Chapter 1: Introduction and Summary
Appendix 1.1: Q and A on No Wars Between Democracies
Chapter 13: Why Does Power Kill?
Other Democratic Peace Documents On This Site
For once, desirable means would serve desirable ends.
----Erich Weede, "Democracy and War Involvement"
Before trying to explain the democratic peace, I must be clear about what is in fact to be explained. We have seen that democracies not only do not make war on each other, but the more democratic two nations the less likely they will commit violence against each other. Moreover, the more democratic a regime, the less likely it will commit violence overall, have domestic political violence, or murder its own people. One should note immediately that what is being expressed is not a dichotomous relationship, such that democracies are nonviolent and other regimes are violent. Rather, as the empirical evidence shows, there is a continuum. At one extreme we have totalitarian regimes committing the greatest violence, with democratic regimes the least, and authoritarian regimes in between. What then needs to be explained is the nature of this underlying continuum and why it should suppress all forms of violence as it does.
But this continuum cannot be well isolated unless I pay more attention to the nature of democratic, authoritarian, and totalitarian regimes than I have so far.
In social and political relations power is the ability to get someone to do something they would not otherwise do.
When dealing with societies, however, there are three basic powers (coercive, authoritative, and bargaining) or some mix of them, that function to structure social relations.
Most of these coercive societies are commanded by totalitarian regimes.
Then there are authoritative societies. These are societies that are structured traditionally, according to customary rules and laws. They are spiritually rather than sensory oriented and define truth more in terms of core books and the sayings of great men rather than empirical knowledge. There is surely some coercion and perhaps minorities or dissidents that are forced to obey, but people largely do what they do because they believe they ought to. The regime is seen as legitimate, with a right to rule. Its laws codify what is seen as moral and proper. Thus the major power structuring the society is the authority of the national culture and the regime. They base their rule on the past, allow little competition for or interference with the rightful power of the regime (although they may negotiate and compromise over issues important to one or two lesser sources of power, such as the church or military), and try to enforce cultural, and often religious, norms.
Authoritarian societies are not totalitarian. Although cultural and religious laws may be enforced on all, such as forbidding women from driving automobiles or wearing "immodest" clothing like shorts, there are still large areas in which people or groups are free. Businesses can operate, profit can be made, the arts can be pursued, science can be studied, and sports enjoyed. All this is regulated by the regime (one could not paint a blasphemous portrait of Mohammed, for example), but there is still a wide area of nonpolitical freedom.
Then there is the third type of national society, that is largely regulated by bargaining power. There is a central government, to be sure, along with the coercion and authority that are the essential attributes of any regime, but most relations between the regime and the society, and especially in the society as a whole, are based on exchange.
We thus have three types of regimes that in effect form a political triangle, as shown in Figure 8.1.
With this triangle we thus have three sides and thus three continua along which regimes can vary.
There is one type of regime yet to locate in this space. That is the oligarchic republics that the historian Spencer Weart found almost as likely not to make war on each other as democracies.
As I pointed out in Chapter 2**, Weart discovered that beginning with the classical Greeks, well established democratic republics never clearly made war on each other, and that well established oligarchic republics almost never did, while each type fought wars against the other and against nonrepublics. How do these political types fit into the political triangle? His democratic republic largely corresponds to what I am calling here a democracy. His oligarchic republic would then comprise a band of regimes separating democracies from totalitarian and authoritarian types. This is all I need say about them at this point, but I will return to his distinction at the appropriate place in trying to account for the democratic peace. For I am now convinced that any explanation of this fact must also be able to account for both peace among oligarchic republics, and for wars between them and democracies.
Now, with the three basic types of regimes (democratic, authoritarian, and totalitarian, and recognizing what we might call the Weart addenda--oligarchic republics) and three continua, what then is the theoretical continuum along which violence moves from the least to the most. I will argue that the most relevant continuum is that along the democratic to totalitarian side of the triangle, which is that from the freedom of an exchange society to subservience to coercion and its corollary, force. As people have more freedom from government coercion the severity of political violence decreases; as the coercion exercised by government increases, violence will increase with it. Indeed, violence in part becomes the mechanism for extending the regime's power and eradicating competing centers of power within the society, as in the extensive democide of virtually all new communist regimes.
Theoretically, then, in the propositions relating democracy and violence where more or less democracy is related to foreign or internal violence or democide, what is meant is more or less freedom from coercion and force, as I will make more precise in chapters 12 and 13.** The proper scale for measuring this, therefore, would have regimes at one end that are constitutionally limited, with a competitive and open political system, the election of top leaders with a secret ballot and universal adult franchise, and civil liberties such as the freedom or speech, religion, and organization. At the other end of this scale would be the totalitarian regime, with no true competitive elections for political power, with no independent sources of power allowed (such as of a church or military), and no civil liberties. In the middle, then, would be the authoritarian regime, where people are excluded from competition for political power and their civil liberties are curtailed to what is allowed by tradition and, often, the only legal religion; but outside these spheres people may have considerable freedom, as in business. All the democratic, authoritarian, totalitarian scales that I used to test out the propositions in Part 1 were based on this continuum.
Few measurements of democracy in the literature are consistent with this continuum, and indeed, the more popular way of measuring democracy can be misleading when correlated with violence. In particular, it is not unusual to consider democracy a dimension defined by electoral related characteristics, such as the extent of the electoral franchise, whether regular elections are held, the percent of the population voting, whether the legislature is elected and the degree of its independence from an executive, whether the executive is elected and their power over the legislature, and the like. A result of such measurements is that the usual democracies (Sweden, Switzerland, Belgium, Great Britain, Canada, the United States, etc.) will be found at one end of the scale, with the totalitarian regimes in the middle, and authoritarian regimes like Saudi Arabia and Kuwait at the opposite end.
However, the most authoritarian regimes and especially those with hereditary absolute monarchs usually have no political parties, no elections, and no legislatures with meaningful power. This means that basing a measurement of democracy on the existence of such structures will necessarily scale these regimes at the opposite end from democracies.
Defining this continuum still does not give us an explanation of violence. It is simply to say that the relationship of more or less democracy to violence means in essence more or less freedom from coercion, more or less civil liberties and political rights. Why this connection?***
* From the pre-publisher edited manuscript of Chapter 8 in R.J. Rummel, Power Kills: Democracy as a Method of Nonviolence, 1997. For full reference to Power Kills, the list of its contents, figures, and tables, and the text of its preface, click book.
** See the table of contents.
*** For the summary answer, see Chapter 1 and Chapter 13.
1. Of the many works on political systems, I have found Finer (1971) most helpful, while still pertinent to contemporary regimes. My discrimination of political characteristics reflects his typology and survey, without following it in detail. Moreover, in my view one cannot well divide these characteristics into different regime types without careful attention to the underlying political ideologies. On this, see O'Sullivan (1989), Minogue (1985), and in particular Ebenstein and Fogelman (1985).
2. In Rummel (Vol. 2: The Conflict Helix, Chapters 19, 20, and 21) I have considered a variety of definitions of power, including those by Russell (1938, p. 18), Bierstedt (1950, p. 733), Simon (1957, p. 5), Dahl (1957, p. 202), Cartwright (1959, p. 193), Kuhn (1963, p. 317), Schermerhorn (1961, p. 12), Jouvenal (1962, p. 96), May (1972, p. 99) and, of course, the most conceptually systematic of treatments, Lasswell and Kaplan (1950, p. 76). As a result I have treated power as a family of physical and social powers. The social are defined as intentionally producing effects through another's will, and involves coercive, authoritative, bargaining, intellectual, altruistic (or love), and manipulative powers, which I will discuss in the text. For an excellent general treatment of power and related questions, see Barry (1989) and Baldwin (1989). For a recent radical view of power, see Wartenberg (1992).
3. Social exchange is not only the basis of bargaining power but a basic type of social interaction. In my work it is one of a handful of central variables. On social exchange, see Baldwin (1989), Cook (1987), and Gergen, Greenberg, and Willis (1980). For an analysis of events in Maoist China in terms of social exchange, see Nee (1991).
4. I have developed this idea in Rummel (Vol. 2: The Conflict Helix, Chapter 30). As to whether these societies in fact exist, I have summarized the relevant empirical analyses done by myself and others in Rummel (Vol. 2: The Conflict Helix, section 34.3 of Chapter 34). I concluded that three separate clusters of nations well representing these three types of societies is delineated by the cross-national factor analysis of hundreds of socio-political, economic, cultural, and behavioral characteristics. See particularly Russett (1967), Cattell (1950), and Rummel (1972). [On the nature of factor analysis, see "Understanding Factor Analysis"]
5. The distinctions to be made among totalitarian, authoritarian, and (liberal) democratic regimes are discussed more fully in Rummel (Vol. 2: The Conflict Helix, Chapter 31). They are based on four fundamental political characteristics: the degree to which regimes are open or closed (allow competitive elections for top offices and have freedom of speech and organization); allow groups outside of politics to be free; follow traditional or positivistic norms; and follow backward looking (e.g., maintaining past customs), forward looking (e.g., mobilizing the population to achieve some goal), or presentist policies (e.g., satisfying public demands). A variety of factor analyses show that the three major types of regimes-democratic, authoritarian, and totalitarian-derived from these four characteristics also emerge from these and a variety of other characteristics on which regimes can be measured. For a summary of these see Rummel (Vol. 2: The Conflict Helix, section 34.2 of Chapter 34). Also see the subsequent factor analyses simultaneous done over years, nations, and variables (called super-p factor analyses) in Rummel (1979b, pp. 42-44). [On factor analysis itself, see "Understanding Factor Analysis"]
6. Totalitarianism has been a disputed concept in political science, although not among those who have suffered under such regimes, such as Aleksandr Solzhenitsyn (1973). Among the best treatments if this type of regime, see Arendt (1966) and Friedrich and Brzezinski (1965). For a 20th century history of totalitarianism, see Taylor (1993). For a treatment of totalitarianism consistent with my democratic, authoritarian, totalitarian typology, see Ebenstein (1962).
7. In Sorokin's (1937-1941) useful and important distinction between sensate and ideational cultures, authoritative societies are ideational. For particularly useful works on authoritarianism as a type of political regime, see Germani (1978) and Perlmutter (1981).
8. The exchange and coercive societies are alternative types of what Sorokin (1937-1941) calls sensate cultures in his great work.
9. Von Mises (1963), one of the top economists of this century, has generalized the exchange basis of the free market to all social relations. I am doing the same here, but for a particular kind of society. The term "exchange" also captures the essence of the type of society Friedrich Hayek, Adam Smith, John Stuart Mill, and Jeremy Bentham had in mind, and which is the ideal of contemporary libertarians (the inheritors of classical liberalism), such as Friedman (1962), Rothbard (1962), and Hospers (1971).
10. I have much discussed democracy as a type of regime in Chapter 2 and described the various definitions used among those testing in particular for the relationship between democracy and war. For recent and helpful theoretical discussions of democracy going beyond definitions, see Barry (1989), Benjamin and Elkin (1985), and Stankiewicz (1981).
11. This political triangle is empirical and has been delineated by the factor analyses cited in endnote 4 and endnote 5. Moreover, in Rummel (Statistics of Democide, Chapter 17) I did a component factor analysis of seventeen political scales and regime characteristics on about half of the over 400 regimes that existed or have existed in this century. A large number of sources were used for the scales, including Banks 1971, Arat (1991), and Gurr (1990). I could not use the scales referenced below in endnote 14 because they only were for regimes for years around 1960. The largest orthogonally rotated factor, accounting for about 50 percent of the variance, was a democratic versus totalitarian dimension; the second was an authoritarian versus totalitarian dimension accounting for 20 percent of the variance and, and the third, taking an additional 14 percent of the variance, was monarchical. In effect, these dimensions empirically verify further the political triangle, since the first two dimensions that together reflect 70 percent of the variation among all the political indices and characteristics comprise the two sides of the political triangle. Because the political triangle exists in a two-dimensional political space, where the two dimensions are democracy versus totalitarianism, and democracy versus authoritarianism, a line drawn from totalitarianism to authoritarianism in the two-dimensional space completes the triangle. That is, the third side is implicit in the other two.
12. There are of course many political mixtures that fall within the triangle, without being at the corners. For example, utilizing Finer's (1971) classification of political types, we could define direct or indirect military, facade-democratic, and quasi-democratic mixtures. Or using an alternative classification by Coleman (1960) we could locate in the triangle systems that are political, tutelary, terminal, or colonial democracies; and modernizing, colonial and racial, conservative, or traditional oligarchies. For an excellent more recent analysis of political types, see Bebler and Seroka (1990). There is no point here is getting so specific about the mixtures, examples of which are given in Chapter 13. It is sufficient to designate the corners of the political triangle and the connecting continua.
13. Weart (1994, 1995)
14. For example, for a political democracy index developed by Bollen (1980), the country most opposite from democracy is Yemen, then followed by in decreasing order Saudi Arabia, Afghanistan (before the communist coup), Cuba, and Iraq. See also his latest scales (Bollen, 1991). For Smith's (1969) index of democracy, the most opposite country is Saudi Arabia, followed by Ethiopia (under Emperor Haile Selassie) and Afghanistan (before the communist coup of 1978). For Cutright and Wiley's (1969, appendix C) political representation index, the most opposite to democracy is Saudi Arabia, followed by Nepal, Afghanistan (before the communist coup), and Iraq. There is also the liberal democracy scale of Coulter (1975), which has as the most opposite from the democracies the UAR, Iraq, Sudan, Kenya, and Pakistan. In spite of the apparent similarity of these scales, their correlations are low. For example, the variation in common among scale values for Bollen's index and that of Smith is 69 percent (this is the product moment correlation squared times 100), with Cutright and Wiley it is 72 percent, and for Coulter it is 62 percent. This is for scales presumably measuring the same thing for about the same years. [On the nature of correlation, see Understanding Correlation]
For my purposes, the best scale is that developed by Raymond Gastil (1991) for Freedom House. It measures the degree of freedom by civil liberties and political rights and ends up placing totalitarian regimes at the opposite end from undoubted democracies. The amount of variation in common between the Freedom House scale values for freedom and Bollen's index of democracy is 70 percent, for Smith it is 47 percent, for Cutright and Wiley 51 percent, and for Coulter 51 percent. Clearly, these scales are all measuring different aspects of political systems, and one therefore must be theoretically clear about the choice of scale.
15. As in fact has happened on a number of scales developed in the literature. See note 14.