Agency, Personhood and the 'I' of Discourse in the Pacific and Beyond
Rumsey, Alan, Journal of the Royal Anthropological Institute
Sahlins bases his account of Polynesian 'heroic history' partly on the fact that chiefs used the pronoun 'I' in reference to their whole group. Mosko (1992) argues that Sahlins's consequent emphasis on 'encompassment' as the modality of chiefly action is diametrically opposed to Strathern's on 'partibility', the effacement of parts of the person as a condition of action. Drawing on comparative material from the New Guinea Highlands, where big men also use 'I' for their whole group, and on Benveniste's and Urban's accounts of the meaning and use of personal pronouns, I argue instead that moments of both encompassment and partibility are inherent in language, corresponding to two distinct dimensions in which the pronouns are meaningful (the 'direct indexical' and the 'anaphoric'), and that close attention to the interaction between the two can yield new insights into the nature of personhood and social agency.
Though anthropology has left behind the bad old days, when 'the person' was admitted as subject matter only in the form of an abstract configuration of social roles, we still suffer from a tendency to separate concrete human action in the here and now from questions of 'personhood' as a cultural category. I will try to help bridge that gap by focusing closely on one of the things that people do with language, namely posit themselves as agents with the first person pronoun I (and related forms). My ethnographic focus is upon the Pacific. This is an especially fertile field for such a venture because the region has seen some of anthropology's most challenging work in the cross-cultural study of personhood (e.g., Leenhardt 1979 ; Lutz 1988; Read 1955), and also because it is one in which specific patterns of pronoun use have already been drawn upon as evidence for one of the most prominent theoretical accounts of social actorhood, in Sahlins's recent work on Polynesian 'heroic history'. In this article I w ill be discussing Sahlins's model in relation to what is currently one of the most prominent anthropological accounts of personhood, a Melanesianist one by Marilyn Strathern. I will attempt to resolve some apparent contradictions between the two models by introducing Melanesian evidence concerning pronominal usage and reconsidering Polynesian usage in light of it. My conclusions and the forms of analysis I introduce along the way will, I hope, provide a new and useful perspective on questions of personhood and social action in general.
Of chiefs, big men, detachability and encompassment
Strathern develops her model of the partible person as a self-conscious displacement of Western notions of the individual and society. From within the terms of that presumed Western orthodoxy (1988: 12), she draws out an image of the Melanesian person as relationally constituted: a composite of detachable parts identified with the separate male and female contributions which have produced it.
Although Strathern intends The gender of the gift as a 'cultural description' (1988: 244) rather than a 'sociological analysis' (Strathern 1992:150), it engages centrally with questions of social action and agency. While the person in this view is composite, the act is unitary  (1988: 274), and is premissed upon the externalization or 'eclipse' of parts of the person who acts. So, for example, a woman in giving birth to a child eclipses the composite male and female origins of her own body and evinces her own agency as a kind of 'constitutive singleness' (1988: 278). An agent in this view is 'one who from his or her own vantage point acts with another's in mind' (1988: 272). So, when a man makes a prestation to his exchange partner, the partner is the cause. In this view agents 'are not the authors of their own acts. They simply do them. Agency and cause are split' (1988: 273).
A very different account of a form of personal agency has come out of Sahlins's model of 'heroic history'. Drawing upon Hawaiian, Fijian and Tongan ethnography and upon Dumont's (1970) understanding of hierarchy as encompassment, Sahlins argues that chiefs acted expansively, summing up in their own person the lives of many. …