veloppement du capitalisme se
poursuit
avec cette rigueur qui avait tant frappe?
Paul-de-Man-Material-Events
If such a closure were achievable, no hegemonic event could be possible and the political, far from being an ontological dimension of the social--an "existential" of the social-- would just be an ontic dimension of the latter.
Second, however, the hegemonic suture has to produce a retotalizing effect, without which
no hegemonic articulation would be possible either. But, third, this re- totalization cannot have the character of a dialectical reintegration. It has, on the contrary, to maintain alive and visible the original and con- stitutive heterogeneity from which the hegemonic articulation started. How is a logic that can maintain these two contradictory requirements at the same time possible? Let us approach this question through the exploration of its possible presence in de Man's texts. We will start from the analysis of Pascal's Re? flexions sur la ge? ome? trie en ge? ne? ral; De l'esprit ge? ome? trique et de l'Art de persuader that de Man carries out in "Pascal's Allegory of Persuasion. "2
Pascal starts his study of the esprit ge? ome? trique from the distinction between nominal and real definitions--the first resulting from con- vention and being thus exempt from contradiction, the second being axioms or propositions to be proved--and asserts that the confusion between the two is the main cause of philosophical difficulties. Main- taining the separation between the two--as the geometrician does--is the first rule of philosophical clarity. However, the argument runs quickly into difficulties, as geometrical discourse includes not only nominal definitions but also "primitive terms"--such as motion, num- ber, and extension--which are undefinable but, nonetheless, fully intel- ligible. According to Pascal, these undefinable words find a universal reference not in the (impossible) fact that all men have the same idea concerning their essence, but instead, in the fact that there is a relation of reference between name and thing, "so that on hearing the expres- sion time, all turn (or direct) the mind to the same entity" (56). But, as de Man shows, this brings back the real definition into the geometrical camp itself, for
the word does not function as a sign or a name, as was the case in the nominal definition, but as a vector, as a directional motion that is mani- fest only as a turn, since the target toward which it turns remains un- known. In other words, the sign has become a trope, a substitutive rela- tionship that has to posit a meaning whose existence cannot be verified, but that confers upon the sign an unavoidable signifying function. (56)
As the semantic function of the primitive terms has the structure of a trope, "it acquires a signifying function that it controls neither in its ex- istence nor in its direction. " Ergo, "[s]ince definition is now itself a primitive term, it follows that the definition of the nominal definition is itself a real, and not a nominal, definition" (57).
This contamination of the nominal by the real definition is still
The Politics of Rhetoric 231
232 Ernesto Laclau
more visible when we move to the question of double infinitude, which is decisive in establishing the coherence and intelligibility of the rela- tionship between mind and cosmos. Here, Pascal deals with the objec- tions put to him by the Chevalier de Me? re? , according to whom--given the Pascalian principle of homogeneity between space and number--it is possible to conceive an extension formed by parts that are not them- selves extended, since it is possible to have numbers made up of units that are devoid of number. With this, the principle of infinite smallness would be put into question. Pascal's answer has two steps. He asserts, in the first place, that which applies to the order of number does not apply to the order of space. One is not a number, there is no plurality in it; but at the same time it belongs to the order of number for, given the Euclidean principle of homogeneity ("magnitudes are said to be of the same kind or species when one magnitude can be made to exceed another by reiterated multiplication"), it is part of the infinity postu- lated by that principle. On that basis, Pascal can distinguish between number and extension, but only at the price of grounding the distinc- tion in real and not nominal definitions. As de Man asserts:
The synecdochal totalization of infinitude is possible because the unit of number, the one, functions as a nominal definition. But, for the argument to be valid, the nominally indivisible number must be distinguished from the really indivisible space, a demonstration that Pascal can ac- complish easily, but only because the key words of the demonstration-- indivisible, spatial extension (e? tendue), species (genre), and definition-- function as real, and not as nominal, definitions. (58-59)
But--second step--if the order of number and the order of extension had to be separated to answer Me? re? 's objection, the rift between the two had also to be healed if the principle of homogeneity between both was to be maintained. This homology is restored by appealing, as far as number is concerned, to the zero--which, on the difference with one is radically heterogeneous with the order of number--and by finding equivalences in the order of time and motion, such as "instant" and "stasis. " This appeal to the zero, however, has dramatic consequences for the coherence of the system, which de Man describes in a passage worth quoting in full:
the coherence of the system is now seen to be entirely dependent on the introduction of an element--the zero and its equivalences in time and motion--that is itself entirely heterogeneous with regard to the system
and is nowhere part of it. . . . Moreover, this rupture of the infinitesimal and the homogeneous does not occur on the transcendental level, but on the level of language, in the inability of a theory of language as sign or as name (nominal definition) to ground this homogeneity without hav- ing recourse to the signifying function, the real definition, that makes the zero of signification the necessary condition for grounded knowl- edge. . . . It is as sign that language is capable of engendering the prin- ciples of infinity, of genus, species, and homogeneity, which allow for synecdochal totalizations, but none of these tropes could come about without the systematic effacement of the zero and its reconversion into a name. There can be no one without zero, but the zero always appears in the guise of a one, of a (some)thing. The name is the trope of the zero. The zero is always called a one, when the zero is actually nameless, "in- nommable. " In the French language, as used by Pascal and his inter- preters, this happens concretely in the confusedly alternate use of the two terms ze? ro and ne? ant. The verbal, predicative form ne? ant, with its gerundive ending, indicates not the zero, but rather the one, as the limit of the infinitely small, the almost zero that is the one. (59)
It is important to give serious consideration to this remarkable passage-- remarkable, among other things, because de Man does not pursue later in his essay all the implications of his own de? marche--for it contains, in nuce, all the relevant dimensions of the problem we are exploring. Everything turns around the role of the zero. The zero is, we are told, something radically heterogeneous with the order of number. The order of number, however, cannot constitute itself without reference to the zero. It is, in this sense, a supplement to the system that, none- theless, is necessary for constituting the latter. Vis-a`-vis the system, the zero is in an undecidable tension between internality and externality-- but an internality that does not exclude heterogeneity. The zero, in the second place is "innommable," nameless, but at the same time it pro- duces effects, it closes the system, even at the price of making it hope- lessly heterogeneous. It retotalizes the system, incurring, however, an inconsistency that cannot be overcome. The zero is nothing, but it is the nothing of the system itself, the impossibility of its consistent clo- sure, which is signified by the zero and in that sense, paradoxically, the zero as empty place becomes the signifier of fullness, of systematicity as such, as that which is lacking. The semantic oscillation between ze? ro and ne? ant that de Man observes is the result of this dual condition of the moment of closure: being an impossible object, but also a necessary
The Politics of Rhetoric 233
234 Ernesto Laclau
one. Finally, if the zero as moment of closure is impossible as an object but also necessary, it will have to have access to the field of represen- tation. But the means of representation will be constitutively inade- quate. It will give to the "innommable" a body, a name, but this can be done only at the price of betraying its true "nonbeing"; thus the tropological movement that prolongs sine die the non-resolvable dia- lectics between the zero and the one. In the words of de Man just quoted: "There can be no one without zero, but the zero always ap- pears in the guise of a one, of a (some)thing. The name is the trope of the zero. The zero is always called a one, when the zero is actually nameless, 'innommable. '"
Now, this succession of structural moments coincides, almost step by step, with the logic of hegemony that I have tried to describe in my work and which I see operating in the texts of Gramsci--to whom I will return later on. To start with, the condition of any hegemonic su- ture is the constitutive nonclosure of a system of political signification. The systemacity of a system, its closure--which is the condition of sig- nification in a system, such as Saussure's, whose identities are merely differential--coincides with the determination of its limits. These lim- its, however, can only be dictated by something that is beyond them. But, as the system is a system of differences, of all actual differences, that "beyond"--which should be heterogeneous with the system in order to fulfill its function of truly closing it--lacks the condition of a true heterogeneity if it consists in one more difference. The latter would be, in some way, undecided, suspended between being internal and external to the system. This jeopardizes the role of the "beyond" as limit and, as a result, the possibility of constituting the differences as truly intrasystemic differences. It is only if the beyond the limit has the character of an exclusion that its role as limit is restored and with it the possibility of emergence of a full system of differences. 3
However, this fullness of the system (obtained, it is true, at the price of a dialectical retrieval of its negation) has a shortcoming. For, vis-a`- vis the excluded element, all differences within the system establish relations of equivalence between themselves. And equivalence is pre- cisely that which subverts difference. So the "beyond" which is the condition of possibility of the system is also its condition of impossi- bility. All identity is constituted within the unsolvable tension between equivalence and difference.
So, as with the Pascalian zero, we arrive at an object that is at the same time impossible and necessary. As impossible, it is an empty place
within the structure. But, as necessary, it is a "nothing" that will pro- duce structural effects, and this requires that it has access to the field of representation. And, as in the dialectics between the zero and the one, this double condition of necessity and impossibility will be constitu- tively inadequate. The fullness of the system, its point of imaginary satu- ration, will be, as in the example of de Man, a nothing that becomes a something. What are the possible means of this distorted representa- tion? Only the particular differences internal to the system. Now, this relationship by which a particular difference takes up the representa- tion of an impossible totality entirely incommensurable with it is what I call a hegemonic relation.
There are only two differences between the hegemonic logic and the Pascalian dialectics between the zero and the one as described by de Man. The first is that--given the numerical nature of Pascal's case-- the zero can only be embodied by the one, while in the case of the hegemonic logic any element within the system can be a bearer of a hegemonic function. 4 The second difference is that, given de Man's in- terests, the determination of the heterogeneous character of the tropo- logical displacement from the zero to the one is the point of arrival of his analysis, whereas for a student of hegemonic logics the analysis of the exact nature of this tropological movement becomes imperative. In de Man's detotalizing discourse, what matters is to show the hetero- geneity out of which the tropological movement operates. This is also a vital part of a hegemonic analysis. But what is decisive for the latter is the determination of the partial retotalizations that the tropological movement makes possible. 5 This is the dimension that we now have to take into account. I will do it through a reference to the opposition metaphor/metonymy as presented in de Man's essay on Proust in Allegories of Reading. 6
II
The text on Proust deals, as is well known, with the discourse of young Marcel on the pleasure of reading and with the way in which such pleasure is constructed through a set of metaphoric substitutions that are, however, persuasive only through the operation of a series of con- tingent metonymic movements. De Man asserts:
The crossing of sensory attributes in synaesthesia is only a special case of a more general pattern of substitution that all tropes have in com- mon. It is the result of an exchange of properties made possible by a
The Politics of Rhetoric 235
236 Ernesto Laclau
proximity or an analogy so close and intimate that it allows the one to substitute for the other without revealing the difference necessarily in- troduced by the substitution. The relational link between the two enti- ties involved in the exchange then becomes so strong that it can be called necessary: there could be no summer without flies; no flies with- out summer. . . . The synecdoche that substitutes part for whole and whole for part is in fact a metaphor, powerful enough to transform a temporal contiguity into an infinite duration. . . . Compared to this compelling coherence, the contingency of a metonymy based only on a casual encounter of two entities that could very well exist in each other's absence would be entirely devoid of poetic power. . . . If metonymy is distinguished from metaphor in terms of necessity and contingency . . . then metonymy is per definition unable to create genuine links, whereas no one can doubt, thanks to the butterflies, the resonance of the crates, and especially the "chamber music" of the flies, of the presence of light and of warmth in the room. On the level of sensation, metaphor can reconcile night and day in a chiaroscuro that is entirely convincing. (62-63)
We see that this passage establishes the distinction between metaphor and metonymy on the basis of the two oppositions contiguity/analogy-- the dominant opposition in classical rhetoric--and contingency/necessity. As far as the first opposition is concerned, the difficulty is that the dis- tinction between analogy and contiguity is rather slippery. Contiguity, in rhetorical terms, cannot be equivalent to mere physical contiguity, for the latter can be the basis of a metaphoric relation. And analogy can depend on such a variety of criteria that we are actually faced with a continuum in which analogy fades into mere contiguity. De Man himself points out, for instance, that "Synecdoche is one of the border- line figures that creates an ambivalent zone between metaphor and metonymy and that, by its spatial nature, creates the illusion of a syn- thesis by totalization" (63). And in one of the essays included in Blindness and Insight, he asserts that it is notoriously difficult, logically as well as historically, to keep the various tropes and figures rigorously apart, to establish precisely when catachresis becomes metaphor and when metaphor turns into metonymy; to quote an apt water-metaphor to which an expert of the field (Lansberg) has to resort precisely in his discussion of metaphor: "the transition (of one figure to another, in this case, from metaphor to metonymy) is fluid. "7
We could say that the frontiers between figures and tropes in classi-
cal rhetoric are ancillary to the main objective distinctions of ancient ontology. This is evident from Aristotle to Cicero and Quintilian. It is precisely the close character of this system of distinctions that is put into question by the deconstructive turn. Both de Man and Ge? rard Genette, for example, have shown how Proust, great defender of the creative role of metaphor, had to ground his own metaphors in a gen- eralized system of metonymic transitions. 8
The distinction between necessity and contingency is more promis- ing. In this case, without being entirely able to avoid the continuum by which one figure fades into the other, we have at least a less ambiguous criterion of classification: a discourse will be more or less metaphoric depending on the degree of fixation that it establishes between its con- stitutive components. De Man attempts to show how all metaphorical totalization is based on a metonymic textual infrastructure that resists this totalizing movement. In Hegemony and Socialist Strategy we have asserted that hegemony is always metonymic. 9 In the light of our previ- ous analysis we see why this has to be the case. What is constitutive of a hegemonic relation is that its component elements and dimensions are articulated by contingent links. A trade union or a peasant organi- zation, for instance, can take up political tasks that are not related by necessary links to their own corporative specificity. The hegemonic links by which those political tasks become workers' or peasants' tasks are metonymic displacements based on relations of contiguity (on the simple availability of those forces in a certain context in which no other social force could take up those tasks--which involves no rela- tion of analogic necessity existing between task and agent). In that sense, if there is going to be hegemony, the traces of the contingency of the articulation cannot be entirely effaced.
The type of relationship involved in a hegemonic link can be further clarified if we return for a moment to the Pascalian zero. As in the case of the hegemonic relation, the heterogeneous character of the element that brings about whatever totalization can exist--the zero--is a con- tingent remainder that cannot be eradicated. But there is a crucial dif- ference between the latter and that inhabiting the tropological move- ment which is at the root of hegemony. Where in hegemony there is free variation as far as the element that occupies the hegemonic posi- tion is concerned, in the case of the zero we do not have such a latitude of maneuver: the zero can only be a one. In that case we are not deal- ing, properly speaking, with a metonymy but with a catachresis. 10 Now, in the field of rhetoric, catachresis occupies a very special position. At
The Politics of Rhetoric 237
238 Ernesto Laclau
the time of the last codification of classical rhetoric in the early nine- teenth century, it was even denied the status of a figure. Thus, Fontanier defined it in the following way:
La Catachre`se, en ge? ne? ral, consiste en ce qu'un signe de? ja` affecte? a` une premie`re ide? e, la soit aussi a` une ide? e nouvelle qui elle-me^me n'en avait point ou n'en a plus d'autre en propre dans la langue. 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.
no hegemonic articulation would be possible either. But, third, this re- totalization cannot have the character of a dialectical reintegration. It has, on the contrary, to maintain alive and visible the original and con- stitutive heterogeneity from which the hegemonic articulation started. How is a logic that can maintain these two contradictory requirements at the same time possible? Let us approach this question through the exploration of its possible presence in de Man's texts. We will start from the analysis of Pascal's Re? flexions sur la ge? ome? trie en ge? ne? ral; De l'esprit ge? ome? trique et de l'Art de persuader that de Man carries out in "Pascal's Allegory of Persuasion. "2
Pascal starts his study of the esprit ge? ome? trique from the distinction between nominal and real definitions--the first resulting from con- vention and being thus exempt from contradiction, the second being axioms or propositions to be proved--and asserts that the confusion between the two is the main cause of philosophical difficulties. Main- taining the separation between the two--as the geometrician does--is the first rule of philosophical clarity. However, the argument runs quickly into difficulties, as geometrical discourse includes not only nominal definitions but also "primitive terms"--such as motion, num- ber, and extension--which are undefinable but, nonetheless, fully intel- ligible. According to Pascal, these undefinable words find a universal reference not in the (impossible) fact that all men have the same idea concerning their essence, but instead, in the fact that there is a relation of reference between name and thing, "so that on hearing the expres- sion time, all turn (or direct) the mind to the same entity" (56). But, as de Man shows, this brings back the real definition into the geometrical camp itself, for
the word does not function as a sign or a name, as was the case in the nominal definition, but as a vector, as a directional motion that is mani- fest only as a turn, since the target toward which it turns remains un- known. In other words, the sign has become a trope, a substitutive rela- tionship that has to posit a meaning whose existence cannot be verified, but that confers upon the sign an unavoidable signifying function. (56)
As the semantic function of the primitive terms has the structure of a trope, "it acquires a signifying function that it controls neither in its ex- istence nor in its direction. " Ergo, "[s]ince definition is now itself a primitive term, it follows that the definition of the nominal definition is itself a real, and not a nominal, definition" (57).
This contamination of the nominal by the real definition is still
The Politics of Rhetoric 231
232 Ernesto Laclau
more visible when we move to the question of double infinitude, which is decisive in establishing the coherence and intelligibility of the rela- tionship between mind and cosmos. Here, Pascal deals with the objec- tions put to him by the Chevalier de Me? re? , according to whom--given the Pascalian principle of homogeneity between space and number--it is possible to conceive an extension formed by parts that are not them- selves extended, since it is possible to have numbers made up of units that are devoid of number. With this, the principle of infinite smallness would be put into question. Pascal's answer has two steps. He asserts, in the first place, that which applies to the order of number does not apply to the order of space. One is not a number, there is no plurality in it; but at the same time it belongs to the order of number for, given the Euclidean principle of homogeneity ("magnitudes are said to be of the same kind or species when one magnitude can be made to exceed another by reiterated multiplication"), it is part of the infinity postu- lated by that principle. On that basis, Pascal can distinguish between number and extension, but only at the price of grounding the distinc- tion in real and not nominal definitions. As de Man asserts:
The synecdochal totalization of infinitude is possible because the unit of number, the one, functions as a nominal definition. But, for the argument to be valid, the nominally indivisible number must be distinguished from the really indivisible space, a demonstration that Pascal can ac- complish easily, but only because the key words of the demonstration-- indivisible, spatial extension (e? tendue), species (genre), and definition-- function as real, and not as nominal, definitions. (58-59)
But--second step--if the order of number and the order of extension had to be separated to answer Me? re? 's objection, the rift between the two had also to be healed if the principle of homogeneity between both was to be maintained. This homology is restored by appealing, as far as number is concerned, to the zero--which, on the difference with one is radically heterogeneous with the order of number--and by finding equivalences in the order of time and motion, such as "instant" and "stasis. " This appeal to the zero, however, has dramatic consequences for the coherence of the system, which de Man describes in a passage worth quoting in full:
the coherence of the system is now seen to be entirely dependent on the introduction of an element--the zero and its equivalences in time and motion--that is itself entirely heterogeneous with regard to the system
and is nowhere part of it. . . . Moreover, this rupture of the infinitesimal and the homogeneous does not occur on the transcendental level, but on the level of language, in the inability of a theory of language as sign or as name (nominal definition) to ground this homogeneity without hav- ing recourse to the signifying function, the real definition, that makes the zero of signification the necessary condition for grounded knowl- edge. . . . It is as sign that language is capable of engendering the prin- ciples of infinity, of genus, species, and homogeneity, which allow for synecdochal totalizations, but none of these tropes could come about without the systematic effacement of the zero and its reconversion into a name. There can be no one without zero, but the zero always appears in the guise of a one, of a (some)thing. The name is the trope of the zero. The zero is always called a one, when the zero is actually nameless, "in- nommable. " In the French language, as used by Pascal and his inter- preters, this happens concretely in the confusedly alternate use of the two terms ze? ro and ne? ant. The verbal, predicative form ne? ant, with its gerundive ending, indicates not the zero, but rather the one, as the limit of the infinitely small, the almost zero that is the one. (59)
It is important to give serious consideration to this remarkable passage-- remarkable, among other things, because de Man does not pursue later in his essay all the implications of his own de? marche--for it contains, in nuce, all the relevant dimensions of the problem we are exploring. Everything turns around the role of the zero. The zero is, we are told, something radically heterogeneous with the order of number. The order of number, however, cannot constitute itself without reference to the zero. It is, in this sense, a supplement to the system that, none- theless, is necessary for constituting the latter. Vis-a`-vis the system, the zero is in an undecidable tension between internality and externality-- but an internality that does not exclude heterogeneity. The zero, in the second place is "innommable," nameless, but at the same time it pro- duces effects, it closes the system, even at the price of making it hope- lessly heterogeneous. It retotalizes the system, incurring, however, an inconsistency that cannot be overcome. The zero is nothing, but it is the nothing of the system itself, the impossibility of its consistent clo- sure, which is signified by the zero and in that sense, paradoxically, the zero as empty place becomes the signifier of fullness, of systematicity as such, as that which is lacking. The semantic oscillation between ze? ro and ne? ant that de Man observes is the result of this dual condition of the moment of closure: being an impossible object, but also a necessary
The Politics of Rhetoric 233
234 Ernesto Laclau
one. Finally, if the zero as moment of closure is impossible as an object but also necessary, it will have to have access to the field of represen- tation. But the means of representation will be constitutively inade- quate. It will give to the "innommable" a body, a name, but this can be done only at the price of betraying its true "nonbeing"; thus the tropological movement that prolongs sine die the non-resolvable dia- lectics between the zero and the one. In the words of de Man just quoted: "There can be no one without zero, but the zero always ap- pears in the guise of a one, of a (some)thing. The name is the trope of the zero. The zero is always called a one, when the zero is actually nameless, 'innommable. '"
Now, this succession of structural moments coincides, almost step by step, with the logic of hegemony that I have tried to describe in my work and which I see operating in the texts of Gramsci--to whom I will return later on. To start with, the condition of any hegemonic su- ture is the constitutive nonclosure of a system of political signification. The systemacity of a system, its closure--which is the condition of sig- nification in a system, such as Saussure's, whose identities are merely differential--coincides with the determination of its limits. These lim- its, however, can only be dictated by something that is beyond them. But, as the system is a system of differences, of all actual differences, that "beyond"--which should be heterogeneous with the system in order to fulfill its function of truly closing it--lacks the condition of a true heterogeneity if it consists in one more difference. The latter would be, in some way, undecided, suspended between being internal and external to the system. This jeopardizes the role of the "beyond" as limit and, as a result, the possibility of constituting the differences as truly intrasystemic differences. It is only if the beyond the limit has the character of an exclusion that its role as limit is restored and with it the possibility of emergence of a full system of differences. 3
However, this fullness of the system (obtained, it is true, at the price of a dialectical retrieval of its negation) has a shortcoming. For, vis-a`- vis the excluded element, all differences within the system establish relations of equivalence between themselves. And equivalence is pre- cisely that which subverts difference. So the "beyond" which is the condition of possibility of the system is also its condition of impossi- bility. All identity is constituted within the unsolvable tension between equivalence and difference.
So, as with the Pascalian zero, we arrive at an object that is at the same time impossible and necessary. As impossible, it is an empty place
within the structure. But, as necessary, it is a "nothing" that will pro- duce structural effects, and this requires that it has access to the field of representation. And, as in the dialectics between the zero and the one, this double condition of necessity and impossibility will be constitu- tively inadequate. The fullness of the system, its point of imaginary satu- ration, will be, as in the example of de Man, a nothing that becomes a something. What are the possible means of this distorted representa- tion? Only the particular differences internal to the system. Now, this relationship by which a particular difference takes up the representa- tion of an impossible totality entirely incommensurable with it is what I call a hegemonic relation.
There are only two differences between the hegemonic logic and the Pascalian dialectics between the zero and the one as described by de Man. The first is that--given the numerical nature of Pascal's case-- the zero can only be embodied by the one, while in the case of the hegemonic logic any element within the system can be a bearer of a hegemonic function. 4 The second difference is that, given de Man's in- terests, the determination of the heterogeneous character of the tropo- logical displacement from the zero to the one is the point of arrival of his analysis, whereas for a student of hegemonic logics the analysis of the exact nature of this tropological movement becomes imperative. In de Man's detotalizing discourse, what matters is to show the hetero- geneity out of which the tropological movement operates. This is also a vital part of a hegemonic analysis. But what is decisive for the latter is the determination of the partial retotalizations that the tropological movement makes possible. 5 This is the dimension that we now have to take into account. I will do it through a reference to the opposition metaphor/metonymy as presented in de Man's essay on Proust in Allegories of Reading. 6
II
The text on Proust deals, as is well known, with the discourse of young Marcel on the pleasure of reading and with the way in which such pleasure is constructed through a set of metaphoric substitutions that are, however, persuasive only through the operation of a series of con- tingent metonymic movements. De Man asserts:
The crossing of sensory attributes in synaesthesia is only a special case of a more general pattern of substitution that all tropes have in com- mon. It is the result of an exchange of properties made possible by a
The Politics of Rhetoric 235
236 Ernesto Laclau
proximity or an analogy so close and intimate that it allows the one to substitute for the other without revealing the difference necessarily in- troduced by the substitution. The relational link between the two enti- ties involved in the exchange then becomes so strong that it can be called necessary: there could be no summer without flies; no flies with- out summer. . . . The synecdoche that substitutes part for whole and whole for part is in fact a metaphor, powerful enough to transform a temporal contiguity into an infinite duration. . . . Compared to this compelling coherence, the contingency of a metonymy based only on a casual encounter of two entities that could very well exist in each other's absence would be entirely devoid of poetic power. . . . If metonymy is distinguished from metaphor in terms of necessity and contingency . . . then metonymy is per definition unable to create genuine links, whereas no one can doubt, thanks to the butterflies, the resonance of the crates, and especially the "chamber music" of the flies, of the presence of light and of warmth in the room. On the level of sensation, metaphor can reconcile night and day in a chiaroscuro that is entirely convincing. (62-63)
We see that this passage establishes the distinction between metaphor and metonymy on the basis of the two oppositions contiguity/analogy-- the dominant opposition in classical rhetoric--and contingency/necessity. As far as the first opposition is concerned, the difficulty is that the dis- tinction between analogy and contiguity is rather slippery. Contiguity, in rhetorical terms, cannot be equivalent to mere physical contiguity, for the latter can be the basis of a metaphoric relation. And analogy can depend on such a variety of criteria that we are actually faced with a continuum in which analogy fades into mere contiguity. De Man himself points out, for instance, that "Synecdoche is one of the border- line figures that creates an ambivalent zone between metaphor and metonymy and that, by its spatial nature, creates the illusion of a syn- thesis by totalization" (63). And in one of the essays included in Blindness and Insight, he asserts that it is notoriously difficult, logically as well as historically, to keep the various tropes and figures rigorously apart, to establish precisely when catachresis becomes metaphor and when metaphor turns into metonymy; to quote an apt water-metaphor to which an expert of the field (Lansberg) has to resort precisely in his discussion of metaphor: "the transition (of one figure to another, in this case, from metaphor to metonymy) is fluid. "7
We could say that the frontiers between figures and tropes in classi-
cal rhetoric are ancillary to the main objective distinctions of ancient ontology. This is evident from Aristotle to Cicero and Quintilian. It is precisely the close character of this system of distinctions that is put into question by the deconstructive turn. Both de Man and Ge? rard Genette, for example, have shown how Proust, great defender of the creative role of metaphor, had to ground his own metaphors in a gen- eralized system of metonymic transitions. 8
The distinction between necessity and contingency is more promis- ing. In this case, without being entirely able to avoid the continuum by which one figure fades into the other, we have at least a less ambiguous criterion of classification: a discourse will be more or less metaphoric depending on the degree of fixation that it establishes between its con- stitutive components. De Man attempts to show how all metaphorical totalization is based on a metonymic textual infrastructure that resists this totalizing movement. In Hegemony and Socialist Strategy we have asserted that hegemony is always metonymic. 9 In the light of our previ- ous analysis we see why this has to be the case. What is constitutive of a hegemonic relation is that its component elements and dimensions are articulated by contingent links. A trade union or a peasant organi- zation, for instance, can take up political tasks that are not related by necessary links to their own corporative specificity. The hegemonic links by which those political tasks become workers' or peasants' tasks are metonymic displacements based on relations of contiguity (on the simple availability of those forces in a certain context in which no other social force could take up those tasks--which involves no rela- tion of analogic necessity existing between task and agent). In that sense, if there is going to be hegemony, the traces of the contingency of the articulation cannot be entirely effaced.
The type of relationship involved in a hegemonic link can be further clarified if we return for a moment to the Pascalian zero. As in the case of the hegemonic relation, the heterogeneous character of the element that brings about whatever totalization can exist--the zero--is a con- tingent remainder that cannot be eradicated. But there is a crucial dif- ference between the latter and that inhabiting the tropological move- ment which is at the root of hegemony. Where in hegemony there is free variation as far as the element that occupies the hegemonic posi- tion is concerned, in the case of the zero we do not have such a latitude of maneuver: the zero can only be a one. In that case we are not deal- ing, properly speaking, with a metonymy but with a catachresis. 10 Now, in the field of rhetoric, catachresis occupies a very special position. At
The Politics of Rhetoric 237
238 Ernesto Laclau
the time of the last codification of classical rhetoric in the early nine- teenth century, it was even denied the status of a figure. Thus, Fontanier defined it in the following way:
La Catachre`se, en ge? ne? ral, consiste en ce qu'un signe de? ja` affecte? a` une premie`re ide? e, la soit aussi a` une ide? e nouvelle qui elle-me^me n'en avait point ou n'en a plus d'autre en propre dans la langue. 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.
