Part One: Smurfs and The Logic of the Democratic Peace Thesis (Academic Clickbait)

The first part of my research landed me in a myriad of tangents, some of which were eventually relevant to the big picture (did you know that Britain and Iceland had militarized conflicts over fish for 20 years, with Iceland threatening to leave NATO each time? And that Britain also helped the Soviet Union attack Finland?), which is my excuse for publishing this blog post so late. See my abstract here for a sense of what I am trying to accomplish. I have added endnotes for irrelevant authorial commentary and to explain logic jargon, of which there is a small amount. Because of the methodology I selected, this report has evolved into a creature longer than any academic paper I have ever written. I was forced to cut corners when I realized that if I continued to review every explanation of the DPT I came across, I would probably not be able to finish my project on time. This is by no means a perfect paper. In fact, it is probably riddled with holes. Consider it an extremely rough draft of a working paper.

I will start by saying that I was never a fan of the Democratic Peace Theory (DPT), which claims that it is highly unlikely for democratic states to go to war with each other though they are just as likely as non-democratic states to fight wars in general, because it has no inherent logic, is often explained with weak inductive models, is not falsifiable, and has a reputation for inspiring foreign policy decisions that have often led to undesirable consequences. Because the cogency of the DPT is questionable, I will follow the example of many other critics to refer to it as the “Democratic Peace Thesis” from here on when not using the abbreviation. Though there is no inherent logic for why pairs of democratic states would be less likely to fight each other, the Democratic Peace (the observation that less wars have occurred between democracies than between other types of dyads) is indeed an empirical observation. It is the practical significance of this correlation between democracies and peace that is debatable. The search for an explanation for the Democratic Peace has spawned an abundance of research that supports, discredits, or offers alternative theories for the DPT. In this report, I will be evaluating the logic of the different strands of arguments within this ongoing conversation using the relevant logic models.

Close examination of the DPT discussion reveals that democracy and peace are often both arbitrarily defined. Most studies appear to only consider “mature representative democracies” in the realm of democracies that do not attack each other. The arguments for why such democracies do not attack each other reveal other underlying criteria that are not shared by all such polities and may apply to polities of other types. To avoid confusion, I will be using the abstract term “Smurf” to indicate a “state which fulfills certain criteria for being treated as a ‘democratic’ state by an explanation for the Democratic Peace Thesis”. (Note that the criteria a “smurf” must meet may be different for each explanation for the Democratic Peace)1. The goal of this endeavor is to find out 1) what makes a state a smurf, 2) What mechanisms allegedly cause peace between smurfs, and 3) How effective these mechanisms are for preventing war with other smurfs. During the data collection part of my research, I came to realize that I had vastly underestimated the amount of interest Western scholars had in the DPT. There are simply too many studies and too many explanations2, so I have limited my range of analysis to the following (note that “limited” is a relative concept): Democratic nature, normative and structural, audience cost/selectorate theory, mutually beneficial institutions, and constructivism.

The Democratic Nature Explanation

The democratic nature explanation for the democratic peace says that democratic states do not fight each other because they are inherently more peaceful than non-democratic states. This argument is easily the most easily disproven, but is also the most well known in “mainstream” politics. It has roots in Immanuel Kant’s original theory of perpetual peace, which says that republics, which are polities with a system of representation and separation of government and legislature, are not prone to initiating wars (Kant 1795). The logic is as follows;

1)Smurfs are states with representative governments and separation of executive and legislature that require the consent of the people to make decisions (True)

2)All people are negatively affected by having to bear the consequences of war (True)

3)If a war happens, all or most people must bear the consequences (False)

4)The public will is rational (Dubious)

5)Because people do not want to bear the consequences of war, they would not consent to engaging in military conflict except in response to threats of potentially being attacked(False)

 Conflict between two or more smurfs would most likely not turn violent for the lack of an initiator of violence (Uncogent)

Premise 1 is true because of the previously given definition of “smurf”. Premise 2 can be proven true via reduction to absurdity)3 :

  1. Let F be a person who bears the consequences of war but is not negatively affected
  2. No consequences of war, such as economic downturn, death, and destruction affect F or all consequences affect F positively
  3. 2) is absurd because F cannot be unaffected or positively affected by all the consequences of war while also bearing consequences.

There cannot be a person who lives with the consequences of war who is not negatively affected

Note that this does not contradict the assessment that Premise 3 is false. Not all wars consume a large amount of resources and some could end up with consequences that only affect a small percentage of the population. Small border wars, especially when victorious, are a good example of this.

Premise 4 is an interesting case. In international relations, like in economics and most other social sciences, there is a base assumption that actors are rational and self-interested. At the systemic level of analysis this is not debated. At the individual level, this is not considered strictly true because individuals suffer from perception biases and may be unable to optimize every decision based on available information. Groups, on the other hand, are supposed to balance out individual biases but suffer from the danger of “groupthink” (the tendency to go along with what one thinks the group is most likely to think), higher risk tolerance, vulnerability to partisan cues, and so on. This is the case with the public will. A recent article that talks about how public opinion is shaped mentions that partisan media is polarizing public opinion, though each subset of the public forms rational opinions based on the information they are exposed to through their selected sources (Gelpi 2017). From this one can see that the public will can be divided, but is mostly rational within each division, though that would mean the public will in general is not rational if it is divided, because each subset of the public holds selective perceptions that they failed to balance out. Premise 4 could still potentially be true if the publics in question are usually exposed to non-partisan media or are able to come to a consensus that eliminates polarity.

If we assume that Premise 4 is false, then Premise 5 is automatically false because it presumes Premise 4 to be true. But if we assume that Premise 4 is true, Premise 5 is still false. The public makes rational decisions from cost-benefit analysis. If the (perceived) benefits of going to war outweigh the costs, people would likely support a war. People also support going to war for many reasons other than self-defense, including territorial expansion/recapture, ideological crusades, and aiding allies.

This is not to mention that historically, smurfs have engaged in plenty of wars for reasons other than responding to threats of potentially being attacked (Example: The war of the Eight-Nation Alliance against Qing Dynasty China), which completely debunks the argument that smurfs are more peaceful in general (which holds true even though it is possible that I have overlooked more logically robust versions of this argument)

Norms and Structure

The normative and structural arguments are usually considered inseparable. The normative explanation claims that two conflicting smurfs would recognize and each other as smurfs and respect their shared preferences for peaceful resolution. The structural explanation claims that due to “democratic institutions” such as division of power or the constraint of public opinion, smurfs have a longer decision-making process than non-smurfs, which other smurfs know and recognize, giving smurfs more time to seek out a peaceful resolution while not expecting the other smurf to make a sudden attack. This is also supposed to explain why smurfs behave differently towards each other than towards non-smurfs. A previous study conveniently maps out this logic in diagram form (Zinnes 2004).

dpt1 dpt2

The relevant characteristic of a “democracy” that Zinnes identifies is the convention of bargaining and compromise between multiple actors, including the public, which is used to settle all conflicts, including external conflicts. Non-democracies, on the other hand, do not employ this convention (Zinnes 2004). Though these characteristics may not be true of every “democracy”, it is the definition that can be used for “smurf” in this context, regardless of whether all smurfs are considered democracies.

The simplified  logic evaluation of the normative-structural argument is as follows:

  1. All states that employ the convention of bargaining and compromise between the government and the people or their representatives to settle conflicts are smurfs (True)
  2. Smurfs recognize that other smurfs have the same conventions for conflict resolution, which creates long processes of decision making that buys time for resolving conflict peacefully (Dubious)
  3. Because of Premise 2, smurfs do not feel threatened by other smurfs (Likely true if Premise 2 is true)

Smurfs do not fight other smurfs (Cogent if all premises are true)

Premise 1 is true because it is a definitional premise and Premise 3 and the first conclusion are likely to be true if Premise 2 is true. Premise 2 is dubious because it requires that all smurfs recognize other smurfs as smurfs, which may not always be possible in the real world, especially if one smurf already has the perception of a non-existent threat from another smurf. It also implies that smurfs always decide to resolve conflicts peacefully through the process of bargaining and compromise. What can guarantee reciprocity for using such conventions? Premise 2 should have these conditions added to it to make it true.

The argument can be changed to:

  1. There is a circle of smurfs that recognize and accept each other as smurfs who are committed to peace(True)
  2. Smurfs within this circle of trust do not perceive other members of the circle as threats if there is an open declaration of commitment to peaceful resolution whenever there is a crisis that could escalate to war, which can be checked by other members of the circle of trust. (True)

Smurfs that mutually recognize and accept each other as smurfs committed to peace will most likely not fight wars among themselves (Cogent)

This circle of trust is beginning to look like a security alliance, and the argument now seems to have constructivist overtones.

So how does this argument perform in the real world? The popularity of the DPT among NATO and its allies suggests the existence of such a circle of trust among these states, all of which fulfill the criteria for being “mature” democracies by the most widely accepted definition. However, as Mearsheimer and Walt point out in an article about theory testing, if the normative-structural explanation is true, there should be evidence of conflict between smurfs that was deescalated for the reasons provided in the argument (Mearsheimer and Walt 2013), but this evidence has not been found.

Audience Cost, Selectorate Theory, and Democratic Constraints

James Fearon popularized the audience cost explanation, which says that leaders of smurfs bear a higher audience cost (possibility of being punished by the public) than leaders of non-smurfs for backing down from threats or reneging on a promise, which makes them less likely to bluff, and in turn, makes their threats more credible. Their credibility is known by other states, which includes other smurfs. When conflict occurs between two smurfs, the leaders of smurfs, knowing the credibility of the other party, would avoid escalating threats to avoid a war (Fearon 1994).

The selectorate theory explanation (Bueno de Mesquita, Morrow, Siverson and Smith, 1999) points out that the public often feels more strongly about policy success than leaders backing down from threats. It offers that all state leaders must satisfy a “winning coalition”, or the people who were responsible for putting the leaders in power. Smaller coalitions are easier to satisfy than larger coalitions. Thus, the bigger the winning coalition, the more effectively a leader’s imprudent actions can be punished by the winning coalition. This argument claims to be “agnostic” about democratic ideals, and does not conclude that democracies avoid fighting each other, but rather that democratic dyads have fought less wars than other types of dyads, which is attributed to the lack of support from the public unless there is a large probability of victory.

Democratic constraints, or the constraint of public opinion, which I mentioned earlier in the evaluation of the generic explanation, is a related topic that focuses on the degree to which the public can effectively punish leaders for unpopular policies, which depends on the willingness and ability of the public to obtain and analyze credible, non-partisan information and form rational opinions about it (Gelpi 2017). So in the case of audience cost and selectorate theory, criteria for identifying a smurf should include the availability, transparency, and non-partisanship of information about foreign affairs to the general public.

From the above information I have obtained the following:

  1. Smurfs are all states where the public has some means of punishing their leaders for decisions they can rationally conclude were harmful to their interests from a large amount of transparent and non-partisan information on foreign policy. (True)
  2. To remain in power, leaders of smurfs must satisfy the public (True)
  3. The people of smurfs punish their leaders for wasting resources on wars which they believe are unlikely to result in victory (True)
  4. For fear of punishment, leaders of smurfs do not escalate conflicts when they do not think they have a good chance of victory (Likely true)

In conflict involving two smurfs, war can only happen when at least one smurf sees a good chance of victory (Cogent)

This argument passes the logic test, but for Premise 4 to be true, there needs to be empirical evidence of leaders of smurfs refraining from escalating a security crisis because the chances of victory in a war were very low. An example that comes to mind is the Cuban Missile Crisis, though that may not be a good example because of the complicated dynamics that nuclear threats present.

The main criticism for selectorate theory is that it requires too much of the public and their sources of information. While it sounds reasonable on paper, it is difficult in real life to find states where the government and its linkage institutions are transparent and neutral enough to not confound public opinion with misleading cues. It is more difficult to find a population of people who are so well trained and interested in politics that they are able to form rational opinions about every action their government takes.

But how is all this relevant to the Democratic Peace? From the previous evaluation, where I found that the argument for smurfs in a security alliance not fighting each other was logically cogent, I derive:

  1. If a security alliance is sufficiently beneficial to a smurf, the people of that smurf will hold their leaders responsible for maintaining amiable relations with other members of the alliance to preserve benefits. (True)
  2. Leaders of smurfs face punishment for acting aggressively toward other smurfs in the alliance (True)

Leaders of smurfs avoid escalating conflicts with other smurfs to avoid punishment (Cogent)

The Constructivist Perspective

IR scholars in China are almost unanimously skeptical of the DPT. The consensus is that while the Democratic Peace may be an empirical observation, the fact that some Western governments are using the DPT to guide foreign policy is turning the DPT into a self-fulfilling prophecy, which also poses a security problem for polities that are not “mature representative democracies”. In other words, democratic values do not create peace between democracies. Rather, the appeal of the DPT led “mature representative democracies” to adhere to a custom of not attacking other democracies, creating a bias towards the “out-group” and an incentive to convert the out-group (to create world peace) in the process. One article warned of “crusade mentality” (Liang and Ding 2005).

International institutions of all types run on the liberal ideal of cooperation. Whenever there is some form of cooperation, there is also some form of commitment to maintain non-hostile relations, though the strength of this commitment varies. NATO is a particularly strong security alliance because it was held together by a common threat for decades (the USSR). When that common threat disappeared, NATO and its allies found a new source of unity in the DPT, and the Democratic Peace custom was created (Pang 1995) . The people of each “mature representative democracy” then came to believe that democracies should behave peacefully toward each other, effectively providing a public opinion constraint against aggression towards other mature representative democracies.

The logic of a constructed peace is consistent with the results of the previous logic evaluations. Basically:


  1. Smurfs are states that are “mature representative democracies” (True)
  2. Smurfs recognize each other as smurfs (True)
  3. Smurfs accept the DPT as a norm of interstate behavior (Possibly true)
  4. Citizens of smurfs see the alliance with other smurfs as beneficial (Possibly true)
  5. Citizens of smurfs believe that war with other smurfs is wrong, and will punish leaders for aggression towards other smurfs (Possibly true)

Smurfs are not likely to fight other smurfs (Strong and cogent if all premises are true)


Though this argument is logically strong, statistical evidence is needed to verify some of these premises, which exceeds the scope of this project. For now, this is where I will conclude.


The next update will probably be a lot shorter. 


  1. I chose “smurf” for two reasons: “Smurf” can substitute any concept in the Smurf language, and no known wars have been fought between two communities of smurfs (the little blue elves), excepting one civil war that resulted in zero casualties and was quickly resolved. Many factors could be responsible for the peace between smurfs, just like there are many factors responsible for the peace between arbitrarily defined democracies. 
  2. In logic, explanations and arguments are separate concepts. Arguments give evidence to prove that a conclusion is true, while explanations give reasons for why a conclusion that was already proven/assumed to be true is true. I will be using the two terms interchangeably because the explanations I evaluate are presented in argumentative form.
  3. Reduction to absurdity is a technique that proves or disproves a hypothesis by showing that the opposite hypothesis is ridiculous in some way. I included this proof for the purpose of structure. I know perfectly well I could have just said that having to bear consequences is almost by definition a negative effect.


Works Cited:

Bueno De Mesquita, Bruce, Michael T. Koch, and Randolph M. Siverson. “Testing Competing Institutional Explanations of the Democratic Peace: The Case of Dispute Duration.” Conflict Management and Peace Science 21, no. 4 (2004): 255–267.

Ding Liang(丁亮) and Liang Tao(梁涛). “The Negative Effects of the Democratic Peace Theory” (民主和平论的负面影响). Guoji Guanxi Xueyuan Xuebao (Beijing)no.2 (2005): 1-4

Fearon, J. D. “DOMESTIC POLITICAL AUDIENCES AND THE ESCALATION OF INTERNATIONAL DISPUTES.” American Political Science Review 88, no. 3 (1994): 577–592.

Gelpi, Christopher. “Democracies in Conflict.” The Journal of Conflict Resolution 61, no. 9 (2017): 1925–1949.

Kant, Immanuel. Perpetual Peace: A Philosophic Essay, 1795.

Mearsheimer, John J., and Stephen M. Walt. “Leaving Theory behind: Why Simplistic Hypothesis Testing Is Bad for International Relations.” European Journal of International Relations 19, no. 3 (September 1, 2013): 427–57.

Pang Zhongying (庞中英). “Several Suggestions for the ‘Democratic Peace Theory” (对“民主和平论”的若干意见). Ouzhou (Beijing) , no. 6 (1995): 62-65.

Zinnes, Dina A. “Constructing Political Logic: The Democratic Peace Puzzle.” Journal of Conflict Resolution 48, no. 3 (2004): 430–454.


  1. cmaufderheide says:

    Wow, this is a really interesting piece! I love the logical, clear-cut arrangement of your work – it makes it really easy to read. It’s very intriguing that empirical evidence suggests the DPT, but the individual pieces of the logic which would prove the DPT true, as you laid out, are so difficult to definitively nail down. I’m looking forward to reading the rest of your posts!

    I was wondering while reading this if you know of any studies (in China or otherwise) that address possible application of/evidence for the DPT prior to NATO? In my readings on revolutionary France I came across references to the idea that all republics (all smurfs, so to speak) – referring to France and the US specifically at the time – would, just by virtue of being such, naturally be allies that would be at peace with each other or come to each other’s aid. Then again, this certainly fits your security alliance observation, given both’s common enemies at that time. On that note, the assertion that the DPT will be self-fulling if governments use it to guide foreign policy is deeply intriguing, and certainly relevant to NATO. If this has always been something of the case (smurfs labeling their allies as other smurfs and vice versa) how can the DPT ever actually be itself assessed/proven or disproven?

  2. Artemis Fang says:

    Hey Chela! Thanks for commenting.
    When were these readings you speak of published? If they were published after 1795, I am willing to bet they were influenced by Immanuel Kant’s theory of Perpetual Peace, which states exactly that two republics would not fight each other by virtue of being republics. (If they were published before 1795, then my sources would be wrong to say that Kant was the first to suggest that there is no war between smurfs). Kant was of course quite influential during his time, so it would be understandable for French revolutionaries to use his ideas as a call for solidarity. And reasonable for other smurfs (such as the US) to reciprocate largely due to their similar historical experiences (see, hyper-realists would say this assessment is ridiculous because states don’t have feelings. But people do).
    There are other observations of peaceful resolution of conflict between “democracies”, but a lot of the critiques dismiss these as trivial figures because there were plenty of militarized conflicts between other alleged democracies (and the proponents respond by saying “they weren’t real wars” or “they weren’t real democracies”). The true decrease in chance of wars between smurfs came around the early 20th century, and the first modern DPT theorists cited the lack of smurfs opposing each other in WWII as why they thought smurfs had some sort of internal mechanism that made their relations with other smurfs more peaceful, but they also said the sample size was too small to conclusively say anything about it.
    (I would actually argue that the alliance effect was preestablished in the case of the Allied Powers because of economic ties formed from imperialism and such. Germany and Italy certainly would have had this relationship with the Allies, but their economies were destroyed by WWI, which is the entire reason they went Fascist in the first place)
    So that leaves the Cold War. And even while the Cold War was going on there were 3 militarized skirmishes (DPT proponents:NOT REAL WARS!) in the course of 20 years between Britain and Iceland because of codfish of all things, with Iceland winning each time after threatening to leave NATO if Britain didn’t comply with their wishes. This can be explained as Iceland not really caring about NATO because the USSR didn’t really threaten them, but staying because having allies is nice and NATO status could be used as a bargaining chip since the rest of NATO cared about Iceland as a strategic heartland against the USSR.
    I don’t really know if the DPT can be proved or disproved at all, but the Constructivist perspective is so far the most logical. If there can be a non-constructed peace between smurfs, it can only be proven in a hands-off, non-interventionist world, because otherwise there would be too many confounding variables. And there would also need to be proof that non-smurfs are in some way completely unable to interact with other states the way smurfs interact with other smurfs. But I don’t know if there can ever be such a future.

  3. cmaufderheide says:

    Thank you so much for taking the time to write such an in-depth response! There are clearly a lot of factors at play which make the issue of proving the DPT based on past/present politics highly complicated. You really sum it up when you point out that proponents of the DPT can always respond to proof of conflict between smurfs with “they weren’t real wars” or “they weren’t real democracies”.

    I’m reading back through the mention of France and the US – it’s in the play by Desbarreaux, when the American soldier shows up (not a very serious source I’ll grant you) – and I’d say the references are roughly half “we are allies” and half references to shared freedom, both being free citizens, both countries fighting for people’s rights, etcetera. Not anything like articulating the DPT along the lines of Kant (though this does predate his work by a year), but definitely a strong sense that the two republics/democracies are allied intrinsically by being such (which makes sense in this context, where both are still threatened by Europe’s monarchies.) I’m pretty sure that such DPT-esque discourse regarding France and the US continued well into the 1800s. (Of course, this could also be framed as similar political systems being likely to ally…?)