1. Perspective And Summary
15A. Phasing Propositions and Their Evidence on International Conflict
Democratic Peace page
The first column defines the types of cause or condition, as described in Appendix 16A: necessary and sufficient cause, sufficient cause, necessary cause, trigger cause, aggravating condition, inhibiting condition, and relative aggravators.
The relative aggravators refers to propositions describing the operation of an aggravating condition relative to other such conditions. Thus, of all attributes of nations, (coercive) power attributes are theoretically most highly correlated with conflict behavior (Proposition 16.31); of all differences between nations, differences in power, wealth, and politics are most highly correlated with conflict behavior (Proposition 16.33); among these three, power is most correlated (Proposition 16.32); and the more totalitarian a state, the more correlated its conflict with political differences (Proposition 16.30).
Each proposition in the Table 16B.1 is given a number and name for easy identification and each is given both an abstract and empirical description. The "empirical" (or operational) description is specific about how observed causes or conditions should relate to conflict. It is the description that a social scientist would probably prefer.
The "abstract" description is the general and compelling way to describe the proposition. It is more intuitive, in a sense poetic. It would probably appeal more to the historian, social philosopher, or polemicist.
The abstract and empirical descriptions of the same proposition are not just alternative statements. They are meant to describe the causes and conditions through different perspectives, each of which is believed true, and both of which together give a fuller appreciation of the cause or condition.
In some cases there is virtually no difference between the empirical and abstract description, except the latter is shorter, as in Proposition 16.1. In other cases, the difference is such as to apparently involve dissimilar propositions, as with Proposition 16.31.
As in Appendix 15A, each proposition will be discussed from the perspective of theory, prediction, evidence, and conclusion. The evidence for all the propositions is presented in Table 16C.3 and Table 16C.4. Appendix 16C also discusses this evidence in general terms, and considers several sources of possible bias. Appendix 16C would be useful reading, before moving into the individual propositions here.
Proposition 16.1 (Disrupted-Expectations): Conflict behavior if and only if disrupted expectations.
Conflict Behavior occurs if and only if a mutual structure of expectations is disrupted.
Theory: The character of Understanding Conflict and War as a whole underlies this proposition, and particularly Part VII of Vol. 2: The Conflict Helix dealing with the conflict helix. In short, peace and order exist within a structure of expectations based on a balance of powers. As this underlying balance becomes incongruent with the structure of expectations, the probability increases that some trigger will disrupt expectations. Manifest conflict then will occur as a new balance of powers and associated structure of expectations are determined.
This disruption of a structure of expectations is a necessary and sufficient cause of international Conflict Behavior. Necessary, in that I assume by theory that Conflict Behavior could not take place without a relevant structure of expectations having been disrupted.
I am assuming that all international relations now take place within mutual or multiple structures of expectations; that new structures are transformations of previous ones. Even new states adopt existing structures as a first approximation and develop, in the process of negotiating independence, new structures to fit their situations.
In social relations within states it is possible for people to first meet, develop initial expectations from their perceptions of each other's field of expression (Part II of Vol. 2: The Conflict Helix), yet have not worked out relevant structure of expectations. This will come out of subsequent conflict which may or may not involve Conflict Behavior. Thus, the conflict helix in Figure 29.1 is shown with a structure of conflict leading to a situation of conflict, to balancing, and to a structure of expectations. In this case Conflict Behavior could occur without the disruption of a structure of expectations, because none was yet in place.
In international relations we no longer have the "strange new boy on the block." The globe has become integrated through modern communications and trade; international elite are acquainted with each other's cultures and societies through travel, foreign residence, education, and involvement in the United Nations. Diverse nonintergovernmental organizations involve in some way all countries, whether independent or not. And some multilateral international legal and economic arrangements are initial structures within which new states must operate. Past is the historical meeting and confrontation of societies and cultures completely new to each other. The age of discovery and exploration is over: international conflict now functions to reorder structures of expectations, rather than establish them de nova
The disruption of structures of expectations also is a sufficient cause of Conflict Behavior: disruption implies some kind of international Conflict Behavior.
While in personal relations tacit adjustments may take place in a balance of powers without Conflict Behavior--one accepts the other's advantage or submits to the changed situation, fearing the consequences--states are not people. They are collections of balances of powers among diverse interests. Therefore, no breakdown in a structure of expectations will be accepted without some kind of manifest conflict and adjustment--accusations, protests, offers, promises, alliances, troop movements, negotiations, agreements. Nor does it mean that the behavior is necessarily antagonistic or violent. It does mean, however, that some disruption-related, conflict-reflecting behavior will occur.
Prediction: Because disruption is both necessary and sufficient, the occurrence of any International Conflict behavior implies that some structure of expectations has been disrupted; the disruption of any structure of expectations implies Conflict Behavior will follow.
This is the ideal causal relationship for correlational analysis, because the product-moment correlation or its variants best reflect necessary and sufficient causation, if it exists (see Understanding Correlation). While a high correlation between x and y does not imply causation, of course, if x is a necessary and sufficient cause for y, however, then they should have a high (or perfect, if one assumes no measurement, collection, and manipulative errors) correlation.
Therefore, the prediction is of a high (or moderate, if much operational slippage exists) positive correlation between disruption and Conflict Behavior in samples involving some members without Conflict Behavior.
If in the mixed sample the conflict variable is measured in any other way but as a dichotomy, then the correlations are predicted to be near zero or moderate positive. The reason for this in the case of disrupted expectations is discussed in Appendix 16C with regard to Figure 16C.l .
In samples in which all members have Conflict Behavior (e.g., a sample of war dyads), then disrupted expectations should be present in all cases. Technically, disrupted expectations would be a constant and the product-moment correlation coefficient would then be indeterminate. But conceptually the correlation should be zero: the two variables (vectors) are orthogonal.
Similarly in samples involving no Conflict Behavior. There should be no disrupted expectations, and therefore in this sample the correlation should be zero, conceptually, between disrupted expectations and lack of conflict behavior.
Evidence: As shown in Table 16C.3, there are seven analyses bearing on the proposition. The most powerful support is given by Sorokin's work on historical movements in war and peace (1957), and Blainey on the causes of war (1973). Melko's (1973) analysis of 52 peace societies also lends considerable indirect support. Indeed, the sum total of both Melko's and Sorokin's findings cannot be well understood without assuming the correctness of the Disrupted Expectations Proposition.
All except one of the sources of evidence are important studies (see Table 16C.4), and over half bear directly on the proposition. Moreover, most analyses did not violate the Model II and distance vector assumptions of the social field. And five of the seven relevant analyses provide strong positive support for the proposition; and all five were done by others than myself
Finally, special attention should be given to Sorokin's historical-cultural analysis of war (see especially Sorokin, 1937-1941), and his overall analysis of cultural change. His theory is that revolution and war manifest the breakdown in the values and norms of sociocultural systems; the shift to a new system. Peace then is a condition of crystallized values and norms. On this theory he brings to bear a diverse collection of positive historical evidence. Clearly, as I discussed in Volume 3 (1977: Section 8.1) Sorokin's theory is easily rewritten in terms of conflict helix: the nature of his whole work provides strong support for the Disrupted Expectations Proposition.
Conclusion: While only a comparatively small number of analyses relate to the proposition, the importance of these analyses and their strong support for the proposition give it solid credibility. Therefore: the evidence strongly supports the proposition.
Theory: No disruption may occur in the structure of expectations, but the balance of powers and expectations can have become so incongruent that the climate of relations--the atmosphere--can be altered. In short, tension is in the air, hostility is felt. There may be no specific conflict behavior to point to, but relations feel strained.
Change should satisfy two conditions before it produces such tension and hostility. First, it must be in the balance of interests, capabilities, and will. One party may have lost interest in maintaining a status quo, its will may have weakened, or the other party may have grown considerably in military power. The old balance of powers has been altered; but the old structure of expectations remains in place. Thus, tension.
This leads to the second condition: the rate of change in the structure of expectations and balance of powers must differ significantly. Structures of expectations evolve as the parties interact. If this evolutionary change is in line with changes in the underlying balance of powers, then no tension or hostility should result. But if the change in expectations does not keep pace with that in the balance of powers, then eventually this growing incongruency will cause tension and hostility (Section 8.5 of Chapter 8 in Vol. 3: Conflict In Perspective). Such change is thus a sufficient condition.
What about Conflict Behavior generally? Such change is not a sufficient condition for Conflict Behavior generally, nor for antagonistic Conflict Behavior, because a disruption in the expectations has not yet, or may not at all, occur. 'Me incongruency may be there, tension may be in the air, but no suitable trigger may have yet occurred to disrupt the expectations, or adjustments in expectations or the underlying balance may lessen the incongruency. Interests may be revitalized; will reaffirmed; armaments cut back.
Conflict may not always be thus avoided, nor may it always be desirable to do so. But this does point out that there is no inevitability to incongruency producing conflict behavior.
Prediction: A significant change in the balance of powers should be positively correlated with the occurrence of tension and hostility. But what about Conflict Behavior generally?
Now, tension and hostility can be present without specific, manifest Conflict Behavior; moreover, manifest conflict behavior can occur without tension and hostility (e.g., a mild protest of the United States to Canada). However, in general, high or intense antagonistic conflict behavior will be associated with tension and hostility. Therefore, in a mixed sample a positive correlation should be expected, as shown in Table 16C.1. Similarly for a conflict scale.
For nonviolent conflict behavior, however, tension and hostility may be present or absent for cases above or below the mean, thus meaning the predicted correlation is random: it hinges on the peculiarities of the sample.
For all other samples and conflict variables, the results are irrelevant. This is because high violence and war transcends tension and hostility, and these terms are no longer applicable.
Evidence: Four analyses are relevant (Mahoney, 1976; Abravanel and Hughes, 1973; Buchanan and Cantril, 1953; M. Sullivan, 1970). Three have strongly favorable results, one is ambiguous. All the positive evidence is indirect; and the studies are not important ones relevant to the propositions. Nonetheless, there are no negative results.
Conclusion: the evidence supports the proposition.
Conflict Behavior occurs only if there is a significant shift in the balance of powers (interests, capabilities, will).
Theory: International Conflict Behavior implies that a relevant structure of expectations has been disrupted. This could occur only if there were some incongruence between the balance of powers and the structure of expectations. And this incongruence is produced by significant change in associated, relative interests, capabilities, and will. Therefore, conflict behavior implies change. Or, to turn this around: a significant change in the balance of powers is a necessary cause of conflict behavior.
Why not sufficient? A structure of expectations may be incongruent, but no trigger may have yet occurred to disrupt it or steps may have been taken to readjust the balance, thus reducing the incongruence (by a gradual increase in arms, reemphasis on relevant national interests, revitalized will).
In social science parlance, disruption is an intermediary variable whose presence is necessary before change causes conflict behavior.
Prediction: Assessing this necessary condition is empirically complex. It must be present in all conflict behavior, so that if measured as a dichotomy it should easily show up in cross-tabulations of presence-absence against the conflict variables. But even here there is a problem. What constitutes significant change?
First, the change must be in interests, capabilities, or will.
Second, the change must be significant to the parties involved: it should be such that (1) one or both parties are dissatisfied with the formal or implicit contract governing their rights and obligations--the structure of expectations--and (2) believe that the shift in the balance of powers favors their revising the contract to their advantage.
Third, the change must be relative to that in some structure of expectations. Absolute change may be irrelevant, because a structure of expectations itself may have slowly evolved congruent with this change in the balance of powers.
Clearly, empirical studies must be designed with this proposition in mind to adequately test it. For this reason I cannot specify the correlations that should occur for relevant variables in a mixed sample, except when Conflict Behavior is measured as a dichotomy as shown in Table 16C.1.
By theory, all Conflict Behavior manifests significant change in the balance of powers. Therefore, for all cases with Conflict Behavior significant change should have occurred. Thus, as shown in Table 16C.1, such change will be present for all conflict variables in the conflict, violence, and war samples.
But in the mixed sample for the nonconflict cases, some may have significant change but Conflict Behavior has yet to occur, or some may have no significant change in their balance of powers. The tendency of the correlation, then, should be moderate positive for the dichotomous conflict variable.
For all other conflict variables of the mixed sample, conflict behavior could occur beneath the mean (see Figure 16C.1), thus complicating the prediction of what correlations to expect. If change has occurred, but the associated conflict behavior is beneath the mean, then this would contribute to a negative correlation--even though both significant change and conflict are present. For this reason, I simply assumed a random correlation for all other conflict variables in the mixed sample.
Because for the conflict, violence, and war samples, change should be present in all cases, the correlations then should be near zero between significant change and all conflict variables in the samples.
Evidence: The evidence is all positive (four analyses) or strongly positive (four analyses). Almost all are direct and more than half come from important studies; most are in design consistent with field theory; all are done by others than myself.
As with the Disrupted-Expectations Proposition, there are also some studies here whose character and importance give them special weight, namely the analyses of peace societies by Melko (1973) and Sorokin's (1957) comprehensive historical analysis.
Conclusion: Considering that the evidence without exception is consistent with the proposition and weighing the nature of the studies involved, then the conclusion: the evidence unanimously supports the proposition.
Geographic distance is negatively correlated with Conflict Behavior.
Theory: Contact is a necessary cause of conflict; it is an empirical concept in latent conflict (see Table 10.1). It generates the awareness necessary to transform a conflict space into a structure. Awareness is present in all conflict.
One dimension of contact is the physical or functional geographic distance between states: their contiguity; their air distance; the type of terrain between them; the transaction distance (the time it takes or the relative expense of shipping goods between them); the number or size of states between them. Geographic distance is thus a complex concept referring to the ease and likelihood of two states directly communicating with each other.
It is also necessary to keep in mind that contact is a necessary cause with a gradient. This is like gravity, which is necessary for a fall, but which also varies in degree (by height above sea level, or by planet). Contact should not only be present in all Conflict Behavior, but the more diverse and intense the Conflict Behavior between parties, the more likely (operationally) the parties are close; and (theoretically) the more their mutual awareness.
That is, the scope and intensity of awareness, while not sufficient for conflict, enhances the operation of other causes, and makes conflict behavior more likely in occurrence, scope, and intensity once the appropriate conditions are present. The stronger the pull of gravity (within a certain gravitational range), the more likely one will slip on ice, if not careful. The more international contact, then the more: important and diverse the opposing interests, likely diverse structures of expectations, likely associated conflict in their formation and disruption, diverse and intense the manifest conflict.
Degree of contact therefore also acts as an aggravating cause (as indicated in Table 16A.1), while contact per se is a necessary cause. Contact is always present, and even more so in the upper ranges of Conflict Behavior.
Geographic distance is one operational dimension of contact. A second is joint power, which is formulated in the next proposition.
Prediction: Because contact is a necessary cause that can vary in degree, contact should not only be present in all Conflict Behavior, but the more diverse and intense the Conflict Behavior between parties, the more likely (operationally) the parties are close; and (theoretically) the more their mutual awareness.
In the mixed sample for the dichotomous conflict variable, eases with conflict should be, on the average, closer geographically (where closer means less difficult passage between them) than states without Conflict Behavior.
This should not be a high correlation, however, because some states without Conflict Behavior between them may also be close. They may have stable structures of expectations, some of the other necessary causes for conflict may be absent, or they may have an incongruent structure of expectations yet to be triggered.
Therefore, at best, the correlation between geographic distance and conflict behavior should be moderately negative, as shown in Table 16C.1. The same logic applies to the conflict scale.
There also should be a moderate, negative correlation of geographic distance with the conflict scale in the conflict and violence samples. Above the mean conflict behavior on the scale will be more diverse in scope and type and more intense and this should be correlated with the scope and degree of contact. This may not be true in the war sample, however, where parties with war and intense conflict of other kinds can be below the mean on the conflict scale. There is a threshold beyond which degree of awareness may no longer discriminate in Conflict Behavior, and I assume that is in discriminating between intensity of wars and violence.
For this reason, and because the below-the-mean Conflict Behavior on the other conflict variables can be itself intense or diverse in scope (see Figure 16C.1), the resulting correlations are assumed random with regard to the proposition: they will depend on the particular sample's peculiarities.
Evidence: The evidence is usually positive (17 out of 25 analyses) and direct (all but three). Important studies split for and against, while those studies which do not violate the field assumptions (Model II and distance vectors) are usually favorable (9 to 2); and most of my own studies have also been favorable (8 to 4). Some of my results are given in Appendix 9A.
This is the first proposition for which negative evidence has appeared, and in their light the expected correlations should be reconsidered. How serious is a negative result here?
Important, but not critical. The reason is that geographic distance has usually been measured as air distance between capital cities, state boundaries, political boundaries (including colonies), contiguity or not, or as the number of political boundaries. In some of my studies I have also measured it as transaction distance (cost of shipping).
In all those measurements, geographic distance roughly reflects contact and awareness. One should expect a negative correlation; but it need not be high (as argued in the prediction Section, above), and it may be low.
Moreover, contact and awareness are not purely a matter of geographic distance, no matter how measured. For distance also depends on the resources of states to transcend distance: their joint power, as formulated in the next proposition. Therefore, the specific characteristics of cases included in a sample can also affect the relationship between geographic distance and Conflict Behavior.
Theoretically, correlations should still be negative in general, otherwise there would be no Geographic Proposition. However, the occurrence of some positive correlations is understandable.
Conclusion: The weight of the evidence in Table 16C.3 is positive and the qualitative nature of this evidence lends further support. Therefore, the conclusion: the evidence supports the proposition.
Joint power is positively correlated with Conflict Behavior.
Theory: Conflict Behavior assumes contact. True. But, it also assumes salience, a mutual importance of the states.
Saliance is an aspect of awareness. Not only are states in contact, their leaders mutually aware of each other, but they each have a mutual salience: a mutual importance of interests, capabilities, and will. This salience qualifies the impact of geographic distance, for states at a considerable distance may be more salient than those closer to home.
Salience is reflected along many dimensions of behavior: the degree of transactions, the existence and degree of diplomatic relations, prominence in the media, state visits and conferences, and the like. There is a more fundamental measure of salience, however, that relates to all these and should be positively correlated with conflict behavior. This is the joint power of two states.
Power in terms of combined national income, resources, area, population, and defense expenditures reflects the interest of national leaders in other states and the interest of other states in it (as a possible aggressor, if for no other reason), plus the ability of leaders to transcend geographic distance.
Power bases (or coercive power potential) is thus by itself a gradient of a nation's global salience, and should be related to Conflict Behavior. This is one source of the State Power Proposition 16.31, to be considered below. The mutual salience of two states is thus measured by their joint power: the power of one plus that of the other.
Conflict behavior assumes salience. It assumes that states have sufficient power to transcend their distances or are given such power by other parties (such as in the American aided South Korean military action against North Vietnam in the Vietnam war); or Soviet aided Cuban military action against Somalia in the Ethiopian-Somalia War.
As with geographic distance, salience is also a gradient. And the more salient two states, the more it can worsen a conflict between them and increase the intensity of Conflict Behavior. Therefore, salience can also be considered an aggravating cause, as shown in Table 16B.1.
Prediction: In my research I have measured power as either the factor scores on the power (bases) component, or by national income or population indicators. Joint power was measured then as the power of the actor plus that of the object state. Or, as this sum divided by geographic distance (sometimes squared). This latter operationalization also measures what is called "the social gravity" of two states.
Studies measuring joint power as any reasonable variant of the above types of measures (such as the power of one times that of the other) are considered relevant to the proposition.
The expected correlations shown in Table 16C.1 and their supporting arguments are the same as for the Geographic Proposition; except for joint power the correlations should be opposite to geographic distance.
Evidence: The evidence is mostly supportive (10 out of 12 analyses), and is mainly direct (6 out of 7). Only two important studies bear on it, however, but both are positive. Excluding studies inconsistent with Model 11 and the distance vector assumptions of field theory, five are for, none against. My own analyses have been positive, except for one ambiguous result.
In addition, I have found that joint power has been an empirically strong correlate of dyadic conflict behavior, sometimes appearing as the best predictor or strongest correlate in regression and factor analyses (on such use of factor analysis, see "Understanding Factor Analysis").
Conclusion: the evidence supports the proposition.
Conflict Behavior is most correlated with sociocultural differences and similarities (distance vectors).
Theory: The medium of the international field is the meanings, values, and norms of international actors. Seated in this medium are the potential field forces of international behavior: the sociocultural distance vectors between actors. This is the analytic level.
Substantively, actors have different and often opposing interests and capabilities. Interests provide the field's energy; they are the actor's need-connected goals and means. Stimulation of needs generates the motivational energy, and interests direct this energy.
All this has been developed in Part IV of Vol. 1: The Dynamic Psychological Field; Chapter 6 of Vol. 2: The Conflict Helix; and in Chapter 8. As pointed out, the axes of opposing interests and capabilities between actors lie along their distance vectors. Component distance vectors are therefore forces because they reflect interests and capabilities; the vector forces are seated in meanings, values, and norms because the component distances are on the sociocultural components of the international sociocultural space-time, such as wealth, power, totalitarianism, Catholic culture, and the others discussed in Chapter 7 (Table 7.1).
By theory conflict is a balancing of opposing interests and capabilities, and by theory this opposition should be reflected in the international space-time distances between actors. In other words, of all types of independent variables (distance magnitudes; internal variables; system variables; actor characteristics alone), distance vectors should best account for the variance in Conflict Behavior.
This does not mean that distance vectors are necessary and sufficient for Conflict Behavior. Distance vectors are necessary causes. They may reflect the structure of conflict and, in relation to perceived situation (see Table 10.1), the situation of conflict; but they still may be held in check by a congruent structure of expectations or by a lack of a trigger if the structure is incongruent with the associated balance of powers.
Distance vectors in their effect are analogous to a tight web of rubber bands, tied together and stretched between many sticks. The web is a balance of pulling forces along the taut rubber bands. The forces are there, but are latent until one band is cut. Then, there will be a sudden readjustment of the web in line with the pulling forces. Likewise, when an international structure of expectations is disrupted, conflict occurs in line with the distance vectors.
Prediction: The prediction is that distance vectors will best account for the variance in Conflict Behavior. In analyses using many different types of independent variables, distance vectors should be the most effective; in studies using only distance vectors, they should account for a significant (practical or statistical, depending on the analysis) amount of variance.
A high correlation may occur but cannot be generally expected: distance vectors are only necessary, not sufficient, for Conflict Behavior. Appropriate vectors may reflect a structure or situation of conflict, but for lack of other necessary causes, no conflict may occur.
When there is conflict behavior, however, it should occur in line with the distance vectors (opposing interests and capabilities).
Direction of correlations is irrelevant: This proposition only states the potency of distance vectors. Other propositions (e.g., Propositions 16.12, 16.14, 16.30, and 16.32) give direction in correlation and substantive interpretation to certain distance vectors.
The variance in distance vectors should be most correlated with the variance in each conflict variable listed in Table 16C.1, regardless of sample.
One source of confusion. In the analyses of a dyadic sample involving the same actor, variation in distance vectors is equivalent to variation in the object state's attributes. Therefore, analyses of object attributes provide evidence for the Distance Vector Proposition.
Evidence: On balance the evidence favors the proposition: overall there are 18 analyses for (with five strongly positive), six against. Most of the positive evidence is inferential, however, and when excluded, there remain six for and five against.
The important studies are generally positive (6 to 1); and excluding model I studies leaves a favorable balance of evidence (8 for to 3 against). Most of the favorable analyses are my own (12 out of 15), and so are half the negative analyses (3 out of 6).
Thus, 75% of the evidence supports the proposition and 21% is strongly positive; 25% is negative, 8% strongly so. Two out of eight important analyses are negative.
Moreover, there is a feel I have developed for distance vectors from my own research and that of others. Often, distance vectors added to an analysis make the difference between accounting for much or little variance in conflict behavior; between having highly significant regression coefficients or barely significant ones.
Conclusion: The proposition is consistent with most of the evidence, but there is still sufficient negative evidence to qualify the conclusion. Therefore: the evidence tends to support the proposition.
The potency of interests (sociocultural distance vectors) depends on an actor's individual perceptions and expectations.
Theory: This is the basic assumption of Model II of field theory: states are influenced by distances (opposing interests and capabilities) in the light of their own experiences, domestic systems, leadership, and so on. In short, perceptions and expectations will vary by state and differently weight distance and behavioral disposition vectors (see Chapter 10). The important role of actor dependent perceptions and expectations has been a theme of the previous volumes. Conflict promoting perceptions and expectations are necessary, but not sufficient, for conflict behavior, however. They can create a situation of conflict, but a structure of expectations based on a stable balance of powers may still maintain a stable peace. When conflict does occur, however, perceptions and expectations should be the parameters of its direction and intensity. This is detailed in Propositions 16.9, 16.28, and 16.29.
Now, I should be clear that I am asserting something more than that perceptions and expectations are an important necessary cause of Conflict Behavior. This is:
Prediction: The prediction closest to theory is that the trace correlations (canonical analysis) of a Model II fit of international behavior (including conflict behavior) will be significant (as shown in Appendix 9A), and more so than alternative models of international behavior. And this should hold regardless of sample or conflict variable.
Those analyses involving expectations or perceptions as variables across actors are also relevant, and should find a moderate correlation (because perception and expectation are necessary causes) between them and the conflict variables.
Evidence: With the exception of one ambiguity, the evidence is uniformly positive; the majority (11 out of 21) strongly positive. Indeed, strongly positive results predominate also for the direct and important studies as well. Moreover, most of this evidence is from analyses other than my own.
This evidence is congruent with my experience in testing Model II on a variety of data (see 1977:Chapter 16): it has generally accounted for about 50% of the variation (canonical trace correlation) in international behavior (including conflict), and first canonical correlations were often greater than .90.
Conclusion: the evidence strongly supports the proposition.
Conflict behavior will occur only if there is a will to conflict.
Theory: There can be no antagonistic conflict behavior without the will-to-conflict. There can be appeasement, acceptance, submission, abnegation, surrender.
Will is an explicit part of the process of conflict. It initiates the uncertainty phase, the point at which the decision is made to confront, or oppose another's interest, or pursue one's own in spite of opposition. And will (or credibility) is one of the three elements, along with interests and capabilities, involved in the balancing of powers; and in the balance underlying the structure of expectations.
While the will to conflict may be there, however, interests may be confused and capabilities insufficient. Will is not sufficient, therefore, but it is necessary.
Prediction: Strength of will should be positively correlated with the conflict behavior variable or conflict scale for the mixed sample shown in Table 16C.1. But, otherwise, because strength of will may vary across types of conflict behavior and along conflict behavior components (e.g., negative sanctions), and because much of this behavior can lie below the mean for the other conflict variables and samples, the resulting correlations are random.
A serious problem lies in measuring will. It cannot be measured by the occurrence of conflict behavior, for then the analysis would be tautological with regard to the proposition. But will could be measured by a complex of previous behavior showing a tendency, say, to withdraw or appease when confrontation occurs; by party platforms, public statements, official appointments, or expert ratings. In any case, its measurement must have been kept separate from the dependent variable to be relevant to the proposition.
Evidence: No evidence is available other than in Blainey's (1973) nonquantitative, historically "soft" systematic analysis of war and peace since 1700.
There may be three reasons for this. One is that will is considered too difficult to measure. I doubt this, however, because I do not recall the problem even being discussed and there has been a variety of attempts to measure other equally soft concepts, like perception.
A second reason may be that will is considered obviously a part of conflict, too self-evident to mention or require analysis. Perhaps. But I suspect that it is less a matter of obviousness than of perceived irrelevance. To many social behaviorists, will is a metaphysical concept; not quite respectable for "scientific" interest. Moreover, conflicts are often assumed to be occurrence, happenings, like highway accidents, or sickness, while win implies a conscious choice to conflict. A third and related reason may be that will is nondeterministic, implying freedom of choice. Most behaviorists are conscious or unconscious determinists who leave no room in their philosophy for a will as first cause.
In any case, whether because it is obvious, irrelevant, or rejected, I am aware of no systematic evidence on the role of will in conflict.
Conclusions: there is insufficient evidence.
Intense violence will occur only if there is an expectation of success.
Theory: The expectation of success is a necessary cause of intense violence, whether large scale military action or war. Escalation of violence, launching a war, or defense against such, is a subjective cost-gains calculation. It is rational in the sense that in the actor's calculus, violence will succeed in preserving, enhancing, or achieving some subjective stake.
This stake may be territory, independence, self-determination, a homeland, religion, national identity, an ideology, honor, credibility, domestic power. It may appear irrational to outsiders. But regardless of content, the actor believes or feels that the stake is worth the violence, and that violence will succeed.
Violence is a threshold in conflict behavior. It indicates an intense opposition of interests; a struggle over the status quo. Negative sanctions, accusations, protest, threats, and the like, can occur between states for whom mutual military violence is presently inconceivable, such as the United States, Canada, Great Britain, Australia, and New Zealand. This threshold is crossed when border clashes, discrete military actions (shooting of a border guard, downing a surveillance plane, capturing an intelligence ship) occur. But the violence and costs are minor: few if any deaths occur, and the violence can often be shrugged off officially as caused by mistaken identity, local commanders, or hotheads.
Such actions may be probes to assess the success of more direct or intense action, they may be limited actions of low risk and possible high gain, or they may be spontaneous local engagements. Nonviolent conflict behavior can also be probes, sanctions, retortions, communications, and the like. The status quo may not even be involved.
However, when intense violence is selected, when large military formations are engaged for more than a day and the killed range upward from the hundreds, a second threshold has been crossed. Now, the costs can be severe, major national values can be lost, a leader's head may be on the tine, and only major stakes and a belief in success can justify the danger.
Overt conflict begins in a situation of uncertainty. But by the time war is chosen, at least the uncertainty of success is dispelled. For both major parties.
Prediction: Expectations of success may or may not be present in nonviolent conflict or low-level violence, while always present in high-level violence. Therefore, it should be correlated positively with the intensity of violence and conflict scale for the mixed, conflict, and violence samples.
For the frequency of war variable, for both the mixed conflict samples, however, war and intense violence will appear below the mean (see Figure 16C.1) and can thus cause the correlation to range from a near zero to a moderate plus.
For the war sample, expectation of success should be present in all cases.
Evidence: The evidence amounts to three analyses, two positive and one strongly positive. One is Blainey's (1973) analysis, a second is Wright's analysis of escalation (1965), and third was my own (Project 6) systematic analysis of Wright's data, augmented by attribute and distance data. All are important studies; two of the three are direct.
Conclusion: Although the number of relevant analyses is small, they are consistent. Thus the conclusion: the evidence unanimously supports the proposition.
Violence is over a status quo.
Theory: The status quo is the division in rights and duties, what is ours and theirs. It is the hard core of any structure of expectations, usually well-defined, and the major values in any interaction.
In international relations the relevant status quo is defined by who has sovereignty over, control over, or the actual or implied allegiance of the people living on a territory. A conflict may be over a border area (e.g., Vietnam versus Cambodia in 1977-1978), the independence of a colony (e.g., the Algerian War, 1954-1962), the independence of a state (e.g., the Korean War, 1950-1953), and the control of a state (e.g., the Soviet invasion of Czechoslovakia in 1968). Moreover, a territoriality defined status quo is often the source of conflict, as over the rights of Greek versus Turkish Cypriots.
Ours and theirs may refer to the allegiance or ideology of territorially defined peoples. We "lost" Cuba to the Soviet Union; they are "gaining" Angola. The "free world" now consists of ... ; NATO will be "weakened" if Italy has a communist government. And so on.
The disruption of the status quo may not cause violence. A new status quo may be determined nonviolently, as in the Panama Canal Treaty negotiated by the United States and Panama (although there were anti-American riots and demonstrations and potential violence was an argument used by the Carter Administration in seeking Senate ratification, the settlement was achieved nonviolently), and the granting of independence to many former colonies.
However, the status quo defines the high stakes, and only the status quo is worth violence.
What comprises nonstatus quo expectations? This is the whole complex of implicit and formal understandings, agreements, and treaties governing the relationship and interaction between parties, which do not define who has what rights and duties over what territory. Trade agreements, counselor understandings, disarmament treaties, diplomatic norms, and the like, are the nonstatus quo part of our international structure of expectations.
Prediction: A disrupted status quo is a necessary condition of violence. That is, all violence should be over some territorial sovereignty, control, allegiance, or territorial-based rights (as over ethnic group or national group rights in a territory).
For the nonconflict behavior sample, nonconflict behavior implies that a disrupted status quo is absent, as shown in Table 16C.1.
For the mixed sample and conflict variable, no disrupted status quo should appear below the mean, while above the mean cases can also have nondisrupted status quos. The correlation can therefore range from near zero to moderate positive.
For nonviolent conflict, however, status quo disruption can appear above and below the mean in frequencies depending upon the sample's specific characteristics. This is because nonviolent conflict may or may not reflect a disrupted status quo, while below the mean occur violent cases with disrupted status quo and nonconflict cases without such disruption.
For other conflict variables, there should be a moderate correlation.
The logic is the same for the conflict variables in the conflict sample, while for the violence and war samples disrupted expectations should be present in each case.
Evidence: All ten sources of evidence support the proposition, four strongly and directly so. Moreover, most of the evidence are from important studies, are consistent with field theory, and all are from studies other than my own.
Conclusion: the evidence unanimously supports the proposition.
Violence will occur between states only if at least one is nonlibertarian (nonliberal democratic).
Theory: Totalitarian and authoritarian states are the sources of military violence. Power is centralized, and in the hands of a small group or leader. Control over the communication media can be used to direct and intensify public opinion against another state: domestic interest groups can be controlled; and dissidents in the elite can be jailed.
In open systems, however, initiating a war or military action is usually precluded by the restraint of public opinion and opposition of interest groups, unless there is a clear and present danger to the state or its security by military aggression or a unilateral change in the status quo (e.g., Egypt's takeover of the Suez Canal in 1956). Then public opinion and the elite unify behind defense, or at least the party in power will support its leadership in the near term.
However, between libertarian states there is a fundamental sympathy of their peoples toward each other's system, a compatibility of basic values, an existence of cross-pressures and overlapping groups and organizations, and a suffusion of power and interests.
Between exchange societies and their libertarian political systems exist the same conditions minimizing violence (Proposition 16.24) that operate within societies.
Therefore, a necessary condition of violence between two states is that at least one be totalitarian or authoritarian.
Prediction: The prediction is narrowly specific: only jointly libertarian states will be without violence. This makes assessing correlation especially difficult. For simplicity, I will call two jointly libertarian states a libertarian dyad. Nonlibertarian dyads are any pair of states, at least one of which is nonlibertarian in the sense developed in Chapter 31 of Vol. 2: The Conflict Helix.
For the nonconflict sample, both libertarian and nonlibertarian dyads can be without conflict.
For the mixed sample and conflict variable, both libertarian and nonlibertarian dyads can appear randomly above and below its mean, because both types of dyads may or may not have conflict behavior. Therefore, a random correlation is expected. Similarly for the nonviolent conflict variable.
However, for the frequency of violence or war variables, libertarian dyads should be at the bottom of the scale and the correlation should therefore be negative. At this point, what independent variables are correlated with the conflict variable should be carefully considered.
There are three types of potentially relevant independent variables. The closest to the proposition would be a dichotomous variable separating libertarian dyads from nonlibertarian. Correlation between this independent variable and the frequency of violence or war for the mixed sample should, by theory, be moderately negative. The reason that a perfect negative correlation cannot be assumed is that the frequency of violence or war variables would have violence also occurring below their means (see Figure 16C.1), and nonlibertarian dyads also may be without violence.
If the independent variable is the absolute political distance between actor and object state, the correlation could be random regarding the Joint Freedom Proposition, because this measurement does not distinguish between libertarian and nonlibertarian dyads, and the proposition does not imply that similar political systems (e.g., two totalitarian states) should be involved in violence more or less than dissimilar ones (e.g., totalitarian and authoritarian or authoritarian and libertarian states).
If the independent variable is a vector distance, the same point applies, except when all dyads have the same actor (Model II). Then, in effect the distance vector is measuring the variation in the object's political system. And when the actor is libertarian, then this variable should be correlated moderately and negatively with the violence in the dyad.
Keeping in mind, then, that the independent variable is a dichotomy or vector distance in Model II for libertarian actors, we should also find that the violence intensity and conflict scale variables are moderately and negatively correlated.
The same logic applies to the conflict sample. For the violence and war samples, however, there should be no cases of libertarian dyads having violence or war, and therefore, these dyads should be absent from the samples.
A final consideration. Violence here refers to low or high-level violence. In a crises or war situation involving other nations a discrete military action or clash may occur between libertarian states, such as the Israel attack on a U.S. surveillance ship during the 1973 Yom Kippur War. Usually the attack will not be given official sanction and apologies will be made. Such events should be rare and ought not to be given significant weight in assessing the proposition.
Evidence: Nine analyses (three strongly) support the proposition; three are negative. Highly significant is that no important analysis is negative, and two of the three negative results were indirect-inferences I made from rather complex statistical manipulations involving samples and variables which, themselves, only indirectly measure what is of concern here. This is, of course, true for the indirect positive evidence as well. However, the importance of this point is that in contrast to these weak negative results some of the direct positive results are clear and absolute, especially regarding wars where with one ambiguous and explicable exception (the War of 1812 between the United States and Great Britain), wars simply have not occurred between libertarian or partially libertarian systems (see Project 48; Babst, n.d., 1972).
Conclusion: the evidence supports the proposition.
Sociocultural dissimilarity is positively correlated with Conflict Behavior.
Theory: The more distant states are in the international space-time, the more likely their interests will conflict and the more difficult communication and negotiation will be (see Chapter 6 of Vol. 3: Conflict In Perspective).
Such absolute distances are not sufficient or necessary for conflict, however. Nations close in space-time may have territorial, ideological, or theological disputes (as between Catholic states in Europe during the Middle Ages or communist states today); moreover, states quite distant may have little contact and salience. But given the occurrence of the conditions embodied in Proposition 16.1 to Proposition 16.11, then sociocultural distance aggravates the conflict, making resolution more difficult.
Prediction: Dissimilarity (as absolute sociocultural distance) should be positively and moderately correlated with the conflict behavior variable in the mixed sample.
However, for nonviolence, the frequency of violence, or frequency of war, and considering the other conditions operating, dissimilarity alone would appear not to have any uniform correlation. For the violence variables this is in effect saying that while conflict is acerbated, there is no necessary relationship between the degree of dissimilarity and the intensity of overall conflict as embodied in a conflict scale.
If this is confusing, consider again Figure 16C.1. While more intense conflict in general will appear above the mean for the conflict variables, for the violence and war frequency variables, cases with intense nonviolent conflict, or intense war (for the violence variable), or intense low-level conflict, can appear far below the mean (possibly at zero).
The predicted correlations for the other samples follow the same logic.
Evidence: All the relevant analyses bear directly on the proposition and ten are supportive versus four against. Of the important studies the balance is almost even, four to three (but two of the three are strongly negative).
Conclusion: the evidence supports the proposition.
Cognitive dissonance is positively correlated with conflict behavior.
Theory: I have considered cognitive dissonance, or imbalance, theoretically in Chapter 12 of Vol. 1: The Dynamic Psychological Field; and particularly with regard to status and behavior in Section 18.2 of Chapter 18 in Vol. 2: The Conflict Helix. In short, there are two aspects of cognitive dissonance relevant to national leaders.
First, leaders will tend to adjust their policies to maintain a psychological balance in their perceived relations with other states. If, for example, leaders perceive an ally is also allied to an hostile state, then they will tend to readjust their policies accordingly. But such readjustment may provoke conflict, for it will involve different behavior and a new structure of expectations.
Second, leaders will tend to see their state in terms of its highest status and other states in terms of their lowest. Because leaders of different states are then emphasizing different components of their relationship, this makes communication more difficult, conflict more likely, and settlement more difficult. Therefore, cognitive dissonance aggravates conflict.
Prediction: Cognitive dissonance can be operationalized as either status incongruence (Vol. 2: The Conflict Helix, 1976: 142) between states or a triadic imbalance in relations (dissonance: Vol. 1: The Dynamic Psychological Field, 1975: 124-128). So measured, cognitive dissonance win be positively correlated with the conflict behavior variable and with the conflict scale for the mixed, conflict, and violence samples, as shown in Table 16C.1. For the other variables,. I see no theoretical reason for cognitive dissonance being correlated with one type of conflict intensity versus another, and therefore expect correlations to vary from one study to another.
Evidence: The evidence is almost uniformly supportive (ten analyses for to one against), but usually indirect. Six out of seven of the analyses were from important studies.
The one negative study was also direct. It involved a time-series analysis (Mogdis, 1969, 1970) of Sino-Soviet interaction and perception and analyzed the effect on their behavior of a cognitive imbalance in the triangular Soviet-China-United States relationship.
Conclusion: Although the evidence is almost unanimous, it is mainly inferential and there is one important negative study. Nonetheless, I can say simply that: the evidence supports the proposition.
Differences in status (wealth, power, prestige) are positively correlated with Conflict Behavior.
Theory: Status vectors are forces toward conflict. The theoretical justification for this was the burden of Chapter 17 and Chapter 18 in Vol. 2: The Conflict Helix.
But these forces are neither sufficient nor necessary. They act within the context of the other forces (causes) and conditions given in Table 16A.1. They are forces which worsen conflict, making conflict behavior more likely and more intense.
Status is defined by wealth, power, and prestige. These are esteemed, desired components. Power, however, has also particular importance in its own right.
Absolute difference in (coercive) power, or the degree of power parity, indicates the likelihood of escalation in war (Proposition 16.21). The centralization of (coercive) power stimulates extreme violence (Proposition 16.20) and inhibits low-level conflict (Proposition 16.23). Together, these propositions imply that of all attributes, power is the most correlated with conflict (Proposition 16.31), and that power is the most potent force toward conflict (Proposition 16.32).
Here, however, the interest is in status distance. Parity or polarity win be considered irrelevant (but not power distance vectors) for this proposition.
Prediction: Status distance (vector) in some sense should be positively correlated with all the conflict variables in the mixed and conflict samples, except for the frequency of low-level conflict.
For this variable other kinds of intense conflict will occur below the mean and this should not generally be compensated for by a high relationship of status distance to the intensity of low-level violence. Status is too gross a force to make a great difference in whether states have, say, few or many border clashes. Therefore, the expected correlation is random.
For the violence and war samples, much of the variance in the effect of status on nonviolent conflict behavior has been eliminated. Therefore, I expect that the correlation win also be random for this variable in those samples.
Evidence: Status has been a focus of many studies, and thus relatively much evidence exists. And virtually all analyses (16 to 1) support the proposition. Most evidence is strongly positive, all of which is direct. Even for important studies, the evidence is mainly positive, most strongly so.
Conclusion: the evidence strongly supports the proposition,
Cross-pressures are positively correlated with nonviolent Conflict Behavior.
Theory: The theoretical function of cross-pressures (and crosscutting) is presented in Vol. 2: The Conflict Helix (1976: passim). There are two aspects to consider: the impact on nonviolent conflict, and on violence. Cross-pressures act to increase one and reduce the other. The former impact is considered here; the latter is embodied in the Cross-Pressures/ Violence Proposition (16.24).
Simply put, diverse, overlapping meanings, values and attitudes in society create a variety of opposing interests and the necessity to achieve many different, associated structures of expectations. There will thus be many small scale conflicts--a hubbub of Conflict Behavior.
For example, consider a society-wide market in which prices are determined by free competition. There will be a variety of cross-pressures operating in the market and many associated conflicts between buyers and sellers, management and labor, and producer and consumer groups, not to mention the competition between businesses. A free market is a field of such conflicts.
In an authoritarian or coercive system, however, issues are generally decided by tradition and authority or by coercion (as in socialism). The function of diverse, competing and opposing interests and capabilities--of cross-pressures--is therefore subdued. People are harmonized by class, ethnic group, region, religion, or coercion.
At the international level, the more cross-pressures can operate in the international system, the more nonviolent conflict will be present. Highly polar systems (coercive systems) or those divided into a few alliances or religious groupings (authoritative systems) should have relatively less nonviolent Conflict Behavior overall.
Between states (dyadically) cross-pressures should operate similarly. The more cross-pressures, the more confused interstate relations and the more likely nonviolent conflict as multiple balancing is required. But the very existence of these cross-pressures also inhibits the formation of a single, dominant dispute along which interests could polarize and cause intense conflict and extreme violence.
See also the theoretical discussion for Proposition 16.20, Proposition 16.23, and Proposition 16.24.
Prediction: The expected correlations for each sample must take into account tile complementary Cross-Pressure /Violence Proposition 16.24 and is discussed in general terms for that proposition. The expected correlations win be the same for both propositions.
Evidence: Tile only evidence I could find was in the Vasquez (1975, 1976) sifting of quantitative results, which was negative.
Much of the explanation for this lack of evidence is that research on or involving cross-pressures is almost always on violence, and that it is not widely appreciated that cross-pressures have a dual role in inhibiting violence and aggravating nonviolent conflict.
Conclusion: The one negative result, especially because it is from a summary, suggests the conclusion: insufficient evidence.
Involvement of a state's honor is positively correlated with violence.
Theory: At the core of interests lies the self-sentiment, our self-image. We act to enhance and develop self-esteem. One of the sources of power for state leadership is that citizens generally link their self-esteem to their nation. Its gains are their gains; its losses they suffer, psychologically. In this sense, nations are like hometown teams. We identify with them, cheer them on to victory and are ecstatic when they win; depressed, sometimes violent, when they lose. That about a state which is linked to the self-esteem of its people defines its honor.
Because self-esteem is the superordinate goal of every one (Section 21.5 of Chapter 21 in Vol. 1: The Dynamic Psychological Field) honor is a central variable in a state's relations with others and in the genesis of violence. It is an aggravating cause.
A festering humiliation over past defeat may aggravate a dispute or stimulate another war, as in the Egyptian-Syrian launched attack on Israel in 1973. Outrage over a surprise attack may crystallize war sentiment and make negotiated peace impossible, as in Japan's attack on Pearl Harbor in 1941. A people may feel insulted by another state, their sense of dignity at stake.
States may surely have material interests in a conflict, power and credibility may be at stake. But the strength and nature of a conflict will not be well understood unless the engagement of honor (self-esteem) is also considered.
Two possible sources of misunderstanding should be clarified. First, honor is an interest, as discussed in Section 21.5 of Chapter 21 in Vol. 1: The Dynamic Psychological Field in terms of self-esteem. The reason I emphasize the term honor here is because of its general use and understanding in international relations. Therefore, honor is not simply an emotion (although pride or a feeling of injustice will be involved) to be opposed to national interests. Honor is a national interest of high order.
Second, honor is sometimes confused with credibility. Clearly, there is an overlap in meaning. A nation's reputation for justice, fulfillment of promises, and discharge of military commitments is an aspect of its honor. But honor as self-esteem and credibility as will to meet promises or threats are distinct. More on credibility below, regarding the next proposition.
Prediction: Honor operates at the higher levels of conflict, specifically to aggravate violence. It is generally irrelevant to nonviolent conflict in all samples. Also, it should have no general correlation one way or another with the frequency of low-level violence, simply because honor may be highly correlated with the violence that appears below the mean for the variable (see Figure 16C.1).
The other conflict variables, however, should be positively correlated with honor being at stake.
Evidence: The only evidence (which is direct, important, and strongly positive) is the nonquantitative, semisystematic analysis of disputes since 1945 by Northedge and Donelan (1971). Honor has simply been ignored by quantitative researchers.
Conclusion: Insufficient evidence.
Involvement of a state's reputation for power (will) is positively correlated with violence.
Theory: The more involved in a conflict is a state's credibility for defending its interests or meeting its commitments to defend others, then the more intense and prolonged the conflict may become.
Now, the status quo is certainly involved in violence (Proposition 16.10). That is what it is being fought over. However, a state's reputation for power, the image a state projects of a will to protect its interests and follow through on threats is also at stake. For leaders realize that what a state does in a current conflict is recorded on an international ledger to be consulted by others interested in pushing their own status quo interests. To show weakness of will, like the boy in the schoolyard who submits, crying, to the taunts and slaps of a bully, is to invite a broad assault by many on one's interests.
State's fight not only to win a current conflict, but to avoid or reduce the costs of the next. For example, the credibility stake became a major variable explaining U.S. military involvement in Vietnam under President Johnson.
Prediction: The predictions for the involvement of credibility are the same as for honor. Both operate similarly and at the same conflict levels.
Evidence: There are two sources of evidence. One is direct and important, and is the Northedge and Donelan (1971) nonquantitative analysis. It strongly supports the propositions.
The other source is also direct and important, and is Barringer's (1972) rigorously systematic work. His results support the proposition and show that the perception of some public commitment at stake is one of the variables discriminating between hostilities and lower-level conflict.
Conclusion: Because there are two important analyses, both supporting the proposition, the conclusion: the evidence supports the proposition.
An actual or growing weakness of the status quo party compared to the antistatus quo party is positively correlated with violence.
Theory: A structure of expectations has at its core the status quo which defines rights: what is ours versus theirs.
Over the status quo may grow a latent conflict between that state which benefits most from the status quo and believes it is just or right, and one who opposes it as unjust or undesirable. However, the antistatus quo party will continue to acquiesce in the status quo as long as it feels attempts to change it will be unsuccessful or too costly. In other words, as long as the original balance of powers undergirding the status quo does not shift to favor the antistatus quo party, the structure of expectations is stable.
However, if the status quo party significantly weakens in will, interests, or capability for coercive power, then the balance in powers will shift. And a disruption of the status quo will likely occur as the antistatus quo party is triggered to pursue a more favorable status quo.
This does not make the weakness of the status quo party a sufficient condition for violence, however. For nonviolent accommodations may take place; both parties may be libertarian (which would preclude violence, as stated in Proposition 16.11); in spite of a disrupted status quo and extreme nonviolent conflict, the antistatus quo party may be deterred from violence by an uncertainty over success or the costs (as in the deterrent effect of nuclear weapons on both the United States and USSR behavior during the 1962 Cuban Missile Crisis).
Nonetheless, the weakness of the status quo party is a potent aggravator of conflict, making violence more likely and more intense. This because the issues deal with territory, core rights, and perceived justice.
Prediction: Operationally, there are three primary variables involved in a balance of powers: interests, will or credibility, and capabilities. And the overall strength of the status quo party is a multiple of these variables. Thus, a status quo party can become weakened by a shift in one, two, or all three variables. Thus, to lose will through the election of a pacifist leader would create weakness; to reduce the salience of a specific status quo interest (such as American loss of interest in defending freedom elsewhere) would create weakness; and, as everyone knows, to reduce military power would create weakness.
What should also be clear is that weakness is comparative: the status quo party could actually remain as strong or be increasing in strength. But if the antistatus quo power increases power (interests, will, capabilities) faster, then a weakness of die status quo party would still be produced; a dangerous gap between the status quo and the supporting balance of powers would occur.
Weakness is an aggravating cause of violence, as are die involvement of honor and credibility.
Thus, the expected correlations shown in Table 16C.1 follow the same logic.
Evidence: Six different analyses bear on the proposition as tabulated in Table 16C.3, four of them are important and five direct. All six support the propositions; three do so strongly. And all these analyses are by others than myself.
Conclusion: the evidence unanimously supports the proposition.
Big Power intervention in a conflict is positively correlated with violence.
Theory: Local conflicts are a balancing of power among non-Big Power states contiguously or regionally involved. A Big Power then intervening to aid one side will likely inject into the local conflict the question of the status quo between the Big Power and its status quo adversary.
Certainly, what a Big Power gains is not necessarily the loss of another. But when Big Powers intervene they raise the stakes involved in the local conflict. The outcome will determine what is within the Big Power's territorial or ideological sphere of influence and support. Therefore, such intervention usually does not go unanswered: The Big Power's adversary will counter-intervene covertly or directly. A local conflict then becomes a balancing of the status quo between Big Powers.
Even were another Big Power not to counterintervene, intervention is usually on behalf of the losing side and will forestall its defeat and encourage an escalation of the local conflict.
Simply, intervention aggravates local conflict, risking an escalation in violence.
Prediction: Big Power intervention is an aggravating cause of violence. As a variable, its expected correlations are the same as for honor and credibility (Proposition 16.16 and Proposition 16.17) and the justification is similar.
Evidence: Two direct analyses strongly support the proposition. One is important and is Barringer's (1972) systematic analysis of variables discriminating the phases of conflict. The other is Kende's (n.d., 1971) statistical survey of 97 local wars, 1945-1969.
Conclusion: the evidence supports the proportion.
Polarity is positively correlated with intense violence.
Theory: Polarity refers to the centralization of coercive power in a system. It is a state (as in "state of the system") variable, without application at the dyadic or nation-state level.
Polarity as a variable does not distinguish the type of international societies involved. Unipolar systems may be authoritarian or coercive, but are in any case nonlibertarian. Bipolar (two major competing centers of power) or multipolar (more than two competing centers of power) systems may be a mixture of societies. Each pole maybe one type (as the American alliance since 1950 has been largely libertarian, while the Warsaw Pact has been coercive) and the overall system can be libertarian, as it is now (Chapter 2).
Polarity is a distinct concept from cross-pressures. The latter may operate at the dyadic or nation-state level, and describe the state of the system as well. It means simply that for an individual state the existence of diverse, contending interests will usually work against any one being the basis of all out conflict. Overlapping and pluralistic groups and relationships define this essentially psychological variable.
Theoretically, however, there is a complementary relationship between polarity and cross-pressures. As a system is centralized (from multipolar to bipolar to unipolar), there is a decrease in the diversity of independent groups and relationships. Interests become aligned towards the poles; issues and disputes increasingly become part of the status quo--a class division--between the poles. If the system is unipolar conflicts still occur along the front between classes (the status quo "ins" and the antistatus quo "outs").
However, while increasing polarity decreases cross-pressures, decreasing polarity may not be followed by increasing cross-pressures (it may be a traditional, authoritative social system with a culture aligning interests along, say, religious lines as in Islamic society). Moreover, an increase in cross-pressures does not mean that polarity is decreasing. The increase may be due to increased economic and technological development.
Polarity and cross-pressures, in other words, are theoretically and conceptually distinct and need not be highly correlated empirically.
Nonetheless, they have opposite effects. Cross-pressures aggravate nonviolent Conflict Behavior (Proposition 16.15) while inhibiting violence (Proposition 16.24); polarity inhibits nonviolent Conflict Behavior (Proposition 16.23), while aggravating intense violence (Proposition 16.20).
Now to focus on the proposition of concern here. Polarity reduces cross-pressures and links otherwise ordinary conflicts into the more general status quo. Conflicts become a question of class power. Increasing polarity generalizes class conflict (in my terms, as developed in Chapter 24 and Chapter 25 in Vol. 2: The Conflict Helix) to the system and increases the salience and scope of this division to each state's interests (see also Proposition 16.22).
Therefore, more becomes at stake in a conflict, the issues are more burning, and the potential intensity and scope of violence is increased. That is, polarity aggravates violence. It makes intense war more likely. It makes a general war more probable. See also the theoretical discussion for Proposition 16.23.
Prediction: The expected correlations for each sample must take into account the complementary Polarity /Conflict Proposition 16.23 and is discussed in general terms for that proposition. The expected correlations will be the same for both propositions.
Evidence: There are three direct, strongly positive, and important analyses; one direct, positive, and important, as shown in Table 16C.4. But there are also two direct, important negative analyses, and three ambiguous ones. And all studies are consistent with the field theory Model II.
Why the negative evidence and unusual ambiguity? A major problem lies in the measurement of polarity. As a theoretical system variable, polarity varies with the centralization of the system, as shown in Figure 16B.1. At the point a system becomes unipolar, there may still be considerable dispersion of power. Consider that under the Articles of Confederation the American states formed a unipolar system during the Revolutionary War and under the Constitution of the United States which followed in 1787, the federation was much more centralized. Then compare this initial unipolar system with the centralized dominance of the American federal government today. Then compare this current unipolar system, which still has considerable libertarian dispersion of power, with the extreme centralization (polarity) of theocratic, authoritative Saudi Arabia or the totalitarian Soviet Union. Yet, all these systems (including the U.S. Confederation) would be unipolar.
Similarly counting the Roman Empire, the ancient Chinese Empire, the Persian Empire, the Holy Roman Empire, the German Confederation, the Napoleonic Empire, the Soviet Empire, and the like as unipolar is to collapse important variance related to conflict and violence.
Unfortunately, most quantitative studies measure polarity by a count of poles of power, and this may well account for the ambiguous and negative results. For when polarity is considered within states and is measured on a scale capturing the relevant variance, the relationship between polarity and violence (antiregime and that of a regime against the people) is positive: authoritarian and totalitarian states manifest more violence (Chapter 35 of Vol. 2: The Conflict Helix).
Conclusion: On balance the evidence supports the proposition, but the three ambiguous and two negative analyses, which are direct and important, suggest caution. Therefore, the conclusion: the evidence tends to support the proposition.
Power parity is correlated with war.
Theory: That power is related to war is one of the central beliefs of international relations. But on the precise nature of this relationship there is much dispute. Some argue that relative power parity stabilizes relations and is a force for peace. Others argue that power parity is a cause of war. The proposition argues the latter position: power parity is an aggravating cause of war.
A conflict that has reached the stage of violence (but not yet war) is over deeply held, status quo interests (Proposition 16.10). Already costs have been incurred, the stakes are high, and further success or losses depend on the balance of coercive power or force. If one side is clearly superior in will and existing military capability and potential, then the other side is likely to avoid escalation, if necessary by negotiating a resolution to the conflict. Of course, if the stakes warrant it, the stronger side is likely to push its war threat chip forward to pressure the other, and in the expectation of success if called.
Power parity is an ambiguous situation of coercive power in which both sides simultaneously can believe that war will be victorious. War clears up this ambiguity.
Note that power here is not simply existing military capability. It is also military potential, which reflects economic power and probable third-country support. And it is strength of interests and will. Outweighed on paper by the combined Arab military forces, Israel nonetheless balanced the equation to her favor by a sharply focused and powerful desire for national and religious survival; and a national will rooted in Jewish history.
Note also that parity tends to convert low-level violence to war, but in theory is systematically unrelated to nonviolent Conflict Behavior or low-level violence. War involves the core national interests, possibly survival. Relative power becomes critical. Other forms of conflict are not so central and indeed may be peripheral to major national interests. Weak states can often pursue conflict with a much stronger one without worrying about the other's full power being brought to bear. The issue may not be that important for the stronger or it may be restrained by concern over provoking the involvement of a Big Power.
Finally, parity is an absolute distance variable.
Prediction: Parity should be positively correlated with the intensity of war: the more ambiguous the relative power, the more both parties can believe in success (Proposition 16.9); and the more ambiguous the relative power, the more intense tends to be the war. At that point when relative power becomes clear to both, when one of the sides who thought they would win now realize that defeat is a near certainty, it is the beginning of the end. Third-party conciliation and negotiation may now be fruitful.
As a cause of war, power parity then is irrelevant for all the nonviolent and low-level violence variables and the expected correlation is random in all samples, as shown in Table 16C.1. This is true also for the war frequency variable, which may be surprising. But consider. For the mixed, conflict, and violence samples, the zero point on the war frequency scale (as shown in Figure 16C.1) may involve cases with high or low parity (because parity is theoretically unrelated to conflict, except for war), and some cases of war below the mean may be intense conflicts which by theory should be highly related to parity. Each war is counted as one in the war frequency variable, thus equating such conflicts as the 1977-1978 Cambodian-Vietnam war, Vietnam war, 1973 Yom Kippur War, and World War II. The combination of nonwar cases and possible intense wars beneath the mean implies that in general the correlation will be random.
However, only war is involved in the war sample and although intense wars still may appear below the mean, the higher frequencies above the mean with the likelihood of their involving more of the intense wars, make probable a positive correlation with parity.
For the violence intensity and conflict scale in all samples, a positive correlation is predicted because of the increasing weight of intensity above the mean.
Evidence: As totaled in Table 16C.4, a large variety of analyses--37 to be exact--have been done involving parity and war. Of these one-fourth are negative (only one is strongly o), all but two of which are unimportant studies that I have done.
My studies have usually involved factor scores for military violence, with parity measured as absolute differences in national income or population. Moreover, the analyses were usually done for one year, especially for 1955, during which no intense war occurred in the sample. Obviously, the sample and measures are gross regarding the proposition and not too much should be made of the negative findings.
On the other side, 65% of the evidence is positive; 30% is strongly positive. Virtually all this evidence is direct, and most from important studies. Indeed, for important studies, which are usually those which have analyzed wars over a long historical period, 78% of the evidence is positive, most of which is strongly positive.
Conclusion: the evidence supports the proposition.
Class conflict is correlated with general war.
Theory: A class is a latent or manifest group that stands in either a consistent superordinate or subordinate position regarding the status quo--in a command or obey position regarding the law norms of a society, as defined and developed in Chapter 24 and Chapter 25 in Vol. 2: The Conflict Helix.
Historically, international relations has been a sequence of orders (general structures of expectations) based on a balance of powers. Each order is a particular distribution of rights, of a status quo, distributing command and obey status. The command is usually an explicit: "These are my territory, my people, my sphere of influence, my privileges. Keep off and beware!" In each order a particular state or group of states most benefit and in effect form the superordinate, upper class. These are the status quo powers.
As the same distribution of rights and privileges covers more international relations states, as the same states in more and more relationships are similarly up or down with respect to each other, an international class division is increased and enhanced.
The division of an order into classes is an aggravating cause of general war. Small wars escalate; more states are drawn in. This class division is a political fault line between the major international forces, and once the fault begins to slip at one place, pressures will be exerted to cause it to shift all along the line.
A general war is one involving the Big Powers, and a number of lesser states as well. The war is nearly unlimited in territory and weapons. Examples of general wars are the Thirty Years War, the Napoleonic Wars, World War I, and World War II.
Theoretically, class formation is closely related to polarity and cross-pressures (Propositions 16.20 and 16.24). Increasing polarity--the centralization of coercive power--increasingly divides society between those who command and those who obey. And increasing polarity decreases cross-pressures in the system, making the class division a societal wide front between opposing interests.
Prediction: The postulated relationship is of class formation with general war, a subcategory of intense wars (a very intense war may not be general, but limited to small states, such as the 1973 Yom Kippur War), where intense wars form a subcategory of Conflict Behavior. Much caution is indicated in predicting empirical correlations.
Table 16C.1 includes no general war variable, for the simple reason that analyses usually focus on war frequency or intensity. Singer and Small (1974) do consider the number of nations involved in a war, which does loosely measure generality, but this and other work by Singer and Small are an exception.
Therefore, simply given the variables included in Table 16C.1, what are the predictions? Assuming that absolute intensity of conflict behavior will itself have a correlation with general war (as an unnormed variable, e.g., number of killed is not divided by the populations involved), then class will operate correlationally like power parity. The expected correlations will be the same in both cases, except that they should be weaker for class, because we are dealing with a subcategory of intense wars.
Evidence: Only three different analyses (Denton, 1969; Kende, n.d., 1971; Barringer, 1972) are relevant. Two are important, one is direct, all are positive.
Conclusion: the evidence supports the proposition.
Polarity is negatively correlated with nonviolent conflict behavior and low-level violence.
Theory: Polarity has already been discussed with regard to the Polarity /Violence Proposition 16.20, which states that polarity aggravates intense violence by creating a system-wide schism between opposing interests. The Polarity/Conflict Proposition, however, points out that polarity inhibits nonviolent and low-level behavior.
The centralization of power into a few poles, or into a unipolar system, suppresses the diversity of interests that lead to frequent nonviolent conflicts and occasional low-level violence. The Big Powers which form the poles, as have the United States and Soviet Union since World War II, try to avoid conflict behavior among allies and satellites, and especially conflict behavior between members of the opposing blocs which could escalate into a confrontation between the bloc leaders. Polarity creates continents of relative surface harmony.
Prediction: In predicting the correlations for polarity and conflict behavior there are three theoretical considerations. These are brought together in Figure 16B.2. The curves are theoretically standardized, which means that each is hypothetically plotted as a variation from its own mean.
First, polarity increases the likelihood of extreme violence (Proposition 16.20). This is violence measured, say, as the total killed in the system from war, other military violence; and in increasing importance as polarity increases, a regime's violence against individuals and groups under its control. Thus, the theoretical, standardized curve for violence in Figure 16B.2 is at its highest for unipolar systems and decreases to approach some minimum as multipolarity increases.
Second, a component of violence is war, especially intense wars like World Wars I and II. Intense wars are increasingly likely as there are fewer poles to the point of a bipolar system of rough power parity. A unipolar system is less likely to have an intense war due to domination by one center of power. This war intensity curve is similar to the empirical
curve found for domestic political systems, where group violence is greatest at the middle range of political coerciveness.
And third, another component of violence is the number of wars. Now, war constitutes violence which to the participants may be quite intense. But in system terms, the war may be minor. For example, a war between two tribes of 500 members each may leave 25 dead on a field and live in the memory of generations, but in the international system such a war does not merit notice. The short border war between Libya and Egypt in 1977 killed less than a thousand, perhaps. The declared war of Rhodesia's neighbors against her white-ruled government has by 1978 cost less than a thousand killed. And then in 1978 there are the current wars between Cambodia and Vietnam, and Ethiopia and Somalia, where the death toll runs at least into several thousand. Yet in the international system these wars are still relatively minor. None of these has cost as many lives as the mass killing of its people by the Cambodian Khmer Rouge regime in the period 1975-1977, for example.
And then at the other end of the scale are the very intense, general wars costing millions of lives.
Now, theoretically, the number of wars should be least in a polarized system, but if any wars do occur the polarity almost guarantees that they will be extremely intense guerilla, revolutionary, or civil wars. As polarity decreases, power is decentralized, more independent states are formed and more wars are likely. But because there will be an increasing plurality of interests, the likelihood of these wars being intense in the system, or generalized to many states, decreases. Thus, the theoretical frequency curve shown in Figure 16B.2.
Polarity aggravates intense violence; but it inhibits lower level violence and, of course, nonviolent conflict behavior. With this background, the expected correlations are shown in Table 16C.1. What has to be kept in mind is that polarity has opposing effects on different conflict variables. For this reason, I predict the correlations with the conflict behavior and conflict scale variables to be random in all samples.
The frequency of low-level violence and war is inhibited by polarity and therefore should be higher in multipolar systems, indicating a negative correlation for all samples. The intensity of violence should be higher as polarity increases, and the expected correlation should therefore be positive.
One final consideration. Studies have used diverse measurements of polarity, which creates a problem in assessing evidence. Fortunately, these different measurements can be classified into two types. There are those measuring polarity by some power variables, such as the number of poles of power or number of Big Powers. And there are measurements of interstate relations from which is inferred polarity. For example, the distribution of trade, intergovernmental organization memberships, or diplomatic ties, may be taken to measure polarity.
The fault with measuring polarity by such relations, however, is that they are reflecting cross-pressures and not necessarily polarity. While increased polarity should decrease the diversity of such relations, it does not follow that decreasing polarity will increase this diversity (see the theoretical discussion for Proposition 16.20). Moreover, different systems at the same level of polarity may have more or less dispersed relationships between members, depending in part on whether the political systems are authoritarian, libertarian, or totalitarian.
Therefore, I have treated all measurements of polarity in terms of behavioral relationships as really measurements of cross-pressures and have related the results to the cross-pressures propositions.
Evidence: The only evidence is in the studies by Haas (1974) and Weede (1975), the former being important and direct. Both are positive.
Conclusion: The evidence supports the proposition.
Cross-pressures are negatively correlated with intense political violence.
Theory: This is the counterpart to the Cross-Pressures/Conflict Proposition 16.15. Multiple overlapping interests generate variegated conflicts. But, the cross-pressures and segmented interests restrain conflict from coalescing, inhibit the formation of a societywide conflict front, and therefore constrain the intensity of violence.
At the international level, cross-pressures cause (aggravate) conflict and inhibit intense violence.
Prediction: Cross-pressures and polarity have a theoretically complementary relationship, although they are distinct concepts (see the discussion for Proposition 16.20). The positive or negative correlations expected for cross-pressures arc the opposite of those expected for polarity.
Evidence: There are six direct, positive analyses (one strongly so), five of which are important, versus one direct and important negative analysis. And there are three analyses with ambiguous results.
Conclusion: In spite of there being only one negative analyses, this plus the ambiguous analyses make up 40% of the total. Therefore, the more restrained conclusion: The evidence tends to support the proposition.
Opposition of world opinion is negatively correlated with violence.
Theory: World opinion consists of the public and elite opinion in other states opposing a conflict and the interests their leaders express in seeing the conflict ended. This opinion tends to dampen conflicts for two reasons.
First, world opinion provides spiritual support and political ammunition to those interest groups and that public working against their state's involvement in a conflict, as happened in the United States during the Vietnam war. In the short run, however, the government may ignore domestic clamor if the leadership perceives the dispute to be critical, as did British Prime Minister Eden in his approval of an invasion of Egypt in 1956 after President Nasser took over the Suez Canal.
Second, world opinion indicates possible difficulties a state will have in its future relations with other countries if it continues in or escalates a conflict. Statesmen are sensitive to the interdependence of all international relations, to its field nature. Pushing a dispute here may negatively affect the negotiation of a treaty there. Or request for economic aid. Or an arms sale agreement. Or trade.
Thus, affronting world opinion, especially of those states that matter in some way, introduces a different set of costs into a conflict which a leadership must weigh against the potential gains (including a reputation for not knuckling in under world opinion, which may sometimes be generated by the propaganda agencies of the opposing states). The stronger and more vociferous this opinion, the more salient it becomes and the more attention will be paid by the parties to the implied costs.
Prediction: World opinion operates best once violence has occurred and at the higher levels of violence. Border clashes or minor military engagements often occur over one or several days and usually are not that important at a global level. They will escalate to a war before usually attracting world attention and before opinion can be worked up. Therefore, there should be no systematic relationship between world opinion and the occurrence of low-level violence, meaning that the predicted correlations will be random in all samples, as shown in Table 16C.1.
However, once war is underway, it always attracts world attention; and because it is usually ongoing for months, war gives world opinion a chance to crystallize, and to thus inhibit escalation. Therefore, both the intensity of violence and the conflict scale should be negatively correlated with world opinion in all samples.
Because world opinion focuses on war, world opinion should have no systematic relationship to conflict behavior in general or to nonviolent conflict behavior: the expected correlations are random for these variables.
Evidence: The evidence is all positive, direct, and important, but only consists of three analyses. One is Wright's (1965) analyses of escalation. A second is my reanalysis of Wright's data with my own added (Project 6). And the third is the Northedge and Donelan (1971) nonquantitative, but systematic analysis.
Conclusion: the evidence supports the proposition.
A stable status quo is negatively correlated with the intensity of nonviolent conflict behavior.
Theory: A structure of expectations is composed of both core status quo agreements, treaties, understandings and surrounding norms, practices, and implicit conventions. Conflict may occur over these nonstatus quo expectations, while the status quo itself remains stable. The dispute may be over airline landing procedures, the terms of a trade agreement, payment for damage done by a tanker spilling oil, lack of consultation or warning about a change in foreign policy, hostile propaganda, and so on. Territorial issues are not involved; in a widely evident sense, what is ours versus what is theirs is not at stake.
As long as the status quo remains stable, conflicts over other issues will be inhibited from escalation. The reason is that a stable status quo-one which is congruent with the balance of powers--is mutually satisfactory and thus both sides will tend to inhibit the intensity and range of their dispute so as not to endanger the status quo.
This proposition is complementary to the Status quo Disruption Proposition 16.10, which states that status quo disruption is a necessary condition for violence. So long as the status quo is stable, nonviolent conflict is inhibited and violence should not occur. But if the status quo itself becomes part of the dispute and is disrupted, then the door to extreme conflict and violence is opened.
Prediction: As a variable in this proposition, status quo ranges from stable to very unstable, with disruption at the extreme end. The prediction is, therefore, that status quo stability will be negatively correlated with nonviolent conflict behavior and the conflict scale for the mixed and conflict samples. For the violent samples, by theory (Proposition 16.10) the status quo has been disrupted and therefore a stable status quo should be absent for all variables.
For the frequency of violence and war and the intensity of violence variables, disruption should have occurred for all nonzero values (see Figure 16C.1), while the zero values can involve a stable status quo. Therefore, the correlation should be a moderate negative for each variable in the mixed and conflict samples.
Evidence: The Barringer (1972) analysis, which is direct and important, is the only one relevant to the proposition. His results are strongly positive.
Conclusion: Because there is only one relevant analyses, the conclusion: insufficient evidence.
The more libertarian a state, the less it tends to be involved in violence.
Theory: Libertarian states are involved in warfare, military intervention, and other kinds of international violence. This is usually reactive violence, a response to perceived aggression from nonlibertarian states or movements, where aggression also means unilateral attempts to change the status quo, as in Egypt's 1957 seizure of the Suez Canal; or attempt to stabilize a political situation from which opposing ideologies could benefit, as in the U.S. intervention in the Dominican Republic in 1965, or Lebanon in 1958.
Nonetheless, in comparison to other states, libertarian states have natural inhibitors on Involvement in violence: the responsiveness of elected leaders to domestic interest groups and public opinion, which ordinarily will oppose violence, tax increases, and conscription, unless there is a clear and present danger. Libertarian states will defend their basic interests, and with violence if need be, if the aggressors are nonlibertarian (as defined in Joint Freedom Proposition 16.11). But domestic interests set limits and libertarian leaders often lack the power or will to take violent initiatives or make moves escalating violence, unlike their authoritarian and totalitarian counterparts.
Prediction: A prior question has to do with the kind of violence limited by libertarian systems. Libertarian systems are the natural enemies of authoritarian and totalitarian states. By their example and the products of freedom they are naturally subversive of authoritarian or totalitarian systems; and these freedoms seem to make libertarian states defenseless against unilateral changes in the status quo. Thus, libertarian states are often involved in reactive and defensive violence against the initiatives of nonlibertarian states. Therefore, in general, I do not expect that there will be a correlation between libertarianism and the frequency of involvement in war or violence. Nor should there be for the conflict behavior variables. The predicted correlations for these variables are therefore random, and for all samples as shown in Table 16C.1.
However, once a libertarian state is involved, domestic forces will usually begin to coalesce against increased violence and for a settlement of some sort. The growth in anti-Vietnam war sentiment and its impact on the American leadership's war policies and decisions are a paradigm case of the Freedom Proposition. It follows that the intensity of violence variable (which measures the scope, occurrence, and degree of violence) and the conflict scale (which has intense violence at the extreme) should be negatively correlated with libertarianism for both the mixed and conflict samples.
For the violence and war samples, however, it is not clear whether there should be a general negative correlation or a random correlation. The prediction is therefore indeterminate.
Evidence: There are 25 relevant analyses, all but two of which are direct. Thirteen analyses support the proposition with 10 opposed; four are strongly for, six strongly opposed. The proportions change much more in favor of the proposition when only important studies are considered, or when limited to those which are not inconsistent with the Model II distance-vector assumptions of field theory.
Conclusion: The negative evidence is substantial and does not allow for unqualified support for the proposition. Therefore, the conclusion: the evidence tends to support the proposition.
Theory: "Surprise" is a characteristic of a class of trigger events (not all trigger events) whose occurrence suddenly focuses perception of opportunity, threat, or injustice. These are the crises creators, the disruptors of the structure of expectations, the stiffeners of will. They are the immediate, the proximate causes of conflict or escalation.
What such a spark might be is unpredictable. It could be an assassination, a terrorist hijacking, a ship blown up, a sighting of dangerous weapons secretly mobilized in neighboring territory, a unilateral change in the status quo, a coup d'état, and so on.
Prediction: Surprise should be positively correlated with crises, with the initiation of conflict behavior or its escalation. There may be considerable conflict as part of the process of balancing without there being surprising trigger events correlated with it. Surprise is initiating or escalatory. Thus, the positive correlations should be with nonviolent conflict behavior, violence frequency, violence intensity, and the conflict scale, in all samples.
Evidence: Two analyses (Phillips and Hainline, 1972; Brady, 1974), directly support the proposition, one strongly so.
Conclusion: the evidence supports the proposition.
Perception of opportunity, threat, or injustice is positively correlated with Conflict Behavior.
Theory: This is a broad class of trigger events whose occurrence disrupts an incongruent structure of expectations (Proposition 16.1).
The distribution of obligations, benefits, and rights locked into a structure of expectations become incongruent with the reality of power. Dissatisfaction develops. And then some event which involves a perception of opportunity (such as a mutiny and breakdown in law and order in an opposing state which indicates weakness) a threat (such as the other side unveiling a new weapon in a military parade), or injustice (such as discovery that the other side has been systematically violating a treaty) brings an increasingly unhappy situation of conflict to a head. And Conflict Behavior may ensue.
The trigger is thus a proximate or immediate cause of Conflict Behavior.
Not all such triggers occur abruptly, as in the Surprise Proposition 16.28. Some may be the looked for event in an increasingly tense situation in which the parties had gradually grown to expect some kind of outbreak. The trigger is like an excuse. Some triggers may be a final straw in a growing list of grievances, or the last bit of evidence to conclusively prove duplicity or subversion.
Triggers which appear abruptly (Proposition 16.28), surprising the parties and catalyzing conflict, are not only a subclass of disruptors, but also can escalate conflict once underway. Disruptors, however, act only on the structure of expectations.
Prediction: A trigger disrupts expectations because it causes a perception of opportunity, threat, or injustice. The trigger event occurs and disappears proximate to the outbreak of Conflict Behavior, but the perception stimulated by the event will remain through the early stages of conflict, if not throughout.
Thus, this perception should be present for all conflict variables and all samples. And the more intense the conflict, the more intense should be this perception of opportunity, threat, or injustice. Thus, the correlation with all conflict variables should be positive.
Evidence: The evidence overwhelmingly and directly supports the proposition, with 20 analyses for (and eight strongly); two against (one strongly); none ambiguous.
Conclusion: the evidence strongly supports the proposition.
The more totalitarian a state, the more correlated its Conflict Behavior with political difference.
Theory: Totalitarian states are coercively unified by an enforced definition of the true and just--by an ideology. By a political formula. Competition with this formula is not permitted, critical assessments are not allowed. The people are taught one truth, one justice.
Therefore, totalitarian leaders are most sensitive to the ideologies of other states, whether these be revisionist or heretical deviations from the one truth, or competing ideologies. And whether those be actively competing ideologies or those competing by example. In any case, the mere existence of alternative political formulas is a threat to the one truth.
Thus, totalitarian states will be particularly affected by political distance, for this is the axis along which opposing ideologies lie. For these states more than others political distance is an aggravating cause of conflict.
Prediction: The more totalitarian a state, the more its conflict will be correlated with political distance vectors. This positive correlation should appear for all samples and conflict variables.
To be clear, this is not to say that this correlation will be higher than for other distance vectors. It is to say that ignoring other distance vectors, that correlation for political distance will be higher for totalitarian states than it will be for authoritarian or libertarian states.
Evidence: Six analyses are relevant, five of which support the proposition; four strongly so. All direct studies strongly support the proposition; all important studies are supportive. The only negative evidence is my own (Project 45), some of which is shown in Appendix 9A.
Conclusion: the evidence strongly supports the proposition.
National coercive power is positively and, among attributes, most correlated with conflict behavior.
Theory: Power breeds conflict. Power shapes conflict. Such were conclusions of Vol. 2: The Conflict Helix. For interstate relations, no less than for their internal affairs, power is the most significant variable in the genesis and form of conflict behavior.
The balancing of power underlies conflict dynamics; the balance of powers undergirds a structure of expectations; the change in power creates incongruence with expectations and the likelihood of disruption and a new outbreak of conflict (Proposition 16.2); the power to transcend distance makes conflict possible (Proposition 16.5); the decentralization and pluralism of power in a state inhibits violence (Proposition 16.27), while such between two states precludes violence between them (Proposition 16.11); the role of power in status is a force toward conflict (Proposition 16.14); the reputation for weakness in power risks escalation in violence (Propositions 16.17 and 16.18); the polarization of power stimulates intense violence (Proposition 16.20) while inhibiting lower-level conflict behavior (Proposition 16.23); the relative distribution of power defines a class division, which is the cauldron of general war (Proposition 16.22); the perception of power can trigger conflict (Proposition 16.29) and escalation (Proposition 16.28); and the power parity between states can make the escalation to war more likely (Proposition 16.21).
Of course, power takes many forms and should be understood as the product of interests, capabilities, and will. But between the governments of states, which are the agents of coercion, coercive power capability is the primary ingredient. Therefore, among states (coercive) power breeds conflict.
Prediction: National coercive power should be both positively correlated with conflict behavior, and among a state's attributes, the most highly correlated. This correlation should hold for all samples and all conflict variables, as shown in Table 16C.1.
Evidence: There are 46 analyses bearing on the proposition, 40 of them positive (18 strongly so). Most of these provide direct evidence, but are mainly from unimportant studies. Of the important evidence, 13 favor (8 strongly) the proposition; 4 are strongly negative.
There is only one negative study which does not violate the Model II, distance-vector assumption of field theory.
Conclusion: The evidence strongly supports the proposition.
Difference in coercive power is most highly correlated with conflict behavior.
Theory: This is a counterpart proposition to Proposition 16.31. Power breeds conflict and it is relative state power that is the force involved. Absolute power is largely irrelevant in international relations. The total income, resources, armaments, and the like, of states have meaning only in time and space relative to other states. The power of Egypt in 1978 would have made it the supreme power of 1878 and would make it now the dominant regional power in Latin America, were it located there. The power of China relative to Vietnam is one thing; relative to the Soviet Union is another.
This relativity itself is not absolute. The magnitude of the difference in power (i.e., power parity) is only relevant to behavior in making the escalation to war more likely (Proposition 16.21). It is the direction of difference and its magnitude that is generally a force toward antagonistic conflict behavior. That is, a state's conflict behavior towards another state is affected by whether the other is stronger or weaker in coercive power and by how much--by the distance vector.
Prediction: Among all distance vectors, the power distance vector between states should be most highly positively or negatively correlated with their conflict behavior. This should hold true for all samples and all conflict variables.
Power refers to coercive power. It may be measured by power-in-being, such as existing armaments and deployment. Or it may comprise, in addition, the power potential of a state--the totality of its resources, national income, population, area, leadership, morale, political unification and control.
Evidence: Twenty-one different analyses are relevant, 19 directly. Fifteen are strongly positive and three positive; only one is negative. Ten important studies are strongly positive; two ambiguous; none negative. Overall, this is the strongest relative support that any proposition has received. Some of my own important and strongly positive evidence is given in Appendix 9A.
Conclusion: the evidence strongly supports the proposition.
Differences (distance vectors) in wealth, power, and politics are of all differences the most correlated with conflict behavior.
Theory: Power influences conflict in a variety of ways, as mentioned in the State Power and Power Vector Propositions 16.31 and 16.32 ). Wealth affects conflict through its role in status (Proposition 16.14), especially in relative status (e.g., rich-poor gap); in sociocultural distances (Proposition 16.6); and in dissimilarity (Proposition 16.12). Moreover, relative wealth defines the exchange system between states, the potential dependency of one state's trade on another.
Political distance is potent for totalitarian states (Proposition 16.30), but as with wealth, also affects conflict through its contributions to sociocultural distances. Moreover, political distance for libertarian states defines the limits of conflict (Proposition 16.11).
Together, power, wealth and political distance reflect the primary lines of opposition among interests and the relative capabilities of states. Together, they define the major axes of the interstate balancing of powers.
Prediction: The wealth, power, and political distance vectors between states together should be more highly correlated with conflict behavior than any other sociocultural distance vectors, singularly or in the aggregate. This should be true for all conflict variables and all samples.
Evidence: Of 15 analyses, all but one direct, 13 support the proposition--strongly so. The one negative I inferred from a less than important (in this context) study by Salmore and Hermann (1970). Some of the positive evidence from my own work is shown in Appendix 9A.
Conclusion: the evidence strongly supports the proposition.
First, note that the percent distribution of evaluations do not vary much by category. What variation there is argues against serious bias. The direct evidence hardly varies in percent positive or negative from that overall, and the strongly positive evidence even increases proportionately. Rather than evidence being skewed positively by unimportant studies, it is the other way around: important studies are more strongly positive.
Moreover, as should be the case if field theory provides the best framework for understanding conflict, studies which do not violate the Model II and distance vector assumptions of the field, are of all most proportionately positive. Out of 202 such ratings, only 8% are negative; 85% are positive.
Finally, if I have been biased toward the proposition in my own analyses then my results should be most favorable. But in fact, my results are least favorable. Of 99 ratings of my results, 23% are negative (compared to 16% overall); 72% are positive (compared to 79% overall).
Also, note that in Table 16C.3 even when the one-level decrement is applied, all percents still favor the propositions; about three to one for field theory consistent studies; over two to one for important studies.
The conclusion about bias seems clear. If bias is influencing the results, it must be nonsystematic or operate in ways not measured by the division in Table 16C.3. However, both as a result of what is shown in Table 16C.3 and my evaluation procedures, I feel confident that the conclusions are not unconsciously distorted. Indeed, I feel that I was probably too cautious in the evaluation and understated the positive evidence.
Now, for the overall conclusions. As tabulated in Table 16C.3, 79% of 390 ratings support the propositions; 35% of the total do so strongly. Quantitatively, recognizing all the dangers and assumptions in treating each analysis as the same, the total evidence overwhelmingly supports the propositions.
To avoid dependence on overall counts alone, qualitative distinctions can be introduced, as was done in classifying the evidence by the categories in Table 16C.3. But, as can be seen, each category by itself leads to the same positive conclusion.
Another qualitative-quantitative way of assessing the overall evidence is by separately assessing each proposition. The resulting conclusions are listed in Table 16B.2. Each of these conclusions took into account the quantitative evidence and my subjective evaluation of the analyses involved, their research designs, adequacy of data, and the like. As can be seen, overall the evidence is supportive, with no negative conclusions.
Therefore, the overall conclusion: the empirical evidence overwhelmingly supports the propositions. And thereby, social field theory and the conflict helix.
* Scanned from Appendix 16B 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. Technically, upon standardization the vector (variable) with constant values disappears into the origin of the space defined by the cases. Because any null vector is orthogonal to any other vector, the orthogonality condition for zero correlation is met.
2. To be clear, I am referring to the "cut of the data," and discriminating among system, dyadic, and state levels (research designs). This should not be confused with the applicability of polarity to the societies within states. See Chapter 32 of Vol. 2: The Conflict Helix for such an application.
3. This is violence in terms of riots, revolutions, guerrilla war, and the like (see Section 35.3 of Chapter 35 in Vol. 2: The Conflict Helix). The systematic violence of a regime against its people has not been included in empirical studies. Were such violence included, then the violence curve of Figure 16B.2 is what should be found.