative, it seems here that we see, as I believe we do in de Man's discus- sion of Nietzsche, the literalization of the trope of performativity.
I want to suggest another way of approaching this question that re- fuses the reduction of linguistic construction to linguistic monism, and which calls into question the figure of language acting unilaterally and unequivocally on the object of construction. It may be that the very term construction no longer makes sense in this context, that the term deconstruction is better suited to what I propose to describe, but I con- fess to not really caring about how or whether these terms are stabilized in relation to each other or, indeed, in relation to me. My concerns are of another order, perhaps in the very tension that emerges as the prob- lem of discursive construction comes into tension with deconstruction.
For my purposes, I think it must be possible to claim that the body is not known or identifiable apart from the linguistic coordinates that establish the boundaries of the body without thereby claiming that the body is nothing other than the language by which it is known. This last claim seeks to make the body an ontological effect of the language that governs its knowability. But this view fails to note the incommensura-
bility between the two domains, an incommensurability that is not pre- cisely an opposition. Although one might accept the proposition that the body is only knowable through language, that the body is given through language, it is never fully given in that way, and to say that it is given partially can only be understood if we also acknowledge that it is given, when it is given, in parts; it is, as it were, given and withheld at the same time, and language might be said to perform both of these operations. Although the body depends on language to be known, the body also exceeds every possible linguistic effort of capture. It would be tempting to conclude that this means that the body exists outside of language, that it has an ontology separable from any linguistic one, and that we might be able to describe this separable ontology. But this is where I would hesitate, perhaps permanently, for as we begin that description of what is outside of language, the chiasm reappears: we have already contaminated, though not contained, the very body we seek to establish in its ontological purity. The body escapes its linguistic grasp, but so too does it escape the subsequent effort to determine on- tologically that very escape. The very description of the extralinguistic body allegorizes the problem of the chiasmic relation between language and body and so fails to supply the distinction it seeks to articulate.
To say that the body is figured chiasmically is to say that the follow- ing logical relations hold simultaneously: the body is given through language but is not, for that reason, reducible to language. The lan- guage through which the body emerges helps to form and establish that body in its knowability, but the language that forms the body does not fully or exclusively form it. Indeed, the movement of language that appears to create what it names, its operation as a seamless perfor- mative of the illocutionary persuasion, covers over or dissimulates the substitution, the trope, by which language appears as transitive act, that is, by which language is mobilized as a performative that simulta- neously does what it says. If language acts on the body in some way, if we want to speak, for instance, of a bodily inscription, as so much cul- tural theory does, it might be worth considering whether language lit- erally acts on a body, and whether that body is an exterior surface for such action, or whether these are figures that we mobilize when we seek to establish the efficacy of language.
This leads to a converse problem, namely, the case in which lan- guage attempts to deny its own implication in the body, in which the case for the radical disembodiment of the soul is made within language. Here it is a question of the way in which the body emerges in the very
How Can I Deny . . . ? 257
258 Judith Butler
language that seeks to deny it, which suggests that no operation of lan- guage can fully separate itself from the operation of the body.
Language itself cannot proceed without positing the body, and when it tries to proceed as if the body were not essential to its own op- eration, figures of the body reappear in spectral and partial form with- in the very language that seeks to perform their denial. Thus, language cannot escape the way in which it is implicated in bodily life, and when it attempts such an escape, the body returns in the form of spectral figures whose semantic implications undermine the explicit claims of disembodiment made within language itself. Thus, just as the effort to determine the body linguistically fails to grasp what it names, so the ef- fort to establish that failure as definitive is undermined by the figural persistence of the body.
The way in which I propose to show this chiasmic relation is through a reconsideration of the opening Meditations of Descartes, the ones in which he seeks to bring the reality of the body into question. Descartes seeks to know whether he can deny the reality of his own body and, in particular, the reality of his limbs. Suspended and inscrutable limbs reemerge in de Man's essay "Phenomenality and Materiality in Kant" in ways that suggest a metonymic relation to the problem that Descartes poses. For de Man, the body within the third Critique is understood, if we can use that word, as prior to figuration and cognition. In Descartes, it emerges as a particular kind of figure, one that suspends the ontologi- cal status of the term, and this raises the question of any absolute sepa- rability between materiality and figuration, a distinction that de Man on some occasions tries to make as absolutely as possible. I hope to make this chiasmic relation between those two texts clear toward the end of my remarks.
