Spencer Weart advances the theory that democracies do not fight each other, a conclusion that has obvious implications for modern foreign policy as well as providing justificationfor the promotion of democracy around the world. Indeed, John McCain has advanced the idea of a League of Democracies as a foreign policy objective for his first term, apparently inspired largely by this book, along with Kagan's The Return of History and the End of Dreams. Critics have pointed out that historians have investigated this question before, and that to some degree, Weart has reinvented the wheel. Most historians who previously investigated the question did so with statistics - for example, by asking whether the fraction of wars between democracies is lower than what would be predicted by countries going to war independent of form of government. Weart instead proceeds based on exhaustive case studies of wars from classical antiquity to the modern day, searching for exceptions to falsify his assertion. Weart searches hardest for those cases which would falsify his claim that democracies do not fight each other and finds almost none out of hundreds of cases.
My issue is therefore not with his claim that democracies don't fight each other, but rather why. His argument boils down to a claim that democratic citizenries identify one another as fellow members of an in-group, and therefore that they should not fight each other. Weart's theory begs the question of a mechanism. That is to say, there are many steps between the perception by democratic citizenries of competitor states and the prosecution of wars against those states (or lack thereof) by those citizens' governments. One break in the chain of citizen perception-to-government action is that democratic citizenries, even modern ones with independent mass media, do not always have access to (or care about) the facts about those competitor governments, whether they be democratic or not (as Americans have learned recently and bitterly). Even in true democracies (as Weart defines them, with 2/3 of adult males able to participate and dissent tolerated), the will of the people does not immediately translate directly to an impact on foreign engagements and the prosecution of wars. That is to say, even in democracies, internal dissent can be extensive but still unable to stop a war, and the democratic process can be imperfect so that one group within a government can go to war despite the wishes of the rest. Weart's theory is also interesting in that it makes a case that, based on in-group thinking, democracies don't fight each other, but not the related claim that autocracies and democracies should not be allies. In fact this does occur, and frequently at that (e.g., the strange bedfellows of the Soviet Union and democratic allies in World War II, and the alliance of the U.S. and any number of foul Latin American dictators in the twentieth century). If in-group comradeship is really the mechanism, why don't democracies always find autocracies unacceptably repugnant?
Previous historians have treated this same problem in game theoretic terms. This seems the clearest way to ask the question, and here is why. Even in the most democratic nation, there is not an immediate translation from the will of the people to the actionsof the politicians responsible for such decisions - especially in highly non-direct democracies with millions of franchised citizens. Looked at from this angle, it becomes difficult to argue that the process of going to war in classic Athens (where you can show up to the assembly and give your two cents) is the same as going to war in the modern United States (where you're most likely stuck arguing with friends and posting to your blog). Therefore, even in democracies, wars result from the behavior of small groups of government officials over weeks or days. The leaders of the democracy may behave differently because they know their party can be voted out next year, but in the meantime the war has begun. Looked at this way, patterns of war and peace between types of government are likely to arise from the way each type plays the game. Type A makes assumptions about Type B's behavior, informed largely by its own decision-making processes, and therefore often wrong - but in the same ways throughout history.
Weart's in-group theory is meant only to explain the lack of wars between democracies and oligarchies, and cannot apply to autocratic aggression. In autocracies, the will of the people has no role in war decisions. Consequently he does in fact resort to a game theoretic explanation for one pattern of autocracy-on-democracy aggression: "appeasement traps", which describes democracies that foolishly assume they are bargaining with autocracies in good faith. The autocracy interprets the concessions as a sign of weakness, they attack the democracy, and the democracy retaliates with much more force than the autocracy anticipated. Interestingly, in a work focused on bad decision-making in foreign policy fiascos (Groupthink by Irving Janis), due to classic "groupthink", the author describes not only the colossal miscalculation of the U.S. of its own security against Japanese attack, but also the reasoning of the (autocratic) 1941 Japanese government in deciding to attack Pearl Harbor. As an autocracy, Japan knew how it would react if one of its own extreme outlying territories were surprise-attacked: it would have withdrawn. Why? If the territories were not critical for military or resource reasons, there was not strategic reason to distract the Imperial navy defending them; and yes, Japanese subjects would have been outraged at such a concession, but why would Tojo have cared? The Japanese government did not understand that the outrage of the American populace translated into a mandate for FDR's government to respond, even if it may have been more militarily expedient at the time to pull back and defend the west coast of North America, and avoid the subsequent entanglement with the war in Europe.
Another question this theory raises is whether "type of government" really forms the primary basis of in-group/out-group perception for most democratic citizens, even today. Except in the minds of the most enlightened moderns, it is hard to imagine that this characteristic is more salient than race, religion or language. Furthermore Weart makes a distinction between oligarchies and democracies, and points out that oligarchies and democracies frequently fight each other, but neither type fights its own. If Italian city-states in the fourteenth century fit the pattern of oligarchy vs. democracy but not oligarchy vs. oligarchy or democracy vs. democracy, and that the citizenry's perception of in-group/out-group based on form of government originates this pattern, then we must believe that fourteenth century Italians (who were not nearly all literate and did not have nearly as reliable access to information as we do) were able to somehow discriminate between oligarchies and democracies, a distinction that is subtle even for educated modern readers of political writing. Again, it seems safer to say that the decision-making bodies of oligarchies and democracies behave differently in their approaches to war based on the future consequence to themselves from their citizenry. In this sense, an autocracy is a form of government where the leaders assume there will never be any such consequences.
Weart concedes that the democratic peace theory is complicated by several other historical tendencies, among them the colonial-era tendency to dehumanize non-European and un-Christianized people. This my-tribe-at-all-costs tendency could also explain why states that undergo revolutions fight to keep their colonies, even when the new government is supposedly motivated by a very different ideology than the one which originally conquered those colonies (as in France, Russia and China, the latter of which in particular is a case study in the regression from universal Enlightenment economic philosophy to overt nationalism). It should also be said that the dataset that Weart has at his disposal consists almost entirely of a) classical Europe, b) post-Italian Renaissance Europe and c) post-Columbian America, if only because a theory about democracies based on historical data requires written records and of course the existence of at least two democracies. Therefore, there has been little chance for contact between democracies where both are not white and predominantly Christian. Though we would hope that both enlightened citizenries really would see past race and religion to the common bond of liberty, the experience of history (and Weart's own concession) suggest that this is not necessarily the case. Weart does mention the Korean War as an example of miscalculation by autocracies dealing with democracies, but after the war began it was prosecuted vigorously by the U.S. (in concert with the U.N.). Realistically, in the Korean War - was the American government - and people - really allied with South Korea because of shared (if newly inculcated) democratic values? Or was it just a strategic calculation by the small group of actors responsible for decisions in the United States government?
Perhaps where Weart could profitably expand his theory is by viewing war as one of many activities that states engage in to achieve objectives; and like war, all of these activities result from the decisions of a small group of actors that behave differently based on to what degree they are beholden to their populations, and who make consistent kinds of assumptions (sometimes erroneous) about each others' actions. Even expanding the definition of war to include proxy actions by client states expands the data set to include Cold War activities of the U.S. and Soviet Union. Indeed, it may occur to many readers that it is tempting from a vantage point so close to the end of the global, ideologically-framed struggle called the Cold War to view all wars as somehow stemming from ideological differences. The fact is that such overarching ideological framing is rare and perhaps the only other comparable instance is the Crusades. For example, it is difficult to see how the growing rivalry between the U.S. and China (the second Cold War) is ideologically driven at all, but rather a struggle for resources. It is also difficult to clarify exactly how the U.S. shared democratic values with the people of South Vietnam or Tibet (to whom the CIA provided extensive support, which was at the time a Buddhist theocracy). This kind of support is difficult to explain as in-group comradeship rather than convenient strategic alliances. Pre Cold-War examples of alliances that fly in the face of ideology are the British aiding the Confederacy after the Atlantic slave trade had been outlawed on humanitarian grounds, and the Roman Republic supporting the Numidians' provocations of civilized Carthage. Weart's argument as it stands does not consider these cases, but would be much more useful if it did.
Weart's conclusion that democracies do not fight each other is indisputable, but his mechanism for democratic peace is suspect. Whatever that mechanism is, it is still strongly in all of our interests - and therefore, over the long term, in the interest of our democratically elected leaders - to encourage demographic transition and the spread of liberal democracies.