Professor Garfinkel Visits the Soothsayers

Ethnomethodology and Mambila Divination

David Zeitlyn

University of Oxford


Garfinkel's techniques of ethnomethodological analysis permit a focus on 'moments of crisis' in dialogue.  It is at such moments that the 'negotiation of meaning' is clearest.  These ideas are applied to the analysis of Mambila 'spider' divination.  Only binary questions may be posed.  It is shown that in their interpretations to answers received, diviners redefine contradictions as calls to explore new possibilities.

Contradictions in discourse serve to give us 'pause for thought' and constitute question rejecting moves.  The article demonstrates how conversational analysis can be used as a powerful analytical tool of 'processual' anthropology.


I shall argue that a paper first published in Scher (1962) which does not mention divination at all is the most important contribution to the study of divination since Evans-Pritchard's Witchcraft Oracles and Magic among the Azande (1937). This paper is Garfinkel's 'Commonsense knowledge of social structures' (1984) .  I shall use some of Garfinkel's conclusions to explain how contradictory answers are understood to be providing additional information.  I hope to show that the means by which this is achieved is particularly relevant to understanding divination types which allow only yes/no answers.  For illustration, I shall describe one instance of divination from my fieldwork among Mambila in Cameroon.

Garfinkel intended the paper to demonstrate that the 'documentary method' is inescapable in both everyday and academic life.  Even the most rigorously statistical and formal sociological analysis will contain a lacuna between the evidence and the conclusions which are drawn from it.  Garfinkel identified the use of 'the documentary method' as a means to bridge that gap (see Heritage 1984: 159 sqq.).   Its use is now widely acknowledged in the humanities, although usually under rubrics other than the 'document' of Mannheim's coinage.  Anthropologists may be more familiar with concepts of hermeneutics, interpretation or, in Peirce's terminology, 'iconicity'.  Suffice to say that behaviour, whether verbal or not, is assumed to be connected to a hidden state of affairs.  This assumption enables us to orient ourselves in the world (and thereby affects our actions).  Such arguments are familiar when applied to intentional attitudes, but Garfinkel points out that they apply equally to the presumption of shared notions of social structure which allow actions to be coordinated.

Indeed the phenomenon of politeness (as analysed by Brown & Levinson 1978, and Strecker 1988) is an example

Page 655

of one of the means whereby such shared notions are maintained and their 'commonality' tested,

The analysis of politeness is an instance of the sort of analysis which I am advocating; however, I shall not labour that point here.  My main concern is to demonstrate the 'documentary method' at work in divination.  I shall begin by examining the examples which Garfinkel gives to illustrate the existence of the documentary method in everyday life.

Garfinkel's examination of 'Commonsense knowledge of social structures'.

Garfinkel's paper, presents an account of an experiment and an analysis of it, followed by his reflections on this analysis. The experiment involved American university students being counselled through an intercom. The students were told to pose questions in an either/or format.  They received yes/no answers.  The questions, and the students 'reflections upon the answers they received', were tape-recorded.  What the students were not told was that the yes/no responses were supplied in a predetermined random sequence. Each student received the same sequence of answers.  Garfinkel quotes two sessions in full and then summarizes the results in a set of aphoristic comments.  McHugh (1968) describes the same experiment and gives further examples.

I quote a short passage from the first case, in which a male Jewish student debates whether he should continue to date a non-Jew in the face of his family's displeasure:

'... My question is, do you feel under the present circumstances [which he has just explained] that I should continue or stop dating the girl?  Let me put that in a positive way.  Do you feel that I should continue dating this girl?

Experimenter: My answer is no.

Subject: No.  Well that is kind of interesting.  I kinda feel that there is really no great animosity between Dad and I but, well, perhaps he feels that greater dislike will grow out of this.  I suppose, or maybe it is easier for an outsider to see certain things that I am blind to at this moment. I would like to ask my second question now.

Experimenter: Okay

Subject: Do you feel that I should have a further discussion with Dad about this situation or not? Should I have further discussion with Dad over this subject about dating the Gentile girl?

Experimenter: My answer is yes.'

This brief extract suffices to show how the subjects constructed for themselves a meaningful dialogue. The subject has taken the random responses of the experimenter as answers to the questions put and has imputed an argument to the experimenter.  Subjects behave as if they were negotiating meaning in the manner found in ordinary conversation1. However, in this experiment, just as in divination, the responses are not intentionally produced answers but are 'mere events'.  It is this similarity which makes Garfinkel's discussion relevant to the analysis of divination.  My concern is with the process of constructing an argument, or of imputing meanings.

Although few anthropologists would wish to use such cold-blooded 'experimental' methods of research, Garfinkel's techniques can be usefully applied in analysing the praxis of divination2.  Eleven of his conclusions which are applicable to the analysis of divination are listed below (some are quotations, others have been rewritten in an attempt to render them more readily comprehensible)3.  The 'answerer' to whom I refer is not Garfinkel's counsellor but the divination.

Page 656

After Garfinkel: eleven principles of divinatory interaction  (from Garfinkel 1984: 90-91, sections D and E)

1) Questions may be retrospectively redefined in the light of later answers.

2) 'The identical utterance is capable of answering several different questions simultaneously and of constituting an answer to a compound question that in terms of the strict logic of propositions does not permit either a yes or no or a single yes or no'.

3) One response can be understood to be answering several previous questions simultaneously.

4) 'Present answers provide answers to further questions that will never be asked'.

5) 'Where answers are unsatisfying or incomplete, the questioners are willing to wait for later answers in order to decide the sense of previous ones'.

6) Incompleteness is attributed by questioners to deficiencies in the method (of yes/no answers), or to incomplete comprehension on the part of the answerer (i.e. the divination) which can in turn be attributed to poor question-setting.

7) Reasons are assumed to exist for 'inappropriate' answers.  Those reasons explain the answer which is given, and determine its sense.

8) 'When answers are incongruous or contradictory, subjects are able to continue by finding that the 'adviser' has learned more in the meantime,  or that he has decided to change his mind,  or that perhaps he is not sufficiently acquainted with the intricacies of the problem,  or the fault was in the question so that another phrasing is necessary'. In other words, explanations or excuses can always be constructed ad hoc.

9) 'Incongruous answers are resolved by imputing knowledge and intent to the adviser'.

10) Contradictions force reinterpretation of the questions: further meanings to the questions are imputed which explain the answers, thus removing the contradiction.

11) Contradictory answers lead to a 'review of the possible intent of the answer so as to rid the answer of contradiction or meaninglessness, and to rid the answerer of untrustworthiness'.

Points 7-11 are particularly relevant to the analysis of contradiction in divination which follows. Garfinkel also provides some maxims concerning the questioners' suspicions of the system.

12) The possibility of random answers may be considered by the subjects but is not tested.  Suspicions are allayed if the answers 'make good sense'.

13) Suspicion turns the 'answers' into 'mere events' and there is then no point in continuing. Therefore:

14) Those who became suspicious are unwilling to continue.

During my fieldwork with Mambila in Cameroon, about which more below, I never encountered such express doubts about divination.  I suggest that this is due to the intellectual protection given to the basic assumptions,  a suggestion in accord with the argument presented by Evans-Pritchard (1937).  These points are therefore complementary to Evans-Pritchard's twenty-two reasons why the Zande do not perceive the futility of their magic (1937:475-478).  These are a catalogue of ad-hoc hypotheses used

Page 657

to protect the validity of the central tenets, in this case underwriting the validity of divination itself, whereas the results of any individual divinatory session remain open to question .  The creation and interpretation of a sequence of questions and answers is independent of any actions employed to protect divination, such as those listed by Evans-Pritchard.  There is neither methodological nor logical contradiction between Evans- Pritchard's analysis, and that of Garfinkel.

Contradiction and reinterpretation

The second case used by Garfinkel to illustrate his argument includes an example of reinterpretation; remarkably he does not comment upon this.  The 'experimenter' (Garfinkel's term) should give yes/no answers to the questions put.  A student of physics is considering whether to leave college or to change his subject.  His first question is whether he should change subjects (answer: 'No').  Eight questions later he asks: 'Will I get a degree?'.  Answer: 'No'.  I quote the questioner's response to this in full:

'According to that I won't get a degree.

What should I do?

Are you still there?'

Experimenter: 'Yes, I am.'  (p 87)

Consider the likely reactions of the subject if the experimenter had replied 'My answer is no'.  It is clear that the experimenter has answered only the second of those two questions, and that he has stepped out of his 'experimental' framework.  It is important to understand why the crisis of faith which occasioned this exchange occurred when it did. When the question 'Will I get a degree?' was answered in the negative the subject perceived the experimenter to be contradicting himself.  By his negative answer, the experimenter was understood to be committing himself to the proposition

A - You will not get a degree in physics.

This was perceived to contradict an earlier answer wherein the experimenter was understood to commit himself to the proposition

B - You should not change your subject.

If one assumes that

C - If you will not get a degree in physics then you should change your subject then B entails

D - You will get a degree in physics [if you persevere]

But from above

A - You will not get a degree in physics.

By asserting B the subject is allowed (granted C) to infer D.  But A and D are contradictories.

To invoke a Gricean relevance principle: if the questioner is not going to get a degree then it is irrelevant what subject he studies.  The 'relevant' answer to the question 'Should I change my subject?' would be: 'You will not get a degree'. This would constitute a rejection of the question.  However, the schema of the dialogue explicitly permitted only yes/no answers.  In order to make sense of the answers received, the subject interpreted contradiction as a rejection of the question. This is a common conversational ploy.  For example:

Shall we go to the pub?

Have you looked in your diary?

Page 658

A question is rejected by answering it by another question.  In order to do this the subject had to disregard the rules which had been presented.  The fact that only yes/no answers to polar questions were possible was ignored, or perhaps subtly reinterpreted.  The subject is now in the position of saying 'When I get answers like this they don't really mean yes and no, they mean something altogether more complicated...'

Conversational analysis provides the means to examine in detail the way in which meaning is negotiated between speakers. Such negotiation is particularly apparent where misunderstanding or disagreement occurs.  We can draw on the literature on 'repairs4' for examples of the negotiation of meaning. Conversational analysis can reveal instances in divinatory discourse where, in order to make sense of an utterance, context is strongly implicated5.  It is clearly necessary to consider this context in order to understand the utterance.  The perspective of the analysis is perforce widened to include not only divination but also the circumstances of the participants and the social structure within which the divination is practised.

An ethnomethodological focus on the negotiation of meaning between speakers is of more assistance than formal logic in understanding the manoeuvres adopted, for example, in response to contradiction. Formal logic identifies a contradiction but allows no other solution than the rejection of a premiss.  It cannot, however, suggest which premiss is at fault. Ethnomethodology, on the other hand, identifies the redefinition of a premiss as a constructive solution to the problems caused by contradiction.  The empirical techniques of conversational analysis allow for the identification of the premiss at issue, and for the study of the process of its redefinition.  This seems more consistent with the questioners' responses as evidenced in the data.In the cases of both divination and Garfinkel's experiment, analysis of the negotiation of meaning is facilitated since in both, one of the parties to the perceived conversation is not in fact negotiating. Paradoxically, divination can be seen to say so much precisely because it is mute.

Studying divination

Many reasons may be given to warrant  the study of divination.  I shall briefly present two of these, and go on to distinguish between two major classes of divination.  A consideration of the limitations of classical sociological analysis of divinatory practice will serve as a prelude to my presentation of the Mambila material.

Evans-Pritchard sought to convince European readers that the Azande were rational to persist in their beliefs, and that their actions were therefore subject to rational explanation.  In particular, he expressly addressed himself to 'political officers, doctors and missionaries in Zandeland, and later to Azandethemselves'.  Divination provided an excellent subject with which to challenge colonial prejudices.  More recently, divination has featured as a leitmotif in what has become known as the 'rationality debate'.  Of its contributors however, only Beattie (1964, 1966 and 1967) and Horton (1967 and elsewhere) have published works about divination per se.  Divination has figured so importantly because it is perceived as a paradigm for 'rationality in irrationality': that is, belief in divination is held to be irrational, but its practice is extremely rational according to the ethnographies (especially Evans-Pritchard 1937).  It  therefore serves as an amenable synecdoche of religious belief and practice.

A second reason for studying divination is that it reveals the actors' understanding of their social structure.  Evidence about indigenous models of the world is provided

Page 659

by the process of posing the questions, by their phrasing, and by the range of possible solutions proposed. Divination thus ranks alongside dispute as a social activity whose study can provide information about, and understanding of, matters of much broader concern than the stated topic of analysis.  Thus, although divination may occupy relatively little of the attention of a group, as with the Azande, it may  provide a rewarding starting-point from which to launch a wider analysis.

Classes of Divination

Classifications of different types of divination have been produced by almost everyone who has written about this subject. Some typologies cover more comprehensively than others the wide range of activities which may be glossed as divination. Whether they are sociologically revealing is another matter.  Without entering into the arguments here I shall follow Cicero (who attributes the idea to Quintus) in making a distinction between 'artificial' and 'natural' divination (De Div. Later authors prefer the terms 'mechanical' and 'emotive' (see Devisch 1985, Vernant 1974 and Zeitlyn 1987 for further discussion)6.

Natural, or emotive divination (sometimes also glossed as 'aleatory') depends on the recognition of a direct relationship between the operator and some occult force or spirit, such that truth is achieved through contact with spirits or by exercise of the 'intuition'.  It typically involves some sort of 'possession' (this is treated in more depth by Lewis 1971). By contrast,  artificial divination aims to reveal truth through the performance of a variety of technical operations, all of which are mechanical in nature.  The divination practices used by Mambila are exclusively of this kind, and are the subject of my discussion here7.  Unlike emotive divination,  mechanical divination appears to involve much clear ratiocination, and its results are open to question in quite different ways.  Although practitioners of any type of divination can be accused of deceit and fraud, only mechanical divination can be performed 'incorrectly',  thus allowing for the possibility of mistaken practice.  For in the case of emotive divination, the truth of the divinatory results is guaranteed by the possessed state of the diviner.  Since possession is an unequivocal state, mistaken practice is impossible8. Any divinatory techniques associated with possession are employed simply as preliminaries necessary to attain this state; they are not means by which the results are obtained9.

The incompleteness of sociological analysis

Sociological analyses of divination which have treated it as a procedure either for legitimating decisions (Park 1963) or for providing therapeutic benefit to the consultants (Beattie 1964) may sufice to explain emotive divination.  However their focus is on the social consequences of the use of divination rather than on the divination per se.  They do not, therefore, consider the possibility, admitted only in mechanical divination, of mistaken practice.  While such analyses may reveal important aspects of a divinatory system, the theoretical standpoint adopted allows for no detailed analysis of the praxis of consultation.  Neither the interaction between diviner and client, nor, more importantly, the interaction between diviner and divination can be understood from this perspective. Conversational analysis, however, provides techniques to understand these interactions.

An analysis of this kind neither precludes nor invites functionalist arguments, rather it precedes and anticipates them. This is possible since, as I have suggested, conventional theories have little or nothing to say about the details of divinatory practice. The results

Page 660

of our analysis  of these details could be subjected to further interpretation in terms of any theory of society, for example, in a functionalist, or a Marxist or a dramaturgical, for the analysis itself is independent of whatever theory adopted.  Thus, I regard as a neutral boon that which Gellner criticises as sociological myopia in ethnomethodology (Gellner 1975).

Why study divination as situated dialogue?  

Ethnomethodology provides productive techniques with which to study divination.  Both the ratiocination involved in producing an answer and the contextualisation of question and answer may be examined in the detail they deserve. Moreover, by analysing the process of divination we can avoid the reliance on abstract accounts which has previously limited descriptions, for example, of the Ifa divination practised by Nigerian Yoruba and neighbouring groups10. Garfinkel lists some of the ways by which (objectively random) utterances are endowed with meaning by listeners so as to construct a sensible dialogue.  The study of divination reveals how a similar process occurs when the participants pose questions and receive answers.

Mambila Divination

Divination systems allowing only yes/no answers involve similar processes of interpretation, particularly when the divination apparently contradicts itself.  Following a summary introduction to Mambila society I shall illustrate this point with an account of Mambila spider divination (further details may be found in Zeitlyn 1990).

The Mambila lie on either side of the Nigeria/Cameroon border, the bulk of them living on the Mambila Plateau in Nigeria. A smaller number (c. 12,000) are to be found in Cameroon, especially at the foot of of Mambila Plateau escarpment, on the Tikar Plain.  My fieldwork was restricted to these latter groups, and in particular to the village of Somié.  Somié had a population of approximately one thousand  (based on the official 1986 tax census) at the time of fieldwork.  Self-sufficient in food, the villagers have grown coffee as a cash crop since the early 1960s.

Cameroonian Mambila on the Tikar Plain have adopted the Tikar institution of the chiefship, yet their social structure otherwise closely resembles that described for the Nigerian village of Warwar given by Rehfisch (1972) based on fieldwork in 1953.  Nigerian Mambila did not have the same type of institutionalized chiefship as is found in Cameroon. In Nigeria villages were organised on gerontocratic principles, and largely lacked political offices.  The system of exchange marriage described by Rehfisch (1960) has now vanished, and with it the two sorts of named group which recruited through different combinations of descent, marriage type (exchange or bridewealth) and residence.  Most people in the village are members of either the Catholic or Protestant church.  However both men's and women's masquerades are still performed, and cases heard at the Chief's palace are regularly concluded with a ritual oath (sua, see below).

Most married men know how to divine, but have varying degrees of confidence in their own skills. Hence if a problem is serious, it is likely to be taken to one of the acknowledged experts.  In the case considered below a man, named Wong, in his late thirties went, by arrangement, to divine with Bi, the head of Njerup hamlet, and an important elder.  Bi is one of the two elders who choose the new chief.  He is also well known as an accomplished diviner.  In order to be confident of the results of divination

Page 661

Wong came to Bi in order to have Bi's sanction.  Bi could correct any mistakes of interpretation, and thereby help ensure the accuracy of the results.

Mambila spider divination11 entails the posing of questions couched as binary alternatives, often requiring yes/no answers.

A hole in the ground inhabited by a spider is covered by an enclosure, usually an inverted pot.  A stick and a stone are placed within this enclosure, near the entrance to the spider's hole.  A set of marked leaves is placed over the entrance to the burrow.  When questions are posed the pot is tapped; in response to the knocking the spider emerges from its hole. In doing so it disturbs the leaves.  The resulting pattern of the leaves in relation to the stick and to the stone is interpreted as an answer to a question.  Questions allow one of two responses, one is explicitly associated with the stick and the other with the stone.

Several different spiders may be consulted simultaneously.  This enables a faster rate of questioning since some twenty minutes elapse before the diviner can check whether the spider has responded to a question. It also allows a consistency check to be made by asking the same question of different spiders.  Diviners admit that ambiguous or unintelligible answers are possible but few such instances were observed.

The following table shows the response to contradiction which arose during a six-hour divination session recorded in January 1987.  The divination concerned a child (Wong's daughter) suffering from malaria.  The main points at issue were whether the illness had been caused by witchcraft, and whether the taking of a sua-oath12 would be sufficient to protect the child from further attack.  The table shows the questions addressed to two different spiders, and the answers received.  In each case the alternative said by the diviners to have been chosen has been marked with an asterisk.  Forty-two questions were posed during the session; their order is indicated by the question numbers.

TABLE 1 Divination Questions

1 The kare  rite is sometimes referred to as a variety of sua.  It is a domestic version of the sua-oath.

2 'Something buried' refers to some witchcraft treatment, which unless detected and removed would continue to act although its perpetrator might be caught by su a

Page 662

Responses to contradiction:

Table 1 contains answers which directly contradict one another.  The acceptance of direct contradiction is, according to the canons of traditional logic, a symptom of 'illogicality'.13  Further comment is warranted.

The sequence starts with Question 33, which was addressed to Pot 1: will sua end the problem or not?

A straightforward yes/no response was sought.  Another pot (Pot 2) was asked a similar question (Q 34) before the response was obtained from Q 33.  The response to Q 34 was taken to advocate the use of sua, as opposed to other sorts of treatments.  This was immediately followed by Q 35 which repeated Q 33.  The response to Q 35 was that sua would not end the problem.  This is a flat contradiction to Q 34.  The problem was then assumed to be one of witchcraft.  In response to this Q36 sought to identify the sex of the witch. The divination was taken to have identified the witch as female.  The response to Q 36 was understood to be identical with an earlier diagnosis of 'problems among the women in Wong's house' (Q 26 and Q 29).  There had in fact been a long-standing quarrel between Wong's wife and his classificatory sister about the usufruct of a field.  Disputes over land tenure are typical of cases in which sua-oaths are taken.  The parties to the dispute swear that, although they may be quarrelling, they bear no malice and will not seek to win their case through witchcraft, for example by causing illness among the children of the other litigant.  The identification of the female witches as being those women embroiled in that dispute was assumed, and therefore not tested further by divination.  This resolved the dilemma posed by the contradiction.  Once the diviners are assured that the witchcraft referred to is only that connected with this quarrel, then sua becomes  an appropriate and sufficient course of action.  When that reassurance has been given they can return to the previous line of questioning.  The earlier question was then repeated in a modified form: will sua end it, or is there other witchcraft to be dealt with, for example: in the form of buried treatments which remain active until discovered and destroyed (Q 37)?  After putting this question the response to Q 33 was sought by inspecting the pot.  The answer found was 'sua will not solve the problem'.  This was immediately pursued in the light of the question which had just been put  (Q 37),  as to whether any witchcraft was to be dealt with.  The diviners understood this to indicate that there might be further witchcraft.  Hence Q 38 draws the distinction between buried witchcraft substances and the ending of the affair by sua.  The responses to both Q 37 and Q 38 indicated that performance of the sua rite was the appropriate action to be taken.  Thus a believable, because consistent, result was obtained.

The contradictory results preceding this were thenceforth ignored.  They had, however, forced the diviners to examine the possibilities of more complicated problems.  Once these possibilities had been eliminated the diviners could return to the main strand of the enquiry as if no contradiction had occurred.

We must take seriously the diviners' assumption that the sequence of questions is a dialogue between divination and diviner.  Mambila diviners talk of asking (bie ) questions of divination (nggam ) as if it were a single entity.  Looking at a result they say 'Divination says..' (Nggam je...).  I was always given  inductive and empirical justifications for the veracity of divination.  Even formal initiation into the technique of spider divination did not contain any instruction about the origin of the knowledge discovered through divination, nor was any account given of what divination was, or

Page 663

how it worked.  The truth of the belief in the efficacy of the technique is held to be warranted by the success of the diviners.

If the process of divination is, in part, to be regarded as a dialogue between diviner and divination, then contradiction may be regarded as a rhetorical device used by the divination to make the diviner cast the net of his questions more widely.  In the example illustrated, the divination has forced the diviners to consider the possibility, previously not addressed by them, that buried witchcraft substances may be responsible for the child's illness.  Garfinkel's methods may be used to reveal the way in which diviners construct the dialogue. In essence: contradictions were understood as question-rejecting moves. They give pause for thought, and lead to changes in tack.  If we return to Garfinkel's maxims listed above, numbers 7 to 11 may be summarized in a single maxim:  'the problem of contradiction may be defused by treating it as a rejection of the question'. Evidence for the validity of this position comes from our success at reconstructing the observed dialogue, which otherwise remains obscure.

It may be objected that the contradictions are only there from our analytical point of view.  However, 'crises of faith' or 'changes in tack', or other breaks in the flow of dialogue, occur at or following the points where we identify a contradiction.  This demonstrates that they are more than etic constructions.  From an emic point of view no contradiction may be perceived, but it is clear that the actors have been given 'pause for thought'.  I would suggest that since conversation assumes14 the absence of contradiction, speakers then strive to preserve the conversation by removing the contradictions.

A cynical account of this divination would be that since the performance of sua is the standard response to many problems, it is only to be expected that a sua-oath will be taken when there is concern about an illness.  The process of divination would then be seen as an empty validating act whose outcome is known in advance.  According to such a view divination resembles a game of 'Twenty Questions', except that play continues until the desired result is obtained15.  The form of the questioning and answers is identical.  But in divination all the participants know the background and the likely results.  Although I am sure that the participants would have admitted that sua was a likely result, I nonetheless reject such an approach.  It allows no room for analysis of the actions and, most particularly, the ratiocination of the diviners is not considered. A similar objection applies to those analyses which see divination as a means to increasing psychological comfort by reducing stress (Park 1963).  That the actors believe in what they are doing is clear from the attitudes expressed, and from the manner in which divination is practised.  The analyst has a responsibility to be faithful to their beliefs.

Contradictions and inference merit consideration which is not possible with conventional sociological analysis.  Both chains of reasoning and the consideration of hypothetical possibilities are involved, and these are capable of reconstruction, as I have attempted to show above.  That some outcomes are highly probable may be regarded as a measure of the predictability of the world.  The fact that time-worn techniques are repeated does not mean that they are not chosen with care and deliberation each time they are adopted.

Ethnomethodology and the techniques which have grown out of it, such as discourse or conversational analysis, provide means by which the care and deliberation exercised in making choices may be brought to light and analysed.  The propositions by Shaw (1985), Parkin (1979) and Werbner (1989) that divination is best analysed as dialogue,

Page 664

can thus be realized.  In so doing we may be able to unpack what Devisch (1985:77)calls the 'purposeful articulation of meaning'. This seems to be what is meant by a praxeological approach (see also Jackson 1978).


In conclusion I should like to note two consequences of my argument.  First, it opens up new areas of data which permit a re-evaluation of existing theoretical positions.  Thompson's re-appraisal of ideology is an instance of this (1984, especially Ch.3).   Discourse analysis provides a new route from which to approach the topic of ideology and to unpack Habermas's notion of an idealised, neutral conversation between equals which acts as a yardstick with which to measure reality.

Secondly, new theories need to be developed in order to comprehend the results of conversational analysis.  As I have argued above, conventional anthropological theories are not sufficiently fine-grained to have anything to say about the details of conversational interaction.  Brown and Levinson (1978) have provided a model of politeness which relates detailed linguistic strategies to sociological factors connecting the conversants.  Thus the domains of politics and sociolinguistics are linked in a manner hitherto inaccessible to analysis.  Strecker (1988) has extended this work by reminding us that any theory of politeness is ipso facto a theory of rudeness, thereby the theory applicable to new ranges of data.  It is this sort of theory, I contend, that is needed if we are to be able, as anthropologists, to examine the tensions and dialectic between normative accounts of what should happen and what actually goes on in the world. In short, I claim that  ethnomethodology enables us to move beyond mere announcements of the desirability of processual anthropology to its actual practice.


Drafts of this paper have been presented at Departmental Seminars at the Universities of St Andrews and Edinburgh, as well as at the Centre of West African Studies, University of Birmingham.  I am grateful to those institutions for their invitations, and the discussion which they provided.  In particular I thank Ladislav Holy, who helped me to clarify several of my ideas.  Drafts have been read by Jeremy MacClancy and Anna Rayne, the latter has also helped immeasurably in improving the clarity of the final version.

The research on which this study is based was funded by the E.S.R.C. (grant no. A00428424416) and by a scholarship from Trinity College, Cambridge.  My research in Cameroon could not have been conducted without the research permits granted by His Excellency the Minister for Higher Education and Scientific Research (R.P. 13/85 and 62/86), and the help provided by his staff.

1There is a telling similarity between this and the manner in which gamblers interact with croupiers and understand roulette as a game of skill rather than a game of chance (Oldman 1974).

2This is quite apart from the reflection that psychiatry and other therapies could be regarded as the major type of divination used in America.

3It would be interesting to attempt a sociological explanation of the opacity of the prose of those working in ethnomethodology.  Sadly, this seems to have contributed to the sidelining of the approach, at least within anthropology. Nor has the prose style improved with the years.  A recent paper by Garfinkel and some colleagues is a masterpiece of recondite impenetrability, a  fact to which the comments made by Holton, who had the unenviable job of commenting on the paper when it was presented, mutely attest (Garfinkel et al 1981, Holton 1981).

4A 'repair' occurs after normal turn-taking rules of dialogue break down.  This is often caused by the speakers having different assessments of 'what is being said.'  Achieving sufficient consensus to continue the dialogue then necessitates negotiation of meaning (see Levinson 1983:339ff; Schegloff, Jefferson & Sacks 1977).

5The process of such implicature was first outlined by Grice.  There remains uncertainty as to its detailed workings (see the discussion in Levinson 1983 Ch3).  Sperber and Wilson (1986) have produced

Page 665

a general theory of relevance which they claim can account for such  implicatures.  This claim remains controversial (see Sanders 1988,  Levinson 1989).

6Confusion may be caused by inconsistencies of usage between different authors.  Evans-Pritchard (1937:10-11) contrasts oracles with divination in  order to express the distinction made here between mechanical (or technical) divination and emotive divination.

7An anti-witchcraft cult called 'Makka' existed around the time of the Second World War.  Some practitioners of 'Makka' appear to have been possessed when detecting witches, although others relied on the administration of ordeals.

8A Sudanese counterexample has been pointed out to me in which possession at the wrong time of day leads to wrong predictions which can be attributed to a mistake about the time.  Nevertheless, I believe that the broad contrast drawn here is helpful in the analysis and comparison of different divinatory systems.

9While focusing on ratiocination I do not intend to imply that mechanical divination is the sole means to the end of truth-seeking (Esther Goody pers. com).  I should also stress that there are many kinds of mechanical divination which do not pose polar questions, and which have answers in other than yes/no forms.

10I know of no work on Ifa or on Ifa-type divinations which devotes much attention to the details of praxis.  Maupoil (1943) concentrates on the mathematics, Bascom (1969) on the verses.  Only de Surgy (1981) includes some case material, albeit principally as an introduction to the sacrificial sequences which follow.  It is surely the case that sufficient is now known about both the ese  verses and the cult of Ifa.  But what actually happens, the way in which the documented theory is put into practice, remains unstudied.

11A discussion of the comparative literature on spider divination, together with a detailed account of Mambila divination and the rules and processes of its interpretation, are found in Zeitlyn (1987).

12Sua  names a variety of ritual oaths, often accompanied by the ritual killing of a chicken as well as some masquerades.  It is the nexus of Mambila religion and has been extensively analysed in Zeitlyn (1990). It is sufficient here to note that it is both an oath which binds the oath-takers not to cause illness, and a death threat to any other persons seeking to do evil.

13A possible response is to abandon standard logics and to use some of the variants (Haack 1978). Their use has been advocated as a solution to long-standing anthropological problems (e.g. Salmon 1978, Evens 1983).  This, however, is a council of despair.  Despite not having explored all the possibilities (Zeitlyn 1983), the adoption of non-standard logics raises as many problems as it (purportedly) solves.  Even in quantum mechanics, where its use was proposed by Reichenbach as long ago as 1944 (Reichenbach 1944), it has not succeeded in solving the philosophical problems (see the discussion in Jammer 1974, and the comprehensive bibliography therein).

14This is the essence of Grice's co-operative maxim.

15This analogy was first suggested by Ernest Gellner.


Bascom, W. R. 1969. Ifa Divination: Communication between Gods & Men in West Africa. London:
Indiana University Press.
Beattie, J. H. M. 1964. Divination in Bunyor, Uganda. Sociologious 14: 44-61.
___. 1966. Consulting a Diviner in Bunyoro: A text. Ethnology 5.2: 202-217.
___. 1967. Consulting a Nyoro Diviner: the Ethnologist as Client. Ethnology 6: 57-65.
Binsbergen, W. van and Schoffeleers, M. (Eds.) 1985. Theoretical Explorations in African Religion,
London: Routledge & Kegan Paul.
Brown, P. B. and S. C. Levinson. 1978. Universals in Language Use: Politeness Phenomena. In E.
N. Goody. 1987
Cicero, Q.  1923 De Senectute, De Amicitia, De Divinatione (Loeb Classical Library). Trans. Falconer,
W.A. London: Heineman
Devisch, R. 1985. Perspectives on Divination in Contemporary Sub-Saharan Africa. In W. van
Binsbergen and M. Schoffeleers. 1985.
Evans-Pritchard, E. 1937. Witchcraft Oracles and Magic among the Azande. Oxford: Oxford University
Evens, T. M. S. 1983.  Mind, Logic and the Efficacy of the Nuer Incest Prohibition.  Man (n.s.)  18:
Garfinkel, H. 1984 (1967). Commonsense knowledge of social structures: the documentary method of
interpretion in lay and professional fact finding. In Studies in Ethnomethodology, pp 76-103. Cambridge:
Polity Press.

Page 666

Garfinkel, H., M. Lynch and E. Livingston. 1981. The work of a discovering science construed
with materials from the optically discovered pulsar. Philosophy of the Social Sciences 11(2): 131-158.
Gellner, E. 1975. Ethnomethodology: The Re-enchantment Industry or the Californian Way of
Subjectivity. Phil. Soc. Sci. 5: 431-450.
Goody,  E. N. (Ed.) 1987 Questions and Politeness. Cambridge: Cambridge University Press.
Haack, S. 1978. Philosophy of Logics. Cambridge: Cambridge University Press.
Heritage, J. 1984. Garfinkel and Ethnomethodology. Cambridge: Polity Press.
Holton, G. 1981. Comments on Professor Harold Garfinkel's Paper. Philosophy of the Social Sciences
11(2): 159-161.
Horton, R. 1967. African Traditional Thought and Western Science. Africa 38: 50-71 &155-187.
Jackson, M. 1978. An Approach to Kuranko Divination. Human Relations 31(2): 117-138.
Jammer, M. 1974. The Philosophy of Quantum Mechanics. New York: Wiley.
Levinson, S. C. 1983. Conversational Structure. Ch 6 In Pragmatics, Cambridge: Cambridge University
Levinson, S. C. 1989. A review of Relevance (Review Article). Journal of Linguistics 25(2): 455-472.
Lewis, I. M. 1971. Ecstatic Religion. Harmondsworth: Penguin.
McHugh, P. 1968. Defining the Situation. The Organisation of Meaning in Social Interaction. New York:
Bobbs Merill.
Maupoil, B. 1943. La Géomancie à l'ancienne Côte des Esclaves. Trav. et Mem. de l'Inst. d'Eth., 42,
Paris: Institute d'Ethnologie.
Oldman, D. 1974. Chance and Skill: A study of Roulette. Sociology 8: 596-603.
Park, G. K. 1963. Divination and its Social Context. JRAI 93: 195-209.
Parkin, D. 1979. Straightening the Paths from Wilderness: the case of Divinatory Speech. JASO 10(3):
Rehfisch, F. 1960  The dynamics of multilinearity on the Mambila Plateau. Africa 3:246-261.
Rehfisch, F. 1972  The Social Structure of a Mambila Village. Zaria: Ahmadu Bello University  Sociology
Department Occasional paper 2.
Reichenbach, H. 1944. The Philosophical Foundations of Quantum mechanics. Berkeley: University of
California Press.
Salmon, M. H. 1978. Do the Azande and the Nuer use a Non-Standard Logic? Man (n.s.) 13: 444-455.
Sanders, R. E. 1988. Review of Sperber and Wilson 'Relevance: Communication and Cognition'.
Language in Society 17: 604-609.
Schegloff, E. A., G. Jefferson and H. Sacks. 1977. The preference for self-correction in the
organisation of repair in conversation. Language 53: 361-382.
Scher, N. J. M. (ed.) 1962. Theories of the Mind, New York: Free Press.
Shaw, R. 1985. Gender and the Structuring of Reality in Temne Divination: an Interactive Study. Africa
53(3): 286-303.
Sperber, D. and Wilson, D. 1986. Relevance: communication and cognition. Oxford: Blackwell.
Strecker, I. 1988. The Social Practice of Symbolisation: An Anthropological Analysis. LSE Monog. on
Social Anthropology, 60, London: Athlone Press.
Surgy, A. de. 1981. La Géomancie et le culte d'Afa chez les Evhe du Littoral. Paris: Publications
Orientalistes de France.
Thompson, J. B. 1984. Studies in the Theory of Ideology. Cambridge: Polity Press.
Vernant, J. P. 1974. Paroles et Signes Muet. In Vernant, J. P. 1974.
Vernant,J. P. (Ed.) 1974 Divination et Rationalité.  Paris: Editions de Seuil.
Werbner, R. P. 1989. Tswapong Wisdom Divination. Ch. 1 In Ritual Passage, Sacred Journey, pp 19-
60. Washington D.C.: Smithsonian Institution Press.
Zeitlyn, D. 1983. Contradictions and alternative logics. Man (N.S.) 18(4): 788-789.
___. 1987. Mambila Divination. Cambridge Anthropology 12(1): 21-51.
___. 1990. Mambila Traditional Religion.  Sua in Somié. (PhD thesis submitted June 1989. awarded
February 1990) University of Cambridge, Cambridge.  

Page 667

Return to top of page - Return to main ERA Page - Return to main Divination Page