* An earlier version of this paper was presented at the annual meeting of the Canadian Political Science Association meetings held in Calgary, June 5–7, 1968. I am indebted to Ole R. Holsti, Robert C. North and John O. Field for comments on perceptual analysis, to Hugh Thorburn for suggestions on reliable sources of behavioural data, to Stanford University Computation Center for assistance in processing attitudinal data, to Queen's University for a grant which made possible the collection and analysis of action data, and to Peter Jewett for research assistance.

¹ This paper focuses on the early period in the development of nonalignment and not on current expressions of the policy; consequently, any generalizations from the findings of this to current conditions should be undertaken with considerable caution.

² We are aware that the term “policy” is used here in a more limited sense than is general in the literature.

³ The relationship of the nonaligned nations to the cold war ranges from low involvement, exemplified by Chad, Gabon, and Niger, to moderate involvement, as seen in the positions of Morocco, Sudan, and Ceylon, to high involvement, indicated by most of the other Afro-Asian states.

⁴ Throughout this study the term “West” refers to the major power members of the Western alliance system. The term “East” refers to major powers in the Communist alliance system and does not include non-Communist states located geographically in the eastern hemisphere.

⁵ This is basically a working or tentative hypothesis. The results should shed light on the general issue regardless of the particular outcome.

⁶ The final communiqué of the Bandung Conference includes the famous ‘Ten Principles of Peaceful Coexistence.” Briefly these comprise the following directives: respect for human rights and the Charter of the United Nations, for sovereignty and territorial integrity, for the right of each nation to defend itself, for justice and international obligation; recognition of the equality of all races and nations; abstention from the use of arrangements organized by the major powers; refraining from acts of aggression; settlement of international disputes by peaceful means; and promotion of mutual interests and obligations.

⁷ United Nations, General Assembly Official Records 1960, 444–6.

⁸ “The Conference of Heads of State or Government of Non-Aligned Countries,” 1961, 256.

⁹ Ibid., 27.

¹⁰ Ibid., 26.

¹¹ The Sino-Indian conflict in 1962 illustrates the relevance (or irrelevance—depending on one's viewpoint) of the nonalignment policy. India could enlist the assistance of the Western powers without formally joining their alliance system.

¹² Content analysis refers to systematic and objective methods of determining the characteristics of the written word. In the broadest terms, content analysis can be undertaken qualitatively or quantitatively. Our analysis is based primarily on automated content analysis, that is, the analysis of the verbal record through the use of computers from rapid data processing. For a detailed discussion of this method see Holsti, Ole R.et al., “Theory and Measurement of Interstate Behavior,” unpublished ms., 1964.Google Scholar

¹³ Although this is not a random sample of all speeches we shall attempt to generalize from this purposive and selected sample to these states’ perceptions in general. For a more detailed discussion of this problem and related findings see the author's doctoral dissertation, “The Perceptual Base of Non-alignment,” Stanford University, 1967.

¹⁴ Nehru, Jawaharlal, India Today and Tomorrow (New Delhi, 1959), 15.Google Scholar

¹⁵ Ibid., 13–14.

¹⁶ Mideast Mirror, April 1955.

¹⁷ “Routine” and “formal” behaviours are both significant indicators of behavioural orientation and are taken into account in the large study of which this paper is only a brief introduction. This section is designed primarily as a pilot study of the behaviour of non-aligned states along a co-operation-conflict dimension. The analysis is in a preliminary state and is presented in the most direct manner. In further analyses we intend to examine the data using more detailed and sophisticated methods. At this point our purpose is to present the preliminary and tentative findings.

¹⁸ We are not testing for degree of integration with the major powers but for differences in terms of implications for conflict.

¹⁹ The specific period considered extended from January 1954 to December 1962.

²⁰ These sources are primarily Western and as such are likely to introduce some bias in the data. To compensate somewhat for this “slant,” we intend to examine Indian, Egyptian and Indonesian sources, primarily to assess the degree of consistency in the reporting of actions but also to complete our data collection wherever omissions in Western sources are evident. It should be noted, however, that we have no reason to believe that the data include biases which would distort the findings to a significant extent. Only future analyses will validate or reject this supposition.

²¹ Moses, Lincoln E.et al., “Scaling Data on Inter-Nation Action,” Science, 156, no. 3778 (1967), 1054–9.CrossRefGoogle ScholarPubMed

²² The 30-item scale is presented in Appendix A. Each data item is ordered along this scale according to the appropriate level of intensity, and our analysis is then based on the frequency of data items along each level of intensity. See below. This method is different from the Q-sort technique in that the latter allows for meaningful comparisons only within the one universe of actions being considered and cannot take into account inter-situation comparisons.

²³ A subsequent experiment on the interaction scale recently conducted by Edward Azar at Stanford University indicated that masked data do not yield statistically significant different results from unmasked data. See his “Scaling Data on Inter-Nation Action: A Follow-up,” 1968. It seems, then, that for future analyses of the nonaligned nations' actions, it is not necessary to undertake the masking procedure.

²⁴ It should be noted that Moses et al. suggest that a 15-item scale is functionally equivalent, and just as effective, as the 30-item scale used here.

²⁵ The non-aggregated scores for each state are presented in Appendix A. We have not conducted any reliability checks since the intensive pre-testing of this method indicated that a high reliability was found in the judges’ rankings of different sets of data (Moses, et al., 1967).

²⁶ Our intent is, in future analyses, to analyse the unaggregated items for each state using analysis of variance as a means of testing for significant differences in the means of the respective distributions, comparing the three states and evaluating changes over time. This technique would take into account both the frequency and intensity of actions in terms of conflict value.