RITUAL, HISTORY AND POWER SELECTED PAPERS IN ANTHROPOLOGY M. BLOCH * BERG LONDON SCHOOL OF ECONOMICS MONOGRAPHS ON SOCIAL ANTHROPOLOGY 18 Ritual, history and power Here too the amount of social structure, of the past in the present, of ritual communication, is correlated with the amount of institution-alized hierarchy and that is what it is about. Please note, however, that I am not proposing a simple connection with the degree of inequality. Some inequality is often manifested as unadorned oppression, but, as Weber pointed out, it is then highly unstable, and only becomes stable when its origins are hidden and when it transforms itself into hierarchy: a legitimate order of inequality in an imaginary world which we call social structure. This is done by the creation of a mystified 'nature', consisting of concepts and categories of time and persons divorced from everyday experience, and where inequality takes on the appearance of an inevitable part of an ordered system.12 For example, Marilyn Strathem, in her study of ideas concerning women in the New Guinea highlands (1972) shows how women are sometimes seen for what they are, producers of food and children, while at other rimes as polluting creatures spoiling the creative activities of men. It is not surprising that in order to maintain two such theories simultaneously two cognitive systems are needed. Unfortunately many anthropologists, fascinated as usual by the exotic, have only paid attention to the world as seen in ritual, forgetting the other conceptualization of the world which their informants also hold, and which is denied by ritual communication. They have presented as cultural variation what are in fact differences between the ritual communication view of the world of the people they study and our everyday practical one. In doing this, and unlike Malinowski, they have confounded the systems by which we know the world with the systems by which we hide it. 2 Symbols, song, dance and features of articulation: Is religion an extreme form of traditional authority?1 Introduction Recent studies of symbols in ritual share two features. First, they isolate symbols from the ritual process; second, they interpret symbols as units containing meaning. In this paper I want to argue that symbols in ritual cannot be understood without a prior study of the nature of the communication medium of ritual in which they are embedded, in particular singing and dancing, and that once this has been done we find that symbols cannot any more be understood as units of meaning simply on the Saussurian signifier/signified model, however subtly this model is handled. Such varied writers as Bettelheim (1962) and Turner (1967) are to my mind examples of writers ultimately using this model for the study of meaning in ritual. Because so much of communication in ritual is linguistic, a fact which has been strangely minimized (with the notable exception of Malinowski, Leach and Tambiah) and because so much more has been written about meaning in speech, I shall concentrate on the linguistic aspects of ritual, especially since some of the work done recendy by linguists seems to be helpful for the kind of argument presented here. Borrowing from linguistics is perhaps a familiar thing to do and it could be argued that in some ways the practice has been the hallmark of structuralist explanations for a long time. This is, in fact, an illusion for two reasons: 1) structuralists have borrowed from linguistics at a time when linguistic theory declared itself uninterested or at least incompetent in dealing with semantics; 2) they have used linguistic models in an analogical way, not in a direct way. By contrast I am talking of language used in ritual and I am therefore applying linguistic theories to the data they are designed to handle; secondly, I am using linguistic semantic studies. In particular 20 Ritual, history and power I am concerned with the theories of such writers as McCawley, Fillmore and G. Lakoff, among others, who have stressed the identity of syntax and semantics: that meaning is primarily transmitted by the way lexical units can be combined in utterances. For these writers, semantics becomes the study of the rules of combination of speech, the propositional content of speech, rules which in this paper I shall call features of articulation. If semantics and syntax are the same, the logical potency of language depends on the creativity of syntax. This does not constitute a restriction, precisely because syntax is so creative. Working freely, the syntax of a particular language can articulate almost any argument. However, it also follows from such a conclusion that if we are dealing with a language use where syntax does not articulate freely, the potential of language for carrying arguments becomes reduced and the proposirional force of language is transformed. Ritual is an occasion where syntactic and other linguistic freedoms are reduced because ritual makes special uses of language: characteristically stylized speech and singing. Ritual is therefore a place where, because the ordinary forms of linguistic communication are changed, we cannot assume the semantic processes of more ordinary communication. The semantic processes of ritual, whether they relate to the way units of meaning (words, for example) are joined, or to the nature of these units, can only be understood after we have studied the significance of the mode of communication used in ritual, i.e. stylized speech and singing. By extension we can also assume that the parallel modifications in non-linguistic communication which occur in ritual also affect the nature of semantics in these fields. An obvious parallel modification is the way the bodily communication of non-ritual life is so often replaced in rituals by dance. We have, of course, much description of singing and dancing, and recent ethnographers have been at pains to improve the quality of the ethnographic record, but the vital question of what is the effect of singing something rather than saying it normally for the way meaning is carried has rarely been faced. Malinowski (1935) and, following him, Leach (1966) and Tambiah (1968) have noted certain aspects of the significance of religious language, especially its archaism and its 'set apart' character. Yet, however valuable their work in stressing the reality of the situation, it still seems to leave open this basic question. Symbols, song, dance and features of articulation 21 If little has been written on religious language, hardly anything has been written of theoretical import on dance, although here again techniques of description have dramatically improved. The questions about the significance of the dance raised by Evans-Pritchard in 1928 still remain as open as ever. This lack of theoretical interest by anthropology in ritualized speech, singing and dancing is particularly strange for two reasons. First, I very much doubt that an event observed by an anthropologist which did not contain these three elements would ever be described by him as a ritual. In other words these phenomena have been implicidy taken as the distinguishing marks of ritual.2 Second, the history of the social anthropology of ritual has been dominated by the influence of Dürkheim, who stressed how through participation in ritual singing and dancing collective representations were made to appear as external to the individual, as having a force of their own. How this process happened was never actually explained by Dürkheim. But for him the transformation of everyday ideas into sacred ones took place during rituals involving singing and dancing. I too shall take as my starting point the Durkheimian problem of how ritual makes its statements appear powerful and holy; but far from considering song and dance as a secondary aspect of ritual, I make it my main concern. A sufficient justification for such a procedure seems to me to be found simply in the realization of just how much of what is going on is left out of account by the kind of formal analysis of ritual to which we have been accustomed. (Beidelman, 1966; Leach, 1972; La Fontaine, 1972; Ardener, 1972.) If we take as a fairly typical example of a ritual the circumcision ceremonies of the Merina of Madagascar (Bloch, 1986), an event of great complexity in which a bewildering number of symbols come into play, what I mean will become clear. This ritual normally begins at 3 o'clock one afternoon and goes on until 6 o'clock the next morning when the actual operation is performed. During nearly all this time the bodies of the participants are not used in the same way as they would be in a non-ritual context, that is to say the participants dance. The voices of the participants are not used in an everyday way, that is, the participants use their language in a particular way: formalized speech and singing. A purely formal analysis of the symbols of the ceremony would simply miss out this central fact. What I want to argue is that the examination of this 22 Ritual, history and power Symbols, song, dance and features of articulation 23 mode of communication should^by contrast with formal analyses, be made not as an afterthought but as the centre of our examination of religious, ritual or mythical phenomena and that only when we can understand its nature can we then understand the nature of the argument to be obtained from a formal analysis. By understanding the nature of argument I mean especially the understanding of the nature of the link between the parts of the argument, the features of articulation. Only when it is clearly realized that the features of articulation in the argument in a ritual are fundamentally different from the features of articulation in logic or in ordinary speech will we be able to understand what sort of link-up there can be between ritual structure and the structure of concepts and thoughts in a non-ritual context. The significance of the formalization of language To identify language uses in ritual we can use again the example referred to above since it is fairly typical. In trie Merina circumcision ceremony we find three fairly distinct uses of language: 1. We have formal oratory by elders explaining the purpose of the ceremony and asking for blessings from God and the ancestors. This oratory is identical to the fomal speeches made by Merina elders at political meetings (Bloch, 1971b). It shares with political oratory such features as the use of a restricted archaic vocabulary, the use of only certain syntactic forms, usually the ones considered the most polite and impersonal, the use of a wealth of illustrations from a given traditional body of sources - proverbs, traditional history, etc. - the use of a special' style of delivery and finally the use of a rigid traditional structure for the whole speech. The speeches of elders at the beginning of a ritual are identical to the speeches of elders in political councils. Such speech-making is referred to by the Merina as 'speaking the words of the ancestors'. By this is meant that the elders are speaking not on their own behalf but on behalf of the ancestors and elders of the whole descent group whether living or dead. They are repeating the general truth which they have had passed on to them by previous generations. They are speaking in the style and often indeed with the very words which they believe were used by the dead ancestors at similar councils or at similar religious rituals in the past. If the speeches of the elders at the beginning of the ceremony are identical to those of the elders in a secular council, as the ceremony goes on a change occurs. The speeches which take place later during the ceremony, under the influence of rhythmic music, drink and general excitement, will, in several instances, appear to the external observer to be of a very different kind and would immediately be classed by him as the speeches of persons under possession. The elders appear to be in a trance-like state and it is said that it is not they who are speaking but their ancestors. The actual utterances are similar in their formal aspects to any other speeches made by elders but they are delivered in a different manner, a different tone, and are often indistinct. Anthropologists would class such speech-making as completely different from the more sober speech-making at the beginning of the ceremony but the participants describe the two kinds of speaking by the same phrase 'speaking the words of the ancestors'. For them possession is an extreme form of saying what would have been said by the elders of the past, or repeating their words. Instead of the ancestors speaking indirectly through the memory of the living elders they speak directly through their person. 2. The second linguistic form used during the ceremony is intoning, what Malinowski calls 'spells' (1935: 240). This occurs at one stage in the ceremony and consists of the repeating again and again by the elders of a set formula in a chanting voice accompanied by whoops and other non-lexical shouts. 3. The third linguistic form is singing. This is really the dominant form in a Merina circumcision ceremony, as indeed in so many other rituals in many other parts of the world. Singing goes on almost uninterrrupted throughout the whole length of the ceremony. The songs are repeated again and again. Their texts are not created but have been learned previously and, as is the case in all songs, the rhythm, the phrasing and the pitch are fixed. Formal speech-making, intoning spells and singing are but different steps in the same process of transformation from secular discursive language, and so for the sake of simplicity we shall first consider the implications of the formalization employed in speech-making. To do this only for speech-making in religious rituals would, however, be misleading, since for the Merina, as indeed for many other cultures, the same type of linguistic modification occurs in religious oratory and in political oratory. As a first step in the 24 Ritual, history and power Symbols, song, dance and features of articulation 25 analysis I shall therefore summarize the results of my study (Bloch, 1975b) of the implications of speech formalization in a political context. This will serve both to make some of the points in the argument and to introduce the significance of formalization in a religious context. In a study of political speech-making our interest inevitably turns first to social control and in particular to the often noted phenomena where in a highly formalized or ritualized political situation there seems no v/ay whereby authority can be challenged except by a total refusal to use the accepted form which is compulsory for this type of occasion, i.e. a total refusal of all political conventions. The ceremonial trappings of a highly formalized situation seem to catch the actors so that they are unable to resist the demands made on them. This situation raises the two following questions: 1) How is it that formalization can become a form of power or coercion? 2) In what types of political systems does one find this 'power through form' and in what type of political system is formalization a less important part of the political process? If we attempt to answer the second question first it is immediately apparent that the extreme formalization of language with its accompanying exercise of power is characteristic of the traditional authority situation as defined by Weber (1968: 226). We find it as part of the political process in Polynesia or Madagascar while it is not found in systems where the powers of leaders are continually dependent on their manipulation and are at every moment challenged. The oratory of a New Guinea highland Big-Man, for example, does not show the features of formalization which I consider below, while in Samoa and in Tonga these features are developed to an extreme extent (Sahlins, 1963). If we can, without much difficulty, answer the question 'In what type of political events do we expect formalization of communication?', the much more difficult and logically prior question remains: Why is formalization a kind of power? In order to answer this our first task is to define what is involved by formalization. To do this we can contrast 'formalized* with what might be called 'everyday' speech acts in the following way: EVERYDAY SPEECH ACTS Choice of loudness Choice of intonation All syntactic forms available Complete vocabulary Flexibility of sequencing of speech acts Few illustrations from a fixed body of accepted parallels No stylistic rules consciously held to operate FORMALIZED SPEECH ACTS Fixed loudness patterns Extremely limited choice of intonation Some syntactic forms excluded Partial vocabulary Fixity of sequencing of speech acts Illustrations only from certain limited sources, e.g. scriptures, proverbs Stylistic rules consciously applied at all levels The first point to notice is that formalized language, the language of traditional authority, is impoverished language: a language where many of the options at all levels of language are abandoned so that choice of form, of style, of words and of syntax is less than in ordinary language. Formalized language corresponds well with what Bernstein has called a restricted code (1972: 474), although I am considering a wider range of linguistic phenomena than Bernstein and I am drawing different sociological conclusions from his. The contrast described above is not to be seen as a dichotomy between two types of speech acts but as a continuum between two extremes. In any particular culture not necessarily all the restrictions of formalized speech acts listed above will come into play when formalized speech is used. This will vary both with the cultural tradition and the type of language. How far formalization goes will also vary in degree both within a culture since there might be a number of more or less formalized codes to choose from (Gumperz, 1961), and also between cultures, since as mentioned above the relative importance of traditional authority in their political system also varies. We are therefore considering degrees of formalization, though of course the nature of the effect of giving up options in language is of the same kind, however far this process goes. This giving up of options on several levels simultaneously is especially significant because increasing restriction of communication has a geometrical effect on the restriction of forms of speech. If we think of the sentence 'the cat sat on the mat' and assume that 26 Ritual, history and pomer Symbols, song, dance and features of articulation 27 there are three alternative nouns for each noun locus we find that there are nine possible sentences. If we say that the verb can be in one of three tenses there are then twenty-seven possible sentences and then that these sentences can be said in one of three intonations, there are then eighty-one possible sentences. Now if we look at this backwards we see that by just specifying one restriction, say intonation, saying that only one intonation is suitable, the number of possible sentences drops from eighty-one to twenty-seven. If, however, we bring in two restrictions occurring at the same time, say that with a certain intonation only certain nouns are acceptable, the effect on the choice of sentences is dramatic. In fact it falls from eighty-one possible sentences to three. This sort of effect is present when we note that in a given political system strict rules of 'politeness' are enforced. In these real-life instances the effect is far greater because of the greater number of restrictions but, of course, the original potential of choice is also very much greater. The reason for this dramatic impoverishment in linguistic choice is the fantastic creativity potential of natural language, which has been so much stressed recently and comes from the relative independence of the rules of combination one from another and their further independence from the units which they operate. If, however, some of these rules are removed or, even more drastic, if rules are attached one to another or rules to units, this creativity very rapidly vanishes. In other words, as soon as you have accepted a form of speaking in an appropriate way you have begun to give up at a bewilderingly rapid rate the very potential for communication. The process does not stop with formal linguistic aspects but also influences those fields usually referred to as style in the literary sense. One of the features of everyday speech is that it can be enriched by comparisons and cross-references to other events of an extremely wide range. When, however, we look at the language of traditional authority we find that the power of cross-references becomes more and more restricted to a body of suitable illustrations, often proverbs or scriptures. The effect of always comparing particular events to the same general illustrations further reduces the specificity of utterances so that all events are made to appear as though they were all alike. In other words the ability of language to communicate messages concerning particular events and its ability to convey specific messages leading to particular action disappears. This is so because the ability of the particular units of speech act to relate closely to the experiential world, and the sequencing of speech units to relate closely to a particular experiential process is greatly reduced, as the number of words, illustrations and grammatical sequences that can be chosen to ft reality is reduced. The individuality and historicity of event disappear since irrespective of minor differences these events are all like the scriptural examples. The formalization of speech therefore dramatically restricts what can be said, so the speech acts are either all alike or all of a kind and thus if this mode of communication is adopted there is hardly any choice of what can be said. Although the restrictions are seen usually as restrictions of form rather than of content, they are a far more effective way of restricting content than would be possible if content were attacked directly. Formalization therefore goes right through the linguistic range. It leads to a specially stylized form of communication: polite, respectful, holy, but from the point of view of the creativity potential of language, impoverished. We may also note that the restrictions implied by formalization are all the more powerful because the abandonment of the freedom implied by natural discourse is in the direction of rare forms of everyday discourse. This is itself due to formalization. The relative fixity of formalized language isolates it from the processes of historical linguistics and so a secondary result of formalization becomes the typical archaism of the language of traditional authority and even more the language of ritual (Aitchison, 1966). What are the implications of abandoning the creativity potential of speech? The first and the most obvious of the implications of abandoning linguistic choice is that an utterance instead of being potentially followed by an infinity of others can be followed by only a few or possibly only one. Because of the linear nature of speech acts this can best be represented in the following way: if we start with a hypothetical situation where all the systemic freedoms of language are retained and where there are no extra linguistic restrictions on speech we can see that any speech act A may be followed within the formal rules of language by a very large number, an almost infinite number, of possible speech acts B, since from a purely structural point of view there is an almost infinite number of grammatical utterances which can follow a particular sentence. We can say that A in no way predicts by its form B. If, on the other hand, we are 28 Ritual, history and power Symbols, song, dance and features of articulation 29 considering a situation where language is greatly restricted, where its creative potential is abandoned by the fact that the speakers have accepted 'appropriate' or formalized forms of speech, we can see that the number of speech acts B which can follow a speech act A has become very seriously restricted. Indeed in the extreme situation where we are dealing with traditional discourse repeated from a traditional body of knowledge only one speech act B can follow speech act/4. In other words with increased formalizations predicts to an ever greater extent B. In terms of the experience of an individual it means that as he uses formalized language he very largely implies his last words by his first, since once he is speaking in the right way there is only one predetermined line along which he can proceed (Schegloff, 1972: 365). In fact even more is implied by the acceptance of formalization, because if the utterance of a speaker predicts what sort of things he will say it also predicts the answer of the other person so long as this other person is also accepting the code. The reason for this is that groups of people may find themselves in situations where a certainform is appropriate: this is the experience of everybody at any political occasion in a society where traditional authority is dominant. In such circumstances not only will one part of a speech act of a speaker predict the next but the speech act of one man will predict that of another. If a superior addresses an inferior, the latter, if he accepts to answer within the formalized code imposed on the situation (and he will rarely be in a position to do anything else) will find himself in a position where he cannot say 'no'. The reason is to be found in the fact that because the formalized speech of the superior has meant that one speech act predicts the next, this predictive power has jumped the gap from one speaker to the next. There is nothing strange in this since just as much as one part of a speech act needs to be appropriately joined to the next so must the speech act of one person be appropriately joined to those of another. In a situation where the features of articulation function freely there is a very wide choice of acceptable answers. In formalized speech the features of articulation have been, as we have seen, rendered arthritic and so the possible answers are dramatically reduced perhaps to one. This means that an inferior, in a political situation with a high degree of formality, usually finds himself handling a communication code where he can only accept what is said if he is to stay within the rules of appropriate behaviour. There is another way we can see this, and this is by considering the possibility of arguing back. As we are moving towards a more formalized code the possibility of contradiction becomes less and less. This is simply because contradiction implies a potential choice of speech act B following speech act A. There are intermediary situations where formalization has not totally restricted the possibilities of speech act B but this is so only to a certain extent and this extent is the limit within which contradiction is possible. It is because the formalization of language is a way whereby one speaker can coerce the response of another that it can be seen as a form of social control. It is really a type of communication where rebellion is impossible and only revolution could be feasible. It is a situation where power is all or nothing and of course in society total refusal is normally out of the question. There is, however, a drawback from the point of view of a political superior in using formalization for social control. If accepting a code (which appears to control only the manner of speaking but which in fact actually also controls the content of what is said) is a way whereby a superior can coerce the response of an inferior, this is done at the cost to the superior of losing his own freedom of manipulation. He can enter into a mode of speaking which compels the hearers, but since the mode predicts his speech at the same time, he also severely restricts himself in what he can say. We can see this dilemma in actual situations where political leaders are at the same time using the 'traditional1 mode of communication to bolster their authority, while themselves occasionally trying to break it down in order, as individuals, to obtain a greater freedom of manipulation over the situation, a freedom of manipulation which they usually need in order to fight off potential rivals to the status they are assuming. This is a very real problem for the political leader who, by using the formalized code, is taking on the role of acting as the 'father of his people' within a well-ordered system where all responses follow each other, but who, at the same time, needs to act the astute politician, thinking of new ways of discrediting his rivals. This he cannot do within the formalized code because the very fixity of the features of articulation make it inappropriate for the supple use of language he needs. There are two alternative codes present in the language a political leader uses. These are usually very easily distinguishable on 30 Ritual, history and power Symbols, song, dance and features of articulation 31 formal grounds. One code, the formalized one, is used to establish his traditional authority, but it is interlaced by a totally different code (used cautiously since it allows for contradiction) for the dirty work (ComarofT, 1980). The reason why the formalized code is unsuitable for practical day-to-day manoeuvring is because formalization creates an uncharted distance between specific things or situations and the communication. Language in its arthritic state cannot espouse the specific shape of reality due to its loss of suppleness in its form and in its manner of reference. Language which cannot espouse the specific shape of reality cannot be used to comment on it or to suggest alteration. In other words, in political situations where jobs still have to be done, formalization can never be used completely. Indeed, as we shall see, the only difference that can be shown to exist between religious communication and the communication of traditional authority is that for the latter a non-formalized code is still necessary while for the former it has been completely abandoned. If the relation between formalization and traditional authority can be accounted for in the way suggested above, it is important to consider further the effect this has on semantics. It was noted above how formalization led 10 increasing fixity in the features of articulation in language so that a part of a speech act predicted the next. Now if propositional force is to be equated, at least in part, with syntax (McCawley, 1971) it is clear that a radical change in syntax, in this case a severe restriction, implies an equally radical change in the nature of the propositional content of language. In other words, once the processes of formalization are under way any notion of meaning derived from the way ordinary language works will be misleading unless heavily qualified, J. Lyons has put this general point in an extreme form: In accordance with the principle 'meaning implies choice', I shall say that any linguistic form, up to and including the complete utterance, whose occurrence is not determined by the context has meaning in that context and, conversely, that any linguistic item whose occurrence in a given context is completely determined has no meaning in that context. It is important to realise that having meaning, as the notion is here defined, is a matter of how much meaning items have in context - that is, their contribution proportionately to the meaning of the whole utterance - not of what meaning they have. (Having no meaning is merely the limiting case of complete predictability) (J. Lyons, 1963). This statement seems to me to go too far, in that fixity does not seem to me to imply no meaning but, as we shaD see, a different kind of meaning: illocutionary force (see below). However, to look for 'meaning' in a fixed utterance without qualification is clearly misleading. In order to understand what kind of qualification is necessary it is useful to look again at the assumption made in the work of several linguists (McCawley, 1971) that semantics can adequately be represented by symbolic logic, the implication being that the articulating properties of language are what makes language capable of carrying a logical argument, and that logical presuppositions (e.g. see the studies in Fillmore and Langoenden 1971) are the very stuff of semantics. Logic implies that one postulated connection between units is more right than another because of the innate relation of the parts of the logical argument; one can therefore say that an argument to be logical must be couched in a form within which contradictory or alternative arguments are possible but excluded, not because of the way they are said but because they are untrue: an argument to be logical must be formally contradictable in order to show its logical nature. Normally any statement is open to contradiction and replacement and since this is so in ordinary situations argument and reason are possible. By contrast, formalized language rules out the two prerequisites for logic, the potential of one statement to be followed by a large number of others and the possibility of contradiction. Formalized language is therefore non-logical and any attempts to represent it as such, whether by a paraphrase into ordinary language which implies 'explanation' or by the use of tabular representation containing a logical form, is misleading. If the connection between the units in formalized language is not logical, that is, if their connection does not come from their fit, we may well ask what is the nature of their connection. The answer is easily found in what has been said above. The connection between units is given by the medium of expression, since the features of juncture predict the order of the argument. There can thus be no true or false explanations, no explanations that are better or worse, 32 Ritual, history and power Symbols, song, dance and features of articulation 33 since the potential for substitution which is employed in the notion of true or false or better or worse has been eliminated by the way the proposition has been put. To put it simply, we can say that logic depends on the flexibility of the features of articulation in language and if there is no such flexibility there can be no argument, no logic, no explanation, and in one sense of the word, no semantics. The effect of removing the possibility of alternatives from the mode of communication, as is done by formalization, makes what is being said beyond logic: its force is traditional authority, but disguised in that it has been accepted unconsciously before the event by the acceptance of the proper, or the polite, or the appropriate way of behaving. What is being said is the right thing because by the acceptance of the formalization of language it has become the only thing. To say this, of course, is not to say that such language is meaningless. We can distinguish two kinds of meaning. One is the propositional force of language, the ability of language to corner reality by adapting communication to past perception and connecting this with future perception. This is the power of language which linguists have been most concerned with (propositional force). This meaning potential of language is lost by formalization, but there is also the aspect of meaning which we can refer to as 'illocutionary force', or perhaps 'performative force' - 'not to report facts but to influence people' (Austin, 1962: 234). A speech sequence of the traditional political oratory type which we consider above is one where something is being communicated by one speaker to the other; at least the relationship in which the speakers stand. This is due to the illocutionary force of the speech. In any specific example of political communication we do not normally find that we are dealing either with speech communicating only propositional force or speech communicating only illocutionary force. In all cases it is a mixture of both. What I am arguing is that with increasing formalization propositional force decreases and illocutionary force increases; in other words that the two types of meaning vary inversely. To say that speech acts of a high degree of formalization have little propositional force and great illocutionary force does not explain what this illocutionary force is. It does, however, mean that we can rephrase what we are looking for when trying to understand formalized language. How does it communicate without explanation? And why should it be that traditional authority and religion tend to use a type of communication which both excludes explanation and hides this exclusion? These questions stand in sharp contrast to the question so often asked in studies of ritual, whether secular or religious: What does it explain? To ask this question assumes a function of traditional authority and religion which it cannot have and draws attention away from other possible functions which it may have, e.g. to hide reality. From the communication of traditional authority to ritual If formalized language diminishes the propositional force of language we can then ask the following question: What happens when formalization is more pervasive than we have considered it to be up to now, and when consequently propositional force is further reduced and illocutionary force is further increased? For this we turn to ritual and religion. Why has this discussion of the communication system of traditional authority been relevant to the study of ritual? The answer could be given at several levels, but first of all we can note that many of the features listed above as characteristic results of the formalization of communication in traditional authority have been noted by several writers as typical of ritual. Rappaport (1971) especially, whose conclusions are very similar to some of the ones in this paper, notes among others the following aspects of ritual: that it is the opposite of explanation, that it is 'fused' communication (following Bateson), that it is non-discursive, that its source is seen by the participants as being outside themselves, and that it is invariant. Secondly, we can note how many ethnographic studies dealing with traditional leaders find it difficult to class their activities as political or religious, whether they are to be thought of as chiefs and kings or priests.3 Thirdly, and here we move to a much more empirical level, we find that religious ritual and traditional authority are the only two major anthropological categories where formalization of language occurs and that this formalization is usually of the same kind in both. To return to the same example, we find in the Merina circumcision ceremonies that an intrinsic part is the 34 Ritual, history and power Symbols, song, dance and features of articulation 35 speeches of the elders, and that these speeches are identical to the speeches of elders in political meetings. As die ceremony proceeds these speeches become more and more formalized to a degree that would be unusual (though not unknown) in a political context, but the nature of this formalization is only a more extreme application of the principles which we have already recognized outside religious ritual. This fact is familiar for many societies and the presence of formalized oratory in a religious context occurs even in our own society in the case of sermons. However, if communication in religion begins at the same point where communication of traditional authority ends, religion carries on the process of formalization and all it implies very much further; much further, but still using exactly the same principles, restricting the range of choice of intonation, restricting the range of choice of vocabulary, restricting the range of choice of syntactic forms, restricting the range of acceptable illustrations, restricting the body of knowledge from which suitable illustrations can be drawn. Because religious rituals push the same techniques of the formalization of language further we must face what this means before we discuss the implications of the similarity. In the circumcision ceremonies we noted three types of linguistic phenomena. The first was the speeches of the elders, and we need not analyse this any further since what has already been said about the speeches of traditional leaders applies entirely here too. The second phenomenon is intoned formulae and prayers. This can be seen to be of a similar nature to the formalized speeches of elders, but having gone further along the line. The characteristic sing-song nature of such intoning is of particular importance in that it forms a bridge between singing and speaking, a bridge which, it is interesting to note, we find in many religious rituals throughout the world. What has happened in the transformation from formalized oratory to intoning is that the severely restricted choices of vocabulary, syntax, style, and speech-act structure which might have remained have now completely gone. What is being intoned is therefore nothing else but repeating what had been said before. Of course, this was true to a certain extent of many highly formalized speeches which were little else than that, but still they were not totally learned 'by heart'.4 Secondly, the choice of intonation and the choice in the rhythm of the delivery, which had only been slightly limited in oratory, is here much more restricted. This produces the sing-song effect. Intoning is therefore but a further move in the process of formalization of speech but it will readily be seen to be very close to the third linguistic manifestation of religious rituals: song. The difference between singing and intoning is very slight. It consists first of the ever more complete denial of choice of intonation and rhythm, and secondly in moving towards rhythms and intonations more remote from the forms of normal speech (a process noted above as a general tendency in formalized communication but which in the case of speech-making manifests itself principally in the fields of vocabulary and syntax). It may seem strange to think of singing as a special kind of language because we are so used to considering art as something set apart from the everyday concerns of life. This sort of point of view has, however, been shown not to apply to the visual arts of those societies without our particular type of politico-economic system. This must surely also be true for oral art. Song is, therefore, nothing but the end of the process of transformation from ordinary language which began with formalization. From the point of view of ritual this transformation is, however, a most important stage since singing is so often and so prominently an integral part of religious action. In the Merina circumcision ceremony it would be fair to estimate that much more than nine-tenths of the time is taken up in singing. The most immediate fact which follows from looking at song and formalized oratory together is that we can see that the communication of traditional authority does not differ in kind from the communication of religion. This assimilation of song to ordinary language may seem less strange if we bear in mind how little break there is in a continuum from, let us say, the language of party political broadcasts to the language of politicians making public appeals in times of emergencies, to the language of sermons, to the way prayers are spoken in church, to the way psalms are intoned, to full-scale singing. This continuum in our own culture is indeed so gradual that it would not be possible to draw the line and say clearly at any point: here speech ends, and song begins. All the implications of formalization which have been discussed above for formalized political speech also apply to song in an extreme form but the case of song is so extreme that I want to add here a few 36 Ritual, history and power other considerations. The almost total lack of individual creativity which is involved in singing a song means that the fixity of the feature of articulation that we had noted for political oratory is present to a much greater extent. Of course not all song completely rules out creativity; there are songs where a degree of innovation is possible, for example, in the Eskimo song contests. There is also the possibility of putting more or less 'expression' in a song. The degree of creativity means that the various forms of singing do not form just one point at the end of continuum from creativity to formalization, but a little of its length at the formalized end. Nonetheless the fact remains that the prepositional force of all song is less than that of spoken words in an ordinary context. The songs sung by groups of people in unison which characterize so much of ritual are particularly extreme examples of lack of individual creativity. These songs almost completely predict the linguistic journey the singer undertakes.5 To engage in a song in this kind of ceremony implies one moment of will: taking part, followed by a period where the linguistic action of the song is so passive that it is as though the singer were experiencing language from outside himself. (I use here, purposefully, language reminiscent of Dürkheim). The way one part predicts another applies clearly in one song but it also applies to a very large extent to the ceremony as a whole, since the singer is plunged by the ritual into a sequence where one song is quasi-automatically followed by the next, so that from the initial act of will in taking part in the ceremony at 3 o'clock in the afternoon until 6 o'clock next morning, the choice of one's words and acts is not exercised. (This, of course, is putting the matter slightly too strongly since the precise order of the song is not pre-ordered, although songs can only be chosen from a very limited body and certain rules of order do exist and cannot be ignored). The implications of the abandonment of the power of creativity throughout the ritual are, of course, fundamental, because if we go back to the earlier argument and remember what is implied by such notions as reason, argument, explanation, we realize that where there is no suppleness of language, where the features of articulation are not points where alternatives could be joined, where A totally predicts B, there can be no propositional force. There is no force of argument when Bs, other than the one given in the ritual, are ruled out by the connection between A and B which is essential since in a song everyone must Symbols, song, dance and features of articulation 37 sing together. In a song, therefore, no argument or reasoning can be communicated, no adaptation to the reality of the situation is possible. You cannot argue with a song. It is because religion uses forms of communication which do not have propositional force, where the relations between the parts cannot be those of the logic of thought, that to extract an argument from what is being said and what is being done in ritual is, in a sense, a denial of the nature of religion: as misleading as it would be to argue that relations of traditional authority are the result of rational arrangements between superiors and inferiors.6 It is therefore misguided to argue, as so many anthropologists have done, that religion is an explanation, a speculation about such things as man's place in the world (Levi-Strauss), an explanation of the place of evil in the world (Evans-Pritchard and Lienhardt) or even a statement of man's place in his social structure (Radcliffe-Brown, Dürkheim, Fortes, etc.). Religion is the last place to find anything 'explained' because as we have seen religious communication rules out the very tools of explanation which, when reintroduced, are considered sacrilegious or irreverent. Dance So far we have been saying that we cannot understand what is being said in a ritual if we do not bear in mind that it is being said in a way which denies what we commonly understand by communication. To do this we have discussed what is implied by saying something in a song rather than in everyday language. I have concentrated on language because so much work has been done on this, but as was noted above it is not only song that characterizes ritual but also dancing. In order not to repeat too much of what has gone before I shall only apply the same argument to bodily movement in a summary way. There is some sense in saying that bodily movements are a kind of language and that symbolic signals are communicated through a variety of movements from one person to another. There is also much evidence that the combination of bodily signs and their order is used to convey more complex messages. In this way there is a sense in which the control of the body is a 'language', albeit a rather poor one in comparison to the language of speech. The restrictions on this 38 Ritual, history and power Symbols, song, dance and features of articulation 39 language in traditional authority situations exactly parallels the formalization of speech and indeed usually accompanies it. The restriction of the free use of bodily language in such situations is actually even more strictly enforced. Again, as with language, we find that these restrictions are also present, and in a more systematic manner, in religious rituals. Messages carried by the language of the body also become ossified, predictable, and repeated from one action to the next, rather than recombined as in everyday situations when they can convey a great variety of messages. As with speech, the formalization of body movement implies ever-growing control of choice of sequences of movement, and when this has occurred completely we have dance. We therefore find dance, as well as formalized body movements, typical of religion. The implications of this transformation from ordinary bodily control to dance are the same as they are for language: argument and bargaining with bodily movements are replaced by fixed, repeated, fused messages. The acceptance of this code implies compulsion. Communication has stopped being a dialectic and has become a matter of repeating correctly. This treatment of dance, like the treatment of song above, seems to go against the generally accepted view of art as a kind of super-communication, a supreme occasion for creativity. By contrast I am arguing that art is, in fact, an inferior form of communication. The reasons for my argument follow from my acceptance of a generative view of language. Nonetheless, it raises the question why the folk view of art should be the opposite of what is the case. The reason for this view probably lies in the fact that the generative processes of language are normally unconscious and that they are so complicated that they cannot usually be raised to a conscious level. However, when nearly all this generative potential of language (or bodily movement) has been forbidden, removed, the remaining choices left are so simple that they can suddenly be apprehended consciously. Creativity has suddenly become controllable, hence enjoyable. This, however, is an illusion of creativity; in fact this is the sphere where it occurs least. The meaning of symbols I have up to now considered the significance of the type of articulation in formalized language and song in its relation to meaning and power as a sequential process. I now want to turn to the effect on the units of meaning, themselves effected by these transformations. The units are, of course, most often words referred to in the texts of songs or other verbal behaviour. There are also non-linguistic symbols to which reference is made sometimes through dance or other formalized bodily communications. Both these types of symbols are easily amenable to the linguistic analysis that has been used so far. Finally there are sometimes material 'symbols' used in ritual performance, the bread and the wine of the communion ceremony for example. The nature of the 'meaning' of these symbols can, however, also be better understood in terms of what has already been discussed. As is well known, Malinowski, despairing of being able to translate the language of Trobriand ritual (actually, usually this was not ordinary language but incantation, a fact of which he made little), concluded that this language was meaningless except in the full context of ritual situation. Each of the words used could not be understood unless we plunged ourselves in sympathy with the Trobriander into the whole field of the ceremony. This theory that the meaning of words only existed in 'context of situation' was extended by him to a general theory of meaning and as such it ran into considerable trouble right from the first. It was finally shown to need modification by semantic theories deriving from transformational grammar, which, in contradiction to Malinowski, stressed the needs of the individual units of meaning to carry their semantic load in themselves, so to speak irrespective of context, because only in this way could these units of meaning be recombined to create an infinity of fitting utterances, the most famous potential of language (Langoenden, 1968). The criticism of generative linguists is well taken for language in the abstract. However, as Hymes (1971) and Halliday (1973) have pointed out, stressing the potential of language as a tool for the speaker, by which he can say anything he likes at any time, is something of an abstraction, and semantic theories which assume this do so at their peril. In real life people rarely find themselves in a situation where they really can say anything they like. In the kind of situation considered here, situations where the power relations are in evidence, or where religion is involved, the notion of the speaker being able to say anything he likes is plainly ridiculous. As we have seen, the very rules of politeness, of appropriateness, of 40 Ritual, history and pomer Symbols, song, dance and features of articulation 41 formalization, reduce and almost eliminate this potential of language. It is no accident that Malinowski developed his theory of context of situation when considering precisely such phenomena, because here the necessity for recombination does not exist, indeed it is forbidden. Now, since the need for meaning to be contained in the unit depends on the need for recombination, it follows that where recombination is not necessary the semantic load need not be carried in the units themselves. In ideal intellectual discourse the contextual associations of meaning are continuously being sheared off as the units are being re-used in different contexts; but in religious contexts these are allowed to grow and intertwine with each other. When we are dealing with a system of communication which has largely given up the power of creativity, words keep their contextual halo and these, fused into each other, form solidified lumps of meaning within the religious framework. Lyons, in the passage quoted above, would deny meaning to units so permanendy joined. J. R. Firth, when discussing collocation, the regular joining of words together, stresses how this drains the parts of prepositional force, giving them a different total semantic effect which cannot be separated from the context. This is illocutionary force. It also protects the collocated words from the same processes of historical linguistics as the constituents undergo outside the ritual field (Aitchison, 1966). This produces a disjunction between words inside and outside the religious sphere, producing words which have little or no prepositional force, even in themselves. There are many examples of this in nearly all religious rituals and the familiar phenomenon of archaism in religious language can be seen as a product of precisely this phenomenon. Words, there, are continually in the process of'drifting out of meaning'. To say that words are 'drifting out of meaning', that they have little lexical value as communicators, again does not mean that they have little illocutionary force. Indeed it can be argued that it is precisely the imprecision of such words which gives them their social and emotional force. An effect of words and speech acts being formalized is their increasing ambiguity. This is due to the detachment of what is being said from the particular speaker so that he speaks for the kind of person he represents (e.g. the eternal elder) (see above), and because of the detachment of the communication from specific time and place (see above). Ambiguity has much social usefulness and the kind of illocutionary force it" might have has been suggested, among others, by P. Leis (1970) and R. Keesing (1972), but what it does not do is help in an 'explanation'. If words in ritual have little explanatory power but much socially useful ambiguity and are little separated from their context, they begin to perform less as parts of a language and more as things, in the same way as material symbols. It is not surprising, therefore, to find that rituals seem to show a predilection to use not only song and dance but also material things for communication. Turner has repeatedly written about this aspect of ritual and has drawn attention to the 'polysemy or multi-vocality' (Turner, 1967: 50) of material symbols. He further suggests that this explains their value for ritual communication. If, however, we look at communication with material symbols objectively, rather than subjectively, and compare this communication with the more normal communication that it replaces, viz. natural language, then its 'richness* requires major qualification. Material symbols are like a series of words with no, or very little, syntax to articulate them. Symbols in ritual not only follow each other in a relatively (compared to words in language) fixed sequence, but the sequence itself is relatively (again compared to language) meaningless. Material symbols are therefore of their nature like words in formalized communication. They can only be part of a message with very weak propositions! force but as a result gain in ambiguity and hence their illocutionary and emotional force. It is not surprising, therefore, to find that the 'explanations' extracted with such labour by anthropologists from their studies of religion produce the kind of gobbledygook which has caused so much controversy, such as the Nuer statement that 'twins are birds*. In other words, any attempt to explain this sort of statement by putting it into a logical system in the way that Levi-Strauss or Evans-Pritchard or Winch have tried to do is doomed to failure, because it overlooks the fact that these statements are couched in a language where reason has been from the first excluded by the effects of formalization and the subsequent transformation of the units of meaning which it produces. Such representations are wrong because they consciously or unconsciously reintroduce a logical structure of a type which is, of its nature, secular. The effect of formalization and the impossibility of linguistic creativity means that ritual is a kind of tunnel into which one plunges, and where, since there is no possibility of turning either to right or 42 Ritual, history and power left, the only thing to do is to follow; but the reason why this direction has been taken is only misleadingly explained in terms of a conscious choice between equally possible directions. Units in ritual do not follow each other logically, but sequentially, since there is no power in the articulation which links them. The sequence of concepts is given and not accepted; it has no 'truth' conditions. It is in this light that another often noted feature of ritual -repetition - becomes understandable. If one thinks of what is going on as an argument, as is implied by formal or logical analysis, repetition is mere redundancy. If, however, we are not dealing with an argument but with a total bonded experience, repetition is the only possibility for emphasis. A frozen statement cannot be expanded, it can only be made again and again and again. Repetition reminds us that we are not dealing with an argument, since an argument is a basis for another argument, not the basis for the same argument again. In some ways the conclusions reached above appear mainly negative. However, they contain by implication their positive side. The first of these is that a self-contained study of ritual is still possible so long as it avoids two things: 1) jumping from the inside of religious discourse to everyday speech when producing an explanation, and 2) either directly or indirectly using logical forms. These restrictions are, however, not very formidable and a straightforward interpretative technique of a traditional kind is perfectly acceptable. Such a study must simply treat the fantastic statements of the ritual as such and accept them willingly, since the way they are put makes their probability or improbability irrelevant. There is no hidden code to crack, only the examination of the given code in which communication takes place. One problem is that the units of the ritual have to be treated with extra care, since they are, in a way, in a process of'drifting out of meaning' as a result of their isolation from normal communication. In other words these units are losing their propositional meaning to a differing degree. In historical terms this drifting out of meaning of religious symbols is a dialectic since the process is regularly reversed as new units are reintroduced from outside by revivalist movements.7 It follows, from what has been said above, that the interpretation of ritual and religion in terms of what it does is again of primary relevance. This is because as the propositional force of language Symbols, song, dance and features of articulation 43 diminishes, 'what it does' comes to the fore (illocutionary force). The experience of the ritual is an experience fused with its context and therefore only an attempt to explain what this event as a whole is for is an explanation of the content. An attempt to link the context of ritual to the world or society directly does not work because in secular terms religious rituals are mis-statements of reality. Circumcision ceremonies do not, as van Gennep (1909) was at pains to point out, make adults out of little boys, curing ceremonies do not cure, etc., and any attempt to pretend that they do (as is done in the work of so many anthropologists) is wrong from the first. By contrast, the interesting question is the disconnection between the religious statement and the real world, a disconnection which is produced by the mode of communication of ritual. This is our material for study in the same way as the distortion of events and situations by oratory in political situations is our subject matter and we can only understand it when we see how it serves to hide the actual situation and preserve authority. In other words, because the units of meaning are changed so that they can serve purposes only for religious events, and since they are inseparable from the context, there can be no explanation without an explanation of what the event or the context is. There can be no understanding other than an ethic, functional, understanding. There is, however, at first a difficulty in simply treating religious statements in the same way as we treat political statements. While we have seen what is the social effect of the formalization of communication for traditional authority, this might appear less clear for religion. I would argue nonetheless that if we look again at the argument above we will find that it perhaps contains an answer to this much more fundamental problem. Is religion an extreme form of traditional authority? In the case of political oratory, we saw that the sign and the tool of traditional authority was formalized communication and that in the case of religious rituals this formalization is pushed even further. Does that mean that we should consider ritual, and perhaps religion, as a special variety of the political process, an extreme form of traditional authority? The question is much less strange if we bear in mind an example such as that of the circumcision ceremony. In it, 1 m m 1 m 44 Ritual, history and power speech-making by the elders is at first little different from speech-making in a political context, but as the ceremony advances and the formalization of communication increases a change takes over the elders so that they become depersonalized from themselves and transformed into representatives of the dead, gradually less a particular elder than the 'eternal elder'. Another example of this could be given from the way Luapula elders use the first person singular when talking for their dead predecessors (Cunnison, 1959). This process reaches its culmination with possession when the particular individual elder ceases to exist, so to speak, and is replaced by 'another'. The reason for this depersonalization seems to me an automatic result of formalization because, as we noted briefly for traditional oratory, the increase in repetition and illustration from a body of sources and the increasing ambiguity of terms means that the statements of a particular event take the form of a statement for all events of this kind. The effect of a comparison back to a set body of references which are very general is that the person is dissolved into the past, and the future is made to seem an inevitable repetition of this past. Since in this situation all events are nothing but repetitions and examples of these same illustrations, it means that the illustrations themselves, to be effective in this removal of events from time, must be of a very general kind, applicable as reducers of any situation. It will also mean that the authority of the elder, although never greater than when he is employing formalized communication, becomes less and less his own. This problem we already saw when we were discussing power in traditional authority. We saw that as the leader turns to formalization his individual will disappears as he transforms reality in a timeless placeless zone in which everybody is in his right place. In the political context the elder has to fight off rivals. However, in the role of religion this is not necessary any more. Indeed one can say that a political event becomes religious when individual power struggles have become unnecessary. Formalization thus not only removed what is being said from a particular time and a particular place, it has also removed it from the actual speaker, and thus created another supernatural being which the elder is slowly becoming or speaks for. The creation of this other supernatural being is best seen in possession, where the notion that two beings are present, one supernatural and one natural, is explicit. The elder is transformed into an ancestor speaking eternal truth; this trans- Symbols, song, dance and features of articulation 45 formation seems to me the articulation between traditional authority and religion. Indeed, while anthropologists would have no doubt in counting possession as religion and counting an elder speaking in a traditional way in a council as politics, we have seen that it is impossible to draw a sharp line between these two social phenomena. The significance of this is that we should perhaps see the origin of religion in this special strategy of leadership, the use of form for power, which we have found in a lesser form in our study of the communication of traditional authority, we would then see the performance of religion as serving a special form of authority. It might at first be objected that religion is not led and organized by political leaders in a society but seems the possession of everybody. This very problem, however, is itself a feature of traditional authority too and can be understood if we note the two following points: 1) It is precisely through the process of making a power situation appear a fact in the nature of the world that traditional authority works. We have already examined how this is brought about. 2) We have noted how the formalization of communication forces political leaders to adopt two conflicting roles which can be distinguished by the types of language they use. This tendency manifests itself even more clearly where this duality is resolved by the leader employing an orator to speak for him on formal occasions. In such a situation we are very close to the relationship of the secular leader and the priestly leader described for so many societies (it is often the priests who are the orators). This suggestion leads us right away from the present argument but it shows why it might not be for methodological reasons alone that it is fruitful, first to look at politics and then at religion, which is seen as the exercise of a particular form of power rather than look at religion out of the political context and see it as a form of explanation.