Reactions to Reflections on the Arab Uprisings

Yesterday, Marc Lynch posted a thoughtful and candid set of reflections on how political scientists who specialize in the Middle East performed as analysts and forecasters during the Arab uprisings, not before them, the subject on which most of the retrospectives have focused thus far. The background to the post is a set of memos Marc commissioned from the contributors to a volume he edited on the origins of the uprisings. As Marc summarizes, their self-criticism is tough:

We paid too much attention to the activists and not enough to the authoritarians; we understated the importance of identity politics; we assumed too quickly that successful popular uprisings would lead to a democratic transition; we under-estimated the key role of international and regional factors in domestic outcomes; we took for granted a second wave of uprisings, which thus far has yet to materialize; we understated the risk of state failure and over-stated the possibility of democratic consensus.

Social scientists and other professional analysts of world affairs should read the whole thing—if not for the specifics, then as an example of how to assess and try to learn from your own mistakes. Here, I’d like to focus on three points that jumped out at me as I read it.

The first is the power of motivated reasoning—”the unconscious tendency of individuals to process information in a manner that suits some end or goal extrinsic to the formation of accurate beliefs.” When we try to forecast politics in real time, we tend to conflate our feelings about specific events or trends with their likelihood. After noting that he and his colleagues over-predicted democratization, Marc observes:

One point that emerged in the workshop discussions is the extent to which we became too emotionally attached to particular actors or policies. Caught up in the rush of events, and often deeply identifying with our networks of friends and colleagues involved in these politics, we may have allowed hope or passion to cloud our better comparative judgment.

That pattern sounds a lot like the one I saw in my own thinking when I realized that my initial forecasts about the duration and outcome of the Syrian civil war had missed badly.

This tendency is probably ubiquitous, but it’s also one about which we can actually do something, even if we can’t eliminate it. Whenever we’re formulating an analysis or prediction, we can start by ask ourselves what result we hope to see and why, and we can think about how that desire might relate to the conclusions we’re reaching. We can try to imagine how someone with different motivations might view the same situation, or just seek out examples of those alternative views. Finally, we can weight or adjust our own analysis accordingly. Basically, we can try to replicate in our own analysis what “wisdom of crowds” systems do to great effect on a larger scale. This exercise can’t fully escape the cognitive traps to which it responds, but I think it can at least mitigate their influence.

Second, Marc’s reflections also underscore our tendency to underestimate the prevalence of inertia in politics, especially during what seem like exceptional times. As I recently wrote, our analytical eyes are drawn to the spectacular and dynamic, but on short time scales at least, continuity is the norm. Observers hoping for change in the countries touched by the Arab uprisings would have done well to remember this fact—and surely some did—when they were trying to assess how much structural change those uprisings would actually produce.

My last point concerns the power of social scientists to shape these processes as they unfold. In reflecting on his own analysis, Marc notes that he correctly saw how the absence of agreement on the basic rules of politics would complicate transitions, but he “was less successful in figuring out how to overcome these problems.” Marc aptly dubs this uncertainty Calvinball, and he concludes:

I’m more convinced than ever that moving beyond Calvinball is essential for any successful transition, but what makes a transitional constitutional design process work—or fail—needs a lot more attention.

Actually, I don’t think the problem is a lack of attention. How to escape this uncertainty in a liberal direction has been a central concern for decades now of scholarship on democratization and the field of applied democracy promotion that’s grown up alongside it. Giuseppe di Palma’s 1990 book, To Craft Democracies, remains a leading example on the kind of advocacy-cum-scholarship this field has produced, but there are countless “lesson learned” white papers and “best practices” policy briefs to go with it.

No, the real problem is that transitional periods are irreducibly fraught with the uncertainties Marc rightly spotlighted, and there simply are no deus-ex-machina resolutions to them. When scholars and practitioners do get involved, we are absorbed into the politics we mean to “correct,” and most of us aren’t nearly as adept in that field as we are in our own. After a couple of decades of closely watching these transitions and the efforts of various parties to point them in particular directions, I have come to believe that this is one of those things social science can help us understand but not “fix.”

The Ethics of Political Science in Practice

As citizens and as engaged intellectuals, we all have the right—indeed, an obligation—to make moral judgments and act based on those convictions. As political scientists, however, we have a unique set of potential contributions and constraints. Political scientists do not typically have anything of distinctive value to add to a chorus of moral condemnation or declarations of normative solidarity. What we do have, hopefully, is the methodological training, empirical knowledge and comparative insight to offer informed assessments about alternative courses of action on contentious issues. Our primary ethical commitment as political scientists, therefore must be to get the theory and the empirical evidence right, and to clearly communicate those findings to relevant audiences—however unpalatable or inconclusive they might be.

That’s a manifesto of sorts, nested in a great post by Marc Lynch at the Monkey Cage. Marc’s post focuses on analysis of the Middle East, but everything he writes generalizes to the whole discipline.

I’ve written a couple of posts on this theme, too:

  • This Is Not a Drill,” on the challenges of doing what Marc proposes in the midst of fast-moving and politically charged events with weighty consequences; and
  • Advocascience,” on the ways that researchers’ political and moral commitments shape our analyses, sometimes but not always intentionally.

Putting all of those pieces together, I’d say that I wholeheartedly agree with Marc in principle, but I also believe this is extremely difficult to do in practice. We can—and, I think, should—aspire to this posture, but we can never quite achieve it.

That applies to forecasting, too, by the way. Coincidentally, I saw this great bit this morning in the Letter from the Editors for a new special issue of The Appendix, on “futures of the past”:

Prediction is a political act. Imagined futures can be powerful tools for social change, but they can also reproduce the injustices of the present.

Concern about this possibility played a role in my decision to leave my old job, helping to produce forecasts of political instability around the world for private consumption by the U.S. government. It is also part of what attracts me to my current work on a public early-warning system for mass atrocities. By making the same forecasts available to all comers, I hope that we can mitigate that downside risk in an area where the immorality of the acts being considered is unambiguous.

As a social scientist, though, I also understand that we’ll never know for sure what good or ill effects our individual and collective efforts had. We won’t know because we can’t observe the “control” worlds we would need to confidently establish cause and effect, and we won’t know because the world we seek to understand keeps changing, sometimes even in response to our own actions. This is the paradox at the core of applied, empirical social science, and it is inescapable.

%d bloggers like this: