Adeno Addis at Tulane Law argues in a newly released article that:
After examining the two well-known approaches - consociation [Ed., i.e., "power sharing" arrangements such as constitutional federalism with divided government] and integration [Ed., i.e., rule by interethnic moderates that excludes extremists secured with electoral engineering] - that have dominated studies of, and prescriptions for, severely fractured societies, the article concludes that each unwisely underemphasizes one or another of the two necessary conditions for long-term stability in these societies: institutions that are both highly inclusive and have the capacity to foster interethnic dialogue.
The article then outlines and defends a version of deliberative democracy that it argues responds to the needs of inclusion (pluralism) and the cultivation of interethnic dialogue. A well-structured deliberative process in the context of a highly inclusive institutional environment has the best prospect of transforming the hard parameters of ethnic identity into the soft parameters of diversity that this article argues will lead to a more sustainable form of pluralistic solidarity.
In short, getting everyone to talk to each other is critical to overcoming barriers to democratic governance in fractured societies.
I agree with the article's observation that "democracy defined as majoritarian rule is often a problem rather than a solution to the extent that it may (and often does) lead to the perpetual domination of one ethnic group by another." I'm not sure, however, that extended domination is always a bad thing. The United States endured reconstruction by excluding dissidents from the conversation and imposing what amounted to martial law, rather than through deliberative discussion. This wasn't an ideal resolution -- the nation still bears the scars and young Southerners still grow up thinking of their Northern compatriots as damn Yankees -- but, it isn't obvious that the United States could have remained a single country any other way.
While few would agree to submission ex ante in a Lockean social contract negotiation, peoples can be quite willing to put up with that situation in a stable way once it is upon them, if they are left with some basic dignity and self-respect, and if they have no reasonable hope of ever becoming dominant (and hence no real incentive to stir the pot trying to secure that kind of majoritarian power). Islam's doctrine of respected minority status for "people of the book," and the American First Amendment's freedom of religion rights are both strategies by which eternally minority ethnic groups have been reconciled to their state in democracies.
The model advocated by Addis might be best summed up as government by Quaker meeting supported by faith that a common pool of reason will lead people in the right direction if they talk long enough and sincerely enough. It argues that the discussion be so inclusive that it must include expatriates.
The ideas are interesting, but the proposal is dreadfully mushy and the conclusion seems to have precious little by way of persausive argument to support it. I don't share the comfortable faith there reason is succeptible to providing consensus after enough deliberation. Also, while there is danger in offending the witch not invited to the party and inviting her wrath, there is also virtue in marginalizing extremists if one can create a perception that even they share that they are indeed marginal. Germany's practice of excluding parties that can't marshal at least five percent support at the polls (largely fascist parties) is one of the better examples of this approach. Italy's post-World War II Pentaparte coalitions, in which all political parties in the relative middle settled on a common principal of excluding fascists and communists from government, even if that meant sacrificing ideals on the political right or left through center-left/center-right coalitions, had a similar effect. The exclusions, notably and importantly, unlike those in Iran, for example, were mostly exclusions imposed through a political rather than a legal process, but they are excclusions nonetheless, and Germany still has laws on the books prohibiting certain kinds of public speech advocating fascist ideas, as do many other countries, while still maintaining a fairly robust democracy.
Indeed, Addis doesn't even squarely address the fundamental question of when and why an effort should be made to preserve fractured socities as political units. The trend of the late 20th and early 21st century has been to encourage fractured societies to break into separate nation-states or largely autonomous regional governments when geography coincides with ethnic distributions sufficiently to make this possible, even if this geography is a fait accompli result which is the result of unconscionable genocide and the forced exodus of peoples. Iraq, Bosnia and Sudan are all moving headlong towards autonomous regional government and/or outright division. Ethiopia has broken up, as have Czechoslovakia, Yugoslavia and the Soviet Union. The Caucuses are fracturing, not caucusing.
Similarly, he doesn't address the approach of defusing fractures in societies though mergers that turns polar divisions into multipolar divisions. The European Union's overarching unity is part of what makes it safe for strong regional autonomy movements to arise in Europe. The Pan-African movement had a similar strategy.
He also doesn't address perhaps the strongest empirical evidence for his assertion -- the deliberative process within one party states -- that led those political structures, typically nominally communist nations, to endure longer and more often in former Western colonies than Western style parliamentary democracies did.
Addis also doesn't really discuss an "integrationist" approach in the way I thought that he meant it when I read his abstract, which is an attempt to form a unitary common national culture either out of whole cloth, or in some sort of melting pot fashion. Yet, this is a viable strategy apart from those he idnetifies. It is at the heart of the American idea, and the existing of a common elite culture and language created by the British is part of what has made India, which was at least as divided as Europe in the pre-colonial era, governable. This was also an essential element of the Soviet Communist empire. It is what Attaturk did to distinguish Turkey from the Ottoman Empire which preceded it. It was how the Islamic empire was forged, and how the Roman empire before it was made into some semblence of a single political entity.
Indeed, a common deliberative culture, at least, is a precondition to the kind of deliberative democracy that Addis proposes, and honestly, my instinct is that it takes quite a bit more than that for it to work. The reason that government by Quaker meeting (a deliberative style I try to replicate when I can while running meetings myself), ever works for anyone, is that its participants have a strong measure of cultural unity and its facilitators have a gift for listening to the meeting in order to draw from the cauldron the subtle threads of consensus.
It is rare that anyone but members of elites have enough time and inclination to devote to deliberations to an extent necessary to break through the barriers that create fractured societies, and elites that do so soon find themselves out of touch with their own constituencies.