1 Even a careful statement like Cardoso’s and Faletto’s saw only two alternatives: military nationalism, unlikely to achieve more than a ‘momentary success’; and working‐class pressures, also ‘modest in the context of authoritarian‐corporatist domination’. ( Cardoso, F. H. and Falleto, E., Dependěncia e Desenvolvimento na América Latina, Rio de Janeiro, Zahar Editores, 1970, pp. 135–38Google Scholar.)

2 Schwartzman is careful to avoid naive culturalism. Between the first (1975) and the second (1982) editions of this major work ( Schwartzman, S., Bases do Autoritarismo Brasileiro, Rio de Janeiro, Editora Campus, 1982 Google Scholar) he greatly attenuated his scepticism toward the electoral mechanism: ‘bad with it, worse without it’ (p. 147).

3 Schmitter ( Schmitter, P., ‘The “Portugalization” of Brazil?’, in Stepan, A., Authoritarian Brazil, Yale University Press, 1973 Google Scholar) was obviously wide of the mark when he compared Brazil’s electoral institutions to those which existed in the Portuguese ancien régime: ‘and yet, elections persist. They are held irregularly at the will and whim of the executive, but the electorate has been permitted a voice. There is nothing peculiar about this in authoritarian regimes. Portugal has elections as regular as clockwork’ (p. 212). The passage of time, or a different perspective, or both, enabled Roett ( Roett, R., ‘The Political Future of Brazil’, in Overholt, W. H., The Future of Brazil, Boulder, Colorado, Westview Press, 1978 Google Scholar) to provide a more balanced account: ‘The regime’s continuing commitment to the niceties of constitutional government confuse many observers of the Brazilian political system. In other Latin American nations, the congresses are closed, parties abolished, and national politics prohibited. In Brazil, although the latitude given to the civilian political process is severely compromised, it does exist. The commitment to political participation ‐ limited, elitist, and manipulable as it is ‐ is strongly rooted in Brazilian constitutional history. Geisel’s efforts at decompression were part of that historical belief that there should be a more open system. The difficulty is drawing the line between acceptable political behavior by civilian political elites and behavior that threatens to weaken the legitimacy of the regime or to open it to political subversion. When the issue is between an open system (understood in Brazilian terms as an elite and not a mass system) and national security, the latter always wins’ (p. 82).

4 See J. Linz, ‘The Future of an Authoritarian Situation or the Institutionalization of an Authoritarian Regime: the Case of Brazil’, in Stepan op. cit.

5 ‘Interestingly’, as Schmitter (op. cit., p. 210) puts it, ‘they continued the practice of permitting local level conflict through the device of sublegendas (competitive lists) within the same party in state and municipal elections’.

6 Up to 1964, the main issue concerning the electoral process was the degree to which electoral behaviour had become free from patron‐client constraints. Today, it is generally recognized that it does meet the necessary requirement of uncertainty, in this sense, and the issue has become essentially the fairness or unfairness of the prevailing rules.

7 This hypothesis is suggested by Souza and Lamounier in their study of government/labour relationships. ( Souza, A. and Lamounier, B., ‘Governo e Sindicatos no Brasil: A Perspectiva dos Anos 80’, Dados, vol. 24, p. 139–59Google Scholar.)

8 On the elections since 1974, see Lamounier, B. and Cardoso, F. H., Os Partidos e a Eleições no Brasil, Rio de Janeiro, Editora Paz e Terra, 1975 Google Scholar; Reis, F. W., Os Partidos e o Regime, São Paulo, Editora Símbolo, 1978 Google Scholar and Lamounier, B., Voto de Desconfiança: Eleições e Mudança Política no Brasil, Rio de Janeiro, Editora Vozes, 1980 Google Scholar.

9 In addition to the works cited in footnote 8, see Soares, G., Sociedade e Política no Brasil, So Paulo, Difel, 1973 Google Scholar and Schwartzman, op. cit.