POST HUMANISM, THE PROBLEM OF THEORISING `COHESIONS' OF POWER, AND THE FALSE SOLUTION OF `DISCURSIVE COHERENCE'

Abstract code: ZIP98209

Originally titled: Concepts of `discourse' in critical educational research:

Some tensions and primrose paths

Lew Zipin, University of Wisconsin-Madison

Paper presented as a work in progress

at the Australian Association for Research in Education Annual Conference

Adelaide, South Australia

29th November - 3rd December, 1998

Lew Zipin

14 Bandjalong Cres

ARANDA, ACT 2614

phone: 02 6251 9639

email: ldzipin@rocknet.net.au

In A. S. Byatt's novel Possession: A Romance (1990), Roland, a British doctoral student in literature, and Maud, a feminist literature professor, come together through odd career convergences. Roland's library research of unsent letter drafts by Randolph Henry Ash, a Victorian poet on whose work Roland's doctorate is based, prompts him to pursue the identity of an unnamed woman to whom the letters were intended. Fortuitous clues lead him to guess that the woman was Christabel LaMotte, also a poet of that era. From a fellow student who once had a conference affair with Maud, Roland learns that she is an interpreter of LaMotte's poetry. He makes an in promptu visit to Maud's campus, consults her, and gains her interest in the matter. Together, they discover some locked-away letters LaMotte had kept: an exchange between the two poets that makes evident their clandestine affair, as yet unknown to biographers of the poets. Searching further pieces of the delicious story, the two scholars travel together to various scenes of the poets' lives, until they are surprised by their own intimate feelings.

Tantalised and confused by signs of possible romance with Maud, Roland one day recalls sentiments Ash wrote in a letter to LaMotte. Byatt portrays Roland's musings:

This passage, from Byatt's roman a clef about academic characters, jogs my thinking about how `coherency' in social and subjective life has indeed been made problematic by certain trends of `postmodernist' academic theory. In my readings of influential authors such as Michel Foucault (1979, 1990), Judith Butler (1990, 1992, 1993), and Ernesto Laclau & Chantal Mouffe (1985), I find great emphasis on the "free" play of "aleatory," contingent, "multivalent" possibilities. There is stress on diversity across dispersed sites of social practice and subjective experience. There is insistence on the lack of any necessary connecting principles which might order this diversity into coherent social "structurings". Most crucially, these authors make it philosophically unfashionable to posit or assume any essentially `human' imperative or capacity -- what Byatt signifies as "deep human desires" -- due to which, we might theorise, all the "random tangles" that proliferate across dispersed sites and contingent events might comb into "coherent plots".

And yet, these authors all pause on the brink of the logical conclusion that there is no basis whatsoever for coherence in practical or subjective life, and so social history should appear as a runaway process of dispersed fragmentation. They recognize that their own research projects engage in sociological and historiographic critique of power formations: of institutional systems, and structural asymmetries of social relation. They thus seek ways to recuperate a notion of "ordering principles" which prevent their own emphases on "connections [that] proliferate apparently at random" from getting "out of hand".

In this paper, I argue that the possibility of theorising `structurally' coherent formations -- and thus avoiding a logic by which social history devolves utterly into random dispersions -- is gravely hamstrung by philosophical commitments of the sort I call `post-humanist' (they could also be called `post-essentialist,' or `post-foundationalist'). Such a commitment strictly refuses to assume any trans-historically essential, a prioristically `human' needs or capacities of people to comb out "random tangles" and generate "coherent plots". Thus, if "ordering" effects are to be discerned in history, they must not be thought to proceed from any "human desires" which, at a level too "deep" to be "controlled by conscious intention", are nonetheless founded at an unconscious level of `human species being', from whence they operate in social history.

Among advocates of such a strict post-humanism, there are those who have sought to resolve the logical problems for envisioning salient formations of cohesive power. There have been efforts to theorise how social `structurings' can emerge and effectively operate across diverse social contexts, but as historically contingent effects: i.e. in no way founded in any trans-historic natures of human being. Perhaps the most forthright of such efforts is by Laclau & Mouffe, in their influential book Hegemony & Socialist Strategy: Toward a Radical Democratic Politics (1985). In what follows, I present these authors' rejection of any and all assumptions of anthropologically-based natures which might explain the origin of cohesive forms of social-historical power. I render their recognition of consequent difficulties for conceiving how extensive and durable forms of `structural' power are possible. I show how they take female subordination (in relation to male power) as illustrative, and seek to theorise how gendered power asymmetry might originate, spread and endure as a historically contingent effect.

Laclau & Mouffe's theoretical efforts rely on a concept of `discourse', derived from Foucault (1972), which I examine. I argue that this conceptual device does not, and cannot, enable a non-essentialist theorisation of `structural' effects in social history. Rather, Laclau & Mouffe fetishise a `discursive' substrate (or foundation) for those effects, which is every bit as essentialist as the anthropological substrate they renounce. I conclude that critical researchers cannot avoid essentialising any and all capacities to generate the forms of historic power that they analyse. It is therefore advisable carefully to define such capacities. But however conceptually defined, I argue, it is philosophically necessary to assume such coherence-generating capacities to inhere in human agents, rather than in discourses (which human agents produce). I end the paper with a brief `neo-humanist' gesture toward viably defined capacities of people to generate `structural' coherence, which I derive from Pierre Bourdieu's concept of a habitus (1977, 1992).

In Hegemony and Socialist Strategy, Laclau & Mouffe affirm that their theoretical efforts aim to preclude certain commonly held assumptions about essentially `human' attributes:

In other words, no anthropologically-founded natures predicate the existence of any forms of power asymmetry which we may discern in social history. Let us say we are researchers of power along a gendered axis of patriarchal asymmetry. We must refuse to assume any essential natures -- in human beings in general; or in men in general, as against women in general -- which predicate the emergence, extensiveness or persistence (in always transforming ways) of women's subordination and struggles in relation to male advantage.

Do Laclau & Mouffe indeed recognize power as taking social-historical forms along relational axes such as a gender? Do they recognize a systemic advantage of men over women that is socially extensive and historically durable? I will soon show that they do accord an `empirical' saliency to gender as a `social-structural' power axis; and I will examine how they theorize the historic possibility of this relational asymmetry. But I want need first to clarify further how, in their efforts to theorise such formations of structural power, they seek a rigorous refusal of any and all assumptions about human essences which could help explain such formations. That is, their rejection of "any ...principle or substratum of an anthropological nature" goes further than structure-specific principles, such as essences of specifiably gendered asymmetry. It precludes also what we might think of as more rudimentary capacities of `human agency'; e.g. capacities to resist, to choose, etc. Say Laclau & Mouffe:

I read this passage to suggest the following logic: 1) In "social agents" there inhere no trans-historic essences that may be thought to originate or sustain the forms of social order which happen to arise in history; and 2) consequently, any historic emergence of a "certain order" can only be a contingent convergence across diverse conditions and events. Since no underlying principle holds all the divergent elements of the convergence together, any appearance of cohesive regularity is always a highly precarious and temporary form of "fixation".

As will later become apparent, when Laclau & Mouffe contemplate the `structural' regularity of gendered power asymmetry, they back off considerably from this suggestion of highly precarious instability in any convergence of "a certain order". But before I turn to their efforts to theorize how gender asymmetry gains cohesive continuity in social history, I want to render some further statements which show how strictly they reject any anthropological ground for the origin of certain orderings of social practice and relation. Citing Foucault, Laclau & Mouffe declare that they join him in "breaking with the category of an `originative subject', which continues to creep into the very conceptions that seek to implement the rupture with it" (115). They assert:

I will soon look closely at what Laclau & Mouffe mean by "discursive conditions of possibility" for the structural coherence of certain forms of social relation. At this juncture, I want to underscore that, whatever `discourse' means to Laclau & Mouffe, they invoke it to eliminate any last notion that structural regularities of social relation could originate from any intrinsically "originative" capacities of social agents. They insist that no capacities which appear to manifest in the practical lives of social agents -- not even a basic capacity to "experience" -- can be thought to inhere originally `in' those agents (i.e. as ontological natures of their `being'). "Subjects" capacitated to experience and act in social practices and relations themselves originate only and utterly as constituted byproducts of "discursive conditions of possibility".

Laclau & Mouffe recognize that the philosophical break from "an `originative subject'" -- and thus, from all assumptions that social agents are ontologically capacitated to originate forms of power relation -- poses tense problems for theorising socially extensive and historically persistent `structurings' of relational power (e.g. institutionalised gender, class and race asymmetries). I noted earlier their own suggested logic that, if "social agents lack any essence," then convergent forms of social regularity which appear in history should only be of a most precarious sort, always on the edge of dispersion. An acute emphasis on dispersion is indeed logically compelling for those who renounce anthropologically-founded originative capacities, observe Laclau & Mouffe:

That is, the post-humanist break from originary essences refused not only all "essentialism of the totality" of a structural formation, but all "essentialism of the elements" (116) within the totality. For example, it became difficult to conceive how `female' and `male' social positionings could cohere as macro-elements in an op-positional `whole structure', when those very macro-categories must themselves comprise multiple and contradictory micro-situations of gender construction. In "the critique of feminist essentialism," note Laclau & Mouffe, feminists themselves rejected unifying principles of "women's oppression" (117). It could be questioned whether women were `oppressed' in every context; and when so, in relation to whom? Do women not form identities in terms of multiple and shifting social-positional categories, in which oppression is often more saliently felt along axes other than gender? If subordination `as female' nonetheless seems widespread, should it not still be "denied that there is a single mechanism of women's oppression" (117)? Ought not the focus to be on specific, locally contingent, diverse and dispersed conditions in which "the category of `the feminine' is constantly produced" (117)?

Laclau & Mouffe largely endorse this appreciation of female and male "experience" as produced in "the field of dispersion of subject positions" (117). However, such postmodern stress on "random tangles," as Byatt says, can get "out of hand". As Laclau & Mouffe put it:

I want to note that this passage affirms a systematic stability of "sex/gender" as a power-relational asymmetry across diverse sites of social practice. Laclau & Mouffe say that "every construction of sexual differences" -- no matter how multiple and heterogeneous -- "invariably constructs the feminine as a pole subordinated to the masculine". The stress on "invariability" suggests a quite "fixated" regularity of outcome. We might ask what happened to the earlier-quoted assertion that "[s]ociety and social agents lack any essence, and their regularities merely consist of the relative and precarious forms of fixation which accompany the establishment of a certain order" (98). Indeed, I find it brave of them to state an empirical recognition of such a remarkably "invariable" fixation of structural asymmetry, which presumably applies broadly across social space, and durably over historic time. That is, I find it brave that they take, as a theoretical target, such a difficult social regularity to theorise, especially given their stricture that "[s]ociety and social agents lack any essence" which can help explain how ordered regularities hold with any extensive or durable fixity.

I will shortly examine how Laclau & Mouffe specifically theorize the gendered effect of an "invariable" subordination of `female' in relation to `male'. But I must first render their approach to theorising the formation of `structurings' in general. In the next section, I examine their argument that, while historic structures of social relation cannot originate in whole or part from any essentially `human' capacities, structures can emerge as `discursive' effects.

As cited earlier, Laclau & Mouffe characterise the break from assuming anthropologically-based originative capacities of `subjects' also as a shift to conceiving `the subject' strictly "in the sense of `subject positions' within a discursive structure. Subjects cannot, therefore, be the origin of social relations" (115) in which they are structurally positioned. The capacity of social agents even to experience `themselves' as socially positioned "depends on precise discursive conditions of possibility" (115).What, then, is the meaning of the adjective "discursive" by which Laclau & Mouffe stress the utterly derivative existence of `subjects' with capacities? By what definition does this `discursivity' make possible what social agents cannot make possible: i.e. the emergence (or origination) of relational structurings in the practical course of social-historical events?

According to The Fontana Dictionary of Modern Thought (1988), an early and elementary definition of `discourse', as developed in linguistic analysis, is "a stretch of language ... The term discourse analysis is often applied to the study of linguistic effects -- semantic, stylistic, syntactic -- which ... take into account ... sentence structure" (232). In later developments of sophisticated critical and sociological approaches to "discourse analysis", focus has not been strictly on linguistic practices per se (i.e. stretches of speech or writing). For example, notes the Fontana Dictionary, "Foucault placed more stress on relationships between discourse and other social practices" (232). And in researching the `structural' effects of discourse in relation to practical social contexts, says the Dictionary, Foucault stressed how particular discourses embed selective norms "which privilege some and punish others", and so "are in fact `discourses of power'" (232).

That is, Foucault stressed how all socially contextualised practices -- even the most silently physical -- are always infused by certain historically contingent ways of knowing (or `epistemes') in which are encoded certain norms of `proper' versus `deviant' perception and behaviour. Thus, all the material practices of any social-institutional domain or context (say, schools in an education system) are always inextricably pervaded by certain forms of historically associated "power-knowledge" (Foucault, 1979), or "discourse", which make `self', `other' and `world' structurally intelligible. Only in this very broad sense of `linguistic', then, can the discourses that tacitly infuse social practices be called "stretches of language". And it is at this un(der)statedly subtle level of `discursive practices' that the norms they encode are effectively communicated to, and constitutive of, subconscious structures of cognition and identity. In this manner, social agents take up subject positions as `normal' or `deviant' selves, in relation to others. As Nancy Fraser puts it (1989): "For Foucault, the subject is merely a derivative product of a certain contingent, historically specific set of linguistically infused social practices that inscribe power relations upon bodies" (56).

Laclau & Mouffe deploy this conception of the `discursivity' of social practices. They say:

The distinction is incorrect in the sense that no situations of social practice are ever disconnected from cultural infusions of `linguistic' meaning. No matter how mutely behavioural, practices are always latently bound up with systems of sensible intelligibility (i.e. `discourse'). Only as an interpretative device of `discourse analysis' is it apt to abstract a `linguistic aspect', so as to decode normative rules that structure meaningful sense. But in actual practice, behavioural and linguistic aspects are always inextricably fused in the social production of `subjects' with coherent senses of `self' in relation to `others', as "structured under the form of discursive totalities".

Hence, "social practices" are always also "discursive practices". There simply are no "non-discursive practices". But I here want to press the question: How do structured forms of discursive totality come about, in and across local sites of social practice? This repeats the question with which I began this section: How does a concept of `discursivity' explain what cannot be explained as caused by social agents: i.e. the emergence of relational structurings in the practical course of social-historical events?

Consider a local context of social practice. According to the theory that there are no non-discursive practices, this social site never lacks the infusion of certain forms of discursive structuring. However, suggest Laclau & Mouffe, this particular site is to be conceived as having a free-radical indeterminacy with regard to how it, given sets of practices within it, and the social agents who inhabit it, may take up positions within various discourses. This is because there is no one discourse formation to which any particular context, practice or social agent `naturally' belongs. We can think of "discursive totalities" as broad-based formations of power/knowledge-in-practice, which run across and incorporate many sites and agents. But a given site or agent may shift into and out of alignment within different discursive structures, and will inevitably take positions in multiple discourses at once. Thus, given sites and agents (or `subjects') are to be conceived, simultaneously, as 1) always positioned in certain discursive structures; and 2) always potentially independent of those very same "discursive totalities", in their capacity to shift and re-align. In the first of the above statuses -- as "differential positions" which "appear articulated within a discourse" -- "we will call [them] moments," say Laclau & Mouffe; while in their simultaneous status as "not discursively articulated", "we will call [them] element[s]" (105)

My question of how forms of social/discursive structure come about thus becomes: How do different social practices, sites and agents -- in their "elemental" status, as "not discursively articulated" -- come into articulated connection as "moments" within discursive totalities? How do free-radical elements not only align within already circulating forms of discursive structure, but also organise in new forms of structural coherence? That is: How do elements actively "articulate" in newly emergent historic forms of discursive structuring? Laclau and Mouffe's answer deploys a certain concept of "discursive articulation" (adapted from Gramsci, 1971). They say:

Laclau & Mouffe here name an active structure-forming process. They call "articulation" those "articulatory practice[s]" which modify relations among elements, resulting in new forms of structurally articulated "discourse". I will recall here that, according to Laclau & Mouffe, all practices are "discursive practices". Thus, articulatory practices are also discursive practices; i.e. they are themselves already "structured under the form of discursive totalities". Do Laclau & Mouffe thus offer a circular explanation of the formation of new "discursive totalities" through the action of "articulatory practices"? That is, do articulatory practices bring about forms of relational totality which are actually not new, but already formed and carried in the articulatory practices? If so, we still require a theoretical explanation of how articulatory practices themselves come to carry certain forms of discursive structuring.

I suggest we need a more elaborated theory of the active articulatory generation of articulated effects. We need a theory that starts at the level of "elements", and elaborates how these elements come into relational connection as "moments" within discursive structures. Stuart Hall (1986), in his reading of Laclau & Mouffe's concept of "articulation", amplifies what a "theory of articulation" must help us understand. Says Hall:

Hall's reading of the double meaning of "articulate" confirms that "articulatory practices" are "discursive"; i.e. they encode `language'-infused intelligibilities. Their action tacitly conveys a `languaged' coherence which goes into the forging of cohesive connections among elements. Hall also confirms the free-radical status by which "different, distinct elements ... have no necessary `belongingness'" to any of the discursive `unities' within which they form relational links. Moreover, Hall broadens the definition of "elements". He suggests a sequence of "elements" of progressively more inclusive composition. The sequence begins at the micro level of "different practices", which, as I earlier suggested, we might read to indicate different local sites of practices wherein social agents take up `subject positions' in `discursive structures'. A "theory of articulation" thus must help us understand how locally situated agents come to cohere in larger "social groups", which are in turn structurally related within broader "social movement[s], etc." I suggest that Hall's "etc." indicates a structuring of relations between ever-broader "elements", extending to the level of macro-positions in historically salient social-relational `structures' (e.g. `male' and `female' genders, structured in an op-positional power relation whereby, say Laclau & Mouffe, "the feminine ... pole is [invariably] subordinated to the masculine".)

But the affirmation that fairly extensive and durable structural `unities' are "not necessary, determined, absolute and essential for all time" still does not explain how connections among the elements which compose them "can... be forged or made". Hall suggests we can address this question by attending to the "circumstances" -- to "certain conditions" -- under which "elements ... come to cohere together". I read this to indicate social-historical contingencies of circumstance, which can be elucidated in empirical research, and which may be thought to facilitate a greater likelihood (but not necessity) of the structures one sees as having emerged in those conditions.

However, I want to trouble any sense of the sufficiency of such `empirical' answers. Are not the contingent historical "conditions" which favour a given structural emergence precisely the "elements" -- the diverse sites and instances of social practice -- which compose the structural formation? The suggestion that certain structures emerge because of certain "conditions of possibility" thus may not say much more than that these structures emerge "because they do". Moreover, should we not expect that, over any broad extent of social space and historic time, contradictions will accumulate across "different, distinct" practical sites and events, as conditions more favourable for dispersive divergences than for cohesive convergences of elements within reasonably extensive and stable formations? This is, after all, the problem of "one-sided ... dispersion" that Laclau & Mouffe seek to solve through a theory of discursive "articulation".

Perhaps a theory that `solves' the problem of dispersion requires another kind of enabling factor to help explain how coherent connections "can be forged or made". We may need to appreciate that "elements" cannot simply induce themselves to cohere in a structural formation. Or rather, we may need to appreciate that only the living human agents who inhabit `elemental' sites of social practice can be sources of articulatory, meaning-giving, cohesion-forging capacities to originate structural `unities'. Perhaps we must make a philosophical-anthropological assumption that such capacities inhere originally in social agents, as essential capacities of anthropological being. Although Hall does not specifically postulate such a cohesion-generating agency, he does argue that "the theory of `a history without subjects,' a language with no speakers'" -- while a fruitful thought experiment -- can "only [be] a stopping point on the route to something else. It's just not possible to make history without subjects in quite that absolute way" (56).

Laclau & Mouffe, however, will take no route that recuperates an `originative subject'. They stipulate that any and all original `essences' must be precluded from efforts to theorize how forms of structural coherence can emerge across different and distinct elements:

This forecloses any possible locus or substrate wherein a priori (original; essential) forces which generate coherence might inhere. There is no `logical' principle of mutual attraction that, to begin with, inheres in-and-across the elements of a discursive formation, by which they magnetically draw together. Nor does any "meaning-giving" capacity inhere, transcendentally, in the people who inhabit social practices, such they can initiate the forging of intelligible discursive formations across distinct practical situations. This reiterates the earlier-discussed assertions that "[s]ociety and social agents lack any essence", and "cannot, therefore, be the origin of social relations".

Having thus stipulated what can not bring about a discursive coherence, what might Laclau & Mouffe suggest can induce intelligible cohesions across a dispersion of elements? Here is their most definitive statement of how coherent formations are possible:

No matter how many times I read this explanation, I remain baffled by it. After stipulating that a "discursive formation is not unified in the logical coherence of its elements", Laclau & Mouffe come back with the assertion that "dispersion itself [is] the principle of unity, insofar as it is governed by rules of formation". The forming of different elements in a cohesive ensemble thus seems to depend on the existence of "rules", i.e. regulatory logics of formation. But if these rules neither inhere among the elements a priori to the forming of `totalities', nor do they abide in any external locus of underlying formational principles, then how do "the complex conditions ... of the dispersed [elements]" get so far as to be "governed by rules"? How do rules come about? Where, in their theory, do Laclau & Mouffe elaborate a capacity to generate rules that operate in the `language'-infused articulatory practices which `signify' discursive coherences?

In this section, I have tried to chase down such a cohesion-generating capacity in Laclau & Mouffe's theory of `discursive articulation'. I find it unelaborated, yet needed, as the missing link without which their explanations of social-structural cohesion entail a slippery circularity. The discursive intelligibility of social practices results from the action of `articulatory practices' which seem already infused with discursive intelligibility. The `signification' of coherence across dispersed elements occurs by virtue of "rules of formation" which either exist in advance (despite stipulations to the contrary), or are of a most mysterious origin: the product of a magical alchemy, in which "regularity" spontaneously converges in-and-of "dispersion itself".

If, as I argue, the elaboration of a capacity to originate coherence is theoretically necessary, it is also, I argue, a prohibitive obstacle for post-humanist projects. Some post-humanist authors acknowledge a need to elaborate such a capacity. They affirm that, given their refusal to proceed from the founding assumption that meaning-giving, coherence-forming capacities exist as a priori essences of anthropologically-based agency, it is necessary to theorise how such `agency' can emerge contingently: i.e. not as an original capacity of `subjects', but as a derivative byproduct of discursive practices which constitute `the subject'. Elsewhere (Zipin, 1998), I have examined two such efforts (Butler, 1990; Davies 1990), and concluded that they do not, and cannot, succeed in theorising a historically contingent, derivative capacity to make coherence in the course of social practices. Rather, these theorists stress the `discursivity' of practices to a point that goes beyond the sense that all social practices are always tacitly `languaged'. They fetishise the linguistic aspect of practices into an Idealistic substrate -- every bit as essentialist as the anthropological substrate it displaces -- wherein capacities to forge coherence have an a priori basis of existence.

In the next section, I unpack how, despite their denials, Laclau & Mouffe engage in just such a fetishisation of the `languaged' aspect of practices, in their efforts to explain how gender assymetry can come about as a `discursive' effect.

As cited earlier, Laclau & Mouffe state an empirical recognition that, across a "multiplicity and heterogeneity" of practical sites where gender is made meaningful, "[e]very construction of sexual differences ... invariably constructs the feminine as a pole subordinated to the masculine" (117-118). The challenge, then, is to theorise how a dispersed multiplicity of gender constructions come to cohere in a broad-based relational structuring that characteristically positions `female' as subordinate to `male'. Laclau & Mouffe offer the following theoretical explanation:

To begin, let us consider the `elements': "the diverse social practices" (or diverse "concrete" situations of practice). Laclau & Mouffe observe an "autonomy and uneven development of the diverse practices"; that is, local sites are distinct from each other in the specificities of their practical conditions. As such, suggest Laclau & Mouffe, the multiplicity of dispersed sites yields a heterogeneous variety of "significations constituting `femininity'" (and `masculinity'). Indeed, considered as `elements' -- i.e. not already articulated as `moments' in a discursive ensemble -- diverse practical conditions ought, I argue, to yield quite divergent meanings of `feminine' and `masculine' gender. We should have no reason to expect that local constructions will always or even typically produce a `subordinated' significance of `female' in relation to `male'. And yet, Laclau & Mouffe refer to "the autonomy and uneven development of the diverse practices which construct the concrete forms of subordination". This repeats their recognition of an "invariable" construction of "the feminine as a pole subordinated to the masculine". But why should this pervasive typicality of `subordination' be the case? How does it come about?

Laclau & Mouffe answer that an over-arching "imaginary signification" of "`subordination', as a general category", somehow comes to characterise "the symbolism which is linked to the feminine condition in a given society". This generalised connotation of `feminine' as `subordinate' enters into "close correlation" with the specific sites of "diverse practices which construct the concrete forms" of gender significance. As such, it "plays a primordial role", informing the signifying processes in all the local and uneven sites, unifying these elements in a negative feedback loop whereby they "react, in turn, by contributing to the maintenance and reproduction of this symbolism". Hence, even as local conditions change over time, they retain and reproduce the pan-elemental "common element", which, across the dispersion, "has strong over-determining effects in terms of the sexual division". The distinct elemental sites "mutually reinforce and act upon one another", co-relating according to regulatory rules of a subordinated `femininity'. We have the structural effect of a discursively articulated formational ensemble.

But the explanatory viability of an "over-determining" feedback loop hinges on the provision, stated by Laclau & Mouffe, that such can abide "once female sex has come to connote a feminine gender" in which an "imaginary signification" of `subordination' is generalised. Thus, the explanation becomes circular. On one hand, the symbolic "common element" which is so crucial to a homeostatic "maintenance and reproduction" across "the diverse practices" must first come into existence. But on the other hand, this generalised symbolism needs in the first place to exist, in order to play its "primordial role" in inducing all the specific situations to contribute "in turn ... to the maintenance and reproduction of this symbolism". To escape such circularity, it is necessary to elaborate how a pervasive category of `subordination' emerges as a historically contingent, yet `primordially' deep-structuring, first instance. It must be elaborated how all the dispersed conditions give social-historical birth to such a `totalising' symbolism.

However, I argue that a general symbolism of `female subordination' can not plausibly be explained as having a contingent social-historical genesis, if Laclau & Mouffe stick to their post-essentialist guns. If autonomous and uneven conditions really lack any a priori cohesive tendency -- and no "meaning-giving subject" exists with agency to create coherent connections -- then we should expect exactly what Laclau & Mouffe themselves suggest as a corollary of their assertion that "[s]ociety and social agents lack any essence": i.e. that any social "regularities [will] merely consist of ... precarious forms of fixation" (98). Perhaps, across a certain range of sites, conditions might be sufficiently congruent for a mutually reinforcing convergence of gender constructions. However, we should expect any such `regularity' to be small-scale across social space and historic time. The greater the extent of space/time, I argue, the more we should expect multiple and diverse convergences, with contradictions across them: they should not share a connotation of `female' as `subordinate'. Such contradiction should disrupt all possibility of a vast convergence, let alone its "maintenance and reproduction". The broad and durable "invariability" that Laclau & Mouffe attribute to `female subordination' could not conceivably emerge as an effect of "dispersion itself".

There are only two theoretical tacks, I argue, by which one might be able to escape this problem that Laclau & Mouffe recognise as a "one-sided emphasis given to the moment of dispersion" (117). Both require abandoning a post-essentialist gun. That is, at least one of the two kinds of "essence" that Laclau & Mouffe shoot down must be recuperated: either 1) structure-specific a priori principles, which "would assign a character of inevitability" to given "forms of subordination" (152); or 2) non-structure-specific capacities of originative subjects, such as "the a priori of a ... meaning-giving subject" (105). I certainly agree with Laclau & Mouffe about the undesirability of the former. But I would argue for the desirability of the latter, which, as a non-specific agency to create intelligible meaning (or coherence; structure), does not predicate that any given form of relational structuring will necessarily emerge in social history.

I consider the latter option in the next section. I here want to argue that, in their explanation of "over-determining effects in terms of the sexual division", Laclau & Mouffe actually do invoke an essential principle of gender division. Since they do not explain how "dispersion itself" can give historic birth to `subordination' as "a general category", we might infer that the circularity of their theory secretly relies on the assumption of a trans-historic existence of that category. It its fair, I argue, to read their reference to "the diverse practices which construct the concrete forms of subordination" as implying that diverse practical sites all share a tendency, to begin with, toward constructions of `female' as `subordinate'. That is, the "common element" always already inheres in each elemental site, as a mutual principle informing the diverse "concrete forms of construction" that feed into a general "symbolism ... linked to the feminine condition". By my reading, Laclau & Mouffe thus implicate what they explicitly say their theory must exclude: an "original sexual division" which explains why this division is "represented a posteriori in social practices" (117).

No matter how Laclau & Mouffe may assert the opposite, the circularity of their theory, I argue, finesses just such a gender-specific principle of original structural division. And if this principle must not be founded in "any ... substratum of an anthropological nature" (152), then it can only inhere in some substrate of a `non-human' nature. I argue that their explanation of gender division as an effect of "imaginary signification" fetishises a metaphysical stuff of `signification' as the ontological basis of gender essences that they deny as having an anthropological basis. The `linguistic' infusion of practices is fetishised into more than an adjectival "aspect" of practices: it becomes the prime mover of practices. In their explanation of `feminine subordination', Laclau & Mouffe conjure noun-like symbolic entities which have verb-like powers to enact effects: an "imaginary signification produces concrete effects"; a "general category inform[s] the ensemble of significations"; a "symbolism ... plays a primordial role". I argue that these entities assume the proportion of an Idealist metaphysic. They play no less essentialist a role in explaining gendered power asymmetry than if the term "originative subject" were to replace them all.

Laclau & Mouffe anticipate the charge that their stress on the `discursivity' of practices is Idealist: They argue: "The fact that every object is constituted as an object of discourse has nothing to do with whether there is a world external to thought, or with the realism/idealism opposition" (108). They insist it is wrong to read them as defining a merely "mental character of discourse. Against this, we will affirm the material character of every discursive structure" (108). This makes the same point as their claim, discussed earlier, that "any distinction between what are usually called the linguistic and behavioural aspects of a social practice, is ... incorrect" (107). Their concept of "discourse" accords with what Wittgenstein (whom they quote) defines as a "`language game'": a "`whole, consisting of language and the actions into which it is woven'" (108).

I agree that one can affirm a tangible materiality of worldly objects and actions, regardless of how one `thinks' them, and at the same time appreciate that one always `thinks' objects and actions, in terms of available forms of `linguistic' intelligibility. But this answer to charges of Idealism ignores what I insist is the crucially challenging question, i.e: By what `articulatory' agency do available forms of `linguistic' coherence -- and their wovenness into actions -- come into existence? What is the materialising capacity that, in and across practical activities, makes certain meaningful connections? Laclau & Mouffe run theoretical circles around such questions. Their mere assertions of the historic contingency of given "language games" do not, I argue, address the dilemma put by Hall (cited earlier) that a "theory of `a history without subjects'" is a theory of "language with no speakers". If theory stipulates that social agents have no original capacities to `speak' into existence (originate) the forms of practiced coherence that gain cultural circulation in social history, then given "language games" -- and "rules" of the games -- can only derive their `articulation' from some Idealised substrate of `symbolic' power. I argue that a post-humanist stress on the `discursivity' of practices cannot help but put primary stress on a `linguistic/mental' aspect, thus fetishising a `primordial' stuff of language-in-itself, from which concrete forms of language-in-practice somehow originate into history. I thus suggest that what Pierre Bourdieu (1993) says of Foucault is apt, and applies as well to Laclau & Mouffe:

Bourdieu suggests that, when theory refuses to attribute to social agents a productive capacity to originate both 1) structured forms of social-relational opposition; and 2) cultural products which `signify' those structural formations, it cannot avoid implicating in the latter the structure-originating power it denies the former. If cultural `stretches of language' -- and the rules of differential positioning they encode -- cannot be said to derive from any productive agency of people, then a cultural `stuff' is implicitly fetishised, i.e. treated as if an original ground of rules of ordered regularity. I would add to this that, even if theory stresses tacit, practice-infused modes of `power-knowledge', the `knowledge' aspect of practices -- the episteme -- is given primary stress, and thus treated as if a transcendent universe (substrate) of primordial being: a `paradise of ideas' from whence forms of power relation emerge into social-historical practice.

Thus, a post-humanist stress on "discursive" conditions of structural possibility cannot avoid being every bit as essentialist as theories which explain structural coherence by recourse to essences "of an anthropological nature". In their efforts to avoid the latter, I have argued, Laclau & Mouffe end up explaining gendered power division by implicit recourse to an original principle of gender division, inhering in a fetishised substrate of `primordial symbolism'. Such explanation is as tautological as any Marxist theory of the a priori necessity of given historic transformations in economic class relations (which Laclau & Mouffe explicitly criticise; see pp. 75ff).

Of course, Laclau & Mouffe do not mean to implicate an a priori principle which necessitates a specifically gendered structural power division in social history. They nonetheless do, I argue, despite efforts to the contrary. I now want to argue that, if the aim is to theorise the historically contingent origin of a relational asymmetry such as `feminine subordination', the only other option is to assume a non-specific essential capacity to originate coherence, such as the "meaning-giving" capacity that Laclau & Mouffe also wish to preclude. As Bourdieu suggests (in the above passage), such a capacity to produce "unexpected" (non-necessary; contingent) yet coherent (trans)formations can only be of a mystical `nature', when attributed to "the `paradise of ideas'": e.g. the "immanent capacity suddenly to transform itself" that Hegel attributed to a Self-begetting World Spirit. But as an anthropologically based nature of human biology, such a meaning-giving, cohesion-begetting capacity seems a far less mysterious attribution. In the next section, drawing on Bourdieu's concept of a habitus, I briefly outline this option for a theory of coherent structural formations in social history.

No critical analyst wants to assume that forms of `structural' division, along axes such as gender, class or race, are inevitable historic outcomes of trans-historic essences. For those who construe `female subordination' and/or other relational asymmetries as salient empirical phenomena, there is proper ambition to theorise such asymmetries as non-necessary contingencies of social history. Yet, to the degree that such forms of structural `opposition' are seen as socially extensive and historically durable, they resist theorisation as contingent phenomena. It becomes difficult not to lapse into subtle invocations of a priori principles which tautologically explain the a posteriori appearance of given power asymmetries. I have yet to see a fully satisfying `historicist' explanation of any of the power divisions which critical analysts treat as salient. But I will argue that a neo-humanist recuperation of certain carefully considered `essences' enables more viable historicist explanation than a post-humanist/essentialist rigour permits.

My readings of Laclau & Mouffe's circular arguments and tricky finesses prompt the conclusion that, short of Idealist alternatives, Hall is right to say: "It's just not possible to make history without subjects in quite that absolute way". Post-humanist efforts to define `discursive' conditions of `articulated' possibility can only finesse the question of the articulatory, coherence-originating agency. I argue that there can be no non-essentialist assumption of such agency. And I suggest that if, in trying to avoid all `essences', we find that our theoretical explanations of social-historical phenomena are running in circles, we should consider whether suppression of certain essentialist concepts costs more than it gains. If done carefully, I argue, such a process leads to certain core philosophical starting points, accepted as necessary for theorising the social-historical phenomena we take to have`empirical' saliency. By this, I mean that we arrive at certain concepts about `essences' which we must assume as first principles. We must assume them because, if we do not, we cannot theoretically explain much; but we can only assume them as first principles because we cannot theorise them as derivative of anything else. We can elaborate our conceptions of such `essences'; but we must appreciate that our conceptual elaborations refer to `natures' which are underivable (original), yet necessary for viable explanation.

I further suggest that two rules of thumb apply here. One is that we should be as minimal as possible in adopting essential first principles. They should be embraced only after we have been thwarted in rigorous efforts to derive them as non-original capacities (yet still find them necessary in order for our theoretical explanations of social-historical phenomena not to lapse into slippery circularities). The other rule of thumb is that, in our attempts at derivation, we should know when to stop, and accept that the non-derivable a priority of certain concepts about `essences' is not a theoretical failure, but a philosophical necessity. (For fuller argument about the philosophical necessity and theoretical roles of such starting assumptions, see Strike, 1989: 5-11.)

But a post-essentialist purism knows not when to stop. It either 1) pushes the effort at derivation beyond the possibility of success; or 2) avoids naming and elaborating concepts that it cannot derive (yet still needs in order to avoid circular explanation). Thus, Laclau & Mouffe circle evasively around the avoided concept of an agentic capacity to make articulated coherence. An undesired consequence is that they end up implicating an a priori principle of gender division. But if they had been willing to adopt, instead of reject, the agency of "a meaning-giving subject", this non-specific capacity to make coherence might have facilitated their effort to theorise a historically contingent origination of specifically gendered division.

Bourdieu postulates just such a coherence-making agency, in elaborating his concept of a "habitus". Bourdieu defines "habitus" as a locus of cognitive structures, embodied `within' each social agent. The agent encounters and internalises these epistemological structures in the course of social practices. At more-or-less unconscious levels of disposition, the agent `takes in' these ways of knowing, as schemas for perceiving how to identify and interact as a positioned `self', vis a vis `others', in the social relations of practical situations. We thus might say -- in the vein of Laclau & Mouffe -- that habitus is a constituted byproduct: a `subjective' inhabitation of the agent by structures of differential positioning which infuse the practical social spaces the agent inhabits.

However, Bourdieu's agent is only on one hand a constituted product of received forms of practiced intelligibility. The agent is also, dialectically, a constitutive producer of epistemes which infuse practice. In the course of social practices, the agent not only takes in the cognitive legacies of history, but also (re)creates them. Says Bourdieu (1977):

That is, Bourdieu aligns with the Marx who, in the first of the Theses on Feuerbach (1991), asserts that a proper "materialist" philosophy must conceive "the active side" -- the principle of active construction; the materialising agency in social history -- "as sensuous human activity, practice" (121). In elaborating his own "theory of practice as practice", Bourdieu does not avoid Marx's suggestion of an essentially human basis of the `sensuousness' -- the `discursivity', if you will -- that infuses practices. As far as possible, Bourdieu seeks non-essentialist derivations of the active capacities he attributes to social agents. But in order to conceive how structured forms of practical sense can actively be constructed -- can emerge and transform in history -- Bourdieu finds he must endorse certain core philosophical assumptions about anthropologically-based essences of truly human agents. He says (1992) that his "[p]raxeology" -- his theory of "habitus" as both actively produced by, and active (re)maker of, practiced forms of cognitive structure -- entails

A "true agent", then, is more than a derivative product of received cognitive structures. The agent is also a historical (re)materialiser, a creative transformer, of the forms of tacitly practiced meaning which inform habitus. Laclau & Mouffe define "agents" who, "lack[ing] any essences", can act in history only as constituted byproducts of "discursive structures", and so "cannot ... be the origin of social relations". But Bourdieu's agents are indeed originative. They embody a universal (essential; trans-historic) capacity "continually [to] make and remake" the relational structures which organise practices and identities. It is in their anthropologically-based nature that social agents have capacity to make coherence across dispersed of social practices. Bourdieu recognizes that diverse social life situations place any given agent in multiple and contradictory positionings, even in the relatively homogeneous conditions of early childhood, let alone over the heterogeneous life trajectories of increasingly `postmodern' biographies. But while habitus thus internalises much contradictory complexity, Bourdieu (1977) postulates a capacity of habitus -- which I read as biological -- to make and hold coherence against precarious destabilisations:

Bourdieu's reference to "overdetermination" strikes resonances with Laclau & Mouffe's invocation of "over-determining effects". But I would underscore two key differences. One is that Bourdieu conceives "an essential overdetermination", i.e. the effect of a universal human capacity to make "coherence and necessity" across diversities of "accident and contingency". The agency that induces overdeterminations is that of an "integrative organism": i.e. a capacity organic to the neuro-biology of agents (at subconscious rather than voluntaristic levels). I read Bourdieu to suggest that this integrative agency is both a capacity and a need to make coherence. (But it is not a need for rigidly totalised coherence. Bourdieu defines a strategic flexibility and inventiveness of habitus that thrives on indeterminate complexity within its `closures'; see 1977: 109-111).

The other key difference is that this essential agency to produce coherence does not implicate any principle of specific structural division. The over-determining capacity of habitus to make coherent connections entails no a priori blueprints for specific outcomes in terms of sexual division, labour division, etc. Bourdieu does suggest, however, that the non-specific capacity to forge integrations is at least a key factor (it of course cannot be the only factor) in explaining the origin and maintenance of those forms of social-structural relation which do appear in history. Indeed, the need of habitus for coherence also plays a key role in establishing "equivalences" across the intersections off multiple structural divisions, where incongruent positionings might otherwise overwhelm an agent's practical sense of coherent `self'.

As noted above, the concept of a non-specific agency to make coherence is necessary but not sufficient for explaining specific forms of structural emergence in social history. It is no simple matter to elaborate sufficient additional factors -- empirical and conceptual -- for plausible theories of the historic contingency of salient power divisions. Much more needs to be thought through, which I must leave for later projects. In bringing this paper to a conclusion, I suggest that the non-specific agency of Bourdieu's "integrative organism" is at least a viable start toward a non-essentialist theory of given social-relational asymmetries. It elaborates the `articulatory' agency that Laclau & Mouffe avoid in their discussion of the `discursive' conditions that supposedly enable coherent and cohesive forms of connection across dispersed "elements".

However, in adopting Bourdieu's integrative organism, we recuperate the philosophical anthropology of "the Marx of the Theses on Feuerbach". We agree with Hall that the strict post-humanism of Laclau & Mouffe, Foucault, Butler and others "can only [be] a stopping point on the route to something else", since it is "not possible to make history without subjects in quite that absolute way". I advocate this neo-humanist (re)turn back to Marx and newly forward. I have argued that there can be no non-essential ground for the `articulatory' agency that is crucial to viable theory; and, short of what Bourdieu calls "idealist intellectualism", there can be no non-human ground for this agency. The post-humanist preclusion of even a "meaning-giving subject" leaves no viable option for imagining how anything approaching pervasive `feminine subordination' could ever come about. It leaves no option for stemming theoretical "one-sidedness given to the moment of dispersion". We need to be able to theorise cohesion and dispersion together, in proper balance. We need a viable theory of the formation of complex `structures', within which dispersive volatility is indeed present -- indeed, (theoretically) desirable -- but not to a degree that, as Byatt says, "gets out of hand." We need viable theories of how, in Byatt's words, "connections [that] proliferate apparently at random" still comb out in "coherent plot[s]". Our best option is to reclaim fashionability for neo-humanist `postmodern' theory, through careful elaborations of what Byatt indicates as "deep human desires" for coherence, in the sensuous practical life of social agents.

REFERENCES