In recent years research has pointed to the possibility that democracies are more peaceful than other kinds of regimes. If true, such a finding would lend support to liberal theorists who argue that it is domestic politics and institutions shape outcomes of inter-state relations and not, as realists maintain, the material dimensions of the international system. Even the narrower claim that democracies are not more peaceful in general but do not go to war with each other has important implications for the practice as well as the study of international relations. For American policymakers, the premise of a 'democratic' peace is already embedded in US foreign policy.

This article will not assess the evidence and arguments for and against the existence of some form of democratic peace which have emerged in addressing this issue.² Instead attention will be directed to a major conceptual that needs to be clarified and its implications for additional research addressed. Specifically, the issue of how we define peace is of fundamental importance but has not received adequate attention. It is obvious that refinement of the democratic peace thesis will require a better specification of different types of peace to replace the simple distinction between 'war' and 'peace'. I have in mind here a more profound conceptual and causal inquiry than the standard quibble over whether 1000 battle deaths, or any other arbitrary cut-off point, should be used to define 'war' in quantitative data sets.³ It may be of some interest in this connection to indicate the development of my interest in this issue.

In 1992 when Professor Shimon Shamir (Tel Aviv University) and I were fellows at the United States Institute of Peace we discussed how best to characterize the state of peace that had emerged in Israeli-Egyptian relations. was clear that some way of identifying different types of peace was needed to replace the simple distinction between war and peace.

---

* Alexander L. George, Department of Political Science, Stanford University, Stanford, CA 94305, USA.

Professor Shamir's preferred typology at that time was a four-fold distinction between 'adversarial peace', 'restricted peace', 'rapprochement', and 'cooperative peace'. Somewhat dissatisfied with this typology, I suggested an alternative a three-fold distinction between 'precarious peace', 'conditional peace', and 'stable peace'.⁴ Indeed over time a number of other typologies have been advanced. Clearly there is a need to bring together the different concepts and terms being used by different scholars to distinguish of peace in order to show the extent to which they overlap and to expose various ambiguities in their definitions. If a common, shared set of concepts can be developed, it will benefit research by providing a basis for systematic, cumulative research on this important question. It has not been entirely clear when investigators were in agreement or disagreement in applying elements of their different typologies to empirical cases.

My own typology was influenced by important writings of scholars, in particular Karl Deutsch (1957) and Kenneth Boulding (1979). Deutsch's concept of 'peace' pointed in the right direction. Deutsch's classic description a 'security community' emphasized that the peace it brought with it was based on, among other things, 'the real assurance that the members of that community will not fight each other physically, but will settle their disputes in some other way'. This identifies a core element of the definition of 'stable peace'. And his emphasis on the importance of developing 'community' remains of prime importance, though whether it is either a necessity or a sufficient condition for the emergence of stable peace in all situations needs to be subjected to empirical testing (Deutsch 1957, 5). I found Deutsch's concept of peace to be somewhat ambiguous, however, and his various definitions of it inconsistent. Boulding's concept of 'stable peace' was quite useful but in need of clarification and additional specification. defined peace as a situation in which the probability of war is so small that it does not enter into the calculations of any of the people involved (Boulding 1979, 13). The full research program envisaged by Deutsch and his colleagues was never completed, though several books were published after his major publication.⁵

Balanced assessments of Deutsch's seminal contribution have been provided a number of scholars, quite recently by Emanuel Adler and Michael Barnett (1998). Perhaps the most systematic follow-up to Deutsch's book was the important study by Stephen R. Rock, Why Peace Breaks Out (1989). But I find inconsistency and some ambiguity in his treatment as well.

As Erik G. Yesson reminds us, all these definitions evoke Kant's insistence his classic work, Perpetual Peace, that peace is not a 'suspension of hostilities' but, rather, 'an end to all hostilities', which means the nullification 'all existing reasons for a future war' (Yesson 1995, 5).

I would like to clarify now my own three-fold typology which will

emphasize the extent to which peace depends upon deterrent and compellantthreats.

'Precarious peace' is a relationship of acute conflict between two states who have already engaged in warfare in the past and/or have been and still are on the verge of major war. At least one state is dissatisfied with the status quo and one or both see the use of military force as legitimate for either defending or changing the status quo. 'Peace', therefore, means little more then the temporary absence of armed conflict. Such a peace depends not merely on 'general deterrence', a term introduced into the literature some years ago by Patrick Morgan. Maintenance of a 'precarious peace' between two adversaries may require frequent use of 'immediate deterrence', is military alerts and deployments, issuance of deterrence threats in war-threatening crises. Arab-Israeli relationships until recent times and the India-Pakistan relationship over many years are examples of 'precarious peace'.

'Conditional Peace', on the other hand, describes a substantially less acute, less precarious conflict relationship. General deterrence plays the predominant role in maintaining peace except in quite infrequent crises or pre-crisis situations in which one or both sides feel it necessary to resort to activities that provide immediate deterrence to avoid outbreak of war. The US-Soviet relationship during the Cold War qualifies as an example of 'conditional peace'. During that era there were occasional but infrequent diplomatic crises, over Berlin and Cuba, for example, in which general deterrence was supplemented with immediate deterrence.

Neither in precarious peace nor in conditional peace does either side rule
out initiating military force as an instrument of policy, and deterrent and
compellant threats of doing so do occur.

'Stable Peace' is a relationship between two states in which neither side considers employing force, or even making a threat of force in any dispute, even serious disputes between them. Deterrence and compellance backed by threats of military force are simply excluded as an instrument of policy. Two states that enjoy stable peace may continue to have serious disputes but they share a firm understanding that such disputes must be dealt with by non-military means. For example, in the Suez Crisis of 1956 President Eisenhower made strong, credible threats of economic sanctions to pressure the British to withdraw their forces from the Suez.

This typology and indeed other typologies are conceptual in the first instance. As in any typology, it can only be the starting point for attempting to characterize actual relationships between states and to undertake empirical research. Types should not be reified; they should not be imposed on historical cases in a mechanical, simplistic way that obscuresrelevant and complexities. The test of a typology should be whether it facilitates empirical research and development of

theory. A comprehensive research program on this fundamental aspect of
international relations entails a number of questions and problems which I
will now address.

One of these is the task of determining whether two states - whether or not they are democracies - enjoy a genuine stable peace. This may be difficult discern for various reasons and it is a matter of finding ways to distinguish between the existence of conditional peace and stable peace. The continued absence of war and war-threatening crises in a relationship, however in and of itself, is not sufficient to establish the existence of stable peace. Peace between two such states may not yet have been subjected to tough tests, disputes severe enough to stimulate one side or the other to consider or make use of immediate deterrence. In fact, if one sees beneath the surface of peace that the military on one or the other side is still preparing contingency plans of a serious kind for possible use of force, then one must question whether stable peace really exists. In such cases general deterrence may still play a role, though not a conspicuous one, in backstopping what appears to be stable peace.

And though the peace appears to be stable, leaders and publics on both sides may feel that it is not a sufficiently cordial relationship including all desired forms, activities, and institutions of a cooperative nature such as confidence-building measures, cooperation on non-security issues, and dispute-resolution Thus, Israeli scholars have felt it necessary to distinguish between 'cold peace' and 'warm peace' to call attention to the fact that Israel and Egypt have never managed to develop the kinds of interactions with each other that include the full repertoire of warm, friendly relations between neighbors. Can one say, nonetheless, that stable peace exists between Israel and Egypt? Has peace between them been subjected tough tests? Does either side have contingency plans for possible use of force or for purposes of backing up immediate deterrence threats should they become necessary in a future crisis?

One may take note of the possibility, too, that although the dominant leadership on both sides may enjoy what appears to be stable peace and believes in and acts in accord with the requirements of stable peace, important elements of the elites or counter-elites and of the public still regard the other side as posing a latent threat to its security. When this suspicion prevails, stable peace may be vulnerable. Such a state of affairs may characterize US-Russian relations since the end of the Cold War. Certainly leaders and elements of their publics have moved from the conditional that characterized US-Soviet relations during the Cold War towards stable peace, but important elements of the political-military elite and of the publics on both sides evidently question whether general deterrence no longer necessary and whether the possible need for resort to immediate deterrence in the future can be safely excluded.

A better example of stable peace is the relationship among most of the Western European countries embraced by the European Union and NATO, a development in the post-1945 era which engaged the interest of Karl Deutsch and his colleagues, and many others.

The research agenda should also include study of the conditions under which and the processes by which states move from a relationship of precarious conditional peace to one of stable peace. There may be many paths to stable peace: negotiated settlements, regime transitions (especially democratization), demographic changes, changes in military, economic and transportation technologies, social or normative changes, and so on. There are few studies of this kind as yet and many historical examples of such a development that should be studied and compared. A leading example, of course, is the already mentioned emergence of the security community in Western Europe.

Some years ago I asked Magnus Jerneck, then visiting Stanford University, Swedish or other Scandinavian scholars had studied the transition to stable peace in relations between Scandinavian countries. Jerneck, a political scientist at Lund University, checked with colleagues in the History Department at Lund and was told that, although the phenomenon well known, no systematic studies of it existed. Accordingly, Jerneck and several of his colleagues formed an interdisciplinary research team that has undertaken research on this problem.⁶ Other possible examples Argentina-Brazil, South Africa and its neighbors after the end of apartheid, and US relations with Canada and Mexico.

Obviously, the interest in stable peace - its emergence, what it is based upon, how it can be recognized, etc. - overlaps with the democratic peace thesis, which has received a great deal of attention and discussion, particularly the United States. Much of the scholarly attention has focused on efforts to explain what it is about being a democratic polity that is the basis for the absence of war between two states. Not enough attention has been given to study of historical transitions in relations between democratic states which have resulted in stable peace between them. It may matter, for example, whether one state in the dyad in question became a democracy through civil war, international war, revolution, occupation, or gradual political development.

In fact, much of the research addressed to explaining what it is about democracy that explains why democratic states do not engage in war with each other does not appear to distinguish clearly between conditional peace and stable peace. The task of unravelling the blurring of the difference between these two types of democratic peace would be facilitated if more attention were given to historical studies of transitions to stable peace. An exemplary study of this is Stephen Rock's recent study (Rock 2000) of how the British employed a strategy of appeasement of the United States

towards the end of the nineteenth century to remove the serious warthreateningdisputes
their relations, thus paving the way from conditional
to stable peace in their relationship.

Another example of study of such a transition is offered by the earlier Deutsch study and by many others that have traced the development of stable peace among Western European countries. Perhaps more so than the Deutsch study itself, these other studies have focused on deliberate efforts after World War II to create the attitudes, policies, and structures for a new peaceful relationship between France and Germany.⁷ Such studies are important because they indicate that efforts and strategies can be adopted to bring into being a relationship of stable peace.

Studies are needed of many other cases of transition to democratic peace. For example, the considerable research already available on relations of the United States with Canada and Mexico should be reviewed in order to identify and explain critical turning points that led to what seems clearly to have become stable peace.

Broad generalizations about conditions and processes that have led to stable peace in different situations may be possible, but it would be well to act on the presumption that this process, like so many other phenomena in international relations, is subject to equifinality (referred, to as 'multiple causation' by some scholars). That is, similar outcomes (i.e. stable peace) can occur through different causal processes. It would be counterproductive to good comparative research to assume that similar outcomes in different cases must have a common explanation. Even when a common factor can be identified, in many cases the question remains whether that is either a necessary or sufficient condition for the emergence of stable peace and how much causal weight can be attributed to it.

What I have been suggesting is that it is best to regard the 'democratic peace' phenomenon as a subset of the broader general phenomenon of stable peace. In this connection I would like to raise the question of whether stable peace is possible only and has occurred only between countries that are democracies. A more comprehensive research program would look for historical cases of stable peace between countries that are not democracies or between states only one of which is a democracy. Some of the research on 'zones of peace' by Arie Kacowicz, reported in his earlier publication and referred to in the forthcoming volume he is editing, moves in this direction. It is important to apply the distinction between conditional and stable peace in such studies also.

Finally, I believe it is important that a full research program should include efforts to judge whether lessons can be drawn from historical studies which may be of relevance for efforts to move relations between adversarial states to stable peace, or at least to something approximating it. Several years ago when I was preparing a foreword to James Goodby's book Europe

Undivided, I was struck by the fact that he was addressing the desirability and feasibility of moving US-Russian relations from conditional peace to stable peace. Following the publication of his book the United States Institute of Peace has set up a working group to develop further Goodby's ideas. This study will examine several alternative future developments affecting the nature and scope of security in the Euroatlantic community. It will consider whether and how a democratic Russia can become a member of a Euroatlantic security community, all members of which enjoy stable peace.

In sum, there are ambiguities and inconsistencies in definitions of the concept of 'peace' in research that addresses the possibility of 'democratic peace'. This major conceptual issue needs to be addressed and clarified given its important implications for scholarship and policy. This paper proposes a distinction between three types of peace: 'precarious peace', 'conditional peace', and 'stable peace'. It is important to clarify whether presumed instances of 'democratic peace' blur the important distinction between conditional and stable peace. This paper also identifies a number of important problems that should be part of a research program on democratic peace, which is usefully regarded as a subtype of stable peace.

---

NOTES

1. The author wishes to thank the Carnegie Corporation of New York for funds that supported the research reported here, and Professor Andrew Bennett for helpful comments suggestions on an earlier draft.

2. For a detailed summary and commentary on theoretical and methodological issues associated with this research program, see Bennett & George (1997). A revision of this paper will appear in the book they have in preparation on the use of case studies for theory development.

3. The concept of 'democracy' - its defining characteristics and how to distinguish between different types of democracy - requires specification, since this influences the selection of historical cases for testing the peace thesis. This issue has received increasing attention from scholars, for it is clear that how it is resolved has a distinct bearing on the extent to which research findings support the peace thesis.

4. For Shamir's typology, see his remarks as reported in the U. S. Institute of Peace Journal (Shamir 1992). My own typology is briefly noted in the same issue. This paper draws also on my presentation on the Grawmeyer Panel at the annual International Studies Association Meeting in Toronto in March 1997 and from the foreword I wrote for Ambassador Goodby's Europe Divided (1998).

5. See Lindgren (1959), Russett (1963) and Katzenstein (1976).

6. An early study was prepared by Magnus Ericson (1997). Other studies will appear in the forthcoming volume being edited by Arie Kacowicz, Yaacov Bar Simon Tov, Magnus Jerneck and Ole Elgstrom, provisionally entitled Stable Peace: Dimensions, Conditions, and Prospects for Success.

7. This literature is too numerous to cite here (see, for example, Willis 1965). It includes, of course, the initiatives of Jean Monnet and others to bring about reconciliation between and Germany and to create structures for cooperative economic development. it also includes the contribution of non-governmental actors and organizations such as Frank Buchman's moral rearmament movement.

REFERENCES

Adler, E. & Barnett, M. 1998. Security Communities. Cambridge: Cambridge University
Press.

Bennett, A. & George, A. L. 1997. 'Case Study Methods and Research on the Democratic
Peace', paper presented at the American Political Science Association Convention,
Washington, August.

Boulding, K. 1979. Stable Peace. Austin: University of Texas Press.

Deutsch, K. W. et al. 1957. Political Community and the North Atlantic Area. Princeton:
Princeton University Press.

Ericson, M. 'The Liberal Peace Meets History: The Scandinavian Experience', paper presented
at the annual meeting of the International Studies Association, Toronto, March 1997.

Goodby, J. 1998. Europe Undivided: The New Logic of Peace in U.S.-Russian Relations.
Washington: USIP Press.

Katzenstein, P. J. 1976. Disjoined Partners: Austria and Germany since 1815. Berkeley:
University of California Press.

Lindgren, R. E. 1959. Norway-Sweden: Union, Disunion, and Scandinavian Integration.
Princeton: Princeton University Press.

Rock, S. R. 1989. Why Peace Breaks out. Chapel Hill: University of North Carolina Press.

Russett, B. M. 1963. Community and Contention: Britain and American in the Twentieth
Century. Westport, CT: Greenwood Press.

Shamir, S. 1992. U.S. Institute of Peace Journal 5(6). Washington: USIP Press.

Willis, F. R. 1965. France, Germany, and the New Europe, 1945-1963. Stanford: Stanford
University Press.

Yesson, E. G. 1995. 'Sovereignty, Domestic Politics, and Stable Peace', revised version of
paper presented at the APSA annual meeting, September.