Focusing on the areas of South Africa now encompassing the Eastern Cape and KwaZulu-Natal, Divided by the Word traces the construction of Xhosa and Zulu as ethnicities defined largely by languages heavily influenced by missionary translation. Throughout the book, Jochen Arndt challenges ‘the widespread assumption that language and collective identity… are by nature coterminous’ (6). Instead, Arndt argues that from 1500 to 1990, Africans and foreigners brought perspectives on race, tribalism, nation, ancestry, and linguistic proficiency to bear on a single long-term process that inseparably constructed Xhosa and Zulu as major ethnicities and languages. A dense historiography supports the idea of ethnic and tribal identities lacking deep precolonial origins, but Arndt's exploration of the role of language adds an important dimension.Footnote ¹

The first chapter introduces a secondary argument that Arndt frequently returns to: before 1800 CE communities living in southeasternmost Africa recognized but did not mobilize language difference in salient social categories. The ‘outsider’ perspective of linguists and archaeologists tends to take for granted that modern languages existed in their current form in the past and that language was at the center of change over long periods of time (18–26). His critique of emphasis is accurate, as reconstruction of the past in both disciplines demands drawing boundaries around discrete groups of people, language, and materialities that are (usually) rooted in evidence. Arndt constructs an alternative ‘insider’ perspective suggesting southeastern Africans recognized speech variability only as ‘soft’ geographic markers (27) irrelevant to identity or status. He admits, however, that the full significance of language before 1800 remains unclear, something underscored by a paucity of temporally relevant evidence.

Pre-1800 history is unclear and drawing sharp lines between disciplines and evidence is more provocative than useful. The dismissal seemingly serves the secondary argument that linguistic and cultural categories could not exist in the past as the disciplinary perspectives demand because language difference was not stable or socially relevant in the past. This contention misunderstands the ways other disciplines reckon with the accumulation of change and the purpose of hypothesizing categories, including languages. Historical linguists prioritize language not because we consider it inextricable from or encompassing of identity, but because it offers a lens into changing beliefs and behaviors, including around group formation. Not without criticisms, archaeologists similarly foreground pottery and other material culture correlated with identity. Working backwards from existing languages does not suppose that those languages are unchanged by colonial history but rather expects that they are. Tracing language change, including by outside influence, is central to our work. The pre-1800 synthesis appears to offer a point of departure with no clear intervention or new research, allowing Arndt to hold up the ‘pre-colonial past’ as different from the subsequent history of language. The history he subsequently crafts, however, is not predicated on the fluidity of language before 1800 and foregrounds European missionaries, not Africans.

Arndt's central story shows that European notions of ‘Caffre’ sat at the center of ethnolinguistic construction before the late nineteenth century. Those familiar with South Africa will recognize ‘Caffre/Kafir’ as racialized terminology, but Arndt demonstrates early iterations to be taxonomic, as well as racial. The formulation of the idea of a single race of Africans along the southeastern coast from the Great Fish River to Delagoa Bay began with sixteenth century European mariners and continued with late eighteenth century naturalists. All used ‘Caffre’ as a broad category of race, culture, and language that spanned the region. Early nineteenth century missionaries thus entered an ontological universe in which ‘Caffre’ was widely legible.

Chapter Four's coverage of early nineteenth century missionary translation projects unfolds convincingly. Motivated by existing categories and buoyed by local contexts of linguistic permeability, missionaries documented a consensus language they called ‘Caffre', which would later come to be known as Xhosa. Arndt offers abundant evidence that missionaries met little resistance as they began to ‘harmonize’ variations in language. Nearly coeval with this process, American missionaries in nearby Natal and Zululand were preoccupied with imagined Zulu culture and the purity of Zuluness, supporting his argument that missionaries centered Zulu as a majority language in the region, as discussed in parts of Chapters Five and Six. Together, these missionaries divided a landscape of similar languages into ‘Caffre’ and Zulu.

Arndt argues that southeastern Africans recognized variations in vocabulary and phonology as geographic markers or quirks that did not impede mutual intelligibility of dialects. The argument sits uncomfortably within the sections on the Zulu kingdom and Zulu language, however. Despite being post-1800, the contradictions are hard to ignore. Zulu elites famously used language differences to marginalize low-status subjects and manipulate shared identities (159–69). Arndt contextualizes this history within the long trend of using language as a soft geographical marker identified in the first chapter. But minimizing African claims about identity neglects the seemingly separate history of crafting language as a referent to status in the Zulu kingdom and thus the making of ‘Zulu’ language without and beyond foreign categories. Further, by treating the construction of Zulu as the product of manipulation rather than African agency, Arndt elevates the significance of missionary interventions over Zulu perceptions of themselves. It is a move that undermines his pursuit of an ‘insider’ history.

In a final chapter and epilogue, Arndt brings the narrative into the twentieth century and weaves it into the well-documented history of emergent ethnic and tribal identities. He here addresses Mbongiseni Buthelezi's long-standing call to complicate Xhosa and Zulu by reasserting his claim that both are primarily language-based identities. He also explains the stakes of fashioning majority languages, including the diminution of other ethnicities and their own concomitant languages. Divided thus offers an imperfect history that will surely be controversial, while also becoming a valued reference in future debates about language and identity in Southern Africa and beyond.