|Harold Wolpe Memorial Trust
|88 Belvedere Rd
|PO Box 44907
|Tel: +27 21 674 0361
|Fax: 086 670 6772
Wolpe H (1980) ‘Towards an analysis of the South African state’, International Journal of the Sociology of Law, 8:399 – 42
Towards an analysis of the South African state
Over the, past few years, the problem of developing an adequate approach to, and substantive analysis of, the role of the state has become a central concern in some Marxist accounts of the South African social formation. In fact, two apparently radically different strands have emerged in these Marxist analyses but both expressly set out to depart, firstly from a simple instrumentalist approach with its economist undertones and its reduction of the state to a “thing/instrument” available to and wielded by an undifferentiated dominant capitalist class and, secondly, from the identification of the state with government. These attempts to provide an analysis of the South African state have led to the introduction of concepts which are drawn, in the one case, from the work of Nicos Poulantzas and, in the other, from the “state derivation” approach of some German writers. (See, for example, the essay by Blanke, Jurgens and Kastendiek in Holloway and Picciotto, 1977.) In both approaches, albeit in different ways, the attempt to overcome the limitations of the instrumentalist approach is conducted through a discussion of the social bases of state power, the relation of the state to capital accumulation, the autonomy of the state from the dominant classes, forms of representation in the state, the form of state and so on. On the basis of such considerations each approach claims that it is presenting a more adequate and more complex explanation of the functioning of the South African state than has hitherto been available. There is no doubt that the various contributions to the debate have provided penetrating analyses of particular episodes in South African history and have sharpened some of the issues pertinent to an understanding of the South African state. Paradoxically, however, this has been achieved despite the fact that in this literature, very little advance has been made towards formulating a more adequate framework for the analysis of the state.
The central problem in this literature is the failure, on the one hand, to develop concepts, which would enable us to grasp the structure of the state itself as a differentiated and contradictory unity and, on the other hand, properly to conceptualize the relation of classes to the state. The effect of this is not only the reintroduction of instrumentalist concepts of the state but also, and this is the fundamental point in the contemporary context, an extremely restricted economistic view of the parameters of the fundamental political conflicts in South Africa.
The object of this, paper is to subject the Poulantzian and “state derivation” based work to a critical examination in order to open the way towards a radical reformulation of the theoretical means through which the substantive analysis of the South African state and politics may be more fruitfully approached.
Fractions of Capital and the Relative Autonomy of the State
The first systematic attempt to theorize and account for the changing role of the state in the specific conditions of capitalist development in South Africa is to be fougid in a body of literature which bases itself on the work of Poulantzas .
It is important to note, however, that although the concepts purport to be drawn from Poulantzas, in the neo-Poulantzian South African writing they are not always given the same meaning as he attributes to them nor are they brought into relationship with one another in a way which is strongly suggested by this argument. This applies in particular to the concept of the “form of state”.
In Kaplan (1977), Davies et al. (1976)and Davies (1979) much importance is attached to the concept “form of state” and indeed, a periodization of the political is developed on the basis of changing forms of the South African state. For the neo-Poulantzians the form of state is conceptualized in terms of the concept of the power bloc. The power bloc is conceived of as the contradictory unity of the dominant classes or fractions in a social formation. The power bloc is the site of struggles between the dominant classes and fractions and its unity depends on one class or fraction becoming hegemonic. As Davies expresses it:
The hegemonic fraction is that fraction which has become, through struggle within the power bloc, (1) the most dominant of all the dominant classes and fractions and (2) the fraction which has been able to set up its own particular interests as the “general interests” of the body politic (Davies, 1979, p. 28).
Now, according to Kaplan (1977) and to Davies et al. (1976) the form of state is to be understood in terms of (a) the composition of the power bloc, and (b) the class or fraction, which is hegemonic in the power bloc. That is to say, the form of state is defined by these elements of the power bloc.
The changes with regard to the composition of the power bloc and/or hegemony, denote changes in the form of state (Kaplan, 1977, p. IV).
Thus, since the form of the state is derived directly from the class composition of the power bloc and/or the identity of the hegemonic class, the question of the structures, institutions and so on which compose the state as such are not made the object of investigation. The effect of this, as I will show, is to reduce the state to an instrument of the hegemonic class which is precisely the conception Poulantzas wished to avoid.
Despite the limitations of Poulantzas’s work and, in particular, the slide between a “structuralist” and an instrumentalist conception of the state [which, however, he began to correct in his later work. See NLR, 95, and State Power and Socialism (1979) ] his concept of the state as structure does not permit a simple reduction of the form of the state to the relation of classes in the power bloc. This is made clear. in the following passage:
Institutions, considered from the point of view of power, can be related only to social classes which hold power. As it is exercised this power of the social classes is organised in specific institutions which are power centres: in this context the state is the centre of the exercise of political power. But this does not mean that power centres, the various institutions of an economic, political, military, cultural, etc., character are mere instruments, organs or appendices of the power of social classes. They possess their autonomy and structural specificity which is not as such reducible to an analysis in terms of power (Poulantzas, 1973, p. 115).
Fundamental to Poulantzas’s argument is the notion of the relative autonomy of the state from the dominant classes in capitalist social formations. In one line of his analysis, Poulantzas bases this autonomy on the distinction between structure and class practices. It will be recalled that for Poulantzas, classes can be conceptualized only in struggle, that is as practices. Those practices, however, are determined by an articulation of economic, political and ideological structures (the structural determination of classes), which do not themselves entail practices.
How is the structural determination of classes conceptualized in relation to the political structure to the state?
Firstly, it must be noted that Poulantzas distinguishes the state as structure from classes and class practices:
... the juridico- political state superstructure ... is not the same as the political class struggle ... since the state apparatuses are no more classes than are relations of production (Poulantzas, 1973, p. 18).
Now, the form of the relationship between the state as political structure and social class is identical to that of the relationship between the economic and the ideological structures and classes, although, of course, the content of that relationship differs. The general form of the relationship between structure and class is expressed by Poulantzas in the following way :
... the concept of power specifies the effects of the ensemble of these (i.e. structural) levels on the relations between social classes in struggle. It points to the effects of the structure on the relations of conflict between the practices of the various classes in “struggle”. In other words, power is not located in the level of the structures, but is an effect of the ensemble of these levels ... (Poulantzas, 1973, p. 19).
It follows from this that for Poulantzas the specific form of the state is a question which has to be determined at the level of the political structures. By contrast, questions concerning the composition of the power bloc, the relations of hegemony and the struggles within it, have to be considered at the level of class practices. These latter phenomena are, in Poulantzas’s argument, conditioned by the structures . Thus for Poulantzas, the form of the state and the power bloc are seen as clearly separate yet articulated “spheres”. In no sense is it possible, in his analysis, to reduce the form of the state to the composition of or to the relations of hegemony within the power bloc. The effects of the political structure on the class struggle are related to the specific form of the state, in particular, to the concrete distinction and relationship between the executive and legislature, but also to the institutional and organizational character of the various state ideological and repressive apparatuses and the relationship between them.
If this is understood, then the formulation of the relation between power bloc and the state apparatuses as a simple instrumental one, becomes extremely problematic. On the other hand, to the extent that the state is assumed to constitute a simple unity, it becomes the basis for instrumentalist conceptions of the relations between classes and the state. The latter is precisely the consequence of the concept of the form of state utilized in the South African work, a concept in fact which precludes the analysis of the form of state as such. Thus, it is argued, the coexistence of the dominant classes which comprise the power bloc is “ensured and organised through the capitalist state” (Davies et al. 1976, p. 5). However,
... the state and state power cannot be divided and parcelled out amongst the classes/fractions within the power bloc. State power is a unity, requiring organizational direction to be effective. Thus, there is always a struggle within the power bloc to assume this organizational role, and thereby ensure the primacy of this class/fractions particular interests. (Davies et al. 1976, p. 5). (Emphasis added).
That is to say, the state as unity is wielded by the hegemonic class as the instrument by means of which its particular interests and the cohesion of the power bloc are simultaneously achieved. It makes no difference in this regard that, in the historical accounts of state intervention in the development of South African capitalism, the roles of specific state institutions or organizations are focused upon (e.g. Board of Trade, Industrial Conciliation Tribunes etc.) depending oil the sphere of inquiry, since each of these is nothing other than a functionally differentiated part of the state conceived as a unified instrument.
Indeed, it is in the historical analyses, produced by the neo-Poulantzians, that the relationship between the dominant classes/fractions and the state as instrument receives its clearest expression. Thus, in Davies et al. (1976), the periodization of the state is presented through an analysis of the changes of the hegemonic class/fraction within the power bloc. It is argued that foreign (mining) capital favoured a policy of free trade whereas national capital supported a policy of protection. In the periods in which mining capital was hegemonic, the state maintained a policy of free trade and, by contrast, when national capital was hegemonic a policy of protection was pursued. Indeed, the hegemony of each fraction is “proved” by reference to the state’s trade policy and it is clear that, whether explicit or not, the state is conceived of as the instrument of the interests of the hegemonic class.
Similarly, in Davies (1979), an account is given of the manner in which the _ state is utilized by-the hegemonic class to ensure, inter alia, “... assignment of white rather than black agents to particular places in the wage earning classes ...” (p. 32). And Kaplan (1977) provides a comparable analysis of the state’s intervention in the economy.
How are we to understand the mechanism by means of which the hegemonic class in the power bloc exercises its directive role over the state? Davies et al. (1976, p. .13-14) draw a sharp distinction between the concept of form of regime which encompasses the representation of class interests through political parties and the form of state. The. latter is a wider concept which governs the “whole field of political class struggle” (ibid., p. 14) and, of this, the relations between political parties constitute only one aspect. Nonetheless, the analysis of the state is restricted, almost exclusively, to the party political scene. The argument runs along familiar lines. Classes are represented in the political scene by political parties which pursue conflicting policies. The party which represents the hegemonic class/fraction constitutes the government. The policy it implements may be moderated by the pressure of other parties. Government ‘policy is realized through the instrumentality of various state apparatuses. In this analysis, the state tends to be reduced to the government and the apparatuses are conceived as its simple instruments.
It is true that Davies et al. (1976, p. 15 ff:) point to the fact that a dislocation may occur between the state and the power bloc, but the state is conceived of merely in terms of the government and the party political representation of the hegemonic class./fraction:
The Herzog Government’s stubborn refusal to abandon the (gold) standard when all fractions of capital were strongly advocating such a policy, represented a disjuncture between the generalised interests of the power bloc and the State and between the interests ofthe hegemonic fraction and their party political representation (ibid., 3, p. 16-17). .
In this formulation, the disjunction between state and power bloc is overcome by re-establishing “... the identity of the interests ... of the hegemonic fraction with their political representatives” (ibid., p. 17). Davies et al. fail to analyse any political practice or struggle outside parliament and government. Thus, the state as a complex and contradictory structure, of which the “political scene” is one element, is never analysed. The total explanation of the forms of intervention of the state are given in the relations within the power bloc and the relation of the hegemonic class/fraction to its party political representatives. The totality of state power is reduced to political representation.
The concept of representation of class interests used in this approach raises another set of problems. The concept of class utilized in this work is again purportedly derived from Poulantzas . In Poulantzas, classes are constituted primarily at the level of the relations of production but include also political and ideological relations. The structural determination of classes is to be distinguished from the political position which a class may take up in any particular conjuncture. Class position is not the automatic outcome of class structural determination but is conditioned by conjunctural factors. However, in Davies et al. (1976) and Davies (1979), a straightforward economism is immediately introduced into the concept. In both these texts, classes are endowed with a unity and a commonality of interests which is pre-given to the political and ideological and has its source in the sphere of the economy. This is so, despite several formal assertions to the contrary. Firstly, the contention that the ideological and the political are elements of the structural determination of classes is shown neither at the theoretical level nor at the level of the historical analyses. Secondly, the distinction between structural determination and class position is asserted, but either not specified, or simply ignored in the historical analyses. Thirdly, and perhaps most importantly, the argument that classes are not “reproduced as a unity” is contradicted both at the theoretical level and in the historical accounts.
In relation to tile latter point, it is contended in Davies et al. (1976, p. 5) that:
Within capitalist social formations, however, classes are not reproduced as a unity, but arefractured and divided. Several dominant classes co-exist (dependent on the articulation with other modes) and, more critically, the dominant class is itself divided into several fractions (resting on their differing roles in the expended reproduction of capital).
But this is immediately followed by:
The dominant classes and fractions share a common interest in the maintenance ofthe relations of exploitation in general, but simultaneously have contradictory interests corresponding to their particular place in the relations ofexploitation.
Again Davies (1979, p. 7) argues:
It is the objective place of a class in the social division of labour which establishes its objective class interests and thereby marks out the boundaries of its class practices.
What this means is that the common interests of a class or fraction are given at the level of the economy and become reflected in the political/ideological practices which have as their objective the reproduction of the pre-defined economic interests. For example, in discussing the distinction between foreign and national capital, Davies et al. (1976, p. 5) state:
... the significance of political/ideological criteria should not be underplayed. Reflecting their base in different sectors of production it is here that the English/Afrikaner traditions are to be located ... (First emphasis added).
This economism has extremely important consequences. In the first place if the interests of classes are conceived as given at the level of the economy, itself operating free of ideological/political relations, then the specific structures of the political and ideological become submerged to the point of obliteration. Since these structures exist as nothing other than the reflection of the economic and as the instrument of classes whose interests are defined at the level of the economic there is no need to analyse them in terms of their forms of organization and contradictory relations. In part, at least, this accounts for the reduction of the state to a simple instrument of economic class interests and the failure to analyse the structures of the state as such, which I discussed above.
In the second place, although this may not be the inevitable consequence of the economist conception of class which is employed, exclusive attention is paid, in the historical analyses, to the state’s intervention in the sphere of the economy. That is to say, whatever point in the political sphere the analysis is concerned with - party political scene, the power bloc or the forms of state interventionism, the focus is entirely on such questions as the allocation of agents in the division of labour, free trade versus protection, taxation, economic crises of the depression and the problem of the gold standard. This is not to suggest hat these and similar questions are not of fundamental importance to an understanding of the trajectory of capitalist development and political struggles in South Africa. What is astonishing, however, is that there should be no analysis of (i) the structures of the state and (ii) the central political struggles revolving around the national question, the Afrikaner/English conflict and so on. Is it possible to understand political interventions in the economy apart from an understanding of the ideological and political structures which explain the way in which the economy is grasped at the level of class struggles?
Capital-in-General and the State
The analysis of the South African state which is based on Poulantzas has been subjected to a critique from the standpoint of the concept of capital-in-general (Clarke, 1978; Innes & Plaut, 1978).
The starting point of the attack is the radical separation between economics and politics which is the foundation of the Poulantzian accounts. According to Clarke [all page references following are from Clarke (1978) unless stated] in the Poulantzian approach, the economic level is conceived as a sphere in which individual economic interests are pursued. Economic interests are those defined by the factor of production which is a source of income. Classes are constituted on the basis of, or are defined in terms of a coalescence of such individual economic interests. However, although defined in this way, classes “... only appear at the political level.” That is, classes or class fractions only become a social reality at the level of the political. The link between the economic and political is established through the economist notion of the political representation of economic interests:
Hence the relationship between the economic and the political is necessarily treated as the simple and direct representation of economic interest by political groups (p. 38).
Clarke draws out the implication of this position. Firstly, capital, the capitalist class, is conceived as a collection of individual capitals each of which is constituted independently at the level of the economy. As such, capital has no unity “prior to the definition of particular capitals” (p. 37) and indeed comprises, at the economic level, not a social unity but rather a mere aggregation of capital. Secondly, it is precisely at the political level that the reality of classes and class fractions is created. However, the political organization of classes and class fractions is based on the conflict between capitals and is incapable of constituting the “unity of capital as a whole” (p. 45). It is at this point that the role of the state enters into the fractionalist account.
The state is the factor of unity of capital as a whole. It follows that outside of the political level, there is no capital-in-general (pp. 46, 47).
Clarke argues that a number of consequences follow from this. First, the reduction of the state to an institution representing and expressing economic interests constituted at the economic level, involves, on the one hand an economist reductionism coupled with an “overpoliticization” of social processes and, on the other hand, the explanation of the states’ role in functionalist terms - the states’ intervention is explained by the function it serves for capitalist interests defined at the economic level. Secondly, by abolishing the notion of capital-in-general except in so far as the state functions to constitute the unity of capital as a whole, the fractionalist approach “condenses everything in the state” (p. 45) and is thus unable to explain the limits within which the state is constrained to operate by capital-in-general. Thirdly, the state is deprived of any autonomy in relation to capital-in-general, despite he elaborate rhetoric of autonomy in relation to particular capitals.
The fractionalists’ emphasis on conflicting capital fractions, on the one hand, and on the conceptualization of the state as expressing and uniting capital as a whole, on the other, leads, Clarke argues, to two further criticisms of their approach. First, it concentrates on the conflicts between capital fractions rather than on the fundamental capital-labour conflict. The state, thus, is presented as the factor of unity of capital fractions and not as another form of the capital-labour contradiction (Innes & Plaut, 1978). Secondly, the question of which fraction of capital is hegemonic becomes central to the explanation ofthe form and policies of the state. But establishing which fraction is hegemonic, is rendered problematic on two grounds. For one thing, according to the fractionalists’ account, in the period of monopoly capital the power bloc becomes monolithic, due to the interpenetration of capitals and it becomes impossible to distinguish separate fractions of capital with identifiably conflicting interests. For another thing, there are no clear criteria of hegemony and the argument tends to become tautological or to reduce to an identification of hegemony with election victories (Clarke, 1978, p. 15).
As against this, Clarke offers a set of propositions which, it is claimed, are adequate to explain the actions of the South African state and which avoid the rigid separation of the economic from the political and the consequent slide into economism, over-politicization, and functionalism. The possibility of escaping from these rests, for Clarke, on the uncovering of the connections between the economic and the political.
The starting point for Clarke, is the concept of capital-in-general. Neither individual capitals nor capital fractions are conceivable as capitals outside of an analytically prior concept of capital-in-general and class relations (p. 41).
The concept of the capital-labour relation, however, does not merely provide the necessary analytic condition for the identification of individual capitals or fractions; it is at the same time the analytic precondition of the analysis of the form of existence of class relations :
... Class relations are the necessary starting point for a Marxist analysis, without which re1ations of exploitation (as the economic form of these class relations) or the relations of domination (as the political form of these relations) cannot be conceptualized. It is the concept of class relations as being analytically prior to the economic, political or ideological forms taken by those relations (even though class relations have no existence independently of those forms) that makes it possible for a Marxist analysis to conceptualize the complexity of the relations between economic and political, their interconnections as complementary forms of the fundamental class relation, without abandoning the theory for a pragmatic pluralism (p. 42).
It is to be stressed that the concept of the class relation, specified only as the capital-labour relation, is the analytical starting point. Yet the capital-labour relation only exists in specific economic, ideological and political forms. Now, presumably, it is unthinkable that Clarke would wish to adopt the idealist position that the forms of existence of a class relation may be derived from the concept of that relation. Nor, presumably, is it open to him to argue that the capital-labour relation as a real existence at the level of capital-in-general determines the specific economic, ideological and political forms in which the relation exists or appears. To argue this would amount to an economism and would make it impossible to account for the different economic, political, and ideological forms in which class relations exist at different times and places in capitalist social formations. It follows, therefore, that even if we accept, as I do, Clarke’s specification of the starting point, that starting point does not, of itself, provide the means for analysing the different forms of existence of the class relations let atone of other social relations which may be pertinent to but not reducible to the relationship between classes. The question that we then have to pose is whether or not Clarke provides a means for analyzing the political forms of existence of class conflict which does not reduce the political merely to the forms of representation and implementation of interests established in the economy.. The question can be approached through Clarke’s analysis of the state and involves both the relation of capital-in-general and that of particular capitals to the state.
For Clarke, the fundamental question is “... how capital-in-general imposes itself on the state?” (p. 63) . At first sight this arises because the state is simply a specific institution that has abrogated to itself certain powers and that is inserted in a particular society (p. 64).
However, the passage, it seems, has some other meaning because Clarke continues immediately:
As such the state is subordinate to the domination of capital that characterizes that society, adomination that does not have to be produced anew at the level of the state since it already exists … (p. 64).
Now, clearly this is somewhat mystifying since if the state is already subordinate to capital, it would seem that the question of how capital imposes itself on the state is superfluous. Nonetheless, Clarke pursues this question for, despite the above passage, the domination of capital-in-general has to be secured. The first point tonote, however, is that, according to Clarke, capital-in-general does not itself actually “choose” instruments to secure its domination. Indeed, it only secures its domination through the action of particular capitals (see p. 64).
Interestingly enough, this appears to be remarkably close to the fractionalists formulation except in so far as Clarke adds that capital-in-general makes it possible to theorize the limits within which specific capitals are able to struggle successfully, vis-à-vis the state, for their particular interests. The limit is given
... by the role of surplus value in the expanded reproduction of the capitalist mode of production (p. 55).
It is difficult to reconcile the proposition that capital-in-general can only secure its dominance (which presumably includes the imposition of the limits in question) through the intervention of particular capitals with the further proposition that it is capital-in-general itself which limits the extent to which particular capitals are able successfully to press their demands. At one point Clarke seems to attempt to by-pass this difficulty by asserting that while capital-in-general exists only at the interrelation of particular capitals, nonetheless, the former must not be understood “as distinct from those capitals …” (p.53).
What the relation is between the analytical priority of capital-in-general, the differentiated forms of particular capitals, and the social whole, remains, therefore, obscured. Yet the fact remains that particular capitals cannot be reduced to capital-in-general (p. 54) and it is through the action of particular capitals that the relation between particular capitals and capital-in-general can be established. That .is, too say that relationship can, only be established concretely (p. 56).
Thus far, Clarke’s argument amounts to this: within the limits established by the valorization process, individual capitals press the state to act in their immediate, particularistic interests and the state’s interventions have effects upon capital-in-general. Clearly, this line of reasoning does not take us beyond the instrumentalism and economism which Clarke set out to overcome, the state acts at the behest of particular capitals and this has effects for total social capital.
Yet Clarke does not maintain a consistent instrumentalism, for in his actual analysis of the South African state, he also sets up the state as the constitutive source of particular policies. He argues that where there are barriers (as, in the case of gold-mining, barriers set by physical accessibility and large capital requirements) to the entry of competing capitals into a particular branch of the economy, state policies which favour that branch will also favour the capitals in that branch. In these conditions, the state may come the instrument of a particular capital (or of opposing capitals):
For this reason the temptation for mining houses to use the state to secure particular advantages was genuine, while the attempt to do so generated the political opposition to mining capital that has limited their access to state power for over fifty years (p. 60).
On the other hand, where there were no barriers limiting the entry of other capitals into a branch, measures which favour that branch will be self-defeating as far as the particular capitals are concerned as these measures will attract other capitals, hence:
We would then expect it to be more likely that such measures were introduced by the state in order . . . to attract capital to the favoured branch in order to alter the material relationship between branches, perhaps because of the failure of the competitive process to achieve this (p. 60), (Emphasis added).
In the first case, the state is the instrument of particular capitals; in the second, it acts in terms of its own conceptions of the interests of capital-in-general. In neither case, furthermore, does Clarke’s conception of the state accord any effectivity to the political other than as the instrument of a dominant class or capital-in-general and then only in relation to interests already established in the economy.
In fact, Clarke does refer to certain institutional structures of the state, the system of representation, the administrative structure and the “tax state”, but primarily in order once again to reiterate that these institutions arc the means by which capital asserts its domination over the state and that it is able to exercise its domination because its dominance has already been established at the economic level (Clarke, p. 65). Only in respect of the system of representation does he hint at specific political institutional contradictions:
At the same time, the forms of popular representation express the domination of capital, in a contradictory form: while they provide a mechanism that can create the political and ideological conditions for the execution of particular policies, they can only do so at the expense of providing access to the state apparatus, of a limited kind, for groups which may carry their resistance to the state and to the rule of capital in that apparatus (p. 65).
Despite this, Clarke’s insistence upon the state as nothing more than an expression of an already constituted dominance and his restriction of the state to an instrument for the management of political contradictions which grow directly and exclusively out of the accumulation process, preclude the type of analysis towards which the above passage points. Instead, Clarke is driven to characterize the recently developed institutional forms of political domination merely as sterile and, as such, more or less irrelevant to the political conflicts in the social formation. This is shown quite clearly in his analysis of the crisis in South Africa. His account focuses on the South African state’s problem of regulating the balance of payments and the necessity for manufacturing industry to become competitive on the export markets. This involves the imposition of a tight work discipline and an intensification of the labour process. This in turn, intensifies the class struggle at the point of production and in relation to the reserve army of labour. These changes in the economy produce a crisis in the political sphere:
The specifically political problem the South African state confronts domestically is that of containing the political struggle generated by the intensification of class conflict in the economic sphere. On the one hand increasingly open and pervasive repression is being used to prevent the class struggle from taking a political form. On the other hand, to the extent that the development of political struggles cannot be prevented, the state is seeking to deflect and divide the political initiatives that emerge. It is here that the South African state is attempting to make limited “concessions” to the black working class, creating sterile forms of black political representation in the Bantustans and in the urban ghettos and attempting to make selective concessions, especially to those with permanent urban residence, in order to intensify the divisions in the working class that the recent struggles have begun to overcome (p. 71).
This analysis is notable in a number of respects. Firstly, its economism: the political conflicts emerge directly out of and as the direct expression of conflicts at the level of production. Clarke totally ignores the political conditions (existing institutional structures, political organizations, etc. and the ideologies which inhere in them) in which the economic changes are occurring, changes which, of course, are themselves conditioned by the political/ideological, and which affect the manner in and extent to which changes in production become defined politically. To read the above quoted passage is to be left with the impression that the politicization of the workers struggle emerged for the first time in the 1970’s as an emergent property of a new economic situation.
Secondly, for all the talk of the economic, political and ideological forms of existence (and presumably struggle) of particular classes, Clarke remains always at the level of capital-, and indeed, labour-in-general.
Thirdly, as I have already argued, Clarke pays no attention to the effectivity of the specific political institutional forms of the state and his analysis is restricted to the contradictions in the economy. The state appears as the unified and harmonious instrument of capitalist domination, despite the earlier reference to contradictions within the state.
These positions flow from Clarke’s failure to define the limits of the conception of capital-in-general and the general concept of the capital-labour relation. While he acknowledges both the different forms of existence of class and the differentiated political institutional structures of the capitalist state, these always become absorbed back into the capital-labour relation in general.
Thus, capital and labour appear only as already unified entities and the differentiated forms (economic, political and ideological) in which they exist are seen to be functionally complementary to one another. In the same way, the differentiated state organizations and institutions are reduced to the simple instrument of capitalist rule.
So, despite very different starting points, the fractionalists and Clarke lead us into the same difficulty, namely a failure to specify the arenas of political struggle. Such a specification involves breaking from economism and establishing the contradictory relations and contradictory trajectory of the political and ideological as well as the economic.
Towards an Analysis of the South African state
My object in the preceding sections was to identify two basic inadequacies which are common to different approaches to the analysis of the South African state. These common inadequacies are, on the one hand, a conception of the state as a simple political instrument of power already established in the economy (although, as I have already pointed out, the state is, quite inconsistently, also presented as the autonomous regulator of the social formation), and, on the other hand, related to this, an economic reductionism which frequently slides into a functionalism. Although these economic features take on different forms in each approach, nonetheless, they have similar analytical and, in certain respects, political effects.
The homogenization of the state into a simple, integral instrument and the reduction of politics to economics has these effects. Firstly, it denudes the state of its complex, contradictory, specificity and denies the possibility that the state structures are themselves the site of political struggles and class conflict. Secondly, it, therefore, places all struggles outside of the sphere of the state structures. Thirdly, and as a consequence, class power is seen as first being constituted in “society” and only then does the dominant class deploy the state/instrument to give effect to its interests. The state thus enters the class struggle as the instrument of the already dominant class. Now, not only does this render the state entirely unproblematic, it always operates to give effect to the interests of the dominant class, and therefore, unnecessary to analyse, except in so far as its interventions are described, it also defines a particular conception of the political struggle. The objective of the political struggle is to take the state/instrument out of the hands of the dominant class, but since the state apparatuses (however conceived) do not constitute the sites of struggle, that struggle takes place only in “society”. The state is exclusively the objective of the class struggle, it is never a site of that struggle.
It is obvious from the above, that the elaboration of an alternative approach calls for a discussion of the relationship between economics and politics, a conceptualization of class, of the relation of classes to the state and ofthe nature of the state. In this concluding section, I will touch on these issues in a rather peremptory way and only in so far as is strictly necessary in order to point in the direction of a more adequate approach to the analysis of the South African state.
Economism and economic reductionisrn appear within Marxism in variant forms. Of central importance in the present context is that variant of economism which asserts that the constitution (definition) of classes in terms ofrelations of production carries with it two further necessary attributes: (i) that the objective interests of classes are given at that level and that classes appear as social forces unified by their place in the relations of production and pursuing those objective interests ii) that class power is constituted in the economy and that state power is a mere appendage of class power.
The analytical starting point for a Marxist analysis of class struggles in capitalist social formations is the concept of the relation of production between capital and labour-power in the capitalist mode of production. It must be immediately emphasized, however, that it is not possible to assume on the basis of this starting point, that classes necessarily appear as unified social forces or that the specific economic, ideological and political content of class struggles is derivable directly from the specification of the relations of production. That is, the unity of a class as a social force is not given at the level of its definition in terms of the relations of production, although the differentiated forms in which classes exist can only be understood in relation to those relations of production. To put this in another way, the relations of production provide the necessary but not the sufficient conditions for the analysis of class struggles. Similarly, it is important to add, while the unity of a class is also not given at the level of capital-in-general or of total social capital, nevertheless the organization and struggle of classes can only be understood in the context of and in relation to the determinate conditions of capital-in-general - relations between different departments and branches of production, balance of payments, rates of profit and so forth.
It follows from this, that the internal unity of classes cannot be presupposed on the basis of their definition in terms of relation of production or on the basis of total social capital. Classes appear in multiple forms of organization. Classes, that is, exist in the form of economic organizations or apparatuses (for example, the organization of workers in the division of labour in capitalist enterprise; the associations of capitalists in particular branches of industry); in politico - economic organizations (e.g. Confederations of industry, trade unions); in political bodies (political parties, campaign organizations, state apparatuses) and so on. It is clear that these forms of class organization will be differentiated on a number of bases - political, ideological, functional and so forth. Classes defined at the level of relations of production do not constitute a unity which is represented by these forms of organization, rather classes, so defined, exist in these forms. What this suggests is that the unity of a class as a social force has to be “constructed” through the articulations and unification of the organizational form in which that class exists. The conclusion to be drawn from this, is that the relationship between the organizations through which a class exists inside the state apparatuses (as a political party or otherwise) and outside of those apparatuses (for example, in a branch of production) cannot be conceived of in instrumental terms, but must be seen as a relationship of articulation or disarticulation. This means that a shorter or longer term unity may come into existence between say, state forms of organization of a class and other, non-state forms of organization of the same class and in this restricted sense it is possible to talk of the political representation of the interests of a fraction of a class. But that representation is the outcome of class struggles and is dependent on the balance of forces; it is not given.
The state must be conceived as comprising determinate forms of organization of social classes, that is of social relations. As Poulantzas correctly argues all power exists “only in so far as it is materialized in certain apparatuses” and, in this sense, state apparatuses “... are no more appendages of power, but play a role in its constitution; the state itself is organically present in the generation of ‘class powers” (Poulantzas, 1979, p. 44-45). As such it is necessary to conceive
... of the capitalist State as a relation, as being structurally shot through and constituted with and by class contradictions, (and this) means firmly grasping the fact that an institution (the State) ... cannot really be a monolithic fissureless bloc, but is itself, by virtue of its very structure (the State is a relation), divided. The various organs and branches of the State (ministries and government offices, executive and parliament, central administration and local and regional authorities, army, judiciary etc.) reveal major contradictions among themselves, each of them frequently constituting the seat and the representative-in short, the crystallization-of this or that fraction of the power bloc, this or that specific and competing interest (Poulantzas, 1976, p. 75).
In this passage, Poulantzas appears to restrict the conflicts and contradictions within the state to class fractions of the power bloc and this implies that only the dominant classes have access to the state apparatuses. It seems to me that it is necessary to qualify this argument in an important respect.
It is undoubtedly the case that, in part, the materialization of class powers in the apparatuses of the state is expressed in the limitation of access or exclusion from the state of particular classes or fractions of classes. Yet, the process of exclusion and access operate in an uneven way and are by no means unchanging over time or uniform across all state apparatuses. I want to approach this, and a related question, through a discussion of ideological and repressive state apparatuses.
There has been much debate in the recent literature concerning the distinction between repressive ideological state apparatuses. In part the debate has focused on the validity of distinguishing apparatuses on this basis particularly as the notion of ideological apparatuses seems to carry with it functionalist implications. It is not my intention to enter into this debate at length and I will restrict my remarks to those aspects which are pertinent to the present paper. In his essay Ideology and Ideological State Apparatuses (1971), Althusser argues as follows:
Whereas the(Repressive) State Apparatuses constitute an organized whole whose different parts are centralized beneath a commanding unity, that of the politics of class struggle applied by the political representatives of the ruling classes possession of State power, the Ideological State Apparatuses are multiple, distinct, “relatively autonomous” and capable of providing an objective field of contradictions which express, in forms which may be limited or extreme, the effects of the clashes between the capitalist class struggle and the proletarian class struggle, as well as their subordinate forms. (pp. 141 and 142).
Whereas the unity of the (Repressive) State Apparatuses is secured by its unified and centralized organization under the leadership of the representative of the classes in power executing the politics of the class struggle of the classes in power, the unity of the different Ideological State Apparatuses is securred, usually in contradictory forms, by the ruling ideology, the ideology of the ruling class (p. 142).
There are a number of problems with this formulation, not least the simple instrumentalism attributed to repressive state apparatuses and the assumption that both sets of apparatuses constitute a given unity - in the case of Repressive State Apparatuses (R.S.A.’s) given in their organization; in the case of the Ideological State Apparatuses (I.S.A.’s) given in the ruling ideology. Althusser does stress that the I.S.A.’s are the site of contradictions which are the effects of capitalist-proletarian struggles and implies that these contradictions are absent from the R.S.A.’s. Of course, by conceiving of the R.S.A. as an instrument, Althusser obliterates both the specificity of the mode of operation of R.S.A.’s and simply reduces them to the control of the political representatives of the ruling class, thus failing to give any weight to their organizational and institutional structures, and excludes the possibility of struggle within and between the different R.S.A.’s. Yet, it is clear that the unity, coherence and functionality of the R.S.A.’s cannot be assumed - contradictions between different coercive state apparatuses are by no means exceptional in capitalist social formations.
Nonetheless, Althusser’s formulation is suggestive of an important feature of the differentiated mode of organization, institutionalization and operation of state apparatuses - namely, the question of “access” by different classes to various apparatuses.
In another context, Poulantzas recognized this aspect when he wrote:
The contradictions most directly and acutely reflected within the state are those among the dominant classes and fractions and between these and the supporting classes, far more than the contradiction between the power bloc and the working class. The latter contradictions are basically expressed in the bourgeois state “at a distance” (Poulantzas, 1973).
Davies et al. (1976) interpreted this to mean that the analysis of the South African state has to concentrate on the contradictions between the dominant classes, since the capital-labour contradictions took place outside of the state, “at a distance”. This, as Innes and Plaut correctly point out, leads to the neglect of the “fundamental” contradictions in South African society. However, the neglect flows not from the incorrectness of Poulantzas’s formulation, but from a particular interpretation of it. The point is that in so far as the proletariat is excluded from the State apparatuses qua state apparatuses, capital-labour contradiction and conflicts can only occur “at a distance.” This is not to suggest that the conflicts within the state can be understood independently of the contradictions between capital and labour. The point at issue is the characteristic differentiated organization of classes in and through tile apparatuses. The concept of access to the state has a far wider significance than Clarke (op. cit.) and Kaplan (1980) recognized. Indeed, to restrict the discussion of access to the state to party political representation is precisely to set up the entire state as the instrument of the dominant classes, wielded on their behalf, by their party political representatives. The crucial aspect is that the entire range of state apparatuses (but more particularly those that are not directly concerned with the coercive sanctioning and imposition of the law and policing) is, in principle, open to become a means of organization and definition of specific classes and class interests. However, not all state apparatuses are equally accessible to the organizations of contending classes. Party political representation is the most direct and open in situations where bourgeois democratic rights exist, but through myriad institutional and organizational forms “parts” of the fundamental classes (and, indeed, others) may become organized either directly, or through representation in a vast variety of state apparatuses. For example, the representation of trade unions and business on school and university councils, trade union representation on industrial conciliation tribunals, capitalists and trade unions on economic planning committees etc.
The access of classes or elements of classes to state apparatuses such that these apparatuses become the means of defining and organizing the interests of those fractions of classes, quite obviously depends on the outcome of struggles which take place in already given structured political and economic conditions. It is clear that there are a number f possibilities, certain apparatuses may become monopolized by a particular class fraction, others may be the site of struggle between opposing classes or class fractions and in this event these classes may have equal or asymmetrical access to the apparatuses in question etc.
It is necessary to emphasize that in so far as a class gains access to a state apparatus, there can be no assumption that the apparatus in question then serves the interests of “the class” defined at the level of relations of production. It follows from our earlier argument that the means (institutional and organizational arrangements) by which a particular fraction of a class, organized in a particular way and engaged in specific forms of struggle, gains access to or is excluded from state apparatuses may well result in a highly restricted and partial representation of the interests of a section of a class. Representation is as much a site of struggle as any other sphere of class relations.
Secondly, from this standpoint it is perfectly possible to conceive of a class or a section of a class achieving, under certain conditions, a situation in which a particular state apparatus operates for a shorter or longer period, and to a greater or lesser degree, so as to contradict certain conditions of reproduction of a capitalist social formation. Here we can think of regional and local states, particular state institutions in education and so on. Apart from the fact that such situations tend to be short-lived, the important point to stress is that the possibility of such an occurrence in no way suggests the parcelling up of the state; for the contradictory operation of a sector of the state no more amounts to this, than the existence of revolutionary parties, engaged in political practices outside of the state apparatus, entails a pluralistic conception of power.
These considerations have an obvious relevance to the analysis of the South African state since the operation of mechanisms of exclusion from and access to the state apparatuses have here taken a particular form and have had far reaching consequences.
Kaplan (1980) argues that the distinction between black and white in capitalist South Africa is based on the fact that most blacks retained possession of land while whites were proletarianized. He says:
This continued attachment to the land provided the material basis for the maintenance and transformation by the South African state, of a differentiated juridicial-political structure, broadly under the rubric of tribalism (p. 21).
By contrast :
... the structures of the capitalist state were extended to all whites. All whites were formally equal, free individuals and citizens, irrespective of their class. Individual equality was formally enshrined within the state’s institutions - from the law to parliament (p. 24).
Notwithstanding the unacceptable reductionism of Kaplan’s account, the differentiation between black and white subjects is not only explained simply on the basis of their respective access to or separation from means of production in the land but is also presupposed in the explanation, the distinction he points towards is important. The difficulty, however, is that the content he gives to the two types of subjects is unclear and insufficiently specified - he conflates, for example, rights of contract with citizenship – and, most importantly, he largely restricts their relevance to the function of achieving the “isolation” of black and white workers.
Access to and exclusion from the state apparatus must be understood in terms of the categories of black and white subjects. The category of the subject refers here to the qualities, rights, obligations, disabilities etc. which are attributed to and hence form (constitute) agents placed in specific institutional and organizational matrices. Clearly, the law plays a central role in this process, yet the formation of specific categories of black and white subjects is not reducible to a single origin, but must be seen as the outcome of class struggles and social conflicts within differentiated but inter-connected spheres (economic, political and ideological) of the social formation.
Whatever the social relations through which black and white subjects are constituted, the important question is the manner in which such categories once formed in the practices in one setting, are appropriated, transformed and put into effect in different institutional and organizational settings. What this implies is that the categories of white and black are not once and for all given, but are continuously formed and transformed. In turn, this suggests the possibility of contradictions not only between white and black, and not only between the black subjects in one sphere (say, the economy) and black subjects in another sphere (say, the political) but, indeed, also within the same sphere (e.g. the black petit-bourgeoisie and the black proletariat).
Kaplan treats the existence of the category of the black subject solely as providing a means of their exclusion from the state apparatuses and the existence of the category of white subjects as providing the means for the extension of the capitalist state to whites. This, however, is inadequate. It is undoubtedly the case that the category of the white subject constitutes the foundation fur white workers to obtain access to state apparatuses as one important form of their organization - parliamentary representation, industrial conciliation tribunals, various government commissions and state committees and so on. But in no sense can white workers’ access to the state apparatuses be equated with that of capital. It is precisely the asymmetrical organization of classes within state apparatuses which enables the apparatuses, whatever the contradictions arid conflicts and transformations, to operate to reproduce both determinate privileged conditions of white subjects and of capitalist production. On the other hand, the category of the black subject provides a condition for the exclusion of black workers (e.g. they are excluded from the definition of employee in the Industrial Conciliation Act) from such important arenas as industrial conciliation and popular representation.
Nonetheless, the category of the black subject is at the same time a condition of access to certain particular types of state apparatuses. I have in mind here the Bantustan legislative councils and administrations, Bantu Education, local councils and similar apparatuses. In order to avoid any misunderstanding, it must be emphasized that there is no intention, in dealing with these state apparatuses, to minimize the massive role played by directly coercive state apparatuses or to suggest that the state organizations in question amount to “real” reforms of the political system. The point being made is that the category of the black subject gives the popular masses access to and constitutes particular state apparatuses as forms of the organization of these masses. In this sense, these state apparatuses arc constituted in a quite different manner from those in which the representatives of specific organizations (say the registered trade unions) are placed on functional committees or in which government nominees serve. This is riot to argue that the latter apparatuses will be free of conflicts and contradictions but is to suggest that the former provide, by the nature of their structure and the forms of access to them, sites of mass struggles.
The argument so far has been that the state apparatuses are, differentially, the means by which certain class powers are organized and that, by this token, these apparatuses become the sites of determinate contradictions, conflicts and struggles. Contradiction and conflict is not, however, merely encapsulated within particularapparatuses but occurs equally between apparatuses. Obviously, this may be due, in part, to the pattern of access/exclusion which, as I pointed out above, operate unequally in relation to different apparatuses, but it may also be attributable to other conditions including the way in which the functional activities of different state branches come to be defined.
Be that as it may, the aspect I wish to lay stress on, is the interrelationship between functionally differentiated state apparatuses. This question has been much neglected, particularly in the discussion of the South African state because of the restrictions inscribed in the instrumentalist conception of the capitalist state. In South Africa, it has simply been assumed that the state apparatuses unproblematically give effect to a single unified policy or set of policies. However, if the state is conceived of not as an instrument but as a complex structure of contradiction and conflict then what becomes significant is precisely the non-complementary and, indeed, competing policies which may be pursued by specific state apparatuses. These relations are significant, of course, because they reveal points of possible instability and emergent struggles. There is obviously much room for research in this area, but its importance can be indicated by a brief reference to the South African judiciary.
Over a considerable period the judiciary enjoyed a considerable autonomy in South Africa. At the level of the law, autonomy was inscribed not only in the constitutional guarantees of independence (tenure of judges and so forth) but also in the procedures and modes of proof laid down in the Criminal Procedure Act, Judges’ Rules, etc., procedures which gave considerable discretion to the judges. Similarly, in constitutional issues, the courts were allotted certain powers. The autonomy accorded to the courts did not remain at the formal level of the law and, indeed, the history of South Africa since 1910records numerous important “legal landmarks” in which the courts operated against the positions pressed for by other state apparatuses. These include early court decisions excluding African women from the provisions of the laws requiring Africans to carry passes; “the rejection of constitutional amendments aimed at depriving coloured voters of the franchise, and, importantly, the obstacle the courts placed in the path of administrators, law-enforcing agencies and the executive through the enforcement of habeas corpus rights and the most narrow interpretation possible of statutes which affected the rights and status of black subjects. The analyses which have been attempted of the relationship of the judiciary to other state apparatuses have tended to reduce the explanation to the assumed ideologies of individual judges. What is required, however, is an analysis of the organizational and ideological structures of these institutions and their articulation within the complex of class and other struggles in the social formation.
I referred earlier to Althusser’s distinction between repressive and ideological state apparatuses, a distinction which was stimulated by, although it is clearly not the same as, Gramsci’s notion of the state as equalling political plus civil society. Encompassed in the concept of ideological state apparatuses is that while range of institutions and organizations, whether public or private, which are united by their role in the constitution of an ideological hegemony in the social formation. The argument for ignoring the distinction between public and private apparatuses rests firstly, on the ground that the distinction public/ private is “internal” to bourgeois law and secondly, on the ground that, in any event, the ideological apparatuses serve as the means for the reproduction of the social formation and they are, therefore, unified by this function.
This argument has been criticized on a number of grounds, and, in particular, the refusal to grant any effectivity to the public/private distinction has been questioned. This objection has much force, for to fail to recognize the real effects of the shifting public/private dichotomy means that the state as a specific complex of structures simply becomes absorbed amorphously into “society”. At the same time, the fundamental strategic importance of the extended conception of the ideological state apparatuses must be recognized - it brings to the centre of the political struggle the institutions and organizations of all kinds in which power is materialized. Politics is not simply the “external” struggle of total confrontation at the barricades, but finds its location in many sites of the materialization of power.
In this paper, however, I have focused on the public apparatuses of the state in South Africa, but this is riot to suggest that the arena of political struggle in the private ideological apparatuses - cultural organizations, residents associations, religious bodies etc. - is not of fundamental importance. The analysis of, and the f6rrnulatiun of strategies and tactics within both the state and the private apparatuses, as a complement to the armed struggle, remains a priority in South Africa.
My thanks are due to Colin Mercer who read and commented on an earlier draft.
- The main texts are to be found in:
Davies, R. (1976), (1979); Davies, R. et al. (1976); Kaplan, D. (1975), (1976), (1976a), (1977), (1980) ; Morris, M. (1975).
- There are, of course, acute problems in Poulantzas’s earlier formulations of the concept of structure which derive from the fact that social relations are excluded from the definition of the concept. I touch on a related problem in the final section of the paper.
- This passage is somewhat confusing because the government’s policy seems to be equated with the state and the state distinguished from the power bloc, the composition of which denotes the form of state. It is, of course, precisely this type of theoretical confusion which precludes an adequate analysis of the state and its articulation with the class struggle.
- I use the term “purportedly” since both in Davies et al. (1970: p. 415) and in Davies (1979: p. 7 and 8) there are clear departures from Poulantzas. Thus in the latter text, structural determination of class is reduced to “the place occupied in the relations of production”, whereas for Poulantzas structural determination includes the ideological and political instances. In the former text, the distinction between structural determination’ and class struggle is simply dissolved.
The reproduction of social classes is never the mere unfolding of the internal logic of capital, but occurs only under concrete conditions of class struggle. It is in this sense that the structural determination of class is specified (p. 5).
- It is true that Clarke also tries to argue that “capital-in-general” also has a real existence, but this is restricted to the money form and the relations between Departments I and II and III take it that he is not wanting to argue that it is capital-in-general in this restricted sense that “imposes” itself on the state.
Althusser, L. (1971) Ideology and Ideological State Apparatuses. In Lenin and Philosophy. New Left Books : London.
Clarke, S. (1978) Capital, fractions of capital and the state: ‘neo-Marxist’ analyses of the South African state. Capital and Class 5, (Summer).
Davies, R. (1976) Mining capital, the state and unskilled white workers in South Africa, 1901-1913. Journal of Southern African Studies, 3, (1).
Davies, R. (1979) Capital, State and White Labour in South Africa, 1900-1906. Humanities Press.
Davies, R., Kaplan, D., Morris, M. & O’Meara, D. (1976) Class struggles and the periodization of the South African state. Review of African Political Economy, 7.
Holloway, J. & Picciotto, S. (1978) State and Capital: A Marxist Debate. Edward Arnold.
Innes, D. & Plaut, M. (1978) Class struggle and the state. Review of African Political Economy, 11, (January-April).
Kaplan, D. (1975) The State and Economic Development in South Africa. Mimeo.
Kaplan, D. (1976) The politics of industrial protection in South Africa, 1910-1939. Journal of Southern African Studies, 3, 1.
Kaplan, D. (1980) An Analysis of the South African State in the “Fusion” Period 1932-1939. I.C.S. Seminar Paper.
Kaplan, D. (1977) Class Conflict, Capital Accumulation and the State: An Historical Analysis of the State in Twentieth Century South Africa. University of Sussex Ph.D. Thesis.’
Kaplan, D. (1980) Relations of production, class struggle and the state in South Africa in the inter-war period, Review of African Political Economy 15-16.
Marks, S. & Trapido, S. (1979) Lord Milner and the South African state. History Workshop, 8, (Autumn).
Morris, M. (1975) Periodization, Class Struggle and the State in South Africa, Mimeo.
Poulantzas, N. (1973) Political Power and Social Classes. New Left Books : London.
Poulantzas, N. (1975) Classes in Contemporary Capitalism. New Left Books: London.
Poulantzas, N. (1976) The capitalist state: A reply to Miliband and Laclau. New Left Review, 95, (January-February).
Poulantzas, N. (1979) State Power and Socialism. New Left Books: London.
Slova, J. (1976) South Africa - No Middle Road. In Southern Africa - The New Politics of Revolution. (Davidson, B., Slovo, J., Wilkinson, A. R.,) Pelican Books: London.