Wednesday, June 4, 2014

Evans-Pritchard's mistake

Part I of Winch's "Understanding a Primitive Society" (American Philosophical Quarterly Volume I, Number 4, October 1964, pp. 307-324) raises "certain difficulties about Professor E. E. Evans-Pritchard's approach in his classic, Witchcraft, Oracles and Magic among the Azande." What I want to do is to get clear on the nature of these difficulties.

In a critical tone, Winch notes that in Evans-Pritchard's work, "There is more than one remark to the effect that "obviously there are no witches"..." (p. 307). This might make Winch appear to be criticizing E-P simply for disagreeing with the Azande. But I don't think that's quite right.

On p. 308 Winch writes that:
Evans-Pritchard, although he emphasizes that a member of scientific culture has a different conception of reality from that of a Zande believer in magic, wants to go beyond merely registering this fact and making the differences explicit, and to say, finally, that the scientific conception agrees with what reality actually is like, whereas the magical conception does not.
If E-P goes wrong it is not, according to Winch, simply for rejecting Zande belief in witches, oracles, and magic but for doing so on the grounds that such belief supposedly fails to agree with reality. One apparent problem here is that it is hard to see a difference between saying, as I surely may, "I do not believe in witches etc." and saying, as Winch finds problematic, "belief in witches etc. does not agree with reality." Surely (at least sometimes) belief that p = belief that p is true = belief that p agrees with reality. Winch sounds as though he is saying that it is a mistake to go beyond saying "I believe this and you believe that" to saying "I am right and you are wrong." But that can't be right. It can't simply by definition be wrong to think oneself right. If 'right' and 'wrong' mean anything then it is possible that I am right and others with different beliefs are wrong. And if 'right' and 'wrong' mean nothing then it is meaningless to say that I am wrong to say that I am right (or wrong).

Presumably Winch is not making that mistake. He is not talking about what anyone might say but about what Evans-Pritchard says. The mistake, whatever it might turn out to be, is a mistake given Evans-Pritchard's particular goals and commitments. And what it will turn out to be has to do with his making certain claims finally and beyond certain other claims. That is, as a kind of justification. Winch would not object, I take it, to a Zande person assuring us that he believes in magic, nor to a European insisting that he believes in science. Nor would he object, I believe, to these assurances being made in the idiom of agreement with reality. The problem he sees is with trying to claim in a non-circular way that one's preferred conception of reality is justified by reality itself.
     Evans-Pritchard [...] is trying to work with a conception of reality which is not determined by its actual use in language. He wants something against which that use can itself be appraised. But this is not possible; and no more possible in the case of scientific discourse than it is in any other. (p. 309)
Here is another problem. What is not possible? If this allegedly impossible thing is conceivable then don't we need proof, rather than mere assertion, that it really is impossible? And if, as I suspect, Winch means that it is not really conceivable because it is incoherent, then what is the 'it' that Evans-Pritchard wants but cannot have? What is the 'it' that we cannot conceive? It seems a little uncharitable to attribute such incoherence to Evans-Pritchard. But then I suppose it would be equally uncharitable of me to attribute it to Winch. Let's say for now that Winch seems to see in Evans-Pritchard an apparent confusion. We get onto firmer footing very soon, thankfully.

On p. 310 Winch quotes a series of definitions that E-P provides of terms such as 'mystical,' 'scientific,' and 'ritual'. It is clear enough, at least with the help of Winch's italics, that E-P has built pro-science and anti-mysticism, anti-ritual bias into the definitions. If we define the relevant terms in such a way then it is not reality that will justify our preference for science over magic but simply our own uses of language. And that is no justification of such uses at all.

Does consulting oracles make sense? Winch considers this question starting on p. 311. Azande do not treat oracular pronouncements as hypotheses but as guides to action. Is it possible to make decisions about what to do in this way? Yes. Does it make sense to do so? To those who do so, yes. To us, probably not. The disagreement between Winch and E-P arises because:
it is clear from other remarks in the book to which I have alluded, that at the time of writing it he would have wished to add: and the European is right and the Zande wrong. This addition I regard as illegitimate and my reasons for so thinking take us to the heart of the matter. (p. 313)
So here we are, at the heart of the matter:
Evans-Pritchard is not content with elucidating the differences in the two concepts of reality involved [in the language of the Azande and in our language]; he wants to go further and say: our concept of reality is the correct one, the Azande are mistaken. But the difficulty is to see what "correct" and "mistaken" can mean in this context. (p. 313)
Another difficulty might be to see what "illegitimate" can mean in the context of the previous quotation from p. 313. I don't think Winch can really mean that E-P has no right to prefer the European 'language' to the Zande one. The mistake that Winch has in mind is that of thinking that one language can be more correct than another in a purely objective, value- and preference-free way. Or at least that the European, scientific language is more correct in this kind of way than the Zande language. Perhaps some possible language would be riddled with problems, but the Azande got along with theirs for a very long time (apparently they have moved over to ways of thinking more like ours now, but that's beside the point). In what sense could their language and way of life be (not immoral or ugly or unpleasant but) incorrect or mistaken?

One way that Winch considers would be if it involved contradiction. But merely possible hypothetical contradictions are not a real problem, and it is not irrational to ignore them. Apparent contradictions that do arise must be dealt with in some way, but if there is a way to deal with them then the system as a whole, including this method for dealing with what might otherwise be a problem, is not irrational. Theoretical contradictions matter for theoretical systems, but not everything that might look like such a system really is one. This is Winch's claim about Zande ideas about the inheritance of witchcraft. If witchcraft is inherited, as the Zande (used to) say, then if one member of a clan biologically-related through the male line is a witch the whole clan must be, and if one is not then none can be. There is a post-mortem test that shows whether a person really was "a witch" or not, though, so this could seemingly be used either to show that everyone or no one is a witch. One positive result here or there need not brand the whole clan as witches because the witch might have been illegitimate or swapped at birth or whatever, but enough results one way or the other would surely clear the matter up. Is it irrational, or less rational, of the Azande not to carry out the relevant tests? Winch thinks not. They simply are not interested in knowing about global witch-statuses. This is not a mistake. It is a mistake, though, if we think that they somehow ought (in some supposedly absolute sense) to have such an interest.

This, I gather, is what Winch takes Evans-Pritchard's mistake to be. He wants our standards of rationality, the norms that govern our assessment of ways of thinking and behaving, to justify themselves and delegitimate those that differ from them. He wants, as it were, a kind of neutral or non-evaluative judgment of value. This is incoherent. That's one problem.

Another problem, it seems to me, is that E-P wants to understand the Azande but goes about it in a way that will not work. Later in the paper (p. 319) Winch writes:
Since it is we who want to understand the Zande category [of magic], it appears that the onus is on us to extend our understanding so as to make room for the Zande category, rather than to insist on seeing it in terms of our own ready-made distinction between science and non-science.
The issue here is not what makes sense or what is in some sense legitimate but simply what will get us what we want. We will not understand those whose concepts are very different from ours if we insist on an overly simple application of our concepts to theirs or translation of their concepts to ours. Of course we do need to relate their concepts to ours in some way, but we are not stuck, as Winch points out, with a fixed stock of concepts or expressions. We can add to what we have, and part of the point of studying other cultures is to grow in this way.

In the end, then, Winch appears to have identified two errors in Evans-Pritchard's thinking: a pragmatic one and a logical one. The logical one is to try to use your own concepts (or way of life) to justify themselves (or itself). The pragmatic one is to try to understand the behavior of another people without the necessary flexibility and open-mindedness, and, indeed, from a starting point that defines that behavior as irrational.


  1. Your ongoing discussion of Winch is very interesting, and I agree with much of what you say. I’m sorry I’m so late in commenting.

    You write: “Winch would not object, I take it, to a Zande person assuring us that he believes in magic, nor to a European insisting that he believes in science. Nor would he object, I believe, to these assurances being made in the idiom of agreement with reality.”

    Would he not object? I hope it won’t simply be found pedantic to raise the question in what circumstances the Zande or the European might be imagined to be offering such assurances. These are hardly the kind of thing that the Azande, for instance, would go around telling each other in the normal course of affairs. Perhaps we could imagine the following situation: the Zande guy is tired of all the missionaries, Chinese businessmen, safari tourists, etc, who keep shaking their head over the practice with the poison oracle. Finally he blurts out: “Well, I *believe* in magic, and that’s that!”

    What this situation seems to presuppose is that he has discovered that what he and his kinfolk are practising is “magic”, i.e. that that is a word Westerners would apply to some of their proceedings. (We may of course wonder what that means to him. Does he also know, for instance, that Westerners apply the same word to Caribbean voodoo practices, to Cheyenne shamanism, etc? If he does know, how does he react to that fact?)

    It seems a natural assumption that if the Zande is to become aware of himself as “someone who believes in magic”, this will come about through his discovering that certain activities and ways of thinking that he shares with his tribe come under a certain kind of (critical? curious?) scrutiny by outsiders. This, then, raises certain further questions. What does he make of the fact that outsiders find their practices curious? Is he able to identify what piques their curiosity without sensing what they consider problematic about it?

    On the other hand, what about the European? In a discussion with a climate skeptic she might say, “Well, *I* believe in what science tells us.” Her interlocutor would probably not agree that it’s a question of believing or not believing in science. He might say, “Of course I believe in science, I simply don’t think those results are unambiguous”. She might call him irresponsible, accuse him of wishful thinking, of denialism, etc. Neither of them would think that they had different “conceptions of reality”.

    “Believing in science”, it seems, is not so much a matter of holding certain beliefs as of belonging to a culture in which certain forms of discourse (appeals to “what science has shown”, etc) have an important role. How is this different from the Azande culture? Of course, they have no such profession as scientist. But on the other hand, they too make careful observations of nature, draw conclusions and argue about them, derive lessons for the future, etc. This seems more a difference of degree than of kind. It is just that, in some cases, what may stare them in the face is an obvious case of witchcraft, whereas for us it never is. (Our “believing in science”, one might suggest, is primarily a matter of the kinds of thing we *don’t* believe in.)

    What one should beware of, I would argue, is to think of “conceptions of reality” in abstract terms, as if, say, belief in science and belief in magic formed two unified, and contrasting, intellectual systems.

    1. Hi, just some stray thoughts --

      first of all, Evans-Pritchard seems pretty nuanced in what he writes about Zande magic and he goes probably further than the anthropologists writing before him to understand 'magic', not as substitute science but as something woven into everyday life in a different way. Perhaps Winch got interested in the work because it after all gives the reader a chance to see why Zande ideas about witchcraft are not simply confused.

      Thinking of the question what Westerners 'believe in' and the Azande don't, from Evans-Pritchard's description one can see that the Azande don't 'believe in' *chance* (things happening by accident). Ideas of witchcraft are a way of figuring out who is to blame in case things go wrong. Nothing ever happens by accident, but there is always someone responsible. Now *this* is not simply a question of what science would say because of course many scientists would also say there are no genuine accidents. Perhaps one thing this shows that the 'western' point of view (supposing something like that can be identified) is not simply a (or *the*) 'scientific' point of view because other issues come into this contrast. Winch of course devotes the last part of his paper precisely to the discussion about different possible roles that contingencies may play in people's lives.

    2. Thank you, Lars. No, that doesn't seem pedantic at all.

      I wanted to explore the possible meaning of "agreement with reality" and suggest that one possible use of these words, as meaning truth or something close to that idea, is all right. As you say, in a discussion about climate change both sides are likely to insist that they believe in science. And I think it's reasonable to imagine a Zande person insisting that he believes in magic in a parallel way. If asked why, by Evans-Pritchard perhaps, such a person might claim that reality itself justifies this belief.

      Perhaps there might be talk of "agreement with reality" in this kind of discussion (from Evans-Pritchard's book): A boy knocked his foot against a small stump of wood in the centre of a bush path, a frequent happening in Africa, and suffered pain and inconvenience in consequence. Owing to its position on his toe it was impossible to keep the cut free from dirt and it began to fester. He declared that witchcraft had made him knock his foot against the stump. I always argued with Azande and criticized their statements, and I did so on this occasion. I told the boy that he had knocked his foot against the stump of wood because he had been careless, and that witchcraft had not placed it in the path, for it had grown there naturally. He agreed that witchcraft had nothing to do with the stump of wood being in his path but added that he had kept his eyes open for stumps, as indeed every Zande does most carefully, and that if he had not been bewitched he would have seen the stump. As a conclusive argument for his view he remarked that all cuts do not take days to heal but, on the contrary, close quickly, for that is the nature of cuts. Why, then, had his sore festered and remained open if there were no witchcraft behind it? This, as I discovered before long, was to be regarded as the Zande explanation of sickness.

      Here is a case of a Zande person being made aware that he is "someone who believes in magic" and defending this belief by reference to various facts or features of reality. There is a sense in which reality supports (without proving correct) his belief in magic, it seems to me. It supports belief in science too, of course.

      What Winch calls "the scientific universe of discourse" is something like an orientation towards reality (as is the magical universe of discourse), and such an orientation cannot be justified by reality itself. If we take the "western" view then we will attribute the boy's injury to carelessness and the location of the wound. If we take the Zande view then we will attribute it to witchcraft. In either case we make assumptions that would be impossible (or so it seems) to prove right or wrong. And so reality does not show one view of the reasons for the boy's injury as being correct and the other incorrect. But of course it is possible and intelligible to insist that one is correct (as the Zande boy does), and if one were to use the words "agreement with reality" or their Zande equivalent as part of (but not as a justification of) such an insistence then I think both that this would be all right and that Winch would think it was all right too. That is, he would think it was OK to speak that way, not that he would agree with the Zande boy.

      (continued below)

    3. I think I agree with your point that: "What one should beware of, I would argue, is to think of “conceptions of reality” in abstract terms, as if, say, belief in science and belief in magic formed two unified, and contrasting, intellectual systems." It's not as if the Azande ignore empirical evidence or all kinds of reasoning. But their thinking about why certain events happen is clearly different from someone like Evans-Pritchard's. In a way I think this does reflect the existence of two (more or less) unified, and contrasting, intellectual systems. But Zande magic, like science, is not just a unified intellectual system. Neither magic nor science is probably completely unified or systematic, and both are tied up with ways of living, of dealing with things like foot injuries and crop failures. They aren't abstract theories, and it would be a mistake to treat them as if they were.

    4. Hi Olli, and thanks. Yes, I think Winch might have chosen to discuss Evans-Pritchard's work in part because he agrees with so much of it.

      I don't know whether there is such a thing as the 'Western' point of view, but there is a difference between Evans-Pritchard's point of view and that of the Azande, and E-P's view is not his alone. So there is a contrast to be made. On the other hand, right before the story about the Zande boy that I quoted above E-P writes that: I found it strange at first to live among Azande and listen to naive explanations of misfortunes which, to our minds, have apparent causes, but after a while I learnt the idiom of their thought and applied notions of witchcraft as spontaneously as themselves in situations where the concept was relevant. So the contrast is not absolute. That is, it is apparently quite possible to 'believe in science' (as E-P presumably still did, although what this means now seems at least slightly problematic) and to spontaneously apply notions of witchcraft. Is such spontaneous application of notions belief? I suppose not. I suppose that if asked whether he believed in witchcraft E-P would have said No. But if he really spontaneously applied those notions then he thought and behaved as if he believed. And that is surely close to belief.