This book is about the complicated interaction between two East African coastal communities, the Muslim Swahili and the ‘traditionalist’ Giriama. It deals particularly with the latter's ongoing conversion to Islam. Stylistically it is a deeply tantalizing work, consisting of a curious mélange of straightforward, eminently readable, narrative passages of descriptive text, and eruptions of ponderous jargon. These latter bulk large in the author's analysis of Swahili hegemony over the subservient Giriama – who, she says, contribute to their own subjection. Giriama subordination and Swahili authority both have their origins, McIntosh claims, in the two ethnic groups' contrasting models of ‘personhood’. This style of analysis suggests a throwback to the ‘culture and personality’ school of anthropology. It is thus not surprising to find the author declaring that ‘customary Giriama models of personhood undergird religious practices that reaffirm both the potency of Islam and the distinction between Giriama and Swahili ethnoreligious groups. Prevailing Swahili models of personhood, in contrast, tend to undergird negative judgements of Giriama that reaffirm the high status of Islam and the same ethnoreligious distinctions.’

What might be called the book's ethnodynamics focus on the conversion of the Giriama to Islam. This turns out to be a complicated and painful process but, in the author's view, one that throws profound light on the Giriama perception of their ‘polluting life-ways’ and ‘ontological presuppositions’. Giriama individuals are forcibly converted to Islam by invasive Muslim spirits which take possession of them and force them to ‘regurgitate the food and drink considered quintessentially Giriama’ (rats and palm wine). These routine parts of their diet are thus extruded and cast aside, enabling Giriama, after submitting to Islamic spirits, to become Muslim Swahili by adopting the latter's food prohibitions. Such possession, contrary to its common depiction in the literature on the subject, is not an act of resistance, McIntosh contends, but an ‘embodiment of a hegemonic premise’. But, of course, acceptance and resistance in spirit possession are not necessarily mutually incompatible, since the latter behaviour can be taken to signify a force that, ultimately, has to be accepted and is, consequently, all the more powerful. Resistance, thus, can be very ambivalent and a sine qua non of the authenticity and power of invasive force. This is certainly a common and widely recognized theme in the possession literature, although one ignored by McIntosh. She concludes rather lamely that religion can be a force ‘that alights on the human body, sweeping it into social currents whose force is more powerful than the will of the individual’. That it is also, and above all, an intervention beyond the control of the possessed individual, as comparative studies of possession and its treatment endlessly demonstrate, enables it to serve as a fundamental criterion of experiential authenticity.

The gender organization of what McIntosh calls ‘diviner healers’ is clearly important here. Females, usually illiterate, work primarily with spirits, whereas literate male diviners and exorcists conform to the pattern general in Muslim societies, relying essentially on spiritual power associated with the Qur'an and such famous Arab magical compendia as the widely distributed Ilmi-Falak (literally ‘magical knowledge’).

A crucial sociological factor, insufficiently emphasized in this essentially culturological study, is that in Swahili culture there appears to be an intriguing connection between madness, money and Muslim spirits. This linkage McIntosh interprets, rather unconvincingly in my view, in terms of a significantly higher incidence of madness among Giriama males than females. This she blandly reports without comment, and connects with the fact that males rather than females serve as labourers under Muslim (for example, Swahili) employers. (She does not, however, seem to take account here of female house servants.) As she reports, jinn are believed to be particularly malevolent Muslim spirits, purchased by wealthy Arabs and Swahili from other Muslims and used to secure financial gain. They may even bewitch a Giriama servant who picks up (and appropriates) loose change off his employer's floor. Most interestingly, these jinn can be used to import gold or currency from distant places, suggesting the themes of labour migration and trade, and more grandly, what are seen as the pernicious effects of the capitalist economy. So as McIntosh not implausibly argues, the ambiguities surrounding Islam are connected to the ambivalent relationships between Giriama and their more powerful neighbours, including the stark socio-economic hierarchy binding the two. At the macro level her arguments, while hypothetical, are not unconvincing, but at more immediate levels her general treatment would have benefited from a deeper immersion in the comparative literature on the sociology of possession. This, in my view, has more to contribute to her analysis than cultural anthropology.