Steve Saideman, a professor at Carleton University who studies ethnic conflict, responded to the maps on his blog, expanding on some of my points, pushing back on others and offering different takes. You can read his thoughts in full here and here. I've pulled out a few of his insights:
(1) It's tough to gauge racial intolerance through just one metric. Saideman makes the point that there are different ways to express racial intolerance. The metric we used suggested, for example, high tolerance among Americans but low tolerance among Indians. It's entirely possible that other metrics could return very different results; Saideman asks, for example, whether the map might look different if we asked people whether they would be okay with a member of a different race marrying into their family. He reproduces this chart showing survey results from Romania, which drives his point home:
(2) Different people might hear the question differently. Saideman writes, "In some places, when one is asked this question, they may think of a single race, perhaps the Vietnamese think of the Chinese but not of other races. So it may not be that the people are very racist in general — they just hate one group that is defined by race." In other words, if Vietnam scored as particularly intolerant (they did), that might just be because they're less tolerant toward the race that popped into their heads first — e.g. the Chinese — than they are of other races in general. This makes it tougher to compare across countries.
(3) What's the link between diversity and conflict? The landmark 2002 study on ethnic diversity, the source of my map on diversity, noted a correlation between countries that are ethnically diverse and that experience internal conflict. But what's the relationship between the two? I offered two theories: that ethnic diversity can lead to competition and thus conflict or, alternatively, that conflict creates more diversity by fracturing big groups into smaller ones.
Saideman offers this: "In societies that have very little diversity, there is no opportunity for [ethnic] violence. For societies where there is a great deal, there is no threat of dominance. But in places where there are a few groups that rival each other, then the threats they pose to each other or at least one to the others can be severe. Ethnic violence may not be about fractionalization/diversity but about polarization."
(4) Ethnic conflict comes from "concentration," not "fractionalization." Here's where Saideman really shows off his expertise. "The one consistent finding for ethnic conflict is not about fractionalization but about group concentration," he writes. "That where ethnic groups have distinct areas apart from each other within a country, there is more conflict. Why? Well, partly because it facilitates separatism. Partly because groups that are separate have a secure base from which to launch attacks. Partly because intermingled groups may be deterred from attacking since they themselves are vulnerable (kind of like mutual assured destruction)."
(5) Getting Somalia right. I cited the example of Somalia from the 2002 Harvard paper, which argues that the country actually became more ethnically diverse as a result of conflict rather than the other way around: As fighting and resource scarcity divided the country, people narrowed down their identities (i.e., someone who might have said "I'm ethnic Dir" before the war might have said "I am Issa" after; Issa is a sub-clan of the Dir). Saideman says we got this wrong:
Fisher cites the article discussing how Somalia's identity politics changed after the civil war in 1991. This would be fine except that Somalia has always been poorly understood. People think that Somalia could be the one African country that could support secession and even be irredentist (seeking to annex neighboring territory inhabited by kin) because of its homogeneity (unless they have read my stuff). Somalis speak the same language, are all Muslim, are of the same race, and so on. But they were always divided by clan identity (kinship), which meant that the irredentism was always inconsistent. The leaders in Mogadishu would support the claims of some of the kin in some of the neighboring countries, depending on whether the kin had ties to politically relevant kin in Somalia. Lots of irredentism, targeting Somalis in Kenya, Ethiopia and Djibouti in the early 1960s because the electoral system required playing to a wider audience. In the mid to late 1970s, the irredentism only focused on the Ogaden clan residing in Ethiopia because its kin was a key partner in the authoritarian regime while the clans tied to the Somalis in Kenya and Djibouti were not in the regime.Fisher and the folks he is citing are probably correct in that the focus since 1991 has probably moved to sub-clan identities. It is still kinship but smaller fractions so you get fighting between different sub-clans that belong to the same clan or clan-family rather than competition between clans or clan-families.