Stathis Kalyvas tweeted an interesting observation, which led Christian Davenport to jump in, and twitter’s character limit has me moving to this space to share my thoughts.
I think Kalyvas makes a nice observation; one which I have had vaguely kicking around in my brain, but he puts quite nicely: an imperfect, but interesting and insightful, covariation.
Davenport’s point is that when we move away from binary classification (autocracy v democracy), we can observe some otherwise hidden variation in the democracy category. Kalyvas then asks whether anyone can think of post WW II examples that weaken the association, and offered the Paris massacre of 1961 as the only exception he could think of.
I offered the 2012 Lonmin Mines example from South Africa, though wondered whether we should consider single party rule governments (especially those with a super majority, as the ANC has had since the transition) in the same category as coalition governments. That thinking reveals the emphasis in my work upon constraints against majority rule: like James Madison, I argue that majority rule will not constrain state abuse (e.g., see here [blog post], here [PDF], here [gated]). Kalyvas seems to agree.
I also observed that in Turkey the police response to protest has been rather different than it is (see my post here, as well HRW’s recent post).
Kalyvas proposed a thought experiment, and Steve Saideman jumped in with an interesting observation about repetition.
I would like to offer this contribution to the discussion. In the US Jim Crow south local and state governments did respond very much like the Egyptian government, though with less use of gunfire: police would “stand down” while “local citizens” beat, bombed, and shot at civil rights marchers. And while there were democratic elections throughout the Jim Crow South (albeit absent black participation), there was only one party that had any chance of winning: it was single party, super majority rule at both the local and state level. As for the types of events I am referencing, the freedom-riders experience is but one well known example. Unfortunately, I cannot find the story I am looking for documenting one such incident in Tallahassee, though the 50 years after-the-fact apology by the local newspaper and this story provide some interesting context.
I envision two reasonable responses to argue that neither the Jim Crow South nor Northern Ireland make good examples. With respect to the Jim Crow South, the Egyptian state was (1) very public in its call for “public support” (aka vigilantism), and (2) it state agents much more liberally used gunfire than did the state in the Jim Crow South.
Turning to Northern Ireland, one might observe that the UK government is occupying that territory. In other words, it would be more akin to US troops in, say, Puerto Rico, than US national guard troops in Little Rock, AR or Oxford, MS. We can add Israeli military repression in the West Bank and Gaza Strip to the Northern Ireland example.
To summarize, I agree with Kalyvas’s that it is an insightful association, but I am always somewhat dubious about binary conceptualizations, especially of “democracy.” They are seductive due to their simplicity, and like Davenport, I worry that they conceal while they illuminate. I do not worry that scholars like Kalyvas will miss this point, but as the ideas diffuse broadly, that risk rises (likely non-linearly).
@WilHMoo
Will Opines 
Pingback: Weekly Links | Political Violence @ a Glance
Should we really consider modern South Africa and Jim Crow South of the U.S.A. to be democratic? Looking at the politics and tactics it seems more that they could arguably be called competitive authoritarian (or maybe just authoritarian for Jim Crow South).
That depends on how one wants to code democracy, which is one of my points. Is it only about majority rule (elections), or do we want to define democracy more broadly to as liberal democracy and include checks on executive authority? In federal systems, do we want to focus on local government, or only . Much of the academic literature treats democracy as majority rule only, or doesn’t explicitly address the issue, yet tends to use data that cover only executive elections, or primarily those.
I would argue that if violence is regularly used to ensure the political domination of a political party and a racial group, local law enforcement is expected to permit and occasionally outright assist in that political violence, the judicial, executive and legislative bodies act to help maintain that dominance, and the civic associations are expected to at least acquiesce to this then whatever elections it might have make it no more democratic than modern Russia.
This may sound offensive since I’m from the northeast of America while speaking about the American south, but looking at it until relatively recently the local government there (especially during the Jim Crow period) seems to have had more in common with authoritarian sectarian-based governments than any actual democracies.