Stuart Blackburn
Dartmouth College, Hanover


In recent years the study of culture has been increasingly directed toward the examination of systems that order peoples' lives. It is felt that these systems, variously termed belief systems, hypotheses, root metaphors, cognitive maps, world views, cognitive systems (Black 1973), are the assumptions from which culture emanates and that their apprehension would provide new explanatory keys. Concomitantly, there has been also an increased awareness of the need to use native terminology and concepts to gain access to those systems. However, despite this concern to avoid distortion of a system's content, there has been a serious misrepresentation on the more fundamental level of its conceptualization. The coherence of the reality (and often even encountered) and later represented as thought, with the consequences that those systems are assumed to be cognitive only.

Since the publication of the Griaule team's studies of Dogon though and the several expositions of Levi-Strauss, the complexity of the logic of non-western cultures is no longer in doubt. That such cognitive systems for ordering the chaotic mass of perceived reality exists cross-culturally and in diverse formations is established. What I want to suggest in this essay is that there is, in addition, another kind of system which is, primarily at least, not cognized but enacted. In other words, the organization of reality I structured not only in cognitive categories, but in behavioural patterns as well. Accordingly, this essay intends to demonstrate that performance, in a Tamil oral tradition, is interpretable as cultural paradigm.

In advancing the interpretation of oral performance as cultural paradigm, the present argument draws on both performance theory and model analysis. The formulation of a performance approach to oral literature has been developed over the past decade chiefly in the discipline of folklore (Abrahams 1968; Ben-Amos 1972) and secondarily in the allied field of socio-linguistics (Hymes 1974, 1975). Although there is consensus that the fundamental premise of the performance approach is event over text, or process over product, it has been employed with a diverse range of emphases. The most recent and comprehensive statement of performance analysis (Bauman 1977), by one of its original proponents, is written with the "language in context emphasis of the ethnography of speaking school. It thus signals a return to formal analysis and a de-emphasis of the original behavioural slant of performance theory" (e.g. Lomax 1968).

Somewhat independent of this development of folklore, but linked to it by a shared debt to K. Burke's dramatistic perspective, a similar dynamic interpretation of cultural behaviour emerged in anthropology (Geertz 1973; Peacock 1969; Turner, 1974; and earlier, Bateson 1936). One branch of symbolic anthropology, this approach is not concerned with the specific issue of performance in oral tradition, but with the larger question of how behavioural sequencing may be, I and of itself, a symbolic domain. Particularly influential to the present argument is the position developed by Geertz (1973 : 443-53) that certain cultural behaviour, e.g. ritual or festival, may be understood as the way people represent reality to themselves. In this sense, performance is a symbol paradigm, a medium for a culture's interpretation of itself.

The conception of performance forwarded in this essay differs from that proposed by Hymes (1975 : 18) which has general currency in the field: conduct for which the doer assumes a responsibility to be evaluated. Although the notion of responsibility is useful, in living oral traditions one need not be so troubled, as was Hymes, about when performance begins and when it ends. Allowing local definitions to identify performance , the present conception is more restricted and includes only those recurrent behavioural segments, ritual and secular, involving two or more persons which are locally regarded as culturally significant. Its function, and it is thus more akin to Turner's "social dramas" (1974), Singer's "cultural performances" (1955 : 27), and Schieffelin's "cultural scenarios" (1976 : 3) than to Bauman's "interpretive frame" (1977 : 11). It is, furthermore diametrically opposed to the idea of performance in Chomskyian linguistics as an imperfect realization of an ontologically prior meaning. This is inverted to assert that performance itself is the vehicle of meaning; the paradigm lies not in competence, but in performance.

In presenting performance as paradigm, "paradigm" is also conceptualized differently than it is in its standard usage as an abstract set of rules, as a scientific world view directing cognition (Kuhn 1962 : 43-51), or as an exemplar for comparison (Black 1962 : 156-69). The present intention is to demonstrate that in performance the instrumentality and the intelligibility of the paradigm are brought into closer alignment (cf. Schieffelin 1976 : 2). Indeed, the position toward which this essay moves is that the paradigm does not exist outside its enactment in performance. Perhaps that is one reason why performance occurs at all without it the paradigm would not even be operative. Austin (1975 : 6-7, passim) has identified as "performative" those words which by their very utterance, perform an action (e.g. "I beg you"). Correspondingly, a performance paradigm may be understood as one which enacts meaning.

One more point of clarification concerns the "model of/model for" distinction which is frequently invoked in the discussion of cultural models. Although that distinction is useful to indicate a model's complexity, it should not obscure the fact that any model or paradigm is inevitably both a description of social reality (on some level) and a sanction for it (cf. Ortner 1975 : 135). Moreover, these two aspects of a model are not passive as in a stencil extracted from behavioral patterns and for which subsequent patterns are cut out. Instead, we must learn to see that the performance paradigm, like speech, organizes the world while it describes it.

Turning more directly to the study of Indian culture, model analysis in this field carries the same assumption of a cognitive medium that it does in the study of culture generally. In reviewing the field, Das faults an over-emphasis on observable behavior and urges attention to the "way in which people impose order upon the world" (1977 : 4). She then demonstrates how these cognitive systems are extractable from Sanskrit texts. Marriott has presented a sophisticated explication of a transactional model generated from a behavioral base, but it, too, is ultimately concerned with "cognitions of reality" (1976 : 113). Implicit in these studies is the assumption that people organize reality only, or primarily, by means of cognitive systems. While crucial social issues are often ordered cognitively, they are also ,I argue, made sense of in behavioural systems.

An almost extraneous statement from Singer (1959 : 145), is especially germane: "Whenever Madrasi Bramans (and non-Brahmans, too, for that matter) wished to exhibit to me some feature of Hinduism, they always referred to, or invited me to see, a particular rite or ceremony in the life cycle, in a temple festival, or in the general sphere of religious and cultural performances as links in the levels of civilization and not "as performances", his remarks indicate that they do serve to organize meaning for people at least in the city of Madras. While the cultural performance concept was thus first applied in the study of India, paradoxically performance theory has been since utilized in that field only infrequently (Blackburn 1979; Wadley 1978).

In proposing the interpretation of performance as paradigm, this essay seeks to make a contribution not only to the study of Indian folklore, but to the study of oral tradition and expressive culture in general. Both the performance approach and the cultural model approach have been utilized in the analysis of oral tradition, but their conjunction has been rare (Schieffelin 1976; Geertz 1973 : 412-53). By combining the two techniques, performance theory is extended to the discernment of cultural paradigms and model analysis is brought into the interpretation of oral performance. The goal of this theoretical stretching is to demonstrate that meaning is not only expressed in and perceived in behaviour, a proposition most would accept, but also that it is ordered by behaviour.


The bow song, or vil p¡u tradition, is found only in the two southernmost of Tamil Nadu, South India. The material presented here refers to Nancil Nadu, a traditional region within the area, although most statements apply equally to the vil p¡u tradition at large1. The research upon which this paper is based was conducted primarily in Nancil Nadu during 1977-1979 and was supported by the Social Science Research Council, the Fulbright-Hays Foundation and the American Institute of Indian Studies. vil p¡u is found also in the neighboring linguistic state of Kerala near its border with Nancil Nadu. References to "Nancil" as a region exist in classical Tamil poetry (e.g. Puram 137, 138, 140).1. Nancil Nadu has a 2,000 year old history, covers nearly 4.00 square miles and stretches to the southern tip of the sub-continent at Kanya Kumari (Cape Comorin). The tradition takes its name from the enormous how, or vil, six to twelve feet long, that is set up in front of the performing group and upon which the lead singer plays. The group consists of a minimum of five persons, and as many as eight or ten, each with a percussion instrument. Until recently, few women have restricting their public behavior, particularly self-expressive behavior. During the past ten to fifteen years, women have become increasingly popular as lead singers (and always in pairs because a single women on display is less acceptable), but they still never perform other roles in the troupes.

The performers are semi-professional, earning only a part of their income from singing and only during a few months of the year. Most are functionally literate, and a few male lead singers, though not educated beyond primary school, are well-versed in the literature of vil p¡u. Generally, performers are not distinguishable socio-economically from their audience of peasant villagers and townfolk. They come from the middle-level castes in the areas, the same castes among which the tradition is currently active2. In Nancil Nadu, vil p¡u is strongest among the Naar (agricultural, traditionally toddy-tapping and largest caste in the region) and largest caste in the region) and the Pi½½ai (or Ve½½ala, landowning caste); it is also found among the K°nar(agricultural, traditionally shephered), Caliyar (weaver), Ëcari (artisan), Campavar and Pa½½ar (untouchable) and some Malayali castes (Kurup, Ilavar).2. vil p¡u is not patronized by Brahmans, or by some segments of high castes that intimate them, for reasons of social prestige, nor is it patronized by the lowest and most oppressed groups for reasons of financial scarcity.

Vil p¡u songs narrate the history of the gods and goddesses installed in a particular temple and worshipped during its festival. Most of the songs are currently available in palm-leaf manuscript, originals or copies of semi-learned poetry written 150-350 years ago. Some song texts have been printed and sold locally in pamphlet form. Texts in the tradition vary in length from 500 to 5000 lines or more (one text of a local Ramaya¸a is 13,000 lines long), and performance varies in duration from 30 minutes to six hours or more. The singing summons the deity propitiated during the festival and thus has a central ritualistic role, integrally connected with the spirit possession that eventuates when the deity appears.

The only performing context for Vil p¡u is the koai (lit. "offering") festival which is distinctive to the area covered by the tradition. The koai occurs once a year (usually between February and May) at each of the several hundred local temples in the area. These temples are "locally" controlled by family, lineage, or clan groups and located predominantly in villages, but also in the mono-caste quarters in the few towns in the region. The koai is the only function in thetemple to which every constituent unit of the temple community, usually the controlling kin group, contributes materially. Ci¤appu is a variant koai, financed by a single family in fulfillment of a vow, in which Vil p¡u performances are also included.

Documentation of Vil p¡u, excepting a few fragmentary early missionary reports, is absent outside the Tamil academic world and only very sparse within it3. The only article I English is Hameed (1956); see also Singer (1959 : 152). See also Reiniche 1979 for some details on vil p¡u in Tinnevelly District. Recently, also, a full monograph on vil p¡u has been published in Tamil (Gomathinayakam 1979).3. However, for the several million people living within its geographic spread, Vil p¡u performance in the koai festival is the most important public event in their lives. A marriage or pilgrimage to an all-India famous temple site may be more personally consequential and expensive, but those events are idiosyncratic and private. By contrast, in the Koai festival everyone in the temple community contributes financially and theoretically shares in the benefits generated. Either as passive audience, or as active petitioners of the deities worshipped, everyone (with exceptions of course) participates. Roles are differentiated: performers, auditors, spirit possession mediums, and temple officials; but all are face-to-face with one another in the same event. Significantly, that event is remarkably uniform throughout the vil p¡u region and is frequently repeated. Although each temple conducts one koai per year, plus additional ci¤appu, a medium-sized village will contain as many as 10 or 15 temples, and thus hold that number of koai. Moreover, if one wishes to visit koai in neighbouring villages, as many local people do, the number of festivals one may experience in one year is considerable.

In order to demonstrate how the performance of this event operates as a cultural paradigm, the essay is divided into essentially two parts. The first part examines the key taxonomic distinctions that structure the tradition outside performance, i.e., as a cognitive system. The second part describes how those distinctions are re-structured in performance, i.e. as a behavioral system. A concluding section summarizes the argument and draws out its broader implications.


In the conceptualization of vil p¡u as a cognitive system, i.e., in the non-performance mode, there are several planes of classification. Three planes, the local pantheon worshipped in the tradition, the temples I which they are enshrined, and the koai festivals in their honour, form one set that is concerned with the spatial and temporal dimensions of the event, or context. A fourth taxonomic plane organizes the oral contents of the event, the sung stories themselves.

Beginning with the first set, pantheon, temple and festival, the central classificatory criterion is sexual gender, bifurcating each plan into male and female divisions. The local folk pantheon in Nancil Nadu includes over a hundred deities, many of whom are deceased ancestors, each with a separate, though sometimes related, history. This disparate array of figures is catalogued in local speech by a binary set of female goddesses and male gods of amman and madan. Amman is cognate with a group of words in the Dravidian languages meaning "mother" and is used throughout the Tamil areas to mean "goddess". Madan on the other hand, is a peculiar word found only in the vil p¡u region where its usage derives from the name of the principal male deity, Cualai Madan.

Most deities whose cult has developed beyond a small kin group (and many that have not) bear either the explicit title amman or madan in their name, e.g. Muttar Amman, Ponni¤attal Amman, Cualai Madan, Karai Madan. Other deities not nominally designated as amman or madan were, nevertheless, consistently identified by those labels in the solicited and unsolicited speech of local people. When the identity of such a deity was questioned, the answer would be, "Oh, he's one of those madans", or "You know, she's an amman". The only exceptions to the amman/madan bifurcation are those figures assimilated into the vil p¡u tradition from pan-Indian mythology. They include Perumal Cami ( a Tamil name for Visnu), Ka½a Cami (a minister of Yama, god of death), Vairava Natan (local pronunciation of "Bairava", a form of Siva), Kaai Ó¤um Perumal ("he who ascends the funeral pyre", or Harischandra), and Sasta (Harischandra, Aiynar). These deities are all male and, except for Sasta, are only satellites, especially to an amman, and never central figures in a temple.4. Another word camati, denotes a resting site for the soul of a locally esteemed person. See Beck (1972 : 2068) for similar terminology in another Tamil area.

Turning to the temples in the vil p¡u tradition, they all belong to the terminological set koyil (lit. "King's house") which designates all worship sites of supra-human powers in the Tamil area, from a large towered complex to a small stone.

Vil p¡u is not performed I the largest temples, but among those in which it is performed there is still a tremendous variety of physical form. The minimal requirement for a vil p¡u temple is a place in which the deity resides and/or to which it is summoned during the festival5. See Shulman (1980, Chap.2) for a discussion of the locale in Tamil temple myths as a point of differentiation from their Sanskrit models.5. This place may be a shaped icon, a painted stone, a tree, an ant-hill, a painting or smeared spot on a wall, or a spot on the ground marked off by white powder during the festival. The temple may be structural, i.e., the deity's place housed within a permanent building (usually a single room), or non-structural, i.e., the deity's locus exposed to the open air. During a festival, sometimes the addition of sheds constructed of banana tree stalks and palmyra thatch transforms the non-structural temple into a structural one.

The amman/madan distinction is the main classificatory criterion for this diversity of temples, reducing it, like the disparate folk pantheon, to a dualistic system. In the speech of local people, again solicited and unsolicited, a particular temple was consistently identified as either "amman/koyil"), indicating that the plane of deity typology has been added to the frame of reference.

In addition to the gender distinction indicated by their labels, the amman and madan temples, are also partially differentiated by physical and iconographical features. The prototypical amman temple is structural, with a anthropomorphic icon, while the madan temple is characteristically not structural andis without an anthropolorphic icon. The amman icon, if in status form, is known as cilia and is recognizably a variant of folk iconography found throughout India. By contrast, the madan icon, which resembles an obelisk, is known by a separate term (pitam) and is unique to the vil p¡u region. It is made of clay and brick, covered wit lime paste and stands 3-8 feet tall on a small rectangular platform of the same construction.

Finally, the amman/madan contrast informs classification on a third plane in the vil p¡u tradition, the festival. vil p¡u is performed only in a koai, a class of festival distinct from others, vila and usava, in the area. The term koai is a peculiarity of the tradition, although the general elements of the festival are cognate with those of festivals in other parts of South India. The koai is divided into two sub-categories, amman koai and madan koai, depending on the temple type in which it occurs. Significantly, the separation between these sub-categories is not only terminological, but temporal: amman koai may occur only on a Tuesday and madan koai only on Friday.

To recapitulate, a distinction based on sexual gender, amman/madan, is the primary classificatory axis on the planes of patheon, temple, and festival in the vil p¡u tradition. The amman/madan contrast is clearly a binary set as evidenced by the fact that the terms are commonly used in opposition to one another. Thus, for example, "That's not an amman temple, but a madan temple", and the converse are frequent statements. It should be pointed out that although the contrasts on each plane were each recognized and articulated locally, they were never inter-related in the manner presented here. Sometimes in speech solicited situations, those terms were used to reference a specific item, often in contradistinction to its binary partner. The most taxonomic system and that given here is that in the former the amman and madan categories were never described as generically female and male. When that interpretation was made by the researcher in conversation with local people, it was, though never denied, accepted only reluctantly.

The amman/madan distinction, however, is not the only classificatory criterion for the vil p¡u pantheon. These categories of gender conjoin otherwise disparate and separate otherwise similar deities. Thus, cutting beneath their loose grouping is another binary set that organizes deities according to the nature of their birth and their death. In local terminology, every deity in the pantheon is either teyva vamsam ("of divine descent") or veupatta vatai ("a cut-up malevolent spirit"). These terms indicate the definitional role of birth (divine descent) and death (cut-up) respectively for the pantheon. The first distinguishes deities by their birth, through the agency of Siva in his celestial abode or its heroic variant, suicide. Conversely, the first group (teyva vamsam) is also defined by their death - they do no die; and the second (veupatta vatai) are defined also by their birth, which is earthly, human and painful. Each deity has thus either a divine, painless birth in Kailas and does not die (hereafter type A), or has a painful, human birth on earth and suffers a tragic death (hereafter type B)6. Although there are some broad areas of agreement between this classification of a folk pantheon and others reported in India (e.g. Harper 1964 : 183 ff), this particular differentiation by birth and death appears unique. See also Aiyappan (1976) for the human-divine continuum in Kerala.6.

The classificatory power on the birth episode for type A deities is illustrated by the variants of the story of the two most pervasive deities in Nancil Nadu Muttar Amman and Cualai Madan story, performers claim that there are seven variants distinguished by as many different kinds of birth, although they are able to identify only two. Even the terms by which these two variants are known employ the word "birth": viz. "birth in the lamp" and "birth (in the sacrificial fire of) Takka Raja". For the Muttar Amman story, the major variants are actually stories of nominally distinct goddesses who collectively comprise a composite amman. These stories are then known by the name of the specific goddess they celebrate, but it is, in fact, the birth episode which differentiates them. Thus, Muttar Amman is born of the sweat of Parvati or in an egg; Ka½I Amman is born from one of siva's eyes or formed from three women; and Uccinimaka½i Amman is born from Siva's third eye or from his head. Beyond this variation in birth episode, these nominally separated stories follow a similar, sometimes identical, narrative sequence.

Under the rubric of type B (earthly birth and tragic death) stories, there is not a similar systematic sub-categorization based on the nature of death. However, a single such sub-category, cumai ta´ki or "load bearer", does exist. Cumai ta´ki refers to those women who die in pregnancy or delivery and are later deified and housed in stone structures, of the same name, placed by the roadside for bearing the loads of travelers.

Again, in local speech the terms for type A (teyva vamsam) and type B (veupatta vatai) showed consistent usage as generic categories in the identification of deities. Like amman/ madan labels, they were used both independently and as a contrast set. However, their definitional criteria of birth and death, like sexual gender in amman/ madan, were never made explicit; only the terms themselves were spoken and understood. The relationship between these two taxonomic distinctions I the pantheon, amman/madan and birth(A)/death(B) is unclear, and they appear to operate independently. The amman/madan contrast is more frequently verbalized, but the A/B contrast was not less known or understood when mentioned. The later represents a slightly higher level of abstraction and functions on the fourth and last classificatory plane, on which amman/ madan is absent, of the vil p¡u story corpus.

The entire domain of story in the local area is divided into two categories: keka katai and cami katai, or "story to be heard" and "god's story". These labels are context-oriented since the same story potentially belongs to either category. keka katai designates stories told outside the temple festival context. They are not intended as worship, no ritual is involved, no possession ensues and the raconteur is not paid. For each of these statements, the reverse is true for the cami katai. Popular keka katai stories are never performed in veupatta vatai, e.g. the pan-Tamil stories of Cirutondar, Kovalan and Kannaki, Nalla Tankal, and some that are occasionally performed in the tradition, e.g. stories from the Ramayana and the Mahabharata.

All stories performed in vil p¡u belong to the cami katai category since, as a form of worship in the festival, they have at least a minimal ritual function. Within the vil p¡u tradition, stories are classified by the A/B distinction already shown to define deity type (by the birth and death episodes in their histories). That point is extended here to detail how the stories themselves are classified by the A(birth) and B(death) categories. It is most significant that those categories are found within the stories as explicit type-markers. Starting the gross plot structure of each category, it is possible to learn how they, together, form a symmetrical cognitive system.

First, at the end of the fixed-phrase invocatory verses that introduce the vil p¡u story (in palm-leaf), printed pamphlet, or oral performance), the name of that story is mentioned for the first and only time. This name marker includes the name of the deity and either of two modifiers: piranta ("born") or iranta/i¤antappaa ("died"). The invocation thus concludes with either of two lines: "We sing the story of X's birth" or "We sing the story of X's death.

Following this invocation, the first segment to be mentioned I the narrative sequence of each story type is the initial setting. In this section, which initiates the narrative proper (or varalaru), again somewhat fixed verbal formulas serve as clear markers of which story type is to follow. Type A begins with the fixed phrase "in golden/silver Kailas Siva ruled", and the line in type B stories is "in the land of all lands". Stock descriptions of each setting then follow: in type A the power and beauty of Siva's court, and in type B the prosperity of the land. In type B, this general description moves quickly to specific and known details, identifying the particular region, a named village, a caste and usually the personal names of persons in one family. The date of the action, however, is never mentioned.

The birth of the deity in this initial setting further differentiates story sequences A and B. although in both childlessness is the lack that is liquidated by a birth, the nature of that birth, as discussed above, is quite different. Type A involves a painless, miraculous birth and type B is painful, human birth. Furthermore, the events that lead up to the birth in each type distinguish them. In type A, the painless birth is preceded by any number of idiosyncratic events and the deity appears rather abruptly. In type B, by contrast, the painful birth is preceded by a standard series of events, pilgrimage to temples, penance, a month-by-month description of the pregnancy, the summoning and arrival of a mid-wife.

After the birth event, the story types follow different paths to a similar key episode, the central conflict. In type A stories, the deity completes certain tasks set by Siva and then requests, and eventually receives, boons from him. Fortified with these special powers, designed to control humans on earth, the deity descends from Kailas to the earth, an episode known in the tradition as "the descent from Kailas". Once on earth, the action moves toward one central conflict which pits the deity against a powerful enemy whom he subdues. Characteristically, this conflict is male/female, either overtly sexual, e.g. rape, or covertly sexual, e.g. the goddess slaying the demon.

In type B stories, the birth of the deity is followed by more patterned minor episodes, taking him through education, various martial arts training, and optionally, marriage. However, in the case of goddesses in this category, this prefacing material is severely curtailed or omitted altogether. After these fixed sequences, the narrative progresses until a particular problem develops which eventuates in the deity's death, by murder or by suicide. For a male deity, this problem is inevitably an inter- (or intra-) caste conflict or open war. Very often this fatal conflict is created, or is believed to have been created, by a male/female sexual antagonism, e.g. a rejected demand for a bride. In other case, it is not the conflict, but more directly the death of the hero already embroiled in it what is believed to be caused by a woman, e.g. by her underlying cause of her death. She is either the innocent victim of masculine aggression, abducted and killed, or its powerful opponent in a war in which she is subdued and then commits suicide.

In type B stories, then, the incident which precipitates the central conflict and/or leads to the deity's death is either war or sexual confrontation or their conflation. This same homology between war and sexual conflict was briefly noted in type A stories. Although it is there less pronounced because the deity does not die, nevertheless sex and war are the interchangeable instruments by means of which a female both conquers and is conquered.

Following this central conflict in Type A stories, the deity creates havoc and coerces various groups of people to worship him or her. In his way, the cult proliferates and the story ends as the deity comes to reside in his or her most famous temple (in performance this is the particular temple in which the story is being sung). In type B stories following death, the deity goes to Siva in Kailas, asks for and receives boons, returns to earth, and by means of the boons established his or her worship. Specific boons are received to wreck revenge on those responsible for the deity's death. The guilty are killed and relatives or neighbours, to avoid a similar fate, erect a temple to the fearsome spirit and institute a koai festival in its honour.

The outline of the episodic sequence of these two stories type is summarized in a linear format in figure 1 below:

Figure 1

Diagramming the directional movement of the same episodic sequences, the following figure results :

Figure 2

With the aid of Figure 2 it is possible to summarize the vil p¡u narrative corpus with reference to its taxonomic categories. In one category, deities are born on earth and after death go to Kailas, receive boons and return to earth for worship. The composite figure shows that after the death and arrival of the deity in Kailas, type B stories follow the same sequence as do type A stories. There is thus a fit between the two story sequences that forms a cyclical movement between the divine and the earthly realms, between deities and humans. As a consequence, if the past-death segment in type B (although it is less important) were to gain prominence over the pre-death segment, the story could be historically transformed into a type A story. Also, a transition from type B to type A, within this cyclical fit could be effected by what I termed elsewhere (Blackburn 1978) the "substituted birth", i.e., an historically subsequent ascription of a prior divine birth. As a result of this prior divine birth in Kailas, the human birth and life of the deity, the tragic aspects of the story, become ontologically secondary and eventually recede into a background.

Viewed synchronically, for present purposes, this cyclical fit illustrates how the two story types form a symmetrical taxonomy for the vil p¡u tradition. That taxonomy, which describes a flow movement between cosmo-geographical regions (Kailas and earth) and their inhabitants (deities and humans), is pegged to crucial narrative themes. First, categories within that taxonomy are identified by "birth" and "death" verbal markers and as narratives themes: birth dominates type A and death dominates type B. Secondly, in both story types a homology was noted between armed conflict and sexual conflict, is central in both story types, but in A it is less ominous, more auspicious and often connected to a birth, while in B it is the cause of tragedy and death7. The sexual conflict in vil p¡u narratives is not between ascetism and eroticism, as O'Flaherty (1973) has convincingly demonstrated for Hindu mythology, but between the bivalence of its erotic mode which causes both birth and death.7.

The consistent and broad contrast between story types A and B, to which all this material points, is underlined by the difference in spatial setting between them. Type A is set in Kailas, a mythical place, and type B transpires in this world, in known places, involving local groups of people and events. Correlating these differences, in type A the setting is distant, problems are painless, and the general tenor is fictive. By contrast, in type B the setting is proximate, problems are painful, and the general tenor is "real".

To summarize Part-I of this essay, the central points are two: (1) the structure of the non-performance taxonomy of the vil p¡u tradition is binary, and (2) that structure organizes the crucial problems, birth, death and sexual conflict, into a cognitive symmetry between type A (distant and fictive) and type B (proximate and real).


The translation from the first part of this essay to the second part marks a shift in perspective from vil p¡u as performance event. A cognitive model (the organization of the tradition by cognitive categories) is complimented by a behavioral model (the organization of performance by behavioral patterns). Significantly, this shift of perspective corresponds, in fact, to what happens in a vil p¡u performance itself. In part I two types of narrative (A and B) were identified on the level of content. On a higher level of abstraction, however, any story of either content type exists potentially in two models defined by the behavioral criterion of whether or not they are in performance. Succinctly, when a story is not being performed, it is referred to as katai; and when it (the same story) is being performed, it is known as p¡u8. This paper is analyzed in greater detail in my 1979 paper.8. Thus, in Part I the classification presented pertained to the katai mode, the stories outside performance. In this second part, we adopt the indigenous viewpoint and consider the tradition in its p¡u mode, as performance event.

First, concerning the performance context, it was stated earlier that the koai festival occurred only on Tuesdays for amman (female) and only on Fridays for madan (male) deities. It is now necessary to point out that those days refer only to the central day of the festival, when the temple's chief deity is worshipped. The traditional duration of the koai is three days, one before and one after the central day. Thus amman koai begins on Monday and ends on Wednesday, and madan koai begins o Thursday and finishes on Saturday9. Tuesdays and Fridays are traditional days of worship in S. India (Whitehead 1976 : 69, 72, 75, passim) and probably in parts of N. India (see, for example, Mahapatra 1972 : 141, 146).9.

This three day program provides the framework for the sequencing of the major vil p¡u performances in the festival. The center day, on which the main deity is worshipped, is marked by the performance of his or her song. The first and third days are likewise by one major performance. The festival is thus structured syntagmatically as a three slot motifemic sequence in which each major performance constitutes a slot.

The isolation of the three "major" performances from among the total number of performances (as many as ten or twelve) during the festival is facilitated by their very sequential position. The center performance is obvious because it sings the story of the principal deity on the center day. The other two major performances are similarly obvious because one opens the festival and the other closes it.

In addition to this positioning, other criteria serve to distinguish major from minor performances. First, major performances are conspicuous by their extended duration, from three to eight hours, while minor performances last only ½ hour to one hour. Secondly, some performers and other knowledgeable persons, when asked what stories would be performed at a particular festival, would list three, the same three that then fell into the slots identified here, and in the same order.

Furthermore, this tripartite infra-structure of the koai festival, marked by its major performances, is constant even when the festival is less than or more than three days. When low finances permit only a one day koai, the three major performances surface more clearly because sometimes they are the only performances. Similarly, when resources allow an expansion of the koai from three to nine days, the tripartite framework of major performances is still identifiable by their extended duration and the deity they celebrate. In these expanded sequences, amman koai commences on Monday and continues through the following Tuesday. The first slot in the tripartite sequences occurs on the initial Monday, the second a week later on the second Tuesday, and the third follows on the same day. In the less common instance of an expanded madan koai, the same rules apply. The nine days begin on Thursday with the first major performance and end on the second Friday with the center and last performances. Within the expansion and contraction of the koai, in both amman and madan variants, the motifemic sequence remains constant. This is presented in figure 3.

Figure 3

(amman)                       Monday            Tuesday            Wednesday                 Tuesday
(mādan)                        Thursday         Friday             Saturday                      Friday
3 day koai                        x----------------x--------------------x
1 day koai                                           x-x-x
9 day koai                       x-------------------------------------------------------------xx

This linear representation depicts the horizontal ordering of the vil p¡u performance event, but is not an adequate description of the entire event. To its syntagmatic dimension another must be added that records the affective aspect of the behavioural sequence. That dimension is the dimension of depth10. See Geertz (1973 : 412-53) for an extended analysis of depth in the Balinese cockfight.10. "Depth" is used here to refer to the relative intensity of the event, the festival or its constituent units, that eventuates in spirit possession . In as much as spirit possession is believed to be caused by the presence of a deity, induced by worship and song, the term "ritual depth" will be used.

Ritual depth culminates, in its "deepest" instances, in a possession dance, or cami aam (lit. "god-dance"). The depth of the possession dance is manifest in that, for the audience, it is the most meaningful segment of the festival. It receives the most attention, the largest crowd. During festival days, the question is always: "Which deity is coming next?" This means not which song, but which deity's dance is next. It indicates a focus not on the oral behavior (song) summoning the deity, but on the kinetic response stimulated by his presence.

Present concern is however not with the possession per se, but with the fact that it indicates a culmination of ritual depth in the performance event. Suffice it to mention that the dance comes to that person (usually a man) who is designated as the dancer (cami ai, cami ko¸ai, K°marattai) for the particular deity and, not infrequently, to any number of others in the audience. It may not infrequently, to any number of others in the audience. It may take the form of convulsive jerking, more active leaping and rolling, or a more controlled and regular pattern of movement that approximates the English word "dance". The dance may also involve props marking the identity of the possessing deity, particularly weapons, knives, swords and poles for males.

Incorporating this dimension of depth into the representation of the festival sequence, the following diagram results:

Figure - 4

As figure 4 illustrates, ritual depth steadily increases as the sequence approaches the central performance at which point a possession dance ensues. Possession occurs also at other points in the sequence, but that associated with the central performance is the most intense. After this culmination, ritual depth gradually recedes and, by the end of the festival, returns to its pre-festival level.

Before detailing how this basic performance structure (figure 4) is elaborated, its tripartite and ritual nature immediately invite comparison with Van Gennep's (1960 [1908]) classical paradigm, of which it is a variant. Moreover, an explanation of the relationship of the three major performances to the ritual depth that runs vertically and sequentially through them is suggested by his theory of liminality.

Through its association with spirit possession, ritual depth entails contact with suprahuman powers and therefore the element of danger. Those powers are dangerous because they are uncontrollable and because contact with them involves a transition from one realm to another. Van Gennep has shown in his vast array of examples that such a transitional or liminal state in ritual is bounded on either side by separation from social reality and then re-incorporation into it. In other words, the transition from one world to another itself requires antecedent and subsequent transitions. These are boundaries that create fixed thresholds and reduce the ambiguity inherent in liminality. The applicability of Van Gennep's schema to the vil p¡u sequence should now be obvious. The three major performances that define the festival framework are organized so that the first and last provide boundaries for the centre one in which ritual depth and danger is greatest11. Beck (1969) has brought out the bounding function, by color, of ritual "heat" in one part of rural Tamil Nadu.11.

Finally, to restate Van Gennep, it should be noted that the boundaries in ritual operate in both directions from its interface with social reality12. Van Gennep called this the "pivoting of the sacred" (1960 : 12-3).12. not only do the boundaries protect normal reality from dangerous contact with supra-human powers, but by encashing the liminal state they also protect it from the incursions of normal reality.

We are now in a position to examine how the binary typology of narratives worked out in Part I fits into this tripartite performance sequence. The first slot in the sequence, whether amman or madan koai, is always filled by the Sasta story, a pre-eminently type A story. Although it commences in Kailas, like all type A stories, unlike them it does not move to the earthly realm of nau13. In some performances, that segment, the descent from Kailas and life on earth, is sung perfunctorily to complete the narrative, but has no role in the development of ritual depth.13. Rather the story remains in the mythic world, alternating between the home of the gods and that of the demons, emphasizing the fictiveness and psychogeographical distant that characterize type A stories. The other prototypical feature of type A stories, the dominance of the birth episode, is also prominent in the Sasta story. In fact, this story, which in other oral traditions and in printed forms recounts the entire history of the god, in vil p¡u performance ends abruptly with his birth.

Furthermore, the birth event is also the point at which ritual depth is greatest in the performance. It is also when possession occurs; i.e., the appearance of the deity in oral performance is enacted in dance. During the oral performance, preparations for the worship of the soon-to-arrive god are made nearby, in or around the temple. The performing group must synchronize with these preparations bringing their story to the birth of Sasta just as they are finished and possession is to ensue. This possession dance, in the first performance slot, is relatively short, is not intense and is usually restricted to the god's known medium. Also, there is no random occurrence of possession in the audience.

The next major performance, in the center slot, and the possession it engenders, is the deepest point of the push for ritual depth in the entire festival program. The story performed in this slot, because it is dedicated to whoever the principal deity in the temple might be, may be either type A or type B. Nevertheless, if the first slot emphasizes aspects characteristic of type A stories, this slot emphasizes those associated with type B. In the case of a type A story in this slot, mythic action in Kailas is overshadowed by the emphasis given to the episodes that deal with life in nau, or this world.

Performances of type A amman stories in this slot, for example, bring her quickly to earth and elaborate her attempts to coerce worship from various caste groups in the local region. A more specific indication of this "real world" emphasis is that among several possession dances that occur during this central performance, the most intense occurs when the goddess takes up residence in the particular temple in which the festival is being held. Her birth in Kailas, the definitive episode for type A narratives which is also emphasized in the first performance slot, does bring possession in this second slot, but in a mild manner. The other point at which intense possession regularly occurs is the destruction of the city of heretics (Tripuram), a death-oriented episode that further illustrates the type B accent of this slot.

Similarly, the performance of madan type A stories in the center slot focuses on the action after the god descends from Kailas to the earth. For example, the divine birth of Cualai Madan, like amman, does evoke possession, but the most intense dances result when he confronts his riva, a spirit conjurer, rapes that persons, daughter and forces him to sacrifice her to him.

This tendency toward type B aspects of death and tragedy in the real world would suggest that the greatest ritual depth is reached when a type B story is, in fact, performed in the center slot. And, indeed, this is the case. As noted earlier, these stories are local historical legends that recount the life history of a person who died a tragic and/or valorous death and is later apotheosized. Although type B stories involve a cyclical movement between Kailas and nau, the events on earth receive the greatest attention. The portions subsequent to death, that take the soul to Siva and then back to earth as a deity, are not elaborated and are sometimes omitted altogether.

Moreover, in these performances possession breaks out at several points, but predictably it is most violent at the point of the hero's death. This possession dance is the most powerful in the entire vil p¡u tradition, lasting often for several hours and involving forms of self-mutilation, e.g. lashing oneself with a chain or rope and striking oneself with a sword. This possession dance is not confined to the regular medium; it occurs also among persons in the audience (a phenomenon sometimes, but not always seen in other situations).

After this climax, the ritual depth of the festival steadily recedes as it approaches the third and final major performance. At the ebb flow of ritual depth, performance in this slot is not necessarily considered a form of worship and need not sing of a deity enshrined in the festival temple. Consequently, there is no possession dance, and sometimes even a portable cinema (in wealthy villages) will substitute for a vil p¡u performance. A regular exception to this pattern occurs when the koai is conducted at a divine amman temple in which Cualai Madan is enshrined. In this instance, that god's song is sung in the third performance slot as worship, but a mild possession arouses little excitement. Other stories which fill this slot are chosen for their entertainment value. The most popularly sung stories include two from the Ramayana and a local narrative of a Brahmin-untouchable love marriage (Muttuppattan).

The performance of these stories, Cutalai Madan, the Ramayana and Muttuppattan, in the third slot renders them in a fashion that highlights type A aspects. In the cualai Madan story, in contrast to the earthly adventures and death episodes underscored in the center performance slot, events in Kailas, his birth and mischievous escapades, are ventilated I this third slot. Moreover, like performances I first performance slot, the possession dance, albeit mild, occurs at his birth.

The tendency of third slot to emphasize type A aspects is even more apparent in the case of the Muttuppattan story. This story is a characteristically type B madan tragedy of love and is tremendously popular in the vil p¡u area. When this story is sung in the center slot (at temples in which Muttuppattan is enshrined), the performance builds to the tragic death, at which point the intense dance erupts. However, performed not for worship but for entertainment, when it is sung in the third slot it is cut conspicuously short at the marriage of the hero and the sorrowful segments are not sung.

Although stories from the Ramayana are performed only in the third slot and thus no basis for comparison exists within the vil p¡u tradition, considering which segments of that epic are sung, a similar selection process is observable. The only two segments that are sung describe Rama's birth and his marriage to Sita; their suffering and separation are rarely, if ever, performed in vil p¡u.

This ability of performance slot to alter story content was first made clear while recording a festival during field work. In the third performance slot, the singers decided to sing the story of Muttuppattan who, although not worshipped at that particular temple, was worshipped in several temples in the vicinity. The audience was therefore used to hearing the entire story performed in the center slot, in which it would climax at his death and lamentation of his widows. When the performers began to end their story at the marriage of the hero, as is appropriate for the third slot, several persons in the crowd objected and asked them to continue. The singers, however, refused explaining that only if possession ensued would it be proper for them to sing to the tragic end of the story. Some countered that this logic was reversed and that possession would result only if they sang that tragic part others pointed out that it was necessary to close the festival with a "good" event, i.e. the marriage. Eventually this reasoning was accepted and the performance and the festival ended.

In addition to these examples of how the third performance slot transforms story content, two examples adduced above from the other slots are pertinent. In the first slot, it was pointed out, the Sasta story is curtailed at the god's birth. I the second slot, type B stories climax at death and omit or de-emphasize the remainder; and type A stories slight the divine and birth events in favor of earthly and death events. To summarize, the first and third performance slots end on or emphasize the auspicious events of birth and marriage, and the second slot ends on or emphasizes the tragic event of death. In other words, the opening and closing slots shape performance toward elements definitive of type A narratives, while the center slot shapes performance toward elements characteristic of type B narratives.

This positioning of specific narratives and their typological category in the tripartite festival sequence for both amman and madan koai, and their condensation into a single pattern is diagrammed I figure 5 below :

Figure 5

Thus, the performance sequence is structured by the same type A and B categories that structure narrative typology. However, the present argument would stress two important points of divergence between these systems that organize the performance and non-performance modes in the vil p¡u, the fictive, distant and painless world, at either end of a behavioral sequence, encases a real, immediate and painful one in the middle.

That relationship between the fictive and the real in performance may be represented thus

Figure 6(hhjghjgdjgfsd)

a figure that describes the basic performance structure in vil p¡u. in attempting to understand the significance of this basic structure, two of its properties must be made clear. First, since the festival is a temporal sequence, figure 6 describes movement, the movement of the entire koai. Secondly, it is a movement toward, and then away from, the point of greatest ritual depth. A point indicated by a possession dance associated with a tragic human death. The association of ritual depth and possession dance with other events, birth and sexual/armed conflict, has been mentioned periodically in the foregoing discussion. The following table collates information gathered from local people (mostly performers and dancers) and from observation of festivals concerning the narrative events associated with possession.

TABLE 1 : Association Between Certain Narrative Events and
Possession Dance (Number of times recorded)

Event associated with possession

Solicited information

Observed behavior

Sexual conflict /war



The consistent connection between the events of death, birth and sexual/armed conflict and possession, demonstrated in Table 1, confirms what figure 6 summarizes: movement in performance is toward problems in the real world. The evidence suggests that on the broad level of content (though probably not on more detailed levels), there is a rather direct relationship between oral literature and social reality (cf. Fischer 1963 : 262). Pertinent here is Jacobs' Thesis (1959 : 129-30) that content emphasizes in oral Literature unresolved socio-psychological problems. The problems isolated vil p¡u performance are so fundamental that no resolution is possible, and it can be expected that there will always be tensions surrounding them. The steady increase of ritual depth toward certain problematic events may be interpreted as a push or a drive toward an expression of the tensions they represent. Superimposing this explanation upon the basic performance structure (figure 6), the following figure results.

Figure 7

Figure 7 explains, on the level of the entire festival, how the basic performance structure organizes the expression of unresolved social tensions. In this concluding section of Part II, it will be shown that this organizational structure is replicated on other performance levels as well.

Beneath the level of the festival, in order of decreasing temporal span, the next unit for analysis is the single performance. Each performance is introduced by a formal series of invocatory verses that removes it from the plane of normal reality and, as the Sasta story does for the entire festival, establishes a certain psychic distance and fictive tone. Following the introduction, as the narrative proceeds toward its climax in a problematic events, ritual depth increases, reaching its deepest point in the center performance, and usually manifests as possession dance. When the narrative climaxes in a marriage [as some-times happens in the third performances slot] and no possession occurs, the relative increase in ritual depth is still identifiable by obvious paralinguistic [ululation, or kuravai] and musical (rhythmic pattern) features14. A specific tempo called "tuukku" ("haste" or "urging") is used at points in the performance when ritual depth is greatest. Another indication of increased depth, but not necessarily possession, is a combination of a paralinguistic feature (emphasis) and a vocal/musical feature (rising tone). The association of these features with "deep" events is consistent in my tape recordings of live performances and independently confirmed in a few written-texts (hand-written by performers as notes during performance) in which a special diacritical marks is written at the end of certain verses as a musical cue14. After this release, the intensity returns quickly, sometimes almost immediately., to its pre-performance level.

Figure 8 illustrates how the movement of each of the three major performances is the same as that of the festival and how they collectively contribute to it. In other words, these constituent units individually organize the expression of tension just as the festival whole organizes it. The greater ritual depth reached in the Friday (madan) koai represents a tremendous intensity associated with the death event in the performance of a few male type B stories in the center slot.

Figure 8

The next level on which this movement pattern, or organizational structure, is to be isolated is the constituent unit of the single performance, the single episode. Here again the sequence is organized by the movement of ritual depth toward a problematic event, In addition to birth, death and sexual/armed conflict, other events which constitute a single episode and mark an increase in depth include: pregnancy, commencement and completion of learning, commencement and completion of martial arts training, coronation, and women bathing in apool15. Women bathing is also a creative/erotic motif in Hindu mythology (O'Flaherty 1973 : 22 and passim).15. Except for the first, these are rarely associated with the onset of possession, but are marked for depth by the paralinguistic and musical features noted elsewhere. Finally, consonant with the analyses of the levels of performance, ritual depth is greatest in those episodes which centre on birth, death and sexual/armed conflict and which are closest to (1) the climatic event of a single performance and (2) the centre performance slot of the festival.

The deep event in the single episode is, like that on other levels, encased in the centre of the sequence. However, an important difference is that the segment immediately preceding the event is not characterized by the fictive element, but by fixity. As the story nears any episodic sequence there is a marked increase in the usage of fixed verses (as opposed to the less constrained prose commentary that is also used) and fixed series of verses. Thus, for example, it is conventional to precede a sequence that culminates in connection with a series of set verses describing the woman's penance for a child. The actual birth is prefaced by another series of fixed verses detailing her month-by-month bodily condition. Implicit sexual conflict, likewise, is preceded by a formulaic scene in which the man or woman asks for betel nut from the other. Sexual conflict may also be introduced by a fixed description of a woman and her friends going to bathe in a pool and their play while bathing. Death scenes are somewhat more idiosyncratic, but are also preceded by an increase in fixed versification. After culminating in possession and/or paralinguistic and musical features, ritual depth moves back to its earlier level. Figure 9 illustrates how the single episode sequence exhibits the basic performance structure and how it forms a unit within the single performance sequence.

Figure 9

The importance of fixity in the basic performance structure on this level is explained by the short duration of the single episode. In the longer sequences of the single performance and festival, the push of ritual depth is articulated as a tension between expression of unresolved problems (the real world) and avoidance of such expression (a fictive world). In the smaller episode unit, the conflict that is diffused over those larger units is compressed and thereby heightened. The push toward expression of real problems remains constant, but the avoidance in a fictive world is quickened into a fixity, or restraint, against it in this unit, that is closest to the actual point of greatest depth, the terms of the conflict are clarified and the essential role of the fictive element changes from a counter-balance to the real world to a more active force straining against entrance to it.

Although this co-relation between fixity and ritual depth is most obvious in the episodic sequence, it is evident also on the level of the single performance16. It should be pointed out, since the present focus is not a language, that greater and lesser linguistic fixity; (verse and prose) are the two oral delivery modes in vil p¡u, and that verse, associated with depth, is used to deliver an affective message, while prose, not associated with depth, delivers the narrative message. See Kapferer (1977) for an analysis of Sinhalese healing rituals illustrating this point, and Bird (1972) for somewhat different conclusions in the study of West African singing.16. those performances that involve the greatest depth, and most intense possession, also require the most fixity; sometimes they require a verbatim reading. To bring this discussion full swing back to the macro level on which it began, the very koai festival itself is, as is all ritual on one level, a fixity. It is a predictability required to stabilize indeterminancy, to safeguard the danger, involved in its central purpose, contact with supra-human powers.


In the vil p¡u tradition there are two organizational systems in which the categories and the relation between those categories are constant, but in which their configuration differ. Within the cognitive system, the typology of temples, festivals, and the folk pantheon (on one level) is organized by a distinction between sexual gender amman (female) and madan (male). In the folk pantheon there is another level that classifies deities by a contrast in the nature of their births and deaths: one group, type A, is painless and immortal in a fictive world and the other, type B, is painful and mortal in the real world. A similar distinction between birth and death, explicit as verbal tags and as thematic foci, divides the narrative corpus into the same types A and B. Birth, death and sexual gender are the key semantic distinctions which are split into the terms of a contrast set to form dualistic categories. A fictive birth is opposed to a human birth, the absence of death to a tragic death, birth to death, and female to male. The organizational principle of the cognitive system is thus binary and symmetrical.

The vil p¡u tradition, however, is not only apperceived by cognitive categories; it is also enacted in behavioral patterns. In performance, the basic organizational structure differs from that in the cognitive system in two important respects: it is tripartite and it includes a dimension of depth. Conflicts that are neutralized in the bi-polar equilibrium of the cognitive system require another configuration, the tripartite sequence, in performance. The key distinction between the fictive and real worlds, that bifurcates the folk pantheon (at one level) and story typology, is re-structured so that one polar term (type B) is encased in the centre of a sequence that begins and ends with the other (type A).

The performance moves through this sequence not only on a lateral dimension, but also on a depth dimension that indicates relative ritual intensity. The greatest points of ritual depth are marked by possession dance and are coincident with the events of birth, death and sexual/armed conflict in the sung narrative. The push and recession of ritual depth represents a tension between a drive to express these unresolved problems in a real world and a need to avoid them in a mythic world.

The interpretation of the real/fictive distinction in vil p¡u provides a possible understanding of how it operates in other contexts. As native ascriptions of belief and non-belief, the real and the fictive have been standard criteria for the classification of folk literature. In most Western cultures the distinction between legend and ballad, on the one hand, and tale, on the other, is that the former take place in a real world and the latter in a fictive world17. The issue of belief in legend is more complex than this, however, as Degh and Vazsonvi (1976) have shown.17. moreover, in some African cultures all oral narratives are divided into two categories: what is believed to be true and what is thought to be fiction (Bascom 1965).

The foregoing discussion suggests that these categories for classifying oral (and literary) genres may be understood as different but complimentary frames of reference between which a culture alternates myth, which is believed to be true but in a distant, fictive world, would then represent a compromise between them. In a recent article, narrative scholar Degh contrasts legend and tale in terms sympathetic to the present essay: "The tale takes place in the 'once-upon-a-time' 'never-neverland' on a different level of existence, whereas the legend brings its fearsome mysteries down to earth. It happens here and now … The tale gives relief from anxiety; the legend arouses it and leaves man alone with his anguish. The tale … is very useful. The legend … is very dangerous" (1979 : 99).

The particular patterning of the fictive/real conflict in vil p¡u suggests also a reversal of its standard psychological interpretation. The "relief" of the fictive tale and the "anguish" of the real legend implies, as has been historically assumed, that the prominent fantasy element in oral literature is a release-mechanism for unexpressed ideals, wishes and anti-social emotions. However, in the performance of this Tamil oral tradition, release comes in contact with the real world, not with the fictive. Fictive expression of problems does achieve a measure of release, but the greater drive, indicated by greater ritual depth, is toward the expression of problems in their real setting. In addition to the need for fantasy, perhaps this need for reality, when taken into account, will help to explain other performance traditions.

As our analysis of the vil p¡u material demonstrates, the reality/fiction distinction may also be useful to the understanding of a single narrative (or single performance). In his study of the Asdiwal myth, Lévi-Strauss delineated four levels of content and noted that two are exact replications of reality, one totally fictive and one a mixture (1967 : 158). He does not develop this point precisely because he regards these levels as the superficial and variant expressions of a deeper contradiction, marriage with the matrilateral cousin. It is the present argument that, on the contrary, the difference between these levels represents a conflict between two approaches for dealing with social contradictions, a conflict which is at least as important as the specific conflict(s) they address, on the level of content or underlying message.

The central conflict in vil p¡u performance is between these two approaches to unresolved socio-psychological problems, between the need for expression of them and the need for restraint toward them. That conflict is organized by what was termed the basic performance structure and isolated on three levels, the festival, the single performance and the single episode. A composite representation of how the basic performance structure determines the patterns of performance in figure 10 below :

Figure 10

Performance in vil p¡u, as illustrated in figure 10, is a multileveled network in which each level is both constitute by, and is linked to the others by, the basic performance structure. It recurs repeatedly in the shorter temporal sequences which collectively and accumulatively contribute to its recurrence in longer sequences.

This model of vil p¡u performance is essentially one of conflict in motion. But the movement occurs neither at regular intervals nor does it proceed with a measured pace. As a response pattern to unresolved problems, tension build up and release pattern to unresolved problems, tension build up and release produces an irregular pulse that repeats at irregular intervals. The paradigm of performance in vil p¡u, then, is a paradigm of a rhythm, a spasmodic-like alternation between expression and restraint18. Geertz (1973 : 445) has made a similar observation about the "spurts" in Balinese life.18. Performance proceeds by a series of short stop-and-start movements toward a temporary resolution in possession.

Finally, certain core patterns in a culture are laterally inscribed across broad domains, e.g. kinship or art, and others are compressed into a dense event. This attempt to isolate a cultural paradigm has identified a core pattern, the basic performance structure, in a dense event, the performance of a group's principal oral tradition. In the crucible of performance, conflict is quickened into motion and a culture's organizational paradigm brought visible into operation. In vil p¡u performance this paradigm is one of a rhythm, an uneven vacillation between expression and restraint.

In conclusion we return to the methodological concern with which we began. Cultural analysis is increasingly concerned with the models that organize peoples' lives. There is, however, a tendency to assume that those models are necessarily thought processes, and it is often forgotten or considered a less elegant argument that people also create and comprehend behaviourally. This essay is an exposition of a behavioural model, of performance as paradigm, which as Geertz (1973 : 448) has put it, provides a culture with a "reading" of itself. One final point must be clearly made: the performance paradigm is not a surface reflection of a deeper cognitive model. The performance of oral tradition is not only expressive behavior, but is itself an alternative and independent medium for the organization of cultural meaning.

Abrahams, Roger 1968 Introductory Remarks to a Rhetorical Theory of Foklore Journal of American Folklore, 81: 143-58.
Aiyappan, A .1976 Deified Men and Humanized Gods : Some Folk Bases of Hindu Theology.  In Agehananda Bharati (ed.) The Realm of the Extra-Human : Agents and Audiences, 139-48.
Austin, J.J. 1975 (1962) How to Do Things With Words. Cambridge: Harvard University Press.
Bascom, William 1965 The Forms of Folklore : Prose Narratives. Journal of American Folklore, 78 : 3-20.
Bateson, Gregord. 1936 Naven. Stanford: Stanford University Press
Bauman, Richard 1977 Verbal Art as Performance. Rowley : Newbury House Publishers.
Beck, Brenda
Peasant Society in Kongu.  Vancouver : University of British Columbia Press. Colour and Heat in South Indian Ritual. Man, 4 : 553-72.
Ben-Amos, Dan. 1972 Toward a Definition of Folklore in Context.  In Americo Paredes and Richard Bauman (eds.) Toward New Perspectives in Folklore, 3-15. Austin: University of  Texas Press.
Bird, Charles 1972 Aspects of Prosody in West African Poetry, In Braj Kachru and Herbert stahlke (eds.) Current Trends in Stylistics, 207-16.  Cahmpaign: Linguistic Res., Inc.
Black, Mary 1973 Belief Systems.  In John Honigmann (ed.) Handbook of Social and Cultural Anthropology, 509-78.  Chicago: Rand McNally.
Black, Max. 1962 Models and Metaphors.  Ithaca: Corness University Press.
Blackburn, Stuart




The Folk Hero and Class Inteests in Tamil Heroic Ballards. Asian Folklore Studies, XXXXVII, 131-49.  Reprinted in Ved. P. Vatuk (ed.) Contributions to Indian Folklore-I (Meerut: Folklore Institute) in Press.

Performance Context and Semantic Structure in the Study of Oral Tradition. Unpublished paper presented at the Annual Meeting of the American Folklore Society, Los Angeles.
Das, Veena 1977 Structure and Cognition.  Delhi: Oxford University Press.

Dégh, Linda


Grimm’s Household Tales and Its Place in the Household.  Western Folklore, XXXVIII : 83-103. and Andrew Vazsonyi 1976. 
Legend and Belief.  In Dan Ben-Amos (ed.) Folklore Genres, 93-124. Austin: University of Texas Press
Fischer, J.L.1963 The Sociopsychological Analysis of Folktales.  Current Anthropology, 4 : 235-96.
Geertz, Clifford.1973 The Interpretation of Cultures.  New York: Basic Books.
Gomathinayakam,Ti.Ci   1979 Tamil Villup Pattukal. Madras: Tamil Pathippagam.
Hameed, K.P.S.1956 Bow Song : A Folk Art from South Travancore. Tamil Culture, V : 274-84
Harper, Edward1964 Ritual Pollution as an Integrator of Caste and Religion.  In Edward Harper (ed.) Religion in South Asia, 1951-96.  Seattle: University of Washington Press.
Hymes, Dell 1974


The Contribution of Folklore to Socio-linguistic Research.  In Hymes The Foundations of Socio-linguistics, 125-42.  Philadelphia Breakthrough into Performance.

In Dan Ben-Amos and Kenneth Goldstein (eds.) Folklore: Performance and Communication, 11074. The Hague: Mouton.
Jacob, Melville 1959 The Content and Style of an Oral Literature. Chicago: University of Chicago Press.
Kapferer, Bruce 1977 First Class to Maradana : Secular Drama I Sinhalese Healing Rites.  In Sally Moore and Barbara Myerhoff (eds.) Secular Ritual, 91-123. Amsterdam: Van Gorcum.
Kluckhohn, Clyde 1965 Myths and Rituals : A General Theory.  In William Lessa and Evon Vogt (eds.) Reader in Comparative Religion, 144-58.  New York : Harper and Row
Kuhn, Thomas 1962 The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Lévi-Strauss,Claude 1967

Structural Anthropology. Vol-II, New York: Basic Books.
Lomax, Alan ( 1968 Folk Cults of Bengal. Calcutta: Indian Publications.
Marriott, McKim 1976 Hindu Transactions : Diversity Without Dualism.  In Bruce Kapferer (ed.) Transaction and Meaning, 109-42. Philadelphia: Institute for the Study of Human Issues.

O’Flaherty, Wendy Doniger 1973

Asceticism and Eroticism in the Mythology of Siva. New York: Oxford University Press
Ortner, Sherry. 1975 Gods’ Bodies, Gods’ Food : A Symbolic Analysis of A Sherpa Ritual. In Roy Willis (ed.). The Interpretation of Symbolism, 133-69.  New York: Halsted Press
Peacock, James


Society as Narrative.  In R. Spencer (ed.) Forms of Symbolic Action, 167-77. 
Proceedings of the 1969 Annual Spring Meeting of the American Ethnological Society, Seattle.  Seattle: University of Washington Press

Reiniche, Marie-Louise 1979

Les dieux et Les Hommes. Cashiers de L’Homme, XVIII.
Schieffelin, Edward 1976 The Sorrow of the Lonely and the Burning of the Dancers. New York: St. Martins Press
Shulman, David 1980 Temple Myths.  Princeton: Princeton University Press.
Singer, Milton


The Cultural Pattern of Indian Civilization.  Far Eastern Quarterly, xv : 23-36. The Great Tradition in a

Metropolitan Center : Madras.  In Milton Singer (ed.) Traditional India : Structure and Change, 141-82
Turner, Victor 1974 Dramas, Fields and Metaphors.  Ithaca: Cornell University Press
Van Gennep, Arnold 1960 (1908) The Rites of Passage.  Chicago : University of Chicago Press. Trans. Monika B. Vizedom and Gabrielle I. Caffee
Wadley, Susan 1978 Texts in Contexts : Oral Traditions and the Study of Religion in Karimpur.  In Sylvia Vatuk (ed.) American Studies in the Anthropology of India, 309-341. New Delhi: Manohar.
Whitehead, Henry 1976 (1921) The Village Gods of South India. Delhi: Summit Publications.