As will be evident from our introductory essay, the impetus for this Social Thought and Commentary section has come in part from our puzzlement at the co-occurrence within Melanesianist ethnography of two apparently contradictory motifs. One is the widely reported idea that one can never know what is in the minds of others, partly because what they say cannot be taken as a reliable indicator of what they are thinking. The other motif is the rising prominence of various more-or-less institutionalized practices of confession, as exemplified with respect to indigenized Christianity by the Urapmin example discussed by Robbins. These practices of confession would seem to stand in a problematical relationship to claims about the impossibility of knowing what is in the mind of another, since that is precisely what they would seem to be designed to reveal.
The essays by Robbins and Schieffelin address this problem by taking the "opacity" motif as their starting point and then considering new practices of confession and how they impact upon speech communities whose preexisting language ideologies would seem to deny the reliability or appropriateness of such disclosure. Here I approach the problem from the other way around, by looking first at practices of confession and then asking [End Page 455] how they do or do not relate to linguistic ideologies. One point of this will be to show that, in at least some Melanesian locales, practices of confession did not begin only with the arrival of Christianity—that there were earlier forms of it which continue to be practiced alongside, and sometimes in tension with, the church-related ones.
The question of the relationship between those older practices and the Christian ones is of especial interest in the context of comparative studies of confession that have been stimulated by Michel Foucault's well-known (1978, 1988) account of late medieval Christian practices of confession as the genealogical precursor and matrix of more recent, secular regimes that have produced the purportedly self-disciplined, modern subject. So far, these comparative studies have been carried out mainly by sociologists and historians, with little direct input from anthropologists. This is unfortunate, for two reasons. The first is that, sophisticated as some of this work has been when dealing with European historical sources, when drawing on anthropological ones it is often far less so. The other reason is that, as we hope these essays will demonstrate, such practices offer a fertile field for comparative study in connection with issues of interest to anthropologists (cf. Strathern and Stewart 1998:62).
The fullest comparative studies of confession to date, by Hepworth and Turner (1982), and Abercrombie, Hill and Turner (1986:35–72 et passim), while taking issue with Foucault in some respects, have agreed in their finding that " the confessional tradition in Europe laid the foundation for the modern personality as self-reflective consciousness" (Abercrombie et al. 1986:47). These writers are well aware from their reading of ethnography that there are other " confessional traditions" outside of Europe which are of longstanding, but, adopting a Durkheimian position—which incidentally, obviates the issue of mental opacity altogether—they have argued that:
confession outside the Western tradition normally assumes a collective, group nature, that is, the confession is a reflection not of the state of mind of the individual, but a reflection of the character of the social structure. We would suggest that such confessionals are in fact very Durkheimian: they are statements about collective properties and forms of public thought, reaffirmations of public values and communal practices, rather than reflections of an interior mentality (1986:46). [End Page 456]
Let us now consider this position in relation to some cases from Papua New Guinea. Turning first to confessional practices of the more traditional sort, in the Ku Waru area of the Western Highlands of Papua New Guinea where Francesca Merlan and I have worked, there are well-established practices this kind which are referred to by two kinds of related expressions. One of them is: ung kis pára si- where si- is a verb root that takes suffixes indicating (among other things) the grammatical person...