Return to Readings

Towards eliminating the concept of religion


F.B. Welbourn

(Delivered to the Colloquium on the Concept of Religion
Lancaster University, December 1969)


Je donne une oeuvre subjective ici, oeuvre cependant qui tends de toutes ges forces vers l’objectivite’. (Minkowskim, quoted Lang, 1945, p. 12)


When the Kikyuy Karinga’a Association, in 1929, declared its intention of returning to the purity of tribal custom, it decided to have nothing to do with dini for seven years. In the same spirit the Buganda Government used to describe pagans as ‘men who have no dini’. An alternative was ‘those who do not read’; and the equation is significant. It lends point to the statement of a Ganda civil servant who, brought up as a Christian, discovered in his retirement the virtues of the old spirits: ‘There’s no conflict between dini and kusamira. Dini is good. It has brought us education and science. But kusamira – that’s part of being a Ganda!’. It lends point also to the criticism that, as a priest, I had commented on Kabaka Yekka (the ‘king alone’ party): ‘The Kabaka has nothing to do with dini. He is a matter of obuwangwa (“essential nature”).


I use vernacular words because, without circumlocution, it’s difficult to translate them into English. Dini is an Arabic word, used throughout East Africa to describe imported systems of creed, myth, ritual and moral precept – Islam and many variants of Christianity – and, by deprivation, the independent churches and the syncretistic movements which have sprung up in response to the impact of the West. It implies something foreign, something about which choice is possible – just as a man may change his clothes, or choose to go naked, according to his company. Adapting the Ganda civil servant, he might say: ‘Clothes are good. They give respectability. But skin – that’s part of being a man’.


On the other hand, there’s no such choice about kusamira. Like the Kabaka, it is a part of essential nature. The spirits are there (even if invisible), experienced, as we experience atoms, as an integral part of the environment: just as the Kabaka is there, even if you never meet him. But if you do so, you must kusinza (‘pay homage to’) him. In the same way, if you encounter a spirit, you must kusamira. It is the appropriate form of response to a particular class of beings who exist in Ganda society. To neglect kusamira because you become a Christian is not to choose Christian dini instead of Ganda dini, but to cut yourself off, under the influence of foreigners, from a fundamental part of Ganda society. In the same way, the Gikuyu dissidents were not choosing Gikuyu dini in preference to Christian dini. They were saying that to be Gikuyu was the first essential, that this involved by definition such definition such practices as clitoridectomy, and that – even if, as was hoped by another group of dissidents, these could ultimately be accepted by Christians – Christianity was at the best an optional extra, better left alone until pure Gikuyu society had been re-established.


In making this sort of distinction between dini and tribal custom, they were doing no more than they had learnt from at least some of the missionaries, who told them that they had no dini;  and there is no doubt that, in doing so, the missionaries thought they were saying not, ‘You have no revealed religion’ but, ‘You have no religion’. Whether because the meaning of the word has changed, or because we have a better understanding of African society, we could not say the same today. ‘Tribal religion’ is, as much as the ‘world religions’, a subject of concern to scholars, missionaries, and administrators. The question is whether, in the two connections, we are not using ‘religion’ in two separate senses, and whether, therefore, clarity would be better served by eliminating it.


Perhaps I can make my question clearer by reference to a slightly different, though still African, context. When I started studying African independent churches, I started from a missiological interest. But I found that I needed the help of scholars in a wide range of ‘secular’ disciplines. I said, in the preface of my book, that it fell somewhere on the boundary between Church History and Social Psychology. I later found it recommended reading for Political Scientists. On the other hand was a missionary complaint that I had given no picture of these bodies as ‘churches’. I had described them in relation to the world around. I had said too little about their inner life of prayer and worship and Bible reading. Whether or not this criticism was just (I don’t think it was), it no doubt reflects my own idea of what the Church is. But I think the issue is more fundamental; and it was immediately clarified for me when I met the African Israel Church Nineveh. Here was a community (Gemeinschaft) which had ample creed and myth and ritual but would have been grossly misdescribed unless this ‘religious aspect’ was presented as part and parcel of an attempt to establish a whole way of living in colonial Africa. The same surely has to be said about an Early Church History which discusses liturgical and doctrinal developments but forgets that Clement thought fishing an activity suitable to Christian gentlemen.


There is a sharp contrast with a contemporary English Anglican congregation at Easter. It contains men who follow the same ritual of receiving the sacrament, repeat the same creed, accept the same resurrection myth. In terms of a behaviourist (dare I say, a phenomenological?) description, they all fall in the same category. But they include men who value the Easter communion as a social tradition; who find in the sacrament an almost magical potency; who see it as no more than a memorial of a historical event which saves individual souls; who think primarily of establishing a ‘Christian order’ of society. They belong to the same religious institution. But their private affinities may be with agnostics, Roman Catholics, revival sects of Free Church politicians; and they may have institutional affiliations with monarchists, artists clubs, family capitalist enterprises and/or the Labour Party. Thompson (1957), in his study of four Birmingham parishes, found that it was these so-called secular affiliations which, much more commonly than church-membership, had an integrating function – proved something of a Gemeinschaft in contrast to the Gesellschaft of the churches.


It seems to me that, whether or not there are satisfactory German words to describe the two different types of phenomena, represented by AIGN and the Birmingham parishes, it is quite impossible – without gross depreciation of the verbal currency – to use for both of them the word ‘religion’ or even the word ‘church’. To do so is simply to ignore their inner meaning; and it was as long ago as 1922 that Weber (1966 p.1) wrote, ‘The external courses of religious behaviour are so diverse that an understanding of his behaviour can only be achieved from the viewpoint of the subjective experiences…’. This is not peculiar to religion. It is surely the basic difference between the British tradition of Social Anthropology and the attitude of behaviourists. It is also the problem with which Laing (1960, p.p. 28-35) was wrestling in his approach to psychotics: ‘If I compared my experience… with the accounts given in the standard textbooks, I found the authors were not giving a description of the way these people behaved with me’, because they were describing the ‘objective clinical signs’, which for Laing, ‘the therapist must have the plasticity to transpose himself into another strange and even alien view of the world. In this act he draws on his own psychotic possibilities’.


The same sort of thing happens when a European studies ‘religion’ in Africa. He has his ‘clinical signs’ of what religion is. Among other things, he understands it as one institution among others – political, economic, legal and so forth. He probably regards it as a voluntary activity. He finds similar signs in Africa and describes them in terms of his own experience: not, it is true, of voluntariness – the facts cannot be stretched that far – but as an institution separate from other institutions, to be relegated (by anthropologists) to a final chapter: or described (by comparative religionists) without any reference to the social context. The family can be described without mention of the living dead, who are experienced as an integral part of it. Alternatively, Parrinder (1969) describes ‘ancestor cults’ without stating their function as affirmations of clan solidarity, or the differences between patrilineal and matrilineal societies. Neither party, so far as I know, has paid attention to the possibility that in some societies they may be based psychologically on what Klein calls the ‘schizoid position’, in others on the ‘depressive’ (Welbourn, 1969a). Yet, without these factors – the social context and the inner meaning – what is being described is not ‘another strange and even alien view of the world’ but merely its ‘clinical signs’.


There is, of course, a more recent school of British social anthropologists (e.g. Evans-Pritchard, 1937 etc.) who see the understanding of African religion and African society as interdependent. They are, I suspect, drawing on their own religious possibilities, as Laing wished to draw on his psychotic possibilities, to enter into an alien view in a manner denied to those whose experience is limited to ‘Birmingham religion’. Godfrey Lienhardt has even written, somewhere, that an anthropologist will always draw on his own religion, or on what for him takes the place of religion, to interpret his observations. When this happens, it becomes possible to see traditional ‘religion’ not as one social institution among many, but as a dimension of all institutions: not as a set of clinical signs but as a total – if strange and alien – way of life of which the clinical signs are but one mode of expression. Again I do not think that one word is adequate to describe both categories.


In these existential-phenomenological terms there are, it seems to me, at least three different social phenomena in Africa all designated as ‘religion’. There is, first, traditional ‘religion’. This is a given part of social experience. Although it deals with crises of individuals as well as of society, it is in most cases primarily an affirmation of social solidarity. In principle it is involved in every aspect of life. Per contra, it is not a voluntary option; and it cannot, except with gross distortion, be studied as a separate institution. It belongs to what H.W. turner (1969) calls an ‘ontocratic’ society, though I prefer ‘unitary’.


At the other end of the scale is the Birmingham’ type of religion, which is found increasingly – whether in Christian or traditional dress – in the towns. It is voluntary. Although it may be concerned with the affirmation of missionary mores, its primary focus is the salvation of individuals. It has little relation with what goes on outside the church building. It is one element in ‘a modern secular state and a religiously plural society’ (H.W. Turner, 1969) – though it must be remembered that, in ‘Birmingham’, ‘religiously plural’ means the choice not merely of a number of different congregations but a wide variety of beliefs and allegiances within each congregation.


Thirdly, there are some of the independent churches, some rural congregations of missionary origin, and Islam perhaps in most of its East African manifestations. They are voluntary – at least in the sense that choice does in fact occur. In a secular and plural society they provide a strong sense of identity for their individual members and are closer, in ethos, to tribal solidarity than to western individualism. In principle they are involved in every aspect of life except, in some cases, national politics.


It may be that this third type is best regarded as a mixed or intermediate type between the other two. Certainly, between the three, it would be possible to find other mixtures and other mediates. There are obvious resemblances and an observable, if not indeed a predictable, transition from one to the other. But there are obvious resemblances, and no doubt an observable transition, between male circumcision in Kalenjin society (Welbourn, 1968) and in our own. But I do not call the latter ‘initiation’; and I gravely misrepresent the former if I present the circumcision element as any more than the focal point in a six-month rite of passing from boyhood to manhood. I suggest that at least the same difference exists between traditional and ‘Birmingham’ religion. I cannot justify the use of the same word for both. In ‘Birmingham’ both religion and circumcision have become residua of more fundamental, and radically different, forms of behaviour.



On this analogy it might be right to keep ‘religion’ for the residuum and find some other word for the ontological concern of a unitary society. But this won’t do, if only because, as one who professes ‘the study of religion’, I find the residuum as academically uninteresting, and as empirically irrelevant to the western world, as is male circumcision. If I am to draw on my own religious possibilities – and these are an integral part of the evidence – I have to define ‘religion’ in unitary terms:

i.                     to be a Christian is not for me a voluntary option. It is not, of course, a necessary part of my culture; and no doubt it has been partly determined by numbers of particular choices. But fundamentally it is something that has happened to me. It is an ontological condition of a wholly different order from decisions as to whether I continue to practice my priesthood, or go to church, or tolerate extra-marital sex. I cannot choose not to be a Christian. That is not to say I might not cease to be one. But, if that happened, it would be precisely because something had happened, not because I made a choice.

ii.                   To be a Christian is concerned with what it means to be a man. It asks questions about individual behaviour and destiny; and it does not for me find much support in the social institutions called ‘Christian’. But its primary reference is corporate. It is a statement – even if a hesitating statement – about the whole human race.

iii.                  To be a Christian means to ask theological questions about every aspects of the social and natural orders. It is theologically just as important to dig good latrines as to celebrate the liturgy.

iv.                 Negatively, to be a Christian does not mean that I have any special brand of experience called ‘religious’. I meet the Whole as ‘Thou’. But, since this experience is akin to my meeting with other men, it seems to me to more accurate to describe it ‘ultra-personal’ than to put it in a special category of its own.


If this is a type of ‘religion’, then I can study traditional ‘religion’ and AICN under the same head. I have to exclude ‘Birmingham’ religion. But there is more to it. I suppose I first began to conceptualize my own ‘religion’ in these terms when, in the 1930s, I had some rapport with communist undergraduates. I had to recognize not that they had found the ‘religion’ that the churches had lost: but that ontologically, their position was of the same order as my own. They had not chosen Marx, but Marx, as it were, had chosen them. They were concerned, at root, with the nature and destiny of man. They set out to provide clues not only to all social, but to all natural, understanding. Theoretically, it was possible to say, ‘This is faith. This is ultimate concern. But it has no ‘religious’ reference. At best it is a ‘pseudo-religion’. But this is dangerously near to Mr. Thwackum: ‘When I say religion, I mean the Christian religion. When I say the Christian religion I mean the Protestant religion. And when I say the Protestant religion I mean the Church of England as by law established’. Perhaps, indeed, the only way in which to use ‘religion’ is to give it a precise meaning in terms of a particular institution. But then Islam and Ganda religion fall into the same category, as ‘pseudo-religions’ with Marxism. And no ‘student of religion’ is going to stand for that.


In any case, for one who experiences religion ontologically, these differentia are irrelevant when it comes to definition. I cannot say, ‘I am religious. You are not’ in the same sort of way as I say. “I am white. You are black’. The question is, Are you a man? – even if a different sort of man. Just as colour is accidental to manhood, so the overt differentia are accidental to religion; and existentially, it is as improper to call Marxism a ‘pseudo-religion’ as to call an African a ‘pseudo-man’. But the relation is closer. According to my ‘religious possibilities’, to say, ‘I am a Christian’ or ‘I am a Marxist’ is to say, ‘I am particular sort of man (however unsuccessfully I practice my manhood)’ just as to say (in any ontological sense), ‘I am African’ is to say, ‘I am a particular sort of man’.


When I left Uganda in 1964, this issue was clearly focussed for me and provides another part of the evidence. I believed I was returning permanently to England on Christian grounds (not matter what). But it was clear that an essential part of me had become African: in the sense that, if I had followed deep private inclination and conviction, I should have applied for Uganda citizenship and tried to enter fully into the religio-political life of the country. I said, ‘had become African’ because – like becoming a Christian – this had happened to me. Although the discovery compelled choice, the thing itself had not been chosen. It raised the basic identity question, ‘Am I Christian or am I African?’; and it provoked an identity crisis far more emotive and long-lasting than my evangelical ‘conversion’ at the age of eighteen.


I am of course saying that, according to my ‘religious possibilities’, if I call it ‘religion’ to be a Christian, then I must call it ‘religion’ to be African – not because African have any special capacity for ‘religiosity’ (I don’t believe they have) but because to be African, in this sense, entails ontological commitment. If, as I suspect, the latter equation is etymologically improper, the former is improper too. I would prefer to say that, like the Kabake, they are matters of essential nature; and I suggest that, in any important sense, that is what ‘the study of religion’ is about.




I have recently (1969b) suggested that Evans-Pritchard’s (1937) magnificent analysis of Zande medicine can best be illustrated by a continuum in which the x- and y-axes represent respectively the ‘empirical symptoms’ and the ‘mystical cause’ of the disease. In his Type (i) an acute illness is directly attributable to a mystical cause and can be treated only by action against a mystical cause and disease together. This is plotted (0, y1). In his Type (ii) a mystical cause enables empirical symptoms to attack, and persist in, an individual. The dual causation requires dual treatment; and it is plotted (x2, y2). Type (iii), where the disease is minor and the mystical cause can be ignored, is plotted (x3, 0). But in (iii) the mystical cause is always present in theory. In (i) the symptoms, if not the cause, are empirical and specific drugs are used in their treatment. Diagnosis is in terms of not x, or y, nor of x+y but of xy; and, because every point on the diagram must be defined in terms of both axes, it implies that every event is, in principle, both empirical and mystical at the same time.


These terms are, in fact, very unsatisfactory. On the mystical axis, not only do sorcerers use material means but witchcraft is contained in an identifiable organ of the body. On the empirical axis ‘The souls of the drugs go down into the body of the man and destroy the soul of the disease which is destroying the soul of the organ’ (p. 492). The x-axis is as mystical as the y-axis. Moreover, the concept of ‘soul’ is used to explain not only ‘magic’ but any technological gap between action and result, for instance, ‘the gap between planting (eleusine) and its appearance above ground’. The distinction between ritual and empirical actions is not qualitative but lies in ‘in the number of steps in an activity which are, or are not, subject to observation and control’ (p. 463f).


This conceptual analysis is not, of course, made by the Zande. What they recognize is implied by Evans-Pritchard’s basic definition of the function of witchcraft (p. 67): ‘what they explained by witchcraft were the particular conditions in a chain of causation which related an individual to natural happenings in such a way that he sustained injury’. ‘Natural’ must be understood as the opposite not of ‘supernatural’ nor of ‘mystical’ but of ‘social’ (Durkheim would have seen this point at once). The symptoms of disease can be observed, its natural cause stated, its treatment with specific medicines prescribed. ‘Natural’ causes represent the general character of particular cases, what can be abstracted from them in the interests of a theory of action by direct contact.


On the other hand, belief in witchcraft symbolizes failures in the interpersonal field. They represent what is unique in particular cases, what reveals as to the personal relationships of the patient, whatever of unobservable takes place when two people communicate or fail to communicate. The Zande concept of dual causation insists that, in all particular cases of disease, there is likely to be a social, as well as a natural, factor involved: and that, in some cases, the social factor is dominant.


In contrasting ‘natural’ and ‘social’ I am not contrasting the ‘natural sciences’ with the ‘social sciences’. These disciplines apply, to different fields, essentially by the same methods in which a subject observes an object and, by abstraction and classification reduces particular instances to an ordered general theory. In their applied form they manipulate objects in the interests of another’s purpose. I am, rather, making the same contrast as Laing (1960, pp.19-24) when he wrote:

‘If you are sitting opposite me, I can see you as another person like myself without you changing or doing anything differently, I can now see you as a complex physical, chemical system, perhaps with its own idiosyncrasies but chemical none the less for that: seen in this way, you are no longer a person but an organism. Expressed in the language of existential phenomenology: the other, as seen as a person or seen as an organism, is the object of different intentional acts. There is no dualism in the sense of the coexistence of two different essences or substances there in the object, psyche and soma; there are two different experiential Gestalts: person and organism.’


Laing, as he knew, was writing of the difference between I-Thou and I-It relationships; and perhaps he was wrong to say ‘intentional acts’. But the contrast is clear. In the I-Thou relationship, a subject is face to face with another subject. Abstraction, classification or manipulation immediately turn him into an It. The relationship is unique. But it is not a ‘mystical’ relationship. The concept of ‘psyche’ is used (whether rightly or wrongly is another matter) to describe Its. The I-Thou relationship is essentially that of one person committed to another person (and may exist as much in a hand-to-hand fight as in intellectual encounter). At the same time, every event in an I-Thou relationship can be described, by an observer external to the relationship, at any level of analysis from the sub-atomic to the sociological. Every person is, at the same time, both subject and object. There is no dualism, but a complementarity such as is found in the Zande concept of disease; and no account is complete which ignores either the x- or the y-axis. This implies that every event involving persons, whether in the social or the natural field, has this complementarity. Insofar as it is ‘general’ – whether physical, economic, political, religious or secular – it is proper that an attempt should be made to describe and classify it in It terms. Insofar as it happens to a particular person, it is a unique event. It terms are inapplicable; only Thou terms are possible. Since Thou terms are terms not of description but of inter-personal response, and the existential totality of an I-Thou relationship is in any case inaccessible to objective observation, any attempt to describe must be mythopoeic. ‘I love you’ becomes ‘My true love hath my heart’. Anything analytical is hopelessly inadequate.


One example of such myth-making is precisely the word psyche. I try to describe the difference between an I-Thou and an I-It relationship by creating a duality of substances, while it would be more exact to talk of a duality of ‘intentional acts’. But this also is mythopoeic, since the ‘intentional acts’ in their totality are unobservable. Much confusion has been caused by supposing that ‘personification’ is the same as ‘psychomorphosis’ – that to say ‘I believe in God (or the living God, or witchcraft)’ is to say, ‘I believe in extra-somatic psychic forces’. The whole western tradition endorses that equation. My understanding of the Zande (and of the Hebrew) material suggests that it is not a necessary equation. Neither of them can think of psyche and soma in separation. An alternative translation would be, ‘I respond to the whole material universe, or to certain events in it, in ways akin to my response to persons – sometimes as Thous, sometimes as Its. But I believe the I-Thou response to be basic.’. There is no necessary connection between belief in God and ‘psychic experiences’ – whether mysticism, ghosts, ESP, LSD trips or psychotic incidents. In an important sense, if these are ‘religion’ or its analogues, then belief in God is not ‘religion’.


Another source of confusion is the dualism of the either-or attitude which regards descriptions on the x- and y-axes as alternative instead of as complementary. ‘An object is either a person or a thing’, whereas in I-Thou language any subject may be treated as an It, any object emerge as a Thou. ‘A miracle is a suspension of natural law’, whereas in the language of complementarity every event can be fitted into a scientific system, but any event may emerge as the medium of God’s self-revelation. ‘Religion should have nothing to do with politics’, whereas in ontological terms every ritual and every act of social control must be plotted with reference to both axes. Rituals do in fact (in Western as well as Zande culture) have empirical consequences; and independent church, insisting that it is ‘purely religious’, is in fact political because Government regards its very existence as potentially subversive.


But this two-dimensional continuum is not enough. Every event (including pure I-Thou events) has a history. History may be used, as historiography, to deepen understanding of events along the x-axis. Or it may be used, as charter myth, to enhance commitment along the y-axis. Any adequate ontological description requires a z-axis, along which is plotted ‘historical awareness’. This of course provides a standard difference between Christianity and Marxism on the one hand and ‘nature religions’ on the other. But I think it is an important factor in understanding the differing responses to the West of, for instance, the Gikuyu and the Ganda. The former, with a developed creation myth and intense awareness of the present activity of God – but little awareness of intervening history – selectively accepted the West (including Christianity) only as a means to defeating it and were well aware of the ambivalence of their attitude. The latter, with no creation myth – but a detailed oral history wound around the material memorials of the kabakas – were able to see the whole of the western offering as fulfillment of tribal destiny. Again, many contemporary African historians – however rigorously they follow the criteria of historiography – are frank about their primary concern to discover the historical roots of their commitment as Africans. In classical psycho-analytical terms, the ego, engaged in the empirical task of adjusting to reality, is guided by both the super-ego, symbolizing society, and the id symbolizing individual history.


If, then, a ‘student of religion’ is concerned with ‘essential nature’, he must be aware of all three dimensions. The y-axis is, by definition, not susceptible of study in orthodox objectivity. It is there to ‘transpose himself into’ in order to provide the only standpoint from which to understand its x-coordinates. Similarly, he must transpose himself into charter myth. Historiography then becomes an x-axis activity supplementing his studies as economist, philosopher, political scientist or sociologist. Precisely because he is concerned with the whole, he cannot fail to be one of these. He will different from the general run not because he specializes, for instance, in ‘the sociology of religion’ but because he insists that no sociology can be complete which does not pay attention to the y- and x-axes. I am not saying that the study of ‘religious institutions’ – or its analogue in other disciplines – is unimportant. I am saying (i) that a study of ‘Birmingham’ religious institutions does not contribute anything significant to an understanding of ontology; and (ii) that a ‘study of religion’ is gravely incomplete which does not take account of the y- and z-coordinates of all institutions.



I would go further and say that the contemporary isolation of ‘religion’ as a separate subject of study is methodically wrong. African independent churches ought to be studied not as ‘religious’ phenomena but as examples of social schism in whatever form. Political parties cannot be adequately understood without reference to their charter myths and their Thou commitments. Unitary societies have to be analysed in terms applicable to communism and nationalism but not to plural societies. African witchcraft beliefs are strictly analogous not to contemporary British covens but to our attitude to coloured immigrants. African cults of the living dead are comparable not with western spiritualism but with Churchill memorials and the Patrice Lumumba University. ‘Spirit-possession’ is matched by pop sessions, exo-psychic mythology by Freudian concepts (Christological controversies by arguments about the nature of the ego).


The rising generation in this country has little use either for history or for the traditional myths in which Thou experiences have been expressed. It thinks it is concerned with what is practical. But so were the Zande. What for Evans-Pritchard were myth and ritual were ‘practical’ for them precisely because they did not make the distinction. Myth, ritual, and indeed ontological commitment, are always most effective when they are not recognized as such. If it is the latter which we want to study, we have to be able to stand sufficiently outside it to talk about it in x-terms. We can do so, to some extent, in the case of Nazism and of what little we know of Maoist China. We ought to be able to do so – at least to the extent that their y-function is vastly greater than their x-consequences – in the case of student demonstrations, general elections and academic colloquia in this country. If we concerned with contemporary British ontology – with what in contemporary Britain is most akin to ‘traditional religions’, what is needed is a massive study of unrecognized commitments as they are expressed not only in unrecognized myth and ritual but in all x-axis activities (Christmas is not Christmas without the Queen’s message). The study of ‘religious institutions’ and of ‘religious experience’ can be left on one side as largely irrelevant to what we want to know. This, then, ought to be the major emphasis of departments of Religious Studies. We shall have to continue to call it ‘religion’ in order to get it past faculty boards and UGC. But we shall not be doing what they think we mean.