Chapter 3

Functionalism

As we saw in the previous chapter, the theorists of political development were strongly imbued with the 'scientific' spirit of the times, and their first instinct was to contribute to the creation of a truly modern science of politics by finding a comprehensive theoretical framework within which their practical concerns regarding the need for stable, pro-Western regimes throughout the world could be addressed. This was to lead to a number of extremely elaborate if eventually unproductive efforts, nowhere more so than in the saga of the 'functional' approach. As we shall see in this chapter, it failed precisely because of the very high level of generalization and abstraction which it sought to achieve – in precise terms, in seeking to devise an analytical framework which was good for all political systems at all times, it failed both to recognize the key elements of politics in modern capitalist societies in particular, and to address the specific situation of societies in transition.
  This effort to produce a comprehensive functional framework within which all political systems past and present could be analysed as a basis for the scientific study of comparative politics and political development is primarily associated with the work of Almond. One of a number of approaches within political development theory, it was soon to be displaced by the cultural, institutional and comparative historical perspectives reviewed in later chapters. In part this was because it was so ambitious in scope that its proponents themselves eventually despaired of turning it into a useful instrument for the analysis of developing societies. At the same time, the theoretical framework upon which was based proved both too abstract and too specific to provide the universal framework of analysis that was sought – too abstract in its determination to produce a universal model of the political system and the political process divorced from the social context in which each was embedded, but too specific in its direct derivation of the model from the familiar political institutions of Britain and the United States. Finally, the functional approach never entirely freed itself from the static quality it derived from its initially overwhelming concern with the classification of existing political systems, rather than with the dynamic processes of political change. Its classificatory power was limited as a result of its reliance upon a distinction between 'Western' and 'non-Western' states which was too crude to capture the character and variety of political systems across the Third World. And despite the efforts that were made to give it a dynamic or developmental orientation, and to use it to address the issue of political development directly, it proved difficult to derive practical policy-related conclusions from it. As a result, when Almond and Powell (1966) finally turned to the issue of the government of new states at the end of Comparative Politics: A Development Approach, they abandoned their own functionalist premises, and drew their recommendations directly from the revisionist theory of democracy.
This reflects a common pattern in the history of the theory of political development: the failure of a particular attempt to provide a theoretically satisfying basis for it has generally led neither to the abandonment of the enterprise, nor to a questioning of its underlying values and policy commitments. Rather, it has prompted a restatement of both, a return to the fundamental precepts of the revisionist theory of democracy, and a promise to do better next time. Such failures led to the increasingly systematic separation of the doctrine for political development from the theory, and in the end to the decision to retain the doctrine, and to do without the theory. The record of the functionalist approach in the 1960s – what it set out to do, how and why it failed, and how its creators reacted when it did – tells us a great deal, therefore, about the ideological and doctrinal content of political development theory, and its tendency to triumph over the claims of social science as science.

The Functional Approach

The basic idea behind the functional approach, stated in the opening pages of The Politics of the Developing Areas, was a simple one: that all political systems perform the same core set of functions, although these functions may be performed by different structures from one society to the next (Almond 1960: 11). This first sketch of the approach was followed by the more elaborate version contained in Comparative Politics: A Developmental Approach. In order to trace the development to the functional approach and assess its significance, therefore, it is necessary to compare these texts, and the manner in which they addressed the specific issue of political development in the new states.
The first version of the approach was presented in The Politics of the Developing Areas (1960). In this multi-authored volume, a lengthy introduction by Almond and a rather briefer conclusion by Coleman framed the case studies of South-East Asia, South Asia, Sub-Saharan Africa, the Middle East and Latin America which made up the bulk of the text. For Almond and Coleman alike, the immediate purpose was to establish a framework for the comparative analysis of political systems. The application of the framework to the problems facing 'new states' was a secondary and still distant goal. Almond hoped that the effort might eventually lead to an ability to 'predict the trend of political development in modernizing states', but believed that this implied 'a state of knowledge of the performance of modern Western polities far beyond what we have attained today' (Almond 1960: 63-4; emphasis mine). Coleman reminded his readers at the very end of the book that its major purpose had been to make a contribution to the general theory of political systems by improving the capacity to 'order the phenomena of non-Western political systems, and to compare them with the Western ones, and with one another', adding that 'a second purpose, which has been fulfilled only to a very limited extent, is to improve our understanding of the processes of political change or modernization' (Coleman 1960b: 576).
  In addition to this acknowledged shortcoming, there was a conspicuous lack of integration between the individual contributions to the collection, or between the theory and the area studies offered. Almond claimed that if the project upon which he was engaged proved successful it would render area studies obsolete (Almond 1960: 64), but gave little indication that he knew anything about the politics of new states in the areas concerned. With the exception of some brief references to India, his examples were drawn either from the United States and Western Europe or from anthropological studies of so-called 'stateless societies'. He left Coleman, in his conclusion, to apply the framework as best he could to the developing world.
  We noted in the previous chapter that Almond's introduction, one of the first texts in which the issue of political development was directly addressed, rejected the idea of a bipolar contrast between 'traditional' and 'modern' political systems. His decision to dwell upon so-called primitive bands such as the Eskimo on the one hand and contemporary 'Western' states on the other reflected his intention to demonstrate the existence of common features across all political structures. This dictated a concern with the least developed political societies, rather than with the 'modernizing' new nations with which the bulk of political development theory would concern itself. Almond and Coleman shared the assumption, perhaps not accepted by the contributors of some of the area studies, that it would not be possible to address the character of such transitional systems until a comprehensive universal framework had been developed.
The first requirement was to define the political system itself. With reference to Weber, Levy, Laswell and Kaplan, and Easton, it was defined as 'that system of interactions to be found in all independent societies which performs the functions of integration and adaptation (both internally and vis-à-vis other societies) by means of the employment, or threat of employment, of more or less legitimate physical compulsion' (7). At the same time, Almond rejected the association of the political system either with specific institutions of the state or with specific individuals holding political office:

when we speak of the political system we include … not just the structures based on law, like parliaments, executives, bureaucracies, and courts, or just the associational or formally organised units, like parties, interest groups, and media of communication, but all of the structures in their political aspects, including undifferentiated structures like kinship and lineage, status and caste groups, as well as anomic phenomena like riots, street demonstrations, and the like (7-8).

The boundary between the political system and other social systems was conceived not as physical, but as behavioural: individuals acting to affect the 'authoritative allocation of values' within any society were assuming a political role, or entering the political system. Within this framework, Almond asked: 'What are the common properties of all political systems? What makes the Bergdama band and the United Kingdom members of the same universe?' and suggested that there were essentially four such properties: all political systems have political structure; all perform the same functions; all political structure is multifunctional; and all political systems are culturally mixed (Box 3.1).

Box 3.1 Almond's 'common properties of political systems'

First, all political systems, including the simplest ones, have political structure. In a sense it is correct to say that even the simplest societies have all the types of political structure which are to be found in the most complex ones. They may be compared with one another according to the degree and form of structural specialization.
Second, the same functions are performed in all political systems, even though these functions may be performed with different frequencies by different kinds of structures. Comparisons may be made according to the frequency of the performance of the functions, the kinds of structure performing them, and the style of their performance.
Third, all political structure, no matter how specialized, whether it is found in primitive or in modern societies, is multifunctional. Political systems may be compared according to the degree of specificity of function in the structure; but the limiting case, while specialized, still involves substantial multifunctionality.
Fourth, all political systems are 'mixed' systems in the cultural sense. There are no 'all-modern' cultures and structures, in the sense of rationality, and no all-primitive ones, in the sense of traditionality. They differ in the relative dominance of the one as against the other, and in the pattern of mixture of the two components.

Source: Almond 1960: 11.

The four points outlined here underline the priority given to the need to discover the universal features shared by all political systems, and the claim that all political systems have particular structures and functions in common. They also identify two important arguments: that all structures are multifunctional, and that all systems are culturally mixed. Almond did not argue for a rigid one-to-one correspondence in which each function was exclusively performed by a particular structure. Instead, he argued (on the basis of his studies of the United States) that a particular structure regulated the manner in which that function was performed by a number of structures. This linked directly to the claim, discussed in the previous chapter, that all systems were culturally mixed: 'traditional' features where never eliminated, even in the most modern societies, but rather incorporated and to a certain extent transformed by modern features which dictated the manner in which particular functions were performed. Thus, for example, such features as kinship, friendship and 'school ties' still affected recruitment in a modern Western political systems, but 'the more thoroughgoing the political modernization, the more these ascriptive criteria are contained within or limited by achievement criteria – educational levels, performance levels on examinations, formal records of achievement in political rules, and the like' (Almond 1960: 32).
  Almond identified seven functions common to all political systems, of which four were concerned with 'input' (political socialization and recruitment, interest articulation, interest aggregation, and political communication) and three with 'output' (rule-making, rule application and rule adjudication). Political socialization and recruitment operated through a range of institutions including the family and the school, as well as through more obviously 'political' institutions. The remaining functions were identified with specific institutions in existing Western systems – interest groups, parties, mass media, legislatures, executives, and the judiciary respectively – which were assumed to have functional counterparts in all political systems. Almond did not argue that the structures which performed these functions in 'modern' political systems were appropriate to all systems, but he did claim that as political systems approached modernity they would require them, or something very like them.
  Complex modern political systems required autonomous and differentiated structures of interest articulation and aggregation if the necessary functions were to be smoothly performed, regulated and integrated. In order to enjoy the necessary degree of autonomy there needed to be good boundary maintenance between the polity and society, and between the particular structures within the political system itself; otherwise, the political system would be plagued by the 'frequent eruptions of unprocessed claims without controlled direction into the political system' (35). This would be avoided not simply by the presence of interest groups, political parties, and mass media of some kind, but by specific types of these institutions: associational interest groups, secular, pragmatic, bargaining parties, and free and neutral mass media – such as typified the homogeneous political cultures of the United Kingdom, the old Commonwealth and the United States (46). These homogeneous political cultures and specific types of interest groups, parties and mass media contrasted with the fragmented political cultures and relatively undifferentiated structures of interest articulation and aggregation of such continental systems as France and Italy, with their politicized interest groups, and their parties tied to particular ideologies or interests.
  In an ideal modern system, the political system is separated and appropriately insulated from direct societal pressure, and characterized by well differentiated associational interest groups and pragmatic parties. These institutions regulate but do not monopolize the functions of interest articulation and regulation respectively, while 'an autonomous communications system "regulates the regulators" and thereby preserves the autonomies and freedoms of the democratic polity', helping to develop and maintain an active and effective electorate and citizenship' (47–8). The key features of the system are functional regulation – the 'regulation of the performance of the function within the polity by a specialized autonomous structure with a boundary of its own and the capacity to "enforce" this boundary in the system as a whole', and cultural penetration – the fusion of modern and traditional cultural traits through the 'penetration of the "traditional" styles of diffuseness, particularism, ascriptiveness, and affectivity, by the "rational" styles of specificity, universalism, achievement, and affective neutrality' (63). In the effectively functioning modern Western political system, the universal traits of multifunctionality and cultural dualism were reflected in smooth functional regulation by differentiated and relatively autonomous institutions, and by the fusion of cultural traits by the penetration and transformation of traditional styles by modern ones. And as far as the development of theory was concerned, the way to a formal theory of political modernization would be found, if at all, through the discovery of 'reliable indicators' for these key concepts of functional regulation and cultural penetration.
  Six years after The Politics of the Developing Areas appeared, a fuller version of the functional approach was set out by Almond and Powell in Comparative Politics: A Developmental Approach (1966). An elaborate justification for the approach was now provided: it was part of an intellectual revolution taking place in the study of comparative government. Almond and Powell condemned the parochialism and formalism of previous approaches, which they accused of being excessively concerned with the constitutions and government institutions of predominantly European political systems, and announced four related goals for their own approach: comprehensiveness (the inclusion of 'non-Western' cases), realism (the analysis of actual behaviour rather than formal rules), precision (the application of quantitative techniques) and intellectual order (the creation of a 'unified theory of politics' which will bring together the fields of comparative government, political theory and international relations). Much of the detail of the version of functionalism offered was carried over from the first sketch outlined above, but the whole approach was recast, as the book title suggested, in an effort to give it a developmental orientation. This specifically developmental framework of analysis was introduced in response to criticisms that the first version of the functional scheme had been too static. Rustow and Ward, who had studied the process of modernization over time in two specific historical cases in Political Modernization in Japan and Turkey, one of the early volumes in the Studies in Political Development series, cited Almond and Coleman (1960), the Politics of the Developing Areas, as one example (in fact, as the most comprehensive and illuminating of its type) of studies which had searched for 'categories of comparison which would be applicable to political systems as they are', complaining that 'the type of comparison these studies embody is static: they analyze and compare political systems viewed in cross-section at a given point in time, and they conceive of these systems as existing essentially in a state of equilibrium' (Rustow and Ward 1964: 10). Almond and Powell now acknowledged the justice of this criticism, and sought to respond to it (Almond and Powell 1966:13).
  The new developmental version of functionalism had three elements – it introduced the idea of the capabilities of political systems and their development over time, it defined the remaining six functions as conversion processes internal to the political system, and it identified the functions of political socialization and recruitment as developmental processes. These innovations allowed Almond and Powell to examine the political system from three related points of view: its transactions with domestic and foreign environments (capabilities); its activities or internal processes (called conversion functions because they convert inputs from the environments to outputs to the environments); and the mechanisms through which it survived or changed over time (developmental system maintenance and adaptation functions). Six conversion functions internal to the system were identified, each carried over from the earlier version of the approach, and linked as before to familiar 'Western' public or social institutions (Box 3. 2).

Box 3.2 Almond and Powell's conversion functions

Function

Rule-making
Rule application
Rule adjudication
Interest articulation
Interest aggregation
Communication

Associated Institution

Legislature
Executive
Judiciary
Pressure groups
Political Parties
Mass media

To emphasize their concern with the behaviour of individuals and the performance of institutions, Almond and Powell discarded the terms 'office' and 'institution' in favour of 'role' and 'structure' (or subsystem), defining structures in this context as 'sets of role which are related to one another' (21). Finally, they argued that each individual (usually national) political system had its own distinctive political culture (and subcultures), defined as a 'psychological dimension' of 'underlying propensities' which consisted of 'attitudes, beliefs, values and skills which are current in an entire population' (23). As we shall see in Chapter 4, the topic of political culture was treated extensively in other work in the same period. For Almond and Powell, then, the political system as a totality was made up of 'interacting roles, structures and subsystems, and of underlying psychological propensities which affect these interactions' (25).
  Within this framework, the political system was now seen as subject to change and development over time. At the structural level, Almond and Powell argued, change took place through a continuous process of bringing new individuals into new political roles (the recruitment function). This process had a developmental character, in that the system as a whole also experienced structural differentiation or increasing subsystem autonomy. At the cultural level, change took place through the inculcation of political attitudes and values through growth to adulthood and recruitment to particular roles (the socialization function). This process overall had a similarly developmental character in that over time the political culture as a whole experienced increased secularization, defined here as the process whereby men (sic, as throughout) become increasingly rational, analytical, and empirical in their political action' (24). In other words, the functions of recruitment and socialization were now introduced as dynamic processes which shaped the development of the political system as a whole over time, and moved it in the direction of increased subsystem autonomy and rationality. These developmental cultural and structural aspects of the process of modernization provided a general framework within which to locate the six universal conversion functions, and to address the issue of change over time in the structures through which they were carried out. They also provided a general model in which the processes of cultural penetration and functional regulation in efficiently functioning modern political systems could be identified as particular forms.
  The analysis was now supplemented by detailed consideration of the capabilities of the political system and their potential for development over time. These referred to the way that the system performed overall as a unit in relation to its environment, and were closely related to the inputs (demands and supports) from the system itself or the environment, and the outputs to the environment. Three of the output types identified related to specific demands and were associated with an identical capability. These were regulative, extractive, and distributive. In addition a fourth class of symbolic output was recognized, and a somewhat different fourth capability, associated in particular with democracies, which were seen as being responsive to the input of demands from groups in the society (27-9). At a later point in the exposition, five capabilities were listed, including the symbolic (38). The theory of the political system as a whole was to be based upon the 'threefold classification of functions', and would 'consist of the discovery of the relations between these different levels of functioning – capabilities, conversion functions and system maintenance and adaptation functions – and of the relation of the functions at each level' (30).
  Within this framework, Almond and Powell now hoped not only to be able to compare political systems in static terms, but also to address the issue of their development. Political development could be explained and predicted by relating system challenges to system responses, and would be expressed in the dimensions of structural differentiation and cultural secularization. Challenges might arise from the external environment, the domestic society, or the political elites themselves, and it was argued that 'development results when the existing structure and culture of the political system is unable to cope with the problem or challenge which confronts it without further structural differentiation and cultural secularization' (34). Four specific types of problems or challenges were identified: those of state-building (penetration and integration), nation-building (loyalty and commitment), participation, and distribution. Five factors were identified as relevant to the analysis of political development – the type (in terms of intensity and number at one time) of problems faced, the resources on which the system could draw, the capabilities of other systems in the domestic or international environments, the functioning pattern of the system itself, and the response of political elites to political system challenges. Finally it was accepted, with an acknowledgement to Huntington's argument that the outcome of political change was frequently political decay rather than political development (Huntington 1965), that development need not always be progressive. It could be negative or regressive if the political system proved unable to respond to the challenges it faced, in which case its capabilities could decline or become overloaded. Thus Almond and Powell stated clearly that 'when we use the developmental concepts "structural differentiation" and "cultural secularization", we do not imply that there is any inevitable trend in these directions in the development of political systems' (1966:25).
  As this exposition reveals, the second version of the functionalist approach was considerably more elaborate than the first. In particular, an attempt was made to specify its systemic quality, and to address the issue of political change. Significantly, though, it retained its universal ambition and abstract analytical framework; and its authors conceded, in the very act of introducing a developmental dimension to the analysis, that no assumptions could be made about the direction that development would take. Both of these aspects of the approach would have important consequences for the way which the politics of 'developing' states would be addressed.
  In order to examine the manner in which the approach was applied to these states in particular, as a contribution to the theory of political development, we must consider first the approach adopted to structure and function in the transitional or developing states themselves, and the extent to which the theory that first Almond and then Almond and Powell together sought to develop was successfully integrated with the empirical data on developing states which these studies contained. This will bring out the tension between the desire to construct a universal framework of analysis on the one hand, and the specific concern with the politics of countries deemed vulnerable because of their transitional status on the other.

Structure and Function in Developing States

  As noted above, the theoretical and empirical chapters of The Politics of the Developing Areas (Almond and Coleman 1960) were poorly integrated, and the elaborate theoretical framework of Comparative Politics: a Developmental Approach (Almond and Powell 1966) was based upon the kinds of institutions most common to developed Western states. In each volume, theoretical exposition took pride of place, as the principal authors argued that a better understanding of the processes of political change in transitional societies required the elaboration of a universal functional theory. However, there was more to these volumes than the attempt to develop such a theory. The case studies in The Politics of the Developing Areas also provided the first extended accounts of the politics of the new states within the political development literature, while Comparative Politics: A Developmental Approach offered the first attempt to make systematic reference to a wide range of 'Third World' cases in the context of a theory of political development. The treatment of the politics of developing states in each of these accounts has an interest of its own therefore, apart from that which it derives from the theoretical approach to which is linked. In fact, as we shall see, a number of the substantive themes addressed were either separate or in principle separable from the claims made for the functional approach itself. It is just as important to examine the treatments of the politics of developing states themselves and the substance of the particular arguments advanced as it is to reconstruct the more abstract functional arguments which were given pride of place.
  The regional case studies in The Politics of the Developing Areas shared a common framework. Each had successive sections on background, processes of change, political groups and political functions, government structures and authoritative functions, and political integration. In addition, each addressed in turn the seven different functions identified by Almond in his introduction. For the most part, however, the common framework was used largely to organize descriptive material, and little or no account was taken of the particular functional arguments which Almond advanced. First and foremost, these contributions were survey articles written by regional experts, intended to give an overview of the politics of the region with appropriate background information.
  A large part of each essay was taken up with the classification and description of political parties in the region, in accordance with typologies which were broadly consistent with what Almond proposed in his introduction. The majority of the case studies were discussions loosely organized around themes common to the broad literature on political development, rather than direct applications of the functional approach. Blanksten's essay on Latin America insisted on the authoritarian character of its heritage; Rustow, on the Middle East, recounted a history 'punctuated by dictatorships, military coups, riots, and states of siege' (Rustow 1960: 395), stressing the weakness of liberal constitutionalism and the threat posed by communism; and Weiner, on South Asia, contrasted the fortunes and prospects of India, Pakistan, and Ceylon (Sri Lanka). Only Pye on Southeast Asia and Coleman on Sub-Saharan Africa organized their material to address current arguments in political development theory. Coleman, writing on the eve of the major wave of decolonization in Sub-Saharan Africa, characterized the politics of the region as 'highly transitional', arguing that

In most instances this great transformation is not occurring within the boundaries of the framework of established societies; rather, entirely new political systems and new societies are in the process of birth. Politically the essence of this dynamic movement is that power is rapidly gravitating into the hands of entirely new social groups unaccustomed to its exercise. As a consequence there is a generalized instability, and particularly an unpredictability regarding political authority (Coleman 1960a: 312).

In these circumstances, feelings of hopelessness, frustration and insecurity abounded; parties tended to be the instruments of groups concerned with their relative position in the new social and political system rather than with questions of public policy, while specific interests tended to be pursued through traditional groupings rather than through functionally specific interest groups, so that in associational and party development alike racial, tribal, religious and regional interests were mobilized and intensified. Even at this early stage of the process of decolonization, Coleman saw evidence of increased roles for armies, increased pro-government partisanship in previously neutral bureaucracies, and the blocking of career paths for the younger generation as a relatively young Africans monopolized the prestigious positions being hastily abandoned by European colonial officials and rulers. Overall, the situation was fluid, and the prospects for stability in the foreseeable future were unclear.
Finally, Pye's lengthy contribution on South-East Asia, apparently written at about the same time as his essay on the 'non-Western political process', largely replicated the arguments developed there. Pye argued that

many of the Westernized structures do not perform the same functions in Southeast Asian societies as they do in the West, while many of the functions of the political process are not performed by the type of group or organization which would perform them in a Western society (Pye 1960: 109).

  Within this framework, the arguments he advanced were perfectly familiar. The key characteristic of the region was that 'as yet there has not emerged in any of the Southeast Asian countries a distinct sphere of political relations that is clearly separated from the more basic patterns of social and personal relations' (118). Political parties were either followings seeking power and positions for their leaders, or ideological movements seeking to represent 'total orientations' towards life; legislatures played very minor roles in law-making; and associational interest groups were poorly developed. In most cases, authoritative institutional groups dominated the political scene:

Instead of the political parties and associational interest groups making policy decisions for the administrators and soldiers, the latter have generally taken the initiative. In several of the countries, their course of political development has resulted in the bureaucracy and the army becoming the only effectively organized and relatively modernized groups in the entire society capable of political action (115).

  This was largely the case because the majority of the population were motivated to participate in politics by feelings of restlessness and insecurity arising from the disruption of their traditional patterns of life. As they were seeking participation 'in order to resolve intensely personal problems', there was 'no logic that can relate the specific needs of the individual to any specific goals of public policy' (133). Rather, such individuals were drawn towards 'various deviant movements… and particularly to the Communist parties' (134). Thus

the leaders of the Southeast Asian countries are faced with a dual task: first, they are seeking to create a viable system of intra-elite role relationships; second, they are called upon to encourage the development of new systems of non-elite role relationships for the bulk of their peoples (118).

In the light of these contributions, it is evident that there was a significant difference of emphasis between the theoretical exposition on the one hand, and the approach adopted in the major area studies contained in The Politics of the Developing Areas on the other. In the construction of the functional framework considerable attention had been paid to the claim that there was a universal set of political functions carried out in all political systems by a variety of structures or institutions, and its theoretical implications. In the treatment of the existing political systems of the new states, in contrast, greater attention was paid to the empirical evidence for the related but far less ambitious claim that structures or institutions familiar in Western contexts did not function as they did in those same contexts. Valuable as this insight might be, it had nothing at all to say about the possibility or the utility of devising a common framework for the analysis of all political systems. Nor did it require the assumption that in the end there was a single set of functions which all political systems were bound to fulfil in one way or another.
As it turned out, the approach adopted in the case studies carried forward a number of familiar arguments first encountered in discussions of the 'non-Western political process' which could feed into but did not either require or directly support the more elaborate functional argument. The manner in which the particular case of transitional societies was addressed, particularly by Coleman and Pye, showed the same gap between the arguments advanced and the full functional model. For each of them, the key characteristic which gave rise to political instability was the simple fact that the societies observed were in a state of flux. Formerly stable patterns of social and political interaction were being disrupted, and new patterns had not yet been consolidated in their place. As we saw in the previous chapter, this perception was central to the concerns which gave rise to the political development literature. However, because it was the simple factor of disruption arising from rapid change which was said to provoke instability, Coleman and Pye did not need to offer or adopt an elaborate functional theory to justify their claims. Their principal hypothesis, that rapid change had a disorienting and destabilizing effect, did not depend upon a particular analysis of the nature of the stability that had preceded it; nor did it require them to assume that stability would be restored on a new basis at some point in the future. Least of all did it require them to subscribe to the idea that a universal set of political functions could be defined, and associated with different structures in different political systems. In the light of this imbalance between theoretical framework and substantive analytical content we need to look further, in order to see what specific conclusions regarding politics in the new states were advanced in connection with the functional approach. In the light of what we have seen so far, we have good reason to be sceptical that they rested upon any of the grander claims of the functional approach. If they did not, the approach itself should be seen as at best purely decorative, and at worst either a self-deluding source of confusion, or a pseudo-scientific attempt to mask and justify highly partisan conclusions.

Theory and Substantive Argument in the Functional Approach

  The conclusions derived from the functional approach are best understood through a reading of Coleman's conclusion to The Politics of the Developing Areas (Almond and Coleman 1960) and the parallel classification of political systems and concluding chapter in Almond and Powell's (1966) Comparative Politics: A Developmental Approach. As we shall see, the classification of developing states presented particular problems, and the conclusions drawn with regard to them did indeed owe more to the conservative elitism of the revisionist theory of democracy plan to the particular claims of the functional approach.
In his conclusion to The Politics of the Developing Areas, Coleman briefly summarized Almond's functional argument, then classified the states covered in the text as competitive, semi-competitive or authoritarian, and modern, mixed or traditional. Only three (Chile, Israel and Uruguay) of the 74 were classified as both modern and competitive. The ensuing discussion emphasized the extent to which the great majority felt short of the characteristics felt to be appropriate for effective modern systems. Coleman next assessed the degree of correlation between democracy and economic development, concluding that the connection was established when countries were were compared in groups, but weakened by the appearance of negative correlations when countries were considered individually. He then offered a six-category typology of African and Asian political systems as political democracies, tutelary democracies, terminal colonial democracies, modernizing oligarchies, traditional oligarchies and colonial or racial oligarchies. In a strange departure, considered further in chapter 7, the Latin American states were excluded from this typology and classified separately in accordance with the criterion of the degree of party competitiveness. At the same time, the whole exercise was regarded as provisional: 'Many of the countries concerned could be regarded as marginal in the category to which they have been assigned, and approached with a different purpose, could justifiably be shifted to another' (Coleman 1960: 562). The book concluded with his classification, and the reminder that 'the development of a rigorous theory of political modernization is still unfinished business' (576).
  Whereas The Politics of the Developing Areas concluded with a tentative classification of political systems, however, Comparative Politics: A Developmental Approach moved from a classification of all political systems from primitive to modern to an attempt to sketch out a theory of political development organized around the ideas of structural differentiation, autonomy and secularization, and to assess its descriptive, explanatory and predictive power. The classification of political systems, based upon Dahl's development of Aristotle's sixfold typology of monarchy, tyranny, aristocracy, oligarchy, polity and democracy, distinguished between primitive, traditional and modern systems in accordance with the degree of structural differentiation and cultural secularization. Virtually all the developing states were designated as modern (or elsewhere as 'modern or modernizing'), along with the secularized city-states such as Athens, the Soviet Union, and all the contemporary Western states (Box 3. 3). Only Ethiopia, described as 'unusually undeveloped and isolated' (Almond and Powell 1966: 218 ft 3) was named as falling into the traditional category.

Box 3.3 Almond and Powell's classification of political systems

CLASSIFICATION OF POLITICAL SYSTEMS ACCORDING TO
DEGREE OF STRUCTURAL DIFFERENTIATION AND CULTURAL SECULARIZATION

I	PRIMITIVE SYSTEMS: INTERMITTENT POLITICAL STRUCTURE A. Primitive Bands (Bergdama) B. Segmentary Systems (Nuer) C. Pyramidal Systems (Ashanti)

II	TRADITIONAL SYSTEMS: A. Patrimonial Systems (Ouagadougou) B. Centralized Bureaucratic (Inca, Tudor England, Ethiopia) C. Feudal Political Systems (Twelfth-Century France)

III

MODERN SYSTEMS: DIFFERENTIATED POLITICAL INFRASTRUCTURES

A. Secularized City-States:
   Limited Differentiation (Athens)
   B. Mobilized Modern Systems:
   High Differentiation and Secularization

1.	Democratic Systems: Subsystem Autonomy and Participant Culture a. High Subsystem Autonomy (Britain) b. Limited Subsystem Autonomy (Fourth Republic France) c. Low Subsystem Autonomy (Mexico)

2.	Authoritarian Systems: Subsystem Control and Subject-Participant Culture a. Radical Totalitarian (U.S.S.R) b. Conservative Totalitarian (Nazi Germany) c. Conservative Authoritarian (Spain) d. Modernizing Authoritarian (Brazil)

C. Premobilized Modern Systems:
Limited Differentiation and Secularization
1. Premobilized Authoritarian (Ghana)
2. Premobilized Democratic (Nigeria prior to January 1966)

Source: Almond and Powell 1966: 217.

  Within the large set of modern states (leaving aside the special case of secularized city-states) a primary distinction was drawn between 'mobilized' and 'pre-mobilized' states, and within each group a distinction was made between authoritarian and democratic forms. The key distinction between traditional and modern systems was the shift from a situation in which all functions were carried out by rulers, leaders or officials, to one in which a separate political infrastructure emerged in which citizens played a part. Mobilized modern systems had in common a higher degree of structural differentiation and secularized political culture. Those with autonomous infrastructures and predominant participant cultures were democratic, while those with a non-autonomous infrastructure and mixed subject-participant cultures were authoritarian. Among the democratic systems the degree of autonomy of different elements of the infrastructure (parties, media, interest groups) from one another and the degree of ascendancy of the participant culture could vary, producing cases of high, limited and low subsystem autonomy respectively. Every case, however, exhibited 'both legal and actual pluralism or autonomy in the political infrastructure'.
  In comparison, authoritarian and totalitarian systems were seen not as lacking all autonomy, but as 'systems in which formal autonomy is eliminated but in which some measure of real pluralism and competitive process still persists' (272). The extent to which pluralism and competition survived varied in accordance with the character and intensity of the goals pursued by the state, being lowest in the radical totalitarian case of the USSR, somewhat higher in the conservative totalitarian case of Nazi Germany, and higher still in the case of Spain, where the goal was stabilization in a conservative mould, and no equivalent attempt was made to transform society. Although Brazil was mentioned as an example of a fourth variant – a 'modernizing authoritarian' system – no further comment was offered in support of this passing reference.
  From our perspective, the greatest significance attaches to the category of 'pre-mobilized modern systems' into which the great majority of new states were placed. In contrast to the 'mobilized' modern systems, which were characterized by a differentiated infrastructure, a 'rather widespread secular political culture', considerable social and economic development, and the spread of instrumental and participatory attitudes, the premobilized modern systems were ones in which 'the trappings of political modernity – parties, interest groups and mass media – have been imposed upon highly traditional societies' (284). These were the familiar 'non-Western' or 'new' states, the 'historical accidents' in which the effects of colonialism and the untimely diffusion of ideas and practices from the developed world gave political development its problematic character:

Traditional and modern elites manipulate mass activity in both democratic and authoritarian variants of premobilized systems, yet the fragmentation among the elites and the ultimate inability to satisfy aroused mass aspirations create a continuing spiral of frustration and instability (285).

Definitive classification of such regimes, as attempted for example by Coleman in the Politics of the Developing Areas, was now judged premature, but a contrast was drawn all the same between Ghana under Nkrumah and Nigeria in the First Republic as representing authoritarian and democratic 'stances' respectively, with Almond and Powell (1966: 290) gracefully noting that the decision to use the two as examples of premobilized systems was made 'before the recent military coups (in 1966) in the two nations'.
  Against the background of this classification of political systems, the concluding chapter embarked upon an extended assessment of the descriptive, explanatory and predictive power of the approach to political development organized around the ideas of structural differentiation, autonomy and secularization. It offered as its central theoretical statement the claim that 'the development of higher levels of system capabilities is dependent upon the development of greater structural differentiation and cultural secularization (323), adding that in structural terms this meant a 'rational' bureaucracy, 'something like a modern interest-group or party system', and 'differentiated political communication structures' (323-4). The classification system offered was defended as a basis for prediction on the grounds that as all political systems were to an extent 'prisoners of their past', it grouped them 'according to the kinds of futures they face' (302). The discussion centred upon the question of democracy in modern systems, and suggested that the twin processes of structural differentiation and secularization eventually forced a choice between extending or containing subsystem autonomy. In other words, the primary contrast that Almond and Powell sought to illuminate was between modern democratic and authoritarian systems, and the key argument proposed was that whereas in democratic systems the level of subsystem autonomy and the responsive capacity of the state were high, in authoritarian systems both were low. The contrast was at its most extreme in the cases of 'Anglo-American democracy and Soviet totalitarianism' (312).
  At this point, a serious break occurred in the development of the theoretical implications of the approach, with devastating implications for the supposed connections between the functional approach on the one hand and the politics of the developing states on the other. Almond and Powell now turned to the 'premobilized modern systems' with the comment that 'though in form these may appear to be democratic or authoritarian systems, they are in fact only at the beginning stages of the differentiation and secularization processes' (313). Thus the classification was provisional, and carried no predictive power:

When one calls a political system at this stage of development 'democratic' or 'authoritarian' one refers not to a functioning political system, but rather to what might be thought of as a 'stance' at the beginning of a developmental process, and one that may change quickly or without much prior warning (313–14).

  The same breach between the development of the analytical framework and its application to developing states was repeated in the discussion which followed of the four developmental processes of state-building, nation-building, participation, and distribution. This offered an extended comparison of the cases of Britain, France, and Germany, but it made no reference at all to the implications of the comparison for the developing world.
  With the theoretical status and the developmental dynamics of the new states both left uncertain, Almond and Powell turned to the predictive capacity of the model, and offered three generalizations. First, as noted above, the development of higher levels of system capabilities was dependent upon the development of greater structural differentiation and cultural secularization. Second, all systems would encounter 'the system development problems of nation building, state building, participation, and distribution' (324). And third, systems with low subsystem autonomy would find it difficult to develop a broad responsive capability. At this point a second break occurred in the development of the theoretical implications of the approach. After asserting that these three 'general theoretical statements' made it possible to 'move in a number of directions in formulating an increasingly specific theory of political change' (325), Almond and Powell decided unceremoniously to abandon the line of argument. Rather than pursue it further in relation to the case of the developing societies which were ostensibly at the core of their theoretical project, they opted to invoke the hitherto unconsidered and untheorized variable of leadership, and to apply this to the specific case of their 'premobilized modern systems'. In so doing, they again moved directly back onto the terrain covered by previous discussions of the 'non-Western political process'. Thus they turned their attention to 'political investment strategies' in the new states in the light of the dilemma posed for their leaders by the explosion of participation and 'the image of the modern and democratic state which, given the social and cultural conditions of their societies, is unattainable in the immediate future' (327). Noting the appeal of the Marxist-Leninist model and the need to find a viable alternative, they offered a 'rational choice' perspective focused on the identification of an investment strategy with the highest chance of success. In so doing they reformulated, in suitably neutral language, Pye's more direct assertion that the new states wanted too many things which they could not have:

The approach followed in many of the new nations, which involves simultaneous investments in the development of all the capabilities including the responsive and distributive ones, seems to be a high-risk, low-benefit strategy. The human and material resources simply are not there to produce this kind of simultaneous solution of the problems of state and nation building, participation and distribution. There must be some scheduling, some system of priorities (329).

After all the efforts expended on the elaboration of the functional approach, then, they were back with the core ideas spelled out in the previous chapter. And it was in the light of these considerations, rather than the conclusions derived from the functional framework itself, that they turned to 'elaborating the logic of public policy as it relates to political development', and advised rulers in new states how best to proceed. The advice they offered was familiar: they should stress state- and nation-building in the first stages over participation and welfare, preserve a degree of pluralism in order to keep options open, make compensatory investments to cope with the disruptive consequences of the modernization process, and pursue complementary investment strategies in education, industrialization, family structure and organization, and urban and community planning. Reflecting on the enormity of the task, Almond and Powell opined that 'the modern political scientist can no longer afford to be the disillusioned child of the Enlightenment, but must be become its sober trustee', (332), and it was on this suitably sententious note that the text came to an end.

Conclusion

The general conclusion to be drawn from this sorry tale is that the functional approach was not only rather a mess, but also largely an irrelevance. Beyond its various deficiencies, widely noted by critics of varying persuasions, its most striking feature was that it was declared by its own authors to be of limited value in addressing the specific case of the developing states (or 'premobilized modern systems'). These continued to be addressed within the framework of 'non-Western states', filtered through the core assumptions of the revisionist approach to democracy.
  As a result, the functional approach never quite got off the ground. Almond himself has stated that his theoretical introduction to The Politics of the Developing Areas was written in virtual isolation from the accompanying case studies (Almond and Coleman 1960: viii; Almond 1970: 18); and the amended framework offered in Comparative Politics: A Developmental Approach failed to illuminate the politics of the developing societies which were its intended target.
  One reason for this was that even to its authors the approach seemed both unfinished and hopelessly over-ambitious. In The Politics of the Developing Areas, Almond look forward to the eventual statistical treatment of the key categories of function, structure, and style, but was forced to admit that none of these had yet been appropriately identified:

The set of political functions which we have proposed is most preliminary. We cannot really say that we have developed a set of functional categories which will prove satisfactory for purposes of analyzing and comparing political systems. We must be even more tentative about the structural categories. Here we have simply used the nomenclature of political and social institutions without pretending to have arrived at clearly defined, universally applicable categories of structure. Finally, in comparing styles of performance of function by structure we have relied in the main upon the pattern variable concepts of Parsons and Shils (Almond 1960: 62).

  Even if these problems were resolved, he confessed, calculation of the possible combinations of types of function, structure and style would produce 'a matrix of several hundred cells', and

if we were then to attempt to sample the actions of a number of polities over a given period of time in order to arrive at precise comparisons of these polities in terms of frequencies of performance of function, by structure, by style, we should have set ourselves a research task of ridiculous proportions (62).

Six years later, this impossible dream was, not surprisingly, no nearer realization. The conclusion to Comparative Politics: A Developmental Approach described the goal of a generalizing and predictive political science as 'a most hazardous and difficult task', and claimed no more than to have assembled some 'preliminary exercises pointing the way toward a theory of political development' (Almond and Powell 1966: 300, 322). It is not surprising, therefore, that little light was thrown on the politics of the developing states themselves.
  The theoretical fruits of the exercise were indeed meagre. As we have seen, they were limited to the assertions that all political systems must face the challenges of nation-building, state-building, participation and distribution; that in order to develop the capabilities required to do so they must undergo processes of cultural secularization, and structural differentiation reflected in the emergence of interest groups, political parties and rational bureaucracies; and that systems with low subsystem autonomy would not develop a high responsive capacity. At the same time, it was conceded, in the more refined version of the approach offered in 1966, that there was no reason to believe that political systems in the real world would actually tend to develop in the direction of secularization and differentiation.
Not surprisingly, the approach attracted substantial criticism. Objections were made to its ethnocentricity, its inherent conservatism, its inability to produce explanations for change, its extraordinarily pretentious and cumbersome use of jargon, and its lack of operationalization in terms of core hypotheses (Chilcote 1981: 178–182). From my perspective, concerned as I am with the case of the developing states and with the insights which flow from a critical Marxist approach, these objections, weighty as they are, are not fundamental. The core problem which generated all these failings, from the failure to produce change to the appearance of an ethnocentric fixation on 'Western' institutions, was the treatment as universal of a process which was historical, and specific to a particular form of society. This falsely universalized 'functional' model was superimposed upon a set of assumptions, carried over more or less wholesale from the revisionist theory of democracy, which simply resurfaced and carried the argument at the crucial points at which the 'functional' model was abandoned. The problem with the functional approach was not so much that it was ethnocentric, but that it smuggled into a supposedly universal framework a set of assumptions and requirements which were specific to the logic of capitalist societies.
  This can be seen in three interesting features of the approach which differentiated it from the cultural, institutional and comparative historical approaches reviewed in the following chapters. First, it abstracted away from the particular substantive goals of individuals to the political processes in which they became involved as they sought to secure them, treating those goals simply as presenting themselves to the 'system', ready made, as input. Second, it saw different social systems (the political system itself, the economy, the religious community, the family, and so on) as separate behavioural realms, with individuals passing from one system to another as they performed different roles. In particular, it saw the polity and the economy as separate spheres. Third, although it took the idea of the legitimate use of coercion and hence the standard Weberian definition of the state as its starting-point, it blurred the distinction between the government and the citizen, or the state and society, by adopting a very broad definition of the political, and including in its set of 'conversion functions' the activities of citizens organising politically, as well as the authoritative actions of the government itself, without making any distinction between them. Hence the six 'conversion functions' captured the activity of three 'state' institutions – executive, legislature and judiciary – and three 'social' institutions – political parties, pressure groups and the media. In its deliberately neglect of substantive goals, its separation of the economic and the political in accordance with behavioural rather than institutional criteria, and its blurring of the dividing line between state and society – the functional approach was quite distinctive.
  The significance of the fact that the political systems of the developing countries defied definitive classification, and that the attempt to explore the theoretical implications of the functional model was abandoned before their case was considered, is that it reveals this underlying characteristic of the literature, and confirms the fundamental commitment to a prescriptive model intended to promote pro-Western capitalist development, rather than to a scientific understanding of political systems. For all that they were presented as implications of the functional approach, the prescriptions addressed to leaders in new states with which the volume concluded depending upon the prior assumption that elite-led pro-capitalist liberal democracy was currently unattainable. This prompted the expedient adoption of a 'rational choice' model focused on 'political investment strategies', centred on a familiar strategy – an elite-led process of development focused in the first instance on state- and nation-building, with increased levels of participation and welfare postponed into the indefinite future. In its essentials – and not least in its insistence on the extent to which democratic politics required intensive preliminary social engineering – this reflected the practical orientation of the revisionist theory of democracy.
  This is scarcely surprising. The point of departure of the functionalist approach – the assertion of the need for 'good boundary maintenance' between the political and social systems – was simply a restatement in a jargonized form of the insistence by the revisionist theorists that political institutions should be protected from direct popular pressure. And the major political institutions Almond proposed to achieve this end – associational interest groups, and pragmatic parties able to bargain and compromise within an overarching consensus – were precisely those celebrated by revisionist theory. It was quite natural, therefore, that when the attempt to theorize the developing states through the functionalist approach broke down, Almond and Powell should have returned to the revisionist theory and spelled out its practical implications for leaders in developing societies regardless. For all its strident opposition to institutionalism, the functional approach retained a strong link with the leading institutions of Western political systems, deriving its functional categories directly from them and endorsing them as the only possible framework for system capability in the modern world. Seen in these terms, the 'functional' content was an excrescence, an elaborate but ultimately irrelevant superstructure which could be discarded without damaging the underlying message.
  The functional approach had some importance strengths, but its serious shortcomings limited the contribution it could make to the understanding of political development. On the positive side, it was a genuinely systemic theory which sought to grasp the logic of the working of the political system as a whole, in conjunction with other related economic and social systems. And within this overall systemic approach it rejected the simplistic dualism which remained a constant temptation in modernization theory, and in particular advanced the important argument that apparently 'traditional' practices such as as political patronage or recruitment on ascriptive criteria might persist and play highly functional roles in 'modern' Western societies.
  However, its goal of a truly universal theoretical framework for the analysis of political systems was arguably fundamentally misconceived, and certainly imperfectly realized. If the dilemmas confronting the developing states arose from the pressures of global modernity, it mattered more to analyse those pressures than to establish a transhistorical framework able to encompass the very different situation of primitive and historical societies. As it was, the analytical exercises pursued were marked by substantial discontinuities, focusing at one moment upon comparisons between primitive and modern systems, at another upon 'Anglo-Saxon' and continental European systems, and at yet another upon 'democratic and totalitarian' systems. The modernizing new states that were central to the second agenda of developing a prescriptive theory of political development rarely took centre stage, and when they did, as in the concluding pages of Comparative Politics: A Developmental Approach, Almond and Powell were led to admit that the theory developed did not provide a means of assessing their developmental status, or predicting the direction in which they might develop. As a result, after all the effort expended upon the creation of a universal functional framework of analysis, the politics of the developing areas were addressed in much the same terms as they had been in the mid-1950s when the 'non-Western political process' was first described. It is significant in this respect that despite the effort to build a developmental dimension into the approach, in order to capture the dynamics of change, it was precisely where political systems were undergoing such change, and were in a state of fluidity, that the approach was declared to be incapable of grasping the characteristics of the political systems concerned. Almond and Powell constructed a theory of political development which was at its weakest when dealing with developing political systems. In the absence of a theoretical advance, the pragmatic advice that was offered mirrored the preconceptions already built into political development theory as a result of the assumptions it shared with the revisionist theory of democracy. As a result, the prescriptions offered were independent of the primary theoretical agenda. Instead of addressing issues of structure and function, they focused on questions of priorities and leadership within a given institutional framework. It is not surprising that many of Almond's contemporaries also decided that it was better to approach these issues directly, without the enormous costs and doubtful benefits of the prior construction of a comprehensive functional framework.