Spiders In and Out of Court, or, 'The Long Legs of the Law':

Styles of Spider Divination in their Sociological Contexts

David Zeitlyn


Different aspects of social life require different types of explanation and have different types of history.  The patterns and forces of social history underdetermine cognitive aspects of phenomena whose broader features may indeed be sociologically constrained and therefore answerable to historical explanation.  This poses a challenge: how to produce revealing, helpful accounts which both offer explanations for developments while still remaining true to the complexity of the society being described.  The challenge lies in dealing with the complexity of social causation.  A variety of functionalism is the usual recourse of anthropologists (this will be further discussed in the conclusions below).

I shall focus on spider divination as one aspect of social life in several Cameroonian societies.  As a social phenomenon spider divination is widespread throughout the southern half of the country.  I do not pretend, however, that it is the most important aspect of those societies.  The particular task which I have set myself is to be faithful both to the cognitive work undertaken in divination and to its social role.

At the same time I wish to ask how we should explain both the similarities and the differences between these systems of spider divination.  In the broadest sense I shall indicate two types of explanation, one cognitive and the other sociological. It is my contention that each type of explanation identifies factors which underdetermine but constrain those of the other explanatory type.

Divination in court, examples from Mambila, Banen and elsewhere

The use in divination of spiders which live in holes in the ground is widespread throughout southern Cameroon.  There is a common term, ŋgam, which is used both for the spiders and, generically, for divination.  There seems little question that the technique has spread from a single source over this area. Vansina (1990: 13) takes ŋgam as an example of a recent innovation on the basis of its continuous distribution over a well-defined area. This is further discussed in Blench and Zeitlyn, 1989/90.  The relation of social structure to particular socia forms is thrown into particular focus by such a phenomenon. The fact of dispersion from common origin implies nothing about the social role of divination, nor does it imply much, if anything, about the divinatory practice.

The political systems of the groups in question may be summarised as follows:

Mambila, who are found on both sides of the international boundary, have rudimentary chiefs in Nigeria.  Among Cameroonian Mambila the institution of the chief has been strengthened by borrowing from the

Page 219

FIG. 1. Banen and Mambila territory in Cameroon

neighbouring Tikar.  They lack, however, any form of unilineal descent system.  To the south of the Mambila lies the large city state of Foumban, which dominated the area at the end of the nineteenth century.  An important and highly organised court dominated many smaller vassal chiefdoms.  Farther south still are the Bamiléké and their neighbours, the Banen, who, like the Bamoun of Foumban, have polities based on chiefdoms and lineages.  Mambila may use the results of spider divination as the principal evidence against someone accused of witchcraft.  

The Bamiléké and Banen use spider divination to diagnose the best response to illness but its results are not used as evidence in court (Hurault 1962: 78).

I shall compare the use of spider divination among the Mambila, whom I know best, and among another group to the south for which there is good documentation.  The comparison will be made both of technique and of social role.  In particular I am relying on the work of Idelette Dugast on the Banen.  Dugast worked in the Ndiki chiefdom of the Banen over a twenty-year span from the mid-1930s to the late 1950s.1

Part of the motivation for the comparisons which follow may be seen in the political role of the Zande benge oracle in its classic description by Evans-Pritchard (1976 [1937]).  All decisions made on the strength of benge oracles may be referred up to a higher court.  The ultimate appeal is to the king and

Page 220

thereby to the king's oracle (a benge oracle which is no different in kind or technique from the others).  

In Zandeland [the poison oracle's] verdicts derive an historic sanction from the fact that its verdicts were traditionally backed by the full authority of the king. ... In legal disputes, therefore, the authority of the poison  oracle was formerly the authority of the king, and this in itself would tend to prevent any serious challenge to its veracity' p.343).  

Only princes (in charge of provinces) and the king are aware of the contradictions inherent in the system.  For example, death may be held to be caused by witchcraft following divination performed by that person's kin.  A different kin group may believe that the same death is caused by vengeance magic aimed at a witch.  Royals are the only group to know both diagnoses.  However, they do not discuss these problems, so no consequences follow from such contradictions (p. 7).  I would suggest, and I hope that the examples considered below corroborate the suggestion, that it  is no coincidence that the benge oracle is of political importance and that it possesses the cognitive features which it does.  The cognitive features I refer to are those of independence from operator manipulation (a topic which Evans-Pritchard discusses at length) and the extent to which operator interpretation is reduced by the oracle providing answers to yes/no questions.  

In Cameroon a telling example of the political importance of spider divination may be found in Foumban.  Foumban is the capital of the Bamoun people, who dominated the regional politics of western Cameroon in the nineteenth century.  A city state with imperial pretensions, its rulers exacted allegiance over a wide area.  In the late nineteenth century the famous ruler Njoya converted to Islam and invented his own writing system.  With this script he wrote  his 'History and Customs of the Bamoun'.  By the time (c. 1910) that he began writing he was secure in alliance with the Germans.  This afforded him some protection from competition for the throne so he could repeal some of the laws which entrenched his position.  The history includes lists of laws repealed and new legislation.  Among the old laws which were repealed
are the following (my translation from the French):

16. If one consulted a diviner on the subject of an important man who was ill, if one affirmed that he would not die then were he to die they also killed the person who had consulted the diviner.

Moral: the clients of bad or mistaken diviners are punished!

17. Those who consult the spider die. p 126 [p. 226 in Njoya's pagination]

The latter is not concerned with errors of divination.  Bamoun spiders do not lie.  They are symbols of wisdom in myths, stories and on carvings: another law which Njoya repealed was a restriction on the use of spiders as a decorative motif on bedposts (p 128).  The use of spider divination was banned as treasonable.  The health and safety of the chief could be discovered through spider divination and therefore its use was prohibited in order to preserve State secrets.   It is not necessary to continue with the intricacies of Bamoun history to appreciate the importance that

Page 221

TABLE 1: Banen and Mambila compared
Measure Banen Mambila
Chiefs Yes No
Lineages Yes No
Spiders in court No Yes
Overnight divination Yes No
Degree of constraint Low High

was accorded to spider divination.  Among other groups its place may not have had such a great political significance. One of the tasks of the sociologist is to explain how this variation has come about and what other features accompany it.

Table 1 summarises the comparisons which  will be made in more detail in the remainder of this article.

Generalised account of spider divination

The common elements of spider divination are as follows:

The divination technique consists of posing a question to a ŋgam spider which lives in a hole in the ground.  An inhabited spider-hole is located and the area immediately around it cleared of vegetation.  Alternatively the spider can be dug out of its hole and taken to a more conveniently sited abandoned hole in which it will be kept.  Over the hole is placed an old pot (about  40 cm diameter), the upturned base of which is knocked out.  This is covered with a shard or piece of tin as a lid which can be removed to inspect the entrance to the burrow and its immediate surroundings.  When a question is to be posed various markers are placed inside the pot near the hole, and these are associated by the diviner with  alternatives posed in the question, or with the individuals concerned.  Most systems involve a set of leaf-cards.  These are made either from the leaves of particular trees or may be cut from the skin of raffia palm ribs.  They are individuated with a set of ideograms; often each ideogram is repeated, once on one card, twice on another.  Such pairs of cards are seen as bad and good respectively.  For instance, the card with one symbol meaning palm-tree refers to a problem associated with palms, that with the palm-tree doubled means that good is to be expected from the palms.

Large numbers of such cards are used: Gebauer published four sets of cards used by Yamba diviners.  The numbers in these vary from 206 to 290 (1964: 46).  Contemporary Mambila use fewer cards, thirty eight being the largest number of ideograms being used in one set, giving seventy six cards in all. After the question has been asked, the objects and the set of leaf-cards are placed in the enclosure, and the pot is covered.  It is usual for some cards or other objects to be placed over the entrance to the spider hole.  The diviner now waits (often overnight) for the spider to emerge.  When the spider emerges it disturbs the objects and the cards.  The diviner interprets the pattern resulting from the spider's passage to answer the question posed.  

The meanings of the cards may be referred to in relation to some of the other objects.  For example, where stones are placed round the hole, as

Page 222

FIG. 2: Palm-tree cards: good - bad (approximately actual size).

occurs in the Banen variant described by Dugast,  each stone may be associated with a different individual.  Cards placed near each stone are interpreted with regard to the individual associated with that stone.  Were I the consultant and the card meaning 'bad palm' was found on or near 'my' stone then I would be told not to climb palm trees, that this would be dangerous, and so forth.  Different systems have different rules for interpreting combinations, and diviners have more or less freedom to read a story from the resulting pattern.

Various identifications have been published for the spiders used in spider divination.2  It should be noted that among the Mambila, 'spider' divination is usually performed by land crabs (Sudanonautes (convexonautes) aubryi) although spiders are also used.  In Mambila both spider and crab can be referred to as ŋgamə.  This appears to be peculiar to those groups such as the Mambila and the Wute3 who live on along the escarpment of the Adamawa Plateau.



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 the Mambila Plateau escarpment, on the Tikar Plain.  My fieldwork was restricted to these 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 my 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 chieftaincy, yet their social structure otherwise closely resembles that described for the Nigerian village of Warwar by Rehfisch (1972) based on fieldwork in 1953.  Nigerian Mambila did not have the same type of institutionalised chieftaincyas is found in Cameroon.  In Nigeria villages were organised on gerontocratic principles, and lacked well developed 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.  Marriage is viri-patrilocal, and is

Page 223

increasingly on the basis of courtship, although the provision of bridewealth is still a major factor. However, bridewealth may be paid in instalments over a number of years.  It is not cited as a reason for the failure of young men to marry.  In Cameroon most people in the village are members of either the Catholic or Protestant Church; Islam is of greater importance among Nigerian Mambila.  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 (suàgà).  The results of spider divination are held to be authoritative evidence, and when witchcraft cases are passed from the chief's court to the national legal system the diviners are called as witnesses.

Previous work

My earlier work on Mambila divination has examined the different types of divination used by the Mambila in Cameroon.  The rules of interpretation of spider divination were explained and an example was examined (Zeitlyn, 1987; 1990a: chapter 3).  Further analysis of this case focused on the role of contradictions and the way in which diviners reacted when the answer to one question contradicted an earlier answer (see Zeitlyn, 1990b, c).

ŋgam dù: Earth divination

ŋgam dù  is the most  important form of Mambila divination  In the account which follows the meanings of the leaf cards are scarcely referred to.  This is because the contemporary form of Mambila spider divination uses the relative positions of the cards rather than their meanings.  We shall concentrate upon the actual process of divination and the principles of interpreting the positions of the cards.

Most adult men apparently know at least the basic principles of interpretation even if they have not formally been taught how to divine.  More men have been taught than regularly practise ŋgam.  Amongs those who do, some are widely regarded as experts and attract clients from far afield.  There is no formal requirement that a diviner should be a household
head but most active diviners are of that status.  Since the commonest reason for divination arises from illness there is more incentive for fathers to divine than there is for single men.   One senior man (Wajiri Bi) has clients who have travelled from Nyamboya and Bankim (70 km away) for a consultation.  In Somié centre I know of five regular practitioners, and in immediately outlying hamlets were three highly reputed men (Wajiri Bi among them) whom people from the centre regularly consulted.

The political role of divination, and of ŋgam in particular, is ambiguous.  I was told that it is regarded as an essential skill, expected of all senior men.  However, this is not formalised, and I suspect that the only explicit statement I received to this effect was little more than an inductive generalisation.  Knowledge of ŋgam is neither necessary nor sufficient for political success.  Yet it is a common means of achieving influence, and thus can contribute to the exercise of power. That it is not sufficient is clear since there are successful and acknowledged practitioners who do not take an active role in village politics.  That it is not necessary is illustrated by the Chief of Somié, well-respected and influential chief, who yet neither knows

Page 224

how to divine, nor consults ŋgam dù, although neighbouring chiefs are known as diviners.  

However, when I repeated  Rehfisch's 'opinion poll' asking who are the elders of Somié (Zeitlyn 1990a: 26), all those named (apart from the chief) knew ŋgam even if they were not regular practitioners.

New chiefs are selected through divination by the headman of Njerup, and Papa the headman of Gumbe.  On them alone rests any formal requirement to practise divination, and only for this one purpose.

ŋgam dù plays a crucial role in the detection of witches since it provides an authoritative verdict.  For this, divination must be carried out by two respected diviners who are not personally involved in the case.  It is likely that the chief will have  increased his authority by stressing his non-involvement with this activity, so that his judgements on the basis of divinatory results are accepted as being more neutral.  This is especially important as it concerns the role of the Chief in referring witches accused in the village court to Bankim where they are committed to the national system of justice in which witchcraft is an imprisonable offence (article 251 of the Penal Code).4  At a trial in Banyo the diviners who 'caught' the witch appear as witnesses for the prosecution.  People are sentenced with terms between six months and ten years.  Several people from Somié were remanded in custody for over a year before being acquitted.  Yet ŋgam dù, although powerful, is not merely a technique by which social control is achieved.  It is one of the ways by which men maintain their authority over women, since women are not allowed to divine.  These considerations alone, however, cannot explain the observed practices.  As has already been said in the introduction, such  'functionalist reduction' is inadequate: it can give no account of intellectual behaviour, particularly of the ratiocination of the diviners.

Basic techniques of ŋgam dù

In Mambila ŋgam is generic for divination.   ŋgam dù (lit. divination ground) is practised with both crabs and spiders. No distinction is made in the treatment they receive.  As was described in the general account, the area around the spider (or crab) hole is cleared and covered by an old pot.  To begin divination a stone is rubbed around the top of the pot which the diviner blows into saying 'yuo yuo' (come out, come out).  The procedure for asking a question involves placing a stick and a stone inside the pot, one either side and slightly in front of the hole, usually the stick to the left, the stone to the right.  Opposite the hole, about 10 cm. away, the divination leaf-cards5 are neatly stacked, pointing at the hole.

Two cards are placed over the hole.  These are usually those meaning 'End' and 'Male' although 'Walk' is also used. Their meanings however, are not usually referred to during interpretation.  Some diviners6 put a stone on the stack in the early stages of divination so that only the two cards over the hole can be moved.  This stone is only removed when further
details are needed.  These are then obtained by allowing the crab to disturb the stack.  However, the stack is often undisturbed even when unweighted. The question is posed: a small stone in the right hand is tapped on the pot following the rhythm of the speech which is often muttered.  I was told

Page 225

FIG. 3: ŋgam dù set-up

that actual vocalization is unnecessary.  Moreover, when I stumbled over the phrases in Jù Bà I was told that I could speak English, and divination would understand.  No matter what is at issue, questions follow a fixed schema allowing two possible responses, one associated with the stick and one with the stone.  The general form for a  question is as follows:

My divination, you shape-changer, you witch, if XXXXX then take the stick, my divination.

No, it is not that, not-XXXXX / YYYYY / divine further, then take/bite the stone, my divination.

Mambila  text


Sam ŋ gwə,  ‘NOT XXXXX’ / ‘YYYYY’ / mbɔ mbɔ, wò sie/numa ta, gam   mò.

The choice is between one option (XXXXX) and either its direct negation (NOT XXXXX) or an element from its contrast set (YYYYY) which may be more or less precisely specified.  Commonly the vague alternative mbɔ mbɔ, (divine further) is offered which always has a negative connotation: further divination is about something evil. The opening phrase can be extended to include other sorts of witches and idioms for witchcraft, thus becoming a list of possible sources of danger. The crab is described as being a witch since 'it must be one otherwise it would not know about witchcraft.'  When enquiring further about this I was told that 'it takes one to know one,' and reminded that people who have inherited witchcraft have 'open eyes,' and can detect witches without necessarily practising witchcraft themselves.

Once the question has been put, the pot is re-covered and the diviner(s) retire for ten to fifteen minutes to allow the crab to emerge and disturb the cards, thus giving its answer.  Often another pot is inspected and further questions put while the answer from the first pot is awaited, so a set of parallel questions may be operated.  This provides a consistency check on the veracity of the divination.  (Truth-telling is considered separately below.)   A new line of questioning is marked by the diviner breaking a twig and throwing away the fragments as he states that he will adopt a fresh approach, and that the divination is to follow suit.

Principles of Interpretation.

When the diviners return, if the crab has emerged and disturbed the cards, the resulting pattern is read.  Often an abbreviated version of the original

Page 226

question is spoken over the pot immediately before removing the lid and inspecting the results.   This section outlines the general rules by which the pattern is interpreted.  Expertise in reading the patterns is acquired firstly by divining with elders expert in divination, and especially with one's teacher.  Although the stereotypical cases can be recounted (see below) the proper interpretation of an equivocal response can only be learnt through seeing a similar response and being taught its interpretation.  The success of a particular interpretation can only be evaluated in the light of subsequent events.  While learning to divine, use is made of truth-testing questions whose answers are easily verified for example:  'Will I eat maize porridge today?'  Controlling the question not only tests the veracity of the crab but also exercises the skill of the diviner.  Later one begins to divine alone, but always refining the technique by induction from past cases.  Thus I suspect that more experienced  diviners rarely reject a response as 'saying nothing,' while this is more common among beginners. It should be stressed that these rules were presented to me as such.  In general conversation about divination a circle would spontaneously be drawn on the ground to represent the spider hole, and a stick, a stone and scraps of leaf positioned to illustrate examples.  I asked how the divination gave its answers, how it could respond to the questions asked of it.   The cases illustrated below were presented to me through the use of the diagrams as paradigm cases.   The simplest responses do not involve the whole stack of cards but only the two cards which are placed over the hole (usually 'End' and 'Male').  If a card is moved towards or onto the stick then the stick has been chosen (sie); similarly, the stone may be chosen.  The position of each card is interpreted firstly on its own according to these rules, and secondly with reference to the positions of the other cards.  Thus the two cards left over the hole may contradict one another.

The first complication of this simple system is the possibility of the cards 'looking,' which is illustrated below.  A card may be viewed as an arrowhead due to the symmetry of its shape7 and with this in mind it can be taken to indicate a direction.  Mambila diviners express this by saying it 'looks' in a certain direction.  Hence, if, when on the stick, it 'points' at the stone it is the stone which has been chosen and vice versa.  However, one diviner did not use this interpretation.  He disregarded the 'pointed-ness' of the cards, concentrating instead on whether the cards had been turned over; this distinction is also covered below. Thus far we have  considered the four following possibilities: The idea of a card 'looking' can be used to elaborate on the basic answer which is read from the alternatives attached to the stick and the stone.  If a card on the stone 'looks' outside the pot as in a) and b) above, this can be used to give more information about the evil which threatens.  For example, when trouble in a compound is at issue, a card 'looking' outwards directs the diviners to consider a cause outside the compound.  This sort of detail is often ignored when the answer selected is the alternative which the client prefers. In principle these four basic possibilities may be doubled by further

Page 227


FIG. 4.   ŋgam dù  basic responses

distinguishing whether the cards are upside down (maplim).  Normally the cards are viewed with the rib uppermost, and this is how they are laid over the hole.  In abstract discussion of interpretation I was told that an inverted card was 'bad', possibly warning of unforeseen problems,  so a card on the stick as in a) above, but inverted, is similar to one 'looking' at the stone...  It is possible to use this principle to aid difficult interpretations, although, in observed divination Wajiri Bi ignored this feature.  Bɘbɘ, who does not refer to 'looking,' equated 'maplim on stick' to 'stone' but said that all cards near the stone were bad.  Despite these variations between diviners there is far more consensus than is reported among Bamiléké diviners (Pradelles 1986:311-313). Some responses are portents of death: the pulling of cards down into the hole, the balancing of cards against the pot wall so that they point (or 'look') down into the ground, or the pushing of the cards outside underneath the pot.  Bɘbɘ made the distinction between the simple pulling of cards which remain flat into the hole, signifying a 'bad' situation which must be corrected, and the cards being folded over in so doing, which tells of a death to come.

Further rules of Interpretation

1) If the card(s) placed over the hole are inserted into the stack then the

Page 228


FIG 5.   ŋgam dù Further responses

divination is taken to have selected the card above the place of entry.  The meaning of this card is referred to in the result, usually in the context of the positions of other cards.  This is the only instance in which the meaning of the cards is invoked in Mambila divination.

2) A card balanced on its base against the pot wall augurs well, whereas balanced on its point it portends death.

These basic rules are sufficient to interpret the simple cases.  The skill in divination lies in the ability to interpret equivocal results, for example when one card is on the stick and another on the stone.  Most often, however, such a result will be rejected as saying nothing.

Divination and Truth

The veracity of any particular divination result may be questioned by the participants.  Other types of Mambila divination such as ŋgam (divination of the ŋgam tubu vine) can only be checked inductively by asking easy questions and, unlike its more serious counterpart, no remedial procedures are available if it is found to be lying.  It is possible that this omission relates to the degree of operator-dependence.  Manipulation by the operator is possible in ŋgam tubu in a direct manner unlike ŋgam  Hence the operator can be blamed whenever the divination is proved wrong.  The suspicion of manipulation prompted sarcastic comments at a demonstration of another type of divination8.  On the other hand, a variety of tests and techniques is employed to ensure the truthfulness of ŋgam which is not operator-dependent.

The most routine check is applied during every divination session by repeating the same question in the same pot.  On the second occasion the stick and stone are transposed.  This enables the diviners to reject answers resulting from the leaves being pushed repeatedly in one direction.  The divination must appear to be paying heed to the question being asked.

Other techniques involve administering an ordeal to the spider.  At intervals of approximately a month sér ŋgam; (porridge [of] divination) is prepared by the diviner who puts it into the holes while uttering a variety of encouraging phrases such as:

Two or three days later divination restarts with a set of questions to establish the state of each

Page 229

particular pot.  Truth-telling is tested by asking either 'Am I here?' or 'Will I eat maize porridge today?'9.

The spiders may also be asked whether any witchcraft is attempting to interfere with them.  This is the only instance where any break with orthodox Western logic occurs.  A 'Cretan liar' paradox results if the answer provided is 'yes'. However, I did not succeed in pointing out the fallacy.  If an answer indicates that witches are interfering then the spider is not consulted that day.

A less common treatment is to administer the powdered inner bark of a tree10 which has been scraped onto an old-style Mambila hoe-blade11.  The bark is sprinkled into the holes, using both hands, to the accompaniment of an invocation similar to that described above.  It is unclear whether all practitioners of ŋgam use this technique, but certainly all use sér ŋgam.

Central to the learning of divination is acquaintance with the names of the leaves which are cooked and eaten with a chicken before the remainder is administered to the divination pots.   Subsequent preparation of sér ŋgam repeats the essentials of this initiation.  It is described as being an ordeal for the spider: only truth-telling spiders can eat it and survive. As has been said a consistency check is performed by using several divination pots simultaneously,  or by simply repeating a question several times.  Only if several pots give the same answer will it be believed.  I never witnessed a case where this became an issue; most conflicting results were read as adding detail to a single answer12. Responses which directly contradict one another are either taken to be 'criticizing the question' or are explained away as aberrations.   With respect to any particular pot a highly empirical attitude is taken.  By using the methods mentioned above and by inductive tests, practitioners satisfy themselves that a pot and the spider or crab within it is truthful.  If a pot persistently misinforms, giving wrong answers to the tests and giving answers inconsistent with those received from other pots, then that pot will simply be abandoned.

Banen spider divination

In her monograph on the Ndiki (Banen) Dugast (196013) devotes 43 pages to divination and spider divination is the first of the six divination types described.  Ordeals are described separately under a different heading, a point to which we shall return. The Ndiki are a group of small-scale farmers with segmentary lineages among which is a Chiefly lineage.  In an area in which there was no land shortage a question may be raised about the origin of the chief and his authority.  In the north of Cameroon there is uncertainty about rainfall giving rise to a niche for ritual specialists and the possibility of its monopoly by chiefs. This does not exist for the Banen who are on the margins of the tropical forest, where rainfall is reliable and plentiful.  The situation is very close to that pictured by Kopytoff in his characterization of 'the African Frontier' (1987) in which competition is for followers (qv Goody 1971) and successful groups split into factions some of which go into the frontier, the uninhabited bush, where they settle.  If they are successful immigrants are

Page 230

attracted and the founding group subsequently becomes the chiefly lineage.  A similar social structure is reproduced by the those who cede from the parent group14.  Such a characterization seems to fit the Bamiléké and Banen, and the zone on the edge of the forest.  To summarize: the Chief and his lineage have a limited set of ritual powers but are not, at base, different in kind from any other lineage.  They have come to dominate as a result of a historical process.   They had (in precolonial times) a monopoly over the use of capital punishment or the selling of kin into slavery (often a preferred solution to the problems posed by habitual offenders).  In such a group the management of witchcraft accusations is important.   Witchcraft accusation is a means by which political rivalries may be expressed, and may lead to the fission not only of lineages but also of the group.  Since the chief must retain followers in order to maintain a viable economy such accusations must be managed and resolved.  Spider divination  does not play a role in this process.  Its use is widespread, however, in response to illness.  

The presence of lineages among the Banen should also be remarked upon unlike the Mambila case in which lineages are absent.  Elders have authority by virtue of their seniority in the lineage. Lineage heads arbitrate internal disputes.  The lineage system 'naturalizes' their authority. Compared to the Mambila case there is less need for an elder to refer to an external authority (divination).

When we come to examine the Banen variety of spider divination we must remember that sociologically it is used in response to personal problems but it is not important when personal problems are placed in a wider arena.  It is not an element of political debate.  Hence its operations are not subject to the constraints of public debate or scrutiny as may occur among the Mambila.  The principal Banen responses to witchcraft accusations were to swear oaths while holding tortoises (it was held to be fatal to lie on this oath) or to undergo a variety of ordeals.  The results of spider divination may have led to the initial accusation but they do not figure as evidence in court.  The results of divination are not subject to public scrutiny in court.

The interpretation of Banen spider divination occurs as a two stage process which I will explain next.  

Who are diviners, who are clients?

Diviners are held to have organs (transmitted matri-laterally) located near the stomach which enable them to interpret dreams, to divine effectively and to administer the treatments which may follow.  These organs bibuŋə (sing. ibuŋə) are not necessarily present at birth.  It is possible to acquire them and hence the ability to divine by eating the bibuŋə of chickens together with certain special medicines.

However, there is great ambiguity in making a claim to possess bibuŋə since the possession of these organs is also a necessary requirement for the mulemp (sing. balemp) - those who transform at night and eat the internal organs.  <Mulemp may be of either sex;   diviners are male.  Diviners have 'good' bibuŋə, they are not mulemp, they never transform or eat people.  Dugast illustrates some of the types of questions asked in divination, but not who asks them.  I suspect that only men may be clients, but this may be an over-generalization from the Mambila model.  Dugast tells us that spider

Page 231

FIG. 6 The four batons, cut to represent male and female pairs, good and bad div_crt_1_15.gif

divination is consulted before a mother visits her daughter, before a man climbs his palms to harvest the nuts, or leaves on a journey. It is also consulted about the breakdown of a marriage or about the possible location of a house.  A similar range of questions is asked in Mambila divination.

The timing of questions

A single question is asked in the evening.  The pot is inspected the next morning to discover the answer. If the response is unclear then it may be repeated the next day.  Clearly such a method demands that the questions are not restrictive, since to ask many small precise questions would simply take too long.  In cases of illness it would be inappropriate to spend several weeks asking a sequence of questions to rule out successive possibilities.  Rather, one large, general question is asked.  Dugast gives the following example:

'I have brought all the (divination) leaves cut by me myself because someone has come so that I can ask you if he is all right.  Tell him this.  If he will out live this year, tell him.  If he will have troubles, or if a bad fate is attached to him, show it. Show me a sign by which I will know that his enemies will achieve nothing.  Place them before me in their place so that I can know fully that they don't want ill of him.  If they want to kill him, show it to me.  If he will win over them this month then place him over them.'   (p 50, my translation.)

This type of questioning puts the onus on the diviner.  When the question is vague more interpretation is necessary to arrive at a definite answer.   So, overnight questioning using only one spider at a time, constrains the type of question and increases the work which the diviner must do.


Four batons, cut from the skin of a palm rib are placed over the spider hole, skin side up.   Eight stones are placed about 15 cm from the hole, forming an octagon around it.  Different individuals are associated with particular stones.  The stack of cards is propped upright, wedged between two small sticks stuck in the ground.  The end of the stack is over the hole so that it will be disturbed when the spider emerges.


The four batons are used to indicate the general form of the answer.  The batons are cut with patterns and are taken to represent two male-female pairs, a good pair and a bad pair15. The good pair have notches cut

Page 232

 div_crt_1_15.gif FIG 7 A good sign.  There is nothing to fear from the enemy.

in one edge, the bad pair have notches cut on both edges. Short batons are female, long male.  When the question is posed they are placed over the spider's hole in sequence (from left to right): good female, good male, bad male, bad female. The examples of interpretation as given by Dugast are shown in Figs. 7-8.  If the order of the batons is changed, and the bad-female is covered by the bad-male (fig. 9), then 'the good have taken the place of the bad:' opponents will be overcome and if the man goes hunting he will kill an animal. Further examples are given of the significance of turning over the batons to reveal the 'inside' surface (shaded) in Figs 10-12.  There are clearly far more possible results than have been illustrated here.  What we are not told is whether there are other standard interpretations of the form 'turned-over, and pushed away from the hole = a corpse at the crossroads.'


There are 84 different cards, many (21 plants, 12 animals: 33 in all) relate to items which may be used in treatments to avert, cure or circumvent the problem as revealed by the interpretation of the batons.  The cards covering the stones are taken to relate to the person associated with that stone.  Unfortunately Dugast does not give much information about the possible complexities of interpretation.  For example, we are not told of the significance of the spider turning over a card.


The relative positions of the four batons provide the key to the main part of the answer.  The cards are then referred to in order to produce more

FIG. 8 'A bad sign: an enemy will kill the consultantdiv_crt_1_14.gif

Page 233

FIG. 9 'The good has taken the place of the bad'div_crt_1_14.gif

detail.  In particular they are used to specify the ingredients of any medicine (Dugast talks of 'amulets') which may be prepared.  In fact the preparation of some medicine (or other ritual treatment, such as circling the patient's head nine times with a particular animal which has been identified by divination) is the usual outcome of a consultation.  In principle it seems possible for the leaves to contradict the batons, but Dugast does not consider this possibility.  If priority is given to the batons any direct contradiction may be avoided.  Conflicting diagnoses may be seen simply as a warning of different and various threats.  It is not a contradiction for a threat to come from two people simultaneously, nor for two people to be threatened.  If the batons indicate a different person from the cards then this could be taken as revealing both a major and a minor threat.  Dugast states that the diviner is often unable to interpret the pattern.  In this case the diviner will repeat the question the next night.

Methods and styles of explanation

In the background to this work lie a set of questions about anthropology and history and how different aspects of social life should be explained.  A brief account of these questions now follows:

People bear children and socialise them.  They acquire language and much more.  They learn not just how to make a living but how to act in society.  They find partners and have their own children in turn. Thus, at its crudest, do societies reproduce.  We could elaborate this account in many directions to explain different types of continuity between generations.


FIG. 10 Neither the husband nor the wife should leave their hut during the day because they would be attacked by opponents who would overcome them.

Page 234


FIG. 11 The good-male is said to be 'at the crossroads'. The consultant must avoid crossroads that day since the diviner 'has seen a corpse'.

But change is harder to explain than continuity, and the direction of change even more so.  Recently some authors have shown how different aspects of society change at different rates.  For example, elements of ritual performances may be relatively invariant while its overall significance and political role may undergo radical change (Bloch 1986).  To say that an element is 'relatively invariant' means that it must be recognisably similar on the occasions of different performances.

The requirement of such invariance puts constraints upon those elements.  We may term 'cognitive' such constraints as the need to be learnable in the first place and subsequently to be remembered.  Pascal Boyer has discussed the implications of this (1990).  To go much further, one might ask which cognitive constraints are imposed (on us as a species) by our ability to learn and use language.  Macnamara (1982) discusses some of these issues from the perspective of developmental psychology. At the most general level different sorts of explanations (involving different factors) must be given for different aspects of a society.  Each factor is under-determined by the forces which drive the others.  Yet forces and


FIG. 12  If an opponent is 'at the crossroads' then 'the opponent will die', and the consultant has nothing to fear from crossroads.

Page 235

factors are constrained and constraining upon one another.  The result is a complex web of interaction and feedback which accommodates the vicissitudes of real life and social change.

Functionalism reprieved

Reports of the death of functionalism have been somewhat exaggerated.  True, functionalism provides little assistance in explaining change.  Also, it gives teleological explanations which justify the existence of aspects of society in terms of their unintended consequences.  Agency is among the casualties of such an approach (Fardon 1988).  Notwithstanding these limitations, functionalism has a contribution to make.  For, at the crudest level things do get done.  Lack of fit between social institution and what it does may then be seen as a motor of history.  But if funerary rites do not dispose of the dead, or if the processes of dispute resolution never decide who should be farming which plot of land then they are not what we think they are.  I take it that in every society the dead must be disposed of and that disputes must be resolved, by force perhaps16.  

Here lies the remaining utility of functionalism.  The role played by an institution acts as a constraint upon its past and present form.  Any dispute resolution procedure must be able to resolve disputes.  The analysis of spider divination shows that such naive and crude a starting point may still produce revealing results.  I should stress, however, that I am by no means suggesting that functionalism is the be-all and the end-all.  Rather, it behoves us to start by asking: how much can be explained by this type of minimal functionalism.  The large part of social life which will not be amenable to such an approach must be explained differently.  To give a concrete example Atkinson and Drew (1979) analyse the use of language in court- rooms.  Linguistic usage in court-rooms differs greatly from that of ordinary conversation.  Atkinson and Drew seek to demonstrate how much of the formality of court-room interaction can be explained by the linguistic setting (a barrister asks questions of a witness so as to be understood by a jury which cannot overtly signal that it is following the exchange).  The silence of the jury necessitates departures from ordinary conversational practice.  This alone can explain much of the strangeness and formality of court-rooms which we may otherwise attribute to more general sociological factors.  The linguistic approach can explain part of court- room organisation.  Other more disparate sociological factors must explain the rest. The problems with functionalism are many and need not be further rehearsed here.  Yet as has just been suggested its use remains important.  If we succeed in making a case for some constraints on the possible range of human society then we may be guided to look for more detail in the actual history of an institution, looking within the realm of the possible.  Such guidance may be of great help when attempting to reconstruct the past, particularly when one is doing this with incomplete and very partial records such as are found in colonial or mission

A functional account

Elster (1979:28 & 1982) has formalized and discussed criteria for a functional

Page 236

account which is non-teleological and ideologically neutral.  He identifies five criteria summarized by Thompson et al. (1990:201) as follows:

  1. 'Y (the function of X) is an effect of X (behavioural pattern).
  2. Y is beneficial for Z (the group)
  3. Y is unintended by the actors producing X.
  4. Y (or at least the causal relation between X and Y) is unrecognised by the actors in Z.
  5. Y maintains X by a causal feedback loop passing through Z.'

This formalization may be applied to the accounts I have given above,as follows:


  1. 'Y (the political role of divination - its use in court) is an effect of X (the 'objectivity' of the divinatory style).
  2. Y is beneficial for Z (the senior men)
  3. Y is unintended by the actors producing X.
  4. Y (or at least the causal relation between X and Y) is unrecognised by the actors in Z.
  5. Y maintains X.'  

One of the justifications for such an exercise is that it reveals the limits of our understanding.  Step five is uncertain.  Is it correct to suggest that the use of spider divination in court maintains the 'objective' style of the Mambila divinatory technique?  I am sceptical of the suggestion, and uncertain about the sort of data needed to answer the question.  Is it legitimate to suggest that the since the senior men use the results of divination in court this affects the practice of the diviners?  Since the diviners are the same senior men who discuss and decide court cases it is not an implausible suggestion.  However, above I have argued for a weaker proposition, namely, that the style of divination appears objective and hence does not invite suggestions of manipulation in court.  We now see that this negative cast to the argument is not sufficient to establish the fifth step to a functionalist account which would satisfy Elster's

For the Banen the fifth step of the functionalist argument appears to work although perhaps to less effect.  The results of divination are not produced at court or at moots. They are not subject to the scrutiny which argument may entail. Individual diviners therefore have more freedom to develop their own idiosyncratic styles of interpretation, and to produce holistic accounts based on an interpretation of the overall pattern. The use of spider divination in court would be non-functional. Compared to the use of tortoises or ordeal spider divination is open to the charge of manipulation.


  1. 'Y (the non-political role of divination) is an effect of X (the 'subjectivity' of the divinatory style).
  2. Y is beneficial for Z (the senior men)
  3. Y is unintended by the actors producing X

Page 237

4. Y (or at least the causal relation between X and Y) is unrecognised by the actors in Z.

5. Y maintains X.'X.


Possible explanations for  the features which I have been examining must skirt the impasse of functionalism.  There is a fit between what is done and the form of what is done.  The naive view would be that  it could not be otherwise. Although such naive functionalism is incapable of allowing for change, nonetheless, what is suggested by the data just presented is that a functionalist stance may be used in order to understand what different types of explanation are needed for different aspects of society. Many questions remain outstanding when we examine the ethnographic map and seek to explain why things are as they are.  The example just considered invites the question why spider divination does not stand behind the power and authority of chiefs among the Banen in the way that ŋgam does for the Mambila chiefs and benge oracles did for the Zande royalty?  

The cognitive features of different varieties of divination provide partial explanation to this.   Where divination is done overnight, and only one question is asked at a time, the questions must perforce be general.  General questions produce vague answers, so the diviner must do more interpretative work in order to produce a definite answer.  Such divination systems are open to charges of manipulation and bias on the part of the diviner.  Therefore, they are less likely to be used to develop the power and authority of chiefs, and will hence have a lesser role to play in public disputes and other arenas within which politics is played out.


1An expanded version of this argument will be found in Zeitlyn, in prep.  This will also include detailed discussion of the work of Pradelles de la Tour Dejean who worked in the Bamiléké chiefdom of Bangoua in the 1970s.

2Mambila use Hysterocrates robustus Pocock, 1899.  Gebauer identified (1964:42) the spider used in Yamba divination as Heteroscroda crassipes, and Leiderer found Phoneyusa bidenta Pocock, 1899 in use among the Bekpak (1982:116).  Nicod (1948, facing p65) describing neighbours of the Banen, illustrates a spider which is Hysterocrates sp. and Laburthe-Tolra (1981:469) gives Hysterocrates  sp. for the ŋgam spider among the Beti.

3According to Siran, p.c.

4Rowlands and Warnier (1988) discuss the relation of sorcery, along with its embodiment in law, to the Nation State.

5Gebauer 1964:35 calls them "leaf-cards",  or "cards": a usage I adopt here.

6Nggeyea Abraham, and those that he has taught.

7See diagrams.

8This was ŋgam ŋgɔfɔrɔ lit. divination of snail, a divination type in which a snail shell is threaded on a length of string held by the diviner.  The ease with which the snail shell moves on the string is taken to give answers to the questions posed.

9The Zande tests of the efficacy of benge poison are similar, q.v. Evans-Pritchard 1937:337.

10Bop: Albizia zygia (DZ).

11Formerly used for bridewealth, these are now rare and are only used for rituals. I could not ascertain whether such hoes were once in everyday use.

12This is discussed in greater detail in Zeitlyn 1990b.

13Readers are referred to chapter V 'La Vie psychique', p 27-150.  A further divination type is described in appendix 2,  pp 603-15.

14Turner 1957 is the classic ethnographic description of such a process.

15No dimensions are given in the text.  The Musée de l'Homme have a set collected by Dugast in May 1935 (38.114.33 1-4).  These are c. 1 cm wide, the 'male' batons are c. 15 cm long, the female ones c. 12.5 cm in length.  

16This is not to deny that there may be misleading procedures which mask oppression.  Rather the opposite.  It is the observation that such procedures do not in fact resolve disputes that reveals them as misleading ideological constructs.

Page 238


Atkinson, J. M., & Drew, P. 1979. Order in Court  London: Macmillan.

Blench, R., & Zeitlyn, D. 1989/1990. 'A Web of Words.' SUGIA (Sprache und Geschicte in Afrika) 10/11, 171-186.

Bloch, M. 1986. From Blessing to Violence: history and ideology in the circumcision ritual of the Merina of Madagascar  Cambridge: Cambridge University Press.

Boyer, P. 1990. Tradition as Truth and Communication.  A cognitive description of traditional discourse  Cambridge: Cambridge University Press.

Dugast, I. 1955. Monographie de la Tribu des Ndiki (Banen du Cameroun)  Paris: Institute d'Ethnologie.

Dugast, I. 1960. Monographie de la Tribu des Ndiki (Banen du Cameroun). 2: Vie Sociale et Familiale  Paris: Institute d'Ethnologie.

Elster, J. 1979. Ulysses and the sirens. Studies in rationality and irrationality  Cambridge: Cambridge University Press.

Elster, J. 1982. Explaining technical change  Cambridge: Cambridge University Press.

Evans-Pritchard, E. 1976 (1937). Witchcraft Oracles and Magic among the Azande  Oxford: Oxford University Press.

Fardon, R. O. 1988. Raiders and Refugees. Trends in Chamba Political Development 1750-1950  Washington D.C.: Smithsonian Institution Press.

Gebauer, P. 1964. Spider Divination in the Cameroons  Milwaukee: Milwaukee Public Museum.

Goody, J. R. 1971. Technology, Tradition and the State in Africa  London: Oxford University Press for the International African Institute.

Hurault, J. M. 1962. La Structure Sociale des Bamiléké  Paris: Mouton.

Kopytoff, I. 1987. 'The Internal African Frontier: the making of African political culture,' in I.Kopytoff (ed.) The African Frontier: the reproduction of traditional African societies,  pp.
3-84. Bloomington: Indiana University Press.

Leiderer, R. 1982. Le Médicine Traditionelle chez les Bekpak (Bafia) du Cameroun. (Coll. Act. Anth. 26-27)  Berlin: D. Reimer.

Macnamara 1982. Names for Things.  A study of Human Learning  London: MIT Press.

Nicod, H. 1948. La Vie Mystérieuse de l'Afrique Noire  Lausanne: Payot.

Pradelles de Latour, C.-H. 1986. Le Champ du Langage dans une Chefferie Bamiléké  Paris: Unpublished thèse du Doctorat d'Etat, EHESS.

Rehfisch, F. 1956. The Social Structure of a Mambila Village  University of London unpublished MA Thesis.

Rehfisch, F. 1960. 'The Dynamics of Multilinearity on the Mambila Plateau', Africa 3, 246-261.

Rowlands, M. J., & Warnier, J.-P. 1988. 'Sorcery, Power and the modern State in Cameroon.' Man (n.s.) 23 (1), 118-132.

Tardits, C. 1960. Les Bamiléké de l'Ouest Cameroun  Paris: Éditions Berger-Levrault.

Thompson, M., Ellis, R., & Wildavsky, A. 1990. Cultural Theory  Oxford: Westview Press.

Turner, V. 1957. Schism and Continuity in an African Society  Manchester: Manchester University Press for the Rhodes-Livingston Institute.

Vansina, J. 1990. Paths in the Rainforest.  Toward a History of Political Tradition in Equatorial Africa  London: James Currey.

Zeitlyn, D. 1987. 'Mambila Divination.' Cambridge Anthropology 12(1), 21-51.

Zeitlyn, D. 1990a. Sua in Somié.  Mambila Traditional Religion, a revised edition of 1990 Cambridge PhD thesis 'Mambila Traditional Religion: Sua in Somié'. Oxford: Privately published.

Zeitlyn, D. 1990b. 'Professor Garfinkel visits the Soothsayers. Ethnomethodology and Mambila Divination.' Man (n.s.) 25(4), 654-66.

Zeitlyn, D. 1990c Divination as Dialogue: the negotiation of meaning with random responses Paper presented at Colloquium on the Implications of a Social Origin of Human Intelligence, Wissenschaftskolleg zu Berlin. E. N. Goody (Ed.)

Zeitlyn, D. in prep. The Techniques of Divination.


The initial 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.  Further research has been carried out during the tenure of a Junior Research Fellowship at Wolfson College, Oxford.  Both Wolfson College and the Faculty of Geography and Anthropology, University of Oxford have been generous with assistance towards the costs of fieldwork.  The writing was completed during the tenure of a Junior Research Fellowship at Wadham College, Oxford.

Richard Fardon provided detailed criticism as a reader on behalf of the Editors of Africa. This has improved the argument; the faults that remain are mine not his.


Both Banen and Mambila in Cameroon use spiders in divination.  The intellectual practice of the divinatory technique is related to the different sociological contexts of divination.  Relative freedom of interpretation and the posing of questions overnight among the Banen correlates with its political unimportance.  Conversely among the Mambila the interpretation of the diviner is rule bound, and many simple questions may be asked in quick succession.  The results provide important evidence in the chief's court.  Social context and intellectual content are each constrained but under-determined by the other.

Return to Top of Page - Return to Main ERA Page - Return to main Divination Page