from Africa in Transition

Difficulties of Defining and Mapping Ethnicity

May 23, 2013

Blog Post

More on:

Sub-Saharan Africa

Elections and Voting

Civil Society


Aging, Youth Bulges, and Population

This is a guest post by Brooke Bocast, a PhD candidate in anthropology at Temple University and a visiting predoctoral fellow at Northwestern University. She is currently writing her doctoral dissertation on gender, consumption, and higher education in Uganda.

On May 16, the Washington Post’s Max Fisher published a “revealing map” of global ethnic diversity. This map is indeed revealing, but not entirely in the terms its author suggests. Based on a 2002 Harvard Institute of Economic Research paper, itself based on analysis of secondary data sets, Fisher’s map presents Uganda as the world’s most ethnically diverse country. Presumably, this finding is meant to “reveal” the hitherto unknown cornucopia of ethnic groups within Uganda’s national borders. What this map actually (but also, only potentially) tells us, is that during the data collection time period, residents of Uganda were acutely aware of particular features of their identities.

The data sets used in the Harvard paper—censuses and the Encyclopedia Britannica—are based on self-reportage of ethnic identity. Given that the western concept of ethnic identity does not directly map onto indigenous African ideas about personhood, it is unclear how the self-identification data was gathered in the first place. Within Uganda, the most closely analogous concept to “ethnicity” is “tribe,” which overlaps with matters of linguistics, lineage, clan, gender, lifespan, and geographical location. My Ugandan friend Stella, for example, is currently not quite sure of her tribe, although she has “a few guesses.”

So we have, on the one hand, local slipperiness of ethnic/tribal identification and, on the other hand, data that purports to reveal a diversity of definitive responses to queries of ethnic identity. There are countless possible interpretations of this juxtaposition. But let’s look at what temporality can tell us. Fisher acknowledges that the Harvard data is limited because it is time-bound. I argue that the existing analyses are limited, not because the data is “old,” but because the Harvard paper, and Fisher’s cartography, collapse a decade’s worth of data into an amorphous present.

If we were to disaggregate the data over time, and still found relatively high rates of ethnic self-identification within Uganda in the late 1990s, our question would become not, “how many ethnic groups does Uganda have?” But, “why did Ugandans demonstrate marked ethnic awareness at that particular point in time?” One possible explanation could be that Uganda held national elections in 2001. Uganda’s politicians are fond of “instrumentalising” ethnicity during campaign season, and this translates into a generally elevated salience of individuals’ tribal affiliations. It is possible that, had she been of voting age, Stella would have “known” her tribe in 1999 or 2000, despite not being cognizant of it today.

Dynamic ethnic self-identification is not unique to Uganda—far from it. Other bloggers, writing in response to Fisher, have already emphasized the social science truism that ethnicity is an unstable category. Given this, we would do well to scrutinize Fisher’s other assertions—“European countries are ethnically homogeneous” for example—to see what they might reveal about the forces that shape peoples’ tendencies towards or away from ethnic self-identification.