A History of African Motherhood is an exciting addition to a series of works on the social histories of Bantu-speaking peoples based on historical linguistics as well as other historical methods. The significance of this study is that the author is able to look at a particular region of Bantu-speaking Africa from 700 to 1900 and trace the changes in how motherhood was conceptualized from the periods prior to, during, and after the creation of kingdoms in the Great Lakes Region of East Africa. The author calls her history ‘an imperfect tapestry’, yet this work allows us to see into the minds of people from earlier eras in an attempt to understand how they conceptualized motherhood and how their concepts altered as historical conditions changed.

The use of historical linguistics and oral traditions to recapture African history may not tell the reader ‘what so and so did on such and such a date’, but it can reveal deeper ideological understandings as well as the transformations of these ideas over time. The historical linguistic method is a major theme of Chapter One. Chapter Two examines how the ancestors of the Buganda and South Kyoga centralized societies viewed the lifestage of motherhood from 700 through 1100. During this period, motherhood was a social category distinct from the biological event of giving birth. In Northern Nyanza's patrilineal society, the ideal mother was married and the marriage was sealed by the complete payment of brideswealth.

In Northern Nyanza the people innovated a new verb, *-bayira (a woman marrying). *-bayira is an active form of the verb, which is significant because in many other Bantu-speaking societies a special passive verb form exists for expressing a woman marrying, while an active form exists for a man marrying. This active verb form reflected the belief in the dynamic role of a wife in Northern Nyanza society. This was especially true in regions where there were newcomers to the area. Such immigrants would marry local women with well-established lineages to improve their situation or that of their descendants. Another new word was *-rya (marriage or a married state) and applied only to women. This noun is derived from the proto-Bantu word *-di (to eat). It could reflect the key role women had in food production and preparation, but this was not unique to the Great Lakes region. Stephens argues that, instead, this new word for the married state of a woman may indicate that women now had to secure land through marriage instead of through their own patrilineage or matriclan.

Stephens writes that brideswealth sealed all marriage relations, which in this region is clearly true. But then she makes a much broader claim that *-ko, the word used in the Great Lakes to mean ‘to give brideswealth’, meant the same in proto-Bantu. This is a rare example where she does not provide evidence to support her claim. To give brideswealth requires patrilineal social relations or at least patrilocality. Overwhelming evidence shows that Bantu-speaking peoples have been matrilineal and matrilocal since they left West Africa at least 5,500 years ago. Only in more recent historical times did some Bantu societies start becoming patrilineal.Footnote ¹ Thus brideservice, not brideswealth, would have originally ‘sealed’ marriage in the proto-Bantu world. Additionally, in Benue-Congo languages (the language group of which Bantu is a subset), proto-Yoruba, and Proto-Gbaya, *-ko meant ‘a nongender-specific spouse’ and the Proto-Bantu semantic innovation was a shift of meaning from spouse to a relative of one's spouse, to affines in general, not brideswealth. It is possible after more research that we will discover the original meaning in proto-Niger-Congo was some form of ‘sealing’ marriage, but it could never have been ‘brideswealth’ in proto-Bantu.

Between 1200 and 1500 in the Great Lakes, societies of the earlier periods were diverging, as were their attitudes toward social motherhood. Some people moved east and began to speak South Kyoga, while the peoples that remained in Northern Nyanza began to speak Luganda. The South Kyoga speakers encountered Nilotic-speaking peoples and instituted important shifts in their ideology of motherhood. Elite motherhood was as much a product of social contracts as it was the outcome of biological reproduction. The linguistic evidence of this is a new verb (wereka), which meant ‘to foster a child’. Also there was a new borrowed word, *-piita (to foster), from Southern Luo.

Probably nothing in Stephens' research symbolizes the shifting position of mothers and the institution of motherhood in this region more than the changing meanings for the noun, *-gole. She cites Gonzales's work in Central-East Tanzania where the meaning for *-gole was (and still is in some societies) ‘a girl who has just developed breasts’, which was a life stage prior to the *-yadi lifestage which started with her first menstrual period and ended with her first successful pregnancy. According to Stephens, *-gole indicated a young girl's potential as a mother at an earlier period in this region. In Northern Nyanza, the meaning of the noun *-gole was narrowed to mean ‘bride or newly wed woman with maternal potential’. As the evolution of *-gole indicates, in the Great Lakes Region from 700 to 1900, motherhood transformed from an important multistaged position to a subset of wifehood. While motherhood in its various manifestations in the Great Lakes was still important, it no longer wielded the authority it once did and does in other parts of the Bantu World.

This work is a valuable contribution to early African history, Bantu history, and gender history. Stephens is able to show clear evidence of the changes and transformations in the ideology as well as the social and political authority of motherhood over 1,200 years. I find this work exciting, thought-provoking, and necessary reading for anyone interested in early African history or gender in African history.