[This article summarises the debate, and suggest that the real issue between them is a methodological one. Laclau goes on to specify what he means by an adequate methodology, and rebukes both Miliband and Poulantzas for not developing one].
Poulantzas is right to say that Miliband's whole argument relies on bourgeois notions of facts and reality. What is needed is a theoretical confrontation with Social Democrat views of the state, and a clarification of Miliband's own theoretical perspective. As a result, Miliband is unable to grasp the operation of the state in marxist terms. Miliband is also right to criticise Poulantzas in reply, although he fails to focus on the main issue itself. He suggests that Poulantzas's theoretical commitments lead to structuralism which fails to establish knowledge of concrete states and the ways in which particular states are able to develop a certain relative autonomy -- 'structural super-determinism'. In the second contribution, this critique turns into a charge of 'structuralist abstractionism... a theoretical approach in which an abstractly defined instance seeks its explanatory principles in another, equally abstractly defined instance... the circular procedural game of mirrors in which, finally, nothing has a precise meaning' (92). Circularity leads to the reintroduction of errors such as economism to add actual content: in particular, the power of the state in Poulantzas has to be reduced after all to the power of a definite class.
Miliband did not justify the theoretical basis of his own stance. His argument that he is contrasting social reformist views with empirical findings leads to the question of how he actually defines 'empirical'. The conventional empiricist view is that objects exist outside thought altogether, and can be used somehow to test theories. However, facts are produced by theories or problematics in the first place, which suggests that 'the problems of logical consistency and empirical validity are not substantially different' (93) [the start of Laclau's own drift to idealism?]. If Miliband is attempting to show that the 'facts' of reformist social theory are in contradiction with its own propositions, then his critique is acceptable. However, it looks as if he is also suggesting that some sort of appeal to real objects can also be used -- and this would involve a substantial theoretical difference with Poulantzas and a problem.
Laclau wants to suggest with Althusser that 'theoretical practice takes place wholly on the plane of thought', that theory begins by working on 'concepts, pieces of information and ideas' provided by other kinds of practice. Theory then transforms these raw materials into objects of knowledge. This is quite unlike empiricism, which assumes that knowledge of real objects can be generalised into theory. Techniques and methods of verification are also part of the same theoretical system, thus 'a theory is only false to the extent that it is internally inconsistent' (94). It follows that theoretical problems cannot be falsified by reference to real objects, but they can be superseded. Problems are either finally resolved within the general assumptions of a theory or they lead to a genuine contradiction in a theory. In the second case, the only way forward is to develop a new theoretical system [sounds exactly like the difference between normal and revolutionary science in Kuhn]. However, strictly speaking, a new theoretical system cannot solve the problems of a former one, but supersedes them [after some unspecified theoretical and political struggle?].
An adequate theoretical critique would begin (a) by pointing out the conflict between the empirical and the theoretical system It would then proceed to (b) identify the precise nature of the theoretical problem indicated by the conflict; (c) to demonstrate internal theoretical contradictions which a system cannot resolve; (d) to propose an alternative theoretical system which can 'overcome' [supersede?] the contradictions (95).
Miliband remains at step (a), and Poulantzas wants him to proceed to the other stages. Until he does, he cannot thoroughly challenge social democratic theory. Indeed, he must reproduce some of it, especially the view that the motivations of social actors are important aspects of social change. Poulantzas also makes mistakes, however, and also fails to challenge in detail the theoretical contradictions of his rivals (economism, historicism and the like), although he does describe some of the differences with his own problematic. Again, it is not enough to show that there are different views: the point is to show how one can supersede the other: Poulantzas operates only at (d).
In this sense, Miliband is right to criticise Poulantzas's concepts. It is not actually the case that Poulantzas cannot explain the relative autonomy of the state -- relative autonomy could be 'the result of a particular articulation between the instances corresponding to the mode of production under consideration' (98) [more or less exactly what he does argue in his final reply]. It is Miliband's underlying concept of autonomy that Poulantzas cannot use, because Miliband's concept is still rooted in some notion of the subject: '[for Miliband]... the adjective "relative" constitutes a simple restriction to an autonomy conceived in terms of freedom. For Poulantzas... the "relative" character of an autonomy indicates that it belongs to a world of structural determinations... as a particular moment of it' (98). What Miliband takes to be a central problem -- how the people occupying apparently separate institutions in fact share a common social origin -- is a minor problem for Poulantzas, who is much more concerned with the general theory of the state in a capitalist mode of production, not a specific case.
Miliband is right to criticise the way Poulantzas thinks about the ideological state apparatuses. Poulantzas does tend to see everything in terms of contributing to the cohesion of the social formation. If all ideological apparatuses must do this, so must everything else -- trade unions, political leaders, socialist parties and even 'the mind of every individual' (100). Althusser argues, slightly differently, that the state itself defines the boundary between private and state apparatuses, but the underlying functionalism is still apparent. Given this general role, it is hard to think of the state 'as an instance... [and it]... must simply be a quality which pervades all the levels of a social formation' (101).
Poulantzas does head towards 'formalism, as a result of which the theoretical substance dissolves into a system of verbal antinomies' (101). Form predominates over content. As content diminishes, 'the symbolic functions of that concept within the discourse tends to increase' (102) [all scholastic marxism heads this way, even Laclau?]. Connections between concepts come to dominate an argument -- either the logical relations between concepts, or more metaphorical notions of 'proximity between different concepts' (102). In the latter case, a concept can evoke a theoretical system, or become a symbol of it. This is what describes formalism, which often stresses conceptual taxonomies.
Poulantzas resorts to taxonomy in the face of complexity. His concepts are so abstract that they will lose their content and become symbolic. Perhaps this is because the whole theoretical approach began with the intention to be descriptive. Poulantzas does try to reintroduce historical reality, 'but this result has been achieved despite, not because of his method' (103). Miliband's criticism of the argument about pertinent effects is an illustration. On the one hand, a class becomes distinct and autonomous if it can demonstrate pertinent effects in a social formation, defined as some new element appearing in the typical framework. However, these pertinent effects must also be politically ineffectual, that is lacking a particular impact.
Formalism extends to some of the most central theoretical concepts, such as the 'mode of production'. Poulantzas develops the idea along with Althusser and Balibar [at this stage, see the rather poorly reproduced diagram my colleagues and I have used to illustrate this model of the mode of production].
It would be better to reformulate the argument more explicitly, which
is what Balibar claims to be doing with Marx's original texts, which
are still contaminated by it older ideological problematics. However,
the concepts developed in the model are inadequate. Returning to Marx's
text [Capital Vol III in this
case] suggests a distinction
between 'the economic' and a 'notion of
"extra-economic coercion"' (106). In pre-capitalist economies, this
extra economic coercion takes distinctively political or ideological
forms in order to regulate production. In capitalism, however, labour
has been turned into a commodity, and thus economic forms -- markets,
commodities -- can regulate production or on their own. We need to keep
this distinction -- the outer determining 'economic' refers to
the general theory of historical materialism, which takes a specific
form in capitalism. Marx's achievement was to demonstrate this
theoretical unity, but there is no need 'to use the same
expression to designate both' (106).
Poulantzas does not clearly distinguish these two notions, and thus can
only operate descriptively and symbolically, and use misleading terms
such as 'structural causality'. The ambiguity and neologisms
address 'an artificial problem created by the metaphysical of
instances' (107). For Laclau it is much simpler: 'the economy' in
the general sense of the production of material existence is simply
never determinant even in the last instance, while 'the economy'
in the sense of production of commodities is determinant when it
'becomes identified with the basic productive relations of that
society' (107). Similar clarification is required for the names of the
other two levels. In particular, the separation between the economic
and the political in Marx is characteristic only of capitalism, and can
only really be used to indicate the differences between capitalist and
pre-capitalist social formations: it is not a full theory of other
modes of production.
Similarly, the problem of relative autonomy is a specifically
capitalist one, and cannot be discussed purely abstractly. Abstract
discussion leads to a logical difficulty about articulation, turning on
how separate the levels are in the first place. If they are completely
separate, is only really the model which unites them -- 'verbal
unity' (108). If, on the other hand, they have some elements in common
to permit specific kinds of articulation, this needs to be theorised,
not just glossed by 'a symbolic concept -- "determination in the
last instance" -- which lacks a precise theoretical content' (109).
Poulantzas needs to winnow out descriptive categories and identify
theoretical concepts which can be linked to produce specific modes of
production. Thus Miliband is right to criticise 'structuralist
abstractionism'. Nevertheless, Poulantzas offers much more of a way
forward in grasping the specificity of the capitalist state.
on to Poulantzas's final contribution