HomePersonalDemocratic PeaceDemocide20th C. DemocideMegamurderersLesser MurderersWhy DemocideDimensionsConflictMethodsTheoryPolicyLinks

Volume 4

Expanded Contents | Figures | Tables


1. Perspective And Summary
2. International Relations
3. The International Actors
4. International Behavior Space-Time
5. International Expectations And Dispositions
6. International Actor And Situation
7. International Sociocultural Space-Time
8. Interests, Capabilities, And Wills
9. The Social Field Of International Relations
10. Latent International Conflict
11. International Conflict: Trigger, Will, And Preparations
12. The Balancing Of Power
13. Comparative Dynamics Of International Conflict
15. Empirical Dynamics Of International Conflict
16. Causes And Conditions Of International Conflict And War
17. Ending Conflict And War: The Balance Of Powers
18. The International Conflict Helix
19. Theoretical And Empirical Conclusions On Conflict And War 20. Principles Of Peace And Conflict


15A. Phasing Propositions and Their Evidence on International Conflict
16A. On Causes of International Conflict
16B. Propositions and Their Evidence on the Causes and Conditions of International Conflict Behavior
16C. Evidence on the Causes and Conditions of International Conflict Behavior
17A. Propositions and Evidence on the Causes and Conditions of Ending International Conflict Behavior
19A. Overall Evidence on 54 Social Field Propositions on International Conflict
19B. Primary Propositions on Social Conflict
I. Unpublished Research and Results on International Relations
II. Event Data: Bases of Empirical Conflict Analysis
III. Characteristics of Published Quantitative International Relations Studies

Other Volumes

Vol. 1: The Dynamic Psychological Field
Vol. 2: The Conflict Helix
Vol. 3: Conflict In Perspective
Vol. 5: The Just Peace

Other Related Work

Conflict And Violence page

Democratic Peace page


Appendix 18A

Descriptive Propositions On
International Conflict*

By R.J. Rummel


I offer eight descriptive propositions about international conflict, which are given in Table 18A.1. The format for considering the evidence for each proposition will be the same as in previous chapters.

Proposition 18.1 (Helix): Conflict is helixical.

Within a closed system Conflict Behavior will decrease in intensity through time and structures of expectations will last longer.

Theory: A structure of expectations becomes incongruent with the associated balance of powers (Change/Conflict Proposition 16.3) producing tension and hostility (Change/Tension Proposition 16.2). The greater this incongruence, the more likely the structure of expectations will be disrupted by some trigger (Surprise and Perception Proposition 16.28 and Proposition 16.29). Disruption creates conflict (Disrupted Expectations Proposition 16.1); and conflict negotiates a new structure of expectations more in accord with the changed balance of powers--a structure which provides a framework for cooperation. And this new structure may in time become incongruent, be disrupted, again creating conflict.

Thus, cooperation-conflict-cooperation-conflict ... the process seems circular, periodic. Conflict appears cyclic.

However, no conflict begins nor is a structure of expectations created de novo, without history. Leaders and peoples learn. Each conflict is informed by previous ones. The next structure of expectations is built on the experience of those that went before. Conflict becomes ritualized or institutionalized, as various tacit understandings, rules, and limits govern conflict behavior. Successive structures of expectations tend to last longer as they become based on deeper and wider experience and communication between the parties.

As shown in Figure 18.1 of the text, the conflict-cooperation cycle is therefore theoretically a helix. Each successive turn from cooperation to conflict is at a higher level of experience and insight between the parties; each turn incorporates what was learned before.

But this assumes a closed system. That is, the relationship between parties neither suffers from internal shocks (a revolution or ideological coup d'état, for example) nor external shocks (what were called "disruptors" in Section 29.5 of Chapter 29 in Vol. 2: The Conflict Helix), intervention or challenges. The successive conflict and cooperation between the parties thus can gradually adjust and increasingly find a long-run equilibrium, like a tribe isolated in a wilderness or a couple in a small farming community.

An ideological change in a state's regime, the intervention of a Big Power in a state's affairs, a breakdown in the international economic system, a producer's boycott (as with oil in 1973) affecting all states can create new conditions and require a whole new order of balancing and expectations between the parties.

All this is the conflict helix which has been developed previously in Vol. 2: The Conflict Helix and presented here.

Prediction: As long as the conditions surrounding a relationship change gradually then the periods of cooperation between two states should last longer and the conflicts that do occur should decrease in intensity.

"Cooperation" here means cooperative flows, such as trade, tourists, students, migrants, mail, messages, aid, cultural and scientific exchange; as well as cooperative structures, such as the comembership in governmental and nongovernmental organizations, and the framework of treaties and agreements between the parties.

There is a potential confusion here between these and a second kind of cooperation that which is part of the balancing of powers, the process of conflict. Promises, negotiation, offers, signing a treaty, accepting tacit rules of conflict, and so on, play a role in the process of conflict (as discussed in Section 12.4 and for Proposition 15.2 in Appendix 15A) and do not necessarily reflect the structures of expectations determined by the conflict.

The proposition says nothing about this cooperation which is part of Conflict Behavior. It does say that the cooperation which is part of the structure of expectations established by conflict should, internal and external conditions remaining relatively constant, become more durable.

Evidence: Table 18A.2 presents the evidence on all the propositions; Table 18A.3 tabulates the accumulated evidence by proposition and category, as discussed in Appendix 16C.

As can be seen, the Helix Proposition or hypotheses similar to it have not been tested in the literature, nor do systematic results generally relate to it. Two studies, however, have results close enough to the proposition for me to infer support.

One is by Most and Starr (n.d., 1976). For three different data sets on war, 1946-1965, they apply a Poisson or Modified Poisson approach to analyze the diffusion of war. In their words (1976: 34),

The analyses here indicate that nations which were at peace generally tended to remain at peace. Nations which were at war in one period may have had higher rates of proneness to subsequent new war participations, but they appear to have followed some natural "regression" toward [if not actually to] peace so that they became involved in fewer war participations in subsequent intervals.

The second study is by Azar et al. (1974). They analyzed reactive process models for 11,664 events transformed into a cooperative-conflict scale for the years 1950-1971 and 12 dyads. Were the Helix Proposition correct, then dyads which are acting primarily with each other in view (i.e. major-major, or minor-minor powers acting within a defined structure of conflict) should move towards a closer linkage of their cooperation and conflict through their mutual learning--the helix. That is, their reactivity coefficients should increase in time. Azar, et al., found that eight out of nine appropriate dyads have positive regression coefficients when their reactivity coefficients are regressed on time.

Conclusion: Of the two sources of evidence, both are indirectly relevant and are important. But the level of indirect support is not strong in either case. Caution therefore dictates the conclusion: the evidence tends to support the proposition.

Proposition 18.2 (Cooperation-Conflict): Conflict occurs across diverse structures of expectations.

Cooperative flows and structures are uncorrelated with aggregate Conflict Behavior.

Theory: Conflict Behavior begets a balance of powers and associated structures of expectations. And this structure is the framework for cooperation. But structures of expectations may become incongruent with the supporting balance, some trigger then may disrupt the structure, producing Conflict Behavior through which a new balance and structure of expectations is negotiated. Thus, cooperation and conflict are complementary aspects of a process of mutual communications, adjustment, and cooperation.

But there are many overlapping and layered structures of expectations. For example, a trade treaty between two states may exist within an overarching mutual security treaty and also be within the framework of a multinational agreement on trade. There may be related, subordinate, or superordinate understandings, agreements or treaties on aid, military basing, off-shore fishing, foreign investment, and so on.

Conflict can occur in the formation of, or because of, the breakdown in any of these structures of expectations. Which is to say that conflict in the aggregate (e.g., the number of accusations, warnings, sanctions of one state to another) reflects diverse balancings of powers; cooperation in the aggregate occurs within diverse structures of expectations. There is therefore no necessary relationship between the type and level of aggregate conflict and the type and amount of aggregate cooperation.

Moreover, there is no reason within field theory to expect in general a certain type and volume of cooperation to follow from a balance of powers established through a specific kind or intensity of conflict. In theory, the expectation is to the contrary. Because conflict is situational and functions to communicate and adjust different interests, capabilities, and Wills, the kind and level of conflict should be uncorrelated across dyads or nations with the kind or level of cooperation which it begets.

Putting these last two paragraphs together, then, in the aggregate conflict and cooperative flows and structures should be uncorrelated.

Prediction: Across levels (cases, states, dyads, and systems) and across time, one should find that various dimensions, components, or factors of cooperative flows or structures are uncorrelated with types or subphases of Conflict Behavior.

Evidence: Table 18A.2 lists fifty-one different analyses bearing on the Cooperation-Conflict Proposition. Table 18A.3 divides them by category. As can be seen, sixteen analyses are strongly positive and direct, with six being from important studies; twenty-one are positive, and all but one directly bear on the proposition. Only two of the positive analyses, however, are from important studies. There are eight negative analyses, three of these strongly so.

Conclusion: Since 73% of the analyses support the proposition and only 16% are against, it follows that: the evidence strongly supports the proposition.

Proposition 18.3 (Internal-External): Internal and external conflict behavior are independent processes.

The internal (domestic) and external (foreign) conflict behavior of states is uncorrelated.

Theory: Conflict within states usually reflects the formation and breakdown of diverse, wholly internal structures of expectations (Part IX of Vol. 2: The Conflict Helix). Similarly, a state's foreign conflict usually concerns wholly international structures of expectations. That is, in general.

Of course, there are situations in which internal and foreign structures overlap or mutually influence each other, as will be discussed for the War Peak Proposition 18.6. Additionally, abrupt changes in the ideological character of a government, a shift in domestic interest group power, can affect the global or regional Big Power status quo and balances of powers. The victory of Castro in Cuba in 1959 is a case in point. Moreover, foreign conflict can certainly affect domestic structures, as did the Vietnam war on U.S. politics, especially the balance of powers between Congress and the President.

However, specific linkages a correlation do not make. The fact that numerous examples of a relationship between internal and foreign conflict behavior can be shown does not invalidate the theoretical argument: in general the states' border divides two different and usually independent processes.

There are other considerations. Aggregate internal and external conflict behavior are associated with the formation of diverse expectations. Even if in the formation of a particular structure of expectations there is a relationship between internal and external conflict behavior, other, ongoing balancing independent of these structures will also contribute to the aggregate behavior. Moreover, as with the Cooperation-Conflict Proposition 18.2, there is no theoretical reason to expect that a particular type and intensity of internal conflict behavior should be associated with a specific type and intensity of external conflict behavior, even when across states the same kind of internal-external structure is involved.

Prediction: There should be no aggregate correlation between internal and external conflict behavior and dimensions or types. There is one exception to this prediction. At the dynamic system level there should be a positive correlation between war peaks and the intensity of internal conflict behavior, as explained in the War Peaks Proposition 18.6.

Evidence: There are 46 analyses bearing on the proposition as listed in Table 18A.2 and totaled in Table 18A.3. Twelve are direct and strongly positive, six of which are from important studies; twenty-five analyses are direct and positive, and six of these are also important studies. There are five negative analyses, four of which are direct, one is important.

From the earliest (1963) publication, the facet of my work which has attracted the most attention has been the finding of a statistical independence between internal and external conflict behavior. As can be seen in Table 18A.3, this independence has been a consistent finding in my subsequent analyses, as well.

Conclusion: Of the evidence 80% is positive; 28% is strongly so. Only 11% is negative; none strongly. Therefore: the evidence strongly supports the proposition.

Proposition 18.4 (Conflict-Independence): Conflicts (and Wars) are (statistically) independent.

Within any time period, the occurrence and intensity of conflict (and wars) are independent (random).

Theory: Conflict Behavior is unlike an infectious disease sweeping through the community of states. Nor when conflict breaks out between two states is it emulated by others. Rather, Conflict Behavior is a process of balancing intrinsic to the life stream of two states. Whether Conflict Behavior breaks out between them depends on the congruence between their diverse expectations and associated balances of powers.

Now, it clearly may be that these expectations and balances overlap or are entwined with other states, as the U.S. and South Korean structures are interrelated with the security structure between Japan and the United States; or as the expectations between the United States and Egypt, Egypt and Israel, and Israel and the United States have become entwined under President Carter.

There are scattered throughout the globe nests of expectations and power balances involving several, and sometimes, numerous states, such as in the Middle East, Western, or Eastern Europe. Conflict Behavior anywhere in this nest can spread throughout by its effect on the balances of powers or as a trigger. However, there is no reason for Conflict Behavior to spread from one nest to another, unless, and this is the only reason, it escalates to involve the general regional or global structures or expectations. General war affects all balances and causes multiple outbreaks which, while intrinsic to particular dyads, become absorbed in the larger confrontation.

With the exception of general war, then, the distribution of Conflict Behavior and intensities around the globe at any one period is independent. The reason for this and the exception are given by the conflict helix.

Prediction: One should find that the frequency and intensities of conflict behavior within some time period do not significantly differ from a Poisson distribution.

Alternatively, the average temporal correlations between the conflict behaviors of dyads (e.g., the by-month correlation between the conflict between Libya and Egypt with that of Ethiopia and Somalia, 1970-1975) should be low in general, although for some specific clusters of dyads the correlations will be high, reflecting their interlocking expectations.

Evidence: Overall, there are five directly relevant analyses, all strongly supporting the proposition. Three are from important studies (Richardson, 1960b; Most and Starr, n.d., 1976; Singer and Small, 1972).

Conclusion: the evidence strongly supports the proposition.

Proposition 18.5 (War Probability): War is state specific.

States differ in the probability of their going to war.

Theory: The logic here should be clear. States form different structures of expectations and supporting balances within situations of conflict. Situations will differ by state, as will the kind of balances of power and structures of expectation they form (Actor Proposition 16.7). Moreover, the incongruence of these structures should also differ, as would their probability of disruption.

All this is to say that states differ in their probability of going to war--that war is state-specific.

Prediction: States will differ significantly in the frequency with which they engage in war.

Evidence: Two direct and important analyses strongly support the proposition (Most and Starr, n.d., 1976; Singer and Small, 1972). There is no negative evidence.

Conclusion: The evidence supports the proposition.

Proposition 18.6 (War Peaks): A disruption of the general international structure of expectations is also a disruption of the internal structure of expectations.

Peaks in international war will be correlated with peaks in internal war.

Theory: Usually, internal and external structures of expectations are independently formed--they are separate streams of activity. However, sometimes there is a flood and the international stream crosses over to engulf the other. This occurs when the general international status quo is disrupted--when a general war occurs. Or when a number of regional wars reflect the breakdown in the international order.

Any state-society is a balance of powers between different groups. A breakdown in the international system and the requirements and destabilizing effects of intense war upset this balance and cause power to be reordered. Internal unrest, and revolutionary activity--internal war--is often the result. This may not occur during the external war, when diverse internal groups may be united in defense, in pursuing the war, or be holding back waiting for the uncertainty of war to be clarified. Among victors, neutrals, and losers alike, internal conditions and power balances "I be sharply altered by the war. And new balances more consistent with the new international status quo must be determined.

Prediction: Within the international system or within isolated regional systems of states, peaks in international war tend to cooccur with peaks in internal war.

Evidence: Two important (Denton and Phillips, 1968; and Denton, 1969) longitudinal analyses of wars strongly, and directly, support the proposition. Three other analyses, two of which are important, also directly support the proposition. There is no negative evidence.

Conclusion: The proposition is strongly supported by the evidence.

Proposition 18.7 (War Cycles): War is cyclic.

War's frequency and intensity are correlated with leadership, generational, and system turnover.

Theory: Holding other conditions relatively constant (that is, the system is closed or isolated), wars for a state, dyad, or international system, should decline in number and intensity in time and the periods of peace should increase in duration. This is the Helix Proposition 18.1.

The reason for this is that people learn from previous wars and structures of expectations. No war or structure of expectations begins anew. All are the present thrust of a past stream of conflict and cooperation.

Organizations, through their law norms, rules and procedures; and societies, through custom, practices, and norms; all incorporate past experience and frame it for the present. States are always thus guided. Nonetheless, states do not make decisions. Leaders and peoples do. And there is a constant turnover of leadership and generations of people. Thus, the mistakes of the past tend to be repeated. The lessons of history tend to be ignored.

Therefore, the frequency and intensity of war has certain cyclic patterns. One cycle should be associated with the pattern of turnover in leadership within states. This of course will differ from state to state, but it appears that the leadership of a state changes on the average every four to six years. In 1955 the average age of the previous two governments for 82 states was 4.88 years, with a standard deviation of 4.12 (Rummel, 1972: 182, variable no.74).

Change in leadership is change in psychological fields, in perspectives, in interests. Most Important, it is a change in the sense for a state's role and position--the nuances of its balance with others. But the affect of this change on the state's structures of expectations should not be immediate, except in the case of a revolutionary turnover. New leadership takes time to get in the saddle, and their change in interests, will, and capabilities take time to produce conflict. Probably the most intense conflict associated with a breakdown in the expectations inherited by a new leadership and relearning lessons that passed out of office with the old will occur towards the end of their average term. The new leadership that then comes in will probably have the task of completing the building of new structures of expectations. Thus, the effect of change in leadership should be seen most in conflict behavior about every eight to twelve years.

At first thought, leadership changes would appear across states to be uncorrelated. However, such changes should have a rough correlation associated with two changes in a system of states. One is the occurrence of general or regional wars, which is usually the cow of a massive turnover in leadership among victors and defeated alike (War Peak Proposition 18.6). Second, a breakdown in the international economic system, as in 1932, will cause widespread leadership turnover.

Of course, such associated cross-national turnovers still leaves a bureaucracy in place. But intense regional and general wars deeply and permanently affect all; all within a society learn from them. War creates pacifists as no abstract lesson ever can.

Moreover, the lessons of the war--how it occurred, and the consequent expectations--are also deeply and widely learned. The so-called Munich model, as to how Britain's appeasement of Hitler's demands in 1938 led to World War II, had an iron grip on American thought and policies regarding Soviet behavior for decades after the war.

Thus, a generation becomes imbued with a perspective and an ethos. But generations change and with them goes their sense for the past. As any parent knows, the new generation simply must learn by themselves many hard lessons. This is no less true for the generational shift in leadership throughout the state.

But what constitutes a generation in time? This is an empirical, rather than theoretical question. The UN 1971 Demographic Yearbook tabulates five births by age of mother for most countries in the world (Table 24). The highest proportions of births occurs within 20-24 and 25-29 age groups. The proportions in these two age groups are often close, usually higher in the first age group, with notable exceptions (Switzerland, Ireland, Italy, Netherlands, Albania, Jordan, and Japan, among others). The definition of a generation thereby closest to actual international birth statistics appears to be 25 years.

Leaders change, generations turnover, and eventually, the social and political systems also undergo significant change. Weakened and altered are the lessons of the past which were consciously and unconsciously incorporated into the society's institutions and culture, prevailing world view and sense for foreign policy and defense. This world view sets a framework for policy and opinion.

Two generations, about 10 leadership turnovers, or 50 years appears sufficient for incremental internal and international changes to add up to a major, indeed, revolutionary departure from the past. That is, the general and specific frameworks of expectations formed out of an upsurge of wars and violence about 50 years later no longer reflect the multiple internal and international balances of power.

Prediction: The frequency of wars, the casualties, the number of battles in the international system, should show significant cycles, irrespective of trend. These should be within a few years of 10, 25 and 50-year cycles; that is, cycles in leadership, generations, and system.

Because three superimposed cycles are predicted, what precisely would constitute positive or negative evidence for the proposition? As a procedure, any analysis is negative which does not uncover (when it is possible to do so) at least one of these three cycles, while within the design it is not possible to pick up the other cycles. Any analysis is ambiguous which fails to uncover one or two of the cycles when it was possible to do so, while finding the other cycles. Any analysis is positive which shows at least one of these cycles to be present, while within the design it was not possible to uncover the other cycles.

Evidence: There are six important, positive analyses bearing on the proposition, all but one of which are direct. Two important, direct negative analyses also exist (Richardson, 1960b; Sorokin, 1957).

I have categorized the six cycle studies of Dewey published over nine years (1962-1971) as one analysis. This is because they all deal with the same historical data on the frequency of battles (600 B.C.-1957A.D.), even though Dewey refines his cyclic analyses as he goes along. He finds (1970) a 11.24 year cycle (p = .0018) over the whole period; a 21.98 year cycle from 524 B.C. through 1938 (p = .0008); a 53.5 year cycle from 1700 to 1913 (p = .032). His analysis provides the only strongly positive evidence.

Conclusion: The evidence supports the proposition.

Proposition 18.8 (War Trend): Wars are neither increasing nor decreasing.

There is no long term trend towards an increase or decrease in the number of wars.

Theory: In a closed system where international conditions remain relatively constant, wars should decrease in number and severity. This is the Helix Proposition18.1 .

However, conditions are seldom unchanging, except among isolated tribes with strong, authoritative cultures sufficient to overcome the effect of leadership and generational changes. Otherwise, leadership, generational, and system changes are sufficient to wash out the mutual learning and balances established between states. The iterative process of gradual adjustment through periodic conflict--the conflict helix--must begin anew. Thus, there are cycles in war. Thus, the long-term trend in warfare should not have significantly changed.

Prediction: Regardless of the cycles existing in data on the frequency of war, the long-term secular movement in warfare in international relations should show that the trend neither decreased nor increased.

Any analysis which shows the long-term secular trend had significantly increased or decreased is a negative result. Long term means over a century; that is, over two full 50-year cycles, which are the longest theoretical cycles (the War Cycles Proposition 18.7).

Evidence: Three important and direct analyses strongly support the proposition (Richardson, 1960b; Singer and Small, 1972; Sorokin, 1957). R. Richardson (1966) found no trend in the frequency of 380 conflicts, 1946-1964. Although this finding was limited in time to 18 years and covered a variety of violence, it showed the proposition to have a short-term validity also. However, as evidence this was considered only positive and others may wish to exclude it altogether. In any case, exclusion would not affect the conclusion drawn here.

Conclusion: The evidence strongly supporters the proposition.


In summary, see the total and percent columns of Table 18A.3. They show by category that 33% of the evidence strongly supports the eight descriptive propositions; and about four-fifths is supportive. Only about 12% is negative. When a one-level decrement is applied, most of the evidence is still supportive for each category.

There is little change in the percentages between categories, except for my own analyses, all of which were supportive. The reader may wish to exclude these as possibly showing bias (because they depart from the proportions in the other categories). However, these comprise only 12 out of the 126 analyses and their exclusion would not change the overall conclusions.

Table 18A.4 summarizes the conclusions on the separate propositions. Overall, the descriptive propositions are strongly supported by the evidence.


* Scanned from Appendix 18A 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.

For citations see the Vol. 4: War, Power, Peace REFERENCES

Go to top of document