RELIGION AND RITUAL: THE DEFINITIONAL PROBLEM Jack Goody volved in categorizing actsand beliefsas religious, or ritual, or magico- religious, with the purpose not only of clearing the way for subsequent treatment of my own empirical data concerning the LoDagaa of North- ern Ghana, but also of clarifying certain aspects of the analysis of social systems in general. For some writers such an investigation has appeared a profitless enterprise. At the beginning of Themis: A Study of the Social Origins of Greek Religion, a book which as its subtitle suggests owes much to the work of Durkheim as well as to the English anthropologists, Jane Harrison comments on the erroneous approach of those inquirers who start with a general term religion of which they had a preconceived idea, and then try to fit into it any facts that come to hand. Instead she proposes no initial definition, but remarks that ‘we shall collect the facts that admittedly are religious and see from what human activities they appear to have sprung’ (1912: 29). It is yet more tempting for the in- quirer into societies farther removed from our own tradition than that of ancient Greece to adopt a similar approach, and quietly to overlook the definitional problems. The dangers, however, outweigh the advantages. In refusing to define her field of discourse, Jane Harrison was far from escaping the problem she perceived; she was merely taking refuge in an implicit rather than an explicit judgment of what constitutes the ‘ad- mittedly religious’. It goes without saying that such hidden decisions may influence the investigation of particular events. It might be possible to examine the mortuary institutions of the LoDagaa without raising the wider issues were it not that the analysis of the specific data must depend, to some extent at least, upon the position which the investigator takes with regard to them. Moreover, the difficulties which arise from a failure adequately to delimit one’s universe of discourse become much more complicated when comparative studies are involved. With these matters in mind, therefore, I shall try and deal with some of the general problems connected with the examination of what have variously and 142 [= THIS PAPER I want to explore the problem of what is in-| | | | | RELIGION AND RITUAL somewhat indiscriminately been described as ritual, ceremonial or religious phenomena. In attempting to clarify these concepts for sociological purposes we are not trying to arrive at the fundamental meaning conveyed by the English term ‘religion.’ Not that we are indifferent to what Bohannan has called the folk-categories of European societies. They form the inevitable starting point from which we develop our analytic concepts. But normally they cannot themselves serve as such. In all branches of com- parative social science this process of defining adequate categories has given birth to polemical problems of considerable magnitude, as witness the discussion which has arisen about the nature of the family or of legal, political, and economic institutions. And the progress of studies of kin- ship systems, to take one such example, has to a significant degree depended upon distinguishing among the various connotations of the contemporary English concept of the ‘family’ in the light of investiga- tions in other societies and then giving more restricted technical mean- ings for this and other words when they are used for comparative analysis. The results of such endeavours may make ethnological reports more tedious to the general reader. While this is to be regretted, it can be avoided only at the expense of the development of the study of human institutions. In this analysis of the various approaches to the definition of religious and ritual phenomena, I shall begin with the nineteenth-century con- tributions of Tylor and others who followed the same general direction of interests. I will then consider the views of Durkheim, Malinowski and of some later writers, especially Talcott Parsons. The latter’s perceptive treatment of the main issues will serve as a guiding thread throughout the argument. But although his discussion is most helpful, he arrives at a position held by a number of other recent writers in this field which, in my opinion, places too much weight on the usefulness of the distinction between the sacred and the profane, a deceptively simple dichotomy that has had a distracting effect on the development of a comparative socio- logy of magico-religious institutions. But before elaborating this state- ment, let us return to the beginning. When Tylor writes of the cult of the dead as central to the develop- ment of religion, his meaning is clear because he puts forward a mini- mum definition of religion, the belief in Spiritual Beings, that is, anim- ism. This formulation was attacked from two main directions. In the first place, early reports of the beliefs of non-literate peoples described con- cepts pertaining to mystical forces of a non-personalized kind; typical of these were the mana of Melanesia and the wakan of Dakota. Marett pointed to the similarities between these ideas, which he referred to as animatism, and the animistic beliefs employed by Tylor as the differentia of religion. Although Marett regarded both animism and animatism as in themselves non-religious, regarding emotive factors as critical in this 143JACK GOODY respect, his contribution had the effect of blurring the previous dis- tinction between magic and religien and led to the adoption of com- promise terms such as ritual, sacred, non-logical, or even magico- religious to designate the domain formerly occupied by the non- scientific elements in the tripartite division of the world of belief into magic, religion and science employed by the earlier writers. While Marett worked from basically the same starting points as Tylor and Frazer, the second main objection came from a radically different direction. To the general aspects of Durkhcim’s thesis I shall return later. His specific criticism of Tylor’s own definition relates to the ques- tion of Buddhism. Here, he claimed, was a set of practices and belicfs, usually considered to be one of the great world religions, and yet described by one authority as ‘a frankly materialistic and atheistic interpretation of the universe’. To include Buddhism some alternative formulation had to be devised. Starting from a standpoint put forward by Robertson Smith, Durkheim developed the thesis that all peoples recognized a radical dichotomy of the universe into the Sacred and the Profane. In accordance with this proposition, he offered his famous definition of religion as ‘a unified system of beliefs and practices relative to sacred things, that is tosay, things set apart and forbidden—beliefs and practices which unite into one single moral community called a Church, all those who adhere to them’ (1947: 47). There have been other attempts to define the sphere of religious phenomena, by reference, for example, to emotive criteria such as feelings of awe. But efforts to isolate specifically religious experiences in this way have proven of little value to investigators in other societies. Evans-Pritchard, for example, has recently written: ‘Certainly one cannot speak of any specifically religious emotion for the Nuer’ (1956: 312). In general, those pragmatically concerned with such questions have tended to adopt either the inclusive approach proposed by Durk- heim or else the exclusive definition offered by Tylor. One reason for this is that certain aspects of Marett’s objections have received little support from subsequent writers, and not only because of his introduc- tion of emotive criteria. Malinowski, for example, has denied the re- lationship between magic on the one hand and animatistic beliefs of the mana variety on the other. In support of his argument he quotes a Dakota text to the effect that ‘all life is wakan’ and contrasts this ‘crude meta- physical concept’ with the more specific attributes of Melanesian magic . . . ‘there is little in common between the concepts of the mana type and the special virtue of magical spell and rite’ (1954: 77-8). However, other writers have accepted Marett’s perception of a continu- um of personal and impersonal supernatural powers, and, unlike him, have included both these spheres in the domain of the religious. Ac- cording to this view, Tylor’s minimal definition would have to be re- worded to run: ‘a belief in spiritual [or supernatural] agencies.’ 144RELIGION AND RITUAL The main difficulty here is in distinguishing between supernatural and natural, or spiritual and nonsspiritual agencies. With ‘beings’ conceived as concrete entities on the human model, the distinction is possible to handle. But in dealing with non-human agencies and mystical powers, it is in many cases difficult to say whether the concepts are more akin to the physicist’s force or to the Bergsonian éan vital. Or again the concept may span both the pragmatic-scientific and the philosophical- religious poles of meaning— The force that through the green fuse drives the flower Drives my green age; That blasts the roots of trees Is my destroyer. (DYLAN THOMAS) An example of such an undifferentiated concept is the LoDagaa notion of ‘medicine’, ti, which has its counterpart in many other African societies. The term is used as readily for European medicines which have an empirical effect as for other concoctions such as love-potions which do not; it is applied to many different types of powder and includes gun- powder as well as the dried roots eaten to enable the hunter to shoot straight and true. Of the two main definitions of religion which we have noted, the exclusive and the inclusive, it is the latter which has had the wider circulation, For it is implied by Marett’s thesis, explicitly proposed by Durkheim and subsequently incorporated in the work of Radcliffe- Brown and his pupils, which has been a major contribution to the comparative study of religious phenomena. Another effective source of diffusion has been through the writings of Talcott Parsons, who has utilized this definition as a basic reference point in his analytic schema. I shall therefore begin by considering the implications of adopting the extensive viewpoint as developed by Durkheim, paying particular attention to its implications for the study of the practices and beliefs associated with death. Durkheim’s own definition contains two elements: religion consists of beliefs and practices relative to sacred as distinct from profane things. But so, he says, does magic. To distinguish between these two spheres he introduces a second criterion; here again he relies on Robertson Smith who had said that magic is opposed to religion as the individual to the social. Religion is public and has a Church, whereas the magician has only his clientele, never a congregation. The contrast between the in- dividual practitioner with particularized relations to his clients, a primitive doctor working with his ‘bastard science’, and the sacerdotal leader, the master of ceremonies, is one of some importance in the analysis of many social systems. But the distinction between these private and public roles, though related to other facets of religion and 145JACK GOODY magic, offers a less than adequate focus for the definition of religious phenomena. Indeed in an excellent discussion of this problem in rela- tion to the Murngin material, Warner notes that magic too has its Church, in that the effects of both good and bad magic depend to a considerable extent upon the commitment of the clients to a belief in the efficacy of the procedures they employ (1937: ch. 8). Consequently, the magician and his clientele also constitute a certain kind of moral com- munity, which though not as explicit as in public ceremonials may be both morphologically and functionally very similar. When Durkheim maintains that magic has no Church, he means that it does not bind people together in the same way as do religious cere- monies; and the epitome of solidary procedures he sees in the physical assembling of the members of the group, i.e., in mass ceremonials. It is of course true that assemblies of this kind may and do play an important part in re-affirming certain of the central institutions of a society; this is the case with coronation services, with Red Army parades and with July 14th celebrations, Nevertheless, in making this the differentiating feature of religious action, Durkheim is tending to fall into thé same error which confuses his early treatment of the problem of the individual and society. Initially at least he appears to confuse two sorts of distinction, on the one hand that between society as a collectivity of human beings and the individual as a single human entity, on the other that between society as the social element within the human personality and the individual as the organic or instinctual element. Eventually he settled for the latter distinction,t but not before he had sown the seeds of con- fusion for future readers. In defining religious phenomena, he con- centrates once again upon the fact of the group assembling or, as some social anthropologists would use the term, tts corporate features.? At one time he remarks that the reason for a group getting together is relatively unimportant as compared with the fact of its assembling. But it is one thing to stress the importance of convening groups— the two meanings of convention, an assembly and a custom, are not accidental. It is yet another to equate such groups with ‘moral com- munities’, as Durkheim tends at times to do; a congregation may be less ‘solidary’ and therefore less ‘moral’ for being dispersed, but as Warner’s comments show, the existence of common norms, which is surely the only possible interpretation of the phrase ‘a moral community’, does not necessarily depend upon the existence of general assemblies. To make such an assumption is to fall into the same error as Durkheim does in his reference to Robertson Smith (1947: 45, n. ii), namely, that of confusing the public and the social. Although in any particular society there may be a great turnover in magical procedures, indeed the combination of an empirical end and a non-empirical means makes this to some extent in- evitable, magic is no less a social phenomenon, in the strict sense, than religion. Sorcery, for example, depends for its effects upon a certain 146RELIGION AND RITUAL degree of consensus, upon the acceptance of a set of social norms by a significant proportion of the members of a society. Though in one sense ‘anti-social’, it is pre-eminently ‘social’ in another, for the persistence of the belief in the attainment of pragmatic ends by non-empirical means depends entirely upon. non-environmental factors; there is no direct reinforcement in the external, extra-human world. The tendency to embody the criterion of simultaneous face-to-face interaction in the concept of social groups in general and of religious groups in particular gives rise to certain inevitable difficulties, which Durkheim himself partly foresaw. He had to give special consideration to the position of beliefs in a personal destiny, cults of guardian spirits, and other non-assembling forms of ‘supernaturalism’. His treatment of these is not altogether happy and as a consequence specialists in the religion of American Indians such as Radin, Lowie and Goldenweiser were led toundervalue his overall contribution to the study of religion. In reaction to his formulation of the problem they followed Marett in try- ing to establish various emotive criteria of religious phenomena, a pro- cedure which subsequent investigators have found of little value. What Durkheim did was to admit these ‘private’ cults as truly religious phenomena, but only by relating them to some more inclusive religious system. So the individual cults are seen solely in the context of the collective cult or Church, and a Church he tended to think of not only as the members of a faith, the congregation, but as a group assembling together in one place, the congregation physically united into one body. While these large-scale, face-to-face relations are of great importance, especially with regard to the effervescence which Durkheim associated with religious activities, they are neither the only sort of social relation- ship to produce this type of solidarity, nor can they be satisfactorily identified with religious activities as such, with sacred procedures in general. Indeed, what Durkheim is talking about in most of his analysis in The Elementary Forms of the Religious Life can more precisely be called ceremonial. Not infrequently he allocates to religion the functions and properties which might more properly be assigned to a phenomenon of greater generality, mass ceremonial. Up to this paint I have been considering the adequacy of the criteria Durkheim employed to distinguish religion from magic, namely its association with a Church in the sense of a moral community. I now want to turn to the way in which he attempted to differentiate the whole sphere of magico-religious acts and beliefs by reference to the dichotomy between the sacred and the profane. This sphere is the one referred to by Radcliffe-Brown as ‘ritual’, a term upon which it is necessary to offer some explanatory comment. Generally the term has been used to refer to the action as distinct from the belief component of magico-religious phenomena. But the word is also employed in an attempt to avoid the distinction made by earlier writers between magic and religion. We have 147JACK GOODY seen that for Durkheim religious rites are obligatory within a religious society of a Church, while magical rites are optional. To fail in the observance of the one is sinful, while to omit magical rites is to incur only bad luck. As far as Malinowski was concerned, a magical rite has a definite practical purpose which is known to all who practise it and can be easily elicited from any native informant, while a rite is religious if it is simply expressive and has no purpose, being not a means to an end but an end in itself. This is the basis of Parsons’ classification of rituals as ‘religious in so far as the goal sought is non-empirical, magical, so far as it is empirical’ (1951: 375). While not denying the possible theoretical interest of these and other differentiations between magical and religious activities, in his paper on Taboo (1939, reprinted in 1952) Radcliffe-Brown tries to avoid the ambiguities involved by employing the term ‘ritual’ to cover magico- religious phenomena in general (1952: 136-9). Thus he speaks of ‘ritual values’ in addition to ritual actions. It is the entire compass of ritual or magico-religious activities which Durkheim associates with the sacred as distinct from the profane, a dichotomy which he describes in the following way: ‘All known religious beliefs,’ he maintains, ‘whether simple or complex, present one com- mon characteristic: they presuppose a classification of all things, real and ideal, of which men think, into two classes or opposed groups, generally designated by two distinct terms which are translated well enough by the words profane and sacred’ (1947: 37). This classification is in all aspects a relative one. ‘The circle of sacred objects cannot be determined, then, once for all. Its extent varies infinitely, according to the different reli- gions’ (1947: 37)- It is important to note that Durkheim conceives the sacred-profane dichotomy to exist within the actor frame of reference; he claims to be dealing with concepts which are present in all cultures, which are meaningful to the people themselves. It is for this reason that his defi- nition plays such a central role in Parsons’ valuable discussion of the theoretical convergences in the sociology of religion. One of the most important aspects of this development is the agreement that ‘situations must be subjectively defined, and the goals and values to which action is oriented must be congruent with these definitions, must, that is, have “meaning” ’ (Parsons 1954: 209-10). Parsons sees this position as con- sistent not only with Weber’s view of understanding (Verstehen) in the social sciences, but also with Malinowski’s claim that the inhabitants of the Trobriand Islands recognize the distinction between technological, magical and religious acts. Indeed he considers Tylor and Frazer as part of the same line of intellectual development because the ‘rationalistic variety of positivism’ was marked by ‘the tendency to treat the actor asif he were a rational, scientific investigator, acting “reasonably”, in the light of the knowledge available to him’ (1954: 199). While it is true that 148RELIGION AND RITUAL Tylor and Frazer rightly or wrongly attributed an intellectualist origin to religion, this was surely the result of their devotion to the categories and ways of thinking current in European society in the late nineteenth century rather than an adoption of the actor point of view, to which they had access only through the reports of missionaries and travellers. However this may be, it is clear that their point of departure was less relative than that of Durkheim, whose one fixed point is the universal ‘duality of the two kingdoms’ of the sacred and the profane. But let us turn to the empirical evidence and ask whether the dicho- tomy is in fact a universal or even a recurrent feature of the actor frame of reference. For although Durkheim so phrased his argument that the discovery of any society which did not recognize the division could be used in refutation, just as he employed Buddhism to reject Tylor’s minimal definition of religion, we are not here concerned with the question of the universality of the phenomena as much as with the elucidation of useful analytic concepts. A major difficulty immediately presents itself. If the dichotomy is really as relative as Durkheim main- tains when he speaks of infinite variations, then it is clearly not easy to decide what to look for. Many of the writers who have adopted this approach equate the profane and the sacred with ‘normal’ on the one hand, and with ‘things set apart and forbidden’ on the other, as Durk- heim did in his original definition. But does not this take us far outside the bounds of the admittedly religious, to use Jane Harrison’s phrase? In the absence of any objective criteria might we not equally well fix upon any dichotomy a particular people make, ‘good’ and ‘bad’, ‘black’ and ‘white’, ‘day’ and ‘night’, and declare that this constitutes the equivalent of sacred and profane? The empirical difficulties may be illustrated from two major dis- cussions of this problem, which treat data on a particular society in the context of general theory. I refer to Malinowski’s examination of Tro- briand magic and Evans-Pritchard’s account of Azande witchcraft. In Parsons’ delineation of the main trends in the sociology of religion, Malinowski is of crucial importance as demonstrating the existence, within the actor frame of reference, of the dichotomy between ritual and non-ritual phenomena. His comment runs as follows: Side by side with this system of rational knowledge and technique, how- ever, and specifically not confused with it, was a system of magical beliefs and practices. These beliefs concerned the possible intervention in the situation of forces and entities which are ‘supernatural’ in the sense that they are not from our point of view objects of empirical observation and experience, but rather what Pareto would call ‘imaginary’ entities with a specifically sacred character, Parsons concludes that ‘this approach to the analysis of primitive magic enabled Malinowski clearly to refute both the view of Lévy-Bruhl, that 149JACK GOODY primitive man confuses the realm of the supernatural and the sacred with the utilitarian and the rational, and also the view which had been classically put forward by Frazer that magic was essentially primitive science, serving the same fundamental functions’ (Parsons 1954: 202-3), Malinowski, however, though he certainly maintained that magical acts were recognized as such by the society, did not feel he had rejected Frazer in his first report. Of the terms magic and religion, he writes in Argonauts of the Western Pacific ‘. . . although I started my field work con- vinced that the theories of religion and magic expounded in the “Golden Bough” are inadequate, I was forced by all my observations in New Guinea to come over to Frazer’s position’ (1922: 73n.). The difficulties of reconciling this acceptance of Frazer, for whom magic was bastard science, with a use of Durkheim’s dichotomy, in which magic falls within the realm of the sacred, makes the reader wonder what Malinow- ski means when he asserts that the Trobriand Islanders themselves recognize a distinction between the world of the sacred and the world of the profane. In the above quotation, Parsons, I think rightly, interprets this as a recognition of the distinction between the realm of the super- natural on the one hand and the utilitarian and the rational on the other. But that this is far from a universal distinction in non-literate societies is illustrated by the most comprehensive account we have which bears on this particular problem, namely Evans-Pritchard’s treatment of Azande witchcraft (1937). Here the author accepts for heuristic purposes the distinction between ‘ritual and empirical actions by reference to their objective results and the notions associated with them’ (1937: 463). But he emphasizes the difficulties raised by this acceptance: Even by the definition of ‘magical’ and ‘empirical’ adhered to in this book it is not always easy to classify a simple act as one or the other. A man burns a piece of bark-cloth and, holding some magical plant in one hand, blows the smoke into the opening through which termites come out of their mounds when they swarm after rain. This is said to encourage them to come out. Azande say that the bark-cloth is termite-medicine, but they are probably speaking metaphorically (1937: 464). Again: Azande offer the same explanation of a ‘soul’ acting to produce certain results in those technological activities in which there is a similar gap between action and result to the gap in magical techniques, a gap where nothing can be seen of what is happening—e.g., it is the ‘soul’ of the eleusine which accounts for the gap between planting of the seed and its germination and appearance above the ground (1937: 464). Evans-Pritchard denies that the Azande make the supernatural- natural distinction in the European sense, and there appears to be no confirmation of Malinowski’s claim that the Trobriand Islanders make 150RELIGION AND RITUAL the same distinction as Europeans between technological and ritual acts. In Evans-Pritchard’s discussion of leechcraft the lack of correspondence is evident; ‘natural’ and ‘supernatural’ forces as we conceive them are inextricably intertwined (1937: 478-510). ‘In every primitive com- munity, studied by trustworthy and competent observers,’ writes Malin- owski, ‘there have been found two clearly distinguishable domains, the Sacred and the Profane; in other words, the domain of Magic and Religion and that of Science’ (1954: 17). But it would seem from the Azande material that this is not the case; whatever differing attitudes exist towards the poles of technological and ritual activities as we see them, there is insufficient evidence to establish a universal conception of ‘two clearly distinguishable domains’. When Malinowski declared that the Trobriand Islanders clearly differentiated the magical from the technological aspects of canoe-building, he seems to have meant that they make a distinction between the situations in which they employ a certain type of verbal formula we would call a spell and those situations in which they do not. This is quite a different sort of distinction. What Evans-Pritchard reports for the Azande holds equally well for the LoDagaa. Among these people, there is no recognized distinction between the natural and the supernatural and, as Durkheim himself noted, this seems to be the case for most cultures. But neither do the LoDagaa appear to have any concepts at all equivalent to the vaguer and not unrelated dichotomy between the sacred and the profane which Durkheim regarded as universal. However, although the absence of correspondences in the cultural equipment of the people themselves weakens Durkheim’s case, and has significant implications for certain aspects of Parsons’ action schema, it does not necessarily destroy the utility of these categories as analytic tools if it proves possible to isolate objective criteria for their use. Can this be done? Caillois begins his book, Man and the Sacred, (1939) with the words: ‘Basically, with regard to the sacred in general, the only thing that can be validly asserted is contained in the very definition of the term—that itis opposed to the profane’ (1959 ed.: 13). It isan indication of theextent to which the dichotomy has been uncritically adopted that a recent writer on comparative religion who takes the subject of his study to be ‘hierophanies’ or manifestations of the sacred, starts by quoting this remark with apparent approval (Eliade 1958: xii). Adequate as this may be for theological purposes, it is hardly sufficient as an analytic tool of comparative sociology. And Durkheim himself, despite his statement concerning infinite variations, offers something more concrete than this. Parsons notes that, negative features apart, Durkheim introduces two positive characteristics of ‘ritual’, that is, of acts oriented to sacred _ things, Firstly, there is ‘the attitude or respect . . . employed as the basic criterion of sacredness throughout’. Secondly, ‘the means-end relation- ship is symbolic, not intrinsic’ (1937: 430-1). Although ‘respect’ is often 151JACK GOODY required by performances of a religious or ceremonial kind, it is by no means an invariant feature of such activities. Moreover, such an attitude is perhaps equally characteristic of authority relationships within the family, which could be designated ‘sacred’ only by an overly loose interpretation. The second criterion, the symbolic rather than intrinsic nature of the means-end relationship, is no less problematical. The first question that arises in any discussion of symbolic relationships is the level of analysis on which one is operating. Considerable confusion has been caused by the failure to be clear as to whether a certain act or object is symbolic for the actor, or for the observer, or for both. In Par- sons’ schema, the symbolic relationship should exist within the actor frame of reference. But when it comes to dealing with the symbolic nature of ritual, the actor frame of reference is partly set on one side and the method of interpretation is likened to that of the psychoanalyst. (1937: 419 ni.) In other words, the reference of the sign is necessarily hidden from the actor. This would also appear to be true of the Durk- heimian formulation, according to which the symbolic reference of ritual, anyhow religious ritual, is ‘society’ or, in Parsons’ explanatory phrase, ‘the common ultimate-value attitudes which constitute the specifically “social” normativeelement in concretesociety’ (1937: 433-4). Radcliffe-Brown, whose views on this question also derive from Durk- heim, writes in a similar vein of the symbolic nature of what he calls ‘ritual acts’. ‘. ritual acts differ from technical acts in having in all instances some expressive or symbolic clement in them’ (1952: 143). The term symbolic is explained in the following way: ‘Whatever has a meaning is a symbol and the meaning is whatever is expressed by the symbol.’ For Radcliffe-Brown the meaning ofa rite is variously come by. At times the determination of meaning comes close to the attribution of social effects or social function, and the present passage continues: ‘the method ... I have found most profitable during work extending over more than thirty years is to study rites as symbolic expressions and to seck to discover their social functions’ (1952: 145). At other times he speaks of meaning being determined by the system of ideas with which a rite is associated. On other occasions the symbolic referent is the ‘social structure, i.e, the network of social relations . . .” (1952: 144); at others, objects and actions of social importance. The formulations of Parsons and Radcliffe-Brown are not identical, but both are agreed that ritual is essentially expressive or symbolic in nature. And in each case the interpreter of the symbolic relationship turns out to be the observer rather than the actor. Radcliffe-Brown specifically rejects the explanation of ritual in terms of the conscious purposes of the participants; meaning is not to be found on the surface, though he claims rather lamely that ‘there is a sense in which people always know the meaning of their own symbols, but they do so intui- tively and can rarely express their understanding in words’ (1952: 143). 152RELIGION AND RITUAL As he himself realizes, once this fact is recognized the attribution of meaning or social function to a ritual raises serious problems of evidence. As for the determination of meaning with reference to ultimate values, the situation is even less clear. What are usually referred to as ‘values’ by social scientists are not the specific desiderata evinced by members of a society in their actions or in their beliefs. They are not what R. B. Perry meant by ‘any object of any interest’, nor Charles Morris by ‘preferential behaviour’, but high-level abstractions from such observational data. While writers like Parsons and Kluckhohn concentrate upon what they call ‘value-orientations’ rather than ‘values’, a yet more abstract con- cept. The difficulties involved are brought out when Parsons and Shils define the related pattern variable concept as ‘a dichotomy, one side of which must be chosen by an actor before the meaning of a situation is determinate for him, and thus before he can act with respect to that situation’ (1952: 77). It is difficult to see how this can be so when the actual list of pattern variables is examined; for example, the alternatives of specificity-diffuseness, useful as they may be to the sociologist in- volved in categorizing preferential behaviour, hardly represent concrete choices in the life of an Iowan farmer, let alone a Fulani herdsman to whom the formulation of such concepts would present problems of much greater complexity. Indeed, Kluckhohn in the same volume admits that the schemes put forward by Parsons and Shils, by F. Kluckhohn and by himself are ‘all analyses from an observer’s point of view and with a minimum of content’ (1952: 420-1). An essential preliminary step, he remarks, is to get the ‘feel of the culture’ from a careful reading of classical ethnographies. It might be added that, given the apparent divergencies in the sensory equipment of investigators, even the process of getting the ‘feel’ of a culture contains such a number of imponderables that the wary may well be daunted from proceeding to this further degree of abstraction. The contention of Parsons that the symbolic reference of religious ritual is the common ultimate-value attitudes of a society appears to lie behind the attempt of some anthropologists and sociologists to use Tillich’s definition of religion as ‘man’s attitudes and actions with respect to his ultimate concern’. For example, Bellah, in his analysis of Tokugawa Religion, writes: This ultimate concern has to do with what is ultimately valuable and meaningful, what we might call ultimate value; and with the ultimate threats to value and meaning, what we might call ultimate frustration. It is one of the social functions of religion to provide a meaningful set of ultimate values on which the morality of a society can be based. Such values when institutionalized can be spoken of as the central values of a society (1957: 6). Lessa and Vogt have an identical starting point and make a similar comment ‘. . . religion is concerned with the explanation and expression 153JACK GOODY of the ultimate values of a society . . .’ (1958: 1). The utility of so vague and general a formulation is open to doubt. Quite apart from the ques- tion of the operation involved in specifying ‘ultimate’ or ‘central’ values, this definition would include all purely ‘rational’ pursuits in the econ- omic or political sphere that were of major interest to the members of a particular society. Parsons himself notes that this definition diverges considerably from common usage; it is perhaps significant that in their pragmatic treatment of religious phenomena the above authors adhere much more closely to the ‘traditional’ sphere of discourse, As the positive differentia of sacred acts, those acts which according to Durkheim define the sphere of religion, appear inadequate for our purpose, let us examine what Parsons calls the negative criterion of the definition, namely, the assumption that this category of practices is one which falls outside the intrinsic means-end schema. The means-end schema is intrinsic, according to Parsons, when the means brings about the end by processes of scientifically understandable causation. By ‘falling outside the intrinsic means-end schema’, Parsons does not merely intend to imply, as the nineteenth-century rationalists did, that such beliefs and practices were automatically ‘irrational’. He recognizes a third type of action which is neither rational nor irrational (or pseudo- rational), but non-rational, or ‘transcendental’; that is, it has no prag- matic end other than the very performance of the acts themselves, and cannot therefore be said either to have achieved, or not to have achieved, such an end. This tripartite distinction follows Pareto’s classification of social action into (1) acts which meet ‘logico-experimental’ standards, (2) acts based on ‘pseudo-scientific’ theories either through ignorance or through error, (3) acts based upon ‘theories which surpass experience’ (Parsons 1937: 429 ff.). It is within this last category, he holds, that religious practices fall. Parsons sees this classification as falling within the actor’s own frame of reference; it refers, he writes, to ‘the cognitive patterns in terms [of] which the actor is oriented to his situation of action’ (1954: 200). How- ever, we have already seen that the testimony of a number of expert observers contradicts the view that all non-European societies normally make a distinction of this kind between transcendental and pseudo- scientific theories. If we accept this other view, then neither the negative nor the positive criteria of the ‘sacred’, as employed by Durkheim, constitute a real departure from the distinction made by the nineteenth- century rationalists between science, magic and religion. In this per- spective, the search for universal categories corresponding to the sacred and the profane appears as an unsuccessful attempt to break loose from the earlier position, unsuccessful because of the difficulties involved in taking categories which are ultimately defined by reference to logico- experimental methods and trying to find their equivalents in non- literate societies. 154RELIGION AND RITUAL The point I am making was seen by Nadel in his study, Nupe Religion. There he writes that we have to ‘judge the transcendentality of things by our own way of thinking’ (1954: 3); the separation of the natural from the supernatural can have a precise meaning only in our own system of thought. Indeed, as he rightly observes, ‘the very conflict between supernatural and empirical knowledge on which we base our judge- ments is likely to be absent in a primitive culture’ (1954: 4). Evans- Pritchard provides evidence that this is so among the Azande and also maintains that the separation into natural and supernatural worlds is not to be found in Nuer thought. Among the LoDagaa too, as I have already remarked, the spheres of the technological and the mystical are not clearly differentiated, either in terms of the ends sought or of the means employed. The implications of this argument run counter to the Weberian insistence on Verstehen analysis in the social sciences, analysis in terms of the subjective aspect of action, its meaning to the participants them-. selves.® Indeed, Nadel himself feels uncomfortable at this discrepancy and hastens to assert, like Durkheim in similar circumstances (1947: 26), that while simple societies do not dichotomize in the same way as we do, in terms of credibility, nevertheless they do differentiate between empirical and transcendental. The evidence for this, he declares, must be lin- guistic. But having delineated a Nupe dichotomy between ‘knowledge’ (kpeyé) and ‘ritual’ (kuti), he then undermines his own position by the caveat that: ‘It is probably futile to expect that the native language should always provide precise distinctions of the kind we require, words like “normal” and “non-normal”, “miraculous” or “superhuman”, “sacred” and “profane” ’ (1954: 6). So far this paper has been devoted to a critical examination of existing usage of the concepts of religion and ritual, particularly as developed by Durkheim and those who have followed him. Before attempting to suggest more concrete ways of employing what have become vague, generalized and ambiguous terms, let me recapitulate the argument so far. I have been primarily concerned to reject Durkheim’s assumption that the sacred-profane dichotomy is a universal feature of people’s views of the human situation. The acceptance of this contrary stand- point means that it is no sounder for the observer to found his cate- gorization of religious activity upon the universal perception by human- ity of a sacred world any more than upon the actor’s division of the universe into natural and supernatural spheres, a contention which Durkheim had himself dismissed. The enquirer into the field of religious behaviour is therefore placed in the same dilemma as the student of political institutions. In Western European society we perceive, and participate in, certain organizations which are labelled political, economic, educational, and so forth. In non-European societies the organizations are usually less differentiated; one single system of social L 155JACK GOODY groups may perform a greater variety of functions and the observer is then faced with the problem of stating which of the institutionalized activities carried out by these latter groups correspond to those charac- terizing the differentiated organizations of advanced societies. In other words, he is forced to develop analytic tools out of our own folk-categ- ories. In so doing he has an obligation to respect both our own folk- categories, which form the basis of his analytic concepts, and the folk- categories of the actors themselves, which provide the raw material to which this conceptual apparatus is applied. On the other hand, he cannot be required to limit himself to ‘meaning’ as the actor himself perceives it, that is to the folk-categories of the people he is studying. Indeed it is difficult to see how those who maintain the contrary view could carry the full implications of their standpoint into the treatment of specific sociological data. A philosopher who starts from an acceptance of the Weberian doctrine of Verstehen and but a limited acquaintance with the more empirical contributions of social science, has recently argued that all sociology is impossible since the observer can never get outside the conceptual apparatus of his own society nor, conversely, inside that of any other (Winch 1958). This is not an illegitimate con- clusion, given the Weberian premise. But it implies the unacceptability of that premise in its literal form. To demonstrate that the sacred-profane dichotomy is not a universal feature of the actor’s situation is not of course to render it unserviceable as an analytic tool. In order to determine its possible utility for such purposes, I then examined the criteria of sacredness. Of the two positive features, respect and the symbolic element, the first failed to indicate either a category or polar type of specifically religious relationships. While in the light of the essentially external character of the dichotomy, the attribution of a ‘symbolic’ or ‘expressive’ element to ritual or reli- gious (i.e. ‘non-rational’) behaviour often turned out to be no more than a way of announcing that the observer is unable to make sense of an action in terms of an intrinsic means-end relationship, a ‘rational’ cause and effect nexus, and has therefore to assume that the action in question stands for something other than it appears to; in other words, that it expresses or is symbolic of something else. But what is that something else? This is where the recognition of the external character of the sacred-profane, supernatural-natural dichotomy becomes important. For it points to the fact that the referent of the symbol is supplied by the observer, not the actor. What the former assumes is expressed (or symbolized) is his interpretation of ‘society’, ‘ultimate values’, ‘social order’ or the ‘social structure’. I do not wish to imply that some magico- religious behaviour is not ‘symbolic’ from the actor’s point of view. Clearly it is. But so is much other behaviour. Indeed, in the last analysis, all verbal behaviour is sign behaviour. Hence the category of symbolic action does not in itself mark off an area or polar type of social action 156RELIGION AND RITUAL which it is in any way possible to characterize as ritual or religious. For it can be said, in an important sense, that all social action is ‘expressive’ or ‘symbolic’ of the social structure, because the more general concept is simply an abstraction from the more specific. It is not, however, ‘ex- pressive’ in the way many sociologists implicity assume, that is, it does not express major principles of social behaviour. Indeed such an ap- proach simply involves the reification of an organizing abstraction into a causal factor. What happens, then, is that symbolic acts are defined in opposition to rational acts and constitute a residual category to which ‘meaning’ is assigned by the observer in order to make sense of otherwise irrational, pseudo-rational or non-rational behaviour. And consequently the con- clusion reached by the earlier examination of the positive criteria of ritual acts is reinforced by looking at the negative criterion, that is, the contention that these practices fall outside the intrinsic means-ends schema. Once it is recognized that the accepted criteria for the isola- tion of sacred or ritual or magico-religious phenomena are derived not from the actor’s but from the observer's assessment of what is intrinsic, certain problems in the study of comparative religion fall into place. The way is open for a partial rehabilitation of the usages of the nineteenth- century anthropologists. The conclusion of this summary of the argument leads directly to the final portion of this paper, which consists of an attempt to suggest more definite ways of using not only the concepts of ‘religion’ and ‘ritual’, but also that of ‘ceremonial’, for it is this, I have claimed, that Durk- heim identifies with religious activity. To begin with ‘religion’, it is sig- nificant that in one of the most thoughtful of recent contributions to the study of the religion of a non-literate people, Evans-Pritchard has defined his field of discourse in Tylorian terms. With all its limitations this definition appears to offer the nearest approach to a resolution of our problem. But the acceptance of Tylor’s minimum definition still leaves un- solved Marett’s problem concerning the boundary between spiritual beings and spiritual forces. While beliefs of this kind must inevitably fall along a continuum, the associated practices present the opportunity for a sharper discrimination. We may take as a point of departure Frazer’s definition of religious acts in terms of the propitiation of supernatural powers. Acts of propitiation directed towards supernatural powers consist of sacrifice (food-offerings and especially blood sacrifice), liba~ tion (offerings of drink), gifts of non-consumable material objects, prayer (verbal offerings) and the ‘payment of respect’ by other forms of gesture. We may say then that religious beliefs are present when non- human agencies are propitiated on the human model. Religious acti- vities include, of course, not only acts of propitiation themselves but all 157JACK GOODY behaviour which has reference to the existence of these agencies, Such a formulation does not entirely dispose of Marett’s savage shouting at the thunderstorm; indeed such acts, if fully institutionalized and related toa set of cosmological beliefs concerning spiritual beings, would certainly be considered as religious within the terms of our definition. But how- ever imperfect the instrument may be, it does, I suggest, provide a focus for the comparative analysis of religious institutions which is of greater utility than the extensive definition preferred by Durkheim; indeed it is the one employed in practice by the majority of writers on this field, whatever expressed theory they may have adopted. With regard to ‘ritual’, I earlier called attention to the fact that Radcliffe-Brown and other writers used this term to designate the whole area of magico-religious acts and beliefs. In adopting such a usage, these writers followed Durkheim’s rejection of the admittedly external, observer-imposed distinction of the nineteenth-century intellectualists, and accepted one based upon the sacred-profane dichotomy, which was assumed to lie within the actor’s own definition of the situation. The conceptual difficulties involved in this usage have already been dis- cussed. Of the alternative possibilities, one is to identify ‘ritual’ with the magico-religious in the sense meant by Tylor and Frazer. A problem arises here from the fact that, both in common usage and in sociological writings, the term is frequently given a wider significance. The Oxford Dictionary for example defines a rite as ‘(1) a formal procedure or act in a religious or other solemn observance; (2) the general or usual custom, habit, or practice of a country, class of persons, etc., now specifically in religion or worship’. Not unconnected with the idea of the formality of the procedure (e.g. in ‘ritual intercourse’) is the further implication that an act so described is either not directed to any pragmatic end (‘rituals of the table’) or if so directed, will fail to achieve the intended aim (‘fertility rituals’). Thus the term has often a wider reference than the field of magico-religious behaviour and ‘rituals of eating’ may or may not be connected with such beliefs. The point was recognized by Nadel when he wrote in his account of Nupe Religion: When we speak of ‘ritual’ we have in mind first of all actions exhibiting a striking or incongruous rigidity, that is, some conspicuous regularity not accounted for by the professed aims of the actions. Any type of behaviour may thus be said to turn into a ‘ritual’ when it is stylized or formalized, and made repetitive in that form. When we call a ritual ‘religious’ we further attribute to the action a particular manner of relating means to ends which we know to be inadequate by empirical standards, and which we commonly call irrational, mystical, or supernatural (1954: 99). For Nadel, the category ‘ritual’ is inclusive and relates to any type of excessively formal action, while religious ritual (and in this he includes magic) covers acts where the means-end relationship is deemed inade- 158RELIGION AND RITUAL quate by empirical standards. This aspect of his distinction seems a little tenuous, as presumably, if the means display an incongruous rigidity, they are also to some extent empirically inadequate. Moreoyer, this statement of the situation fails to take account of the view, explicitly developed by Pareto, that much magico-religious behaviour is non- rational rather than irrational. But with these qualifications, Nadel’s position is basically similar to that taken in this paper. What has been said concerning the ‘secular’ nature of much ritual is equally applicable to ‘ceremonial’, Let us first explore the interrelation- ship between two concepts. Wilson recently used them in the following way. Ritual she defines as ‘a primarily religious action . . . directed to securing the blessing of some mystical power ... Symbols and con- cepts are employed in rituals but are subordinated to practical ends’ (1957: 9). Ceremonial is an ‘elaborate conventional form for the ex- pression of feeling, not confined to religious occasions’. Here ritual is equated with religious action, while ceremonial is a more inclusive concept referring to any ‘elaborate conventional form’. Wilson per- ceives that it may be important not automatically to identify ceremonial with religious performances in the way that Durkheim had tended to do. For while ceremonials such as Corpus Christi Day processions which celebrate mystical powers may perform similar functions to those like the anniversary of the October Revolution which have an exclusively secular significance, it is often useful to distinguish between them, particularly when considering the beliefs involved. However, it seems simpler to make the same distinction by using the term ritual in the general sense of what Wilson speaks of as ‘conventional’ action, while referring to the activities addressed to ‘some mystical power’ as religi- ous. Following the formulation of Radcliffe-Brown, ceremonial may then be used to refer to those collective actions required by custom, performed on occasions of change in the social life. Thus a ceremonial consists of a specific sequence of ritual acts, performed in public. In conclusion then, by ritual we refer to a category of standardized behaviour (custom) in which the relationship between the means and the end is not ‘intrinsic’, i.e. is either irrational or non-rational. Within this general category falls magical action, which is essentially irrational, since it has a pragmatic end which its procedures fail to achieve, or achieve for other reasons than the patient, and possibly the practitioner, supposes. This is Frazer’s ‘bastard science’. Then there are religious acts, which may be irrational (as in the case of many forms of sacrifice and prayer) or they may be non-rational, as:in many public celebra- tions, but all of which involve supernatural beings. Then finally there is a category of ritual which is neither religious nor magical; it neither assumes the existence of spiritual beings nor is it aimed at some em- pirical end, though this is not to deny that it may have a recognized ‘purpose’ within the actor frame of reference as well as some ‘latent 159JACK GOODY function’ from the observer’s standpoint. Within this category of ritual fall ceremonials of the non-magico-religious kind: civil marriage cere- monies, rituals of birth and death in secular houscholds or societies, Here too are the acts that we cannot speak of as public ceremonials, the ‘rituals of family living’ (Bossard and Boll, 1950) or ‘rituals of liquida- tion’ (Leites and Bernaut, 1954) and similar types of formalized inter- personal behaviour. The intention of this paper was to examine the ranges of meaning assigned to certain basic concepts in the sociology of religion with a view to clarifying their use as analytic tools. The general conclusion is that it is impossible to escape from the fact that the category of magico- religious acts and beliefs can be defined only by the observer and that attempts to see either this or the sacred-profane dichotomy as a univer- sal part of the actor’s perception of his situation are misleading. Any effectiveness which these terms may have in comparative studies de- pends upon a realization of their limitations and involves a return to the usages of earlier, pre-Durkheimian writers in this field, anyhow as a starting point for further exploration. Postscript The first draft of this paper was given a number of years ago at a seminar in Oxford, and I would like to acknowledge the stimulus of a series of discussions with M. Richter, Hunter College, New York. The paper was rewritten when I was a Fellow at the Center for Ad- vanced Study in the Behavioral Sciences, 1959-60, but some additional comment upon more recent developments seems called for. When I first wrote, commitment to the sacred-profane, ritus-mos dis- tinction dominated social anthropological thinking in the field of com- parative religion. The influence of Durkheim is still much in evidence here, and nowhere more clearly than in the work of Lévi-Strauss and those who have followed his theoretical interests. Leach sees the dis- tinction as referring to aspects rather than types of social action, and ritual as a ‘pattern of symbols’ referring to the ‘system of socially approved “proper” relations between individuals and groups’ (1954: 15). Needham (1960) however employs the radical dichotomy in much the same way as Durkheim and Lévi-Strauss when he comments upon Bernardi’s material (1959) on the Mugwe, the Failing Prophet of the Meru of Kenya. The emphasis which these writers give to this distinc- tion is connected with their general interest in ‘elementary structures’ with two or three constituents (‘binary’ or ‘tertiary structures’), a morphological scheme into which conceptual dichotomies and tricho- tomies fit with an attractive neatness (e.g. Lévi-Strauss 1956, goff.). But apart from the more general criticism made in the body of the paper, there are two further dangers here. Firstly, such radical dis- 160RELIGION AND RITUAL tinctions sometimes appear to be seen as having explanatory power in themselves, especially when an association is made between two or more sets of ‘oppositions’. Secondly, just because the elucidation of these relationships is given explanatory force, there is a tendency to assume the presence of such concepts on evidence of a rather slender kind. The difficulties behind some aspects of Durkheim’s position have become clearer in the course of the last few years and this has led to a change of approach. In the paper itself, I draw attention to Evans- Pritchard’s adoption of the Tylorian definition of religion in his analysis of the Nuer (1956). In his introduction to the recent translation of Hertz’s essays, he also rejects the polarity between the sacred and the profane which he finds, as I do, ‘vague and ill-defined’ (1960: 12). Firth, in his recent Huxley Memorial Lecture, calls upon anthropolo- gists not to be afraid of subscribing to the ‘intellectual, rationalist view’ held by nineteenth-century writers in the field, and puts forward a definition of religion not far removed from that of Tylor. ‘Religion’, he writes, ‘may be defined as a concern of man in society with basic human ends and standards of value, seen in relation to non-human entities or powers’ (1959: 131). Moreover the conclusion which he reaches after examining Tikopian material upon spirit mediumship is broadly in line with Evans-Pritchard’s account of Nuer religion, namely, that while some aspects of religious practices and beliefs are closely related to the ‘social structure’, others are relatively loosely linked and operate as ‘semi-independent variables’. To anyone outside the tradition of academic sociology, such a con- clusion might well pass without comment. But a corollary of the im- mense impetus which Durkheim’s great work, The Elementary Forms of the Religious Life, gave to the sociology of religion has been a tendency to over-determine the relationship of the religious with other social in- stitutions. Moreover, definitions of ritual and religion as ‘symbolic’ of social relations have the disadvantage, not only of being hampered by the ambiguities involved in the term symbolic, but of seeming to assert as a general principle precisely what requires to be demonstrated in each particular case. Forde, in the Frazer lecture for 1957, The Context of Belief (19582), has called attention to other gaps in the Durkheimian approach, emphasiz- ing in particular that other ritual practices such as the fetishism of the Yaké are not simply symbolic expressions of social relationships but are also concerned with environmentally determined conditions, such as the incidence of disease. Here as elsewhere, Forde lays special emphasis upon the manipulative aspects of the ‘supernatural economy’ of the Yaké (19586), a theme also pursued in my own discussion of the inevitable ‘circulation of shrines’ arising from the built-in obsolescence of those ‘irrational’ magico-religious agencies which make specific promises that they are later seen not to fulfill (1957). Some of the broader 161162 JACK GOODY implications for the Durkheimian thesis had been brought out in Warner’s examination of the Murngin material (1937: ch. 8), from which it is clear that the distinction between magic and religion on an instrumental-expressive basis (to use Parsons’ terminology) is not viable. In his theoretical treatment of the subject, Good (1951) suggests a continuum, with magic in general more instrumental, religion more expressive. Recently, in an interesting article entiled A Definition of Religion, and its Uses (1960), Horton has examined variations in religious behaviour along a similar axis, the poles of which he specifies as manipulation and communion, His definition of religion is essentially that of Tylor’s and the central argument is close to the one developed here, except that he is rather more sanguine than I about the immediate profit to be gained from adopting such a position. NOTES 1¢,,, man is double. There are two often with repugnance, the interpreta- beings in him: an individual being which has its foundation in the organism . and a social being which represents the highest reality in the intellectual and moral order that we can know by obser- vation—I mean society’ (1947: 16). * Neither Maine nor Weber intro- duces the criterion of assembling as a requisite of a corporate group. Neither does Radcliffe-Brown in his 1935 essay on Patrilineal and Matrilineal Succession (reprinted 1952). But subsequently he appears to consider this as an essential characteristic (1950: 41). Durkheim’s in- fluence is in evidence here. I regard this usage as making for possible confusions and prefer to speak of corporate groups in Radcliffe-Brown’s earliest sense, i.e. in the legal sense of having an estate. It is also important at times to distinguish groups in which the members gather together for various purposes; these I speak of as ‘assembling or convening groups’. Weber’s term Verband has also been translated ‘corporate group’, but he is referring to cases where the ‘order is enforced by the action of specific indi- viduals whose regular function this is, of a chief or “head” (Leiter), and usually also an administrative staff” (1947:145- 146). Theseare groups with a hierarchical authority system converging on one man or upon a few individuals; I refer to them as ‘pyramidal groups,’ following the usage of Fortes and Evans-Pritchard (1940: 1~23). 3 Ernest Jones writes of the symbol in the Freudian sense: . the individual has no notion of its meaning, and rejects tion’ offered of it (‘The Theory of Sym- bolism’, British Journal of Psychology, 9, 1817-19, 184, quoted by Morris i640, 78). ™ In the introduction to the second edition of The Andaman Islanders (1933), Radeliffe-Brown distinguishes between what he sees as two interrelated concepts, meaning and function. The meaning of a rite ‘lies in what it expresses, and this is deter- mined by its associations within a system of ideas, sentiments and mental attitudes’ (viii), The notion of function ‘rests on the conception of culture as an adaptive mechanism ...” (ix) and concerns the contribution of the part to the continuity of the whole. Although he makes this dis- tinction, he also remarks that ‘In the two theoretical chapters of this book the dis- cussion of meanings and the discussion of function are carried on together. Perhaps it would have been an advantage to separate them’ (x). But in his later essay on Taboo, the Frazer lecture for 1939, the two also tend to get treated together. For instance, in his discussion of ‘the meaning and social function’ of the Andaman practice of avoiding the use of a person’s name, the two operations are dealt with as one (1952: 146) and the custom is seen as a ‘symbolic recognition’ that the par- ticular person is occupying an abnormal social position. In other words, the rite symbolizes an aspect of the ‘social struc- ture’. Again, both the meaning and social function of totemic rites are related to ‘the whole body of cosmological ideas of which each rite is a partial expression’ (1952:145); here the rite is held to ex-RELIGION AND RITUAL press the cosmology, and the cosmology in turn has a ‘very special’ relationship with the social structure. On the other hand, in his analysis of Andaman food taboos, the referent of ritual is seen as objects and customs of social importance, those imbued with ‘ritual value’. 5 The strains involved in attempting to confine sociological analysis to ‘purpos- ive’ action are brought out in Dorothy Emmet’s discussion of Nadel and Par- sons in her book, Function, Purpose and Powers (London, 1938), p. 108. 163 * Similar qualms were expressed by de Brosse in 1760 about allegorical inter- pretations of myths in his work, Du Culte des Diews Fétiches ... ‘Vallégorie est un instrument universel qui se préte & tout. Le systéme du sens figuré une fois admis, on y voit facilement tout ce que l’on veut comme dans les nuages: la matiére n'est jamais embarrassante; il ne faut plus que de Vesprit et de l’imagination: c’est un vaste champ, fertile en explications, quelles que soient celles dont on peut avoir besoin’ (6-7). REFERENCES Betian, R. N. 1957. Tokugawa Religion: The Values of Pre-Industrial Japan. Glen- coe, Tl. Bernarbt, B. 1959. The Mugwe: a Failing Prophet. London. Bossarp, J. H. S, and Bott, E. S. 1950. Ritual in Family Living. Philadelphia. Cantor, R. 1959. Man and the Sacred. (first ed. 1939). Glencoe, Ill. DE Brosse, G. 1760. Du Culte des Dieux Fétiches, Paris. DurxHem, E. 1947. The Elementary Forms of the Religious Life (frst published, 1912; Eng. transl., 1915). Amer. edn., Glencoe, Ill. Eutave, M. 1958. Patterns in Comparative Religion. New York. Evans-PRiTCHARD, E. E. 1937. Witch- craft, Oracles and Magic among The Azande. London. Evans-PrircHarp, E. E. Religion. London. | + Firta, R. 1959. Problem and Assump- tion in an Anthropological Study of Religion. 7. R. Anthrop. Inst., 89, 2. Forpe, D 19580. The Context of Belief. Liverpool University Press Forve, D. 1958). Spirits, Witches, and Sorcerers in the Supernatural Econ- omy of the Yaké. 7.R. Anthrop. Inst., 88, Forres, M., and Evans-Prircuarp, E. E. 1940. African Political Systems. Lon- ion. Goone, W. J. 1951. Religion among the Primitives. Glencoe, Ill. Goopy, J. 1957. Anomie in Ashanti? Africa, 27, 4. Harrison, J. E. 1912. Themis. Cam- bridge. canna 1960. Death and the Right Hand (Engl. transl), London, 1956. Nuer Horron, R. 1960. A Definition of Re- lgion, and its Uses. 7. R. Anthrop. Inst., Rvcxnionn, C. 1952. Values and Value- Orientations in the Theory of Action: An Exploration in Definition and Classifica- tion, in Toward a General Theory of Action (ed. Parsons, T., and Shils, E. A.). Cambridge, Mass. Leacu, E. R. 1954. Political Systems 9 Highland Burma. London, ae Lerres, N., and Bernavt, E. 1954. Ritual of Liquidation: Communists on Trial. Glencoe, Til. Lessa, W. A., and Vocr, E. Z. 1958. Reader in Comparative Religion, Evanston, Ill. Liivi-Srrauss, C. 1956. Les organisations dualistes existentelles? Bijdragen Tot de Taal, Landen Volkenkunde, 112, 2. Mauinowski, B. 1922. Argonauts of the Western Pacific. London. Mainowski, B. 1954. Magic, Science and Religion, in Science, Religion and Real- ity (ed. Needham, J.). London, 1925. Reprinted in Magic, Science and Religion. New York. Morr, C. 1946. Signs, Language and Behavior. New York. NaneL, S. F. 1954. Mupe Religion. Lon- don. Neepuay, R. 1960. The Left Hand of the Mugwe: an Analytical Note on the Structure of Meru Symbolism. Africa, 30, 1. Parsons, T. 1937. The Structure of Social Action. New York. Parsons, T. 1954. Essays in Sociological Theory (rev. ed.). Glencoe, Ill. Parsons, T., and Sums, E, A. 1952. Values, Motives and Systems of Action, in Toward a General Theory of Action (ed.164 Parsons, T., and Shils, E. A.). Cam- bridge, Mass. Rancutre-Brown, A. R. 1939. Taboo. Cambridge. Reprinted in Structure and Function in Primitive Society. London. 1952. Rapcurre-Brown, A. R. 1950. Intro- duction to African Systems of Kinship and Marriage (cd, Forde, D., and Radcliffe- Brown, A. R.). London. Rapcurre-Brown, A. R. 1952. Struc JACK GOODY ture and Function in Primitive Society. Lon- don. ‘Warner, W. L. 1937. A Black Civiliza- tion. New York. ‘Weser, M. 1947. The Theory of Social and Economic Organisation (transl. Henderson, A. M., and Parsons, T.). Edinburgh. Wuson, M. 1957. Rituals of Kinship among the Nyakusa. London. Wincu, P. 1958. The Idea of a Social Science and Its Relation to Philosophy, London, University of CambridgeCopyright © 2003 EBSCO Publishing