Ian Roxborough Unity and Diversity[1]

Embed Size (px)

Citation preview

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    1/27

    Unity and Diversity in Latin American HistoryAuthor(s): Ian RoxboroughReviewed work(s):Source: Journal of Latin American Studies, Vol. 16, No. 1 (May, 1984), pp. 1-26Published by: Cambridge University PressStable URL: http://www.jstor.org/stable/157285 .

    Accessed: 05/03/2012 20:11

    Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

    JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

    content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

    of scholarship. For more information about JSTOR, please contact [email protected].

    Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access toJournal of

    Latin American Studies.

    http://www.jstor.org

    http://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/stable/157285?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/157285?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=cup
  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    2/27

    J. Lat. Amer. Stud. I6, I-26 Printed in Great Britain

    Unity and Diversity inLatin American History*by IAN ROXBOROUGH

    IntroductionThis article has its origins in a generalized feeling of dissatisfaction withcurrent theories about political developmental trends in Latin America.It is an early statement of a series of arguments which will subsequentlybe developed in a forthcoming book.1The veritable explosion of empirically grounded monographs in the lastfifteen or twenty years has made the task of producing a synthetic accountof Latin American development simultaneously more pressing and moredifficult: more difficult because it has made simple explanatory modelsharderto sustain, and has opened up the accepted historiography to seriousand widespread revisionist attack; more pressing because many, if notmost, social scientists accept the need to develop a theory of social changewhich is historically grounded, capable of explaining large-scale socialtransformations. My concern in this article is with the methodologicalissues involved in the formulation of an adequate theory of Latin Americandevelopment, ratherthan with establishing new facts. There is considerablehistoriographical controversy over many of the events discussed in thisarticle, and in these cases I have made my own judgement about where,on balance, the evidence points. This is, naturally, a tentative andprovisional matter, and future research, or sources of which I have notpreviously been aware, may well necessitate a reformulation of some ofthe propositions advanced below. The methodological points which formthe core of this article are, however, a largely separate issue and must standor fall on their own merits.

    The countries of Latin America have a rich and complex history. So* Based on a section of Ian Roxborough's forthcoming book Latin America: Class, Stateand Developmento be published by Macmillan, London and Basingstoke 1985.1 bid.

    LAS I6

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    3/27

    2 Ian Roxboroughrich, complex and varied has this history been, indeed, that many scholarshave seen the task of historians as that of producing twenty separatehistories, eschewing any attempt at generalization for Latin America as awhole. Despite the temptations of what might be termed 'nationalempiricism', the common features of the republics of Latin America - interalia an Iberian colonial past, approximately coterminous achievement ofindependence, integration into the expanding world economy of the latenineteenth century as producers of primary products, and seeminglysimilar patterns of economic and political development in the twentiethcentury - mean that there have been and will continue to be powerfulpressures in the direction of producing a coherent and integrated accountof the pattern of Latin American history as a whole. This article willattempt to survey some of the major approaches to this task, and willconsider some of the methodological and empirical criticisms and con-siderations which arise. The chronological focus will be on the periodsince the mid-nineteenth century.Three principal approaches may be discerned. Firstly, there are historiesof Latin America which claim to see a common theme or principleunderlying the histories of the twenty republics. Overriding any national'peculiarities' are the commonalities of the specifically Latin Americanexperience. One writer has argued, for example, that the Latin Americancountries all share a common Iberian tradition of corporatism, and thatthis is the central factor explaining the distinctly Latin American patternof development.2 Another has pointed to the 'centralist tradition' ofgovernment inherited from the Spanish and Portuguese colonists.3 Froma dependency perspective, various writers have attempted to show howthe nature of the economic system imposed by the Iberian conquest - the'colonial heritage' - has perpetuated a syndrome of underdevelopment.4Other explanations, cultural, racial, geographical, etc., may readily befound in the literature. From such a perspective, the history of LatinAmerica is seen as the unfolding or acting out of some essence ordemiurge, or possibly (though this is less common) of the dialecticalinterplay of two such essences or principles. Deviations from this modelare seen either as variations on a theme or simply as deviations, to be

    2 H. Wiarda, Corporatismand National Developmentn Latin America (Boulder, WestviewPress, 198 ).3 C. Veliz, The CentralistTradition n Latin America(Princeton, Princeton University Press,1980).4 S. Stein and B. Stein, The Colonial Heritage of Latin America (New York, OxfordUniversity Press, 1970); E. Galeano, Open Veins of Latin America (New York, MonthlyReview Press, I973).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    4/27

    Unity and Diversity in Latin American History 3ignored or explained in a purely ad hocmanner. I will refer to this approachto Latin American history as the 'essentialist' approach.

    The second approach discussed in this article is really a variant of thefirst. However, because of its widespread popularity, it merits separatetreatment. This is the notion that the various nations of Latin Americaunderwent a broadly similar and parallel development which may bedescribed as a sequence of historical stages. Following Malloy, this willbe termed the 'modal pattern' approach.5 Typically, it is claimed that theeconomies of the Latin American countries were first orientated towardproduction of primary products for the world market (the phase of'development toward the outside'); as the stimulus of the world marketfaltered in the period between the outbreak of World War I and the closeof World War II, the Latin American economies turned in on themselvesand began to industrialize. By meeting domestic demand for manufacturedgoods from internal production, rather than through export revenues, theybegan the second phase, that of import-substitution industrialization (ISI).Finally, in recent decades, following certain internal problems associatedwith ISI and consequent on the massive penetration of the Latin Americaneconomies by multinational manufacturing corporations, it is held thatsome at least of the Latin American economies are entering a new phasewhich has variously been labelled 'associated dependent development', the'Brazilian model', 'bureaucratic authoritarianism', etc.6As in the essentialist approach, deviations from the modal pattern areseen as precisely that, deviations from the basic theme or pattern of LatinAmerican history. The criticisms directed at both essentialist and modalpattern accounts of Latin American history, as will be seen in greater detailbelow, frequently hinge upon precisely this issue: how many exceptionscan be tolerated before the explanatory paradigm must be discarded? Tosome extent, this must be a matter of judgement. Nevertheless, it ispossible to break out of the dilemma by adopting a third approach. Whatis involved is a conscious attempt to steer a course between the Scylla ofnational empiricism and the Charybdis of essentialism by undertakingsystematic comparisons of national histories. The waters of comparativehistory are poorly charted, and reefs and shoals abound. It is littlewonder, then, that only a few scholars have ventured into this area. Indeed,

    5 J. Malloy, 'Latin America: the Modal Pattern', in J. Malloy (ed.), AuthoritarianismandCorporatism n Latin America (Pittsburgh, University of Pittsburgh Press, 1977).6 F. H. Cardoso, 'Associated Dependent Development' in A. Stepan (ed.), AuthoritarianBrazil (New Haven, Yale University Press, I973); G. A. O'Donnell, Modernigation ndBureaucratic-Authoritarianism(Berkeley, Institute of International Studies, University ofCalifornia, 1973).I-2

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    5/27

    4 Ian Roxboroughto the best of the author's knowledge, there are only a handful ofsystematic attempts to reconcile the diversity of national experiences witha systematic and coherent theory for the region as a whole.7 Of these, themost important is, of course, the pioneering work of Fernando HenriqueCardoso and Enzo Faletto, frequently cited and promptly ignored.8 Thethrust of this article will be that it is to the methodology of Cardoso andFaletto that we must look for an adequate account of Latin Americanhistory but, as will be argued in detail below, their substantive accountis in need of radical amendment.

    Having briefly stated the rival positions, we move now to a moreleisurely examination of their respective merits and demerits.Essentialist approaches: Claudio VeligRather than take a large number of writers, I will concentrate on a recentexample of the essentialist approach to Latin American history. In sodoing, I have attempted to select an author whose centrality and reputationprecludes any possibility of attacking a straw man. In a recent andimportant attempt to synthesize Latin American history, Claudio Veliz hasargued that 'a centralist tradition' has dominated the history of thecontinent.9 Latin America, in sharp contrast to its northern neighbour, wasoriginally colonized by highly centralized monarchies. Colonial rule wasauthoritarian, and the interests and affairsof the colonies were subordinatedto the directing influences of Spain and Portugal. The metropolitansocieties sought to regulate colonial life in minute detail. This centralisttradition was further strengthened with the Bourbon and Pombalinereforms. As a result, Latin Americans learnt to look to the state forsolutions to social, political and economic problems; rather than form freeassociations, the natural political response of Latin Americans was toorientate their action directly toward the state. This tradition, argues Veliz,survives today in the form of the centralized military governments whichdominate the region.There was, however, according to Veliz, a 'liberal pause' in what hasbeen an otherwise unbroken history of centralism and authoritarianism.The liberal pause began in the mid-nineteenth century, and started to

    7 F. H. Cardoso and E. Faletto, Dependency nd Developmentn Latin America (Berkeley,University of California Press, 1979; original Spanish edition, 1969); F. S. Weaver,Class,State andIndustrialStructureWestport, Conn, Greenwood Press, 1980); A. Cueva,El DesarrollodelCapitalismoenAmerica Latina (Mexico, Siglo XXI, 1977); V. Bambirra,Capitalismo DependienteLatinoamericano(Santiago, CESO, 1973).8 The reasons for the practical neglect of the Cardoso-Faletto volume are unclear. It wouldbe interesting to explore this topic from the point of view of the sociology ofknowledge. 9 C. Veliz, op. cit.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    6/27

    Unity and Diversity in Latin American Historycrumble in the depression of the interwar period. However, it experiencedsomething of a revival during the Second World War and its immediateaftermath. This liberal pause finally came to an end in the late 'sixties andearly 'seventies.It is not unreasonable in any account of Latin American history to notethat a dominant trend has been interrupted for a while. However, as PaulCammack has noted, the 'liberal pause' acknowledged by Veliz runs fromthe mid-nineteenth century right up to the last decade or two.10 It coversvirtually the entire period of national independence. Of the last 30 yearsof Latin American history (from 185 to 1980), perhaps some thirty years(a mere 23 % of the total) might be described as fitting Veliz's centralistmodel of political behaviour. As an explanation of post-colonial LatinAmerican history, the notion of a centralist tradition is seriously deficient.In reality, for Veliz, only three periods of Latin American history aremarked by authoritarian centralism: the colonial period; the period ofstate-building immediately following independence and prior to thedefinitive integration of the Latin American economies into the expandingcapitalist world economy; and the recent period of military government.That there should be some similarity between the politics of militarydictatorship and those prevailing under the colonial regime and under theearly authoritarian attempts at state-building will surely cause littledispute. But this formal similarity does not imply the existence of a realtradition,given the century-long hiatus of the 'liberal interlude'. That is,this tradition cannot easily be seen as the cause of these forms ofauthoritarianism, given the existence of a liberal pause. As an explanatorydevice, tradition is only of interest to the extent that its continuity,embodied in institutionalized practices of socialization and cultural trans-mission, acts as an effective cause of system stability. Judged entirely in itsown terms and on the basis of its own interpretation of the historicalrecord, the theory of a centralist tradition must be found seriously wanting.There are, not surprisingly, other criticisms which could be adduced.For example, the colonial period was not one of unbroken centralism.During the seventeenth century and part of the eighteenth, imperialcontrol over the American colonies was both lax and ineffective. Only withthe absolutist reforms of the later eighteenth century was a consistentattempt made to control the colonies from the metropolis. It is, thus, byno means clear that even the period of colonial history can be accuratelycharacterized by reference to the notion of a centralist tradition. Nor isit easy to use that notion to describe contemporary Latin America. The10 P. Cammack, eview of Vliz, 'CentralistTradition nLatinAmerica',DurhamUniversity

    Journal,vol. 75, no. 2 (1983), pp. II8-119.

    5

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    7/27

    6 Ian Roxboroughincidence of military rule has been uneven and, in some countries, thesoldiers have returned to their barracks and allowed civilians to assumepower.

    Taking Claudio Veliz as an example, I have argued that the notion thatLatin America as a whole has some essential unity and persistence, bothin time and space, is an inaccurate reading of the historical record, whichis one of diversity and change. Similar criticisms could be made of otheressentialist theories, though space precludes a fuller discussion in thisarticle.The modal patternThe most widespread interpretation of Latin American history owes muchto dependency theory. It links the stages of growth of the Latin Americaneconomies to their changing relations with the world economy. Thisframework has been utilized by a number of writers, among whom theworks of Celso Furtado, Guillermo O'Donnell, and Osvaldo Sunkeldeserve special mention.ll Here, the short article by James Malloy in hisedited book, Authoritarianism and Corporatism in Latin America, will betaken as a representative statement of the basic position.12

    The expansion of the world economy in the second half of thenineteenth century brought with it an increase in demand for foodstuffsand raw materials. The majority of Latin American countries tookadvantage of this increased demand from the metropolis and began toexport growing quantities of primary products to Europe and the UnitedStates. In turn, they provided expanding markets for the manufacturedgoods of the industrial economies. From this period until the i930s, LatinAmerican countries pursued a growth model orientated toward theoutside.

    As the links between metropolis and satellite were weakened by the FirstWorld War, and then by the depression of the 1930S and the Second WorldWar, the Latin American economies faced sharp drops in export revenues,and, hence, their import capacity was curtailed. This situation, together withtariff barriers and certain economic policies which effectively operated as

    1 C. Furtado, EconomicDevelopmentof Latin America (Cambridge, Cambridge UniversityPress, 1970); 0. Sunkel and P. Paz, El SubdesarrolloLatinoamericanoy la Teora delDesarrollo(Mexico, Siglo XXI, 970); D. Collier (ed.), TheNew Authoritarianism nLatinAmerica (Berkeley, University of California Press, I979).12 J. Malloy, op. cit. Naturally, many writers differ in significant aspects from the positionput forward by Malloy. My claim is that writers in this tradition all share a broadlysimilar framework for the analysis of Latin American history; Malloy has been chosenbecause his work is perhaps the clearest and most succinct example of this approach.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    8/27

    Unity and Diversity in Latin American Historyforms of the Keynesian multiplier, provided a stimulus for what becameknown asimport-substitution industrialization. 13Beginning with consumergoods, items which had previously been imported were now manufacturedin Latin America.

    However, capital and intermediate goods and raw materials for industrystill had to be imported, and this - in the booming world economy of thepost-war decades - led to the increasing penetration of Latin Americaneconomies by affiliates of manufacturing multinationals. This marked thebeginning of a third phase of Latin American economic growth,conventionally dated from the mid-'fifties or 'sixties.

    The key feature of this third phase, which we will call, followingO'Donnell, the bureaucratic-authoritarian (B-A) phase, lies in the changingcomposition of effective demand.14In the ISI phase, locally manufacturedgoods (such as textiles and food-processing) were aimed at a broadspectrum of the population. In the B-A phase, the stimulus to economicgrowth was provided by the consumer durables sector: primarily cars andvarious electrical appliances. Since only the high-income sections of thepopulation could be expected to afford such goods, continued economicgrowth required that increases in market demand be concentrated in thehigh-income strata. The logic of capital accumulation in this phase, it wasargued, would lead to an increasingly regressive distribution of income.The contrast with the ISI phase could not be more stark: the earlier phaseof industrialization depended on continually expanding the purchasingpower of the bulk of the population, particularly lower income groups.15These differences in the distribution of market demand had theirpolitical corollaries. The period of ISI was accompanied by a displacementof the old agrarian oligarchies from state power and by the mobilizationof previously excluded classes and strata. A heterogeneous coalition ofindustrialists, the urban middle class, urban workers and migrants to thecities led this assault on the oligarchical state. Frequently with the aid ofsections of the military, this coalition led to the installation of populist orBonapartist regimes and to a new level of state autonomy from direct classpressures. Industrial expansion, growth in employment, and widespreadrises in living standards were the material bases for the widespread supportenjoyed by these populist governments.13 C. Furtado, The EconomicGrowth of Brazil (Berkeley, University of California Press,

    I965).14 G. A. O'Donnell, ModernizationandBureaucratic-Authoritarianism,p. cit.15 Economic Commission for Latin America, DevelopmentProblems in Latin America(Austin, University of Texas Press, 1970).

    7

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    9/27

    8 Ian RoxboroughWith the advent of the B-A phase, the picture altered radically. Now,

    working class wages were to be constrained and popular mobilizationdampened. As a consequence, the transition from the ISI phase to the B-Aphase was accompanied by an increase in state repression, often in the formof a military take-over of the government. Durable and long-termmilitary-technocratic regimes were installed in one country after another:BJrazil n I964, Argentina in I966, Chile and Uruguay in I973. The modelwas elaborated in the wake of this series of military interventions, and waslargely designed to provide an explanation for what was seen as a universal(or at least very widespread) phenomenon in Latin America. Thus, themodal pattern of Latin American development supposedly discerned bythis influential body of scholars postulates three stages of economicgrowth, each of which has definite political correlates.

    Table i. The 'modalpattern' for Latin AmericaDominant State PoliticalCentre Periphery classes form movements

    1850/70-1930 Industrial Exportof Agrarian Oligarchic Middle-classmaturity primary export parliament- radicalof G.B. products oligarchy arianism challenge1930--I960 U.S. dominance I.S.I. National Bonapartism Populistbourgeoisie expansioni96o Latemonopoly Neo-liberalism Associated Authoritarian Authoritariancapitalism: M.N.C. monopoly corporatism exclusioncrisis of U.S. domination capitaldominance

    Three principal questions may be asked of this model: (i) Is theconnection between the logic of capital accumulation and the economicand political policies actually adopted as determinate as the modelsuggests? That is, are the causal linkages in the chain of theoreticalreasoning adequate? (2) Do any or all of the Latin American economiespass through this threefold sequence, and is the timing that specified bythe model? (3) Is there empirically any association between these threephases of economic growth (assuming they exist) and forms of politicalbehaviour?

    It can be argued that the modal patternmodel is seriously deficient vis-d-visall three objections, and that there are sufficient grounds, both logical andempirical, for rejecting it, even in terms of the countries which served asthe point of departure for the formulation of the model. In this article,the focus will be on the third of these possible criticisms of the modal pat-tern: the existence or otherwise of a correlation between the economic and

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    10/27

    Unity and Diversity in Latin American Historythe political processes.16 Since the interest in any interpretative schema liesin linking economic, social and political processes, this third criticism iscentral.

    The countries which have inspired most theorizing in this vein havebeen Argentina and Brazil. Yet even here there aregrounds for questioningthe utility of the modal pattern model. Elsewhere, of course, there is evenless fit between the model and historical reality. We will concentrate ontwo central issues: does populism occur in the period I930-6o?; and isthere a trend towards corporatist exclusion in the period after I960? Firstit is necessary to define 'populism', a daunting task in itself. The term isoften defined in a purely ostensive manner, namely populism is that whichoccurred during the governments of Cairdenas I 934-40), Per6n (1946-5 5)and Vargas (1930-45). A somewhat broader focus is that of lanni, forwhom everything that happened in Latin America between 1930 and I960may be described as 'populist'.17 However, ostensive definitions areinherently unsatisfactory and ought to be replaced with stipulativedefinitions. There is some considerable controversy on this subject, withtheorists falling into two broad camps.On the one hand, the 'classical' definition of populism sees it as aphenomenon combining a particular form of ideology with certainorganizational and social structural features. For 'classical' theorists,populism is aloosely organized multiclass movement united by acharismaticleader behind an ideology and programme of social justice and nationalism.In their view this linkage of ideology and organization is the strengthofthe definition; it links ideology with a definite mode of political partici-pation. There is an implicit or explicit contrast with the supposedlyclass-orientated nature of politics in the advanced industrial societies ofWestern Europe. Popular participation in populist movements, it isasserted, does not take on a 'class' character. Either the subordinate stratacompose a 'mass' or the working class does not yet have its ownautonomous organizations. It tends, therefore, to be organized and led by16 Other criticismsof the B-A model maybe found in J. Serra, Three MistakenThesesRegardingthe Connectionbetween Industrialization nd AuthoritarianRegimes' inD. Collier(ed.), TheNewAuthoritarianism;. Roxborough,'State, Multinationals ndthe Working Class in Brazil and Mexico', in C. Lewis and C. Abel (eds.), EconomicImperialismn Latin America London, Athlone Press, forthcoming);M. Wallerstein,'The Collapseof Democracyin Brazil:its Economic Determinants',Latin AmericanResearch Review, vol. 15, no. 3 (I980); K. Remmer & G. Merkx, 'Bureaucratic-Authoritarianism Revisited', Latin American ResearchReview, vol. 17, no. 2 (I982);G. O'Donnell, 'Replyto RemmerandMerkx',LatinAmericanResearch eview, ol. 17,no. 2 (1982).17 0. Ianni, La Formaciondel Estado Populistaen America Latina (Mexico, ERA, 1972).

    9

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    11/27

    Io Ian Roxboroughother social classes or political forces in a heteronomous manner. Theabsence or weakness of an autonomous working class is central to theclassical definition of populism.18

    Opposing this definition, there is what might be called the 'discourseanalysis' school of thought. For these theorists, populism is first andforemost an ideologicalphenomenon: populism is an ideology, or strandof an ideology, which asserts that political conflict is between 'the people'(who are seen as the 'nation') and the 'oligarchy' or 'imperialism'.19Hence, whenever a political leader makes an appeal to 'the people' he orshe is a populist, and his or her movement or government may byextension be described as 'populist'. In this definition, nothing is saidabout the causes of populism, about its social base, about its policies, orabout its organizational style. Having defined the phenomenon withreference to a particular form of ideological appeal, it remains a matterof empirical enquiry as to whether this appeal to the people is associatedwith particular features in the genesis, structure, social base, and outcomeof the movement. The 'discourse analysis' definition is, therefore, aminimal definition; while some theorists would argue that populism ismore than this, none would argue that it is less than this. Moreover, ifpopulism is an 'appeal to the people', it does not follow that those to whomthis appeal is addressed actually heed it. Whether the subordinate classesand strata of Latin America support populist leaders because they acceptthe validity of this appeal to the people, or whether they support suchleaders because this is a rational choice among alternatives, given theirinterests, remains an open matter.A corollary of this minimal definition of the term is that one must expectto find 'populism' in a wide variety of contexts, and this is, indeed, the casein Latin America. Populist appeals to the people were made by theChristian Democrat Eduardo Frei in Chile in the i96os, by Mexicanpresidents throughout the twentieth century, by the Peruvian military18 Clearly, a great variety of differing approaches fall under the 'classical' definition of

    populism. Essential to all of them is the contrast between the types of action andorganization supposedly associated with class actors and that characteristic of situationsof low 'classness'.19 What I regard as the 'classical' position is represented by many of the works in G.Ionescu and E. Gellner (eds.), Populism(London, Weidenfeld and Nicolson, I969), and,more recently, N. Mouzelis, 'Ideology and Class Politics', New Left Review,no. 112,(1978). The principal opposing view of populism (which I refer to as the 'discourseanalysis' approach) is set out in E. Laclau, PoliticsandIdeologyn Marxist Theory London,New Left Review, I977); E. de Ipola, Ideologiay Discurso Populista (Mexico, FoliosEdiciones, 1982); H. Osakabe, Argumentafao DiscursoPolitico(Sao Paulo, Kair6s, 1979);G. G. Debert, Ideologiae Populismo(Sao Paulo, Quieroz Editor, I979).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    12/27

    Unity and Diversity in Latin American History 1government of the late 'sixties and early 'seventies, and by a motley hostof civilians and military men in a wide variety of contexts. The proponentsof the modal pattern model, however, adopt the 'classical' definition ofpopulism, and see in the paradigmatic cases of Per6n, Vargas andCardenas attempts to mobilize subordinate strata from above into multi-class coalitions behind an industrial-developmentalist project of the Stateand/or the national bourgeoisie.The available evidence, however, suggests that the fit between the clas-sical definition of populism and these three cases is open to question.Indeed, of the three paradigmatic examples of 'populism' (according tothe ostensive definition), one of them - the 1930-45 Vargas government -fails even to fit the minimal definition. During the period from his seizureof power up to the calling of elections after the Second World War, GetulioVargas' discourse was not noted for the frequency of its appeals to the'people'. His was a conservative, authoritarian and demobilizing regimethat is hard to describe as 'populist' under any definition.20 It was onlywith the advent of competitive politics in the period after 945 that Vargasmade any sustained appeal to the people. In this second period, indeed,Vargas may justly be described as a populist, and the heyday of Brazilianpopulism may perhaps be dated as 1945-64. But Brazilian ISI was wellunder way in the I930S and had begun considerably earlier. The 'fit'between economics and politics in this paradigmatic case is ratherloose.In the cases of Cardenas and Peron there is little doubt that they meetthe minimal definition of populism. But the definition of populism usedby theorists working within the modal pattern framework goes beyondthis minimal definition to include the notion that mass support for populistmovements was not organized primarily along class lines. That is, thesupport for populist leaders was not amulti-class alliance, with independenttrade unions lending the support of an autonomously organized workingclass to a Bonapartist figure, but rather an amorphous mass movement orcoalition with direct ties between individuals and the charismatic leader.2120 T. Harding, 'The Political History of OrganizedLabor in Brazil', Ph.D., Stanford

    University, 1973, p. 127; A. C. Bernardo, Tutela e autonomiasindical: Brasil i90o-I94y(Sao Paulo, Quieroz Editor, I982).

    21 I make a distinction here between 'alliance' and 'coalition' which is seldom madeexplicitby 'classical' theoristsof populism.Most politicalforces arecoalitions in thesense that their supportersare drawn from a variety of classes; to talk about class'alliances', however, suggests that organizationswhich are representative f only oneclass(e.g. tradeunions)deliberatelyandconsciouslyform a compactwith other socialactorsto furthertheir mutual nterests.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    13/27

    I 2 Ian RoxboroughThis analysis springs basically from Weber (charisma) and Durkheim viamass society theory.22For the classic definition of populism to have any utility, it must bedemonstrated that we are dealing with a situation in which subordinateclasses or strata are organized into the populist coalition in a heteronomousmanner. If this is not the case, then we are dealing with an alliance ofclasses, rather than with populism. The available evidence suggests that,at least in the early phases, both Per6n and Cardenas were supported byautonomously organized working class institutions, i.e. relatively indepen-dent trade unions. These movements may, therefore, be analysed in termsof more or less explicit and deliberate alliances of the working class withindividuals holding state power. No reference to the concept of populismis necessary to explain this, nor does the notion of populism add anythingto the analysis. It is only at a later stage that the unions lose their autonomyand the working class becomes subordinate to the state.23 Empirically,neither early Peronism nor the Cardenas government fit into a classicialdefinition of populism.In the case of the earlyPeronist movement, Juan Carlos Torre has shownhow the supposedly 'spontaneous' demonstration of the poor people ofBuenos Aires on I 7 October 1945 to demand the release of Peron had beencalled for and organized by the principal trade-union confederation, theCGT.24Moreover, at the meeting of the CGT's executive committee wherethat decision was taken, the union leaders were deeply divided over theirattitudes to Peron. On the one hand were those who wished to havenothing to do with Per6n; in a slight majority were those who, despiteconsiderable distrust and misgivings, decided to support Peron as a lesserevil.25 Moreover, the notion that it was recent migrants to Buenos Airesfrom the interior - a 'new' working class - which provided Per6n withhis principal base of support has come under sustained attack.26Little, and22 See W. A. Kornhauser,ThePoliticsof MassSocietyGlencoe, The Free Press, i959).23 J. Horowitz, 'The Impact of the Pre-1943 Labor Union Traditions on Peronism',

    Journalof Latin American Studies, vol. 5, part i (I983); T. Di Tella, 'Working-ClassOrganizationand Politicsin Argentina',Latin AmericanResearch eview, ol. I6, no. 2(1981).24 J. C. Torre,'Sobreas Origensdo Peronismo',EstudosCEBRAP, no. i6 (Abril-Junho1976).25 Ibid.On theoriginsof Peronism,seealso W. Little,'The PopularOriginsof Peronism',in D. Rock (ed.), Argentinan theTwentieth enturyLondon, Duckworth, I975), andM. Murmisand J. C. Portantiero eds.), EstudiosSobreLos Or/geneselPeronismo,ol.i (Buenos Aires, Siglo XXI, I97I).26 The classical statementconcerningthe role of the migrantsfrom the interior is G.Germani,PoliticaySociedadnunaEpocadeTransicionBuenosAires, Paidos, 1968).See

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    14/27

    Unity and Diversity in Latin American History 13Murmis and Portantiero have argued that Per6n's support was drawn fromall sectors of the working class; repressed by conservative governmentssince 193o, Argentina's urban workers and trade union leaders saw inPeron a potential, if ambiguous, ally. There was, therefore, considerablecontinuity in working-class political behaviour, at least during the earlyphase of the regime.27 Of course, this was all to change. Union leaderserredin supposing that they could retain control over relatively independentunions in the face of a deliberate strategy by Per6n to establish completestate control over labour. It is at this point, as di Tella has suggested, thatthe mobilization of previously non-participating sections of the workingclass may have some role in explaining the nature of Peronist control overthe trade unions.28 In any case, the evidence suggests that with Peronismwe are dealing with an attempt at an alliance between the working classand the state, an alliance in which the state gradually gained the upper handand increasingly subordinated the working class to its control. Byconflating the later subordination of the unions to the state (1947-5 5) withthe period of genesis of the Per6n-union alliance (I943-7), themodal-pattern model seriously understates the autonomy of working classaction and, by describing this phenomenon as 'populism', implicitlyascribes motives and orientations to the working class which were less thanrational. There is no doubt that Peron's discourse, at least until 1952, waspopulist.29 However, to infer from Peron's massive support that this appealwas successful is to make a series of possibly unwarranted assumptionsabout working-class motives.Similar considerations apply to the Cardenasperiod. Most analysts haveargued that the Mexican unions were in a state of disarraywhen Cardenascame to power in 1934 and that, in any case, the Mexican working class hada long tradition of state tutelage. In a bid to break away from the powerbehind the throne exercised by Plutarco Elias Calles, the new Presidentactively mobilized working-class and peasant support on his behalf. Theoutcome of this was the reorganization of the dominant party alongcorporatist lines and the formation of the Confederacione Trabajadores e

    also G. Germani, Authoritarianism,Fascism and National Populism (New Brunswick,Transaction Books, 1978); G. Germani, 'El Surgimiento del Peronismo', DesarrolloEcondmico,vol. 13, no. 51 (I973). Some of the critics include Little, op cit.; Murmisand Portentiero, op. cit.; T. Halperin, 'Algunas observaciones sobre Germani, elsurgimiento de peronismo y los migrantes internos', DesarrolloEcondmico,vol. 14, no.56 (I975); E. Kenworthy, 'Interpretaciones ortodoxas y revisionistas del apoyo inicialdel peronismo', DesarrolloEcondmico,vol. 14, no. 56 (1975).27 Horowitz, op. cit. 28 di Tella, op. cit., pp. 49-5 I

    29 de Ipola, op. cit., p. 154.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    15/27

    I4 Ian RoxboroughMexico (CTM) in 1936. Such an account is not, however, entirely accurate.Although the earlier union movement had been subordinated to the statethrough Luis Morones' CROM, there had always existed independentdissident unions. By 1933, together with breakaway elements from theCROM, the CGT and the Communist Party unions (the CSUM) wereexperiencing a period of growth and revived militancy. When Cardenaswas elected President, the union movement - after considerable initialhesitation - rallied to his support in the struggle against Calles. This switchin the position of organized labour was in part due to the change inComintern tactics in 935 towards a policy of popular fronts.30 Even whenthe CTM was formed, the Communists and independent leftists retainedconsiderable support within the union movement as a whole, and it wasonly with the purges of 1948 that the union movement was definitivelybrought under state control.31 As in the case of Peronism, we have herean independent labour movement entering into alliance with the state andsubsequently being subordinated to that state. The notions of a' mobilizablemass' or of a coalition involving a heteronomous working class - crucialto the classical definition of populism - are inapplicable.The empirical evidence, then, does not lend strong support to the notionthat the governments of Cardenas, Peron and Vargas can reasonably bedescribed as 'populist'. In all cases the organizational features held to bedefining characteristics of populism (in the classical version of the definition)are absent. In terms of the minimal definition of populism (i.e. a particulartype of ideological appeal), both the Cardenas and the Peronist governmentsmay be labelled as 'populist'. The result, however, of using such a minimaldefinition, as Laclau has argued, is that a great many otherwise diversemovements, including Nazi Germany, must also be labelled 'populist'.32The utility of the notion of 'populism' as a description for a particularstage in the development of Latin American societies is, therefore, opento serious question.On the second major issue, the structural causes of the militaryinterventions of the I96os and 1970S, the evidence is equally problematicin terms of the modal pattern model. The three countries of Latin America30 S. Le6n, 'Alianza de clase y cardenismo', RevistaMexicanade CienciasPoliticasy Sociales,no. 89 (I977); S. Le6n, 'El Comite Nacional de Defensa Proletaria', Revista Mexicanade Sociologia,vol. 40, no. 2 (1978).31 F. Barbosa Cano, 'El charrazo contra el STPRM', in J. Woldenberg et al., Memoriasdel

    EncuentroSobreHistoria del MovimientoObrero(3 vols. Puebla, Universidad Aut6nomade Puebla, 1980), vol. 2; L. Medina, CivilismoyModernigacion elAutoritarismo: Historiade la RevolucidnMexicana, vol. 20 (Mexico, El Colegio de Mexico, i979).32 Laclau, op. cit.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    16/27

    Unity and Diversity in Latin American History 15which have developed sufficiently to enter the B-A phase are Argentina,Brazil and Mexico. Of these three, perhaps only Argentina stands up wellto an examination of the empirical evidence. The Mexican case is a difficultone in that there have been no clear discontinuities either of forms ofpolitical domination or of the model of capital accumulation since the earlyfifties. It is difficult to point to a clear rupture with the ISI phase or model.In Brazil, although there appears to have been a sharp break in thepattern of development in 1964, with the installation of a long-termmilitary regime and the implementation of a drastic and successfulstabilization programme, followed by a period of economic growth of theorder of 9 %perannum, the detailed evidence suggests that the discontinuityis more apparent than real. In a recent article, Jose Serrahas criticized threeof the central postulates of the B-A model in this area.33First, he arguesthat there was no sudden alteration in the rate of economic growth afterthe military coup of 1964; rather, Brazilian economic history displays along-term growth rate of some 6 % per annum since the I930S. Secondly,within this secular trend, there are naturally cyclical movements in theeconomy, and the sharp recession of 1964-7 was followed by a spectacularboom in the economy, giving rise to theories about the 'Brazilian miracle'and the relationship between rapid rates of capital accumulation andauthoritarian regimes. Thirdly, Serra also argues that there was nonoticeable increase in the 'deepening' of capital as claimed by the theoristsof bureaucratic-authoritarianism. In short, the trends in capital accumula-tion date back to the 1930s, and the arrival of the military in 1964 doesnot produce a radical shift in the developmental model.Similar arguments may be made about the behaviour of the wages ofurban workers. One of the implications of the theory of bureaucratic-authoritarianism is that the requirements of capital accumulation in thenew phase of capitalist development necessitated greater control over thelabour force. Real wages must remain stagnant (or even fall), unions mustbe controlled, and strikes reduced in frequency and scope. Once again, thelong-run evidence in the case of Brazil fails to support the model fully.The trend of real wages for the urban working class since the I930S hasbeen one of slow improvement, and the advent of the military coup merelymarks a short-term reversal of the rapid growth of real wages under thegovernment of Joao Goulart (i961-4).34 In Mexico, realwages plummeted33 J. Serra, op. cit.; A. Villela and W. Suzigan, 'Government Policy and the EconomicGrowth of Brazil, 1889-1945 ', Brazilian EconomicStudies,no. 3 (x975).34 J. Wells, 'Industrial Accumulation and Living Standards in the Long-run: The SaoPaulo Industrial Working Class, 193-75 ', Journalof DevelopmentStudies, vol. I9, nos.

    2 and 3 (January and April I983).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    17/27

    I6 Ian Roxboroughduring the Second World War, remained more or less stable until the early'fifties, and then rose steadily until the mid-I970s.35 Although thedistribution of income in both of these countries has steadily deteriorated,rapid economic growth has meant that this could occur without producinga long-run fall in workers' real wages.In summary, then, the empirical record does not suggest strong supportfor the modal pattern model as regards the relationship between stages ofeconomic growth and forms of political activity. This is so in the moreimportant countries, the countries which ought to provide the strongestsupport for the theory. It follows, a fortiori, that the case for the modalpattern model as a general description of Latin American development as awhole is weaker still. The principal reasons for the inadequacies of themodal pattern model seem to be twofold: on the one hand, even in thesupposedly typical cases of Argentina, Brazil and Mexico, the model isinaccurate in its depiction of political and economic events; complexprocesses are glossed over and reduced to pre-defined formulae which,upon careful empirical examination, fail to support the contentions of themodel. In the second place, the attempt to impose a uniform pattern onall of Latin America simply avoids the great range of variation. To presenta model which will fit all Latin American countries, the modal patternmodel necessarily has recourse to a simplified theoretical construct whichcan no longer adequately describe even the countries on the basis of whoseexperience it was first formulated.Multiplepaths: Cardosoand FalettoIn 1969 Fernando Henrique Cardoso and Enzo Faletto published a shortinterpretative account of Latin American history.36It should be noted thatit is by no means clear whether the authors intend to develop a general35 J. Bortz, 'El salario obrero en el Distrito Federal', InvestigacionEcondmica,no. 4 (1977).Since the mid-7os, real wages have tended to fall.36 F. H. Cardoso and E. Faletto, op. cit. In this article I have used the 1979 Englishtranslation. As Robert Packenham, 'Plus Sa Change... The English Edition of Cardosoand Faletto's DependenciayDesarrolloenAmerica Latina', Latin AmericanResearchReview,vol. I7, no. I (I982), points out, there are a number of differences between the 1969

    (Spanish) and the 1979 (English) editions. These consist principally in the addition ofillustrative material to the chapters discussed in this article (chapters 3 and 4) and apost scriptum. As Packenham notes, the addition of this material does not alter theauthors' argument. Moreover, to judge by the footnotes, Cardoso and Faletto do notseem to have assimilated the monographic literature published after 969; only two newsources are mentioned, one from 1972 (dealing with Peru) and one from 1970 (dealingwith Mexico). These do not alter either the authors' arguments or the criticismspresented in the present article.

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    18/27

    Unity and Diversity in Latin American History 7theoretical model of Latin American development, or whether they aremerely applying a particular kind of methodology to a series of concretehistorical cases without any attempt to develop such a general theory. Atthe very minimum, however, DependencyndDevelopmentndicates the keyvariables which, according to the authors, enable the analyst to make senseof the multiplicity of developmental paths followed by Latin Americancountries. Their starting point is a differentiation of types of exporteconomies in the nineteenth century into two basic types: (i) enclaveeconomies and (2) economies with national control of the productivesystem. The enclaves which Cardoso and Faletto have in mind are miningor plantation enclaves. Although the image conjured up by the notion ofan enclave is of a geographically separate entity, the key feature for theseauthors appears to be not the geographical situation of the productiveenterprise but rather whether ownership is foreign or domestic. Now itis certainly the case that most plantations and mining concerns in LatinAmerica have been owned by foreign intersts, and that agriculturalactivities which have been more widely diffused across the nationalterritory have generally been owned by nationals. Nevertheless, it isimportant for the logic of the explanation not to conflate the locationalcharacteristics of the enterprise with the nationality of its owners.This said, Cardoso and Faletto claim that this difference in the structureof the export sectors led to different trajectories for these two groups ofcountries. The key question facing Latin American states in the earlytwentieth century was how the middle classes were to be incorporatedpolitically. In the enclave economies the state was under the exclusivedomination of the oligarchy. As a consequence, the incorporation of themiddle classes could proceed either by a breakdown of the oligarchic order,as in the Mexican revolution, or as a result of conflicts among the dominantclasses, enabling a middle-class challenge to succeed. In this path,characteristic of Chile, the role of the state was a key to the success orotherwise of this project.In the societies where there was national control of the productivesystem, the state reflected a co-existence of the oligarchy and the bour-geoisie. Here also there were two possible routes to the incorporation ofthe middle classes. Where there was a substantial degree of unity amongthe dominant classes, the incorporation of the middle classes would occurunder the hegemony of the bourgeoisie. According to Cardoso andFaletto, Argentina illustrates this case. In countries where the dominantclasses were fragmented, such as Brazil, there was only a partial incorpora-tion of the middle classes. In both enclave and nationally owned export

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    19/27

    18, Ian Roxboroughsector situations, middle-class challenges to the system of oligarchicdomination might fail, due to the resistance of the oligarchy and/or theweakness of the middle classes. Such was the case in Peru andColombia.

    Such, in brief, is the basic schema outlined by Cardoso and Faletto.Variations in the structure of the dominant class are used to explaindiffering political outcomes, defined in terms of the varying modes ofincorporation of the middle classes. Condensing the account presented inDependencyand Development in this manner naturally runs the risk of doing

    Table 2. The Cardoso-Faletto schemeType of dependent Middle-classeconomy challenge Outcome

    Enclaveeconomies Breakdown n oligarchicdomination(Mexico,Successfulmiddle-class Bolivia,Venezuela)challenge Middle-classaccess tooligarchic-bourgeoisdomination(Chile)Unsuccessfulmiddle-class Maintenanceof oligarchicchallenge domination(Peru)

    Nationallyowned Successfulmiddle-class Middle-classncorporationexportsector economies challenge(dominant underbourgeoishegemonyclassunity) (Argentina)Successfulmiddle-class Partial ncorporationofchallenge(dominant middleclass(Brazil)classdivided)Unsuccessfulmiddle-class Maintenance f oligarchicchallenge domination(Colombia)

    serious injustice to the complexities and richness of their historicalaccount. This difficulty is doubly compounded by the fact that Cardosoand Faletto have written a text which, in many places, is elliptical to thepoint of opacity. As I shall attempt to demonstrate below, their accountis open to a number of objections.

    Throughout their account the key terms 'middle classes', 'bourgeoisie'and 'oligarchy' are left undefined and often without any clear empiricalreferent. The result is a rather shadowy portrayal of socio-economicconflict in which individuals, governments, policies or political parties areheld to 'represent' the class forces supposedly in conflict. The existenceof these class forces, and of the conflicts of interest among them, is assertedor inferred rather than demonstrated empirically. This is particularlycrucial since the key issues in the model presented by the authors revolvearound the degree of unity or disunity of the dominant classes and the

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    20/27

    Unity and Diversity in Latin American History 19relationship between dominant classes and the state. If there is noindependent measure of these variables, then we are likely to be presentedwith an analysis which is both reductionist and tautological: reductionistbecause political forces are taken to indicate the presence of social classes;tautological because the evidence for the existence of classes is the existenceof political actors.The signal advance of the Cardoso-Faletto model over both theessentialist theories and the modal pattern model is that, on the basis oftwo dichotomous variables (national v. foreign ownership of exportsectors, and degree of dominant class unity), a number of different oliticalpaths can be traced.37 The essentialist and modal pattern approaches bothpostulate a single developmental path which all Latin American countriesfollow more or less closely. These approaches treat 'deviations' as justthat; and account for them in an ad hoc manner. In contrast, theCardoso-Faletto model is based on the interaction of previously specifiedtheoretical variables which in turn specify alternative paths. In addition,the Cardoso-Faletto model has the advantage of being a social actionapproach: as an integral part of the theory, social actors are faced with realchoices, and hence outcomes are not deterministic. This feature of themodel also moves the analyst toward a consideration of multiple paths.The Cardoso-Faletto model, then, offers a theoretically satisfactoryaccount of the diverse political trajectories of the nations of LatinAmerica. However, as we shall see, while the Cardoso-Faletto method-ology might be acceptable, their empirical analysis suffers from a numberof serious problems.

    Clearly Cardoso and Faletto had a certain body of empirical material inmind when in the mid-sixties they wrote their short book. There is, indeed,evidence concerning the structure of dominant classes in Latin America,and this can be brought to bear on the propositions advanced by Cardosoand Faletto. Interestingly, such evidence as is presently available suggeststhat many of the claims and assumptions concerning the nature ofdominant classes put forward in DependencyndDevelopment re mistaken.(The evidence on this topic is, it must be noted, both sparse and repletewith methodological problems. Until further studies are published,however, we have no choice but to work on the basis of the existingmaterial.)In Argentina, Cardoso and Faletto refer to 'class unity'. While Cardoso37 At variouspoints the authorsallude to otherdifferentiatingactors,such as the degreeof diversity of the export system, though these variablesappearto be of secondaryimportance. bid.,pp. 80-8 .

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    21/27

    20 Ian Roxboroughand Faletto note the existence of a variety of fractions of the dominantclasses in Argentina, they assert that 'the agro-livestock-exporting sector'constituted a hegemonic dominant class.38The work of Peter Smith andof Miguel Murmis and Juan Carlos Portantiero, however, suggests thatthe dominant classes in Argentina were sharply divided.39 In the agrariansector, there was a division between the breeders of the interior provincesand the fatteners of the pampa near Buenos Aires. Similarly, Argentineindustrialists were divided into those closely linked to foreign enterprisesand those concerned primarily with the domestic market.40In their remarks on Brazil, Cardoso and Faletto make the claim that thedivision within the ranks of the dominant classes was one between thecoffee oligarchy and the industrialists. They also assert that there wasconsiderable state autonomy with regard to the industrialization pro-gramme. With regard to the first claim about the disunity of the Braziliandominant class, it does, indeed, appear to be the case that there wasconsiderable fragmentation and conflict among dominant classes. Theconflict, however, was not between the coffee oligarchy and industrialists.As the work of Warren Dean indicates, Brazilian industrialists came fromthe ranks of the coffee planters, particularly in the state of Sao Paulo, andretained close ties with them.41Moreover, it is difficult to point to sectoralclashes between coffee exporters and industrialists during this period.Rather than a sectoral division, the Brazilian dominant classes weredivided along regional lines.42 This had its origins in the changing patternof Brazil's exports, the historical development of which meant thatregions - and with them, regional oligarchies - rose and fell as their38 Ibid., p. 83.39 M. Murmis and J. C. Portantiero, op. cit.; P. Smith, Politics andBeef in Argentina (New

    York, Columbia University Press, I969).40 J. Teichman, 'Interest Conflict and Entrepreneurial Support for Peron', Latin AmericanResearchReview,vol. 16, no. i (I98i).41 W. Dean, TheIndustrializationof Sdo Paulo (Austin, University of Texas Press, I969).42 On the regional oligarchies in Brazil see, interalia, R. Faoro, Os Donos do Poder(2 vols,Porto Alegre, Editora Globo, 1979, st ed. 1957); J. Love etal., 'O Poder dos Estados'in B. Fausto (ed.), Historia Geral da Civili.afao Brasileira,vol. 8 (I977); J. L. Love, RioGrandedo Sul andBrazilian Regionalism, 882-I93o (Stanford, Stanford University Press,I971); J. L. Love, Sio Paulo in the Brazilian Federation,i889-1937 (Stanford, StanfordUniversity Press, i980); R. M. Levine, Pernambucon the Brazilian Federation,I889-1937(Stanford, Stanford University Press, 1978); J. D. Wirth, Minas Gerais in the BrazilianFederation, 889-1937 (Stanford, Stanford University Press, 1977); E.-S. Pang, Bahia inthe First Brazilian Republic,i889-i934 (Gainesville, University of Florida Press, 1979);V. Nunes Leal, Coronelismo (Cambridge, Cambridge University Press, 1977);F. Uricoechea, Patrimonial Foundationsof the Brazilian BureaucraticState (Berkeley,University of California Press, I980).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    22/27

    Unity and Diversity in Latin American History 2exports prospered and declined in the world market. The conflicts withinthe dominant classes in Brazil in the early twentieth century were betweenthe oligarchies of Sao Paulo and Minas Gerais (the famous cafe con laitealliance) and the weaker oligarchies of Rio de Janeiro, Rio Grande do Sul,and the states of the north-east. It was conflicts among these states whichexplain Vargas' assumption of power in 1930 and the subsequent PaulistaCivil War of 1932, alliances and conflicts which cannot easily be describedin sectoral terms.43

    As to the question of the relative autonomy of the state under GetiilioVargas (1930-45), this is a matter of some controversy, and it is somewhatpremature to pronounce a judgement on the ongoing debate on this topic.However, it cannot be disputed that the claim that this was a period ofrelative autonomy for the state has been subjected to sustained attack.44In the case of Chile, Cardoso and Faletto, following the widely knownanalysis of Claudio Veliz, assert that conflicts among dominant groupswere responsible for the particular pattern of development followed in thatcountry.45 However, the bulk of recent empirical evidence suggests thatthe dominant class in Chile was remarkably homogeneous. As both Zeitlinand his associates46 and Kirsch47 demonstrate, there were considerable tiesof ownership between bankers, merchants, industrialists and landowners.As Zeitlin and Ratcliff note:The contradictions between agrarian and industrial capital, and the clashes overstate policies affecting them...did not arise between ontologically 'real' classsegments of large landowners and capitalists. For contradictorynterestsand socialcleavageswithin the dominant class did notcoincide; rather, the dominant agrarian43 The 1932 civil war has sometimes been interpreted in a way which suggested that therewas a clash of economic or class interests between Sao Paulo and the Federal

    Government, e.g. P. Flynn, BraZil: a Political Analysis (London, Ernest Benn, 1978),but such a position is hard to reconcile with the data on the close links between coffeeplanters and industrialists.44 E. Diniz, Empresariado,Estado e CapitalismonoBrasil (Rio de Janeiro, Paz e Terra, 1978);E. Diniz and R. Boschi, EmpresariadoNacional e Estado no Brasil (Rio de Janeiro,Editora Forense Universitaria, 1978).45 C. V6liz, 'La Mesa de Tres Patas', Desarrollo Economico,vol. 3, nos. I-2 (April-September I963); J. Carriere, Landownersand Politics in Chile (Amsterdam, CEDLA,I981).46 M. Zeitlin and R. Ratcliff, 'Research Methods for the Analysis of the Internal Structureof Dominant Classes', Latin American ResearchReview,vol. io, no. 3 (1975); M. Zeitlinet al., 'Class Segments - Agrarian Property and Political Leadership in the CapitalistClass of Chile', American SociologicalReview,vol. 41, no. 6 (1976).

    47 H. Kirsch, Industrial Developmentin a Traditional Society (Gainesville, University ofFlorida Press, 1977); see also T. C. Wright, Landownersand Reform in Chile, ii99-4o(Urbana, University of Illinois Press, I982); and A. J. Bauer, Chilean Rural Society(Cambridge, Cambridge University Press, 1975).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    23/27

    22 Ian Roxboroughand industrialelements were internallyrelated,if not 'fused', in so complex apattern that neitherof thempossessedspecificautonomyor distinctivesocial identity.48[Emphasis n original.]

    Thus, while there is a great deal to commend the Cardoso-Faletto modelon methodological grounds, in the key countries of Chile, Brazil andArgentina, and on the key question of the unity/fragmentation of thedominant classes, it must be found lacking both in its empirical assertionsand in the lack of conceptual clarity in the definition of key terms. Thecentral division made by the authors between enclave economies andnationally owned productive systems obscures what is, in fact, the keyvariable: foreign v. national control of the export sector; and the valuesassigned to the second key variable (unity/disunity of the dominantclasses) are, in a number of instances, empirically incorrect.Towards an adequate modelThe argument of this article has been that an adequate conceptualizationof Latin American history must take into account the diversity of historicalexperiences of the various republics. Of the various models reviewed here,only the Cardoso-Faletto model adequately meets this criterion. Thatparticular model, however, is seriously flawed in a number of ways, bothconceptually and empirically. Its methodology, however, can serve as thefoundation for the construction of a better model. The task of elaboratingsuch a model is beyond the scope of this article. Nevertheless, somesuggestions as to a possible way forward will be made.The key variables of the Cardoso-Faletto model had to do with thestructure of the dominant classes. This derived from the nature of theproductive system, in particular, from the organization of the exportsectors and their linkages with the world market. This seems an indispen-sable starting point. It enables the analyst to incorporate into a singletheoretical framework both a concern for situations of external dependencyand a focus on the productive structure of the particular Latin Americancountry. This means that it is not necessary to treat the Latin Americaneconomies as simple adjuncts or reflexes of the metropolitan economies,as was done with some of the earlier and less sophisticated writing ondependency.49However, this model needs to be complemented with variables relating48 M. Zeitlin and R. Ratcliff, 'Research Methods for the Analysis of the Internal Structureof Dominant Classes', Latin American ResearchReview,vol. Io, no. 3 (1975), p. 54.49 F. H. Cardoso, 'The Consumption of Dependency Theory in the United States', LatinAmerican ResearchReview,vol. 12, no. 3 (1977).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    24/27

    Unity and Diversity in Latin American History 23to the structure of subordinate classes. As Colin Henfrey has noted, withits exclusive focus on dominant classes, the Cardoso-Faletto model is indanger of becoming a one-class model of history.50 On the contrary, socialchange is better analysed in terms of conflicts, tensions, and accommodationsbetween and within both dominant and subordinate classes (and strata,quasi-classes, etc.). Even in those situations where subordinate classesappear to have little or no political input and where politics appears to beentirely an intra-elite affair, a two-class model is still called for.51 Thepassivity or quiescence of subordinate classes and strata is somethingwhich requires explanation. The model, then, calls for two key variables(or sets of variables) defining the structure of both dominant andsubordinate classes.52

    The model proposed here is a simple one; it assumes that thedevelopment of the economy will bring with it the emergence anddevelopment of new classes or subclasses, and that these will press forinclusion in the political arena. Existing power holders will respond tothese challenges in one way or another, and this will constitute the coreof political conflict.53 There is nothing novel about such a model; mostof its features appear in the other models examined in this article. At thelevel of empirical specification, however, this model differs sharply fromthe others.50 C. Henfrey, 'Dependency, Modes of Production and the Class Analysis of LatinAmerica', Latin AmericanPerspectives, ol. 8, nos 3 and 4 (1981), p. 29; cf. also A. Cueva,'A Summary f "ProblemsandPerspectivesof DependencyTheory"', Latin AmericanPerspectives,vol. 3, no. 4 (1976), p. I4.51When I refer to a 'two-class model' I mean to imply that the basic dynamicfactorsin the model derive from the interaction of the two classes. In this sense they areindispensable tartingpointsorbuildingblocks.Naturally,n anyconcreteinvestigationit will be necessaryalso to take into account other social classesand strata,as these

    will have some impacton the outcomes of political conflicts. It is possible to go onaddingextra actors o ananalysis,making t morecomplexandtherebyexplainingmoreof a concretesituation. The addition of these extrafactorsis, however, a theoreticallyquite differentmatterfrom the questionof whetherthe model begins simplywith thedominantclass(es)(a one-classmodel), or with both dominantandsubordinateclasses(a two-classmodel).52 It should not be thought that this is a simple matter. A descriptionof the structureof a classis anexceedinglydifficultandcomplextask. There aremanyfactors nvolvedin the structuration f socialclasses.One would wish to know a greatdealaboutlabourmarkets,social mobility,residentialpatterns,ethnicity,religion, etc.53 L. Binder et al., Crisesand Sequencesn Political DevelopmentPrinceton, PrincetonUniversityPress, 1971); J. S. Valenzuela, Labor MovementFormationand Politics:the Chileanand FrenchCases n ComparativePerspective',Ph.D., Columbia,1979;C.Weisman,Modernizationnd heWorking lass Austin,Universityof TexasPress, 1982);C. Anderson,PoliticsandEconomic hangen LatinAmerica New York, Van Nostrand,I967).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    25/27

    24 Ian RoxboroughFirst, with respect to the emergence of new classes and the nature oftheir challenge, the model presented here differs in a number of important

    ways. For a start, to talk of a 'middle class' challenging the oligarchy maybe misleading. 'Middle class' is a portmanteau term; its object needs care-ful dissection into its component parts and, indeed, serious thinking andresearch on this topic is urgently needed.54 Until very recently also, mostanalyses of Latin American history have tended to accord a very restrictedrole to the urban working class, arguing either that it has been boughtoff in a populist coalition with the middle classes, or that it has effectivelybeen disarticulated as a potential political actor by the corporatist inter-mediation of the state, or by divisions within the working class itselfderiving from the opposing interests of a highly paid labour aristocracyand a mass of unorganized workers.Such an interpretation of the Latin American experience does, however,have a number of problems. As Tom Skidmore has noted, it involvessomething of a paradox: on the one hand, one of the main factors in thegenesis of military intervention in Latin America has been the militancyof organized labour; on the other hand, the accepted wisdom in the fieldis that labour movements in Latin America are generally weak and/orconservative. Skidmore asks, 'How have Latin American workers, onceconsidered eminently manipulable, been able to frighten the guardians ofnational security?'55There is, in fact, a good deal of evidence that the urbanworking classes of Latin America have by no means been so politicallypassive or so incapable of pursuing their own class interests as someauthors would have us believe, though naturally this varies from countryto country.56In the preceding discussion of the modal pattern model, the 'classical'definition of populism was examined in some detail. It was argued thatthis notion failed to describe accurately the nature of political processesin the Cardenas,Vargas and Peron governments. A more accurate account,it was argued, could be derived from a study of the working class, thestate, anddominant classes. The outcomes could be explained by postulatingthe relatively rational pursuit of class interests by the various actors.54 I find the position taken by E. O. Wright, ClassStructureandIncomeDetermination NewYork, Academic Press, 1979), particularly suggestive on this topic.55 T. Skidmore, 'Workers and Soldiers: Urban Labor Movements and Elite Responsesin Twentieth Century Latin America', in V. Bernhard (ed.), Elites, Masses, and

    Modernizationn Latin America (Austin, University of Texas Press, 1979), p. 86.56 This is a point made by T. Skidmore, ibid., and by I. Roxborough, 'The Analysis ofLabour Movements in Latin America', Bulletinof Latin American Research,vol. i, no.i (October, I98 ).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    26/27

    Unity and Diversity in Latin American History 5Differences in outcomes were the result of differences in the nature of thesesocial classes, in terms of their internal unity, etc., and in the varyingrelations between these social actors and the state.

    Attention having been drawn to the importance of the working classas part of the explanation, this needs to be qualified immediately. Theprocess of incorporation of the working class is only of relevance whenthe working class (a) exists as a class actor and (b) appears to pose a seriousthreat ':oelite politics. This does not always occur, and in some countriesor at certain times it is the peasantry rather than the working class thatshould be the focus of attention. However, the working class emerged as asignificant political force at an early stage in the history of Mexico, Brazil,Peru, Argentina and Chile.57Cardoso and Faletto are correct in focusing on crises of incorporationas the key events in a theoretically informed account of historicaldevelopment in Latin America. However, rather than look primarily atthe incorporation of the middle classes, it seems more useful to focus onthe incorporation, first, of the industrial bourgeoisie and second, of theworking class.Since we are dealing with at least two crises of political incorporation,the typology of political outcomes must also be rather more complex thanthat suggested by Cardoso and Faletto. Neither process of incorporationis ever necessarily 'solved' in a definitive manner. Incorporation may bepartial and conditional. It may also be reversed at a later stage by exclusionof an actor from the political process.There is, moreover, a major difference between the two processes ofincorporation. Incorporation of the industrial bourgeoisie into the systemof oligarchic domination need not involve a high level of conflict. Indeed,through family ties and interlocking patterns of ownership, there may besuch a degree of integration between these two classes that they form asingle, fused class of owners. This variable, the degree of unity/disunityof the dominant classes(es) is, as we have seen, one of the key differentiating57 On the origins of Latin American labour movements, see interalia, P. Blanchard, TheOriginsof thePeruvianLaborMovement(Pittsburgh, University of Pittsburgh Press, 1982);B. Fausto, TrabalhoUrbano ConflitoSocial(Sao Paulo, DIFEL, 1977); P. DeShazo, UrbanWorkers and Labor Unions in Chile I902-1927 (Madison, University of Wisconsin Press,

    I983); P. S. Pinheiro and M. M. Hall (eds.), A Classe Operariano Brasil, vol. i (SaoPaulo, Alfa-Omega, 1979); R. Anderson, Outcasts n theirOwn Land: Mexican IndustrialWorkers 90o6-I9II (de Kalb, Northern Illinois University Press, 1976); J. Oddone, ElGremialismoProletarioArgentino (Buenos Aires, Galema, 1949); S. Marotta, ElMovimientoSindical Argentino (Buenos Aires, Libera, I960); H. Spalding, La Clase TrabajadoraArgentina (Buenos Aires, Libera, I960).

  • 7/27/2019 Ian Roxborough Unity and Diversity[1]

    27/27

    26 Ian Roxboroughfactors in the account presented by Cardoso and Faletto, and has beenconsidered at length by a number of authors. Unlike the issue of the natureof dominant classes, the process of incorporation of the working classhas, until recently, been relatively neglected. A rather unsatisfactoryreference to populism has tended to be substituted for a consideredanalysis of the varying modalities of working class incorporation into thesystem. It would, therefore, seem appropriate to have as an integral partof the analysis a study of dominant class reactions to the threat posed bythe growing urban working class. Their reactions to this threat varied, andthis in turn is a function of four factors: (I) the existence (or not) ofdivisions within the dominant classes and the nature of those divisions;(2) the nature of the (real or perceived) threat from the working class,which would depend in part on the historical development of that class,its degree of internal unity, etc.; (3) the relation of various classes to thestate, and the power and coherence of the state. (The extent to which thestate is autonomous of, or directly responsive to, social classes - bothdominant and subordinate - varies from one situation to another. It is amatter for empirical investigation, not apriori pronouncements and finally,(4) dominant class reactions will depend on the actions of other subordinateand intermediate classes, such as the peasantry and the middle classes.Underlying these constellations of class forces are, of course, a numberof economic forces which, over time, have constrained the formation ofsocial classes. The account suggested here need not be an economicdeterminist one. While due weight ought to be given to the economicforces shaping the formation of social classes, there are many other factors,both political and cultural, at play.The purpose of this article has been mainly critical; an alternative historyof Latin America has not been presented. Only the barest bones of anoutline of such an history have been indicated. This should not be takenas an indication that such a task is not possible; on the contrary,scholarship in this area is advancing rapidly. Out of the flood of detailedmonographic studies there are beginning to emerge the materials for acoherent account of the histories of the countries of Latin America. Theprocess is a slow and hesitant one, and will involve constant returns toprevious attempts at synthesis, partly to build on the very realachievementsof previous scholarship, but partly also to free ourselves from themethodological shortcomings that have resulted from attempting to fittwenty separate histories into a single Procrustean bed.