8
Medical Anthropology and the Problem of Belief
Byron J. Good
■       - |
I begin with an intuition that there is a close relationship between science, including medicine, and religious fundamentalism that turns, in part, on our concept "belief." For fundamentalist Christians, salvation follows from belief, and mission work is conceived as an effort to convince the unbelievers to take on a set of beliefs that will produce a new life and ultimate salvation. Ironically, quite a-rcligious scientists and policy makers see a similar benefit from correct belief. Educate the public about the hazards of drug use, our current Enlightenment theory goes, get people to believe the right rbing and the problem will disappear. Educate the patient, medical journals advise clinicians, ami solve the problems of noncompliance that plague the treatment of chronic disease. Investigate public beliefs about vaccinations or risky health  behaviors  using  the  Health Belief Model, a generation of health psychologists has told us, get people to believe the right thing and our public health problems will resolve. Salvation from drugs and from preventable illness will follow from correct belief.
Wilfred Cantwell Smith, a comparative historian of religion and theologian, argues that the fundamentalist conception of belief is a
recent Christian heresy (Smith 1977, 1979). 1 want to explore the hypothesis that anthropology has shared this heresy with religious fun. damentalists, that "belief" has a distinctive cultural history within anthropology and that the conceptualization of culture as "belief" is far from a trivial matter.
A quick review of the history of medical anthropology will convince the reader that "belief" has played a particularly important analytic role in this subdiscipline, as it has in the medical behavioral sciences and in public health. Why is there this deep attachment to analyzing others' understandings of illness and its treatment as medical "beliefs" and practices, and why is there such urgency expressed about correcting beliefs when mistaken? To begin to address this issue, I first describe the general theoretical paradigm that frames what I have referred to as the "empiricist theory of medical knowledge." I will indicate its relationship to the intellectualist tradition in anthropology and to debates about rationality and relativism, showing how the language of belief functions within the rationalist tradition. At the end of this chapter, I review recent criticisms that have shaken the foundations of this paradigm, criticisms that suggest the need for an alternative
Byron |. Good, "Medical Anthropology and the Problem of Belief," pp. 1-24 [modified} in Medicine, Rationality and Experience. Cambridge: Cambridge University Press, 1994. Reprinted by permission o: Cambridge University Press.
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
65
direction in the field. This discussion will serve to frame the constructive chapters that follow.
The language of clinical medicine is a highly technical language of the biosciences, grounded in a natural science view of the relation between language, biology, and experience (B. Good and M. Good 1981). As George Engel (1977) and a host of medical reformers have shown, the "medical model" typically employed in clinical practice and research assumes that diseases are universal biological or psychophysiological entities, resulting from somatic lesions or dysfunctions. These produce "signs" or physiological abnormalities that can be measured by clinical and laboratory procedures, as well as "symptoms" or expressions of the experience of distress, communicated as an ordered set of complaints. The primary tasks of clinical medicine are thus diagnosis - that is, the interpretation of the patient's symptoms by relating them to their functional and structural sources in the body and to underlying disease entities - and rational treatment aimed at intervention in the disease mechanisms. All subspecialties of clinical medicine thus share a distinctive medical "hermeneutic," an implicit understanding of medical interpretation. While patients' symptoms may be coded in cultural language, the primary interpretive task of the clinician is to decode patients' symbolic expressions in terms of their underlying somatic referents. Disordered experience, communicated in the language of culture, is interpreted in light of disordered physiology and yields medical diagnoses.
| ... | The empiricist theory of medical language is grounded in what philosopher Charles Taylor (1985a, 1985b, 1989) calls "the polemical, no-nonsense nominalism" of Enlightenment theories of language and meaning. For seventeenth-century philosophers such as Hobbes and Locke, the development of a language for science required a dcinystification of language itself, showing it to be a pliant instrument of rationality and thought, as well as the emergence of a disenchanted view of the natural world. The development of such a natural philosophy and the attendant theory of language required the separation of "the order of words" from "the order of things," in
Foucault's terms (1970), the freeing of the order of language and symbols from a world of hierarchical planes of being and correspondences present in Renaissance cosmology. What we must seek, Francis Bacon argued, is not to identify ideas or meanings in the universe, but "to build an adequate representation of things" (Taylor 1985a: 249). Thus, theories of language became the battle ground between the religious orthodoxy, who conceived "nature" as reflecting God's creative presence and language as a source of divine revelation, and those who viewed the world as natural and language as conventional and instrumental. What emerged was a conception of language in which representation and designation are exceedingly important attributes, f... |
This broad perspective has the status of a "folk epistemology" for medical practice in hospitals and clinics of contemporary biomedi-cine. A person's complaint is meaningful if it reflects a physiological condition; if no such empirical referent can be found, the very mean-ingfulness of the complaint is called into question. Such complaints (for example of chronic-pain - see M. Good et al. 1992) are often held to reflect patients' beliefs or psychological states, that is subjective opinions and experiences which may have no grounds in disordered physiology and thus in objective reality. "Real pathology," on the other hand, reflects disordered physiology. Contemporary technical medicine provides objective knowledge of such pathology, represented as a straight-forward and transparent reflection of the natural order revealed through the dense semiotic system of physical findings, laboratory results, and the visual products of contemporary imaging techniques. And "rational behavior" is that which is oriented in relation to such objective knowledge.
Rationality and the Empiricist Paradigm in Anthropology
The empiricist paradigm is most clearly represented by the intellectualist tradition in anthropology, which was prominent in Britain at the turn of the century and reemerged under the
66
BYRON J. GOOD
banner of Neo-Tylorianism in an important set of debates about the nature of rationality during the 1970s.' Even a cursory examination will indicate how the rationalist position flows out of the "Enlightenment" tradition of anthropology, demonstrate the critical role of "belief" in this paradigm, and suggest why it has had such power within medical anthropology.
A central issue in the rationality debate has been discussion of the problem of "apparently irrational beliefs" (for example Sperber 1985: ch. 2). How do we make sense of cultural views of the world that arc not in accord with contemporary natural sciences, it is often asked. Do we argue that members of traditional cultures live in wholly different worlds, and their statements are true in their worlds, not ours, or even that they cannot be translated intelligibly into our language? Advocates of a typical rationalist position hold that such relativism is essentially incoherent, and have often argued either that seemingly irrational statements must be understood symbolically rather than literally or that they represent a kind of "proto-science," an effort to explain events in the world in an orderly fashion that is a functional equivalent of modern science. The crucial interpretive problem, for this tradition, is how to answer a question stated explicitly by Lukes (1970: 194): "When I come across a set of beliefs which appear prima facie irrational, what should be my attitude toward them?" Given our claims that other forms of thought are rational, how do we make sense of beliefs that are obviously false?
lor much of this debate, Evans-Pritchard's Witchcraft, Oracles and Magic among the Azattde (1937) serves as the primary source. 1 his book was the first and is arguably still ihe most important modernist text in medical anthropology. It has had enduring influence because of the wealth of the ethnography and the richness of its interpretation of witchcraft as an explanation for illness and misfortune. Which anthropologist can think of cultural responses to misfortune without conjuring the image ol Evans-Pritchard's young lad stubbing his toe and blaming witchcraft for its failure to heal, or of the granary collapsing? To these misfortunes, the Zande explanation was clear. "Every Zande knows that termites eat the
supports [of the granaries] in course of tiniv, and that even the hardest woods decay aftor years of service," Evans-Pritchard reports. Bi,r "why should these particular people have bee^ sitting under this particular granary at the paK ticular moment when it collapsed?" Thus, a|_ though practical reasons explain the immediate causes of illness and misfortune, the Azanc1e turn to witchcraft to answer the "why me?1' question, to find an underlying cause in th« moral universe and a response that is socially embedded and morally satisfying.
The Azande text has been the key for tl\e rationality debate for another reason. Evans. Pritchard in this text was explicitly empiricist; and his work provided examples that serve n.s paradigmatic challenges to relativism. Takt> for example, his analysis of the Zande autopsy to investigate witchcraft, which appears as a substance in the intestine of a witch. Since witchcraft is inherited by kin, an autopsy may be performed on a deceased kinsman to determine whether others bear the unwanted sub-Stance. Evans-Pritchard (1937: 42) describes the scene:
Two lateral gashes are made in the belly and one end of the intestines is placed in a cleft branch and they are wound round it. After the other end has been severed from the body another man takes it and unwinds the intestines as he walks away from the man holding the cleft branch. The old men walk alongside the entrails as they are stretched in the air and examine them for witchcraft-substance. The intestines are usually replaced in the belly when the examination is finished and the corpse is buried. I have been told that if no witchcraft-substance were discovered in a man's belly his kinsmen might strike his accusers in the face with his intestines or might dry them in the sun and afterwards rake them to court and there boast of their victory.
Evans-Pritchard's (1937: 63) interpretation of this dramatic scene is telling.
It is an inevitable conclusion from Zande descriptions of witchcraft that it is not an objective reality. The physiological condition which is said to be the seat of witchcraft, and which 1 believe to be nothing more than food passing through the small intestine, is an objective
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
67
condition, but the qualities they attribute to it and the rest of their beliefs about it are mystical. Witches, as Azande conceive them, cannot exist.
He goes on immediately to argue that although mistaken, the Zande views serve as a natural philosophy and embrace a system of values which regulate human conduct. They are, however, mystical. "Mystical notions," he argues in the book's introduction (p. 12), are those that attribute to phenomena "supra-sensible qualities," "which are not derived from observation" and "which they do not possess." "Common-sense notions" attribute to phenomena only what can be observed in them or logically inferred from observation. Though they may be mistaken, they do not assert forces that cannot be observed. Both are distinct from "scientific notions." "Our body of scientific knowledge and Logic," he says (p. 12), "are the sole arbiters of what are mystical, common-sense, and scientific notions."
Evans-Pritchard assumes in this account that the meaning of Zande "medical discourse" - whether of witchcraft, oracles, or "leechcraft" - is constituted by its referential relationship to the natural order as reflected in empirical experience. Analysis in the rationality literature follows from this assumption; it frames Zande beliefs as propositions, then questions the verifiability and the deductive validity of their inferences. Since we know that witches cannot exist empirically, it is argued, the rationality of Zande thought is called into doubt. It follows that the anthropologist must therefore organize analysis in response to the following kinds of questions. How can a set of beliefs and institutions which are so obviously false (propositionally) be maintained for such long periods of time by persons who in much of their lives are so reasonable? How could they possibly believe that, and why haven't their beliefs progressed, that is come to represent the natural world more correctly? Do such beliefs imply that the Zande have a different "mentality" or different psychological or logical processes than we? Do they simply divide up the common-sense and religious domains differently than
do we (as Evans-Pritchard responded ro Levy-Bruhl)? Are some societies simply organized around views that are reasonable but wrong?
Not altogether obvious in Evans-Pritchard's text is the juxtaposition of "belief" and "knowledge." The book is devoted largely to Zande mystical notions - witchcraft and sorcery - and ritual behaviors, such as resort to the poison oracle. One chapter, however, entitled "Leech-craft," is devoted to their common-sense notions of sickness. The language of "belief" and "knowledge" mirror this distinction. The book begins: "Azande believe that some people-are witches and can injure them in virtue of an inherent quality ... They believe also that sorcerers may do them ill by performing magic rites with bad medicines ... Against both they employ diviners, oracles, and medicines. The relations between these beliefs and rites are the subject of this book" (p. 21; my emphasis). On the other hand, the Leechcraft chapter argues: "Azande know diseases by their major symptoms" (p. 482). "The very fact of naming diseases and differentiating them from one another by their symptoms shows observation and common-sense inferences" (pp. 494-5). Thus, the book is organized around a distinction between those ideas that accord with objective reality | .., ] and those that do not; the language of knowledge is used to describe the former, the language of belief the latter. Evans-Pritchard's text transcends its empiricist formulation, in particular because of the subtlety of its analysis of Zande reasoning and the location of witchcraft in Zande social relations, bur it makes explicit many of the assumptions found more generally in the rationality tradition and shared by much of the medical social sciences.
If Evans-Pritchard's work on the Azande is the classic modernist text on witchcraft and illness, Jeanne Favret-Saada's Deadly Words: Witchcraft in the Bocage (1980), first published in French in 1977, is surely the classic post-modernist ethnography on the topic. Favret-Saada's ethnography is a first-person account of her effort to investigate witchcraft in rural France. In the early months of her work, villagers referred her to a few well known healers who were often interviewed by
68
BYRON J. GOOD
the press, but the peasants themselves refused to discuss the matter with her. Witchcraft? Only fools believe in that!
"Take an ethnographer," she begins (1980: 4). "She has spent more than thirty months in the Bocage in Mayenne, studying witchcraft ... 'Tell us about the witches', she is asked again and again when she gets back to the city. Just as one might say: tell us tales about ogres or wolves, about Little Red Riding Hood. Frighten us, but make it clear that it's only a story; or that they are just peasants: credulous, backward and marginal ...
"No wonder that country people in the West are not in any hurry to step forward and be taken for idiots in the way that public opinion would have them be ..."
The book is an account of how she eventually found her way into the discourse of witchcraft. She was taken ill, beset with accidents, and sought the aid of a healer in the region, an unwitcher. She began to interview a man and his family, whom she had met when the man was a patient in a mental hospital. As they told her the details of his illness and who they suspected might be responsible, she realized that they saw her as a healer and now expected her to act on their behalf. Why else would she ask about such matters so explicitly? Only the powerful would dare to ask such questions. Simply by asking about their difficulties, she was seen to be entering into their struggle with an enemy wishing them harm, a life and death struggle in which she was now an advocate for their interests. Witchcraft, she came to see, was a battle of powerful wills, a fight to the death, a fight through the medium of spoken words. One could only talk about witchcraft from an engaged position - as one bewitched, as a suspected witch, or as one willing to serve as unwitcher. To engage in talk was to enter the struggle.
In Favret-Saada's account, the language oi belief, the position of the ethnographer, and assumptions about the relation of culture and reality are radically different than in Evans-Pritchard's text. Science for Favret-Saada is not the arbiter between the empirically real and the mystical, as for Evans-Pritchard, but   one   of  several   "official   theories of
misfortune," backed by powerful social ag^n cies: the School, the Church, the Medical A^So ciation.  Language is not a set of neu^r, i propositions about the world, which the ^.i nographer judges to be more or less empirically valid, but the medium through which vici^us and life-threatening power struggles are e _ gaged. The world of illness and witchcraft a^u, opens to the ethnographer as she enters t|)e discourse. And much of the text turns on u'onjc reflections on "belief" - the peasants' clajnis not to believe in witchcraft, even as they s^e^. the help of the unwitcher; the mocking vie\\> Qf the authorities about those who do believe; ^nt] Favret-Saada's juxtaposition of the meaning Qf belief in her text and in that of Evans-Pritch;\u| For many ethnographers, as for the Fret-lc|1 press, the question is whether the peas;\Ilts really believe in witchcraft, and if so, how rkev can hold such beliefs in today's world. But for those attacked by a sorcerer, for those pease\11ts - and Favret-Saada herself - whose very liVes were at stake, belief in witchcraft is not r|K. question. Flow to protect oneself, how to won| off the evil attacks producing illness and t\\is-fortune, is the only significant issue to kg addressed.
Much has changed in the world of anth.r0_ pology between that of 1935 colonialist AfVjC;l and  contemporary  post-colonialist ethnography. Evans-Pritchard's confident positioning of himself as observer and arbiter of the rationality of the native discourse is largely unavailable to us today. And throughout the historv and   sociology   of   science,   the confident recording of science's progress in discovering the facts of nature has also given way. 1 \\•[]] return   to  these   issues   as   the discussion proceeds,  but the juxtaposition  of Ev^ns-Pritchard's and Favret-Saada's texts brings into focus the role of "belief"  as  an analytic category in the history of anthropology and in the study of such phenomena as witchcraft, provoking several questions. Why has the discussion of others' beliefs come to be invoked increasingly with irony? What is the role of belief in the empiricist paradigm, and why has that position begun to give way? Where does the disjunction between "belief" and "knowledge," which I noted in Witchcraft, Oracles, and Magic and which serves as the basis for
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
69
bavret-Saada's irony, come from? Why "belief," and what is at stake here?
The Problem of Belief in Anthropology
Rodney Needham's Belief, Language and Experience, published in 1972, is the classic examination of the philosophy of belief by an anthropologist. Needham explores in great detail assumptions about belief as mental state, asking whether philosophers have formulated this with adequate clarity to allow us to use the term in cross-cultural research, and asking whether members of other societies indeed experience what we call "belief." After an extraordinary review, he concludes both that philosophers have failed to clarify "the supposed capacity for belief" and are unlikely to do so, and that evidence suggests the term may well not have counterparts in the ethnopsycho-logical language of many societies. Needham's analysis suggests that Evans-Pritchard's claim that the Azande believe some people are witches may be a less straightforward description of the mental states of Zande individuals than we usually presume. For the moment, however, I want to focus on another dimension of belief as anthropologists have used the term in cultural analysis.
Mary Steedly, an anthropologist who worked with the Karobatak people in Sumatra, tells how when she was beginning fieldwork she was often asked a question, which she understood to mean "do you believe in spirits?" (1993: ch. I). It was one of those embarrassing questions anthropologists struggle to answer, since she didn't, personally, bur respected and wanted to learn about the understandings of persons in the village in which she worked. After stumbling to answer the question for some months, she discovered her questioners were asking "Do you trust spirits? Do you believe what they say? Do you maintain a relationship with them?" Any sensible person helieves in their existence; that isn't even a meaningful question. The real question is how one chooses to relate to them.
Anthropologists often talk with members of other societies about some aspect of their
world which does not exist in ours and which we are comfortable asserting is not part of empirical reality. How is it that "belief" has come to be the language through which we discuss such matters - the Zande witches, or the three humors wind, bile, and phlegm in Ayurvedic medicine, or the four humors of seventeenth-century European and American medicine? Moreover, why have we in Western civilization given such importance to beliefs, such importance that wars in Christendom are fought over beliefs, that church schisms and persecutions and martyrdom revolve around correct belief? How is it that belief came to be so central to anthropological analysis, and what is implied by the juxtaposition of belief and knowledge?
By far the richest discussion of the history of the concept belief is to he found in the writing of Wilfred Cantwell Smith, the historian of religion, whose lectures when I was a graduate student set me to thinking about these matters. In two books completed during the late 1970s, Smith explores the relation between "belief" and "faith" historically and across religious traditions. He sets out not to compare beliefs among religions, but to examine the place of belief itself in Buddhist, Hindu, Islamic, and Christian history. Through careful historical and linguistic analysis, he comes to the startling conclusion that "the idea that believing is religiously important turns out to be a modern idea," and that the meaning of the English words "to believe" and "belief" have changed so dramatically in the past three centuries that they wreak profound havoc in our ability to under stand our own historical tradition and the religious faith of others.
The word "belief" has a long history in the English language; over the course u has so changed that its earlier meanings are only dimly felt today (Smith 1977: 41-46; 1979: 105—27). In Old English, the words which evolved into modern "believe" (geleofan, gele-fan, geliefan) meant "to belove," "to hold dear," "to cherish," "to regard as lief." They were the equivalent of what the German word belieben means today (mciu lieber Ffeund is "my dear or cherished friend"), and show the same root as the Latin libel, "it pleases," or
70
BYRON J. GOOD
libido, "pleasure." This meaning survives in rhe Modern English archaism "lief" and the past participle "beloved." In medieval texts, "leve," "love," and "beleuc" are virtual equivalents. In Chaucer's Canterbury Tales, the words "accepted my bileve" mean simply "accept my loyalty; receive me as one who submits himself to you." Thus Smith argues that "belief in God" originally means "a loyal pledging of oneself to God, a decision and commitment to live one's life in His service" (1977: 42). Its counterpart in the medieval language of the Church was "I renounce the Devil," belief and renunciation being parallel and contrasting actions, rather than states of mind.
Smith (1977: 44} sums up his argument about the change of the religious meaning of "belief" in our history as follows:
I he affirmation "I believe in God" used to mean: "Given the reality of God as a fact of the universe, I hereby pledge to Him my heart and soul. I committcdly opt to live in loyalty to Him. I offer my life to be judged by Him trusting His mercy." Today the statement may be taken by some as meaning: "Given the uncertainty as to whether there be a God or not, as a fact of modern life, I announce that my opinion is 'yes'. I judge God to be existent."
Smith argues that this change in the language of belief can be traced in the grammar and semantics of English literature and philosophy, as well as popular usage. Three changes - in the object of the verb, the subject ol the verb, and the relation of belief and knowledge - serve as indicators of the changing semantics of the verb "to believe." First, Smith finds that grammatically, the object of the verb "to believe" shifted from a person (whom one trusted or had faith in), to a person and his word (his virtue accruing to the trustworthiness oi his word), to a proposition. This latter shift began to occur by the end of the seventeenth century, with Locke, for example, who characterized "belief" along with "assent" and "opinion" as "the admitting or receiving any proposition lor true, upon arguments or proofs that are found to persuade us ... without certain knowledge ..." (Smith 1977: 48). In the twentieth century we have seen a further shift
as beliefs have come to mean "presupp0s_ itions," as in "belief systems."
A second shift has occurred in the subject of the verb "to believe," from an almost exclu$jVe use of the first person - "1 believe" - to t|le predominant use of the third person, "he be-Iieves" or "they believe." In anthropology, tne impersonal "it is believed that" parallels ^ discussion of culture as belief system or system of thought. This change in subject subtly shifts the nature of the speech act involved - from existential to descriptive - and alters the authorization of the speaker.
Third, Smith observes that an iinport^nt and often unrecognized change has occurred in the relation of belief to truth and knowledge, as these are historically conceived. Bacon wrote in 1625 of "the belief of truth," which he defined as the "enjoyment of it," in contrast to the inquiry or wooing of truth and the knowledge or presence of truth. Belief maintains its sense here of holding dear, of appropriating to oneself that which is recognized as true. By the nineteenth century, however, *t0 believe" had come to connote doubt, and today it suggests outright error or falsehood. Knowledge requires both certitude and correctness; belief implies uncertainty, error, or both. [ ... | Smith's favorite illustration of the juxtaposition of belief and knowledge is an entry jn the Random House dictionary which defined "belief" as "an opinion or conviction," and at once illustrates this with "the belief that the earth is flaf\ Indeed, it is virtually unacceptable usage to say that members of some society "believe" the earth is round; if this is part of their world view, then it is knowledge, tun belief!
Smith goes on to argue that our failure to recognize this shift in meaning has led to mistranslation of texts in the Christian tradition and ultimately to "the heresy of believing," the deeply mistaken view that belief in this modern sense is the essence of the religious life rather than faith. Credo, in the Latin, is literally "1 set my heart" (from Latin cordis or heart [as in cordial] and *~do or "'-dere, to put). Credo in unum Deum was correctly translated in the sixteenth century as "I believe in one God," when it meant "I formally pledge my allegiance to God," Whom we of course all acknowledge
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
71
to be present in the world. Today, it is a mistranslation, suggesting that the Credo consists of propositions the veracity of which we assert. This is historically inaccurate and profoundly misrepresents the traditional ritual acclamation. Equally importantly, for the comparati-vist, the misplaced focus on beliefs as the primary dimension of religious life has led to mistranslations and misunderstandings of other religious traditions, and in Smith's view, to the great failure to explore the faith of others in their historical and communal contexts, even to make faith a central category in comparative research.
Smith's argument about the importance of placing the study of faith rather than beliefs at the center of comparative and historical studies of religion has important implications for the study of illness experience. My interest at this time, however, is the place of "belief" in the history of anthropology, and what the use of the term tells us about the anthropological project. In what way does Smith's analysis of belief relate to the use of the term in anthropological writing? What is the history of believing in anthropology? How is the use of "belief" related to the epistemological assumptions of anthropologists?
From my initial explorations, it would appear that the term "belief" as it is employed in anthropology does indeed connote error or falsehood, although it is seldom explicitly asserted. A quick scan of the typical volumes on an anthropologist's shelf will provide many examples. My own favorite, paralleling Smith's discovery in the Random House Dictionary, comes from Ward Goodenough's little book, Culture, Language and Society (1981). In a discussion of "propositions" and the nature of reasoning cross-culturally, he provides the following example from the German ethnologist Girschner, to illustrate the "reasonableness" of members of other cultures.
Consider, for example, the following comment by a Micronesian navigator, defending his belief that the sun goes around the earth (Girschner, 1913 ...)
I am well aware of the foreigner's claim that the earth moves and the sun stands still, as someone once told us; but this we cannot
believe, for how else could it happen that in the morning and evening the sun burns less hot than in the day? It must be because the sun has been cooled when it emerges from the water and toward setting it again approaches the water. And furthermore, how can it be possible that the sun remains still when we are yet able to observe that in the course of the year it changes its position in relation to the stars? (emphasis added] (Goodenough 1981: 69).
Quite reasonable, even if mistaken: that is how the beliefs of others seem to be.
The juxtaposition of belief and knowledge is most evident in the intellectualist writing of turn-of-the-century British social anthropology. An example from a classic text in medical anthropology will be particularly instructive. W. H. R. Rivers' Medicine, Magic and Religion was published in 1924, the first major comparative study of medical systems by an anthropologist-physician/ The book is designed to show how concepts of disease vary cross-culturally, but focuses largely on beliefs about causation of disease. Rivers uses "believe" largely in the third person or impersonally; the object of belief is almost exclusively propositions; and these propositions are, from Rivers' point of view, counter-factual. For example, he writes (1924: 29):
Thus, in Murray Island, in Torres Straits, disease is believed to occur by the action of certain men who, through their possession of objects called zogo and their knowledge of the appropriate rites, have the power of inflicting disease. Thus, one zogo is believed to make people lean and hungry and at the same time to produce dysentery; another will produce constipation, and a third insanity.
His attitude is made clear several pages later, when he discusses the rationality of such beliefs. "From our modern standpoint we are able to see that these ideas are wrong. But the important point is that, however wrong may be the beliefs of the Papuan and Melanesian concerning the causation of disease, their practices are the logical consequence of those beliefs." This view is conveyed more subtly, however, and with far more profound implications at the end of the book. The conclusion is devoted to illuminating the role of belief in the practice of
72
BYRON J. GOOD
Western medicine. Whereas in earlier chapters of the hook, the word "believe," along with "ascribe," "regard," and "attribute," appears on nearly every page of discussion of the medical concepts of others, the word "believe" does not appear in the final fourteen pages of the book. Here the word "knowledge," and cognates "recognize," "realize," "acknowledge," and "awareness," are used to describe Western medicine. Rivers could not have more clearly stated his judgment.
This juxtaposition of what others believe to what we know is not only true of intellectualist writers such as Tylor, Frazer, and Rivers. Close reading of the Evans-Pritchard text shows that he uses "belief" and its cognates to far greater analytic advantage than his predecessors, focusing on the coherence of a set of ideas. "All their beliefs hang together," he writes (1937: 194), "and were a Zande to give up faith in witch-doctorhood he would have to surrender equally his belief in witchcraft and oracles." The study of folk "logics" is an important part of the repertoire of cultural analysis, and Evans-Pritchard was a master of this genre. Nonetheless, his analysis framed culture as beliefs, and these were juxtaposed to knowledge - grossly in the introduction of the book, then in a subtle and nuanced way throughout this classic text.
The subtle or explicit representation of belief and knowledge as disjunct continues to be found in anthropological writing up to the present time. It is most explicit in rationalist writing and subsequent discussions of relativism. A final example from Dan Sperber's book On Anthropological Knowledge (19X5), which proposes to "outline an epistemology of anthropology" (p. 7), will illustrate, I he central chapter in the book is entitled "Apparently Irrational beliefs." It begins with an extract from Sperber's field diary during his research in Ethiopia, when an old man, Filate, comes in a state of great excitement to tell Sperber that he has learned of a dragon - "Its heart is made of gold, it has one horn on the nape of its neck. It is golden all over, It does not live far, two days' walk at most ..." - and asks him if he will kill it. Since Sperber had respect and affection for old Filate, and since Filate was too poor to drink and was
not senile, Sperber was left to puzzle how a person could actually believe in dragons ^ j about how to reconcile his respect for FlLte with "the knowledge that such a belie): j absurd."
Sperber's analysis of this problem K\K]S him directly to the usual arguments al\Hir the nature of rationality. How are we as thropologists to interpret cultural beliefs ke they about dragons or the role of witchcraft jQ causing illness - that are "apparently jr_ rational," that is, not in accord with how we know the empirical world to be? Are Sijch beliefs to be taken as literal or "symbol^."-.. If they represent literal claims about me nature of the empirical world, why have &\jCU systems not given way in the face of empirjcai experience? In Evans-Pritchard's words, \vhv do the Azande practitioners not "percu;ve the futility of their magic" (1937: 475)? \lK\ what is the alternative? A strong relative claim that the Azande world and ours are incommensurable, that so different are thev that we cannot translate between our World and theirs? Sperber follows through th.ese arguments; he ridicules the view that c]ie mind "actively creat|es] its universe" (Dous>l;ls 1975: xviii), as deriving from a "hermeneurj_ co-psychedelic subculture" (Sperber 1 ^85: 38), and develops a detailed analysis of different types of propositional beliefs. In the end, he concludes that old Filate's belief was Only "semipropositional" and was "not factual," that is, that it was not a kind of belief intended to really represent the way the world is and not clear enough to be stated in propositional terms that could be falsifiable. Thus his solution is that the old man really didn't believe in the dragon after all, that it was Only a kind of fantasy to entertain himself and ultimately the anthropologist.
My intent is not to join the rationality debate and the technical issues it raises here, although these questions serve as the stimulus for many of the concerns of this book, nor to speculate on old Filate's motives. Here my intention is to raise meta-level questions about the role of "belief" in anthropology. How does it happen that the "apparently irrational beliefs" provide the paradigmatic problem for a  central  tradition  in  anthropology? Any
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
73
human science, historical or anthropological, must deal with problems of translation, of differing world views and understandings of reality, of course. But how does it happen that "irrational beliefs" becomes the central, paradigmatic issue?
Surprisingly, there seems to be little analysis of the history of the concept "belief" in anthropology. It is constantly employed, a kind of Wittgensteinian "odd job word," but often used with little self-consciousness. The word almost never appears in indexes, even when it is employed throughout a text, and thus its use is not easy to trace. It is beyond the scope of this discussion to attempt such a history, but a brief review of anthropological texts suggests several hypotheses.
First, the juxtaposition of "belief" and "knowledge" and the use of "belief" to denote (or at least connote) counter-factual assertions has a long history in both anthropology and philosophy. This is contrary to what might be expected for both disciplines -for anthropology, because our primary goal has been to make understandable other societies in a non-judgmental way; for philosophy, because much of modern epistemology is designed to investigate the grounds for true belief.
Second, belief as an analytic category in anthropology appears to be most closely associated with religion and with discussions of the so-called folk sciences. "Belief" is most closely associated, that is, with cultural accounts either of the unknowable or of mistaken understandings of the "natural world," where science can distinguish knowledge from belief. In medical anthropology, analysis of "beliefs" is most prominent in cultural accounts of those conditions (such as infectious diseases) for which biological theories have greatest authority, and least prominent for those forms of illness (for example psychoparhology) for which biological explanations are most open to challenge.
Third, the term belief, though present throughout anthropological writing, appears with quite varied frequency and analytic meaning in different theoretical paradigms. For example, it seems far less central in American anthropology, with its background in nineteenth-century German historicist theorizing,
than in British social anthropology, in particular in the rationality literature.
Fourth, the representation of others' culture as "beliefs" authorizes the position and knowledge of the anthropological observer. Though differing in content, anthropological characterizations of others' beliefs played a similar role in validating the position of the anthropologist as the description of native religious beliefs did for missionaries. However, the rising concern about the position of the anthropologist vis-avis members of the societies he or she studies has produced a "crisis" in ethnographic writing (Marcus and Fischer 1 986: 8) and a generalized epistemological hypochondria, and this change in the relationship of anthropologist to the "Other" can be traced in the increasingly self-conscious and ironic uses of the term "belief."
Fifth, despite such post-modern hypochondria in some regions of the contemporary social sciences, the term "belief" and its counterparts continue to be important odd job words nor only in the cognitive sciences, where culture is closely linked with states of the mind, but in fields such as the medical social sciences, where the conflict between historicist interpretations and the claims of the natural sciences is most intense. Examination of the concept thus has special relevance for medical anthropology.
These are rather crude hypotheses. Flow-ever, they reflect my conviction that it was fateful for anthropology when belief emerged as a central category for the analysis of culture. This formation of anthropological discourse was linked to the philosophical climate within which anthropology emerged, a climate in which empiricist theories and sharp conflicts between the natural sciences and religion were prominent. It was also rooted in anthropologists' traditional relations to those they studied, framed by the superiority of European and American science and industrial development and by the colonialist context of research. Given the semantics of the term, that is the meaning "belief" had taken on by the late nineteenth century and continues to have in the twentieth century, the analysis of culture as belief thus both reflected and helped reproduce an underlying epistemology and a prevailing structure of power relations.
74
BYRON J. GOOD
A Shaking of the Foundations
Anthropology's greatest contribution to twentieth-century  sociology  of  knowledge has been the insistence that human knowledge is culturally shaped and constituted in relation to distinctive forms of life and social organization. In medical anthropology, this historicist vision runs headlong into the powerful realist claims  of  modern  biology. Enlightenment convictions about the advance of medical knowledge run deep, and although faith in medical institutions has given way to some extent, medicine is a domain in which "a salva-tional view of science" (Geertz 1988: 146; cf. Midgley 1992) still has great force. No wonder that discussions of "the problem of irrational beliefs" so often cite medical examples.
Nonetheless, the foundations for a comparative, cross-cultural study of illness, healing and medical knowledge which is based in the empiricist paradigm have been profoundly shaken in recent years. Geertz concludes his chapter on Evans-Pritchard in Works and Lives (1988), noting that the confidence that shines through Evans-Pritchard's writing, as well as through   Levi-Strauss's   Tristes Tropiques
language and empirical reality now seem dated. Rationality and relativism no longer neatly divide the field. Increasingly, social scientists and philosophers have joined in investigating how language activities and social practices actively contribute to the construction of scientific knowledge (e.g., Latour 1987). [...]
Second, the normative dimensions of the empiricist paradigm seem increasingly unacceptable. It is nor that any of us doubt that the biological sciences have made astounding advances in understanding human physiology* but we are no longer prepared to view the history   of  medicine   as   a straightforward recording of the continuous discoveiy of the facts of nature. Given the rapidity of change of scientific knowledge, as well as subaltern and feminist critiques of science and its authority, claims to "facticity" have been seriously undermined. The role of science as arbiter between knowledge and belief is thus placed into question. Critical analysis has replaced celebration as the idiom of the history and sociology of science, l-.-l
Third, the place of the ethnographer & objective, scientific observer - both in research and in ethnographic texts - seems less and
(1955), is simply not available to ethnograph-    less available to us today. Evans-Pritcrutrd
ers today. Our relationships with those we study have changed profoundly, and our confidence in our own view of reality, even in the claims of the natural sciences to simply represent the empirical world, has been seriously undermined. This change is represented by increasingly ironic reflections on terms such as "rationality" and "belief" in anthropology, feminist studies, and the sociology of science, and by the proliferation of new approaches in medical anthropology.
Several aspects of the empiricist paradigm relevant to comparative medical studies have become especially problematic, pushing our field in new directions. First, positivist approaches to epistemology and the empiricist theory of language have come under sustained criticism in philosophy, the history and sociology of science, and anthropology. Whichever authors one invokes - Thomas Kuhn, Michel Foucault, Paul Feyerabend, Hilary Putnam, Richard Rorty, or a generation that grew up with these figures -older theories of the  relationship between
could assume such a position in his writing* on the Azande only by ignoring his own relation to the colonial authorities, Favret-Saadu (1980: 10) suggests that even Evans-Pritchard* while conducting field research, could situate himself outside of Zande witchcraft discourse--beyond possible charges of being a witch himself, for example - only because the Azande granted him the title "Prince without portfolio," which served as a kind of exemption from the claims of the discourse and thus protected him. Whatever the case for Evans-Pritchard  and  witchcraft,  the  position t>t today's anthropologist is increasingly contested. [... 1 In medical anthropology, arbitrating between belief and knowledge suggest:^ positioning ourselves within what Favret-Saati a calls "the official theories of misfortune- ^ backed as they are by powerful social agencies Finding a stance both as researcher and in ethnographic text is thus increasingly difficult The position implied by the language of belief j often untenable.
MEDICAL ANTHROPOLOGY AND THE PROBLEM OF BELIEF
75
Finally, a variety of more technical analyses of belief suggests problems with the empiricist program, challenging the utility of "belief" as an analytic category, even questioning the existence in other societies of "beliefs" in our sense of the word (see Stich 1983; cf. looker 1992; Hahn 1 973). A view of culture as propositional, men-talistic, voluntaristic, and individualistic - for example, of medical beliefs as rational propositions about the world, held in the minds (or brains) of individuals, and subject to voluntary control - is an elaboration of a particular folk psychology; such a view reproduces an ideology of individualism that matches poorly with much of what we know about the real world.
1 ... | Thus, despite powerful authorization by biomedicine and the biological sciences, the empiricist program in medical anthropology is deeply problematic j ... | How we situate ourselves in relation to the underlying theoretical issues at stake here is extremely important for how we conceive a program for medical anthropology. Flow we situate our research in relation to biomedical categories and claims, the nature of authority we grant to biological and medical knowledge, the problems we see as central to the field, and the way we define the project in which we are engaged are all strongly influenced by our stance on these issues. Medical anthropology is one of the primary sites within anthropology where alternative responses to the confrontation between historicism and the natural sciences are being worked out.
[ . . . |   All   medicine   joins   rational and deeply irrational elements, combining an attention to the material body with a concern for the moral dimensions of sickness and suffering. In bis Marett Lecture  in 1950, Evans-Pritchard argued that "social anthropology    is    a    kind    of historiography" that "studies societies as moral systems ..." In all societies, even in the modern world with  overarching  moral   orders  no longer intact, serious illness leads men and women to confront moral dimensions of life. It is after all a central task of "the work of culture" (Obeyesekere   1990)   to   transform human misery into suffering, and to counter sickness with healing. Biomedicine, as other forms of healing, is of special interest because it combines the empirical or natural sciences with this  primal  task.  It  is  the  privilege and
the obligation of medical anthropology to bring renewed attention to human experience, to suffering, to meaning and interpretation, to the role of narratives and historicity, as well as to the role of social formations and institutions, as we explore a central aspect of what it means to be human across cultures.
NOTES
1 Key texts in this debate include Wilson (1970), Horton and Finnegan (1973), Hookway and Pertit (1978), Hollis and Lukes (1982), Leplin (1984), and Doyal and Harris (! 986). See also, for example, A. Rorty (1988), Sperber (1985), Shweder (1984), Taylor (1985b: 1 34-5 1) and Tambiah (1990).
2 1 am particularly grateful to Theresa O'Nell for her help in analyzing this text.
REFERENCES
Douglas, Mary
1975   Implicit Meanings: Essays in Anthropology. London: Routledge & Kegan Paul. >oyal, Len, and Roger Harris 1986   Empiricism, Explanation and Rationality. An Introduction to the Philosophy of the Social Sciences. London: Routledge and Kegan Paul, ingel, George L.
1977   The Need for a New Medical Model: A Challenge tor Biomedicine. Science 196: 129-36. Evans-Pritchard, E. E.
1937 Witchcraft, Oracles ami Magic among the Azande. Oxford: Oxford University Press.
1950 11962|   Theories of Primitive Religion (The  1950 Marett Lecture). In  Essays in Social Anthropology. London: Fsber and Faber. Favret-Saada,Jeanne
1980   Deadly Words. Witchcraft in (he Bocage. Cambridge: Cambridge University Press. Foucault, Michel
1970   The Order of Things. An Archaeology of the Human Sciences. New York: Random Flo use. Geertz, Clifford
1988 Works and Lives. The Anthropologist as Author. Stanford, CA: Stanford University Press.