On Not Knowing Other Minds: Confession, Intention, and Linguistic Exchange in a Papua New Guinea Community

Anthropologists working in the Pacific, and particularly in Melanesia, often report that the people with whom they work say that it is impossible to know what is in the mind of another person. One might read these statements as general existential claims about the human condition: we live as we dream alone, as Conrad would have it; or more prosaically, we live in effect as behaviorists, resigned to registering what people do without any interpretive recourse to the inner states that might have shaped their doings. But as much as such statements about the opacity of other minds can point to a general view of human life, in the contexts in which Melanesians make them they are very often more pointed than this, serving as metalinguistic claims about the limits of speech. It is speech in particular that such statements tend to single out, suggesting that speech is rendered problematic by virtue of its inability to convey what is in the minds of those who produce it. Because of its inability to convey what is in the minds of others, Melanesians often say speech is "mere talk" or " just talk"—not something one should treat as of much importance. So often in fact do Melanesians reiterate this point that it almost begins to sound as if they are aware that people might be tempted to interpret speech as a conduit for conveying the inner states of those who produce it, and then go to great lengths to warn them against doing so.

The broadest question I want to pose in this essay has to do with what we, as anthropologists, can do with statements about the opacity of other minds and the inability of speech to counteract this opacity. This may sound like an obvious question—one we as a matter of course ask about any kind of data we gather in the field. But I want to suggest that in the case of opacity statements and their language-ideological correlates, the answer is far more difficult to find than it is in most other cases. The difficulties follow from the extent to which these statements sound unbelievable to those anthropologists who have not worked in cultures in which they are prevalent. The problem is not that our colleagues cannot credit that people might make such statements, but is rather that they find it impossible to believe that people, even if they make them, might in any sense live by them. People who make opacity statements, the standard dismissive criticism has it, must nonetheless in their own practices of listening be making some kinds of assumptions about what speaker's intend to say, otherwise they could not interpret the speaker's speech as meaningful. Anthropologists who raise such doubts draw on the deeply held Western assumption that those mental states we call intentions are crucial to the meaning of speech. It is this fundamental tenet of Western language ideologies against which claims about the opacity of other minds offend. Indeed, such claims offend so badly, it seems, that Western scholars cannot help but recoil against them and assert that whatever else they may be about, they certainly cannot be taken to be adequate self-reports of people's ways of approaching language.

Up against the skepticism of their colleagues, ethnographers who find reports of the impossibility of knowing other minds to be paramount in local understandings of speech also face the additional problem that an informant's assertion that "you cannot know what is in another's mind, even if they claim to tell it to you in speech" comes very close to being a version of the liar paradox. If speech cannot convey what speakers think or feel, how can a spoken assertion that this is the case be taken to be an accurate report of the speaker's own experience—of how the speaker goes about approaching speech in his/her mind? Of course, this is only a paradox from the point of view of intentionality based, mind-reading language ideologies that expect speakers...