¹ See for example the six volumes edited by the Social Science Research Council's Committee on Comparative Politics (published by Princeton Press, 1963–66) and Huntington, Samuel, Political Order in Changing Societies (New Haven, 1968)Google Scholar.

² One exception to this trend is Frey's, Frederick W. The Turkish Political Elite (Cambridge, 1965)Google Scholar.

³ The concept of role, as used here, refers to all the principal aspects of a role: the expectations of others, the orientations of the individual occupying the position, and the actual behavior of the individual. The congruence and consensus among these role aspects is a major focus of my investigation. This essay assumes the reader is generally familiar with the concept of “role” as used in this sense. See for reference, Biddle, Bruce J. and Thomas, Edwin J., editors, Role Theory: Concepts and Research (New York, 1966)Google Scholar. One example of a study of legislative behavior which relied heavily upon role analysis as a tool of research to reveal patterns of politics in four state legislatures is Wahlke, John C., Eulau, Heinz, Buchanan, William and Ferguson, Leroy C., The Legislative System (New York, 1962)Google Scholar.

⁴ See Easton, David, A Framework for Political Analysis (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1965), pp. 39–57 Google Scholar.

⁵ See Gross, Neal, Mason, Ward S. and McEachern, Alexander W., Explorations in Role Analysis (New York: John Wiley and Sons, Inc., 1958), Chapters 10–12Google Scholar.

⁶ This pattern of “closed” politics and the process by which it has developed are described in Hopkins, Raymond F., Political Roles in a New State: A Study of Politics in Tanzania (dissertation, Yale, 1968 Google Scholar; Yale Press, forthcoming). References are to pages in the dissertation.

⁷ For a readable review of the national political history of Tanzania, see Taylor, J. Clagett, The Political Development of Tanganyika (Stanford, 1963)Google Scholar; and for an insightful analysis of the party and state machinery see Bienen, Henry, Tanzania: Party Transformation and Economic Development (Princeton, 1967)Google Scholar.

⁸ One member, Chief Masanja, resigned from TANU to join the ANC in 1962. But this exception has proved unimportant.

⁹ All speeches in the 1961/62 and 1965/66 periods have been coded; approximately 50% of the speeches in the intervening three years have been coded. Twenty-eight different subject categories, many related to ministerial divisions, such as agriculture or education, were used in coding subjects. Orientation was either national or local. The position of a speech could be one of four types: supportive (of the government's position), critical, neutral (if the speech was balanced or there was no official government position), and “demand” question-answer. This last category was used for many of the short questions and answers expressing and responding to demands or problems, usually local, raised by backbenchers. The language was Swahili or English (or in a few instances, mixed). Three length categories were used: short (0 to 1½ minutes), medium (1½ to 7 minutes) and long (more than 7 minutes). The coding procedures and tests for reliability are reported in Hopkins, op. cit., pp. 399–410. Intercoder reliability indicated an average of over 92% agreement.

¹⁰ Mtaki, however, who had attracted attention by his flamboyant style and recitation of Marxist principles, retired from politics in 1967 to enter private business. Another frequent critic, Buhatwa, was also made a Junior Minister but he lost in the 1965 elections and became an administrator in the Ministry of Housing.

¹¹ Anangisye and Kambona lost their seats due to missing three sessions of the Assembly. Anangisye was in detention and Kambona in political exile. Kambona was not an Assembly critic, but in private he was often highly critical.

¹² See The Nationalist, July 11, 1968, and December 10, 1968, p. 1 Google Scholar.

¹³ An analysis of topics discussed in the 1965/66 Assembly revealed that criticism and support were not evenly distributed. For instance, on many of the twenty-eight subjects into which speeches were categorized, not a single critical speech was given. The most heavily criticized legislative topics (in over ten per cent of speeches) were elections, housing and labor unions. In contrast, military affairs received no criticism and was supported in 76% of the speeches on the subject. This is in marked contrast to the first year of independence when only 40% of the speeches supported military policy and 28% were critical. The lack of criticism in these areas reflects pressures, not organized by any individual or even by the Party, but rather, generated by latent political norms which dictate that delicate and controversial subjects are either best avoided or commented upon only in a constructive manner.

A factor analysis of these debates in 1965/1966 revealed the most prominent dimension as a “Security-Sovereignty” factor including issues such as foreign policy, Party, constitutional matters, the military and questions of national unity. This factor explained 27.8% of the variance among the 28 subject variables. High loading subjects accounted for 20% of all speeches. The factor scores for legislators on this dimension correlated highly, .68 with a high frequency of supportive speeches. See Hopkins, op. cit., pp. 229–230.

¹⁴ Informal interviews with a number of other National Assembly Members, including Second Vice-President Rashidi Kawawa and Assembly Speaker Adam Sapi Mkwawa, were also used in this study. A statement of the interview format used in the interviews may be found in Hopkins, op. cit., pp. 358–375. Those interviewed are not identified since anonymity was promised.

¹⁵ The term “intellectual” is applied in the general sense suggested by Shils, Edward in his article, “The Intellectuals in the Political Development of New States,” World Politics, XII (April, 1960), 329–368 CrossRefGoogle Scholar. In order to operationalize these distinctions, which emerged initially from subjective analysis of the interviews, legislators with college degrees or post-secondary education and who read widely were selected as “Intellectuals.” One M.P. with a college degree has been excluded from this typology because he did not fit the basic characteristics of either “Intellectuals” or “Locals.” He did not represent a constituency, had not given a single speech in the National Assembly and had no interest in a political career. His work, in fact, often kept him out of the country.

¹⁶ The twenty-sixth, the Deputy Speaker, is the only nongovernmental member in the sample serving his second term. He was excluded from the typology.

¹⁷ The seventh legislator remaining, a National Member, did not fit this or any other category.

¹⁸ In most cases the N was too small to meet the standard requirements of an expected frequency of five in each cell for the X² test. Rather than collapse cell categories to avoid this, it seemed better to keep distinctions which I believed had analytical importance and to present only the appropriate X²'s. However, many differences are large enough to be statistically significant if the cell size requirement were relaxed. See Blalock, Hubert M., Social Statistics (New York, 1960), p. 221 Google Scholar.

¹⁹ Members were divided into three roughly equal groups. To do this, it was necessary to collapse critical percentages so that all those for whom 10% or more of their speeches were critical became the “most critical” group.

²⁰ Ralph K. Huitt has analyzed the costs of deviancy in the U.S. Senate. See his “The Outsider in the Senate: An Alternative Role,” this Review, 55 (September, 1961), 566–575 Google Scholar.

²¹ Nyerere, J. K., “Address at the Opening of the National Assembly after the General Election,” (Dar es Salaam: Ministry of Information and Tourism, October 12, 1965), no paginationGoogle Scholar.

²² As cited in Tordoff, , “Parliament in Tanzania,” Journal of Commonwealth Political Studies (July, 1965)CrossRefGoogle Scholar.

²³ Nyerere, “Address at the Opening of the National Assembly.”

²⁴ The report blamed practices inherited from the British more appropriate to a two party system for the lack of vigor in Assembly debates. In the British party system, private party caucuses and party whips to enforce discipline were accepted procedures essential to maintain the unity upon which the party in power depended for its control of the legislative process. Thus the TPP and the National Executive Committee provided private forums for prior discussions of policy and legislative matters. These prior discussions, the report concluded, inevitably have “inhibited subsequent discussion on the floor of the house.” Report of the Presidential Commission on the Establishment of a Democratic One Party State (Dar es Salaam: Government Printer, 1965), pp. 20, 21 Google Scholar.

²⁵ Nyerere, Julius K., “Democracy and the Party System” (Dar es Salaam: the Standard Press, 1963), p. 6 Google Scholar.

²⁶ Nyerere, , “Address at the Opening of the National Assembly,” p. 6 Google Scholar.

²⁷ Julius K. Nyerere, “Democracy and the Party System,” op. cit., pp. 6–7.

²⁸ See Hopkins, op. cit., pp. 376–384 for information about the questionnaire and sample.

²⁹ For legislators who were members of the government, and hence constrained in their M.P. role, these questions were phrased in the third person and their expectations about the behavior of a backbencher were solicited. Written versions, shown to administrators, were similarly in the third person.

³⁰ See Hopkins, op. cit., for the use of variance as a measure of consensus, pp. 393–94. This measure has been recommended by Neal Gross, et al., p. 107.

³¹ See Hopkins, Raymond F. and Sherman, Neal, “Students and Politics in Tanzania,” in Lipset, Seymour M. and Hanna, William J. (eds.), Students and Politics in Africa (forthcoming Basic Books)Google Scholar.

³² This is accomplished by the formula

where

NI= the number of items

S_j = the standard deviation of the jth item, and

X_ij = 0, 1, or 2 according to a weak, moderate, or strong role expectation by the ith. respondent.

³³ Samuel P. Huntington, op. cit., Chapter One, discusses the importance of coherence for political institutionalization. Congruence between sets of expectations should be an important indicator of institutionalization. See Hopkins, Raymond F., “Political Roles and Political Institutionalization,” an unpublished paper read at the American Political Science Association Meetings, September 2, 1969 Google Scholar. Moreover, the variance indicative of consensus among role expectations for M.P.'s was markedly lower, .394, than among elite role expectations for the President (.519) or for administrators (.521).

³⁴ One M.P., an Intellectual, now expelled from the Party and the Assembly, claimed that Nyerere's expectations created opportunities for dissent. He stated: “You know, I like to make a lot of noise in the Assembly and I often disagree. President Nyerere is the only one who understands the need for this. If it weren't for him, they'd never let me get away with it.”

³⁵ See Hopkins, op. cit., for rules regarding the role of President and administrator.

³⁶ Standard, June 24, 1966.

³⁷ Nationalist, February 6, 1967.

³⁸ Nationalist, February 15, 1967.

³⁹ Nationalist, December 28, 1966. Kamaliza has since been dropped from his post.

⁴⁰ Nationalist, June 12, 1966.

⁴¹ This study is reported in Cliffe, Lionel (ed.), One Party Democracy (Nairobi, 1967)Google Scholar in chapter 12 by Kenneth Prewitt and Goran Hyden.

⁴² A revealing incident occurred on one occasion when I raised this question subsequent to an interview (since the question was added part way through the interviewing process) in the presence of another Tanzanian. When the M.P. (a Local) declared that the Bunge was more powerful than the N.E.C., the other Tanzanian burst forth with, “No, no! That's the wrong answer.” This other Tanzanian later took me aside and explained, “Some of these M.P.'s just don't understand the way things are.” The deviance of Locals underscores the often unexpressed desire among Locals to increase the power and importance of the Assembly and, as a consequence, their own power and importance.

⁴³ Debate and criticism in the Assembly meetings which occurred after the period covered by the content analysis, for at least the period of September-October 1966 until July, 1968, seem to have continued these trends. Policy initiative and criticism, judging by newspaper accounts only, appear to have increased somewhat in these sessions.

⁴⁴ Nationalist, June 29, 1966. The Afro-Shirazi Party is the only party on Zanzibar; it is not yet integrated with TANU.

⁴⁵ See, for example, statements of the Mwanza Area Commissioner, Nationalist, April 28, 1966 Google ScholarPubMed, and frequent columns of “Pressman's Commentary” in the Nationalist.

₄₆ This term was actually used by one Intellectual describing his efforts to pressure the government on insurance regulation.

⁴⁷ Professor Henry L. Bretton in a private communication, February 22, 1969, suggested that in Ghana “Parliamentary proceedings were a form of shadow-boxing” and the institution was “impotent.”

⁴⁸ Frederick W. Frey, op. cit., p. 395.