1. Perspective And Summary
15A. Phasing Propositions and Their Evidence on International Conflict
Democratic Peace page
The great tragedy of Science--the slaying of a beautiful hypothesis by an ugly fact.
In previous volumes, I have delineated the generic process of conflict within social fields and provided relevant evidence. In Vol. 1: The Dynamic Psychological Field I focused on the psychological and philosophical foundations, and gathered together the systematic, psychological evidence on the dynamic field of motives, attitudes, and goals (see Part IV of that Volume). There, I was especially concerned with a solid, empirical foundation for needs and sentiments, especially the integrated self, against which prevailing theories about the drive for power, aggression, authoritarian personality, and so on, could be evaluated.
In Vol. 2: The Conflict Helix I described the social field and its process that is the conflict helix, and focused my discussion at the level of state-societies. A number of general propositions about social conflict were drawn from the analysis and all the systemic, empirical evidence I could find for and against the propositions was presented (see Chapters 34 and 35).
In Vol. 3: Conflict In Perspective I was concerned with contending theories and views on conflict and presented no systematic evidence. However, the ability of social field theory and its conflict helix to confront or absorb other perspectives as was shown is itself support for the theory.
Finally, in this book I have focused on one exchange society, one social field: international relations. The nature and components of the international field and behavior within the field have been presented and discussed (from 1979: Chapters 4 and 7), and the process of international conflict has been delineated (from 1979: Chapters 8-13). Many empirical results already have been presented, especially bearing on the components of the field and behavior (from 1979: Chapters 4 and 7, Appendix 9A, and Chapter 11). These results were used descriptively and as evidence for the theoretical components and actors of behavior space and the international social field (Chapter 4 and Chapter 7).
Descriptive empirical results are assumed to map some aspect of reality. There is, of course, the question of validity and reliability; and the user of descriptive results should know and present the evidence for and against a particular description, such as that wealth is a major component of states. However the descriptive results by themselves are not meant as evidence for a theory, but as the empirical foundation or framework for such evidence.
For example, to assert that components in a social field are forces towards behavior first requires an empirical description of the components. This description neither confirms or denies the theory, but provides the base for assessing the theory. Once the empirical mapping determines the components of the social field, then the relationship among these field and behavioral components can be used as empirical evidence for and against the proposition that behavior is dependent on the field components.
The test of whether results are used descriptively or evidentially depends on whether in principle the results could falsify the theory, hypothesis or propositions. Any possible, properly determined set of empirical components would comprise components of the international field, but not all possible empirical results would satisfy the proposition that behavior is dependent on field components.
In the following chapters, I will present a number of propositions about conflict drawn from the field theory and its conflict helix. And I will provide the evidence for each proposition. The purpose of this Chapter is to present the variety of considerations that will inform and guide this effort.
However, the result would be a partial picture, an incomplete argument. The "truths" presented might be intuitively convincing and theoretically sound, and the reader might therefore be persuaded. But truth stands on three legs: intuition, reason (theory), and experience (the empirical world). Therefore, a complete and persuasive picture of the field requires a systematically ordered confrontation with reality. Not a descriptive mapping, but hard and comprehensive tests against data with the greatest potential for falsifying a "truth." And this confrontation cannot be selective. All relevant evidence must be found and the criteria for what is relevant should be biased, if at all, towards a negative finding. But this is saying no more than we must be good scientists in dealing with evidence.
Granted the need for such evidence, then, to what is the evidence to be related? Theory, theorems, hypotheses, or propositions? The social field theory presented here and in the previous volumes is a comprehensive perspective involving hard theory (equations and explicit logic) integrated within a philosophical, metasocial and intuitive framework. The theory in its totality cannot be tested empirically. Aspects of the theory, such as the analytic core, can be tested, however, and in their accumulated results add empirical confirmation to the intuitive and rational persuasiveness of the overall structure.
Because of this intuitive-rational-empirical nature of the theory, theorems cannot be derived from the whole and therefore tested; although they can be from aspects of the theory, such as that involving status.
Then why not test hypotheses? Hypotheses can be made precise and related to both hard and soft theory. However, an hypothesis is a statement that is tentative; it is an assumption, a speculation, a starting point for research. And this is not what I will be asserting.
The statements about conflict that can be made on the basis of social field theory are more than provisional suppositions. They are propositions, definite affirmations about conflict grounded in a perspective on our psychology, society, and conflict.
Moving from the empirical descriptions of the previous chapters to empirical propositions will serve a number of purposes. First, the most important, empirical aspects of the international field and its conflict helix can be made specific and indexed as propositions. Second, propositions provide organization to a mass of empirical results and findings of which various dimensions, levels of meaning, and significance might otherwise be lost. And third, propositions serve as foci for arraying all the positive and negative empirical results bearing on a statement, and thus help in avoiding selective bias.
There is a greater tension here than in previous volumes between the scientific requirement to explicitly present methods and evidence and my desire to make these results useful and understandable to the nonspecialist. It is the usual conflict between the need to communicate to both the technical and general reader; the strain between the requirements of evaluation and communication.
In what follows I have tried to resolve this tension by (1) giving in each chapter a nontechnical summary and discussion of the propositions and their evidence, (2) consigning the specific proposition-by-proposition technical considerations of the evidence to a chapter's appendix.
In order to evaluate the evidence for a proposition in the appendices, I will define the proposition, discuss its theoretical basis, present the relevant evidence, and state a conclusion about the proposition's empirical status. The whole purpose is to provide the maximum information in the shortest space. The style may therefore be too systematic and formal--lacking in life--for some, but the reader might consider it penance for the soaring, metasocial prose he may have noticed elsewhere.
A second source of evidence is the systematic literature--books, monographs, articles, conference papers, and so on--on international relations and foreign policy. The literature and the parameters of this relevant research are given in Appendix III. The Appendix also describes the manner in which these sources were selected and the considerations governing their description. It also includes the published reports of the DON project.
As shown by the vertical dimension in Table 14.1, the evidence can be also organized by case, states, dyad, or system. Case refers to evidence based on the analysis of specific conflicts, incidents of violence, crises, or wars. For example, Butterworth (1976) analyzed 310 management attempts in 274 conflicts during 1945-1974. This was a static analysis of cases of conflict.
States refers to the analyses of sovereign or independent societies, whether nation-states, city-states, or primitive tribes. The unit of analysis underlying evidence at this level is the state, its behavior or attributes.
The dyadic level concerns the behavior, similarities and differences, geographic distance, and so on, between states. These usually will be dyadic (involve a pair of states) or triadic relations reducible to dyads. The unit of analysis will generally be a pair of states, such as the relations between China and Japan, Bolivia and Argentina, or North and South Korea.
Finally, there is the system level, which refers to the character of a system of states (e.g., the European system; the international system) and the aggregate, overall behavior within the system. An example of system level (and static) evidence would be the results of Singer and Small's (1966) analysis of the relationship between the number and character of alliances in the international system and the number of wars, 1815-1945.
At the bottom of Table 14.1 a place is given to surveys which have collected, summarized, or given an overview of relevant systematic results.
There are nine cells in Table 4.1, or nine different kinds of evidence for a proposition.
The element concerning the data is of particular importance. Much of the evidence to be presented is based on event data, a count or scaling of occurrences of different kinds of conflict and cooperative behavior. This type of data provokes many questions about data sources, validity, reliability, methods of analysis, and so on. Because of the importance of event data here, in Appendix II I discuss and overview such data for both the general and technical reader.
Regarding the whole list of elements, each source of evidence could be rated on each element and even an overall summary rating could be developed. But this would be misleading. The rating and method of forming the overall scale would be fundamentally subjective and involve an implicit weighting. I prefer to admit that these elements all will be gauged and weighted subjectively and the final judgment about a proposition will be an intuitive synthesis--guided by method, to be sure, but no less intuitive.
Thus, the best I can do here is to make my sources of evidence and their parameters clear and to present my logic as explicitly as possible. Whatever biases or selectivity involved may therefore become evident to the reader.
* Scanned from Chapter 14 in R.J. Rummel, War, Power, Peace, 1979. For full reference to the book and the list of its contents in hypertext, click book. Typographical errors have been corrected, clarifications added, and style updated.
1. See my Field Theory Evolving (1977), especially Chapter 16, for empirical tests of the analytic structure of field theory in international relations.
2. In the years when I considered myself a logical positivist (Rummel, 1976b), I had driven toward an axiomatic field theory from which theorems could be derived and tested. For the first such effort, see Rummel (1965), republished in (1977: Chapter 2). The last gasp of this effort was my status-field theory (1977: Chapter 9).
3. Reviews of the project are given in Hilton (1973), Seidelmann (1973), and Hoole and Zinnes (1976). For the autobiographical background, see Rummel (1976b). After 13 years, the project was terminated in June, 1975.
4. More precisely, the evidence is based on analyses in which time is an explicit or implicit variable. It is explicit if an actual time variable is used, say, in regression analysis. It is implicit if the data is organized and analyzed by time units, such as in a matrix of U.S.-U.S.S.R. negative communications for each year, 1946-1977.
5. There are some studies which do not fit easily into the table. Sample survey analyses, public opinion studies, or content analyses are classified as at the state level if assessing the attitudes, perceptions opinions, and the like of one or a number of states. If, however, these studies have analyzed such variables for a state toward a specific other, they are considered dyadic. The classification is made easier if it is kept in mind that the focus is not on attitudes, perceptions, or opinions, per se, but on what they manifest about cases, states, dyads, or systems.
6. This is more than a statistical degree-of-freedom question, which is only concerned with the number of cases versus the restraints of the statistical techniques used and the number of variables. Keeping techniques and variables constant, a study involving 22 wars spread through three centuries may be far more important than one of 22 incidents of military violence in 197 1.