Drezner vs. Slaughter

[Health warning: very long post with extensive quotes, on issues likely to be of primary concern to IR scholars and foreign policy wonks. Also, that owes a substantial intellectual debt to joint work with Cosma Shalizi, without in any way implicating him in its arguments]

There’s an interesting debate between Anne-Marie Slaughter and Dan Drezner over whether realism or networked internationalism best describes the basic contours of international politics.

Anne-Marie :

Traditional foreign policy continues to assume the world of World War II, the Cuban Missile Crisis, and the first and second Gulf Wars—an international system in which a limited number of states pursue their largely power-based interests in bargaining situations that are often zero-sum and in which the line between international and domestic politics is still discernible and defensible. … But Clinton herself insists that 21st century diplomacy must not only be government to government, but also government to society and society to society, in a process facilitated and legitimated by government. That much broader concept opens the door to a do-it-yourself foreign policy, in which individuals and groups can invent and execute an idea—for good or ill—that can affect their own and other countries in ways that once only governments could. … conference sponsored by Google Ideas, the Council on Foreign Relations, and the Tribeca Film Festival that brought together more than 80 former gang members, violent religious extremists, violent nationalist extremists, and violent white supremacists from 19 countries across six continents. … conference grew out of a vision developed by Jared Cohen, the head of Google Ideas, when he served in the U.S. State Department’s Office of Policy Planning … Personal Democracy Forum, a bottom-up gathering of digital activists from the U.S. and across the Middle East, including bloggers and organizers at the center of the Arab Spring. Other participants again included government officials, corporate executives, and the civic sector. … need for a framework that moves beyond states and addresses both governments and societies. Here complexity theory and network theory offer more answers than game theory; neuroscience, psychology, and sociology often provide better insights and solutions than political science and economics.

Dan :

actual great power foreign policies bear no resemblance whatsoever to that description of “traditional foreign policy.” Slaughter knows this very well, given that she was Clinton’s first director of policy planning. … much of her writing in international relations is about the ways in which traditional governments are becoming more networked and adaptive to emergent foreign policy concerns. … weakness as an international relations theorist is to uncritically observe phenomena like the Summit Against Violent Extremism and then inductively generalize from them to extrapolate the future of world politics. Slaughter believes that these “bottom-up” movements represent the future of world politics—and she may well be right. My own inclination is that DIY foreign policy represents a poor and underprovided substitute for effective state action global governance.

Anne-Marie again :

as realists/traditionalists never tire of repeating, the U.S.-China relationship is the most important global relationship of the 21st century: what matters most is ensuring that as both nations pursue their power-based interests they do not collide catastrophically. Never mind that an avian flu virus that is both fatal and aerosol-borne arising anywhere in Asia could do far more damage to global security and the economy than China ever could—just see the forthcoming movie Contagion. … The second approach tends to make our heads hurt; it often seems like a hopelessly complex system that is impossible to influence or shape. Waltz, rather than pushing us to develop workable models and policy prescriptions based on that complexity, says it’s okay just to white it all out.
many more concrete examples in the coming months, beginning with my next post, in which I will also answer Dan’s charge … that the rapidly increasing examples of social actors coming together, often with governments, to address foreign policy issues are bound to remain peripheral and ineffective.

Dan again

Anne-Marie places great faith in the ability of transnational, networked, non-state actors to bend the policy agenda to their preferred sets of solutions. I think that these groups can try to voice their demands for particular policy problems to be addressed. I think, at the national level, that social movements can force even recalcitrant politicians to alter their policy agenda (see: Party, Tea). Where Slaughter’s optimism runs into my skepticism is the ability of these movements to a) go transnational; and b) supply rather than demand global solutions. I’m skeptical about the viability of transnational interests to effectively pressure multiple governments to adopt a common policy solution, and I’m super-skeptical that these groups can supply broad-based solutions independently of national governments.

This is a useful debate – but one where I think both sides are wrong. Anne-Marie is right to argue that the world is becoming a more complex and messy place, in a way that makes realists’ heads explode. She depicts the world as one where networked relationships are replacing traditional forms of inter-state diplomacy, and argues (to my mind, entirely correctly) that network theory, the cognitive sciences etc can tell us more about what is happening than Waltzian realism.

I think though, that she goes astray in identifying the transition to a more networked world with a transition to a world with less zero-sum interactions. This is because Anne-Marie tends to see these networks in largely functionalist terms. Her previous – and highly influential work – on trans-governmental networks of cooperation between bureaucrats within national administration is an excellent example of this. She talks about the ways in which bureaucrats work across borders to achieve valuable policy goals. She worries about the ways in which this might weaken democratic legitimacy. But she doesn’t really ask whether bureaucrats might be motivated less by efficiency concerns than by power considerations.

Here, the empirics are starting to tell a different story. Abe Newman at Georgetown has forthcoming research which examines how SEC initiatives to encourage international cooperation have quite a lot to do with the SEC’s wish to escape (and, ultimately, to remake) domestic strictures on its authority. Its motivation for international cooperation can be explained well by its domestic struggle with other organization for decision making power and resources. If the kinds of international networked cooperation we see are all about struggles for resources, rather than achieving functionalist imperatives, then we may expect a very different international networked society than if these forms of cooperation are aimed at pursuing functionalist goals and Pareto-improvements. To be specific: we cannot expect one that will be characterized by worthy initiatives aimed at decreasing violence, bringing together nice activists and so on. Nor can we expect (I develop this point below) that such worthy initiatives will have much change of succeeding, to the extent that they come into being through government sponsorship or otherwise.

Dan’s take is, as Anne-Marie says, a state centered one. He’s doing what I’ll describe, after his own post, as the ‘subtle realist two-step.’ First – unlike many realists – he acknowledges that other factors than the international system can shape state preferences. It is certainly possible in his world that networked organizations can shape these preferences so that states will do things that they would not otherwise do. Second – and here he continues to agree with realism – is that he sees states as the crucial chokepoints in the international system. Even if non-state actors may influence states’ initial preferences, states still get the final say over what does or does not happen in international politics. His position is thus rather closer to that of Andy Moravcsik, Anne-Marie’s spouse and a vehement critic of realists, than it is, say, to John Mearsheimer’s. The result is his argument that:

All of the social movements and all of the online networks can agitate for policy solutions, but they’re not going to be able to alter fierce distributional conflicts that exist when trying to address many of the topline issues in world politics show no signs of abating. The kind of non-state actors that Slaughter embraces have not been shy in engaging issues like climate change, Israel/Palestine or macroeconomic imbalances—but I haven’t seen any appreciable change in global public policies as a result.

This is true up to a point. States haven’t gone away, and remain highly important. But what it misses is two things. First that state actions are not the only important vectors of influence in world politics. Second (and here Anne-Marie’s earlier work is correct) that states themselves are not unitary actors, but instead micro-networks of bureaucratic sub-units, each with their own agenda.

Networks demonstrably matter in international politics in ways that are not connected to state actions. The Arab Spring is an obvious example. The Tunisian upheaval’s influence on Egyptian politics, on Bahraini politics, on Libyan politics, on Syrian politics etc was not mediated through the preferences and actions of the Tunisian state. Instead, it occurred through processes of cross-national social emulation – activists in Egypt who both learned from the Tunisian example, and were able to use it as a focal point for their own activities. Obviously, the success of activists (and other groups that may opportunistically have latched onto this political moment) in other countries was mixed, and depended on domestic configurations of state and society. But this doesn’t change the fact that if we are to understand what happened in these countries, we have to put away our state centric theories. Social movements learned from each other through international networks and media, which had demonstrably important political consequences, and states (which would obviously have liked other outcomes) had to run to catch up. Outcomes would have been very different if e.g. al Jazeera, blogs, YouTube and more humble forms of communication had not existed, so that the individual populations of these countries had no very reliable way of learning what was happening across their borders except what their state authorities allowed them to know.

Other examples can be found in Dan’s work on regulatory politics, which also adopts the subtle-realist two-step. His arguments have the advantage of simplifying a complicated international space, but by the same token obscure important causal relationships. For example – Dan argues that regulatory politics are set by great powers, and that when great powers disagree on regulations, the result will be either an effective stalemate, or rival standards, which win or lose depending on how many third party states decide to conform to them. But this fails to recognize e.g. the ways in which bureaucrats within one large state – when confronted with regulatory disagreements with another state that hamper their work but that cannot be forced to back down using power politics – use a variety of non-diplomatic tools (including the widening and exploitation of divisions within the other state’s government, efforts to influence self-regulatory standards, ‘trading up’ tools to reshape business preferences etc) in order to try to change politics within the target country. Nor does it acknowledge the key role of non-state actors in setting standards, which often have repercussions for state laws and domestic enforcement systems. The international order is far more complex than he suggests, offering a larger set of actors a larger set of tools.

So if Anne-Marie depicts the world as one where networked relations are overtaking traditional power maneuvering between states, and argues for a problem-solving networked approach to US public policy. Dan argues that the traditional power politics approach still works just fine with a bit of tweaking. If Anne-Marie gets it wrong on the ‘networks as self-starting problem solvers’ bit, and Dan gets it wrong on the ‘networks don’t count, except insofar as they fit into traditional inter-state politics’ bit, is there a better alternative?

I think there is. Rather than seeing the international sphere as a space for inter-state power politics, or as a space for networked common action, we can think of it as a space for contagion. That is, think of it as a space where ever-multiplying and ever-ramifying sets of networked relationships across border serve not to enable problem-solving DIY diplomatists, but instead to transmit social influences in ways that are difficult to predict ex ante. This would mean taking seriously the kinds of complexity theory and network theory arguments that Anne-Marie mentions, but following them to a quite different set of conclusions than she does.

The world that complexity theory and network theory depicts is one where actions have highly unpredictable consequences. This follows both from theoretical arguments about processes of contagion across large scale networks, and from empirical research conducted via e.g. experiments.

The theoretical work has straightforward implications that are largely unappealing to policy makers. One can draw some useful conclusions about how network topology (the shape of the overall network) makes processes of contagion more or less likely. One can identify nodes that play a much more significant role than others in determining when contagion succeeds or fails (the degree of influence within a network will likely correlate well with more traditional indices of power). However, one cannot come up with anything that even begins to approximate a useful guide to liberal internationalist policy makers about which innovative actions are likely to succeed, and which to fail, or about how to set up or encourage networks that will help carry out useful social tasks. The work of Duncan Watts on music taste formation provides some empirical evidence to back this up (his recent book is highly recommended and very accessible). Watts finds that socially networked processes massively magnify the influence of early ‘chance’ events, in ways that make it very hard to predict which cultural trends will take off, and which will wither on the vine, and which also deviate from what one might think would be the ‘right’ outcome.

This suggests a story something like the following. Thanks to both the Internet and more traditional forms of information diffusion, the world is becoming more interconnected. Cross-national networks play a much more important role in conveying information, and in allowing some forms of social coordination to take place, than they used to. This means that states – although their power has surely not withered – find themselves in a much more complex environment than they used to, where they are far less able to serve as gatekeepers between their domestic space and the world of international and cross-national politics. In particular, this world is much more likely to be subject to contagion, in which political ideas, forms of organizing, and other social influences cross national borders with consequences that may (a) sometimes be substantial, but (b) are extremely difficult to predict in advance.

Such contagion may, or may not, lead to actors actively coordinating across borders, pursuing international goals etc. But coordinated action is not a necessary condition for causal efficacy – even when actors pursue purely domestic goals, their actions may have international repercussions, thanks to cross-national forms of communication and interaction.

Just because the world has become more networked, it does not mean that states can either (a) easily use networks to pursue their policy goals, or (b) turn over responsibilities to networks that will self-organize around socially useful tasks and responsibilities. To the extent that networks’ politics are predictable, they will conform to the same kinds of (frequently unpleasant) politics as do states. That is, they will be characterized by power inequalities (sometimes gross), actors pursuing their self-interest while entirely blind to the needs of others, and the rest of the shebang. To the extent that networks’ politics are unpredictable, they will be unlikely to be useful tools of policy.

This is a story with far fewer helpful policy lessons than either Dan’s or Anne-Marie’s. It points to plausible developments in world politics, without providing any very obvious tools to deal with them. It suggests that states are likely to get less good at managing things because of networked politics, but that networked politics itself offers no very obvious substitute, without far-reaching social and political changes (e.g. in addressing fundamental power imbalances) that networked politics itself is unlikely to achieve.

I think that it’s a story that captures the world that we live in better than either Dan or Anne-Marie’s. But it’s one where the increasing difficulty of of bold, structuring policy measures with predictable results is baked into the underlying theory. This is, perhaps the crucial distinction between it and both Dan and Anne-Marie’s accounts, which together presuppose that IR scholars should help the US government shape the world in ways that are congenial to US interests and/or international wellbeing (the two are usually indistinguishable in DC policy discussions). Both are focused on the purported solution providers – states (in particular the US) or networks – rather than the problems. I think that the interesting and important changes are happening in the set of problems, rather than the set of solutions.

15 Responses to Drezner vs. Slaughter

  1. renegade August 15, 2011 at 3:40 pm #

    Henry, you argue: “this world is much more likely to be subject to contagion, in which political ideas, forms of organizing, and other social influences cross national borders with consequences that may (a) sometimes be substantial, but (b) are extremely difficult to predict in advance.”

    I understand the logic of this but as an empirical matter, is this true? Surely you would need to show a spike in contagion. Otherwise there is nothing to be explained. The 20th Century saw great contagion of ideas through networks, especially through the spread of communism but also fascism, Islamism, and democracy. There have been instances of contagion following the emergence of these new social networking technologies but so far they have not been nearly as consequential as, say, the Russian revolution or 1989. To listen to this whole debate it’s as if contagion of ideas is a truly novel development that never occurred before.

    I guess my question is, is there an uptick in contagion? And if there is, how do we know it is part of these new technologies and not just a continuation of a 20th Century trend?

  2. Justin August 15, 2011 at 4:24 pm #

    The “unlike many realists” passage above is an absurd straw man:


  3. Henry Farrell August 15, 2011 at 4:49 pm #

    Justin – fair enough that I overstated this, but the claim that this is an “absurd straw man” is itself an absurd straw man. As Dan himself acknowledges there are some reasons why people caricature this. More specifically, I would only change the relevant sentence to “unlike many realists, Dan acknowledges that other factors than the international system can shape state preferences _in theoretically important ways.”_ Waltz doesn’t do this – hence his distinction between international relations, which one can theorize about, and foreign policy, which one can’t. Waltz’s claim seems to me to involve a rather theological set of notions about what constitutes theory and what does not – but that is an entirely different debate.

  4. Henry Farrell August 15, 2011 at 4:58 pm #

    renegade – as I note above, this is a “story” which I think captures the world better than the two alternative stories mentioned above. I don’t know of any good and direct empirical tests of it. But it has an excellent basis in theory – _ceteris paribus_, the more ties there are between nodes, the more likely it is that contagion will occur, some considerable secondary evidence in its favor – e.g. the clear causal role that a combination of satellite TV and social media played in the Arab Spring, and an implausibility of counter-factuals – do we think that we would see so much contagion so swiftly if e.g. we were limited to 19th century communication technologies.

  5. Justin August 15, 2011 at 5:06 pm #

    Thanks for the response. But Waltz doesn’t say you can’t theorize about foreign policy. (Why wouldn’t you be able to theorize about something?) He just says his theory isn’t a theory of foreign policy. His exchange with Colin Elman in Security Studies is useful on this score.

    I grant that this is a bugbear of mine, but I think it’s a common misread of Waltz that makes his theory look a bit ridiculous, which I don’t think it is.

  6. Henry Farrell August 15, 2011 at 5:32 pm #

    Justin – Waltz says in the dialogue you cite:

    bq. The most satisfying way [to explain states’ behavior] would be to provide a single theory capable of explaining the behavior of states, their interactions, and international outcomes. Unfortunately, no one has even suggested how such a grand theory can be constructed, let alone developed one. Someone may one day fashion a unified theory of internal and external politics. Until that day comes, the theoretical separation of domestic and international politics need not bother us unduly.

    I don’t think it’s at all unfair to read this as a claim that until the day when the grand overarching theory of domestic and international politics is proclaimed from the rooftops, the angel sounds the seventh trump and all of that, foreign policy making will not be amenable to proper Theory, and it is best for serious scholars of international relations not to try to theorize it.

  7. James Moore August 15, 2011 at 7:16 pm #

    I think there’s an interesting development that’s being missed here – non-government level organizations are starting to get (or possibly have built and used) serious hard-power tools. Take a look at Stuxnet, and think about what it says about international relations if it wasn’t created by a government. If the Iranian government can’t build a reliable air-gap system around critical infrastructure, can anyone? Presumably they could have had (or did, I have no way of knowing) policies like “if you bring a usb device past this point, you will be tortured and killed without exception.”

    And I seriously doubt that anyone thinks something like Stuxnet is government-level hard any more.

    But even the simple stuff, like the DOS attacks from various Anonymous-style pseudo-organizations matter. They’re small scale, but they have real-world effects, and they’re just the beginning.

  8. Zach August 15, 2011 at 11:34 pm #

    I think the notion that networked actors can influence states’ agendas and shape the incentives for different policy actions is perfectly consistent with Drezner’s position; it does not seem to require that states be the only vector of change, only that it function as a chokepoint and that for networked actors to have a meaningful effect it must be mediated through a state.

    For example, in your Arab Spring example, the Tunisian upheaval’s influence on Egyptian politics was still mediated through the Egyptian state. Egyptian protestors defined their goals in opposition to the state and had to persuade or compel the state to undertake reforms in order to achieve them. States still served as chokepoints for change. By the same token, networked actors like MNCs, industry groups, and standard consortia can often influence the international standards process and shape the incentives for adopting a given standard, but ultimately (usually) national governments are required to implement and enforce them.

  9. Henry Farrell August 16, 2011 at 11:30 am #

    Zach – the point of the two-step, as I read it, is that the action with non-state actors all happens at the domestic level, and thus is filtered through states, which remain the key actors in international politics. The point I’m making is that when there are transnational vectors of influence, this assumption breaks down. There are two claims which are bound together in Dan’s post, but which are, I think, analytically distinct. The first is that states are the chokepoint between the domestic and international realm. The second is that states are a chokepoint for ‘change,’ however you want to measure it. The first of these is, I think, pretty clearly contradicted by the Arab Spring events. The second might or might not be, depending on how you define, measure etc, and what happens over the longer term.

  10. Henry Farrell August 16, 2011 at 11:32 am #

    James – the Stuxnet example is not a good one – one of the reasons that techies figured that it had to be a government, is because it relies on three or four zero-day exploits bundled together, which suggests a level of organization and patience that loosely affiliated hackers are unlikely to have. Nonetheless, there is an interesting broader point. The best take I have read on this is actually Charlie Stross’s sf novel, _Halting State._

  11. Bear Braumoeller August 16, 2011 at 2:20 pm #

    I’d agree with much of the above, except that there aren’t tools to deal with it. Zeev Maoz’s book Networks of Nations is a very good one; I’ve got a forthcoming book on systemic politics from Cambridge that’s based on a boundedly-rational variant of complexity theory, which gets at how constituencies and domestic politics influence structural outcomes… but the problem that I found is not one of tools but one of endogeneity, which Henry hints at but doesn’t make as explicit as it could be.

    Let me put it this way. Your argument is that transnational networks of people exist and influence each other. But who influences whom, and to what degree? How do different ideas for a climate change policy, a new idea of Soviet socialism, or whatever coalesce from a universe of possibilities held by multiple actors into a single idea? If the answer, as it often is with transnational networks, is, “everybody influences everybody, at least potentially,” the problem, in a nutshell, is that you don’t actually have any independent variables unless you bring in something other than ideas to explain the outcome. You actually begin to run up against that problem everyone learns in intro statistics courses, that you need enough degrees of freedom to run an analysis. Something has to be exogenous, for us to be able to get a handle on it. Bruce Bueno de Mesquita, for example, uses the influence of the actors to get a handle on how much pull their ideas will eventually have.

    I don’t mean to get too far down this particular rathole… just trying to make a straightforward network-theory point: the outcome of an unconstrained “everybody influences everybody” transnational network story is pretty much a priori indeterminate. It may be as important, when you’re talking about transnational networks, to specify who ISN’T influenced as it is to specify who is.

  12. Henry Farrell August 17, 2011 at 5:27 pm #

    Bear – I’m most familiar with this problem from the work that people like Jure Leskovec are doing with large scale data on diffusion via the Internet. They have a ‘by their bootstraps’ approach that is interesting – study the data over time, to make reasonable imputations about the underlying networks of influence, which one can then use to make predictions about how influence will work in the future. Of course, the problems in going from there to generalizable causal arguments of the sort that social scientists are comfortable with are quite formidable.

    • Bear Braumoeller August 18, 2011 at 1:31 pm #

      Actually, yes, credit where credit is due, and thanks. But doing it inductively, with a ton of data, in some ways a more formalized version of what Dan seems to be complaining about: inductively identifying patterns and then extrapolating from them the future of world politics.

      I’m really just trying to make a plea for basic, uncontroversial research design. Complexity theory isn’t really just a way of saying “sh*t happens” with Greek letters; it’s a form of specifying very precisely what happens to what and how, and under what circumstances. And what I’ve noticed is that, particularly with complexity theory and transnational networks for some reason, we seem to write off that part of what we do—which is really bad, because it’s really important.

      So, to take a concrete example, I don’t think you’re allowed to be called a foreign policy analyst of any kind if you haven’t, at this point, at least given a dutiful nod to the Arab Spring as an exemplar of how transnational networks of ideas have changed the face of modern world politics. But as Anne Appelbaum pointed out in Slate, something very similar happened in 1848, without Twitter… which should make us question how much the world has actually changed, at least a little. Moreover, a comparison of the two sets of revolutions points to enough heterogeneity in goals within each to make us question how coherent the transnational network that arguably fueled them really is.

      Even leaving that point aside, what about the Anglo-Saxon Summer—the riots that started in London and, fueled by anger about the economy and popular outrage about austerity measures, spread to the other capitols of Europe? Those, of course, haven’t happened, despite not too dissimilar circumstances (an unjust death that painted officialdom in a very unpopular light as a catalyzing event; relative deprivation and youth as the tinder). If it had, it would most likely be held up as the next example, and Twitter (or Blackberry) as its means of propagation. Which raises an alarming possibility: does that mean that transnational networks are only likely to be defined as transnational networks in the first place when they succeed in having an impact?

      I’m not saying that transnational networks don’t happen, or that they don’t matter. I’m really fascinated by the subject. I’m just saying that in this area, even more than most due to the complex empirical implications, we have to look particularly hard at the non-barking dogs.

  13. James Moore August 17, 2011 at 7:42 pm #

    I agree that there’s a larger issue, but don’t think there’s much consensus about government origins for Stuxnet – check out Bruce Schneier’s article at http://www.schneier.com/blog/archives/2010/10/stuxnet.html . Basically, it’s hard to find serious sources that have good evidence of government origins. Probably interesting only for historians, though – it’s unlikely to get resolved in the immediate future.

    Personally, I don’t think the argument for losely connected groups is as interesting as the argument for closely-connected small groups acting together. 5-10 _really good_ programmers could be extremely effective for any particular cause.

  14. Hume's bastard August 30, 2011 at 6:16 am #

    @Henry Farrell: what happened to MNCs and the geoeconomic layer – of the triple-layer cake the complex interdependence school posits? In Libya, what role did the oil corporations play in legitmating the legal sovereignty of the national councils? Which brings me to Stephen Krasner on “Problematic Sovereignty”. Just because states’ legal sovereignty is swiss-cheesed by NGOs, IGOs, INGOs, and MNCs, it doesn’t follow that the other three kinds of sovereignty aren’t still important.