Pierre Maranda and Elli Köngas Maranda

Assistant Professors of Anthropology at the University of British Columbia

“Structural Analysis of Oral Tradition.
Introduction.”

Un article publié dans l'ouvrage sous la direction de Pierre Maranda et Elli Köngäs Maranda, Structural Analysis of Oral Tradition, pp. ix-xxxiv. Philadelphia: The University of Pensylvania Press, 1971, 324 pp.

Structural analysis is defined by the rules that govern it. The three rules that follow are criterial to us. The first one is the rejection of eclectic data such as "striking moments" in folk tales (Propp 1958 : 6-7) in favor of total corpora and a holistic approach (Piaget 1968 : 8-10). True enough, a folk tale can partake of several wholes : all other folk tales of the culture, the entire oral literature of the group of its carriers, the culture itself, or the folk tales of all cultures. At present, we are inclined to think that the most important frame of a narrative is its own socio‑cultural background, even in the case of borrowed plots, for plots (and other narrative materials) can only be borrowed if they fit or can be molded to fit the culture, more exactly the level of culture which we would call deep structures.

The second rule is that structural analysis rests on mappings or transformations. These terms are used here in a technical sense, similar to that used in Chomskyan linguistics (for explications, see Maranda 1970, especially on the notions of isomorphism and homomorphism ; Maranda and K6ngds Maranda 1970, cf. Buchler and Selby 1968, and Maranda 1969b, Köngäs Maranda 1969). Instances of narrative communication — which can be captured and archived as texts — then are on the level of performance and can be called surface structures ; under every performance there is a process in which some materials, such as inherited or international tale plots or actual experiences, are transformed to fit the deep ("timeless") structures of the culture.

The third rule bears on the cutting of units. Many efforts have been [x] made in the history of folkloristics — by Veselovskij, Aarne and Krohn, later historic-geographic scholars, and by Boas and his students — to determine narrative units. Let us recall the first revelation in this respect : "Function must be taken as an act of dramatis personae, which is defined from the point of view of its significance for the course of action of a tale as a whole" (Propp 1958 : 20).

It would be well to keep in mind that the basic tenet of structural study in folkloristics (as it has long been in linguistics) is that the units, neither given nor evident, must be found in the corpus itself. The phonemes of one language are identical with those of another only by coincidence ; the analyst must find his phonemes in determining if a change of a phone type makes a change of meaning.

The two last rules are heuristic devices, not necessarily "universals" (Chomsky 1965 : 28-30, 55-59). In the cutting of units, the concepts of distinctive features and substitutability are valuable tools to investigate paradigmatic sets and syntagmatic chains, but their validity can be questioned (Martinet 1955 : 63-75 ; Bar Hillel 1957 ; cf. Lane 1967). Then, in the complementary, synthetic transformational approach, the analogic mathematical use postulates ignorance of the gap between artificial and natural languages (Barbut 1966 ; Piaget 1968). All formalizations in the behavioral sciences are limited by the random nature of history and by the constraints of inadequately controlled environments.

Hubert and Mauss (1897-1898, 1902-1903) and Durkheim and Mauss (1963) started structural analysis in anthropology. In linguistics, the Moscow school and de Saussure made pioneering developments about a decade later. Then Propp (1958) developed his approach independently, inspired by Russian formalism and by B6dier (1893), as he was dissatisfied with Aarne's classification of folk tales. But it is with Lévi-Strauss that structural anthropology and structural folkloristics merged under the catalyst of structural linguistics (1945, 1955, 1958, etc.).

The structural approach seems to have developed faster in linguistics than in anthropology and in folkloristics. In the Prague linguistic circle, Trubetzkoy and Jakobson, among others, gave a strong impetus to the method. A number of studies applied it to delimited cases, and soon structural linguistics was represented by an imposing array of papers and monographs, all related to the same framework. In anthropological [xi] folkloristics, the set of working propositions was neither as economical nor as operational as in linguistics. L'Année sociologique was not the equivalent, from that viewpoint, of the Travaux du Cercle Linguistique de Prague, which came later, and, for that matter, later than Russian formalism. It is probably because the French were concerned with "total social phenomena" that they did not develop a spectacular analytics. Their data were more complex than phonemes.

In contrast to the French, Propp chose to focus on the form of folk tales. Yet he did not overlook their sociological context (1965). His strategy led him to propose a model of greater operational, i.e., algebraic, value than those of Hubert, Mauss, Hertz, and van Gennep.

It would be too long to discuss here the parallelisms, convergences and divergences of structural linguistics, French anthropological folkloristics, and Russian formalism. They are combined, in different proportions, in Lévi-Strauss' mythological works. The interested reader can find substantial if not full coverage in the literature (Lévi-Strauss 1950, 1960b ; Pirkova-Jakobson 1958 ; Evans-Pritchard 1960, especially 1321 ; Todorov 1966 ; Piaget 1968). The Russians, however, deserve special mention since their folkloristics are known almost exclusively through Propp's monograph (1958). They anticipated or adumbrated not only Hjemslev's combinatorics but also transformational analysis and that on the level of discourse.

In the late 1910s until 1928 (when it fell into disrepute "for considerations extrinsic to scholarship" [Jakobson 1965 : 12 ; Pirkova-Jakobson 1958 : vii]), there arose in Russia what was called, derogatively at first, the Formalist school. Concerned with literature and folklore, the Russians stressed the principles on which social anthropology had begun to proliferate in France during the preceding decade and on which de Saussure was to build structural linguistics.

Tynianov, for instance, one of the theoreticians of the group, insisted that functions be defined through what is now known as argument substitutability, on the one hand, and through their interconnectedness in a combinatorial system on the other. This led him to consider a hierarchy of analytic levels — combinatorial, constructive, or sequential, and systemic — which form homogenous ensembles ("series" in his terminology), a definite improvement on the genre‑type‑motif approach (Tynianov 1965 ; cf. Todorov 1965). Along the same lines, Jakobson points out (1965 : 11 ; cf. Harris 1951 : 18) :

[xii]

One could quote the deep insights [of Russian formalists] into the correlation of referential and poetic functions or into the interconnectedness of synchrony and diachrony, and especially into mutability, ordinarily overlooked, in the hierarchy of values. The works applying syntactic principles to the analysis of whole statements and of their dialogic interrelations led to the greatest discoveries of Russian poetics, viz., that of the laws governing the composition of folkloric topics (Propp, Skaftymov) or of literary works.

Propp first showed that "genres" cannot be used as taxonomic principle because of internal inconsistencies. Contents, form, function (in the Malinowskian sense) are not equivalent and thus cannot serve indifferently to classify data. "Plot" is equally unoperational for the same reasons and to rename it "type" when it is summarized does not lead anywhere : identification remains intuitive and unreplicable.

Instead of using "motifs," the Russians classified their documents into what logicians call "function" and "argument" or "term" (this procedure is treated at length elsewhere, Köngäs and Maranda 1962). Accordingly, Propp distinguished dramatis personae (terms) from the dramatic functions they fulfill in a narrative — the latter, he said, defined the former. After a survey of a corpus of one hundred Russian folk tales, he reached the conclusion that thirty‑one functions are enough to account for the composition of the data. The number of dramatis personae is greater, "approximately one hundred and fifty." Finally, there are twenty "classes of transformations which account for mutations, and these rest on sociological factors" (Propp 1965 : 236, 238). The last clause must be emphasized because Propp has been charged with pure formalism (Lévi-Strauss 1960b ; Bremond 1964). In point of fact, the Morphology of the Folktale lends itself to such criticism, but in an article published the same year, 1928, and translated only in 1965, Propp gives evidence of a sure anthropological sense in his propositions on the investigation of transformations. The example on p. xiii is schematized from the paper in question (1965 : 258-259 ; see also 237-244).

Hubert and Mauss proposed models for magic and sacrifice ; Propp followed with one for the folk tale ; then Lévi-Strauss built one for myth. Thus, folklorists have at their disposal the equivalent of de Saussure's langue to which actual informant performances, equivalent to parole, can be referred (cf. Pêcheux 1967 ; Köngäs Maranda 1969 ; Maranda 1969a). This is not to say that the models in question are unalterable.

[xiii]

These are some transformations of the terms, or arguments, of the function abduction" in folk tales ; the transformer is always the sociological context ("milieu humain").

Like la langue, they will be deepened and made more adequate as research goes on.

*   *   *

The reader of this volume will notice the methodological consistency that prevails in structural folkloristics from Hubert and Mauss to most of the papers in this collection. He will also see how the model underlying these studies is indeed refined and tested, and we hope that he will find it a challenge as well as an invitation to go further.

This volume was planned as a review of the uses of the structural method in folkloristic anthropology, and a wide variety of approaches was included. The contributors were asked to begin their articles with brief statements of their operational concepts which were to be used in an analysis of one or more items of oral tradition.

The papers are grouped into three parts devoted respectively to (1) myth, (2) ritual and certain other forms, and (3) myth in culture contact.

Lévi-Strauss, one of the pioneers of the structural analysis of myth, opens the first section with an article which introduces the analytic concept of deduction and which is also a concise illustration of his method. The next three papers put structuralism to a test. Leach and Hymes apply it to materials previously analyzed in functionalist and psychosocial frameworks. Greimas starts with a structural analysis by Lévi-Strauss and tests it in the light of semantic theory.

The second section consists of studies of diverse folkloric genres, [xiv] ordered here according to their decreasing semantic complexity. Rituals string highly condensed and multivalent symbols ; they often share the socio‑cosmological implications and the language of myth. Like rituals, folk dramas use multiple encoding (movement, music, and words), but they do not carry as much information per time unit since they display their components more at length and more explicitly. Folk songs are doubly encoded (music and text), but folk tales and riddles are encoded only verbally. Neither folk songs nor folk tales require the active participation of the audience in the way that riddles do. However, whereas folk song texts are often reduced to a few redundant statements, and whereas riddle components, whatever their intricacy, are resolved in one or a few stock answers, folk tales carry as much information as folk dramas although in a less elaborate fashion. Thus, from the approach adopted in this volume, folk tales are more complex than riddles, which in turn are more complex than folk songs.

The last section returns to myth, this time in the context of culture contact. The data studied so far are traditional messages ; in contrast, these stories are "in the making," seeking a stable form. Both papers in Section VII deal with the G6 of Central Brazil, one with the reaction of the native society to the contact situation, the other with a post‑contact reworking of European messages. The last paper is also an attempt to assess the value of desk studies. The social anthropologist who collected the variants comments on the analytic sketches drawn by three contributors.

Two papers report on the processing of data collected by their authors in the field (Turner, Peacock). Three tackle texts coming from areas where the analysts have conducted ethnographic research (Lévi-Strauss, Köngäs Maranda, da Matta). The remaining six are desk studies. Hymes, Köngäs Maranda, Turner, and Peacock have worked on materials in their original versions and have also resorted to philology. The reader will see that, as expected, firsthand data and the knowledge of the language in which the documents are expressed yield more comprehensive interpretations. On the other hand, as demonstrated by Hymes' reanalysis of Jacobs' sources, knowledge of the language is no insurance against strategic shortcomings. Linguistic and semantic considerations have been voiced that seem to justify work on translated materials. Dundes repeats in another context Lévi-Strauss' view that the contents of oral tradition can undergo translation without distortions, contrary to what happens in poetry (Lévi-Strauss 1958a : Ch.11). The [xv] "Experiment" at the end of the volume throws some light on this issue. Finally, all contributors except Greimas use cross‑references to complementary texts, i.e., contingency analysis. Frequency analysis is almost completely ignored (one reference in Hymes) except by Lomax and Halifax, who rely exclusively on it, and by Maranda in the "Experiment." Doubtless, these two strategies will eventually be brought together as semantic weight (the degree of redundancy of certain actors, symbols, and actions) may be significant in the constitution of folkloric semantic spaces (cf. Maranda 1967a, 1967b, 1968, in press : Ch. 4).

Lévi-Strauss' paper recasts a section of Volume II of his continuing Mythologiques (1964, 1967, 1968). The focus of the "Deduction of the Crane" is on three Guianese myths taken from neighboring peoples (Warrau, Carib, and Macushi) ; this provides an instance of the author's use of the comparative method within culture areas (cf. Lévi-Strauss 1958b).

The Kantian concept of deduction is introduced to investigate the mechanisms of mythopoeic thought. A mythic universe is the product of native deductive processes which the investigator attempts to map out. Transformational analysis is therefore based on the unraveling of mental operations, and it throws light on cognition as well as on its expression. After Kant, Lévi-Strauss distinguishes two types of deduction-empirical and transcendental. "Empirical deduction is based on the perception of both similarities and (cause‑effect) contingencies. As for transcendental deduction, it requires a true reasoning process rather than a simple judgment." Empirical deduction may be direct or indirect. Direct empirical deduction comes from experience and analogy ; indirect empirical deduction from the inversion of analogic contents (cf. Greimas' concepts of "posed" and "reversed contents"). In contrast, transcendental deduction does not stem from observation but from logical requirements ; it rests on a relation between concepts no longer bound to external reality but connected "according to their compatibilities and incompatibilities in the architecture of the mind" (Lévi-Strauss 1966 : 407). Mythic constructs not only display but also harness a conceptual universe by the conjunction of empirical and transcendental deductions. To understand such cognitive processes, the analyst resorts to a critique in which he seeks the conditions, to be fulfilled by a social structure, that would generate a given corpus of myths (the illusion that myths only reflect social structure must be preliminarily dispelled) (Lévi-Strauss 1967 : 294). Actually, a social structure and a mythology may [xvi] follow divergent paths, each evolving according to different determinants, and mutually influencing each other. Lévi-Strauss maintains, however, that, in the last analysis, the decisive factors are technological and economic (1962a : 153).

The mythic deduction is always dialectical in nature, and it is never purely cyclical but rather spiral-like. When the analyst follows a path returning to its point of departure (at the end of a series of transformations), he does not find complete reversibility : he reaches the same longitude but at a different latitude. The distance between the two latitudinal points — beginning and end of the trajectory — is a meaningful semantic fact. The distance may be a matter of framework, code, or lexicon, according to the strategy adopted in the investigation (Lévi-Strauss 1967 : 341).

Here as elsewhere (1958a, 1962a), Lévi-Strauss points out the role of metaphors and metonymies in the constitution of mythic messages. Ki3ngiis Maranda puts the same tropes to use in her analysis of riddles and shows how they generate simple and complex forms. Both authors use a generative approach to test semantic hypotheses and they show how intersecting elements of paradigmatic sets form syntagms which, intersecting in turn with each other according to a combinatorial model, form larger sets or higher orders. Thus, a corpus presents the aspect of a tree structure or of several partly overlapping tree structures, the number of summits depending on how far a culture carries the process of synthesis.

This "degree of order" could therefore be used as a "single internal and formal criterion for a typology." It would consist in a measurement of the breadth and depth of given mythic universes ; it would at the same time throw light on cognitive systems and mythopeic thought in different societies. The point at which "the myths of a region or a population... cease the process of composition" would provide the gauge to fathom the domain. The investigator would then brace (to use Lévi-Strauss' topographic metaphor [1967 : 216]) empirical deductions and semantic fields, upon which the links established by transcendental deduction would become apparent, and ethnoscience would truly "blossom into a logic and a philosophy." Thus, cognitive psychology and ethnoscience would meet, as wished by Brown in his discussion of the conference "Transcultural Studies in Cognition" (1964 : 251 : cf. Goodenough 1964 : 39), for, if the mind is a "template" for ethnoscience and a "transformer" for Piaget, the conjunction of both is operated in several [xvii] of the papers grouped in this collection where the breaking of codes built of templates is done through transformational analysis.

In the conclusion of Mythologiques II (1967 : 405), Lévi-Strauss compares his work to that of deciphering an unknown language by breaking the code with the help of multilingual documents all conveying the same message. The prevailing codes disentangled by the author in South American myths are culinary (techno‑economic), auditory, sociological, and cosmological. Lévi-Strauss adds that these, according to the myths themselves, are not equivalent. The operational value of the auditory code is greater than that of the others since it provides "a common language into which the messages of techno-economic, sociological, and cosmological codes can be translated."

"The Deduction of the Crane" shows the four codes in action. First, a contents/container contrast (cf. Lévi-Strauss 1958a : Ch. 12) gives a key for a first reading of the texts in terms of "fillin-up" (active and passive, from the bottom and from the top) and "emptying." But the first contrast is intelligible only with reference to a dialectic of literal and symbolic meanings related to the alimentary and sexual codes. Then, these and the meteorological and astronomical coordinates also present in the myths are subsumed under an auditory code. Empirical deduction now opens on a native transcendental model, a critique of codes, as it were, which leads to the conclusion that, among all codes, the articulation of sound is still the most efficient. "The disjointed judgments based on empirical deduction which occur in the myths [thus] appear coherent."

Leach gives a dispassionate reinterpretation of Radcliffe‑Brown's data. He draws as much as possible on Radcliffe-Brown's own insights, exploits all the possibilities found in the Andaman Islanders. This enhances the interest of his own analysis. In clear, didactic steps, Leach shows how the pitfalls of functionalism can be avoided. To paraphrase one of his last paragraphs, functionalism concentrates on individual symbolic meanings instead of on the special structural position which give individual symbols their meanings. It cannot thus apply the principle of substitutability, unravel transformations, or disclose rules. Furthermore, only structural analysis allows one to formulate and test semantic hypotheses.

Leach's structural theory of tabu is a useful tool in the analysis of semantic spaces. It shows how systems of classification operate, how [xviii] the continuous is distributed into discrete sets (Leach 1964 ; cf. Lévi-Strauss 1962a, 1964), and it points out the function of margins as border categories. Consequently, the analyst can relate behavioral norms to cognition ; Leach's article illustrates that prediction is also within his reach. Finally, the operational definition of margins is most valuable in the investigation of metaphoric and metonymic processes.

Not all structuralists would agree with Leach that analysis stops where his does. "Why the various agents of transformation should hang together" may be viewed as a semantic problem. The Andamanese agents of metamorphoses (fire, killing, dancing, and ochre painting) may be located on a continuum : they may be expressions of a code which can be broken. For instance, the Andamanese message about the order of the universe is perhaps encoded in kinesthetic terms. Transitive motion and intransitive motion would be the basic contrast. Transitive motion to alter (marking with fire, piercing with arrows) yields order, whereas transitive motion to destroy (crushing the cicada) causes disorder. Concomitantly, planned (dancing) and accidental (falling) intransitive motion leads to similarly opposed results. The relation could then be proposed

If the relation holds, the message is stated in a kinesthetic code whose range is restricted to refrained action and mastered automotion, to altering without destroying and to accomplishing a skilled performance without faux pas. Leach's theory of tabu would account for this impact of negative norms on the structure of the code. Explorations of the Andamanese corpus from this angle would test the hypothesis.

In his defense of the girl charged with murder by Jacobs, Hymes gives a lesson in the uses of evidence. The article is instructive not only on how to conduct an analysis but also on how to gather one's data.

Along the lines of structural ethnology, Hymes raises the question of whether the object of folkloristics should be understood as "traditional" at all, a point made also by Kuusi (1959), and Köngäs Maranda (1963). The verbalizing system of a society does not constitute merely a device to repeat ancient lore, but is a mechanism to interpret events [xix] through an explicitly or implicitly known structure (cf. Boas 1916 : 332-334). As Hymes says :

Folklorists have commonly identified their object of study, traditional materials, as a matter of texts, not of underlying rules.... The occurrence of a reworked European tale in an Indian pueblo... may evoke amusement, or embarrassment, if one thinks of one's goal as autochtonous texts. If one thinks of one's goal as natively valid rules, such a case may be an invaluable opportunity to verify the principles of the native genre through an instance of their productivity [emphasis added].

The point is to lay bare the implicit model underlying a given mythopoeic universe and to specify the formal regularities that characterize it.

The lack of information on actors which Hymes is inclined to consider typical of Clackamas oral literature is a widespread feature of folklore in general. Aristotle believed, and Propp has reemphasized the principle, that dramatis personae are defined by their actions. Relevantly, Hymes shows that proper names are significant and provide concise identifiers which, in place of explicit psychological attributes, are pregnant in terms of a dramatic system (cf. Lévi-Strauss 1960b ; 1962a, Ch. 7).

Finally, an elegant application of Lévi-Strauss' method reveals the structure of the myth, which is at the same time placed in its context. The demonstration advances through the permutations of the components to a dialectical analysis on the basis of the combinatorial model.

The replicability of structural analysis is sometimes doubted by folklorists (cf. Robinson and Joiner). The construction of combinatorial models (Lévi-Strauss 1962a, 1962b) is a simple matter, once the axes of permutation have been defined with the help of contrastive features (Lévi-Strauss 1958b : Ch. 11). Leach points out, in the first part of his paper, how such features can be recognized on the level of the narrative sequence (contrast between beginning and end, cf. K6nglis and Maranda 1962). Greimas' paper is a step toward making the method explicit in order to achieve replicability.

The author of Sémantique structurale (Greimas 1966) reanalyzes the Bororo myth of reference which is the starting point of Le Cru et le cuit (Lévi-Strauss 1964). The paper is based on semantic theory, but explores concepts used by folklorists, such as contents, structure, narrative model, code, and combinatorics.

[xx]

In the first two sections, which are theoretical, Greimas defines "topical contents" with the help of a restrictive interpretation of Lévi-Strauss' canonical formula of the structure of myths (1958b : Ch. 11), restrictive because the formula proposes to encompass a whole series of variants whereas Greimas applies it to only one variant. Thus, instead of considering the inversions the components of a myth undergo in its transformations, this paper presents the outcome of the myth as the inversion of the initial situation. This is parallel to Leach's method.

Greimas uses the contrasting pair "posed"/"reversed" contents to describe the pattern. According to native thought, myths of origins of cultural implements thus imply a Parmenidian axiom that nothing comes from nothing and that if man is now endowed with culture, some other beings must have been deprived of it for his benefit. The present possession of fire by man means that there was a time when he did not have it ("reversed contents"). Accordingly, myths will be found in which the initial situation depicts a fireless humanity and the possession of fire is attributed to some supernaturals (the jaguar among the G8, the gods among the Greeks). Then the action of a mediator (an uninitiated Gê boy — not yet socialized and thus suited to be a link between nature and culture — Prometheus) will reverse the initial situation. "Reversed contents" is the inverse of the existing cultural situation, attained at the end of the myth and named by Greimas "posed contents." Such inversion should not be confused with motifs in the pattern of "the cheater cheated," which are only secondary expressions of inversion mechanisms. "The Dictionary and the Code" exemplifies his dialectical deductions which are first defined in "The Framework."

The interplay of terms and functions in Lévi-Strauss' formula figures in Greimas' approach as that between the lexemic and sememic levels along different isotopies. The concept of isotopy enables Greimas to isolate levels and measure redundancy. Two isotopies are necessary and sufficient. The first one is parallel with signs in linguistics, the second with semes in semantics. Lexemes are shown to be rough units which can be handled only when broken into their semantic components, which constitute the code. The rules governing the use of the code are stated in the combinatorial model.

A mediator is defined as an operator of inversion which effects the passage from reversed to posed contents. The mediator is a lexeme, "a meta‑subject of dialectical transformations," i.e., a dramatis persona capable of reconciling oppositions and thus a significant clue to the [xxi] code (cf. Köngäs and Maranda 1962 ; Maranda 1963 : Appendices). Here Greimas leaves Propp's model in order to be able to account for the internal motivations of such functions as tests and the qualifications of a hero. On the lexemic level only, the analyst cannot reach the code, he cannot interpret, he can only describe the story.

In the reanalysis of the Bororo myth (Sections 3 and 4 of the paper), the priority of the alimentary code over the cosmological code is assessed. There sememic and lexemic isotopies merge. The rest of the paper tests the validity of the interpretation : if it breaks the code, all the components fall into place, and no residues are left.

Turner's paper is based on exegeses elicited from his cooperative Ndembu informants (cf. also Turner 1962, 1964a, 1964b). A metalanguage is first presented, then applied to the description of a ritual.

Superficially, ritual appears as a sequence of operations performed on a subject. Turner, in contrast, gives priority to the semantic fields broached by the ritual components and shows how the "syntax" comes from the conjunction of the prestressed (Lévi-Strauss 1966b) symbols. The analytic dimensions and bases singled out all belong to "content" in formalist terminology : nonetheless, when each is fully displayed, a structure emerges.

As in the myth analyses discussed above, inversion is now shown to be also a basic feature of rituals. In this case, the woman who is at the outset of the ritual too masculine is feminized in the outcome. The mechanism is not a simple inversion, for the initial aggression is converted into fecundity. The woman "is being compared with a male shedder of blood and ... it is desired to convert her into a fruitful wife. ... It would seem, too, from the context that the masculine strivings of the woman are not merely annulled but put at the service of her orthodox role." The differences is one between "cyclical" and "spiral" models (K6ngiis and Maranda 1962). The author ends by interpreting his data as a counterpart of witchcraft. His conclusion exemplifies social anthropology in the holistic tradition.

Peacock's materials, like Turner's, consist partly of verbal, partly of nonverbal behavior. Like Turner, he has also collected the data himself and elicited participants' comments. Although Peacock, inspired by Burke and Parsons, draws on a conjunction of approaches devised for literary and sociological analyses, his working concepts are close to [xxii] those used by most contributors, such as the status o f the characters, outcome, the agency which brings about the outcome, cosmology. In addition, he takes into account setting and time, as commanded by his data.

Peacock's contribution deals in part with the semantic aspect of culture change, like da Matta's paper and the experiment at the end of this volume. He shows that, as in the case of rituals, a linear, sequential analysis of a plot does not reveal its full message. An analysis in depth of the cosmological categories in which the action is rooted is at least as significant as the unfolding of the action itself (cf. Goldman 1955). To quote Peacock's metaphor, "the alus-kasar cosmology chains each class to a separate side of the cosmos and weights each down with a symbolic baggage so that either side would find it a chore to cross to the other side. Can it be fortuitous that such a cosmology should be associated with the I‑plot image of society : a society composed of classes that cannot marry ?" Similarly, the other categories (madju-kuna), which intersect with proletariat and elite, will be more exploited in Peacock's Type 11 society. Cosmological categories that cut across both elite and proletariat build a semantic bridge which will be taken as the way to class permeability.

In his conclusion, Peacock mentions the importance of collecting several variants of the same data to assess typicality. This refers to a practice which anthropologists should learn from folklorists. Informants should not be repressed merely because they want to "tell more of the same" ; only the comparison of different variants lays bare the manipulation of symbols, conscious or unwitting, practiced from individual to individual and from one occasion to another. (See also Dundes' remarks on what he calls "individual relativism," and Newman 1964 : 271.) In agreement with Leach (1954 ; cf. Firth 1960), Peacock draws attention to the sociological impact of symbol manipulations, where basic orientations and social dynamisms are eminently perceptible. More than missionaries, native prophets are effective on that plane because they know how to manipulate their own collective representations — they are intimately familiar with their own cultural codes — while missionary teachings are so encoded that semantic incongruencies are not easily resolved (cf. Lanternari 1960). Here again, the knowledge of the code is paramount and it may well have much more predictive value for the anthropologist and other social scientists than, say, economic analysis or the rate of increase of literacy.

[xxiii]

In structural studies of folklore, two main trends can be distinguished. One stems from Lévi-Strauss, the other from Propp (1958) — see Köngäs and Maranda (1962) ; Leach, pd. (1967), Nutini and Buchler, eds. (in press) ; Maranda and Pouillon, eds. (1970) ; Bremond (1964, 1966) : Greimas (this volume) ; and Dundes (1964). The two approaches converge to a certain degree (cf. Pirkova-Jakobson 1958), but the divergences have been emphasized by Lévi-Strauss (1960c) ; Bremond (1964) ; and Dundes (1964). Propp's study (1958) lays out the linear structure of Russian folk tales which is said to be immutable and uniform. The neatness of this sequential structure has led Proppians to demand that folklorists focus on it and leave aside other levels of investigation even if admittedly structural, such as linguistic style ("texture") or sociological or cosmological contexts. Lévi-Strauss has proposed a canonical formula for the analysis of myth (1958a : Ch. 11) but, contrary to Propp, he finds the structure of a corpus in paradigms of minimal units (mythemes) as well as in their syntagmatic combinations to form sequences.

Much of the controversy will have to be revised in the light of the publication in French of a 1929 paper by Propp (Todorov, ed. 1965). The Russian's sociological concern and his adumbration of transformational analysis (especially pp. 237-244, 258-259) indicate that he paid much more attention to general context and to contents than his critiques and followers have assumed.

In his paper, Dundes refers to his objections to what be calls "the rearrangement school." Such criticisms, however, overlook the fact that Lévi-Strauss considers myth as both parole and langue. As parole, myth exhibits time‑bound events which are essentially irreversible. As langue, myth is an expression of reversible time. Lévi-Strauss compares this distinction to that between musical harmony and melody line. Contrary to Dundes' contention, the semantic nuclei which are found by grouping mythemes are not only depicted in the myth ; they make the myth and its structure. These nuclei help the investigator spot the center of gravity of a story. Propp himself emphasizes the importance of reiterated elements. As he says, one can study constituent parts without reference to the plot (sujet) ; the study of the vertical columns reveals the norms and paths of transformations. Moreover Propp says, what is true of each particular element will also be true of the general design (Propp 1965 : 237 ; on form and meaning, see also Goodenough 1964 : 37 and Lévi-Strauss 1958b : 240-241).

[xxiv]

Dundes formalizes his data in terms of Proppian functions, which he renames, inspired by Pike's terminology, "motifemes." Five motifemes describe the sequential patterns he deals with, and their order is commanded by a move from a state to its converse. Dundes therefore goes further than Propp (1958), who did not carry his analysis beyond the descriptive level, and his operational framework is consequently similar to that of the other contributors — the emphasis on mediating processes (the agents that bring about the passage from a state to another) varying from paper to paper.

The comparison made by Dundes between African and American Indian data of the same type are sociologically and typologically significant. The author contrasts contract (African) /interdiction (American Indian), which are functionally equivalent in his corpus, and the presence (Africa) /absence (American Indian) of friendship in the frame, from which he draws sociological and psychological conclusions. The paper evidences the great interest of comparing structures of narratives of the same type in two broad and unrelated culture areas, which reveals differences between the parameters of different universes.

Metaphor, metonomy, and transformation are Köngäs Maranda's analytic concepts in her study of riddles. With proverbs and jokes, riddles are the shortest and also among the most formalized items of oral tradition. It is as though stylistic intensification compensated for lack of thematic elaboration.

Myths, rituals, songs, jokes, and folk tales are fairly universally found in indigenous cultures. Why is it that so many societies do not know riddles or proverbs (cf. Boas 1916 : 338) ? What kind of an index does this lack provide ? Or could the absence be viewed from another angle, and could it be maintained that proverbs, riddles, and myths are in free variation so that the same cognitive function may be fulfilled by any of them ? Greimas' concepts of reversed and posed contents could be applied to the topical analysis of riddles : Lévi-Strauss (1960a) has also hinted at the isomorphism of myths and riddles ; but fundamental problems are barely beginning to be posed in this domain. Are myths solved riddles on the origin and nature of culture and the world, and riddles a "catechism" of mythology ? Structurally, proverbs, riddles, and jokes are parallel (cf. Greimas [1966] and Morin [1966] on jokes, K6ngds and Maranda [1962] and Maranda [in press : Ch. 4] on proverbs). Studies in depth of these forms would doubtless contribute as [xxv] significantly to the unravelling of world views and cognitive processes as does the study of myth, ritual, and folk tales.

The metaphoric structure of Finnish riddles, brings together man and his tools. Tools are not seen as an extension of man but as a "substitution" ; they are depicted as humans. The metaphoric operator that maps animate humans onto inanimate objects and vice versa often rests on metonymies (body parts and parts of objects). In this sense, riddle imagery is, as it were, a "metasemantics" of the language, reflecting upon and reviving metaphors which are too faded to be sensed as such any longer. If this proposition can be generalized and confirmed by the study of other corpora, riddles would seem to be a form of semantic control (by reinforcement mechanisms) and a critique of the use of words. When the question "What has legs but cannot walk ?" receives the answer "A table," the metaphorical dimension of the language is brought forth (cf. Lévi‑Strauss 1966 on sens propre and sens figuré). The riddler's way of thinking is sensitized to a native model which tolerates extensions of meaning in certain directions and the semantic drill reactivates the bearing of a tradition.

The sexual division of labor plays an important part in the author's corpus. This division is correlated with the contrasts outdoors/indoors and mobility/stability, extremes of the male/female parameter. Meteorological and cosmological codes are subordinate to sociological and technological codes (cf. Lévi-Strauss' paper and those by da Matta and by Maranda on the same codes). It is as if Finnish peasant society, as it explores itself in these riddles, ignored transcendental problems or posed and solved them in sociological and technological terms.

Lomax's and Halifax's paper is the only one in this volume to represent broad crosscultural analysis (Dundes contrasts African and North American dramatic situations but he does not attempt to correlate several factors as do Lomax and Halifax). It also draws on computerization, like Maranda in the experiment. But here, the procedure is inverted in that the computer program has inspired an analysis conducted by hand. The authors have adopted this strategy in order to obviate to program shortcomings with respect to homonymy. If texts are fed into the machine without preliminary editing (Colby's procedure, followed by the authors — cf. the pioneering work of Sebeok 1960 and Maranda 1967a, 1967b), i.e., without preliminary solution of the ambiguities inherent in natural languages, the output will indeed contain "filing [xxvi] cabinets of error" as Lomax and Halifax put it. So, they decided to do the editing as they went along coding their data.

The paper makes the important point that song texts are low entropy messages, repeating concise "cultural statements" more directly and immediately than even proverbs. The paper also exemplifies a quantitative approach not yet recognized by most students of oral tradition. In fact, it may very well be that semantic intersections are more criterial than the number of times a lexeme (in Greimas' terminology) is mentioned. But it is only when the results of quantitative and qualitative analyses of a corpus are compared that the respective bearing of either method can be evaluated (cf. Lévi-Strauss 1964, 1967, 1968, and Maranda 1967a).

Da Matta contrasts a stabilized, traditional myth with a recent one whose ending is still flexible and which seems to be in search of its final form as the society itself is trying to cope with a new historical situation. Hymes emphasizes the importance of studying lore in the making as access to the mechanisms of mythopoesis ; da Matta provides such a case.

The fire myth has all the features of a classical origin story running parallel with the Gê initiation ritual and built according to standard Gê combinatorics (cf. Lévi-Strauss 1964). The Auké "anti‑myth" describes the uninitiated and uninitiatable boy who, rejected by traditional society, becomes the European emperor of Brazil ; the pattern of successful mediation in the myth becomes one of failing mediation in the antimyth. Thus, the socially defined boy and the helpful jaguar contribute to native technology in the fire myth ; in the Auk6 story, the socially undefinable boy becomes jaguar at will, is killed by the fire originally brought to the people by a jaguar, and fails in his attempt to introduce a new technological (European) phase. Cultural achievement is depicted here with the telling lines of a well‑mastered demonstration ; bewilderment leads there to recoil, although available mediators are merged into a multivalent hero and although ultimate sanctions are called on. But the same message is stated in both narratives : Technology brings civilization about and it does it only if society accepts innovation.

Needless to say, the acculturation problems faced by the Timbira are common to all traditional societies in these decades. All have origin stories, a stock of mediators to account for cultural modifications, and [xxvii] a semantic combinatorial model, which together enable them to revamp their conceptual framework. But not all react to extraneous forces like the Timbira. Demographic factors may be responsible for the strategies adopted, small peoples behaving like losers, more numerous ones taking bolder attitudes (cf. Peacock's paper).

The editors' aim in proposing the "experiment" was to assess, on a restricted basis, the value of the long practice of library work in folkloristics. The decision to use European‑inspired Sherente texts was motivated by the fact that (1) European elements offered outsiders some grasp of the data ; (2) native reinterpretations could be perceived ; (3) the data bear witness to acculturation processes, a situation of great interest to modern studies of oral tradition ; and (4) the social anthropologist who bad collected the texts was willing to comment on the contributors' essays. The three who have accepted the challenge represent as many viewpoints ; convergences are all the more significant.

Dundes rightly draws attention to the preliminary problem of identification (but he does not take into account that over one hundred pages are devoted to the structure of Sherente tales in Lévi-Strauss [1964]). The author's Proppian assumptions (Propp 1958) are evident when he declares that these Sherente data are not amenable to structural analysis because the sequential pattern is European. However, he brings the sequential aspect back under his focus when he points out the inversion of the "industry moral" from one of the versions to another.

Dundes deals with the "contents" of the tales. He stresses significant aspects that are congruent with da Matta's interpretation of the Auké story. He also implicitly shows that the tales go further than what da Matta characterizes, after Balandier, as the "first movement of reaction to colonial domination," since they represent an attempt at coexistence accompanied by value contamination.

In contrast with Dundes, Leach looks for structure on the level of the dramatis personae. The main characters, White and Indians, are analyzed into four components that undergo transformations in the course of the narrative. In his second step, Leach takes up broader units, the sequences, which he also contrasts according to their components. On this basis, he provides socio‑psychological interpretations that receive at least partial confirmation from the ethnography.

Dundes' and Leach's queries reveal the type of information they deem relevant to structural analysis. Leach calls for ethnographic in [xxviii] formation in his queries 1 to 4 and 8 to 10 ; 5 to 7 and 11 are philological and exegetical. Dundes' questions show a slightly different emphasis : like Leach, he urges field workers to gather more contextual data and exegeses (cf. Turner's and Peacock's papers). But although Leach's interpretation leans toward psychological sketches, it is Dundes, despite his more sociological interpretation, who raises psychological issues.

Much groundwork was done by hand before Maranda submitted the data to computerized analysis. Ambiguities were resolved, comparable units used ; in brief, the data were translated from the natural language in which they are expressed into an analytic language prior to machine analysis. Episodes and "sentences," defined operationally, were analyzed first, then a few nodes evaluated as meaningful because of their positions in the narrative. The computer in fact applied a data filter to the documents. This represents only one of the several routines available for semantic investigations. Given the time it takes to prepare adequate inputs, one may ask whether computerization really pays off in the end. Maranda's experience suggests that it does, were it only because it compels the analyst to elaborate careful algorithms (cf. Granger 1960).

Maranda's conclusion is congruent with da Matta's interpretation of the Timbira myths. The two G6 groups seem to pose problems of culture contact and acculturation in a technological code which, after all, is not surprising.

The informants' portraits which open Maybury-Lewis' comments do not assess directly Leach's concise suggestions but the latter's inferences can be paired to the features delineated by the ethnographer. Wakuke, an "underdog" who "accepts his position" (Leach) is described by Maybury-Lewis as a "true Sherente" who "did not, like so many Sherente, blame the rapacity of the white man for all the evils which had befallen his tribe" ; and Wakuke insisted that the Sherente were responsible for many of their own troubles. Suzaur6's variants are seen by Leach as coming from "one who is very anxious to disavow his Indian origins which he holds in contempt." According to Maybury‑Lewis, the informant "was well thought of by the local Brazilians" ; his plans were such that they "flew in the face of tradition" ; additionally, he belonged to a group "not regarded as 'true Sherente' but as latecomers who were integrated to the tribe" ; finally, he failed in his attempt to found a rival [xxix] village. Leach sees Tinkwa's variant as Indian oriented, minimizing the role of the civilisados and tending to link "civilized Indians" with "Indians." The ethnographer writes "Tinkwa, unlike Wakuke and Suzaur6, is a hardened white hater who bitterly resents the encroachments and machinations ... of the local civilisados."

The present situation of the Sherente as described by Maybury-Lewis confirms the three analysts' interpretations. These Indians are a small group trying to define and keep their identity in the face of the threatening backwoods Brazilians who surround and outnumber them.

Maybury-Lewis' discussion of the points made or raised by the contributors is clear and need not be reviewed here. Attention may be drawn, nonetheless, to the clarification by ethnographic knowledge of the metamorphosis of unbaptized children into pigs (but cf. Lévi-Strauss 1964 : 92-111 ; 1966 : 293-294). The introduction of an incest component in Variants I and IV is also interestingly interpreted by reference to firsthand data on the Sherente conception of sex.

The ethnographer draws conclusions on the bearing of the experiment. Library analysis is neither impossible nor pointless ; it may even contribute new insights. Its most articulate justification is probably found in Lévi-Strauss (1966 : 302-307). According to him, if ethnographic knowledge is absolutely indispensable at the beginning stages of the investigation, it loses importance gradually as the semantic universe becomes more intelligible. Then, "each myth's context consists more and more of other myths and less and less of the customs, beliefs, and rituals of the particular population from which comes the myth in question" (1966 : 305 ; cf. Boas 1916 : 330).

*  *  *

Two points will be underscored to end this introduction. (1) The papers grouped in this volume show that the structural and semantic anthropological approach to oral tradition is different from traditional folkloristics. Anyone familiar with the literature will agree on this. It is also hoped that the methods and concepts used in the papers will lead scholars in the field to submit their own procedures to a keen evaluation. (2) The materials of folkloristic anthropology, far from being on the wane, are outgrowing the number of field workers. Historical pressures on traditional societies bring about changes whose semantic and cognitive aspects are of great theoretical and practical import. This should be kept in mind by scholars designing field research.

[xxx]

Anthony S. Kroch must be thanked for the translation of L6viStrauss' paper from the French, and Mrs. Kipnis Clougher for her translation from the French of Greimas' paper. Thanks are also due to the Radcliffe Institute of Harvard University for the financial help which has made possible the preparation of the manuscript ; its director, Dean Constance Smith, is entitled to special gratitude.

PIERRE MARANDA AND ELLI KÖNGÄS MARANDA
Harvard University

REFERENCES

BARBUT, MARC

1966 Sur le sens du mot structure en mathématiques. Les Temps Modernes, 246 : 791-814.

BAR HILLEL, YOSHUA

1957 Three methodological remarks on Fundamentals of Language. Word, 13.

BÉDIER, JOSEPH

1893 Les fabliaux. Paris.

BOAS, FRANZ

1916 Tsimshian mythology. Annual Report 31, Bureau of American Ethnology. Washington, The Smithsonian Institution, pp. 29-979.

BREMOND, CLAUDE

1964 Le message narratif. Communications, 4 : 4-32.

1966 La logique des possibles narratifs. Communications, 8 : 60-76.

BROWN, ROGER

1964 Discussion of the conference. In A. K. Romney and R. G. D'Andrade, eds., Transcultural studies in cognition. American Anthropologist, 66 : 243-253.

BUCHLER, IRA R. and HENRY A. SELBY

1968 A formal study of myth. Center for Intercultural Studies in Folklore and Oral History, Monograph Series No. 1. Austin, The University of Texas Press.

CHOMSKY, NOAM

1965 Aspects of the theory of syntax. Cambridge, Massachusetts Institute of Technology Press.

DUNDES, ALAN

1964 The morphology of North American Indian folktales. Folklore Fellows Communications No. 195. Helsinki, Academia Scientiarum Fennica.

[xxxi]

DURKHEIM, ÉMILE and MARCEL MAUSS

1963 Primitive classification. Translated by Rodney Needham. Chicago, University of Chicago Press. (De quelques formes de classification primitive. Année sociologique, 3, 1901-1902.)

EVANS-PRITCHARD, EDWARD E.

1960 Introduction. In R. Hertz, Death and the right hand. Glencoe, Illinois, The Free Press, pp. 9-24.

FIRTH, RAYMOND

1960 The plasticity of myth. Ethnologica, n.s. 2 : 181-188.

GOLDMAN, LUCIEN

1955 Racine. Paris, Plon.

GOODENOUGH, WARD H.

1957 Cultural anthropology and linguistics. In P. L. Garvin, ed., Report on the Seventh Annual Round Table Meeting on Linguistics and Language Study. Washington, Georgetown University Press, pp. 167-173.

1964 Reprint of 1957. In D. Hymes, ed., Language in culture and society. New York, Harper & Row, pp. 36-39.

GRANGER, GEORGES G.

1960 Pensée formelle et sciences de l'homme. Paris, Aubier.

GREIMAS, A. JULIEN

1966 Sémantique structurale. Paris, Larousse.

HARRIS, ZELLIG S.

1951 Methods in structural linguistics. Chicago, University of Chicago Press.

HUBERT, HENRI and MARCEL MAUSS

1897-1898 Essai sur la nature et la fonction du sacrifice. Année Sociologique, 2.

1902-1903 Esquisse d'une théorie générale de la magie. Année Sociologique 4.

JAKOBSON, ROMAN

1966 Preface. In T. Todorov, Théorie de la littérature. Textes des formalistes russes. Paris, Seuil, pp. 9-13.

KÖNGÄS MARANDA, ELLI

1963 The concept of folklore. Midwest Folklore ,13 : 69-88.

1969 Review of Buchler and Selby (1968). Journal of American Folklore.

KÖNGÄS MARANDA, ELLI and PIERRE MARANDA

1962 Structural models in folklore. Midwest Folklore, 12 : 133-192.

KUUSI, MATTI

1959 Kansanperinteen metamorfoosi ("The metamorphosis of folklore.") Suomalainen Suomi.

[xxxii]

LANE, HARLAN L.

1967 A behavioral basis for the polarity principle in linguistics. Language, 43 : 494-511.

LANTERNARI, VITTORIO

1963 The religions of the oppressed. New York, Alfred A. Knopf.

LEACH, EDMUND

1954 Political systems of Highland Burma. Cambridge, Harvard University Press.

1964 Anthropological aspects of language : animal categories and verbal abuse. In E. H. Lenneberg, ed., New directions in the study of language. Cambridge, Massachusetts Institute of Technology Press, pp. 23-63.

LEACH, EDMUND R., ed.

1967 The structural study of myth and totemism. ASA Monographs 5. London, Tavistock.

LÉVI-STRAUSS, CLAUDE

1945 L'analyse structurale en linguistique et en anthropologie. Reprinted in Anthropologie structurale. Paris, Plon.

1950 Introduction à l'oeuvre de Marcel Mauss. In M. Mauss, Sociologie et anthropologie. Paris, Presses Universitaires de France, pp. ix-lii.

1955 L'analyse structurale des mythes. Reprinied in Anthropologie structurale, Paris, Plon.

1958a Anthropologie structurale. Paris, Plon. (English translation : Structural anthropology. New York, Basic Books, 1963.)

1958b La Geste d'Asdiwal. Annuaire 1958-1959, École pratique des Hautes Études (Section Sciences religieuses), Paris, pp. 3-43.

1958c Race and history. Paris, UNESCO.

1960a Leçon inaugurale, chaire d'anthropologie sociale, Collège de France.

1960b L'analyse morphologique des contes populaires russes. International Journal of Slavic Poetics and Linguistics 3 : 122-149. (Also under the title La Structure et la forme, réflexions sur un ouvrage de V. Propp. Cahiers d l'Institut de science économique appliquée, 99 : 3-36.)

1962a La Pensée sauvage. Paris, Plon. (English translation : The Savage Mind. London, Weidenfeld and Nicolson, 1966.)

1962b Le Totémisme aujourd'hui. Paris, Presses Universitaires de France. (English translation : Totemism, Boston, Beacon Press, 1962.)

1964 Mythologiques I : Le Cru et le cuit. Paris, Plon. (English translation : The Raw and the cooked, New York, Harper and Row, 1968.)

1967 Mythologiques II: Du Miel aux cendres. Paris, Plon.

1968 Mythologiques III: L'Origine des manières de table. Paris, Plon.

[xxxiii]

MARANDA, PIERRE

1963 Note sur l'élément de parenté. Anthropos, 58 : 810-828.

1967a Formal analysis and intra‑cultural Variation. Social Science Information, G : 7-36.

1967b Lecture mécanographique des mythes. Social Science Information, G : 213-222.

1970 Anthropological analytics : Lévi-Strauss' concept of social structure. In H. Nutini and 1. R. Buchler, eds., The anthropology of Claude Lévi-Strauss. New York, Appleton-Century Crofts, in press.

1969a Contributions méthodologiques de l'informatique aux sciences humaines. Actes du colloque de Grenoble, Centre d'études pour la traduction automatique, in press.

1969b Review of Buchler and Selby (1968). American Anthropologist, 71 : 521-523.

1968 Analyse quantitative et qualitative de mythes sur ordinateurs. In J. C. Gardin and R. Jaulin, eds., Calcul et formalisation dans les sciences de l'homme. Paris, Centre National de la Recherche Scientifique.

in press French kinship : structure and history. The Hague, Mouton & Co.

MARANDA, PIERRE and ELLI KÖNGÄS MARANDA

1970 Le crane et l'utérus ; deux théorèmes malaitains. In P. Maranda et J. Pouillon, eds., Échanges et communications, mélanges offerts à C. Lévi-Strauss pour son 60^ème anniversaire. Paris-The Hague, Mouton & Co.

MARTINET, ANDRE

1955 Économie des changements phonétiques. Bibliotheca Romanica, Series Prima, Manualia et Commentationes, vol. 10. Berne, Franke.

MORIN, VIOLETTE

1966 L'histoire drôle. Communications, 8 : 102-119.

NEWMAN, PHILIP L.

1964 Religious belief and ritual in a New Guinea society. American Anthropologist, 66 : 257-272.

PÊHEUX, M.

1967 Analyse de contenu et théorie du discours. Bulletin du Centre d'études et de recherches en psychotechnique, 16 : 211-230.

PIAGET, JEAN

1968 Le structuralisme. Paris, Presses Universitaires de France.

PIRKOVA-JAKOBSON, SVATAVA

1958 Introduction. In Propp, 1958, pp. v-viii.

PROPP, VLADIMIR

1958 Morphology of the folktale. Translated by L. Scott. Publication 10 of the Indiana University Research Center in Anthropology, Folklore, and Linguistics.

[xxxv]

1965 Les transformations des contes fantastiques. In Todorov, ed., 1965, pp. 234-262.

ROBINSON, M. and L. JOINER

1968 . An Experiment in the Structural Study of Myth. In E. R. Leach, ed., Dialectic in practical religion. Cambridge, Cambridge University Press.

SEBEOK, THOMAS A.

1960 Decoding a text : levels and aspects in a Cheremis sonnet. In T. A. Sebeok, ed., Style in language. New York, John Wiley & Sons, pp. 221-235.

TODOROV, TZVETAN

1965 Introduction. In Todorov, ed., 1965, pp. 15-30.

1966 Recherches sémantiques. Languages, 1 : 5-43. TODOROV, TZVETAN, ed.

1965 Théorie de la littérature. Paris, Seuil. TURNER, VICTOR

1962 Three symbols of passage in Ndembu circumcision ritual. In M. Gluckman, ed., Essays in the ritual of social relations. Ann Arbor, University of Michigan Press.

1964a The interpretation of symbols in Ndembu ritual. In M. Gluckman and E. Devons, eds., Closed systems and open minds. Edinburgh, Oliver and Boyd.

1964b Witchcraft and sorcery : taxonomy versus dynamics. Africa, 34 : 314-324.

TYNIANOV, J.

1965 De 1'évolution littéraire. In Todorov, ed., 1965, pp. 120-137.