Nations at War

Geller, Daniel S. & J. David Singer (1998) Nations at War: A Scientific Study of International Conflict. Cambridge Studies in International Relations. Cambridge: Cambridge University Press. ISBN 0 521 62119 4 (hb); 0 521 62906 3 (pb).


by Johan M.G. van der Dennen

The authors
Daniel S. Geller is Professor of Political Science at the University of Mississippi. He is the author of Domestic Factors in Foreign Policy and numerous journal articles and book chapters on war, weapons technology, and military and economic power.

J. David Singer is Professor of Political Science at the University of Michigan and Director of the Correlates of War (COW) Project, a database on all international and civil wars from 1816 to 1993. He is the author, co-author or editor of 20 books and well over 100 scholarly articles and monographs. Among the better-known - and by now classical - volumes are Singer (1968), Singer & Small (1972), Singer et al. (1979), Singer (1980), Small & Singer (1982) and Gibbs & Singer (1993).

Nations at War provides an explanation of war in international politics grounded on data-based, empirical research. The book classifies and synthesizes the research findings of over 500 quantitative analyses of war at the analytical levels of the state, the dyad, the region, and the international system. Because wars follow from political decisions, two basic decisionmaking models - the rational and the nonrational - are examined in relation to the explanatory framework of the volume. In addition, case analyses of two wars - the Iran/Iraq War (1980) and World War I (1914) - are provided as demonstrations of scientifically-based explanations of historical events. The primary structural factors responsible for the onset and seriousness of war are identified and the explanations are developed according to the scientific model of 'covering laws.' The conclusion presents discussion of the potential for probabilistic predictions of conflict within the context of war and peace studies.

Some 20 years ago, Vasquez (1976), Rummel (1979), Zinnes (1980), Beer (1981), and Van der Dennen (1981) - later followed by Vasquez (1987), Levy (1989) and Hower & Zinnes (1989) - were the first to review and summarize the results of the vast macroquantitative literature resulting from ambitious data-based projects such as Singer's Correlates of War (COW) and Rummel's Dimensionality of Nations (DON), now complemented by the Conflict and Peace Data Bank (COPDAB), the Militarized Interstate Disputes data set (MID), and the Long-Range Analysis of War Project (LORANOW), among others. The conclusions of these reviews will be compared with the Geller & Singer conclusions in order to check if any serious progress has been made (if at all) in the macroquantitative approach to interstate violence.

Monadic level: war-proneness of states
If states do not act similarly in similar situations, then perhaps variation in foreign policies is due to variation in the internal attributes of states. According to this logic, states with the same critical internal attributes will evidence similar patterns of war behavior, distinct from the patterns produced by states with different attributes. Chapter 3 focuses on quantitative empirical research that relates to the war-proneness of states.
A set of consistent and cumulative empirical uniformities on the onset and seriousness of war are identified at the state level of analysis. Factors increasing the probability of the onset (occurrence/initiation) of war at the level of the state are:

* Power status (major power)
* Power cycle (critical point if major power)
* Alliance (alliance member)
* Borders (number of borders)

The factor identified as increasing the probable seriousness (magnitude/duration/severity) of war at the level of the state is:

* Power status (major power)

At the analytic level of the state, national capabilities - whether measured in terms of military capabilities alone or a combination of military, economic, and demographic variables - reveal strong and consistent linkages to foreign conflict. Most notable are the distinctions among nations in power status and war behavior. Power status has shown a relationship to the frequency of war, the initiation of war, and the severity of war: major powers are much more likely to engage in war than are minor powers, and they are more likely to fight severe (high battle death) wars than are minor powers. The level of a state's militarization also reveals consistent positive associations with violent foreign conflict. The exception to these findings involves the rate of military buildups. However, in this area the most relevant results pertain to dyadic-level arms races rather than to state-level increases. Lastly, critical points on a major state's power cycle appear to be strongly associated with both war initiation and war involvement
At the analytic level of the state, there appears to be a strong connection between the number of alliances and involvement in war. Although it is acknowledged that alliances are frequently formed in anticipation of war and that, as a result, the statistical association may be noncausal, the pattern is nevertheless consistent and cumulative and may be considered substantive.
The evidence on the issue of borders and war is consistent and robust. Irrespective of explanatory model - physical opportunity or the context of interactions - there appears to be a substantive linkage between the number of borders and frequency of war at the level of the monad.
Although little data-based evidence exists to date, the preliminary findings indicate that dissatisfaction with the status quo may be an important factor in patterns of war initiation and in levels of foreign policy hostility.

There is little supportive evidence of a connection between the following national attributes and foreign conflict behavior or war: measures of population size and density (subpopulations have not been examined until very recently, however, vide infra), geographic size, economic development, business cycles, and national culture appear to be unrelated to monadic-level foreign conflict. The is also no evidence supporting a national-level cycle hypothesis. A strong and consistent empirical association between internal and external conflict has not been established.
Evidence regarding a connection between regime characteristics and war can be divided into two categories - regime attributes and regime-type. The former (attribute) category consists of factors inclusive of centralization, bureaucracy, election cycle, and formation/transformation. Results in this area show consistent relationships between regime attributes and foreign conflict behavior. However, the imprecision of some of the concepts and the absence of replications precludes a substantive evaluation based on cumulative evidence. The case is quite different with regard to regime-type and patterns of war. Here, the weight of cumulative evidence reveals the lack of a substantive connection between the form of government and involvement in militarized disputes or wars. At the level of the monad, democracies are as likely to engage in war as are states with other types of regimes.

Dyadic level: war-proneness of dyads
In contrast to state-level studies, dyadic analyses permit the examination of relational attributes for pairs of states that engage in conflict. Some of these elements have a substantial history in realist explanations on the causes of war (e.g., capability differentials, alliances, and arms races), while others have constituted cornerstones of liberal philosophy as elements associated with peaceful interstate relations (e.g., economic development, free trade, and democracy). Chapter 4 focuses on quantitative empirical research that relates to the war proneness of dyads. A set of consistent and cumulative empirical uniformities on the onset of war are identified at the dyadic level of analysis and are noted in that chapter. Factors increasing the probability of the onset (occurrence/initiation) of war at the level of the dyad are:

* Contiguity/proximity (common border/distance)
* Political systems (absence of joint democracies)
* Economic development (absence of joint advanced economies)
* Static capability balance (parity)
* Dynamic capability balance (unstable: shift/transition)
* Alliance (unbalanced external alliance-tie)
* Enduring rivalry

The findings of research on proximity/contiguity and war are cumulative and consistent. At the level of the dyad, the distance between states is inversely related to warfare. Proximate dyads are more likely to engage in war than are nonproximate states. The results on contiguity are even more compelling. War within dyads with a land or (narrow) water border is much more likely than between noncontiguous states.
The review of quantitative empirical research on the possible relationship between dyadic-level capability balances and conflict has focused on patterns of dispute/war initiation and dispute/war occurrence for populations of all interstate dyads, major power dyads, and enduring dyadic rivalries. The analysis suggests a growing and cumulative body of evidence pointing to the salience of both static and dynamic capability balances for the occurrence and initiation of militarized disputes and warfare. Specifically, conditions of approximate parity and shifts toward parity are consistently and significantly associated with conflict and war irrespective of population.
The evidence regarding alliance within dyads and the onset of war is intriguing. Whatever the explanation, dyads with one external alliance are more war-prone than dyads where both parties have external ties.
Rummel (1979) established the 'joint-freedom proposition': 'libertarian' (as opposed to authoritarian or totalitarian) states will not engage in violence with each other. Also Gleditsch (1995:318) concludes a recent study that "double democracy is a near-perfect sufficient condition for peace" (italics in original). The evidence in the area of the joint-freedom proposition is consistent and cumulative. Democratic dyads are less likely to engage in war than are nondemocratic pairs. It has been asserted several times that the absence of war between democratic nations is one of the strongest empirically-based generalizations that exists in the field of international politics. There are two competing explanations of this phenomenon. One explanation focuses on the political culture of democratic states (i.e., nonviolent norms), whereas the other (structural/institutional) explanation involves democratic political structure (i.e., decisionmaking constraints). The latter explanation seems to be logically inconsistent with the evidence at the monadic level.
Based on a small number of recent studies, patterns of both conflict escalation and war initiation appear to be affected by the distribution of military capabilities between status quo defenders and challengers (in accordance with Organski's power transition theory and Galtung's status inconsistency theory). Almost every case of defender initiated war occurred under an unstable military balance. This pattern suggests that stable military balances of either preponderance or parity are generally interpreted by status quo defenders as supportive of deterrence, whereas unstable balances producing capability shifts or transitions are deemed dangerous enough to provoke preventive military action.
The evidence in the area of economic factors is suggestive but limited. The most compelling finding reveals the relative absence of war within dyads characterized by a high level of economic development.

The literature on arms races and war is inconsistent. As Siverson & Diehl (1989:214) note, "If there is any consensus among arms race studies, it is that some arms races lead to war and some do not".
The results of analyses on alliance and war are also mixed. In the area of alliance within dyads, the findings are inconsistent.
Evidence on a relationship between dyadic trade patterns and conflict is also mixed.

Regional level: war-proneness of regions
Chapter 5 examines data-based research on war-prone regions. The initial section of the chapter focuses on simple comparisons between geographical regions in terms of the number of crises, subwar disputes, and wars as well as inter-regional comparisons of processes of militarized dispute contagion. The second section of that chapter explores evidence relating to intra-regional patterns of conflict and war that may have been dependent upon spatial heterogeneity, population pressure, polarity, and normative constraints. The third section of the chapter examines evidence of war contagion as it pertains to patterns of intra-regional behavior. The final section in chapter 5 examines evidence of intra-regional cycles of war. One consistent and cumulative empirical pattern on the onset of war is identified at the regional level of analysis. The factor increasing the probability of the onset (occurrence) of war at the level of the region is:

* Contagion/diffusion (presence of an ongoing regional war)

From 1816 to 1945 most conflicts involved the European major powers and the United States. In the post-1945 period - on a per-state basis - the Middle East, Asia, and Africa become the most conflict-prone. The latter development seems to confirm the observation that the world is currently divided into 'zones of peace, wealth, and democracy' on the one hand, and 'zones of turmoil, war, and development' on the other (see also Gantzel & Schwinghammer, 1995).
The results of a small number of quantitative studies suggest that a war contagion/diffusion process operates at the regional level.

The analyses produce mixed results on the issue of regional cycles of war. Farrar (1977) explored the possibility of cycles of war in the European state subsystem for the period between 1494 and 1973. Drawing on data from Wright (1942), Farrar identifies three types of war: 'probing' war (i.e., minimal violence causing little change), 'adjusting' war (i.e., moderate violence but does not threaten the status hierarchy) and 'hegemonic' war (i.e., massive violence that threatens or alters the hierarchy). Farrar claims that these wars occur in a sequential pattern - probing, adjusting, and hegemonic wars tend to follow in that order. The complete cycle of war-types runs for approximately 100 years and is repeated four times during the period between 1494 and 1973.

Systemic level: war-proneness of systems
Chapter 6 reviews data-based research on war-prone systems. Here, certain system-level attributes such as polarity, alliances, capability concentration, Kondratieff economic cycles, normative constraints, and the presence of intergovernmental organizations are examined for war effects. A set of consistent and cumulative empirical uniformities on the onset and seriousness of war are identified at the systemic level of analysis. Factors increasing the probability of the onset (occurrence) of war at the level of the international system are:

* Polarity (weak unipolarity/declining leader)
* Unstable hierarchy
* Number of borders
* Frequency of civil/revolutionary wars

The factor identified as increasing the probable seriousness (magnitude/duration/severity) of war at the level of the international system is:

* Alliance (high polarization)

The results of data-based studies on polarity and warfare are mixed with no definitive linear pattern evident regarding unipolar, bipolar, and multipolar configurations and the occurrence of war. The only polar structure that appears to influence conflict probability is unipolarity.
Every international system possesses a hierarchy based on relative capabilities, and the extent of the capability differential between the leading state and potential challengers matters. If the hierarchy is clear, with the leading state in possession of a substantial capability advantage over its nearest potential rival, then the probability of action to rearrange the hierarchical order is likely to be low. However, if the capability advantage of the leading state is small or is eroding, other states my choose to attempt to alter the hierarchy. The challenges may be directed against the leading state or lesser states within an increasingly unstable international order. This logic is found in the works of Organski, Modelski, Gilpin, Wallerstein, and others. Changes in the hierarchical arrangement of major powers appear to be associated statistically with the onset of great power warfare.
In sum, shifts in system-level power structure appear to have an interactive effect with dyadic capability distributions in the onset of both major power and global wars. As the international system moves from a high concentration of resources in the leading state toward multipolarity (power diffusion), lower-order conflict among the set of major states will become increasingly probable, due to the weakening of the principal defender of the hierarchy.
The evidence regarding system-level alliances and warfare is clear. The onset (occurrence/initiation) of war is unrelated to either alliance formation (aggregation) or configuration (polarization). However, the magnitude, duration, and severity (or 'seriousness') of war does show consistent and significant correlation with the configuration of alliances. This is not surprising since alliance is a principal mechanism by which small wars become big wars - war spreads through alliances.
Goldstein (1988) presents an exhaustive analysis of 'long waves' in economic factors and wars. Among his findings, Goldstein reports that only the severity of great power war correlates strongly with economic long waves.
Modelski's (1981) basic long cycle theory posits 100-year cycles of increasing power diffusion and subsequent global warfare followed by a high level of military/economic capability concentration in a single state (i.e., unipolarity) resulting in global order.
The application of Kondratieff 'waves' - the expansion and contraction of the global capitalist economy - to the occurrence of major power war has been explored by Modelski, Wallerstein and Mansfield. Mansfield's (1988) data-based study reports that the mean number of wars begun during the upward phase of the Kondratieff cycle is greater than the mean number during the downward phase. This relationship is statistically significant, but the tendency for wars to vary with Kondratieff periods only appears when all wars are examined. The association does not hold for major power wars alone, as is maintained in the principal arguments of Modelski and Wallerstein.
The results of the set of studies relating subsystem (state-level) attributes to system-level patterns of conflict and war are also interesting, with the global counts of civil and revolutionary wars revealing positive associations with global dispute and war frequencies.
The findings on the positive relationship between the total number of borders in the interstate system and the frequency of war participation in the system mirror the results on borders/contiguity and war at the monadic and dyadic levels of analysis. The number of borders and contiguity correlate positively with conflict and war at the state, dyad, and system levels.
Gochman & Maoz (1984:593) conclude that "the apparent increase in the disputatiousness of interstate relations over the past century and a half is, at least in part, a function of the growth in the size of the interstate system". Levy's (1982) analysis of the frequency and seriousness of great power wars between 1495 and 1975 indicates that great power war has been decreasing in frequency but increasing in seriousness over time.
Levy & Morgan (1986:46) report that the "probability of war is independent of the period of time since the last war" (which refutes the war-weariness hypothesis). They also note that there is no evidence supporting the hypothesis "that more serious wars generate a greater inhibition against subsequent war". In short, the systemic distributions of great power wars appear to be random with respect to time.

System-level capability concentration is unrelated to the occurrence of major power war.
A second pattern relating state-level attributes to system-level frequencies of conflict is the finding of a negative linkage between regime homogeneity and international war - global conflict levels vary in relation to the distribution of regime-types.
Positive findings of war contagion/diffusion at the regional level of analysis are consistent and cumulative. However, the results of studies at the systemic level are ambiguous.
The data-based findings on the efficacy of peaceful international norms and the establishment of IGOs as means of reducing system-level violence are not encouraging. Norms constraining the use of force or proclaiming the illegality of war do show some ability to reduce war behavior. However, as Kegley & Raymond (1986:213) note: "the historical evidence indicates that these conditions are met only rarely, and that the formation of such an international security regime is ephemeral".
The results involving the effects of IGOs on system-level conflict are less consistent, but not particularly hopeful. There is an empirical tendency for IGOs to be established at the conclusion of massive wars. However, the historical success of diplomatic institutions and procedures for nonviolent conflict resolution is punctuated by catastrophe. As Vasquez (1993:283) notes: "Some of these formal rule-making efforts, like the Peace of Westphalia and the Congress of Vienna, have been highly successful in resolving the issues that led to war and mitigating the use of violence. Others, such as the Versailles system, have been disastrous, bringing about the very war they tried to prevent".

Case analyses of two wars - presented in chapters 7 and 8 - are provided as demonstrations of scientifically-based explanations of historical events. More specifically, the cases of the Iran/Iraq War (1980) and World War I (1914) are explained on the basis of empirical uniformities established through systematic quantitative analyses. The explanations in chapters 7 and 8 demonstrate that the Iran/Iraq War and World War I are specific instances of a set of patterns which have appeared in a much larger number of cases. In short, both wars were high-probability events consistent with a broad array of empirical patterns.
Wars are contests of power. They may develop over any of a large number of issues such as territory, security, wealth, or ethnicity. However, it is maintained that the presence or absence of certain factors increases the probability of the onset and seriousness of such contests. The analysis presented here has attempted to identify those factors and to provide an account of conflict using findings derived from over five decades of systematic, quantitative empirical research.

Comparison and Evaluation
So far the extensive review of Geller & Singer's Nations at War. In this section I shall attempt to evaluate the major findings of this important, and fascinating, study in comparing these findings with the results of previous macroquantitative research reviews, particularly Singer & Small, 1972; Haas, 1974; Vasquez, 1976, 1987, 1993; Rummel, 1979; Zinnes, 1980; Wilkinson, 1980 (who reanalyzed Richardson's [1960] data); Beer, 1981; Eberwein, 1981; Van der Dennen, 1981; Levy, 1989; Hower & Zinnes, 1989; Cashman, 1993; and Maoz, 1995.

Ad frequency/cyclicity
Richardson (1960) and Singer & Small (1972) found that the onsets and endings of wars per year show a Poisson distribution: "A suggestion made by the Poisson law is that... discontent with present circumstances underlies... peace and war" (Richardson quoted in Wilkinson, 1980: 19).
Small & Singer (1982) found no trends in the indicators of war over time. Unlike Small & Singer, Levy (1983) finds some patterns to the extent that the frequency of war has been diminishing, but its destructiveness increasing (Hower & Zinnes, 1989: 2).
One long-term historical trend in war is that great power wars have been declining in frequency since about 1500 A.D., but increasing in seriousness (destructiveness) (Levy, 1989: 213). The evidence presented in Van der Dennen (1981: 143-53) also tends to support the notion that there exists a long-term trend from many wars with few casualties to fewer wars with many casualties (with the twentieth century the most devastating and bloody) (see also Wright, 1942; Beer, 1974, 1981). Furthermore, since World War II, there has been a decline in conventional inter-state wars and an increase in ethno-national wars fought within the boundaries of a single state (Van der Dennen, 1981: 152).
"Wars tend to be simple rather than complex... the least complex war (one versus one) is the most common type... There are nonetheless a few wars that involve many fighting pairs. Such wars also tend to be unusually long and unusually bloody. Ordinarily, they are gang-ups (many against one, or many against few)... Richardson found that the occurrence, infrequency, and gang-up pattern of complex wars could all be explained by assuming that there exists a general propensity for states to fight anyone, modified by an overriding preference for fighting neighbors and a mild tendency for great powers to fight whomever they conveniently can; and that the propensity to fight one's neighbors increases markedly when war breaks out or spreads among those neighbors" (Wilkinson, 1980: 118-9).
"Research suggests that... long-run trends may exist. In particular, peace today may be more diffused, war more concentrated, and casualties more aggravated than in the past. An early study, published near the beginning of World War I, was already sensitive to such change.
Wars may be less frequent than formerly, yet they may be greater in magnitude, involving larger proportions of the total population. They may be more bitterly fought and subject to less interruption than in the olden times; and also the suffering may be greater even in spite of advancing knowledge and skill in the care of the wounded (Woods & Baltzly, 1915: 2).
As we move through time from 3600 BC to AD 1980, peace diffusion implies that peace periods have become longer and less frequent" (Beer, 1981: 39-40).
Proportional to the size of the system, the number of small state wars has not increased (Levy, 1989: 214).
Richardson (1960) calculated that the percentage of persons who died because of 'deadly quarrels' (i.e., violent interactions from murders to world wars) since A.D. 1820 was 1.6 percent. And he added sardonically: "This is less than one might have guessed from the large amount of attention which quarrels attract. Those who enjoy wars can excuse their taste by saying that wars after all are much less deadly than disease" (Van der Dennen, 1981: 145). Wilkinson (1980: 117) adds that the death toll of war is largely the product of a very few immense wars.
Richardson (1960) concluded that "There is a suggestion, but not a conclusive proof, that mankind has become less warlike since A.D. 1820" (Van der Dennen, 1981: 177).
"There was no clear trend in the frequency of war... Smaller wars are very much more common than larger ones; frequency varies inversely with magnitude" (Wilkinson, 1980: 117)
There do not appear to be any significant cyclical trends in the frequency of war of various types, according to Levy (1989: 299), but there is evidence that the magnitude or severity of war has been somewhat cyclical over the last five centuries. Goldstein (1988) in particular emphasizes the cyclical nature of major warfare (Levy, 1989: 299).
Wright (1942) identified a peace period of about 50 years between major war outbreaks and concentrations, and a war period of 4-5 years for the fighting of such wars (Beer, 1981: 49-50).
Sorokin (1937: 357) discounted both linear and cyclical theories of change in warfare: "All we can say is that the war-peace curve fluctuates, but in its fluctuations, with the exceptions mentioned, no regular periodicity or uniform rhythm is noticeable" (Van der Dennen, 1981: 153). On the other hand, Van der Dennen (1981: 153-4) lists a number of studies which report periodicity and/or cyclical recurrence of international conflicts, including Richardson (1960), Wright (1942), Houweling & Siccama (1977), and Singer & Small (1972). Some of the shorter cycles were interpreted as 'generation effects'.
"There is evidence of a cycle with a 20-40 year period in war frequency and participation. There may also be a longer cycle, or two, with periods on the order of 100 or 200 years" (Wilkinson, 1980: 118).
"Singer and Small's intensive study of world major wars suggests a regular peace period, somewhat shorter than Wright's, "between 20 and 40 years in the fluctuations of the amount of war underway during the 150 years" that they studied (1972: 212). On the other hand, their data show no regular peace periods when they consider "the intervals between the beginnings of successive wars" (1972: 215). New wars seemed to begin at random intervals" (Beer, 1981: 50).
Long-cycle theory, developed mainly by Modelski (e.g., 1987) and Thompson (e.g., 1988, does not attempt to explain all wars in the system, but only a restricted class of global wars, defined as those wars that determine the constitution or authority arrangement of the global political system. They are the result of a structural crisis in the system and are basically succession struggles for leadership in the system. Recent empirical research has provided some evidence in support of long-cycle theory (Levy, 1989: 255).
There is little empirical support for the arguments linking business cycles and Kondratieff long waves to the outbreak of war (Hower & Zinnes, 1989: 6). Goldstein (1985), however, discovered that "upswing periods are characterized not by more wars than on the downswings nor by wars that last much longer but by bigger wars - six to twenty times bigger as indicated by battle fatalities. These bigger wars are more costly, more destructive and more severe in their economic impact" (Goldstein, 1985: 425).

Ad national attributes
Dissimilarity of national attributes between two states seems to have had little influence on the likelihood of their going to war with each other. "Singer & Small (1972), for example, found that most of the wars in the Correlates of War project were between nations in close geographical proximity and similar on most attribute dimensions" (Van der Dennen, 1981: 162; italics in original).
On the other hand, Rummel (1979: 248) found 'sociocultural dissimilarity' to be one of the aggravating conditions of international conflict behavior generally. Similarly, differences in status (wealth, power, prestige) are positively correlated with conflict behavior (280).
Hypotheses that certain political cultures, ideologies, or religions are more warlike than others have found little support in the quantitative empirical literature (Levy, 1989: 311).
Singer & Small (1972: 287) conclude from their study of 19th- and 20th-century wars that "most of the war in the system has been accounted for by a small fraction of nations". Similarly, Zinnes (1980: 335), summarizing from many statistical studies of war, concludes that "some nations seem more prone to engage in this type of behavior than others".
Also Eberwein (1981: 23) reported that "nations differ significantly in terms of the ability or willingness to engage in conflict behavior. The same holds for the conflict nations receive".
In contrast to the finding that some states are more war prone than others, Richardson (1960) found that none can be properly characterized as inherently belligerent or inherently pacific. The problem of war does not arise from the diabolism of one or a few states (Van der Dennen., 1981: 155). Vasquez (1976; as summarized by Van der Dennen [1981: 178-9]) stated that there appears to be no strong relationship between the economic, social-cultural, political, or educational characteristics of nations and violence between nations. Thus, the notion that the national attributes of a nation may be related to international conflict is not supported. Also Zinnes (1980: 336) noted the almost total absence of any relationship (r is less than .3 and not significant) between war and its major potential determinants, such as power, regime type, size, or development and war-proneness (p. 341).
Attempts to trace war to differences between societies in their religions, languages, ideologies, and other characteristics have been slightly more successful, generally finding positive but weak relationships between societal differences and war (Levy, 1989: 311). Richardson (1960), for example, implicated differences of religion (especially Christianity and Islam) in war causation (Van der Dennen. 1981: 155). Wilkinson (1980: 119) concluded: "The propensity of any two groups to fight increases as the differences between them (in language, religion, race, and cultural style) increase. A homogeneous world would probably be a more peaceful one".
"Most of the hypotheses under consideration, however, are basically ad hoc in nature and have not been integrated into a more comprehensive theoretical framework and, as a result, it is not clear how the findings should be interpreted" (Levy, 1989: 311).
"It is hard to escape the conclusion that national attribute theories have done a relatively poor job in explaining the incidence of war... Neither the type of government a state has nor its economic institutions, its economic well-being, its population growth rate, or its previous involvement in war seems to matter much" (Cashman, 1993: 157-8).

Ad (balance of) power: parity versus preponderance
"Because of the absence of a coherent theory specifying the conditions under which parity is stabilizing and those under which preponderance is stabilizing, and because of the failure to incorporate the risk orientations of statesmen, it is not surprising that empirical research has not produced any consistent findings on this question" (Levy, 1989: 234). Singer, Bremer & Stuckey (1972) find that concentrations of military capabilities among the great powers are associated with war in the nineteenth century and with peace in the twentieth century, while parity is associated with peace in the nineteenth century and war in the twentieth century.
Rummel (1968) initially found no relationship between characteristics such as the size and capabilities of a nation on the one hand and its foreign conflict behavior on the other. In contrast, Cashman (1993: 137) reports that a sizable amount of empirical evidence tends to support the thesis that large, powerful states (regardless of the nature of their political or economic systems) tend to be perpetrators of war rather than small states.
There has yet to be a conclusive test of power transition theory or any of these hypotheses relating to the conditions under which power shifts lead to war (Levy, 1989: 258).
"Contemporary statistical studies have not really provided solid support confirming the relationship between [power] concentration and war" (Beer, 1981: 168).
In his survey, Van der Dennen (1981) found that the quantitative literature on power and war yielded inconclusive and ambiguous (sometimes contradictory) results.
Although some studies do support the dyadic preponderance hypothesis, much of the evidence runs against it (Levy, 1989: 242).
Rummel (1979: 249) considers power parity to be a particularly aggravating condition of war: "The more equal in power two states are, the more objectively ambiguous the outcome and the more both sides can believe in success". War clears up this ambiguity. Rummel also notes, however, that war still may occur, in spite of a gross inequality in military forces and resources, because other factors such as honor, credibility, survival, or determination may make the difference, as they have in the Israeli-Arab wars (254).
Also in Organski & Kugler's (1980) theory, war becomes more likely as the power parities of states converge. Also Blainey (1973: 246) asserted that "wars usually begin when two nations disagree on their relative strength".
According to Eberwein (1981: 27) "power parity between nations makes them more willing to settle their disputes by military force".
Vasquez (1986, 1993) suggests that the distribution of power affects the type of war that occurs rather than the likelihood of war.
The hypothesis that military superiority is necessary for deterrence may be rejected (Levy, 1989: 242).
"Another intriguing observation about international violence is the consistent correlation between defense expenditures and international violence... Naroll's (1969) study of 2,000 years of history allows him to conclude that 'armament tends to make war more likely'..." (Zinnes, 1980: 344; cf. Naroll, 1966; Naroll, Bullough & Naroll, 1974).
Cashman (1993: 253), after an extensive review of the literature, concludes "despite the studies by Kim and Ferris, a significant amount of evidence suggests that we should also have serious doubts about the traditional balance-of-power hypothesis that equality of power leads to peace".
Maoz (1995: 47-8) concludes that "parity in military capability, the nature of contiguity and major power presence all have a positive effect on the probability of conflict in a dyad...This suggests that as states grow more equal in military capabilities, they are more likely to be subject to the kind of calculations that are consistent with the power transition theory (Organski and Kugler, 1980). The state losing in relative capabilities invokes the declining power logic (Levy, 1987), while the state gaining in relative capabilities invokes the status inconsistency logic described by East (1972). However, these relationships should not be taken as supporting the realpolitik paradigm".
Later on (Maoz, 1995: 51-2) it is asserted that, in general, "there exists a positive correlation between the relative capability of a state and its conflict and war proneness... On an aggregate level, it is fairly safe to say that the more militarily powerful a state is, the more likely it is to be involved in a large number of militarized dispute both as an initiator and as a target".
In preindustrial (nonstate) societies, all variables which may be conjectured to be indicative of power, such as advanced military capacity, political integration or centralization, socioeconomic development, stratification, size, and 'civilization', all contribute to belligerence (reviewed and summarized by Van der Dennen, 1995).

Ad systemic status inconsistency (rank disequilibrium)
In his survey, Van der Dennen (1981) found that the quantitative literature on rank-disequilibrium and war yielded inconclusive and ambiguous (sometimes contradictory) results.
Dissatisfaction (measured by status inconsistency) does not appear to be related to the conflict behavior of nations. Evidence at the nation-state level does not seem to support the link between dissatisfaction and conflict behavior (Hower & Zinnes, 1989: 7).
In general, systemic status inconsistency appears to be a better predictor of the amount of systemic war than simple indicators of capability distribution. Furthermore, unlike the nation-state puzzle, status inconsistency at the systemic level seems to bear a significant relationship to systemic war (Zinnes, 1980: 353).
On the other hand, Cashman (1993: 232) concludes his survey that "Ultimately, status discrepancy is probably only a subsidiary factor in the cause of war".
"For individual major powers, the correlation between status inconsistency and participation in military conflicts is either moderately positive or moderately negative or completely nonexistent. At the aggregate level of analysis for all major powers only a small positive correlation is found" (Eberwein, 1981: 26).

Ad polarity
Although there has been a modest amount of work on polarity and war, the results are inconclusive, in part because of the ambiguity of the central concept of polarity (Levy, 1989: 235).
Vasquez (1976) concluded that the notion that alliances or the balance of power are related to international conflict is not supported.
In his survey, Van der Dennen (1981) found that the quantitative literature on polarity and war, yielded inconclusive and ambiguous (sometimes contradictory) results. Also Beer (1981: 165) found that the evidence of the influence of bipolarity or multipolarity was still incomplete and contradictory.
Rummel (1979: 249), on the other hand, found that polarity aggravates conflict behavior and violence: "International systems in which power is highly centralized assure that once conflict breaks out, it can easily involve the fundamental status quo and become a test of the organizing power, thus encouraging escalation and extreme violence".
Zinnes (1980) concluded her review as follows: "The pieces of the systemic puzzle are few and ambiguous. The indices of capability distribution are generally poor predictors of the amount of international violence in the system and yet indices of systemic status inconsistency, which contain capability measures, do predict, under certain conditions, some aspects of international violence. It is still more perplexing to recall that status inconsistency measures at the nation-state level were typically inconsequential in predicting the violent behavior of nations. Why should aggregate measures have greater predictive power than the nation-state measures? The results pertaining to alliance structure are even more confusing. Only under very special conditions are any reasonable correlations obtained between alliances and amounts of war. More significantly, exceedingly discrepant results are found for different measures of systemic polarity - discrepant not only in terms of magnitude but also of direction. One can only conclude that the systemic picture, in contrast to the nation-state puzzle, is blurred and sketchy" (Zinnes, 1980: 358-9).
"Mesquita found that the change in the tightness of the alignment structure is a very powerful predictor for the occurrence and duration of wars, especially for major power wars" (Eberwein, 1981: 29).
After surveying the literature on polarity and war, Cashman (1993: 242) states: "Perhaps the only thing we can conclude with any certainty form these studies is that the relationship between polarization and wars is neither linear nor constant over time".
And Maoz (1995: 45) concluded: "... at the very least, we must conclude that there is little systematic evidence that is robust and stable suggesting that alliance patterns, polarity, or power distribution are related to war".

Ad alliance
"Wars appear to rise independently, but to spread contagiously, through alliance structures, to neighbors, and otherwise" (Wilkinson, 1980: 118).
"Ceteris paribus, alliances are more frequently followed by war than by peace" (Vasquez, 1993: 312).
Levy (1981) finds that the formation of alliances over the last five centuries (but not in the nineteenth century) has generally been soon followed by war, but that most wars have not been preceded by alliances. He suggests that the tendency for wars to follow alliances may be explained by the tendency of states to form alliances for protection whenever they perceive that the probability of war is high, so that the causal linkage is from the anticipation of war to alliance formation and not from alliance formation to war (Levy, 1989: 236).
"Alliances seem to increase the propensity of escalation if at least one major power is involved. If one or more nations join an ongoing dispute, the probability of this escalating to war increased considerably. This finding holds for both major and minor power disputes" (Eberwein, 1981: 28).
A systematic survey of historical societies over 2,000 years reveals "no relationship worth mentioning between frequency of warfare and... the organization of a web of defensive alliances" (Naroll et al., 1974: 332-33, 339).
"Some statistical studies suggest that specific alliances are belligerent bonds or 'war welds' that tend to recur over time" (Beer, 1981: 269).
"One of the few patterns that researchers seem to be able to agree on is that once wars begin, the presence of alliances serves to expand war activity throughout the system to more states. Alliances act as contagion mechanisms for war, but alliance is not the only contagion mechanism. (Geographic contiguity is another.)" (Cashman, 1993: 245).
"Davis et al. demonstrate that wars starting initially between two nations tend to have an 'epidemic' effect, i.e. lead to an expansion of war. The entry of major powers into wars instead is randomly distributed over time and does not accord with the epidemic principle... What triggering effect do alliances have upon belligerent controversies? More precisely, to what extent do the obligations incurred in alliances become an automatic mechanism for entry into war? Sabrosky's analysis shows that, by and large, this mechanism does not come into effect. In almost two-thirds of the cases in question, nations did not fulfil their alliance obligations. They remained neutral when an alliance partner became involved in a war... Alliances have a positive effect on the extent of participation in war" (Eberwein, 1981: 30).

Ad arms races
"[T]he evidence, as far as it goes, is that only a minority of wars have been preceded by arms races" (Richardson, 1960: 74).
Having reviewed the arms race literature, Hower & Zinnes (1989: 10) conclude: "Thus, there is evidence supporting both the arms race-leads-to-war or the arms race-prevents-war hypotheses. It all depends on how variables are operationalized and what type of test is used".
The evidence on the question whether arms races contribute to an increased likelihood of war is mixed (Levy, 1989: 238).

Ad contagion and learning
Levy (1982) finds no relationship between war and contagion, leading him to conclude: "... these findings are consistent with a long line of empirical research going back to Sorokin (1937). These earlier studies have repeatedly demonstrated the absence of contagion... in the outbreak of war" (p. 579) (Hower & Zinnes, 1989: 4).
"To what extent is the memory of past wars a relevant factor in accounting for war? Garnham finds little evidence in support of the war-weariness hypothesis he tested for dyads involving major powers... Singer and Small report that previously victorious nations are more likely to engage in new war ventures. But this is only true for the 20th century. The opposite holds for the 19th century. Yet the longer the war abstention, the more severe will be the following war, in which these nations are involved" (Eberwein, 1981: 29).
Referring to the findings of Richardson (1960), Zinnes (1980: 347) concludes that "the longer peace exists between two enemies, the less likely they will fight each other again", as if a slow process of forgetting and forgiving went on".
On the other hand, Hower & Zinnes (1989: 4) summarize various studies that find no significant correlations between various measures of the magnitude or severity of war and the length of time to the next war, casting a shadow over the learning model argument.
Wilkinson (1980: 118) concluded: "Having previously fought as enemies seems to be an incitement to fight again, though the effects of past wars diminish with time".
And Cashman (1993: 156) concludes that empirical studies of the war weariness hypothesis have once again produced a "mixed bag of results" which is considerable less than compelling.
Maoz (1995: 50) reports: "Maoz found that the time span between the termination of one dispute and the outbreak of another tends to increase with the extent to which the victor of the prior dispute is capable of imposing its will on the loser. The less decisive the political and military outcome of the previous dispute, the more likely is it to recur over a short period of time. Also, the winner of the previous dispute - rather than the loser - tends to be the initiator of the subsequent dispute".

Ad internal/external conflict
The internal conflict/external conflict hypothesis has also been subjected to systematic empirical research. Most quantitative studies found that there exists no relationship between the two. More sophisticated studies, which attempt to control for other variables such as the type of regime, have found some positive but relatively weak relationships between internal and external conflict. "This discussion draws attention to several significant puzzles in the literature on societal-level sources of international conflict. One is the gap between the general conclusion of large-N correlational studies that there seems to be little connection between domestic and foreign conflict and the findings of many individual case studies that the scapegoat motivation has an important impact in the process leading up to many wars" (Levy, 1989: 273-4).
"Domestic conflict does not generally lead to foreign conflict behavior" Eberwein (1981: 24) reports, but "the intervention of third parties in ongoing disputes consistently increases the probability of their escalating to war... Domestic unrest seems to some extent to invite other nations to interfere. But that relationship is far from simple: it may either be an occasion to intervene in the domestic affairs of a nation or an opportunity to attack a nation given its present weakness" (27).
Sorokin's (1937: 487-89) research on European major wars led him to conclude that, in general, international and domestic violence each ran a course independent of the other, without either positive or negative association.
On the other hand, Beer (1981: 59) reports some studies which found positive correlations between internal and external violence.
"If we compare these three studies, we see that each suggests that a relationship exists between variables that measure internal conflict and variables that measure violent international behavior when the internal conflict measures are taken together with other attributes of nations, such as governmental structure (Wilkenfeld), population diversity (Hazlewood), or demands and instability (Bobrow et al.)" (Zinnes, 1980: 343; italics in original).
Van der Dennen (1981: 165-175) also found that the quantitative literature on the domestic/foreign conflict linkage yielded inconclusive and ambiguous results, although there might be some evidence that the governmental structure of a state (type of regime such as personalist, centrist and polyarchic) is important in predicting the relationships between foreign and domestic conflict behavior (e.g., Wilkenfeld, 1968, 1973).
Cashman (1993: 151) similarly concludes his survey of the scapegoat theory (or 'diversionary war theory') that it is "difficult to use these studies as evidence to either confirm or disconfirm a theory that domestic conflict causes external conflict".

Ad economic factors
The belief that economic transactions lead to peace is not supported (Vasquez, 1976).
"There has been no convincing empirical test of the liberal hypotheses [especially, free trade] regarding the causes of war... some features of liberal economic structures tend to reduce the level of international conflict, other features are destabilizing" (Levy, 1989: 261-2).
Eberwein (1981: 28): "Trade, presumed to establish cooperative links between nations, does not prevent the escalation of military confrontations where at least one major power is involved: the reverse seems to be the case".
"Economic causes seem to have figured directly in less than 29 percent of the wars since 1820... Relative wealth and poverty seem to have had very little influence... The influence of all these factors together has been less than that of quarrels over territory which may be more political than economic" (Van der Dennen., 1981: 155). Richardson (1960: 197-204) emphasizes (non-economic) revenge and retribution as possible motives for violence.
"Some [studies] suggest that war is more likely between groups of equal size and resources... Some find that countries that are unequal are more likely to go to war... Others find that general foreign conflict, including both war incidence and casualties, are unrelated to differences in size and economic development... GNP per capita is not substantially related to most indicators of foreign conflict... [Haas (1974)] finds that whether states have great internal discrepancies between rich and poor, or narrow differences, seems irrelevant for amounts of international violence" (Beer, 1981: 171-73).
"Neither economic weakness nor prosperity seem to prevent war" (Cashman, 1993: 136).
Maoz (1995: 48), on the other hand, concludes that "the probability of dyadic conflict declines with the level of economic wealth and the rates of growth in the dyad. Wealthy and rapidly-growing states are far less likely to fight each other than expected by chance alone". These wealthy states tend, however, to be Western democracies (vide infra).

Ad contiguity/proximity/territory
Contiguity had already been singled out by Richardson (1960): "States have tended to become involved in wars in proportion to the number of states with which they have common frontiers. Contiguity has been an important factor in war during this period... [but]... The actual occurrence of war has been far less than would be expected from the opportunities for war presented by geographical contiguity" (Van der Dennen., 1981: 155; cf. Wilkinson, 1980: 118).
"In short, contact and contiguity appear to be relevant factors in predicting international violence" (Zinnes, 1980: 349).
Beer (1981: 53): "a number of studies report a strong general relationship between the number of wars and the number of borders with other nations".
Mandel (1980) finds that border disputes are more likely to develop between states of equal power and to occur with greater frequency and severity between states with lower technological development (Hower & Zinnes, 1989: 2).
"... contiguity is the single most important predictor of dyadic dispute involvement" (Maoz, 1995: 46).
Vasquez (1993) holds that border disputes and other territorial issues are the main causes of war in the last five centuries. Vasquez (1993): "Of all the possible issues states can fight over, the evidence overwhelmingly indicates that issues involving territory, especially territorial contiguity, are the main ones prone to collective violence. It is remarkable how many interstate wars (regardless of type) from 1816 to 1980 involve states that are neighbors (i.e. contiguous or indirectly contiguous because of water). These findings provide more than a strong hint that war is intimately connected with struggles over contiguous territory. In this analysis, territory is seen as a general underlying cause of war" (Vasquez, 1993: 293).
This general observation seems to imply that the more states in the international system the more wars in the system.
Thus, most interstate wars are fought or begin between neighbors. Vasquez (1995) presents a territorial explanation of the relationship between contiguity and war, and juxtaposes it with the proximity and interaction explanations. Boundary conflicts constitute a major source of war, he contends, but once boundaries are accepted, peace can reign.
Other explanations of the observed correlation have been opportunity (Russett, Wesley; Starr & Most, Siverson & Starr), uncertainty (Midlarsky), 'willingness' (Starr), proximity and the logistical 'loss of strength gradient' (Boulding, Garnham, Diehl, Starr & Most). Perhaps it is appropriate to conclude, as Starr & Most (1978: 445) do, that "borders may not cause wars, but it seems reasonable to suggest that they create certain structures of risks and opportunities in which war is likely" (quoted in Cashman, 1993: 145).
Additional literature on contiguity and war includes: Richardson, 1960; Wallensteen, 1981; Small & Singer, 1982; Diehl, 1985, 1991; Luard, 1986; Gochman, 1990; Singer, 1990; Bremer, 1992; Goertz & Diehl, 1992; Vasquez, 1993, 1995; (and Van der Dennen [1995] for preindustrial - non-state - societies).

Ad democratic peace
Rummel (1979: 277) offered the 'joint freedom proposition': Libertarian systems mutually preclude violence. War will occur between states only if at least one is nonlibertarian.
"Some contemporary research suggests that communities with military leadership are more likely to go to war than civilian-dominated societies" (Beer, 1981: 284).
Hower & Zinnes (1989: 5) review the studies on regime type and international conflict and note that the evidence is conflicting. "One piece of evidence, however, is consistently supported in many tests. Democratic or libertarian regimes do not engage in violent conflict with one another".
"Although democracies have fought in wars as frequently as have non-democratic states, they almost never fight each other... This absence of war between democratic states comes as close as anything we have to an empirical law in international relations... This finding is particularly interesting because it runs counter to many of the leading theories of international conflict and war" (Levy, 1989: 270).
Also Cashman (1993: 129) notes that Rummel's 'joint freedom proposition' "seems to have garnered overwhelming support among scholars".
And Maoz (1995: 48) states "Perhaps the most important factor that was found as an inhibitor - rather than facilitator of dyadic dispute involvement was the joint democratic regime of the members of the dyad".
Interestingly, neither Vasquez (1976), Zinnes (1980), Eberwein (1981), nor Van der Dennen (1981) mention literature on the democracies-don't-fight-each-other finding (or 'Joint Freedom Proposition'), which now turns out to be both a robust and a rather recent finding (to the best of my knowledge it was reported for the first time by Small & Singer [1976] and Rummel, [1979, 1983], after initial suggestions by Babst [1972]), Jencks [1973], and Haas [1974]).
It, thus, appears that where a pair of states shares a common democratic political culture it exerts conflict dampening impact that overrides ethnic, linguistic, or religious factors (Henderson, 1998).
Surprisingly, the no-warfare effect also holds true among band-level, tribal, and other non-industrial societies having participatory (the nearest equivalent of democratic) as opposed to hierarchical social structures in the Human Relations Area Files (Russett, Ember & Ember, 1993).
Additional literature on democracy and peace includes: Babst, 1972; Small & Singer, 1976; Rummel, 1979, 1983; Chan, 1984, 1997; Weede, 1984; Maoz & Abdolali, 1989; Bremer, 1992; Maoz & Russett, 1992, 1993; Russett et al., 1993; Everts, 1995; Ray, 1995; Russett, 1995; Oneal et al., 1996; Oneal & Russett, 1997; and Mousseau, 1998.

Ad ethnic diversity
Ethnic diversity and heterogeneity are not mentioned by Geller & Singer as polemogenic factors. Yet, there is evidence that these factors are important predictors of the initiation of violent conflict.
"Hazlewood finds that population diversity and ethnic diversity together with Rummel's dimension of turmoil relate very strongly with Rummel's war dimension" (Zinnes, 1980: 343).
"Midlarski and Thomas show that societies dominated by religious and bureaucratic groups engage in belligerent controversies with greater vehemence than others. If military groups are dominant, the wars last longer" (Eberwein, 1981: 29).
Earlier studies by Rummel (in the nineteensixties), and Haas (1974) found that the heterogeneity in composition of a population is consistently associated with the frequency of wars, military actions and foreign conflict casualties. Countries with many different ethnic groups, language communities, nationality groups and religious and racial groups enter wars more often than homogeneous polities (Van der Dennen. 1981: 158).
Rummel (1997) finds that two simple measures, the number of ethnic groups and the number of religious groups a state has, are related to its collective (internal and external) violence: the more groups the more violence.
Vanhanen (1999) and Ellingsen (2000) also found significant correlations between the number of ethnies and domestic conflicts and violence. Vanhanen concludes his extensive research:

The results indicate that, as hypothesized, ethnic conflicts have emerged in practically all ethnically divided societies, which implies that the question is of a universal human phenomenon. Everywhere in ethnically divided societies people belonging to the same ethnic group tended to align with their relatives in social and political interest conflicts. They are following the imperatives of ethnic nepotism, their unconscious impulse to support their relatives in conflicts because it has furthered their own inclusive fitness. Therefore, ethnic conflicts in various forms are very common in ethnically divided societies. I came to the conclusion that ethnic nepotism provides an ultimate evolutionary explanation for the appearance of ethnic conflicts across all cultural boundaries. Furthermore, if the ultimate cause of ethnic conflicts is in human nature, in our evolved behavioral predispositions or epigenetic rules, we have to accept the conclusion that ethnic conflicts are extremely persistent, if not absolutely inevitable, in ethnically divided societies (p. 231).

Using data for two different types of domestic conflict Ellingsen (2000) concludes that multiethnicity does increase the propensity of domestic violence, although less so for large-scale conflicts. However, the country's political regime and socioeconomic level are more important in predicting domestic conflict.

Ad population structure
As we have ascertained, demographic variables are hardly or not at all correlated with the warring behavior of states in Geller & Singer's analysis. Until recently, however, there were hardly any investigations regarding specific subpopulations, especially the proportion of young males in the demographic pyramid.
Vasquez (1976) concluded that there may be a strong relationship between the demographic characteristics of nations and violence between nations.
Similarly, Eberwein (1981: 23) stated that "the level of conflict behavior sent and received varies positively with population size. This holds for individual nations as well as for pairs of nations. Given the moderate relationships, population size is perhaps a necessary but by no means a sufficient condition accounting for the amount of conflict behavior sent/received".
"Size of population, measured both in absolute terms and relative to resources, appeared usually as only a background factor, or a minor irritant" (Beer, 1981: 169).
Mesquida & Wiener (1996) recently argued that collective aggression (meaning insurrection, rioting, interstate war, civil war, ethnic cleansing, genocide) may be viewed within an evolutionary or behavioral ecology context as an attempt by (coalitions or gangs of) young males to acquire otherwise unobtainable resources and thereby gain inclusive fitness (cf. Chagnon, 1990; Daly & Wilson, 1987; Low, 1993; Manson & Wrangham, 1991; Van der Dennen, 1995; Wilson & Daly, 1985; Wrangham & Peterson, 1996). Wilson & Daly coined the term 'young male syndrome' for the fact that homicidal violence and property crimes, in a wide range of - if not all - societies, are committed largely by young males. Van der Dennen (1995) argued that preindustrial societies may live in peace only when they are able to solve the external problem to be 'left in peace' by other societies, and the internal problem of controlling the young, male, 'brave warriors' (especially in societies where warlike exploits are rewarded by status, prestige, and/or access to nubile women).
Mesquida & Wiener examined the influence that different population age structures have on the development of violent conflicts and concluded that the presence of a relatively large number of young men makes coalitional aggression more probable, particularly when resources needed to attract a mate are insufficiently available or poorly distributed. Such explanation, they claim, may account for the existence of tribal as well as modern warfare.
This study thus argues that the age composition of the male population should be regarded as the critical ecological/demographic factor affecting a population's tendency toward peace or violent conflicts (as hypothesized by the French polemologist Bouthoul [1951 et seq.] - who called it a "structure explosive" - and historian Moller [1967]). Analyses of interstate and intrastate episodes of collective violence since the 1960s indicate the existence of a consistent correlation between the ratio of males 15 to 29 years of age per 100 males 30 years of age and older, and the level of coalitional violence as measured by the number of reported conflict related deaths.

Ad misperceptions and miscalculations
The reasons why nations resort to war or military force to settle disputes, Singer (1980: 350) argued, are the foreign policy elites' erroneous predictions about the consequences of their actions. Assuming that no single foreign policy decision-maker is even remotely infected with 'war addiction', the fundamental issue seems to "lie not so much with the stupidity or cupidity of the foreign policy elites as with their ignorance".
Rejecting a purely 'rational choice' orientation, Singer (1995: 229) prefers a hybrid model of decision processes that might be described as 'psycho-political', accepting some of the premises of the bounded rationality approach, but only at the individual level of aggregation. "The psycho-political model begins, then, with the premise that most of the individuals involved in the foreign policy/national security process - across most cultures and over the past two centuries - are moderately competent problem solvers and typically bring rationality to bear in solving their own problems. But it promptly takes cognizance of the forces of the extra-rational...".
Eberwein (1981: 27) implicates perceptions in escalation to military confrontations: "From a number of findings one can plausibly infer that perceptual factors may play a considerable role in triggering the escalation process. These perceptions may either arise from the given conflict situation or be related to longer term grievances between nations. Suggestive for the former is Leng and Goodsell's finding that military confrontations escalate within a relatively short period of time to war, i.e. 14 days, or not al all. A finding one can relate to the latter case is reported by Sabrosky. Rigidity in perceptions prevailing between the revisionist Austro-German alliance on the one hand and the status quo oriented British, French, and Russian triad may finally have triggered World War I".
Cashman (1993) attempts to build a 'model' of the typical scenario for war. In this model (mis)perceptions play a crucial role:

Perhaps the best way to conceptualize the role of phenomena at the systemic, dyadic, and nation-state levels of analysis is as stimuli that trigger individual perceptions (and misperceptions), which then guide policy decisions about war and peace.
Misperceptions about the opponent's actions, his intentions and capabilities- and hence about the degree of threat to one's own security - may create the conditions necessary for the onset of a crisis. Once the crisis begins, these misperceptions may accelerate and exacerbate the level of tension. Particularly important are the combination of an overperception of the rival's hostility and treachery and an underperception of both the capabilities of the rival and the amount of risk involved. Unwarranted confidence in the ability to compel one's opponent(s) to back down short of war, or of one's ability to defeat the adversary with little cost if war does begin, seem to be an important part of the picture. A significant aspect of this is the perception that other states are not committed to assisting one's opponent or are unwilling or unable to carry out their commitments... Ultimately, decisions for war are likely to be predicated on the assumption that war is inevitable, or that it can be waged successfully, or, at the very least, that it can be waged at an acceptable level of cost (Cashman, 1993: 281-2).

Postlude and conclusions
Hower & Zinnes (1989: 11) concluded their summary of the macroquantitative research with the following remarkable sentence: "... if there is one overarching conclusion to be drawn from the foregoing pages it is a bold questionmark". They were undoubtedly grosso modo right, but some macroquantitative findings are very probably here to stay.
All in all, it looks like we are discovering a number of rather simple indices of collective violence: territorial contiguity, the number of ethnic and religious groups, and the number of young males in the population.
Also power, especially totalitarian power, has been consistently related to warfare and other forms of collective violence such as democide and genocide (Rummel, 1979 et seq.). As Rummel is fond of saying (as a variant of the 'power corrupts' theme): "power kills; absolute power kills absolutely".
It also appears that sharing a common democratic political culture - the very opposite of totalitarian power - is an excellent and potent force for peace, although Singer (1999: 5) warns against too much optimism in this respect: "[I]f humans indeed normally prefer relatively hierarchical if not downright authoritarian regimes [Somit & Peterson, 1998], if the putative increase in democracies is little more than a momentary aberration, and if the democratic peace proposition rests on some questionable coding and scaling criteria, we might want to return to some of the more promising strategies by which to move to a less war-prone world".
Finally, the empirical research also shows that no single factor makes war inevitable. "As David Singer is fond of saying, there are many exits along the road to war" (Vasquez, 1993: 196).

Coda in a minor key
It is clear that a macroquantitative approach to war and peace covers a limited domain: it can only occupy itself with hypothetical polemogenic factors which are, in principle, quantifiable. If all possible polemogenic factors could be ordered on a single dimension from purely voluntaristic (or strategic) to purely deterministic (or cataclysmic), it would be obvious that the factors situated at the deterministic pole are the ones most easily quantifiable, so the macroquantitative studies tend to cluster at the deterministic pole. But these factors will at the same time tend to be the culmination of impersonal (structural) forces originating in the dynamics of the international system: power balances or imbalances, unequal economic growth, conjunctural waves, etc. etc. The fact that it is, ultimately, human beings who decide to wage war and that it is human beings who fight in wars sometimes threatens to be obscured in this abstract perspective.
This, in turn, leads to the age-old question whether, and of which impersonal forces the human being is a puppet. The fact that, all in all, so few unambiguous 'forces' have been identified in the macroquantitative research may throw some light on this question. Apparently, ever since Thucydides presented his explanation of the Peloponnesian War - the growth of the power of Athens, and the fear that this inspired in Sparta - there has been next to nothing to be added.
For many readers the results of these many hundreds of macroquantitative studies may be very disappointing. Depending upon data bases, time periods, operationalizations and indices, theories and hypotheses, and methods and procedures, the outcomes of these studies are, for the most part, heterogeneous, ambiguous, not particularly robust, conflicting and sometimes even contradictory. About very few theoretical frameworks and potential polemogenic factors (predictors) some consensus or communis opinio exists.
Started as 'unsophisticated number crunching' (as it was derogatorily called), the macroquantitative research now, partly because of the paucity of unambiguous results, is in danger of degenerating into method-fetishism (much sharpening of tools but very little agriculture). The Journal of Conflict Resolution is unreadable for many peace researchers who were not educated in higher mathematics, multivariate statistics, sophisticated modeling, and the intricacies of chaos theory. It is sometimes very difficult to escape the impression that the debates among the methodological 'whizzkids' are interesting only for themselves, and that even some commitment to the original problem - the explanation or explication of why people/societies/states wage wars - is obfuscated by personal 'claims to fame' and petty, acrimonious altercations. A profound dichotomy within peace research may be imminent: the 'hard-cored' quantitative, narrow-minded superexperts on the one hand, and the 'soft sector' of gratuitous, 'academic' babblers, not hampered by any relevant knowledge, on the other hand.
It may even be argued that the results of the macroquantitative studies is out of all proportion to the arrogance with which the macroquantitative method is claimed (at least by some researchers) to be the only scientific one (in the sense of generating and empirically testing falsifiable propositions, derived from solid theoretical frameworks, resulting in a cumulative body of explanatory and predictive knowledge). At the moment, however, knowledge is still largely confined to isolated islands of theory rather than comprehensive, multivariate theoretical frameworks, as Diehl (1995: 207) acknowledges. He offers an alternative framework: the study of 'rivalries' as a broader orientation than 'wars'. And probably, at the same time, we should be aware that the role of pure chance and random factors in war may be much greater than previously assumed. In the words of Bremer (1995: 18) "war may stem from the improbable concatenation of weak forces rather than from the operation of a few strong forces".
Eberwein (1981: 32) stated a considerable time ago: "Sullivan and Siverson are certainly right in stating that 'the systemic level is not particularly helpful in explaining how decision-makers choose among alternatives once an occasion of conflict arises'. There certainly are theoretical justifications to study international conflict at different levels of analysis" and "... the danger remains that quantitative research may become out of tune with reality".
Now that, anno 1999, interstate wars are being increasingly replaced by horrible genocides, such as in Cambodia, Rwanda and Burundi, and equally horrible massacres and 'ethnic cleansing' within a context of ethnonationalistic and civil wars, such as in the former Yugoslavia, Eberwein's observation may contain more than a kernel of truth. It is a sobering, if not depressing thought.

The following account of the causes of civil war is based on Why They Fight by Byman & Van Evera (1998):

Byman & Van Evera's (1998: 23) main thesis is that, in recent years, civil war has replaced international war as the dominant form of war and has nearly replaced it as the only form of war. The Persian gulf war of 1990-91 has been the only major old-fashioned interstate war since 1989. The war problem is now largely synonymous with the civil war problem.
While the post-cold war world is hardly tranquil, the number of active wars worldwide has fallen markedly since 1989. The number of civil wars has fallen significantly - eleven states experienced civil wars in 1996 compared with seventeen in 1989 - and international conflict has nearly vanished, at least for now.
Byman & Van Evera hypothesize that seven causes of civil violence stand out in importance (that is, in their potence and prevalence) in the cold war and post-cold war periods. They are: 1. The collapse of post-Second World War empires; 2. A lack of regime legitimacy; 3. State weakness; 4. Communal hegemonism; 5. Revolutionary ideology; 6. Aristocratic intransigence; 7. Superpower proxy wars (Byman & Van Evera, 1998: 2).

The Collapse of Empire

The first hypothesis is that the collapse of empire, embodied in the 1991 collapse of the Soviet Union and Yugoslavia, generated a number of recent wars. Imperial collapse spawns successor states that are primed for civil war. These states often have illegitimate regimes, commingling hostile populations, and contested borders. Moreover, no 'rules of the road' define the rights and responsibilities of the former metropole or other external powers in the former imperial zone; hence outsiders often intervene to claim a sphere of influence or to disrupt another power's sphere, sparking civil conflict. These dangers caused considerable carnage after many past imperial collapses. Massive killing occurred after the British, French, Ottomans, and Portuguese withdrew from their empires earlier in this century. Chaos likewise followed the collapse of ancient empires from China to Rome (Fieldhouse, 1966; Doyle, 1986; Snyder, 1991; Rudolph & Good, 1992; Ladinsky, f.c.). The collapses of the Soviet and Yugoslav empires resulted in massive bloodshed for similar reasons: they spawned frail successor governments that ruled hostile intermingled populations and often faced destructive interference by the former metropole (Byman & Van Evera, 1998: 2-3).

A Lack of Regime Legitimacy

A lack of regime legitimacy caused many civil wars both during and after the cold war. Civil-war-causing crises of government legitimacy have sprung from two main causes: the growth of restive middle classes in authoritarian states, and a lack of regime accountability, which leads in turn to corrupt or incompetent state policies. Such legitimacy crises cause violence both when a regimes tries to regain legitimacy and when a regime tries to suppress dissent - a dilemma we label 'the reform trap' (the 'reform trap' is similar to the 'King's dilemma' analyzed definitively by Huntington [1968: 177-91]).
To regain legitimacy, besieged authoritarian regimes may move to democratize. Democratization, however, can spawn conflict as old elites inflame and manipulate hatreds in an effort to gain or remain in power. Democratization can also spawn conflict if majoritarian democratic rules are adopted to cast all power to tyrannical majorities, driving oppressed minorities to rebel. Finally, democratization can spawn war by giving political space to hardened secessionist groups that cannot be appeased by power-sharing and instead exploit democratic freedoms to organize for war. Democracy causes peace between mature democracies, but democratization is a dangerous cause of war in multiethnic authoritarian states; hence crises of regime legitimacy that trigger democratization are also causes of war.
Alternately, regimes that eschew democratization and instead attempt to suppress dissent often 'hunker down' relying on an increasingly narrow core to defend them. Hunkering down may enable a regime to survive a short-term challenge to its legitimacy, buy it can cause violent resistance by excluded social groups and does little to ease the original legitimacy crisis.
This second war-cause grew more prevalent with the cold war's end, as the collapse of communism and the triumph of free-market liberalism discredited repressive and statist ideologies throughout the world. Regimes that embraced these ideologies lost legitimacy and often undertook destabilizing reforms (Byman & Van Evera, 1998: 3-4).
Hunkering-down behavior fueled many current disputes, including both those that began before and after the end of the cold war. In Rwanda and Burundi, for example, regimes have relied increasingly on ethnic kinsmen with no pretense of including others. Similarly, regimes in Burma, Chad, El Salvador, Ethiopia, Guatemala, Iraq, Kenya, Lebanon, Liberia, Pakistan, the Philippines, Somalia, Uganda, and Zaire have relied more and more on one particular ethnic, religious, or tribal group or one sector of society, usually the military, to govern after challenges to their legitimacy arose. This reliance has created a self-sustaining cycle: as dissent increases, regimes fall back more and more on 'trusted' individuals from the same communal group or sector; this reliance in turn increased resentment among the excluded groups and engendered charges of corruption and favoritism (Byman & Van Evera, 1998: 35-6).
Hunkering down often follows democratization when regimes face a legitimacy challenge. Efforts to democratize lead to a growing governability crisis, which in turn leads the regime to hunker down and abort its democratization experiment. This 'reform trap' generates further dissent and leads to civil war.
The democratization process and subsequent hunkering down often share the same cause: a regime attempting to stay in power while gaining popular support for painful reforms. Algeria, for example, began its democratization process after food riots in 1988, and Sierra Leone allowed elections after it adopted unpopular economic reforms under IMF pressure. The pain engendered by the economic reform process, however, often leads the regime to lose the elections, causing it to hunker down in order to stay in power.
The 'reform trap' is a common factor in civil wars throughout the world today. Algeria, Burma, Pakistan, Somalia, and Tajikistan all initiated hesitant democratization and the, when the results were not to the liking of the regime or a powerful group, chose instead to hunker down and ignore the elections (Byman & Van Evera, 1998: 35-6).

State Weakness

A third leading cause of civil conflict is state weakness. The spread of modern small arms among Third World populations has weakened governments' relative position vis-à-vis potential civil opponents over the past few decades. The end of superpower aid to beleaguered governments has further weakened many authoritarian regimes since 1989. This weakening, in turn, has made it easier for groups to challenge governments by force (Byman & Van Evera, 1998: 4).
A decrease in a state's coercive ability fosters civil conflict in two ways. First, if the state is weak, restive ethnic groups or other threats to peace are no longer reassured or deterred from organizing. Predatory groups plot war because they are less deterred by fear of state repression. This alarms other groups who then mobilize for war in self-defense, taking security into their own hands because they no longer trust the central state to provide it. Thus as the weakness of the state became apparent in the early 1970s, various communal groups began forming militias for self-defense.
The second, related, impact of decreased state coercive ability is an inability to defeat groups committed to violence. Even unpopular regimes can stamp out potentially violent opposition when they have enough resources to overwhelm the insurgents directly, arrest their leaders, or otherwise interfere with group organization before the violence becomes widespread (Byman & Van Evera, 37-8).
Until recently, only governments had access to large quantities of modern weapons, giving them a tremendous military advantage over insurgents. Today, however, small arms are available in many states to anyone with a little money. Thus, the state's monopoly on violence is more easily challenged by armed citizens. This cause of weak states is self-perpetuating: a weak state cannot stop the flow of weapons across its border, and this flow in turn further weakens the state (Byman & Van Evera, 1998: 39).

Communal Hegemonism

A fourth cause is communal hegemonism - the aspiration of ethnic, religious, clan or class groups for hegemony over other groups. Violence often results when hegemonistic groups seek to impose their way of life on others, particularly on peoples that have a well-developed group identity of their own. Peace is strongest when groups adopt a live-and-let-live stance toward others.
The war-causing effects of communal hegemonism are catalyzed by the first three causes. Submerged hegemonistic groups are freer to go on a rampage, and are more likely to provoke defensive violence by others, when an empire collapses, regime legitimacy declines, or central power weakens (Byman & Van Evera, 1998: 4).
Some ethnic groups are not content with equality or social recognition. Instead, they demand that their language be the only official language, their religion be enshrined as the only permitted faith, and their culture and customs be followed by the rest of society. Such hegemonic beliefs can come from many sources, ranging from culture to ideology. What these sources share is a common commitment to the idea that one particular cultural system is more worthy than others.
The quest for communal hegemony - and resistance to it - is a major source of conflict in the world today. Concerns about hegemony, both real and imagined, contributed to the post-cold war conflicts that broke out in Georgia, Indonesia (with the Aceh), Russia, and the former Yugoslavia. This follows a pattern common before the cold war. Religious, tribal, or ethnic hegemonism fed communal violence in Afghanistan, Angola, Burma, Burundi, Chad, Ethiopia, India, Iraq, Lebanon, Liberia, Pakistan, Rwanda, Somalia, South Africa, Sri Lanka, Sudan, and Turkey.
Hegemonism causes conflict in to fundamental ways. First, groups often strive for preeminence by attacking, intimidating, and destroying other communal groups. Bengalis in Bangladesh, Hutu interahamwe murderers in Rwanda, and chauvinistic Hindus in India are but a few examples of groups that used violence in their attempts to ensure communal hegemony (e.g., Gray, 1994; De Waal & Omaar, 1995). The cause of conflict - the will to dominance - is readily apparent. When other ethnic groups refuse to submit, the hegemonic group imposes its will by force.
Second, hegemonism causes more conflict when it incites the defensive concerns of other communal groups (Byman & Van Evera, 1998: 39-40).

Revolutionary Ideology

Fifth, revolutionary political ideas have caused civil war by leading groups to adopt extreme goals and tactics that precluded peaceful compromise with others. Marxist-Leninist ideas have fueled civil wars worldwide since 1917, by leading movements of the disenfranchised to embrace extreme communist political programs. Muslim revolutionary movements have likewise triggered civil violence in the Arab world and South Asia in recent years. By delegitimizing Marxist-Leninist political ideas, the Soviet collapse vastly reduced the prevalence of this important cause of war (Byman & Van Evera, 1998: 4.

Aristocratic Intransigence

Sixth, aristocratic intransigence - the refusal of elites in some steeply stratified states to share power and wealth - has triggered several recent civil wars (for example, in Nicaragua in the 1970s and in El Salvador and Guatemala in the 1980s). This danger has diminished with the cold war's end, as democratic norms has spread to conservative elites and as these elites have lost unqualified U.S. Backing. This loss has forced them to moderate their behavior (Byman & Van Evera, 1998: 4-5).

Superpower Proxy Wars

Seventh, the superpower competition for influence in Third World states was a major cause of civil conflict during the cold war. A number of Third World states (Afghanistan, Angola, Cambodia, Nicaragua, El Salvador, Guatemala, and Ethiopia, among others) became cold-war battlegrounds when one or both superpowers backed proxies or directly intervened to support client groups. The cold war's end stopped this competition and slowed or ended the civil violence it spawned (Byman & Van Evera, 1998: 5).

All seven causes operated both during and after the cold war, but some were more prevalent during the cold war while some have been more prevalent afterward. Specifically, the first three causes listed above (collapse of empire, regime illegitimacy, and state weakness) grew more abundant after 1989, the fourth (communal hegemonism) stood roughly constant across both periods, and the last three (revolutionary ideology, aristocratic intransigence, and superpower competition) abated sharply after 1989. The abatement of these last three causes largely explains the net decline in civil violence since 1989 (Byman & Van Evera, 1998: 5).


Babst, D. (1972) A force for peace. Industrial Research, April, pp. 55-8.
Beer, F.A. (1974) How Much War in History? Definitions, Estimates, Extrapolations and Trends. Beverly Hills: Sage.
Beer, F.A. (1981) Peace Against War: The Ecology of International Violence. San Francisco: Freeman.
Blainey, G. (1973) The Causes of War. New York: Free Press.
Bobrow, D.B. et al. (1973) The impact of foreign assistance on national development and international conflict. Journal of Peace Science, 1, 1, pp. 39-60.
Bouthoul, G. (1951) Les guerres. Eléments de polémologie. Paris: Payot.
Bouthoul, G. (1970) L'infanticide différé. Paris: Librairie Hachette.
Bremer, S.A. (1992) Dangerous dyads: Conditions affecting the likelihood of interstate war, 1816-1965. Journal of Conflict Resolution, 36, 2, pp. 309-41.
Bremer, S.A. (1995) Advancing the scientific study of war. In: S.A. Bremer & T.R. Byman, D. & S. Van Evera (1998) Why They Fight: Hypotheses on the causes of contemporary deadly conflict. Security Studies, 7, 3, pp. 1-50.
Cusack (Eds.) The Process of War: Advancing the Scientific Study of War. Amsterdam: Gordon & Breach, pp. 1-33.
Cashman, G. (1993) What Causes War? An Introduction to Theories of International Conflict. New York: Lexington Books.
Daly, M. & M. Wilson (1987) Homicide. Hawthorne NY: Aldine de Gruyter.
Chagnon, N.A. (1990) Reproductive and somatic conflicts of interests in the genesis of violence and warfare among tribesmen. In: J. Haas (Ed.) The Anthropology of War. Cambridge: Cambridge University Press, pp. 77-104.
Chan, S. (1984) Mirror, mirror on the wall... : Are the freer countries more pacific? Journal of Conflict Resolution, 28, 4, pp. 617-48.
Chan, S. (1997) In search of democratic peace: problems and promise. Mershon International Studies Research, 41, pp. 59-91.
Crescenzi, M.J.C. & A.J. Enterline (1999) Ripples from the waves? A systemic, time-series analysis of democracy, democratization, and interstate war. Journal of Peace Research, 36, 1, pp. 75-94.
De Waal, A. & R. Omaar (1995) The genocide in Rwanda and the international response. Current History, 94, 591, pp. 156-61.
Diehl, P.F. (1985) Contiguity and military escalation in major power rivalries, 1816-1980. Journal of Politics, 47, 4, pp. 1203-11.
Diehl, P.F. (1991) Geography and war: A review and assessment of the empirical literature. International Interactions, 17, 1, pp. 11-27.
Diehl, P.F. (1995) Avoiding detours and dead ends on the road to war. In: In: S.A. Bremer & T.R. Cusack (Eds.) The Process of War: Advancing the Scientific Study of War. Amsterdam: Gordon & Breach, pp. 203-10.
Eberwein, W.D. (1981) The quantitative study of international conflict: Quantity and quality? An assessment of empirical research. Journal of Peace Research, 18, 1, pp. 19-38.
Ellingsen, T. (2000) Colorful community or ethnic witches' brew? Multiethnicity and domestic conflict during and after the cold war. Journal of Conflict Resolution, 44, 2, pp. 228-49.
Everts, P.P. (1995) Democratie en vrede. Transaktie, 24, 1, pp. 45-72.
Farrar, L.L. (1977) Cycles of war: Historical speculations on future international violence. International Interactions, 3, pp. 161-79.
Galtung, J. (1964) A structural theory of aggression. Journal of Peace Research, 1, pp. 95-119.
Gantzel, K.J. & T. Schwinghammer (1995) Die Kriege nach dem Zweiten Weltkrieg, 1945 bis 1992. Daten und Tendenzen. Münster: Lit Verlag.
Gibbs, B.H. & J.D. Singer (1993) Empirical Knowledge on World Politics: A Summary of Quantitative Research , 1970-1991. Westport CT: Greenwood Press.
Gilpin, R. (1981) War and Change in World Politics. Cambridge: Cambridge University Press.
Gleditch, N.P. (1995) Geography, democracy, and peace. International Interactions, 20, pp. 297-323.
Gochman, C.S. (1990) The geography of conflict: Militarized interstate disputes since 1816. Paper annual meeting International Studies Association, Washington DC., April 10-14.
Gochman, C.S. & Z. Maoz (1984) Militarized interstate disputes, 1816-1976: Procedures, patterns and insights. Journal of Conflict Resolution, 28, 4, pp. 585-616.
Goertz, G. & P.F. Diehl (1992) Territorial Changes and International Conflict. London: Routledge.
Goldstein, J.S. (1988) Long Cycles: Prosperity and War in the Modern Age. New Have: Yale University Press.
Gray, R.A. (1994) Genocide in the Chittagong Hill Tracts of Bangladesh. Reference Services Review, 22, 4, pp. 59-80.
Haas, M. (1974) International Conflict. New York: Bobbs-Merrill.
Hazlewood, L.A. (1975) Diversion mechanisms and encapsulation processes: The domestic conflict - foreign conflict hypotheses reconsidered. In: P.J. McGowan (Ed.) Sage International Yearbook of Foreign Policy Studies. Beverly Hills: Sage, pp. 213-43.
Henderson, E.A. (1998) The democratic peace through the lens of culture, 1820-1989. International Studies Quarterly, 42, 3, pp. 461-84.
Houweling, H.W. & J.G. Siccama (1977) Geschiedenis in oorlogvoering? Transaktie, 6, 3, pp. 152-78.
Hower, G. & D.A. Zinnes (1989) International political conflict: A literature review. DDIR-Update, 3, 3, pp. 1-14.
Jencks, H.W. (1973) The Great Powers and War: 1902-1971. Paper Conf. Peace Science Society, San Francisco, February.
Kegley, C.W. & G. Raymond (1986) Normative constraints on the use of force short of war. Journal of Peace Research, 23, pp. 213-27.
Levy, J.S. (1981) Alliance formation and war behavior. Journal of Conflict Resolution, 25, 581-614.
Levy, J.S. (1982) Historical trends in great power war, 1495-1975. International Studies Quarterly, 26, pp. 278-300.
Levy, J.S. (1983) War in the Modern Great Power System, 1495-1975. Lexington: University Press of Kentucky.
Levy, J.S. (1989) The causes of war: A review of theories and evidence. In: P.E. Tetlock et al. (Eds.) Behavior, Society, and Nuclear War (Vol. 1). New York: Oxford University Press, pp. 209-333.
Levy, J.S. & T.C. Morgan (1986) The war-weariness hypothesis: An empirical test. American Journal of Political Science, 30, pp. 26-49.
Low, B. (1993) An evolutionary perspective on war. In: W. Zimmerman & H.K. Jacobson (Eds.) Behavior, Culture, and Conflict in World Politics. Ann Arbor: University of Michigan Press, pp. 13-56.
Luard, E. (1968) Conflict and Peace in the International System. Boston: Little, Brown.
Luard, E. (1986) War in International Society. New Haven: Yale University Press.
Mandel, R. (1980) Roots of the modern interstate border dispute. Journal of Conflict Resolution, 33, pp. 3-36.
Mansfield, E.D. (1988) Distributions of war over time. World Politics, 41, pp. 21-51.
Manson, J.H. & R.W. Wrangham (1991) Intergroup aggression in chimpanzees and humans. Current Anthropology, 32, 4, pp. 369-77.
Maoz, Z. (1995) The onset and initiation of disputes. In: S.A. Bremer & T.R. Cusack (Eds.) The Process of War: Advancing the Scientific Study of War. Amsterdam: Gordon & Breach, pp. 35-62.
Maoz, Z. & N. Abdolali (1989) Regime types and international conflict, 1816-1976. Journal of Conflict Resolution, 33, pp. 3-35.
Maoz, Z. & B.M. Russett (1992) Alliances, distance, wealth, and political stability: Is the lack of conflict among democracies a statistical artifact? International Interactions, 17, 3.
Maoz, Z. & B.M. Russett (1993) Normative and structural causes of democratic peace. American Political Science Review, 87, 3, pp. 624-38.
Mesquida, C.G. & N.I. Wiener (1996) Human collective aggression: A behavioral ecology perspective. Ethology & Sociobiology, 17, 4, pp. 247-62.
Modelski, G. (1972) Principles of World Politics. New York: Free Press.
Modelski, G. (1981) Long cycles, Kondratieffs, and alternating innovations: Implications for U.S. foreign policy. In: C.W. Kegley & P. McGowan (Eds.) The Political Economy of Foreign Policy Behavior. Beverly Hills: Sage, pp. 63-83.
Modelski, G. (1987) Long Cycles in World Politics. Seattle: University of Washington Press.
Moller, H. (1967) Youth as a force in the modern world. Comparative Studies in Society & History, 10, 237-60.
Mousseau, M. (1998) Democracy and compromise in militarized interstate conflicts, 1816-1992. Journal of Conflict Resolution, 42, 2, pp. 210-30.
Naroll, R. (1966) Does military deterrence deter? TransAction, 3, 2, pp. 14-20.
Naroll, R. (1969) Deterrence in history. In: D. Pruitt & R. Snyder (Eds.) Theories and Research on the Causes of War. Englewood Cliffs: Prentice-Hall, pp. 150-64.
Naroll, R.; V.L. Bullough & F. Naroll (1974) Military Deterrence in History: A Pilot Cross-Cultural Survey. Albany: SUNY Press.
Oneal, J.R.; F.H. Oneal; Z. Maoz & B.M. Russett (1996) The liberal peace: Interdependence, democracy, and international conflict 1950-85. Journal of Peace Research, 33, 1, pp. 11-28.
Oneal, J.R. & B.M. Russett (1997) The classical liberals were right: Democracy, interdependence, and conflict, 1950-1985. International Studies Quarterly, ??
Organski, A.F.K. (1958) World Politics. New York: Knopf
Organski, A.F.K. & J. Kugler (1980) The War Ledger. Chicago: University of Chicago Press.
Ray, J.L. (1995) Democracy and International Conflict: An Evaluation of the Democratic Peace Proposition. Columbia: University of South Carolina Press.
Richardson, L. (1960) Statistics of Deadly Quarrels. Pittsburgh: Boxwood Press.
Rummel, R.J. (1968) The relationship between national attributes and foreign conflict behavior. In: J.D. Singer (Ed.), pp. 187-214.
Rummel, R.J. (1979) Understanding Conflict and War; Vol. 4: War, Power, Peace. Beverly Hills: Sage.
Rummel, R.J. (1983) Libertarianism and international violence. Journal of Conflict Resolution, 27, 1, pp. 27-71.
Rummel, R.J. (1996) The Holocaust in comparative and historical perspective. In: A.J. Jongman (Ed.) Contemporary Genocides: Causes, Cases, Consequences. Leiden: PIOOM, pp. 17-32.
Rummel, R.J. (1997) Power Kills: Democracy as a Method of Nonviolence. New Brunswick: Transaction Publ.
Russett, B.M. (Ed.) (1972) Peace, War, and Numbers. Beverly Hills: Sage.
Russett, B.M. (1995) The democratic peace: And yet it moves. International Security, 19, pp. 164-75.
Russett, B.M.; C.R. Ember & M. Ember (1993) The democratic peace in non-industrial societies. In: Russett et al., pp. 43-72.
Russett, B.M. et al. (1993) Grasping the Democratic Peace: Principles for a Post-Cold War World. Princeton: Princeton University Press.
Singer, J.D. (Ed.) (1968) Quantitative International Politics: Insights and Evidence. New York: Free Press.
Singer, J.D. (Ed.) (1980) The Correlates of War: II. Testing Some Realpolitik Models. New York: Free Press.
Singer, J.D. (1990) Accounting for international war in the twentieth century. Olin lecture series, US Air Force Academy, Dec. 7.
Singer, J.D. (1995) Metaphors and models in the explanation of war. In: S.A. Bremer & T.R. Cusack (Eds.) The Process of War: Advancing the Scientific Study of War. Amsterdam: Gordon & Breach, pp. 227-32.
Singer, J.D. (1999) Evolutionary bias, democratic regimes, and war 1816-1995 (Abstract). Aggressive Behavior, 25, 1, 4-5.
Singer, J.D. & M. Small (1972) The Wages of War, 1816-1965. New York: Wiley.
Singer, J.D.; S.A. Bremer & J. Stuckey (1972) Capability distribution, uncertainty, and major power war, 1820-1965. In: B.M. Russett (Ed.), pp. 19-48.
Singer, J.D. et al. (1979) Explaining War: Selected Papers from the Correlates of War Project. Beverly Hills: Sage.
Siverson, R.M. & P.F. Diehl (1989) Arms races, the conflict spiral, and the onset of war. In: M.I. Midlarsky (Ed.) Handbook of War Studies. Boston: Unwin Hyman, pp. 195-218.
Small, M. & J.D. Singer (1976) The war proneness of democratic regimes. Jerusalem Journal of International Relations, 1, Summer, pp. 49-69.
Small, M. & J.D. Singer (1982) Resort to Arms: International and Civil Wars, 1816-1980. Beverly Hills: Sage.
Somit, A.O. & S.A. Peterson (1998) Darwinism, Dominance and Democracy: The Evolutionary Basis of Authoritarianism. New York: Praeger.
Sorokin, P. (1937) Social and Cultural Dynamics. Vol. 3: Fluctuations of Social Relationships, War and Revolutions. New York: American Book Co.
Starr, H. & B. Most (1978) A return journey: Richardson, 'frontiers' and wars in the 1946-1965 era. Journal of Conflict Resolution, 22, 3, pp. 141-7.
Stuckey, J. & J.D. Singer (1973) The powerful and the war prone. Ranking the nations by relative capability and war experience, 1820-1964. Paper Mexico City Conference on 'Poder Social'.
Thompson, W.R. (1988) On Global War. Columbia: University of South Carolina Press.
Van der Dennen, J.M.G. (1981) On war: Concepts, definitions, research data - a short literature review and bibliography. In: UNESCO Yearbook on Peace and Conflict Studies 1980. Westport CT: Greenwood Press, pp. 128-89.
Van der Dennen, J.M.G. (1995) The Origin of War: The Evolution of a Male-Coalitional Reproductive Strategy. Groningen: Origin Press.
Vanhanen, T. (1999) Ethnic Conflicts Explained by Ethnic Nepotism. Research in Biopolitics, Vol. 7, Stamford CT: JAI Press.
Vasquez, J.A. (1976) Statistical findings in international politics: A data-based assessment. International Studies Quarterly, 20, 2, 171-218.
Vasquez, J.A. (1986) Capability, types of war, peace. Western Political Quarterly, 38, 313-27.
Vasquez, J.A. (1987) The steps to war: Toward a scientific explanation of correlates of war findings. World Politics, 40, pp. 108-45.
Vasquez, J.A. (1993) The War Puzzle. Cambridge: Cambridge University Press.
Vasquez, J.A. (1995) Why do neighbors fight? Proximity, interaction, or territoriality?. Journal of Peace Research, 32, 3, pp. 277-94.
Wallensteen, P. (1981) Incompatibility, confrontation, and war: Four models and three historical systems, 1816-1976. Journal of Peace Research, 18, 1, pp. 57-90.
Wallerstein, I. (1974) The Modern World-System. New York: Academic Press.
Weede, E. (1984) Democracy and war involvement. Journal of Conflict Resolution, 28, 4, pp. 649-64.
Wilkenfeld, J. (1968) Domestic and foreign conflict behavior of nations. Jounal of Peace Research, 5, 1, pp. 56-69.
Wilkenfeld, J. (Ed.) (1973) Conflict Behavior and Linkage Politics. New York: David McKay.
Wilkinson, D. (1980) Deadly Quarrels: Lewis F. Richardson and the Statistical Study of War. Berkeley: University of California Press.
Wilson, M. & M. Daly (1985) Competitiveness, risk-taking and violence - the young male syndrome. Ethology & Sociobiology, 6, 1, pp. 59-73.
Woods, F.A. & A. Baltzly (1915) Is War Diminishing? A Study of the Prevalence of War in Europe from 1450 to the Present. Boston: Houghton Mifflin.
Wrangham, R. & D. Peterson (1996) Demonic Males: Apes and the Origins of Human Violence. Boston: Houghton Mifflin.
Wright, Q. (1942) A Study of War. Chicago: University of Chicago Press.
Zinnes, D.A. (1980) Why war? Evidence on the outbreak of international conflict. In: T.R. Gurr (Ed.) Handbook of Political Conflict. New York: Free Press, pp. 331-60.

Please send your reactions to