A feature article in Slate last May (note 1) used US census data in creating a map of the most commonly spoken African languages by state. Of the five languages and language cagegories that were shown on the map, two were obvious (Amharic and Swahili), two were actually branches of language families (Bantu and Cushitic), another was a language group (Nilotic), and one was puzzling (“Kru, Ibo, Yoruba”).
This raised questions about the categorization used by the Census Bureau and how they aggregate data from census forms, which I explore briefly in a blog posting (note 2). This is not to say that using language categories in reporting is necessarily inappropriate (it may be a protection of privacy where numbers of speakers of specific languages will be very low), but to ask about the choice of categories (note 3) and their coverage. On the latter point, the way I read the available categories, there is no appropriate one for Wolof or Akan.
Have any linguists specializing in African languages looked at this issue? The next census is in 5 years, so if any changes are to be recommended, now would seem to be a good time to do so.
The maps in Slate are fun, but Osborn raises several interesting questions that might encourage further discussion and could improve the next census.