TY - JOUR
AU - Chernoff, Fred
AB - Abstract This essay argues that the field of international relations has exhibited “progress” of the sort found in the natural sciences. Several well-known accounts of “science” and “progress” are adumbrated; four offer positive accounts of progress (those of Peirce, Duhem, Popper, and Lakatos) and one evidences a negative assessment (Kuhn). Recent studies of the democratic peace—both supporting and opposing—are analyzed to show that they satisfy the terms of each of the definitions of scientific progress. Does the field of international relations exhibit “progress” in the scientific sense, as the behavioral movement promised a half century ago? A decade ago in this journal Jack Levy (1994:452) noted that “the idea that democracies almost never go to war with each other is now commonplace. The skeptics are in retreat and the proposition has acquired a nearly law-like status.” Whatever conclusions one draws about Levy's comment—made in the midst of intensive academic scrutiny of democratic peace studies—an appraisal of progress in international relations would seem to hinge on the notion of “progress” one accepts. There have been several highly influential views of scientific progress advanced in the past century and a half. This essay argues that whichever view the reader prefers, democratic peace studies succeed in satisfying its key conditions, indicating that progress in international relations is possible. Cumulation and consensus in international relations, along with many other fields in the social sciences, have been slow in coming. Three-quarters of a century ago, John Hobson (1926:7) felt the need “to afford some explanation of the slowness of these sciences in producing any considerable body of larger truths, in the shape of generally accepted laws and principles.” Democratic peace studies show that such a body of knowledge is possible. Indeed, many authors have touted the successes of research on the democratic peace. To show that democratic peace studies satisfy a range of definitions of scientific progress—rather than just one or two—the accounts presented here are necessarily brief and focus on the authors' notions of science and progress. This essay does not attempt to adjudicate between the different (often overlapping) accounts of science. Given the normal space limitations for journal articles, at least three different strategies could have been used, each with different drawbacks. One approach would have been to choose just one of the accounts and offer a thorough explanation of it and a detailed argument for how the study of the democratic peace satisfies or does not satisfy its requirements. If this approach were selected, a critic could have argued that some other account of science is, in fact, superior and thus the conclusions are moot. A second strategy would have been to select one of the accounts and present a thorough argument that it is indeed the correct account, followed by an analysis of democratic peace studies against its requirements. Once again, the question of which is the correct account would arise and any debate could not be settled conclusively in a single journal article. A third strategy, taken here, is to provide a brief explanation of the various accounts of science and an argument showing whether democratic peace studies satisfy the criteria for “scientific” and “progressive” fields inherent in each. This approach requires that each of the five explanations discussed in what follows is rendered in outline form and that each of the analyses is somewhat limited. More can be said, to be sure, about the argument that democratic peace studies fit a given account of science. However, because this essay shows that democratic peace studies fit all five accounts, the results, preliminary though they may be, should be of interest to political scientists. Moreover, because this essay does not take sides regarding which account of science is correct, its results should be of interest to all scholars of international relations who have entertained the idea that this field or some areas within the field are progressive, cumulative, and scientific. The five views of science and progress discussed here overlap with one another and, yet, are distinct. The reader might object that the analysis is inherently flawed if democratic peace studies fit all of them. That is, given that they are all different from one another, it would seem most likely that democratic peace studies would fit one or two but not all. However, it is entirely possible that research on the democratic peace—or, for that matter, any other specific area of study—will fit all, or none. The five views were chosen to include certain scientific areas of research—Newtonian mechanics, optics, organic chemistry, for example, as well as quantum theory and relativistic physics. That democratic peace studies satisfy all the definitions should be no more surprising than that they satisfy only some subset. Finding progress in the literature on the democratic peace will be highly significant because this research deals with a core question in international relations. As Harvey Starr (1997:153) puts it, “given that war has been perhaps the single most central concern to students of international relations across history—and certainly to Realists—uncovering one factor, variable, or set of conditions that is associated with the complete (or almost complete) absence of war is a stunning achievement.” The argument to follow seeks to cast doubt on the views of those who are skeptical of science and progress in international relations and who would dispute the claim that democratic peace studies have satisfied the requirements of being scientific and of producing genuine progress, no matter which of the major philosophers the reader prefers. This essay focuses on the views of philosophers of science and attempts to deal with the objections of reflectivist and postmodern social scientists by examining Thomas Kuhn's (1962) views which they generally endorse. The essay is organized as follows: (1) discussion of the major points in the debate regarding democratic peace hypotheses, (2) a brief overview of the various concepts of scientific progress presented by the major philosophers of science under review and some of the criticisms of these notions of progress, (3) evidence that democratic peace studies satisfy the various criteria of science and progress indicated by these philosophers, and (4) a presentation of the problems for considerations of progress raised by the thesis of radical underdetermination of theory by data. Debate over Democratic Peace Hypotheses One of the major controversies in international relations during the pasthalf-century has been over claims that the social sciences are capable of producing theories whose structures and functions parallel, at least roughly, theories in the natural sciences. Such claims emanate from behavioralists, positivists, and proponents of the scientific approach (Hempel 1965; Singer 1969; Rummel 1979, 1981; Rosenau 1980; Russett 1993; Bueno de Mesquita 2002) and have been contested by interpretivists, critical theorists, and postmodernists as well as others (Carr 1964; Habermas 1971; Taylor 1985; Ashley 1986; Wyn-Jones 1999). Key traits that scientific disciplines manifest are the accumulation of knowledge as successive waves of inquiry proceed and the development of a consensus on a question that is carefully examined according to scientific methods of inquiry. Critics have argued that the natural sciences are capable of these traits but not such social sciences as political science and the study of international relations (Bohman 1993; Little 1998). Other critics, especially Kuhn and his followers, have contended that even the natural sciences do not exhibit cumulation and an approach to consensus (Feyerabend 1962; Kuhn 1962; Latour 1987; Knorr-Cetina 1993). They hold that competing paradigms are not commensurable. Defenders of a scientific approach to international relations have had difficulty citing examples of areas of study that exhibit these characteristics. This essay proposes that research on the democratic peace constitutes just such an example. Thus, the conceptions of the natural sciences of Popper, Lakatos, Duhem, and Peirce and those of the critics of natural scientific progress, like Kuhn and Feyerabend, are examined below. It is argued here that recent studies of the democratic peace satisfy all these definitions of scientific progress. To start with, however, what is needed is a careful step-by-step examination of the interactions among the critics and defenders of the democratic peace phenomenon. The central democratic peace hypothesis was first stated at the end of the eighteenth century by Immanuel Kant (1989). Kant's argument was theoretical rather than empirical; the historical record was sparse given that few states actually fit his definition of a republic or democracy at the time. The idea was not widely discussed until Dean Babst (1964, 1972) published the first of a pair of articles arguing that democracies are more peaceful than nondemocracies, although Babst did not cite Kant. Two prominent quantitative international relations theorists, Melvin Small and J. David Singer (1976), published a critique of this democratic peace hypothesis, while R. J. Rummel (1979, 1981, 1983, 1995, 1997) was the first to publish empirical tests supporting and calling attention to the hypothesis. Michael Doyle (1983a, 1983b, 1986) authored a series of widely read papers that further highlighted the debate. Since the late 1980s, an explosion of articles and books has either attacked the democratic peace hypothesis or restated and refined it. Jack Levy (1989), William Dixon (1993), John Owen (1994, 1997), James Lee Ray (1996, 2000, 2003), Steve Chan (1997), Harvey Starr (1997) and others have defended the notion of a democratic peace. The work of Bruce Russett and his coauthors constitutes the epicenter of the debate. They (Russett and Starr 1981; Russett 1990, 1993; Maoz and Russett 1993; Oneal and Russett 1997, 1999a, 1999b, 2001; Russett and Oneal 2001) have offered evidence for democratic peace claims, developing a comprehensive set of tests with generally positive results. Dixon (1993) and Gregory Raymond (1994) have extended ideas regarding the concept of a democratic peace to include methods of dispute resolution. A long list of authors, a number being realists, have attacked the work of Russett and other defenders of democratic peace hypotheses on grounds of substance and conceptualization (such as Layne 1994; Kegley and Hermann 1995, 1997; Oren 1995; Farber and Gowa 1997; and Gowa 1999) and on grounds of research design and statistical testing (Spiro 1994; Beck, Katz, and Tucker 1998; Gartzke 1998; Green, Kim, and Yoon 2001; Henderson 2002). Similarly, the critique that focused on the direction of causality advanced by Manus Midlarsky (1995) and William Thompson (1996) has been directly engaged by authors like Michael Mousseau and Yuhang Shi (1999). These various authors have debated a number of distinct hypotheses related to the democratic peace proposition. Kant (1989) argued that a system of democratic states would create the basis for a peaceful international environment. The most important recent formulations are the claims that democratic states are more peaceful than nondemocratic states (the monadic hypothesis) and that democracies are more peaceful with respect to one another than others are (the dyadic hypothesis). Moreover, a range of corollaries and extensions of the basic hypotheses have been advanced. Russett and Oneal (2001) have taken Kant's view that peace arises in a system that has democratic states, free trade, and properly formed international organizations to formulate a triadic theory of international cooperation and peace. Others have looked at how democracy and peace relate to territorial disputes (Huth and Allee 2003), interventions and extrasystemic wars (Kegley and Hermann 1997; Hermann and Kegley 2001), war initiation (Small and Singer 1976; Reiter and Stam 1998, 2002), geography (Gleditsch 1995; Gleditsch and Ward 1997), crisis escalation (Hewitt and Wilkenfeld 1996; Senese 1997), and third party mediation (Dixon 1993; Raymond 1994). These latter claims form part of what Lakatos (1978) calls the auxiliary hypotheses of a research program (see also Ray 2003). Scientific Progress and Its Critics The debate concerning the proper methods to use in studying international relations is an old one, going back to the early post-World War II years when behavioralists of the scientific school argued that this field and other social sciences could replicate the successes of the physical sciences. Many authors have disputed this view, using two general arguments. One states that the natural sciences show progress, but the social sciences do not. The second contends that neither the social nor the natural sciences exhibit true progress. With respect to the first sort of criticism, philosophers of social science like Charles Taylor (1985), Daniel Little (1991, 1998) and James Bohman (1993) propose that the social sciences are inherently unlike the natural sciences; all agree that whereas the natural sciences are predictive, the social sciences are not. They hold that views of scientific progress such as those of Charles Sanders Peirce, Pierre Duhem, Karl Popper, Imre Lakatos, and others cannot be applied to the social sciences. Little (1991), for example, argues that social scientific laws are too complex; they contain too many variables. Furthermore, the social sciences exhibit phenomenal regularities, not underlying, governing regularities. The second sort of criticism holds that even the natural sciences do not really show progress and so neither, a fortiori, do the social sciences. These arguments are targeted at traditional positivism. They are both descriptive and prescriptive, describing practice in the sciences and arguing that a more traditionally positivist and ambitious notion of certainty is philosophically unwarranted. In the late 1950s through the early 1960s, Norwood Russell Hanson (1958), Thomas Kuhn (1962), and Paul Feyerabend (1962) began to develop positions aimed at undermining the traditional account of the natural sciences (see also Latour 1987; Knorr-Cetina 1993). The implications for the social sciences were obvious, and Kuhn became one of the most oft-cited philosophers of science in the writings of international relations theorists and other social scientists. Scholars argued that the natural sciences are essentially subjective; the process of replacing one theory or paradigm by another, allegedly superior, one can be understood on the basis of sociological rather than rational grounds. Ultimately, scientific progress is an illusion because of the problem of the incommensurability of paradigms. In international relations, various reflectivist and postmodern critics of scientific progress concur with this element of the Kuhnian critique. Peirce Methodology Charles Peirce (1960–1966, 1982) was an active physical scientist most of his life and was the founder of the most important movement in philosophy and the philosophy of science in the United States: pragmatism. In Peirce's view, scientific practice requires induction, deduction, and abduction. Although he supported the use of inference in reaching the best explanation, he did not see it as the end of inquiry, leading to the acceptance of an explanatory statement as true; it was an intermediate step in the process of confirming scientific principles. Progress and approach-to-consensus are closely connected in Peirce's accounts of science and truth. Indeed, consensus is crucial to his view of the scientific pursuit of truth because he holds that “the truth (about any given matter) is what the community of inquirers would converge to in the long run” (Reynolds 2002:67). Peirce's theory of inquiry emphasizes the long-term convergence of investigators. According to Peirce, continued inquiry leads away from false beliefs. How can this happen through a process that includes fallible inductive inferences? For this philosopher of science, the real guarantee of the validity of induction is that it is “a method of reaching conclusions which, if it be persisted in long enough, will assuredly correct any error concerning future experience into which it may temporarily lead us” (Peirce 1960–1966, Vol. 2:para 769). But beliefs may be incorrect at any given point in the process of inquiry and may remain so for generations (Peirce 1982, Vol. 3:79). Critics who charge that Peirce's account of science is overly optimistic about what constitutes progress generally misrepresent his position, overlooking his view that error may remain for generations, in part because there is no unshakable foundation to his epistemology. The enterprise of science, according to Pierce (1960–1966, Vol. 5:589; cited by Misak 1991:123), is metaphorically “not standing upon a bedrock of fact. It is walking upon a bog, and [one] can only say, this ground seems to hold for the present. Here I will stay till it begins to give way.” Democratic Peace Studies and Consensus Democratic peace studies can be seen as scientific from Pierce's point of view because they satisfy at least three of his criteria. First, democratic peace studies use both induction and deduction as well as abduction. Second, the abductive steps, or “inferences to best explanation,” are not the end of inquiry but are actually subjected to further inductive tests. Typical is the inference from the dyadic hypothesis to the explanation that institutional structure accounts for the observed dyadic peace (for example, Maoz and Russett 1993). This explanation is subjected to empirical tests and compared to competing explanations such as the proposition that dyadic peace results instead from democratic norms. Third, the process of hypothesis testing by both realist and liberal authors has moved toward consensus on some key propositions concerning the democratic peace phenomenon, particularly the qualified dyadic hypothesis that mature democracies rarely, if ever, fight one another. Both liberal proponents of the democratic peace (for example, Ray 1996) as well as realist opponents (for example, Farber and Gowa 1997) acknowledge a “strong consensus on two related issues”—the dyadic hypotheses about war and about other lower-level disputes (Farber and Gowa 1997:394). With respect to the growing consensus, a number of authors who began with different conclusions, stemming from divergent ways of operationalizing their terms, have subsequently altered their definitions and methods, coming closer to agreement. One example is the interaction between Maoz and Russett (1993), Thompson and Tucker (1997), and Oneal and Russett (1997), discussed below in connection with Kuhn's views of definitions. Here the conclusions that many international relations theorists themselves have drawn regarding consensus are significant. For example, as Chan (1997:62) sees it, in democratic peace studies “considerable progress and excitement have resulted from an iterative process of criticism and response.” He says, “the democratic peace proposition is one of the most robust generalizations that has been produced to date by the [quantitative] research tradition” (Chan 1997:60). Nathaniel Beck and his colleagues (1998:1274) have observed that the notion “that democracies do not go to war against one another” is “one of the most celebrated propositions in the IR/IPE literature” and that it “has been confirmed by myriad empirical studies.” Lewis Richardson (1960) and Small and Singer (1976) have argued against the monadic hypothesis. However, as noted earlier, Rummel (1979, 1981, 1983, 1995, 1997) argued in its favor. Other authors (Chan 1984; Weede 1984; Garnham 1986; Maoz and Abdolali 1989) have taken up the question empirically and “provided evidence against Rummel's conclusion and reinforced a consensus that democracies have generally not engaged in less war than” nondemocracies (Starr 1997:153, n.1). Starr (1997) notes, however, that some dissent continues (Ray 1996:Ch. 1). With regard to the dyadic hypothesis “that democracies do not (or only rarely) go to war against one another,”Starr (1997:153) says that “the empirical findings across a number of studies have produced a consensus in support of this hypothesis.” Indeed, it is “one of the strongest nontrivial or nontautological generalizations that can be made about international relations” (Russett 1990:123). Even before the intensive debate of the past fifteen years, which has served to increase agreement on some points, Levy (1989:270) argued that the dyadic hypothesis was “as close as anything we have to an empirical law in international relations.” At this point, however, more agreement exists regarding the observational generalizations than about the explanations for them. Normative and institutional explanations compete with one another (Maoz and Russett 1993; Siverson 1995). Another type of explanation offered by Ray (2003), Bueno de Mesquita and his colleagues (1992, 1995, 1999), and others proceeds using rational choice notions: citizens and subjects—rather than presidents and monarchs—fight in wars, die in wars, and pay taxes to finance wars. In most cases, it is not in citizens' self-interest for the state to go to war. Nor is it in the interest of decision makers in democracies to go to war, given that war conflicts with the interests of those who can remove them from office (see Ray 2003:230–235). Democracies in Transition Some authors have argued that nondemocratic states become more war-prone as they undergo a transition to democracy and have asked whether the properties and causal mechanisms attributed to democratic states by proponents of the democratic peace are undermined by the observation that states become more war-prone as they democratize (Mansfield and Snyder 1995a; Zakaria 1997). The increase in war-proneness might be explained by holding that any sort of regime, whether democratic or nondemocratic, becomes vulnerable to attack by nondemocratic neighbors when it is in a transitional phase. The vulnerability explanation has been criticized by authors who show that the war-proneness of democratizing regimes stems from the democratizing states' initiation of crises. Does the debate over the inherent properties of democracies, as those regimes come into being, undermine the claims just made about the scientific nature of the study of the monadic democratic peace hypothesis? Three observations are helpful here regarding this question: the recent development of the debate, the cumulative nature of the debate, and the consistency of even negative evaluations with the existing consensus surrounding the dyadic hypothesis. On the first point, more disagreement probably exists on the question of the peacefulness of emerging democracies than on the monadic hypothesis applied to fully formed or mature democracies. The former cases have been intensively studied for a much shorter period, given that research commenced in full force only in the mid-1990s (more than a decade after serious debate over the monadic and dyadic hypotheses gained momentum). Second, cumulation has occurred in the debate over the peacefulness of transitional democracies. Edward Mansfield and Jack Snyder (1995a, 1995b) have contended that states undergoing a democratic transition have a greater probability of interstate war, and they argue that it occurs because elites in the transitional democracies make use of nationalist appeals to garner mass support that, in the absence of strong democratic institutions, may propel them into war. Critics indicate that Mansfield and Snyder have not considered various possible causes of war and hold that they studied time periods too lengthy to permit inferences about causal effects from start to finish (Maoz 1998). Rather than responding by attempting to discredit their critics, Mansfield and Snyder (2002:311–318) have acknowledged the force of the criticisms and revised their research design to produce new analyses that do not fall victim to the dangers the critics outlined. In response to critics of their earlier papers, for example, they indicate that by “employing a more refined research design than in our prior work, we aim here to identify more precisely the conditions under which democratization stimulates hostilities” (Mansfield and Snyder 2002:298). They also explain how their research design yields results that one should expect to differ from related, but distinct, studies of (1) general regime transition (that is, regardless of whether the regime prior to transition was an autocracy, anocracy, or democracy; see Enterline 1996), (2) full or complete transition to democracy (Oneal and Russett 1997), and (3) escalation to crisis short of war (Enterline 1996) rather than escalation to war. Third, Mansfield and Snyder (2002), who are the most persistent proponents of the claim that transitional democracies are war-prone, argue in their most recent piece that no conflict exists between an endorsement of the dyadic hypothesis and their assertions regarding transitional democracies. They accept many of the major findings in the democratic peace literature, for example, that mature democracies initiate fewer crises (Rousseau et al. 1996), win a larger percentage of wars and suffer fewer casualties (Siverson 1995; Bennet and Stam 1998), and so on. These findings hold, Mansfield and Snyder (2002:300) suggest, because mature democracies have “a mutually reinforcing set of institutional, informational, and normative characteristics distinctive to mature democracies, such as accountability to cost-conscious voters, greater transparency of facts and preferences in policy debates, and respect for the civil liberties that make democracy possible. In relations between mature democracies, these characteristics of each party interact in ways that make war very unlikely.” However, as leading proponents of the dangers of democratic transitions, they nevertheless contend that the monadic properties of transitional democracies, which explain their volatile behavior, are consistent with the monadic properties of democracies that, in interaction with similar properties of other democracies, facilitate the avoidance of war. “We see no conceptual mismatch between our monadic argument and the democratic peace literature,” they argue (Mansfield and Snyder 2002:300), even though that literature includes Russett and Oneal's (2001) rejection of the monadic hypothesis and acceptance of the dyadic hypothesis. Duhem Exactly a century ago, in 1904–1905, Pierre Duhem (1954) developed a very sophisticated version of “conventionalism.” For Duhem, adopting a conventionalist stance toward science was necessary because conventional choice is inescapable in the physical sciences. It is always possible to use different axiomatizations of geometry as the basis for physical theory. Duhem was well aware of the relatively new non-Euclidean geometries. But he proposed that sound, though extra-logical, reasons exist for choosing one over the others. Although fellow conventionalist Henri Poincaré (1905) argued—mistakenly as it turned out—that the simplest geometry, namely Euclidean, would always be preferred by physicists, Duhem contended, in contrast, that physicists would and should prefer the geometry that produces the greatest simplicity if it is conjoined with a set of physical laws; that is, different geometric axioms produce empirically adequate accounts of the observational data when conjoined to different sets of physical laws. The simpler conjunction is what physicists will prefer (see Chernoff 2002, 2004). This stipulation of measures might be interpreted in several ways in democratic peace studies, perhaps the most obvious being the choice of ways of assessing “democracy-nondemocracy” and “peace-war.” In this regard, the literature shows both proponents and opponents of the democratic peace phenomenon adopting the Correlates of War measure and definition of war as 1,000 battle deaths (Singer and Small 1972, 1982) and the general approach of the Freedom House and Polity II, III, and IV definitions of democracy. Over time, there has been a shift in emphasis from Freedom House to Polity, but it has been a shift that has been adopted by most authors in the field and, thus, constitutes conventional choice. Today several widely accepted and largely overlapping typologies have emerged (for example, Freedom House and Polity). However, when Small and Singer (1976:53) published their critique of the democratic peace hypotheses, they said that “there is no generally accepted typology of national regimes or governmental types.” On war initiation, they were more satisfied with published material. They indicated that they were comfortable with the general historical consensus regarding which party fired the first shot or crossed the first frontier in the initiation of a war. Yet, they were not satisfied with judgments regarding “responsibility” for the war, that is, for the provocative behavior, which they considered more subject to dispute. One of Duhem's most profound and lasting contributions to the philosophy of science is his holistic principle limiting the falsification of hypotheses (and his consequent rejection of crucial experiments). According to this now widely accepted principle, the observation of counter-instances (or falsifying observations) cannot disprove a specific hypothesis and, thereby, prove its competitor; counter-observations show only that a false statement exists somewhere in the conjunction of hypotheses, laws, theoretical principles, auxiliary statements, and background beliefs (including those about the functioning of experimental equipment). Although charges that the social sciences have not exhibited progress and cumulation are often well-founded, it is argued here that a Duhemian account shows that democratic peace studies constitute scientific inquiry and exhibit progress. Thus, the conventionalist account suggests that research on the democratic peace can evidence cumulation and approach-to-consensus in a way not unlike that in the natural sciences. There have been examples of conventional choices of definitions and measures in democratic peace studies both in the case of the dependent variable and in the case of the independent variable. In terms of the dependent variable, scholars have debated what should be counted: mere absence of war, stable peace, interdependence, and so on. Should we count threats of the use of force that do not subsequently escalate to violence? Should we include covert actions that are shielded from the public? Studies have largely adopted the convention of “absence of war” as the way of assessing core democratic peace hypotheses, although lower levels of violence have been studied to gain deeper insight into the mechanisms of democracy that produce peace, to understand which specific features of democracy lead to peace, and to use some statistical methods whose tests require a large N (because wars are not frequent enough for the relevant tests). Other questions about the dependent variable have also been raised, such as whether peaceful democratic norms ought to preclude internal civil war, whether conflict should be measured qualitatively or quantitatively, and, if the latter, how much violence is permissible before one counts the event as a war. Indeed, a number of authors have argued in favor of scalar tests. Zachary Elkins (2000:299) has shown that such tests are more valid because “measures of democracy which provide for gradations best fit the behavior that theoretical work on democracy would predict.” The choice, however, of which particular measure to adopt as the dependent variable in tests of the core hypotheses is largely a conventional one. Similarly, the quantitative question of how much violence is required for an incident to be regarded as a war has been answered, following Singer and Small (1972, 1982), by a conventional decision to place the threshold at 1,000 battle fatalities. In terms of the independent variable, one must ask: What sorts of regimes are more pacific? As Owen (1994) notes, some states are liberal but not democratic (Britain before 1832), whereas others are democratic but not liberal (the Confederate States of America). So distinctions must be drawn between states that have constraints on their leaders, that require accountability of their leaders, that are liberal (defined in terms of political liberties and rights), that provide a limited role for a hereditary monarch to choose governmental leaders (Babst 1964), that are republican (equal rights among the—possibly minority—group of decision-making citizens), that are democratic (defined in terms of voting rights), and that are libertarian (defined by Rummel 1983 in terms of a combination of political and economic freedoms). If “democracy” is chosen, as it often is, then in formulating and testing democratic peace hypotheses it is important to clarify what features of a state must be present for it to be considered a democracy—for example, mass participation, protection of the rights of minorities, individual liberties, and so on. Once the distinctions are drawn and the preferred term selected, then, as with the dependent variable, there remains the question of whether it is to be investigated by means of a more qualitative (democracy versus nondemocracy) measure or quantitative (degrees of democracy–autocracy) measure. The two approaches may well yield different results. Moreover, a distinction exists between initializing democracy (building institutions) versus consolidating it (creating a civic culture); some hold that only the latter is relevant for testing the democratic peace hypotheses. The Duhemian view helps explain how authors managed to argue at cross-purposes for many years and then move toward convergence at a particular point in time. The key is the notion of the “measure-stipulation.” As noted earlier, in physical theory everyone must make a conventional choice of a measure-stipulation. One investigator may choose one set of geometric axioms (for example, Euclidean) and adopt a set of physical laws that are consistent with all observations; another investigator may choose an alternative geometry and adopt a different set of physical laws that are consistent with all observations. Progress occurs when a common choice is made in the discipline. Is there a rational basis for choice among the possible measure-stipulations? Duhem argues that rational, even if extra-logical, criteria do guide theory choice, directing all rational investigators to one set of stipulations. The process of conventional adoption of the equivalent of a measure-stipulation appears to be what has happened in research on the democratic peace. If the criteria for theory choice are not widely accepted, then different groups of theorists will use different theories (as realists and liberals have so often done), and there will be no generally accepted theory or laws. In democratic peace studies, however, conventional but rational choices of measures have been made, and cumulative work has resulted. In conclusion, three important points should be considered. First, there is the need for conventional choice in the natural sciences, so it should not be surprising that international relations scholars, especially those doing quantitative studies, also cannot get by without it. Second, there has been adoption of measure-stipulations in democratic peace studies. Both liberals and realists have largely accepted the same variable concepts and methods of measurement. Third, there has been at least modest cumulation as the hypotheses have been intensively debated. This last point will be discussed further below. Popper In his early work on science, Karl Popper (1965, 1968) sought a demarcation criterion that would separate the genuinely scientific from everything nonscientific. Unlike the logical positivists, who were his contemporaries, Popper did not seek a criterion of the “meaningful” but rather of the “scientific.” He argued that the theories of Marx and of Freud were not scientific because any new observation could be fit into the malleable contours of each theory. Popper regarded these theories as pseudo-scientific and found logical positivism's treatment of them to be inadequate. Although there was some common ground between Popper and the logical positivists, he argued against them on a variety of issues—for instance, flaws in the verification of the principle of meaningfulness (Popper 1968:35–40), the role of metaphysics (Popper 1968:35–39), and the necessity for universal laws (Popper 1965:288–289). Popper was heavily influenced by Hume's criticism of induction, as a result of which he denied the acceptability of induction. He was particularly opposed to induction as the foundation of scientific knowledge; indeed, he sought to solve the foundational problem by devising a criterion of “the scientific” that was purely deductive. That criterion was “falsifiability.” The methodology based on falsifiability was structured on the deductive rule of inference, modus tollens. Given a bold theoretical conjecture, one should find observational consequences (set up appropriate experiments) and ascertain whether the expected consequences follow. This process has been interpreted both as a demarcation principle and as a basis for scientific methodology. Progress is a matter of conjectures and refutations. Bold conjectures are offered; refutations are attempted; and if the refutations fail, the conjectures constitute a better basis for knowledge than the refuted propositions that are known to be false. Popper argues that bold conjectural propositions, for which little or no basis exists at the time of their initial proposal, are the core of scientific progress. These conjectures are then subjected to rigorous testing. If they are true, they will stand up to rigorous testing and acceptance of them will advance progress. Weak conjectures are not as likely to move scientific progress forward because they largely overlap the evidence and do not go beyond it, thus failing to add enough of what Popper (1968:41, Appendices VII and VIII) calls “content.” They do not reveal powerful new truths. Popper (1965:vii) says that progress is possible in “our scientific knowledge by unjustified (and unjustifiable) anticipations, by guesses, by tentative solutions to our problems, by conjectures. These conjectures are controlled by criticism, that is, by attempted refutations, which include severely critical tests. They may survive these tests, but they can never be positively justified…even as probable (in the sense of the probability calculus)” [italics in original]. Democratic peace studies exhibit the characteristic of falsifiability as a demarcation criterion and as a methodology. We can see the application of the falsifiability principle in research on the democratic peace by considering, for example, the treatment of the monadic hypothesis. The claim that democracies are inherently more peaceful than other types of states is attributed to Woodrow Wilson and to Kant before him. Small and Singer (1976:50) also attribute this proposition to Machiavelli who lived well before either. Soon after Babst's (1964) first article calling democracy and democracies a force for peace, Maurice East and Phillip Gregg (1967) argued in favor of the monadic hypothesis in a piece in the International Studies Quarterly. Small and Singer (1976) performed one of the first empirical explorations of the mondic hypothesis. In fact, they examined three variations of this hypothesis: (1) that democracies are involved in fewer wars than nondemocracies, (2) that democracies initiate fewer wars (proportionally) than nondemocracies, and (3) that wars involving democracies are less severe (shorter or less violent) than wars fought by other types of political systems. They also considered a fourth proposition, the dyadic hypothesis, that democracies (almost) never fight one another. Small and Singer found that democracies can act very aggressively; the evidence was especially strong when colonial wars were included in the analysis. They concluded that the data did not support three of the four democratic peace hypotheses, and they had reservations with respect to the fourth. Small and Singer (1976) say that they doubt that the absence of war between democracies is truly a result of the democratic nature of the regimes. They argue that most wars are fought between geographical neighbors (direct neighbors or those with neighboring colonies) and that democracies rarely border one another. As time has passed, many other scholars have argued against the monadic hypothesis (for example, Chan 1984; Weede 1984; Garnham 1986; Maoz and Abdolali 1989; Starr 1997). The widespread rejection of the monadic hypothesis based on empirical analyses using commonly accepted sources of data and coding, such as the Correlates of War and Polity III datasets, is about as clear an example, despite some holdouts (for instance, Rummel 1997), for the falsification of a statement as one can demand. Thus, at least some propositions in democratic peace studies are falsifiable and meet Popper's standard of scientific form. Furthermore, some hypotheses (within various sets of competing hypotheses) have been falsified, which moves research on the democratic peace in the direction of Popper's defined notion of progress. Lakatos Imre Lakatos (1978) offers the most widely endorsed methodology by researchers studying international relations. Although Popper and Kuhn are perhaps cited more frequently, they are cited both favorably and critically. Given that Lakatos commands so much loyalty among methodologists in the field, it is worth considering in some detail the criteria he defends. Lakatos (1978:37–47) advances an account of the scientific method, derived largely from Popper, which he terms “sophisticated falsificationism.”2 Whereas Popper relies only on deduction to produce a secure scientific methodology that can escape skepticism, Lakatos (along with Peirce, Duhem, and most other philosophers of science) contends that other, ampliative forms of reasoning are acceptable in science. Philosophers of science, including most of those considered here, have produced rationally defensible, nondeductive criteria of theory choice. So not only are some theories rejected deductively from a finite body of evidence, but other theories, which may be deductively consistent with the evidence, are rejected on the basis of nondeductive criteria. Lakatos, Duhem, and others offer examples of such criteria. One such criterion that Lakatos defends is that of “progressiveness.” According to Lakatos (1978:34), an emendation of an existing theory is “progressive” if it includes all the unrefuted content of its predecessor and “leads to the actual discovery of some new fact.” Most current theorists of note in international relations endorse a Lakatosian approach to theory testing and appraisal, though not all explicitly discuss each of the major components of his philosophy of science. Their adoption of his views is sometimes suggested by their use of distinctly Lakatosian terminology, viewing theories in terms of “research program(me)s,” with “hard core” and “protective belts” of propositions, even if they do not specifically cite his work. Richard Ashley (1986:260, 275–276) uses the Lakatosian understanding of “scientific research program.” Other elements of Lakatos's account, or outright endorsement of his methodology, are evident in a wide variety of writings on international relations (see, for example, Jervis 1976; Russett 1985; King, Keohane, and Verba 1994; Wayman and Diehl 1994; Christensen and Snyder 1997; Elman and Elman 1997; Schweller 1997; Vasquez 1997, 1998; Waltz 1997; Wendt 1999; Hudson 2001; Russett and Oneal 2001). Lakatos's Methodology and the Criterion of “Progress” New Content. Philosophers of science have to account for the processes involved in gathering evidence, formulating theory, and handling new counter-evidence. According to Lakatos's view of the history of science, theories that are found to have anomalies are typically not promptly rejected but rather are revised in light of such counter-evidence and evolve into new theories. He holds that the hard core of the theory is typically not revised; rather other elements of the theory, especially the protective belt of auxiliary hypotheses, may be adjusted to deal with new observations. The original theory is regarded as evolving into a new and more adequate version rather than as having been falsified and discarded, as Popper argued. Had the original theory simply been viewed as falsified and abandoned, the result, Lakatos contends, would have been the rejection of many of the most successful and widely accepted theories in the history of science. So for Lakatos the unit of appraisal is a “series of theories.” When a shift to a new series of theories occurs, it is because the new set is superior in that it meets several crucial criteria. Ad Hoc Adjustments and New Content. How should one evaluate competing sets of laws and associated explanations in a given domain if all of them are more or less consistent with observed data? Theories, or research programs, are to be preferred if they are “progressive” rather than “degenerating.” A progressive research program is classified as such if it (1) is not repeatedly revised in an ad hoc fashion when new evidence arises, (2) includes the unrefuted content of earlier research programs, (3) adds new content to the earlier theories and research programs, and (4) directs researchers' attention to novel facts. That is, progressive research programs not only need less in the way of ad hoc adjustments, but they actually point the way to new facts and explain phenomena that the theory was not specifically formulated to explain. Lakatos argues for the superiority of progressive research programs over degenerating ones. Novel Facts. In his discussions of progressiveness, Lakatos uses the phrase “prediction of novel facts” to point to facts or observed phenomena that may be predicted to occur in the future as well as events that may have occurred in the past. Lakatos (1978:32, n. 3) says explicitly that postdiction (also known as retrodiction) is included in the “wide sense” in which he uses the term “prediction.” That the phenomena may have occurred in the past and also have been previously observed is clear from Lakatos's use of the example of the differentials in astronomical measurements taken during the day versus those taken at night. The possibility of carrying out such measurements was always there. So, for Lakatos, a progressive research program takes older theories and replaces them with new ones that both add content and predict new facts without the help of a continuing series of ad hoc additions. In the case of the Lakatosian methodology, the extensive discussion that has already occurred in the international relations literature eases the task of presenting a brief evaluation of the performance of democratic peace studies. The most recent and sophisticated general exposition of Lakatos's methodology can be found in the work of Colin Elman and Miriam Fendius Elman (2002). They summarize Lakatos's methodology of scientific research programs as including a hard core of propositions, positive and negative heuristics, and a protective belt of auxiliary hypotheses. In a very careful recent paper specifically evaluating democratic peace studies in terms of Lakatos's account, Ray (2003) offers ten examples of auxiliary hypotheses. He argues that the hard core of the research program consists of five stipulations. First, democracies exist. Second, democracy's impact is universal. Third, the probability of war between two states is equal to one minus the product of the strength of democracy in the two states divided by one plus the distance between the states raised to a power of a geographic constant. Fourth, “domestic political processes have important impacts on international interactions, and vice versa” (Ray 2003:213). Fifth, “a primary goal of leaders of states is to stay in power” (Ray 2003:235). Ray ties the fifth proposition to the democratic peace hard core by citing the work of Bueno de Mesquita and his collaborators and by citing one of his own earlier papers in which he has said that “the basic realist argument that foreign policymakers will make decisions that are in the ‘national interest’ has always implied a quite ‘unrealistic’ tendency of national leaders and foreign policymakers to be altruistic” (cited by Ray 2003:233). He points out that Bueno de Mesquita, Siverson, and Woller (1992) as well as Bueno de Mesquita and Siverson (1995) show that regimes that initiate wars and lose very often are removed, particularly if they are in democracies. Bueno de Mesquita and Lalman (1992) posit, as Ray (2003:231) puts it, “that national leaders will attempt to maximize utility within both an international and a national context.” He indicates that the fifth proposition of the democratic peace hard core is consistent with various realist-oriented theories and claims that this rational choice explanation is more in the realist tradition than either the normative or institutional explanations. But one can argue that this proposition is presumably connected to the normative explanation of the democratic peace, given that the norms of a liberal democratic polity will create pressure on leaders to behave according to such norms because the leaders desire to stay in power. Lakatos's Criteria and Democratic Peace Studies Increasing Content. The development of democratic peace studies, as it has played out in the past thirty years, is progressive in Lakatos's sense. Democratic peace proponents treat democratic states as rationally pursuing their own self-interest. The explanations for the pacificity of democracies typically involve such considerations. Democratic peace studies, thus, incorporate into a fuller explanation the unrefuted content drawn from points made by other rationalist international relations theorists, both realists and liberals. Moreover, the modern democratic peace debate, which was initiated by Babst in 1964 and 1972, has increasingly expanded the content on which it focuses. One of the two propositions that Babst advanced is the dyadic hypothesis that democracies have historically never fought one another. (The other, which has not been part of the core democratic peace debate, is that democracies are numerically on the increase. He concluded that when enough democracies are in place, the world will become much more peaceful.) As the reader will recall from our earlier discussion, Small and Singer (1976) examined four propositions, finding only that the dyadic hypothesis stood up to empirical scrutiny. The unrefuted content of these 1960s and 1970s versions of the democratic peace, especially the dyadic hypothesis, were tested in the 1980s and 1990s and built into larger theories by the end of the decade, especially by Russett and Oneal (Oneal and Russett 1999b; Russett and Oneal 2001). No Ad Hoc Adjustments. Defenders of the democratic peace phenomenon (Russett 1993; Ray 1996; Oneal and Russett 1999a, 1999b, 2001; Russell and Oneal 2001) have faced intense scrutiny, primarily, though not exclusively, from realists (Layne 1994; Spiro 1994; Farber and Gowa 1997; Beck, Katz, and Tucker 1998; Green, Kim, and Yoon 2001). In their responses, defenders have made a substantial number of adjustments to their arguments. But these changes do not constitute what Lakatos regards as ad hoc adjustments, and thus do not run afoul of Lakatos's criterion. It is difficult to prove a negative of this sort, that is, the absence of ad hoc adjustments. Still, democratic peace defenders have made technical adjustments as a result of the criticisms of other scholars who have, in some instances, shown that the indicators used by those defending the democratic peace propositions are not as precise as they should be if the goal is to get at the core associations between “democracy” and “peace.” Those defending the democratic peace have offered refinements in the definitions and indicators for both the independent and dependent variables (see below), seeking to get closer to the core meanings of the democratic peace concepts. But there has been no backtracking on, for example, the dyadic hypothesis. The monadic hypothesis has been largely rejected by both sides of the democratic peace debate. This rejection, though, does not constitute a “degenerating” shift because the monadic hypothesis is neither logically nor theoretically bound up with the formulation or testing of the dyadic or other key hypotheses, and it is not part of the Kantian origins of the research program. (At least, it is not obviously part of the origin, given that Kant argued that a system replete with democracies would, if other conditions were met, lead to peace; he did not say generally that good states do or would go to war less.) At the center of the program is an interest in understanding the basis for a peaceful international system, which requires the dyadic but not the monadic hypothesis. Thus, the Lakatosian hard core of the program has not been compromised in the course of criticism and response. Novel Facts. The literature on the democratic peace meets the Lakatosian criterion of “prediction of novel facts,” that is, of highlighting observable phenomena that have not heretofore been observed or explained. The literature on peace and conflict, prior to the explicit injection of the democratic peace hypotheses, focused on other variables, most often balance of power, distribution of capabilities, presence of institutions and regimes, and “soft power” (see Keohane and Nye 1977; Nye 2002). Prior to Kant or Rousseau, it did not discuss the role that democracies or republics played in war and peace. The nature of the democratic peace studies seems clearly parallel to an example cited by Lakatos regarding Halley's comet. Halley used the Newtonian research program to predict that an observed comet would return in seventy-two years; the comet had for ages been passing earth every seventy-two years, but attention had not been focused on that periodicity. Likewise, for centuries, democracies have been avoiding war with one another. The Kantian view predicts such a pattern and the modern interpretations of the democratic peace draw attention to an observable association that other theorists did not observe and even continued to question for a time after the empirical democratic peace claim was made. Democratic peace hypotheses focus on a set of dyadic relationships that can be explained by the democratic regime type. One might argue that democratic peace studies do not at this point constitute a genuine research program in Lakatos's sense, and that may be accurate. But that is not to say that it is not developing the characteristics of a genuine research program. The key characteristics of a proper research program include a hard core “tenaciously protected from refutation by a vast ‘protective belt’ of ‘auxiliary hypotheses’ and a ‘heuristic,’ that is, a powerful problem-solving machinery, which, with the help of sophisticated mathematical techniques, digests anomalies and even turns them into positive evidence” (Lakatos 1978:4). On the surface, defenders of the democratic peace hypotheses seem to push for this type of theoretical structure just about as well as some of the authors of research programs that Lakatos includes in his writing, such as those of Freud and Marx. But a final determination as to whether democratic peace studies do or do not fit the mold of the research programs that Lakatos mentions would take us further into the debatable aspects of the interpretation of Lakatos than is appropriate in a short survey of the sort being presented here. So democratic peace hypotheses have implications and predictions that go well beyond those of other international relations theories of conflict. For example, territorially proximate, militarily powerful traditional rivals in an anarchical international system periodically go to war. Democratic peace hypotheses predict that if two such states are democratic, they will not (or are very unlikely to) go to war with one another during the time they remain democracies. France-Britain, France-Germany, and Germany-Russia are examples. Democratic peace defenders could have predicted (or may retrodict) peace between France and Britain after centuries of war; after World War II they would have predicted (or may retrodict) peace between France and Germany. And, today, with Russia emerging as a democracy, they predict peace between Germany and Russia for as long as the two states are democracies. Careful formulations of the dyadic hypothesis do not absolutely rule out war because the social science laws on which predictions are based are not deterministic. Nonetheless, they predict a very low likelihood of war. Democratic peace defenders do not argue that conflicts between democracies are unlikely, only that the escalation of those conflicts into war is unlikely (Leng 1993). Of course, other competing factors, in the proper combination, could probably override the effects of joint democracy. On the subject of the democratic peace, Elman and Elman (2002) offer only one paragraph, but it concisely delivers its conclusion on the “novel facts” criterion. The authors (Elman and Elman 2002:254–255) say that one interpretation of novel facts is supported by the way that researchers of the democratic peace have recently begun to move beyond the absence of war among democracies by expanding the range of dependent variables explained by both cultural/normative and rational/institutionalist theories. Thus Spencer Weart (1998) argues that his cultural theory can account for peace among democratic republics as well as among oligarchies and durable democratic alliances. Similarly, the institutional explanation of Bueno de Mesquita et al. (1999) accounts for numerous empirical regularities beyond the fact that democracies do not fight with one another (e.g., democracies tend to win the wars they fight; transitional democracies are more likely to fight than are stable democracies; democratic dyads choose pacific means of dispute settlement more often than do other pairs of states). Accordingly, Bueno de Mesquita et al. (1999:792) can argue that their theory suggests “novel hypotheses that do not form part of the corpus of democratic peace,” thus lending further credibility to their institutional constraints argument. Although Ray (2003:241) is careful to qualify the claim, he concludes that: the democratic peace research programme has excess empirical content over realism or neorealism. It has a “hard core” that is compatible with but also capable of subsuming basic realist assumptions regarding states seeking power or security. Much of the excess content of the democratic peace research programme has been corroborated repeatedly and consistently in all the research devoted to the basic democratic peace proposition and a wide array of auxiliary hypotheses. In other words, according to Lakatosian principles, it would be fair to conclude that realism and/or neorealism have been “falsified.” Kuhn Incommensurability As noted above, work (for example, Feyerabend 1962; Kuhn 1962) that aims to refute the possibility of social scientific progress makes use of arguments that seek to undermine the idea of progress in both the natural and social sciences. Chief among these arguments are those that are concerned with the principles of incommensurability of paradigms and radical underdetermination of theory by data. Postmodern critics of progress in international relations and the social sciences have made extensive use of these arguments, among their various lines of attack. They contend that no progress has been made in the field of international relations because the competing paradigms are incommensurable. In what follows, it will be shown that this argument appears unfounded given that, if there are incommensurable paradigms in international relations, they are generally taken to be those of realists and liberals. But the behavior of these two groups in the democratic peace debate does not conform to Kuhnian expectations.3 Lakatos (1978:4) characterizes Kuhn's view as an affirmative answer, in contrast with his own negative answer, to the question of whether we must “agree that a scientific revolution is” just an “irrational change in commitment, that it is a religious conversion”? He adds, “if Kuhn is right, then there is no explicit demarcation between science and pseudoscience, no distinction between scientific progress and intellectual decay, there is no objective standard of honesty” (Lakatos 1978:4). The incommensurability of paradigms thesis is generally believed to undermine traditional accounts of science that conceive of theory choice as objectively—or at least as intersubjectively—valid. The outlines of Kuhn's philosophy of science, including the incommensurability of paradigms thesis, are well known. Kuhn (1962) rejects the logical positivist and logical empiricist accounts of science that dominated the middle of the twentieth century by arguing that they were entirely disconnected from the actual history of science. Theory choice is not a regular feature of scientific activity. Science proceeds quite dogmatically within a dominant paradigm (later termed “disciplinary matrix”) that includes methodological rules and procedures for testing hypotheses. The dominant theoretical exemplars have various anomalies, which are not ordinarily taken as falsifying them. Moments of crisis do occur when the activities of “normal science” give way to the genuine choice of a new theory and paradigm. However, according to Kuhn, the choice is not made on the basis of philosophical, rational, or logical grounds of the sort specified by logical empiricists. Rather, the outcomes of the choices are driven by sociological factors. By Kuhn's account, rationally grounded changes from one paradigm to a better one are impossible; hence the ordinary notion of scientific progress collapses. Progress can occur within a circumscribed sphere only, within a single research paradigm. Two reasons are often cited to explain why paradigm choice is not rational, and thus why progress is misconceived. The first is that the paradigm itself contains all the methodological rules, normative prescriptions, and testing procedures of which use is made. So no rules or standards exist outside the competing paradigms that can be taken as objective and employed as a basis for unbiased choice between them. According to the incommensurability thesis, each paradigm can only be understood in a way insulated from the others. The second, and most frequently invoked, reason why theory or paradigm choice is said not to be based on rational grounds is that the logical empiricist methods of science and comparative theory appraisal require a theory-neutral language in which to express basic observations. (Serious criticisms of this second reason are proffered by Fodor 1984 and Greenwood 1990.) But terms that may appear the same in two or more theories (for example, “mass” in Newtonian and relativistic physics) do not in truth have the same meanings in the different theories because their meanings are given by the theories in which they are embedded. Thus, since no true theory-neutral language exists, the theories are incommensurable. For this reason, scientists must turn on and off a “gestalt switch” as they move from one paradigm to another. Those who advance this position often use the analogy of the “duck-rabbit” discussed by Wittgenstein (1953). One may see the diagram as a duck or as a rabbit, and one may switch back and forth between the two. But one may not see it as both duck and rabbit at the same time. This argument was developed in the philosophy of science by Hanson (1958), Kuhn (1962), and Feyerabend (1962) as noted above. According to the incommensurability thesis, when a particular field exhibits a lack of consensus, it may be the result of the impossibility of evaluating contending theories against one another. When consensus reigns, it is for sociological reasons and because of rational progress. (For more on this point, see Hunt 1994.) Dialogue between Competing Paradigms over the Democratic Peace The actual debate between proponents and opponents of the democratic peace is quite different from what the a priori skepticism of the incommensurability thesis would suggest. Supporters and dissenters engage each other very directly regarding questions concerned with definitions of terms, theoretical principles, and methodology. This section will offer evidence to refute the claim that liberal and realist authors argue past and miss one another using incommensurable terms, even if they are orthographically similar. Just as Kuhn argues that positivist and empiricist accounts should be rejected because they are not empirically adequate to explicate the history of physics, it is proposed here that his account is not empirically adequate to help us understand the history of democratic peace studies. Definitions. The dialogue surrounding the definitions of key terms in the literature on the democratic peace is one example of the progressive and cumulative nature of the field. Scholars have come to agree on certain conventions regarding how the independent and dependent variables are measured. As scholars involved in doing research on the democratic peace have reacted to one another, they have generally adopted previously published definitions, which has resulted in greater agreement on the nature of the hypotheses. Theorists can point to many examples of genuinely progressive interactions among researchers over the definitions of key terms. Russett's work with various colleagues provides significant examples of this process (see also the discussion in Tures 2001). In his 1993 paper with Zeev Maoz, Russett developed an institutional measure of democracy: Regime Type=(Democracy Score−Autocracy Score)×Executive Power Score (Maoz and Russett 1993). The scores referred to in this measure are found in the Polity data (see Jaggers and Gurr 1995). This measure of democracy was generally adopted by other scholars until several years later when Thompson and Tucker (1997:448) pointed out that those states with greater concentrations of executive power are more likely to be coded as democracies, even though such states are commonly regarded as less democratic. As a result, in subsequent work Russett dispensed with this particular measure (Oneal and Russett 1997:274). This example is typical of the type of interaction among those doing research on the democratic peace and is suggestive of the genuinely cumulative and progressive character of the debate. Such examples make it hard to accept the criticism of Kuhnians who doubt progress is possible because theorists in opposed traditions are supposed to argue past one another. The sort of progress that Kuhn denies arises from the previously noted “iterative process of criticism and response” (Chan 1997:62), and the iterations continue year after year. Once again Russett (Oneal and Russett 1999b) acknowledges room for improvement, even in the revised measure of democracy that resulted from the interaction with his critics. Moreover, he has commented on criticisms of his revised definition (democracy score−autocracy score), such as that posed by Kristian Gleditsch and Michael Ward (1997), by observing that this measure has “the virtue of being symmetric and transitive” while also noting that “the relative importance of its components is unstable over time” (Oneal and Russett 1999b:12, n. 30). Oneal and Russett accept the shortcomings in their definition and the need to improve it in future studies. Theoretical Critique: Security Interests versus Joint Democracy. Even critics of the democratic peace use similar definitions and methods, or at least they start with these and build their critiques by arguing for the need for adjustments in such techniques. The realist critique by Farber and Gowa (1997) is an excellent example. These two scholars argue that “interests” account for the pattern of dyadic war and peace better than “regime type.” However, their argument does not proceed as the Kuhnian incommensurability account would expect, that is, by offering a distinct set of terms—or a set of terms that take on different meanings because they are embedded in different theories. Rather, the terms “democracy,”“autocracy,”“anocracy,”“militarized interstate dispute” (MID), and “war” have the same meanings as they do when used by supporters of the democratic peace phenomenon; both sides in the debate start with the same datasets, generally the Correlates of War (Singer and Small 1972, 1982) and Polity II and III (Jaggers and Gurr 1995). Even in arguing for their preferred realist variable (that is, interests), Farber and Gowa make use of the same data set for assessing the presence of war as the proponents of the democratic peace do for many of their tests, in particular, the Correlates of War data. When deviating from the definitions used in the consensually agreed-upon datasets, authors accept the burden of defending their deviations. Indeed, when Christopher Layne (1994) criticized the definition of democracy being used in democratic peace studies without offering an explicit definition of his own, the failure did not go unnoticed (see Russett 1995:167). Even though most proponents and opponents come from two different theoretical traditions in the study of international relations (liberalism and realism), the two groups do not in general misunderstand one another. By using the same terms and measures, both groups advance the debate to greater degrees of precision and depth. Farber and Gowa (1997:404), for example, are careful to make sure that their “data yield results consistent with those of previous studies.” They are clearly building upon the work of liberal democratic peace proponents as they present their critique, arguing that “security interests” account for patterns of war and peace better than “democracy.” Even though they use the same 1820–1980 dataset that Russett and other democratic peace proponents use, Farber and Gowa disaggregate the data by dividing them into three time periods: pre-World War I, interwar years, and post-World War II. These researchers show that the democratic peace hypotheses do well in the post-1945 period but poorly earlier, especially in the pre-1914 period. They point out that so many new states emerged in the system after World War II that the results derived by others, based on the aggregated data, are unduly biased by the particular systemic conditions after 1945. For example, in the dyadic analysis of the 165 years from 1816 to 1980, there are 284,602 dyads for which complete data exist. However, even though the thirty-five years following World War II constitute just over one-fifth (21.2 percent) of the number of total years examined, they included nearly two-thirds (65.7 percent) of the dyads. Thus, when the results are aggregated, any factors stemming from the post-1945 system disproportionately affect the statistical tests. Proponents and opponents of the democratic peace are not talking past one another; they are speaking the same language, building upon one another's methods, and developing new results to which the other side is expected to respond. With their study, Farber and Gowa pose a serious challenge to defenders of the democratic peace phenomenon. The responses offered in Oneal and Russett (1999a, 2001) and in Russett and Oneal (2001) primarily deal with the causal and theoretical issues Farber and Gowa have raised. The latter use alliances as their indicator of “common interests” when testing their argument that such interests explain conflict patterns much better than does “democracy.” Russett and Oneal point out that alliances are not independent of democracy. Ideology is often a reason for states to choose one another as alliance partners, even though geostrategic reasons can also have a bearing on conflict behavior. Erik Gartzke (1998) has also argued that common interests more than democratic character are responsible for peace among democracies. In his view, democracies have an affinity for one another that leads them to have overlapping rather than conflicting interests. Hence, the supposed effects of democracy in constraining the escalation of conflicting interests are greatly overstated given that conflicting interests do not exist and, thus, do not need to be constrained. Gartzke contends that all the proposed causal mechanisms used to explain the low incidence of conflict among democracies mistakenly assume that there have been potential conflicts between democracies. Consequently, because there has been little conflict to prevent, there has been no need to escalate. Gartzke (1998:6) uses the following analogy: “Observations of the democratic peace are not unlike studying incidents of seasickness in Central Asia. There is nothing to report. Still, we cannot then assume that Uzbek culture makes them hearty seafaring folk or that Tadzhik bureaucrates introduce a mysterious regime that makes local villagers immune to the effects of vertigo.” Democracies' rarely conflicting preferences are what lead to lack of war, not anything about their democratic nature. Gartkze uses UN General Assembly voting as a measure of “affinity” or “similar preferences,” which some regard as superior to Farber and Gowa's “alliance” measure given that there are more cases of states working together without formal alliances. It seems obvious that if democracies have greater affinities for one another and states with greater affinities and similar preferences do not fight, then democracy is still playing a crucial role. Gartzke (1998:11) proposes that, “regime type similarity actually leads to similar preferences.” Indeed, the factors that cause democracy, like “ecological, material, or cultural factors,” are actually the causes of the shared preferences among democracies. In dealing with this issue of causation, Russett and Oneal argue that democracy has causal effects on UN voting and alliances. But the important observation here is that Russett and Oneal (2001) take up the challenge of “preference similarity” and, using Gartzke's idea of UN General Assembly voting as the indicator, they respond to the criticism by testing the democratic peace hypotheses using “common interests” as one of the control variables. Methodological Critiques. Spiro (1994) has criticized Russett (1993) and Maoz and Russett (1993) on methodological grounds, especially their analyses of time series data. Russett (1995) responded directly to Spiro. The papers by Oneal and Russett (1999a and 1999b) revised and extended this analysis (for example, by bringing in “trade” as an explanatory variable). These time series techniques have also been criticized by Nathaniel Beck, Jonathan Katz, and Richard Tucker (1998) and later by Donald Green, Soo Yeon Kim, and David Yoon (2001) on different grounds. Russett and Oneal responded to these critics in subsequent publications (Oneal and Russett 1999a, 2001; Russett and Oneal 2001) and made adjustments in their analyses suggesting once more the progressive nature of the debate. Let us examine each of these critiques in more detail. Spiro (1994) argued that Russett's (and Maoz and Russett's) use of the pooled time series method was illegitimate because cases were not entirely independent; peace between the United States and Canada in 1994 is not causally independent of the peace between them in 1993. Russett (1995) replied to Spiro's attack by offering an alternative method, by not using dyad-years as the unit of analysis but instead stable dyadic relationships or “dyad-regimes”—periods during which a relationship remained stable. The entire two-century period of peace between the United States and Canada became a single case. This revamping of the unit of analysis avoids any charge of inflating the set of peaceful dyad-years with large numbers of observations that are not independent. Russett then presents an alternative analysis of “war” for the post-World War II period as well as analyses in which the dependent variable is “use of force” and “disputes of any sort.” In the case of war, there were 169 dyad-regimes in which no war occurred, and there were zero wars between democracies. In the 1,045 dyad-regimes that did not contain a democracy, there were thirty-seven wars (3.4 percent) among them. Russett (1995:175) regards the objections raised by Spiro as constructive, indicating that the critique “forced me to devise new tests. The result is that the evidence for the democratic peace is stronger and more robust than ever.” The second type of criticism was leveled by Beck, Katz, and Tucker (1998). In their 1997 paper in the International Studies Quarterly, Oneal and Russett use logistic regression of a pooled time series of dyad-years by state to argue that “democracy” and “interdependence” are significant factors in reducing international conflict (rather than just “democracy” as in Russett 1993 and Maoz and Russett 1993). Beck, Katz, and Tucker charge that Oneal and Russett encounter problems because time series cross-sectional data, especially those using binary variables, undercut the assumption that the observations are independent. “Since it is unlikely that units are statistically unrelated over time,” binary time-series cross-section observations “are likely to be temporally dependent” (Beck, Katz, and Tucker 1998:1260–1261). They indicate that corrections need to be made for temporal dependencies and suggest a method for doing so. Although Beck, Katz, and Tucker (1998:1275) say that “we were able to replicate Oneal and Russett's (1997) original estimates exactly,” they add that “a different picture emerges, however, when we correct for temporally dependent observations using grouped duration methods.” They find that the dramatic effects are on the “trade” variable. In effect, by showing how the data can be reanalyzed, Beck, Katz, and Tucker (1998:sec. 4) offer a remedy to the problem, although, as they admit, the results may prove to be very different. Their alternative method upholds the effects of democracy but not the effects of liberal trade policies.4 An interesting example of progress here is Beck and his colleagues' exact replication of the democratic peace results using the same methods, and their acknowledgement that even when the corrections are made, the powerful effects of democracy remain intact. Furthermore, in subsequent papers, Oneal and Russett (1999a:22, 2001:470) acknowledge the criticisms and revise their analyses to meet them. Referring later to their 1997 piece, Oneal and Russett (2001:470) say that “we did not consider whether there was heteroskedacticity in the error terms, account for the grouping of our data by dyads, or address the lack of independence in the time series” and add that Beck, Katz, and Tucker “showed us the error of our ways” (Oneal and Russett 2001:470). Russett and Oneal appear to take criticisms seriously and address them in a way that, presumably, the critics would applaud. Oneal and Russett (1999a) correct “for temporal dependence” in the estimation of their democratic peace equation. The serial correlation in their time series could lead to coefficients having lower standard errors, erroneously elevating the measure of statistical significance. To deal with this problem, Oneal and Russett (1999a:22)“report…the results of two analyses in which corrections for temporal dependence were applied in the estimation” of their democratic peace equation. Following the recommendations of Beck and his associates, they estimate the coefficients of their “theoretical variables in the presence of a piecewise linear spline of the number of years since the last dispute” and contend that this and other adjustments take into account recent methodological innovations as well as provide further support for the democratic peace. The third critique is that of Green, Kim, and Yoon (2001), who contend that pooling time series data is inappropriate in a broad range of studies in the social sciences. These researchers single out democratic peace studies because of their prominence and importance. In pooled time series analysis, dyads are put in an undifferentiated pool by year (or some other temporal measure). This type of analysis is very common in studies of the democratic peace phenomenon and is done by both supporters and opponents. Because published results give supporters the upper hand (given the view that a near consensus exists), discrediting this general form of analysis undermines their contentions more than those of the opponents. With a revised “fixed-effects” method, Green, Kim, and Yoon test democratic peace hypotheses and show that neither democracy nor mutual trade have strong pacific effects. Russett and Oneal (2001:311–312; Oneal and Russett 2001) do not dispute this methodological argument entirely. They point out that dyadic analysis shows important effects that neither state-level nor system-level analysis is able to do (Oneal and Russett 2001:483) and argue that “a fixed effects model loses a great deal of information when used to explain militarized disputes…because many dyads have not experienced any armed conflict. Consequently, their experience does not enter into the estimation” (Russett and Oneal 2001:312). They also cite others' support for their methodological position, like Beck and Katz (2001), who, as already noted, have been critical of some democratic peace studies. But most importantly from the point of view of cumulation and approach-to-consensus, they have retested their data using a fixed-effects analysis and show that the democratic peace hypotheses are supported by such tests. They observe that the reason Green, Kim, and Yoon find different results is that they use a much shorter time period. Oneal and Russett (2001:471–472) argue that a study like that of Green and his colleagues, which covers just forty-two years (thirty-nine of which are during the Cold War), is likely to overlook much that an examination of 107 years (like Oneal and Russett's) will find because of the greater variation in the values of the variables in the longer period. Green, Kim, and Yoon's data went from 1951 through 1992 whereas the data that Oneal and Russett analyzed began in 1886 and ran through 1992. It is clear from the dialogue between defenders and critics of democratic peace studies that its defenders do not dismiss criticism. They generally take methodological and theoretical critiques very seriously and respond to them in the most direct possible way. They do not just shift indicators, terms, and definitions to produce the desired statistical results but in most cases they develop definitions that capture more of the substance of the concepts at issue and both sides communicate with each other about what they are doing and why. An analysis of the debate does not reveal each side subtly missing the precise points made by the other side, as the incommensurability thesis might suggest. Although physicists do not fully agree with one another all the time, that discipline is marked by a history of approach-to-consensus when new physical theories are introduced. Such a history is vastly different from the history of continuing dissensus in the study of international relations. Even when most authors in a subfield of international relations share a naturalistic or scientific approach, there has been enduring disagreement. Democratic peace studies, however, appear to constitute an exception; as this discussion suggests, there has been cumulation and approach-to-consensus. Radical Underdetermination of Theory by Data Duhem argues correctly that theory is underdetermined by data, which is consistent with a notion of “scientific progress.” Some have taken the principle much further and have asserted that theory is radically underdetermined, thus undercutting a belief in rational scientific progress. The principle of radical underdetermination of theory by data implies that the data are so far from being able to adjudicate objectively between competing theories that adjustments are always possible in one's preferred theory to the point that the theory can be rendered deductively consistent with the erstwhile anomalous observation. According to the radical underdetermination principle, the available data cannot help the investigator choose between competing theories. The source of radical underdetermination typically cited by advocates is W. V. O. Quine's (1953) influential essay, “Two Dogmas of Empiricism,” particularly his argument that observations do not ever compel the investigator to reject a theory or any particular statements within it. As Quine puts it, any statement can be maintained in the face of new observations, “come what may.” He argues that theories can be adjusted, possibly in radical ways, to accommodate any new observation. This position contrasts with those of Duhem and Popper who regard observations as capable of “falsifying” theories. That is, even though Duhem clearly contends that scientific community standards of evidence can demonstrate the falsification of some theories, Quine holds that any theory can be retained—although he does support “rational” theory on pragmatic grounds (see Quine 1953:46). Some have argued that Quine's view constitutes another philosophical basis, beyond that of Kuhn, for rejecting the claim that science is rational and, thus, to reject the traditional conception of scientific progress. The present essay has focused on democratic peace studies and their relationship to conceptions of “scientific progress.” In the case of the antiprogressive arguments of the radical underdetermination supporters, the best counter-arguments are to befound in the philosophical literature (for example, Gillies 1993:Ch. 5; in international relations see Wight 1996). However, it is worth pointing out that thedemocratic peace literature does, at the very least, cast some doubt on radical underdetermination, even if empirical evidence cannot directly confute the radical underdetermination principle. Two aspects of the literature on the democratic peace raise questions about radical underdetermination. First, convergence between realists and liberals would seem unlikely if their political principles allow them to interpret any pattern of democracy throughout the history of interstate warfare in a way so as to retain their original views. Second, the widespread rejection of the monadic hypothesis by liberals is difficult to square with the acceptability of radical underdetermination. If it is possible to maintain any preferred theory in the face of any erstwhile empirical counter-evidence, then it would not make much sense for supporters of Kantianism to abandon the monadic hypothesis. Even though radical underdetermination does not rule out theorists mistakenly or unnecessarily rejecting favored principles, it does hold that rejection of old favorites is never logically compelled. So the rejection of favored propositions and the concessions each side has made to the other constitute puzzles for supporters of the radical underdetermination thesis. Conclusions The debate over the democratic peace phenomenon appears to fit the criteria for “science” and “progress” offered by many of the most influential philosophers of science, including Peirce, Duhem, Popper, and Lakatos. Democratic peace studies also cast at least prima facie doubt on the principles of incommensurability and radical underdetermination, which postmodern and other skeptics of scientific progress use as a basis for some of their arguments against the notion of progress in the study of international relations and the social and natural sciences. These studies cast doubt on incommensurability and radical underdetermination by failing to conform to incommensurability's expectations about the terms realists and liberals will use or to radical underdetermination supporters' expectations concerning the way in which convergence has occurred on some of the key democratic peace hypotheses. Peirce, Duhem, and others stress the importance of consensus in the progress of science. There has, indeed, been a movement toward consensus in democratic peace studies. In a similar vein, critics of scientific progress based on the incommensurability and radical underdetermination principles have some difficulty explaining the development of consensus. The development of consensus in democratic peace studies is clear in that even opponents of the democratic peace like Farber and Gowa (1997:394) acknowledge a “strong consensus on two related issues,” that is, the dyadic hypotheses about war and about other lower-level disputes. Many democratic peace supporters also comment on the development of consensus regarding some claims (see Weede 1984; Garnham 1986; Levy 1989:270; Maoz and Abdolali 1989; Russett 1990:123; Gleditsch 1995:297; Chan 1997:60, 62; Starr 1997; Beck, Katz, and Tucker 1998:1274). Proponents of the longest-standing competing theoretical approaches in the field of international relations, liberalism and realism, have come to agree on nontrivial, indeed very significant, generalizations such as the negation of the monadic hypothesis and support for the qualified, “mature” dyadic hypothesis. Advocates of incommensurability and radical underdetermination would not expect this sort of convergence. However, such critics of “progress” might, at least at first glance, try to account for the growing strength of the proponents' side in the debate and still deny genuine “progress” by explaining it in terms of sociological and historical rather than rational influences. Such critics could still maintain that the terms used by the proponents and opponents have divergent meanings—the paradigms are incommensurable—and thus the two sides do not truly engage in “dialogue.” Moreover, even if one side gains numerical strength, the two sides are still arguing past one another. To dig beneath the surface, it is necessary to look at the key strands of the democratic peace debate and observe how the two sides interact with one another, that is, how critics attack democratic peace hypotheses and how defenders respond to these attacks. The detailed analysis presented here shows that the Kuhnian expectations (of contending theorists arguing past one another) do not fit the nature of the debate. This statement is underscored by the fact that some of the earlier opponents of the views that have become widely accepted either have publicly changed their position, as Russett notes (see, for example, Weede 1992; Chan 1993; Dixon 1993), or have simply ceased advancing arguments for their earlier views. The dialogue between defenders and opponents of democratic peace hypotheses, outlined earlier, supports the Lakatosian notion of “scientific progress” and undermines (though does not refute) the Kuhnian skepticism about rational progress. Defenders of the relevance of democracy to peace have adhered to the core principles of the democratic peace literature, that “regime type” explains conflict in that democracies (or “liberal” or otherwise “free” states) do not go to war against one another. They have not diluted these principles or shifted them in such a way that they fail to capture the core concepts of the democratic peace. Thus, they have not responded in a way that Lakatos would regard as degenerating (for example, as Vasquez 1997, 1998 has argued in the case of neorealism). The work of Russett and his colleagues, which has been at the center of the defense of the democratic peace phenomenon, has not abandoned the hard core of the Kantian notion of “liberal peace,” with its tripartite character of political freedom, economic interdependence, and international institutions. Even though many corollaries have been tested, they have not been proffered as substitutes for discredited core democratic peace claims (such as the monadic hypothesis). Over the past decade and a half, the work of Russett and his collaborators has shifted from an emphasis on the dyadic hypothesis to include also other elements of the Kantian theoretical prescription such as economic interdependence and, more recently, international organizations. The latest book by Russett and Oneal (2001) specifically refers to all three factors in its subtitle. Their central argument is that each of the three reinforces the others, each promotes peace, and peace reinforces each of the three. Russett and Oneal (2001:337) indicate that “the reciprocal effects” of each of the three “are important in understanding the international system.” They add that “democracies are more likely to trade with one another.…Democracies are more likely than authoritarian states to form and to join many international organizations” and those organizations reinforce democracy (Russett and Oneal 2001:38). They add that “a high level of interdependence among states is likely to create a need for institutions to manage and stabilize their commercial relations.” Given that progress in this area of empirical international relations inquiry has occurred, it is reasonable to assert that progress is possible—ab esse ad posse—in other areas of the field as well. It must be emphasized that the position suggested here in no way excludes a role for interpretive analysis of international relations nor for normative theorizing. Whether statistical, rational choice models, or interpretive analysis is most appropriate depends upon the nature of the question the theorist or policymaker seeks to answer. Normative questions of equity and morals are clearly essential to policymaking. Theories of international relations that are applied to policymaking should explore and clarify the concepts involved and advance the soundest normative principles. Indeed, policy decisions cannot be made without consideration of moral questions; but moral principles cannot be effectively applied unless one has the best possible idea of how the world works and what associations and causal relationships obtain for the policy goals being pursued. The aim of this essay is not to endorse some or all of the democratic peace hypotheses. The author accepts that any adequate metatheory must be fallibilist; hence, however strong the consensus is on the dyadic hypothesis at the present time, this hypothesis may be proven false at a later date. As Duhem would point out, there may be some other factor that explains both the development of democracy and patterns of interstate conflict better than the democratic peace (as the recent article by Rosato 2003 attempts). The best-supported scientific findings throughout history are eventually modified. Furthermore, in the future there may be a spate of wars between democracies, which would certainly dash claims of a democratic peace. (It would take relatively few cases—of established, mature, economically interdependent, liberal democracies going to war—to destroy the statistical basis for the democratic peace hypotheses.) The principle conclusion of this essay is, rather, that quasi-naturalist methodological and foundational approaches that draw strong parallels between the natural sciences and some categories of international relations theory and envision the possibility of scientific progress may well be on the right track, given that, however one defines it, democratic peace studies do appear, indeed, to manifest scientific progress. References Ashley Richard C. . ( 1986 ) The Poverty of Neorealism . In Neorealism and Its Critics , edited by Keohane Robert O. . New York : Columbia University Press . Babst Dean . ( 1964 ) Elective Governments: A Force for Peace . Wisconsin Sociologist 3 : 9 – 14 . Babst Dean . ( 1972 ) A Force for Peace . Industrial Research 14 : 55 – 58 . Barbieri Katherine . ( 2002 ) The Liberal Illusion: Does Trade Promote Peace? Ann Arbor : University of Michigan Press . Beck Nathaniel Katz Jonathan N. . ( 2001 ) Throwing Out the Baby with the Bath Water: A Comment on Green, Kim, and Yoon . International Organization 55 : 487 – 495 . Google Scholar CrossRef Search ADS Beck Nathaniel Katz Jonathan N. Tucker Richard . ( 1998 ) Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable . American Journal of Political Science 42 : 1260 – 1288 . Google Scholar CrossRef Search ADS Bennett D. Scott Stam Alan C. III. . ( 1998 ) The Declining Advantages of Democracy: A Combined Model of War Outcomes and Duration . Journal of Conflict Resolution 42 : 344 – 366 . Google Scholar CrossRef Search ADS Bohman James . ( 1993 ) The New Philosophy of Social Science: Problems of Indeterminacy . Cambridge, MA : MIT Press . Bueno De Mesquita Bruce . ( 2002 ) Predicting Politics . Columbus : Ohio State University Press . Bueno De Mesquita Bruce Lalman David . ( 1992 ) War and Reason . New Haven : Yale University Press . Bueno De Mesquita Bruce Siverson Randolph . ( 1995 ) War and the Survival of Political Leaders: A Comparative Study of Regime Types and Political Accountability . American Political Science Review 89 : 841 – 853 . Google Scholar CrossRef Search ADS Bueno De Mesquita Bruce Siverson Randolph Woller Gary . ( 1992 ) War and the Fate of Regimes: A Comparative Analysis . American Political Science Review 86 : 638 – 646 . Google Scholar CrossRef Search ADS Bueno De Mesquita Bruce Morrow James D. Siverson Randolph M. Smith Alastair . ( 1999 ) An Institutional Explanation of the Democratic Peace . American Political Science Review 93 : 791 – 807 . Google Scholar CrossRef Search ADS Carr Edward Hallett . ( 1964 ) The Twenty Years' Crisis, 1919–1939: An Introduction to the Study of International Relations . New York : Harper . Chan Steve . ( 1984 ) Mirror, Mirror on the Wall…Are the Freer Countries More Pacific? Journal of Conflict Resolution 28 : 617 – 648 . Google Scholar CrossRef Search ADS Chan Steve . ( 1993 ) Democracy and War: Some Thoughts on Future Research Agenda . International Interactions 18 : 205 – 213 . Google Scholar CrossRef Search ADS Chan Steve . ( 1997 ) In Search of Democratic Peace: Problems and Promise . Mershon International Studies Review 41 : 59 – 91 . Google Scholar CrossRef Search ADS Chernoff Fred . ( 2002 ) Scientific Realism as a Meta-Theory of International Relations . International Studies Quarterly 46 : 189 – 207 . Google Scholar CrossRef Search ADS Chernoff Fred . ( 2004 ) The Power of International Theory . London : Routledge . Christensen Thomas J. Snyder Jack . ( 1997 ) Progressive Research on Degenerate Alliances . American Political Science Review 91 : 1919 – 1922 . Google Scholar CrossRef Search ADS Dixon William J. . ( 1993 ) Democracy and the Management of International Conflict . Journal of Conflict Resolution 37 : 42 – 68 . Google Scholar CrossRef Search ADS Dixon William J. . ( 2003 ) Commercial Liberalism on Trial . International Studies Review 5 : 364 – 366 . Google Scholar CrossRef Search ADS Doyle Michael . ( 1983a ) Kant, Liberal Legacies, and Foreign Affairs, Part 1 . Philosophy and Public Affairs 12 : 202 – 235 . Doyle Michael . ( 1983b ) Kant, Liberal Legacies, and Foreign Affairs, Part 2 . Philosophy and Public Affairs 12 : 323 – 353 . Doyle Michael . ( 1986 ) Liberalism and World Politics . American Political Science Review 80 : 1151 – 1161 . Google Scholar CrossRef Search ADS Duhem Pierre . ( 1954 ) The Aim and Structure of Physical Theory . Translated by Philip P. Weiner. Princeton : Princeton University Press . East Maurice A. Gregg Phillip M. . ( 1967 ) Factors Influencing Cooperation and Conflict in the International System . International Studies Quarterly 11 : 244 – 269 . Google Scholar CrossRef Search ADS Elkins Zachary . ( 2000 ) Gradations of Democracy? Empirical Tests of Alternative Conceptualizations . American Journal of Political Science 44 : 293 – 300 . Google Scholar CrossRef Search ADS Elman Colin Elman Miriam Fendius . ( 1997 ) Lakatos and Neorealism: A Reply to Vasquez . American Political Science Review 91 : 923 – 926 . Google Scholar CrossRef Search ADS Elman Colin Elman Miriam Fendius . ( 2002 ) How Not to Be Lakatos Intolerant: Appraising Progress in IR Research . International Studies Quarterly 46 : 231 – 262 . Google Scholar CrossRef Search ADS Enterline Andrew J. . ( 1996 ) Driving While Democratizing . International Security 20 : 183 – 196 . Farber Henry Gowa Joanne . ( 1997 ) Common Interests or Common Politics? Journal of Politics 59 : 393 – 417 . Google Scholar CrossRef Search ADS Feyerabend Paul K. . ( 1962 ) Against Method . London : Verso . Fodor Jerry . ( 1984 ) Observation Reconsidered . Philosophy of Science 51 : 23 – 43 . Google Scholar CrossRef Search ADS Garnham David . ( 1986 ) War Proneness, War Weariness and Regime Type, 1860–1980 . Journal of Peace Research 23 : 279 – 289 . Google Scholar CrossRef Search ADS Gartzke Erik . ( 1998 ) Kant We All Just Get Along? Opportunity, Willingness and the Origins of the Democratic Peace . American Journal of Political Science 42 : 1 – 27 . Google Scholar CrossRef Search ADS Gillies Donald . ( 1993 ) Philosophy of Science in the Twentieth Century: Four Central Themes . Oxford : Blackwell . Gleditsch Kristian Ward Michael D. . ( 1997 ) War and Peace in Space and Time: The Role of Democratization . International Studies Quarterly 44 : 1 – 29 . Google Scholar CrossRef Search ADS Gleditsch Nils Petter . ( 1995 ) Geography, Democracy, and Peace . International Interactions 20 : 297 – 323 . Google Scholar CrossRef Search ADS Gowa Joanne . ( 1999 ) Ballets and Bullets: The Elusive Democratic Peace . Princeton : Princeton University Press . Green Donald Kim Soo Yeon Yoon David . ( 2001 ) Dirty Pool . International Organization 55 : 441 – 468 . Google Scholar CrossRef Search ADS Greenwood John D. . ( 1990 ) Two Dogmas of Neoempiricism: The “Theory-Informity” of Observation and the Quine-Duhem Thesis . Philosophy of Science 57 : 553 – 574 . Google Scholar CrossRef Search ADS Habermas Jürgen . ( 1971 ) Knowledge and Human Interests . Translated by Jeremy J. Shapiro. Boston : Beacon Press . Hanson Norwood Russell . ( 1958 ) Patterns of Discovery: An Enquiry into the Conceptual Foundations of Science . London : Cambridge University Press . Hempel Carl G. . ( 1965 ) The Function of General Laws of History . In Aspects of Scientific Explanation . New York : Free Press . Henderson Errol R. . ( 2002 ) Democracy and War: The End of an Illusion? Boulder : Lynne Reinner . Hermann Margaret G. Kegley Charles W. Jr. . ( 2001 ) Democracies and Intervention: Is There a Danger Zone in the Democratic Peace? Journal of Peace Research 38 : 237 – 245 . Google Scholar CrossRef Search ADS Hewitt J. Joseph Wilkenfeld Jonathan . ( 1996 ) Democracies in International Crisis . International Interactions 22 : 123 – 142 . Google Scholar CrossRef Search ADS Hobson John A. . ( 1926 ) Free Thought in the Social Sciences . New York : Macmillan . Hudson Valerie M. . ( 2001 ) Manifest Agency, Foreign Policy Analysis, and IR Theory: The Other Agent-Structure Debate and Its (Highest) Stakes. Paper presented at the annual meeting of the International Studies Association, Chicago, IL, February 20–24. Hunt Shelby D. . ( 1994 ) A Realist Theory of Empirical Testing: Resolving the Theory-Ladenness/Objectivity Debate . Philosophy of the Social Sciences 24 : 133 – 158 . Google Scholar CrossRef Search ADS Huth Paul K. Allee Todd L. . ( 2003 ) Democratic Peace and Territorial Conflict in the Twentieth Century . Cambridge : Cambridge University Press . Jaggers Robert Gurr Ted . ( 1995 ) Tracking Democracy's Third Wave with the Polity III Data . Journal of Peace Research 32 : 469 – 482 . Google Scholar CrossRef Search ADS Jervis Robert . ( 1976 ) Perception and Misperception in International Politics . Princeton : Princeton University Press . Kant Immanuel . ( 1989 ) Idea for a Universal History from a Cosmopolitan Point of View. And, Perpetual Peace . In Kant on History , edited by Beck L. W. . New York : Macmillan/Library of Liberal Arts . Kegley Charles W. Jr. Hermann Margaret G. . ( 1995 ) Military Intervention and the Democratic Peace . International Interactions 21 : 1 – 21 . Google Scholar CrossRef Search ADS Kegley Charles W. Jr. Hermann Margaret G. . ( 1997 ) Putting Military Intervention into the Democratic Peace . Comparative Political Studies 30 : 78 – 107 . Google Scholar CrossRef Search ADS Keohane Robert O. Nye Joseph S. Jr. . ( 1977 ) Power and Interdependence . Boston : Little Brown . King Gary Keohane Robert O. Verba Sidney . ( 1994 ) Designing Social Inquiry: Scientific Inference in Qualitative Research . Princeton : Princeton University Press . Knorr-Cetina Karin . ( 1993 ) The Realism–Constructivism Debate . In Taking the Naturalistic Turn: Or How Real Philosophy of Science Is Done , edited by Callebaut Werner . Chicago : University of Chicago Press . Kuhn Thomas S. . ( 1962 ) The Structure of Scientific Revolutions . Chicago : University of Chicago Press . Lakatos Imre . ( 1978 ) The Methodology of Scientific Research Programmes: Philosophical Papers . Vol. 1 . Edited by Worrall John Gregory Currie . Cambridge : Cambridge University Press . Latour Bruno . ( 1987 ) Science in Action: How to Follow Scientists and Engineers through Society . Cambridge, MA : Harvard University Press . Layne Christopher . ( 1994 ) Kant or Cant? The Myth of Democratic Peace . International Security 18 : 5 – 49 . Google Scholar CrossRef Search ADS Leng Russell . ( 1993 ) Reciprocating Influence Strategies and Success in Interstate Crisis Bargaining . Journal of Conflict Resolution 37 : 3 – 41 . Google Scholar CrossRef Search ADS Levy Jack S. . ( 1989 ) The Causes of War: A Review of Theories and Evidence . In Behavior, Society, and Nuclear War , edited by Tetlock Philip E. Husbands Jo L. Jervis Robert Stern Paul S. Tilly Charles . New York : Oxford University Press . Levy Jack S. . ( 1994 ) The Democratic Peace Hypothesis: From Description to Explanation . Mershon International Studies Review 38 : 352 – 354 . Google Scholar CrossRef Search ADS Little Daniel . ( 1991 ) Varieties of Social Explanation . Boulder : Westview . Little Daniel . ( 1998 ) Microfoundations, Method, and Causation . New Brunswick, NJ: Transaction Publishers . Mansfield Edward D. Snyder Jack . ( 1995a ) Democratization and the Danger of War . International Security 20 : 5 – 38 . Google Scholar CrossRef Search ADS Mansfield Edward D. Snyder Jack . ( 1995b ) Democratization and War . Foreign Affairs 74 : 79 – 97 . Google Scholar CrossRef Search ADS Mansfield Edward D. Snyder Jack . ( 2002 ) Democratic Transitions, Institutional Strength, and War . International Organization 56 : 297 – 337 . Google Scholar CrossRef Search ADS Maoz Zeev . ( 1998 ) Realist and Cultural Critiques of Democratic Peace: A Theoretical and Empirical Re-Assessment . International Interactions 24 : 3 – 89 . Google Scholar CrossRef Search ADS Maoz Zeev Abdolali Nazrin . ( 1989 ) Regime Types and International Conflict . Journal of Conflict Resolution 33 : 3 – 35 . Google Scholar CrossRef Search ADS Maoz Zeev Russett Bruce . ( 1993 ) Normative and Structural Causes of Democratic Peace, 1946–1986 . American Political Science Review 87 : 624 – 638 . Google Scholar CrossRef Search ADS Midlarsky Manus I. . ( 1995 ) Environmental Influences on Democracy: Aridity, Warfare, and a Reversal of the Causal Arrow . Journal of Conflict Resolution 39 : 224 – 262 . Google Scholar CrossRef Search ADS Misak Cheryl K. . ( 1991 ) Truth and the End of Inquiry: A Peircean Account of Truth . Oxford : Clarendon . Mousseau Michael Shi Yuhang . ( 1999 ) A Test for Reverse Causality in the Democratic Peace Literature . Journal of Peace Research 36 : 639 – 663 . Google Scholar CrossRef Search ADS Nye Joseph S. Jr. . ( 2002 ) The Paradox of American Power: Why the World's Only Superpower Can't Go It Alone . New York : Oxford University Press . Oneal John R. Russett Bruce . ( 1997 ) The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985 . International Studies Quarterly 41 : 267 – 294 . Google Scholar CrossRef Search ADS Oneal John R. Russett Bruce . ( 1999a ) Is the Liberal Peace Just an Artifact of Cold War Interests? Assessing Recent Critiques . International Interactions 25 : 1 – 29 . Google Scholar CrossRef Search ADS Oneal John R. Russett Bruce . ( 1999b ) The Kantian Peace: The Pacific Benefits of Democracy, Interdependence, and International Organizations, 1885–1992 . World Politics 52 : 1 – 37 . Google Scholar CrossRef Search ADS Oneal John R. Russett Bruce . ( 2001 ) Clear and Clean: The Fixed Effects of Democracy and Economic Interdependence . International Organization 55 : 469 – 485 . Google Scholar CrossRef Search ADS Oren Ido . ( 1995 ) The Subjectivity of the “Democratic Peace . International Security 20 : 147 – 184 . Google Scholar CrossRef Search ADS Owen John . ( 1994 ) How Liberalism Produces Democratic Peace . International Security 19 : 87 – 125 . Google Scholar CrossRef Search ADS Owen John . ( 1997 ) Liberal Peace, Liberal War, American Politics and International Security . Ithaca : Cornell University Press . Peirce Charles S. . ( 19601966 ) Charles S. Peirce: Collected Papers . Edited by Hartshorne Charles Weiss Paul . Cambridge, MA : Harvard University Press . Peirce Charles S. . ( 1982 ) Writings of Charles S. Peirce: A Chronological Edition . Edited by Fisch Max H. . Bloomington : Indiana University Press . Poincaré Henri . ( 1905 ) Science and Hypothesis . Translated by George Bruce Halsted. New York : Science Press . Popper Karl . ( 1965 ) Conjectures and Refutations: The Growth of Scientific Knowledge . 2nd ed . New York : Basic Books . Popper Karl . ( 1968 ) The Logic of Scientific Discovery . 2nd ed . New York : Harper Torchbook . Quine Willard Van Ormand . ( 1953 ) Two Dogmas of Empiricism . In From a Logical Point of View . Cambridge, MA : Harvard University Press . Ray James Lee . ( 1996 ) Democracy and International Conflict: An Evaluation of the Democratic Peace Proposition . Columbia : University of South Carolina Press . Ray James Lee . ( 2000 ) Democracy: On the Level(s), Does Democracy Correlate with Peace? In What Do We Know about War , edited by Vasquez John A. . Lanham, MD : Rowman and Littlefield . Ray James Lee . ( 2003 ) A Lakatosian View of the Democratic Peace Research Programme: Does It Falsify Realism (or Neorealism)? In Progress in International Relations Theory: Appraising the Field , edited by Elman Colin Elman Miriam Fendius . Cambridge, MA : MIT Press . Raymond Gregory A. . ( 1994 ) Democracy, Disputes, and Third Party Intermediaries . Journal of Conflict Resolution 38 : 24 – 42 . Google Scholar CrossRef Search ADS Reiter Dan Stam Allan C. III. . ( 1998 ) Democracy, War Initiation, and Victory . American Political Science Review 92 : 377 – 389 . Google Scholar CrossRef Search ADS Reiter Dan Stam Allan C. III. . ( 2002 ) Democracies at War . Princeton : Princeton University Press . Reynolds Andrew . ( 2002 ) Peirce's Scientific Metaphysics: The Philosophy of Chance, Law, and Evolution . Nashville : Vanderbilt University Press . Richardson Lewis F. . ( 1960 ) The Statistics of Deadly Quarrels . Pittsburgh : Boxwood . Rosato Sebastian . ( 2003 ) The Flawed Logic of Democratic Peace Theory . American Political Science Review 97 : 585 – 602 . Google Scholar CrossRef Search ADS Rosenau James N. . ( 1980 ) The Scientific Study of International Relations . Revised edition . London : Pinter . Rousseau David Gelpi Christopher Reiter Dan Huth Paul . ( 1996 ) Assessing the Dyadic Nature of the Democratic Peace, 1918–1988 . American Political Science Review 90 : 512 – 533 . Google Scholar CrossRef Search ADS Rummel R. J. . ( 1979 ) War, Power, Peace . Vol. 4 of Understanding Conflict and War. Beverly Hills, CA : Sage . Rummel R. J. . ( 1981 ) The Just Peace . Vol. 5 of Understanding Conflict and War. Beverly Hills, CA : Sage . Rummel R. J. . ( 1983 ) Libertarianism and International Violence . Journal of Conflict Resolution 27 : 27 – 71 . Google Scholar CrossRef Search ADS Rummel R. J. . ( 1995 ) Democracies ARE Less Warlike Than Other Regimes . European Journal of International Relations 1 : 457 – 479 . Google Scholar CrossRef Search ADS Rummel R. J. . ( 1997 ) Power Kills: Democracy as a Method of Nonviolence . New Brunswick, NJ: Transaction Publishers . Russett Bruce . ( 1985 ) The Mysterious Case of Vanishing Hegemony; or, Is Mark Twain Really Dead? International Organization 39 : 457 – 479 . Google Scholar CrossRef Search ADS Russett Bruce . ( 1990 ) Controlling the Sword: The Democratic Governance of National Security . Cambridge, MA : Harvard University Press . Russett Bruce . ( 1993 ) Grasping the Democratic Peace: Principles for a Post-Cold War World . Princeton : Princeton University Press . Russett Bruce . ( 1995 ) The Democratic Peace: And Yet It Moves . International Security 19 : 164 – 175 . Google Scholar CrossRef Search ADS Russett Bruce Oneal John R. . ( 2001 ) Triangulating Peace: Democracy, Trade, and International Organizations . New York : W. W. Norton . Russett Bruce Starr Harvey . ( 1981 ) World Politics: The Menu for Choice . San Francisco : Freeman . Schweller Randall L. . ( 1997 ) New Realist Research on Alliances: Refining, Not Refuting, Waltz's Balancing Proposition . American Political Science Review 91 : 927 – 930 . Google Scholar CrossRef Search ADS Senese Paul D. . ( 1997 ) Between Dispute and War: The Effect of Joint Democracy on Interstate Conflict Escalation . Journal of Politics 59 : 1 – 27 . Google Scholar CrossRef Search ADS Singer J. David . ( 1969 ) The Incompleat Theorist: Insight without Evidence . In Contending Approaches to International Politics , edited by Rosenau James N. Knorr Klaus . Princeton : Princeton University Press . Singer J. David Small Melvin . ( 1972 ) The Correlates of War Project: An Interim Report and Rationale . World Politics 24 : 243 – 270 . Google Scholar CrossRef Search ADS Singer J. David Small Melvin . ( 1982 ) Resort to Arms: International and Civil Wars 1816–1929 . Los Angeles : Sage . Siverson Randolph M. . ( 1995 ) Democracies and War Participation: In Defense of the Institutional Constraints Argument . European Journal of International Relations 1 : 481 – 489 . Google Scholar CrossRef Search ADS Small Melvin Singer J. David . ( 1976 ) The War-Proneness of Democratic Regimes . Jerusalem Journal of International Relations 1 : 50 – 69 . Spiro David E. . ( 1994 ) The Insignificance of the Liberal Peace . International Security 19 : 50 – 86 . Google Scholar CrossRef Search ADS Starr Harvey . ( 1997 ) Democracy and Integration: Why Democracies Don't Fight Each Other . Journal of Peace Research 34 : 153 – 162 . Google Scholar CrossRef Search ADS Taylor Charles . ( 1985 ) Interpretation and the Sciences of Man . In Interpretation and the Human Sciences: Philosophical Papers . Vol. 2 . Cambridge : Cambridge University Press . Thompson William R. . ( 1996 ) Democracy and Peace: Putting the Cart before the Horse? International Organization 50 : 141 – 174 . Google Scholar CrossRef Search ADS Thompson William R. Tucker Richard . ( 1997 ) A Tale of Two Democratic Peace Critiques . Journal of Conflict Resolutions 41 : 428 – 454 . Google Scholar CrossRef Search ADS Tures John A. . ( 2001 ) Democracies as Intervening States: A Critique of Kegley and Hermann . Journal of Peace Research 38 : 227 – 235 . Google Scholar CrossRef Search ADS Vasquez John A. . ( 1997 ) The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neotraditional Research on Waltz's Balancing Proposition . American Political Science Review 91 : 899 – 912 . Google Scholar CrossRef Search ADS Vasquez John A. . ( 1998 ) The Power of Power Politics: From Classical Realism to Neotraditionalism . Cambridge : Cambridge University Press . Waltz Kenneth N. . ( 1997 ) Evaluating Theories . American Political Science Review 91 : 913 – 917 . Google Scholar CrossRef Search ADS Wayman Frank W. Diehl Paul F. . ( 1994 ) Realism Reconsidered . In Reconstructing Realpolitik , edited by Wayman Frank W. Diehl Paul F. . Ann Arbor : University of Michigan Press . Weart Spencer R. . ( 1998 ) Never at War: Why Democracies Will Not Fight One Another . New Haven : Yale University Press . Weede Erich . ( 1984 ) Democracy and War Involvement . Journal of Conflict Resolution 28 : 649 – 664 . Google Scholar CrossRef Search ADS Weede Erich . ( 1992 ) Some Simple Calculations on Democracy and War Involvement . Journal of Peace Research 29 : 377 – 383 . Google Scholar CrossRef Search ADS Wendt Alexander . ( 1999 ) Social Theory of International Politics . Cambridge : Cambridge University Press . Wight Colin . ( 1996 ) Incommensurability and Cross-Paradigm Communications in International Relations Theory: What's the Frequency, Kenneth? Millennium: Journal of International Studies 25 : 291 – 319 . Google Scholar CrossRef Search ADS Wittgenstein Ludwig . ( 1953 ) Philosophical Investigations . Translated by G. E. M. Anscombe. Oxford : Blackwell . Wyn-Jones Richard . ( 1999 ) Security, Strategy, and Critical Theory . Boulder : Lynne Rienner . Zakaria Fareed . ( 1997 ) The Rise of Illiberal Democracy . Foreign Affairs 76 : 22 – 43 . Google Scholar CrossRef Search ADS Footnotes 1 Thanks are due James Lee Ray, Bruce Russett, and John Vasquez for their illuminating comments on and criticisms of earlier drafts of this paper and to Claire Putzeys and Emily Weedon for valuable research assistance. 2 Lakatos (1978)“reconstructs” three stages in the evolution of Popper's methodology; most of what Lakatos defends can be found, he argues, somewhere in the second and third stages of Popper's work. 3 Of course, one could argue that realists and liberals are not really functioning within different research paradigms at all or that the notion of paradigm cannot be applied in international relations (Ray 2003). However, if these moves are made, then the skepticism regarding progress to which Kuhn subscribes disappears even before it is subject to scrutiny. 4 The trade variable has been the most controversial of those Russett and Oneal (2001) defend. Barbieri (2002) has argued that the relationship between trade and peace is an illusion. Russett and Oneal, in turn, question her data. Part of this debate appears to stem from their different treatment of missing cases and rules for case inclusion (see Dixon 2003). © 2004 International Studies Review.
TI - The Study of Democratic Peace and Progress in International Relations
JF - International Studies Review
DO - 10.1111/j.1079-1760.2004.00372.x
DA - 2004-05-20
UR - https://www.deepdyve.com/lp/oxford-university-press/the-study-of-democratic-peace-and-progress-in-international-relations-DNytl04N0t
SP - 1
EP - 77
VL - Advance Article
IS - 1
DP - DeepDyve
ER -