back to Diskus7

DISKUS Vol. 7 (2006)


Seth Kunin
Department of Theology and Religion
Durham University, UK

Email: s.d.kunin@durham.ac.uk


[to follow]

The Dialectical Relationship Between Theory and Ethnography
Seth D Kunin
Durham University
Durham, UK

During the history of Religious Studies as a discipline, there has developed a tension concerning the relationship between theory and ethnography: though it must be said that each of these has been taken in a very narrow sense.  Interestingly, both sides of the polarized argument have tended to use similar models of the two.  Theory is seen by both as the product of reductionism and arises from disciplinary perspectives outside of the religious studies disciplinary framework and ethnography has been seen as an internal point of view free of external theorizing.  There are a number of scholars who have attempted to bridge this divide, for example Ugo Bianchi, in his dialectical approach, as well as approaches that tend to be particularly associated with other disciplines, for example, sociology or anthropology. In many of these, Bianchi aside, there tends to be no clear view of how theory and ethnography can be placed in creative dialogue rather than in opposition to each other.  Thus, for example, in Steve Bruce’s sociological analysis of religion there is a dialogue of sorts between his theoretical perspective on secularization and the societies he analyses, but no attempt to redefine that perspective in relation to new ethnographic material.  Within the discipline of religious studies, in part as a mechanism of boundary setting and disciplinary justification, the polarization is often retained and can be broadly characterized in terms of the phenomenologist and reductionist approaches.

This paper critiques this opposition on two levels. First, it presents a brief analysis of the two approaches and challenges some aspects of their respective understandings of theory and ethnography.  As much of this critique is well known (and indeed often takes up aspects of the two approaches’ critiques of each other) the presentation here will only present a somewhat abstract and consciously stereotypical view.  The aim of this discussion is both to highlight the problems with their views on theory and ethnography and to offer some suggestions for how these may be understood within a dialogical framework.  Second, the paper presents a somewhat autobiographical discussion tracing my own use of theory and ethnography. Here, I attempt to demonstrate (through this retrospective and therefore artificial means) how the two may be set in a creative dialogue that both enhances the understanding of the societies being analyzed and the theoretical perspective that underlies the analysis.  It should be noted that I am neither attempting to argue for the value of structuralist theory nor indeed the quality or insight of my own work; rather I am providing a model for how theory and ethnography should be dialectically related and which should be used in relation to any theoretical perspective or ethnographic context.  The discussion of the interrelationship of theory and ethnography is also abstracted in an important respect.  Clearly the development of the two and their dialogue also depends on the development of broader theoretical discussions and their influence on the development of a particular application of theory.  This is an important part of theoretical/ethnographic autobiographical development but one, for the sake of space and clarity, that is only alluded to.

For the purposes of this discussion the approach taken by the phenomenological study of religion will be taken as emblematic of the study of religion in a broader sense.  This is justified by the fact that many scholars within the broader discipline have suggested that religion is an object of study in its own right (separable from other cultural domains) and that religion should be specifically studied by theories of its own rather than those from other fields of study.  While this second aspect does not eschew theory, very often these approaches argue for theories that are either religious (depending, as in Otto, on the existence of a transcendent other) or that arise from the specific religious tradition being studied.  Arguably, this position is actually atheoretical as the theoretical approaches used are by definition either specific to an object in general (that is assumed to have an essential basis) or are specific to a particular object.  It is suggested that theory, as theory, needs to be abstractable and therefore non-specific in nature.  If it is used in relation to a category, for example ‘religion’, there must be a clear demonstration that the category is separable and distinct rather than part of a more general category.  Although many approaches to religion depend on an assumption that religion is a distinct category, most if not all rely on an undemonstrable basis for this pre-supposition.  Without such a demonstration religion should be considered one cultural domain among others and analyzable through the same theoretical perspectives as other, non-religious cultural phenomena.

The phenomenological school (with some exceptions) assumes a transcendental basis for religion or the religious experience. <1> This provides the basis for analyzing religion as a category distinct from other domains, or indeed as a category having a single historical trajectory separate from the particular culture in which it is instantiated. <2>   If the transcendental aspect is absent, the approach, in order to remain intellectually coherent, requires some other form of essentialist basis for the category.  Often this leads to a circular argument, i.e. this is the definition of religion, that is, that it must have these specific dimensions, and that only things with these dimensions (by definition) are religion.  There very rarely is any clear demonstration of why these elements rather than any other set make religion a unique category of analysis.  It may be suggested that a functionalist approach (which characterizes many reductionist analyses) also is circular in the same way.  While this is largely true, the nature of the definition at least provides an abstractible basis for the inclusion of specific elements and not others.  Thus, Radcliffe-Brown’s suggestion that religion has the function of social control provides in its definition an abstractible analysis rather than merely, and seemingly arbitrarily, describing/defining the phenomena. <3>

Phenomenology of religion is characterized by three methodological elements: epochē, einfühlung and eidetic vision: epochē is the bracketing out of the observer’s own system of belief; einfühlung is the practice of sympathy or more often privileging internal systems of explanation, and eidetic vision refers to the phenomenological practices of typology or morphology.  Each of these elements has clear implications in relation to theory.  The first two should be understood together.  They emphasize the importance of understanding the ethnographic context in its own terms, but perhaps even more importantly argue against the application of any external theoretical model, privileging instead internal modes of analysis.  This approach is in fact atheoretical as it either is merely the replication of an internal understanding or theology, or the development of an internal theory that is by definition not abstractable – any form of abstraction out of its ethnographic context would be methodologically problematic.  The third element is ironically in conflict with the other two – it is and must be based on a preexistent understanding of religion and of the significant categories into which religious practice can be divided.  There is no suggestion that this is specifically done for each religious tradition according to its own typology, rather all religions are fodder for a pre-existing typology and there is a presupposition that these elements ultimately have the same essence where ever they may be found.

While the third of these methodological approaches is arguably the most problematic, it is the atheoretical bias of the first two that is relevant for the discussion here.  As suggested, these elements (and, as suggested, the third as well) assume that religion has a separable essence that can only be understood in its own terms.  Theories, either external to the study of religion or indeed from a specific culture itself, cannot be used to understand the phenomena.  Ultimately, these elements have the effect (at least in terms of their own argumentation if not their practice) of placing the emphasis on ethnography and denying either the application of theory or perhaps more specifically the application of any theory that is seen as reducing religion to a factor outside of itself.

While aspects of the phenomenological approach may be laudable in and of themselves, it is questionable whether it actually takes one further in the understanding of religion. <4>   The basis of the approach is essentially circular, and thus any analysis is merely descriptive, and on the basis of it being analyzed, conforms to the pre-existing definition.  To the extent that the analyses are ethnographic, they merely replicate the internal point of view and are therefore purely of descriptive interest, awaiting theoretical analysis rather than providing it.  If we take the third element, there is indeed a form of theorizing implicit to it. Eidetic vision, however, often assumes a typology or morphology and with very little argumentative basis finds similarities or meaning that often s reside solely in the mind of the theorist.  The only basis for the eidetic vision (which is as static as the other aspects of phenomenology) is the assumption that all religion is necessarily/essentially the same and any element can automatically be decontextualized and associated with apparently similar elements.

The second side of the argument is stereotypically found in the reductionist camp.  Scholars representing this approach (and indeed my own approach is a form of reductionism) argue that religion must be seen as a cultural/human product and analyzed using the same theories applied to other aspects of human experience.  Although reductionism is often associated with functionalist analysis - for example, the argument that religion can be reduced to the function of social control -  in fact it covers a much wider domain, as for example the structuralist approach suggested below in which all cultural domains, including religion, are reducible to underlying patterns of structuring.   While on the surface the reductionist argument as presented, for example, by Segal (1994) and Strenski (1994), does not have necessary implications for the relationship between theory and ethnography, it necessarily privileges theory over ethnography. Ultimately there must be one abstract theoretical approach that explains religious manifestations within the context of other human domains. <5>

If religion does not have a unique and separable basis from other social phenomena, and we would agree with that proposition, then the principle of an abstractable theoretical understanding is intellectually coherent. This of course leaves open the possible application and testing of different theoretical abstractions.  In some versions of reductionist argumentation, the priority of theory and the implications of reductionism regarding the need for abstraction leads to a version of theorizing that suggests that the form and content of the theory is independent of any ethnographic context, i.e. that ethnography must conform to the theory rather than be in dialectical relationship with it.  The nature of this dialectic relation, however, is significant; we agree that while theory will in principle be abstractable and applicable, at any one time it must be seen as provisional and the process of dialectical relation with ethnography provides the basis for testing and refining the theory.  This process does not result, as some have argued, in a unique theory or analysis for each ethnographic context; rather it leads to a refined theory that should be applicable in any ethnographic context.

In this section we have touched on some aspects of two potentially opposing approaches to the study of religion.  The most fundamental difference between the approaches relates to their understanding of the nature of religion; while the phenomenological approach views religion as unique, based on some essential element that causes it to differ from other human constructs, the reductionist approach considers it to be one among many human constructs and as such amenable to the same theoretical and methodological forms of analysis.  Based on this presupposition the two approaches place differential emphases on theory and ethnography – with the most extreme versions of each suggesting that one of these emphases to a large extent excludes the other (particularly if ethnography is viewed as specific and descriptive, rather than intrinsically and theoretically motivated and if theory is viewed as almost exclusively abstract and independent of ethnographic amendment or validation).

In this paper we argue for a different approach to both of these elements.  Following much recent scholarship it seems impossible to consider any form of ethnography as unmotivated – untheorized material is unconsciously theorized rather than an actual depiction of an internal point of view.  Thus, the initial theoretical positioning is an attempt to make conscious and clearly articulated the inherent theorization.  From the opposite direction, theory can not be independent of ethnography; initially it needs to be understood within its own context and progressively to be challenged by new ethnographic contexts.  While this may not lead to the final abstraction or a ‘theory of everything’, it can make the theorist aware of the provisionality of the theory and reduce its reliance on the originating context.  Such a dialectical use of ethnography should not merely lead to a particular version of the theory, but rather be used as a means of refining a more abstract version, which can then be used in relation to ethnographic contexts already analysed in order to throw upon them new and hopefully revealing light.

This second section examines the practice of anthropological research to demonstrate the arguments relating to the necessary relationship between theory and ethnography presented in the first part of the paper.  The discussion is presented in autobiographical form, relating to the development of my own application of structuralist theory to a range of different ethnographic contexts, both textual and living.  As in all autobiographical material the discussion is somewhat artificial, relating more to present understanding than to my thinking during the course of the research projects discussed. Nonetheless, it provides a good example of the type of dialectic relationship that is being argued for in this paper.

One of the key themes discussed above is the impossibility of the naive observer, and thus the need for an articulated theoretical foundation to all ethnographic research.  My own work was and remains within the anthropological structuralist model first articulated by Claude Lévi-Strauss.  The initial approach was of an orthodox structuralist form closely based on the discussions in The Savage Mind (1966), Structuralist Anthropology volumes one (1963) and two (1963), and the Mythologiques(1970, 1973, 1978, and 1981). <6>   The theoretical basis of this approach included the main aspects of classical structuralist theory.  Thus, for example, although it took a more nuanced view of the basis of the binary nature of underlying structure than did Lévi-Strauss, nonetheless it was based on an essentially binary understanding.  The approach was also orthodox in seeing underlying structure as entirely unconscious, and thus offered little room for agency or transformation on the conscious cultural level.  These elements were the starting point for the ethnographic research which formed the basis of The Logic of Incest and a number of other scholarly publications. <7>

The ethnographic content of this research was twofold: there was a significant focus on the biblical text and a secondary focus on rabbinic texts, particularly those which expanded or interpreted the narratives found in Genesis.  Alongside the theoretical apparatus that formed the basis for interpreting the textual material, the methodology for analyzing the material was also that described by Lévi-Strauss. <8>   The results of this analysis both confirmed and expanded my structuralist theoretical expectations.

The material included in The Logic of Incest was largely narrative myth, though there is also some brief discussion of non-narrative cultural domains.  Analysis demonstrated that the narratives contained a pervasive oppositional structural form, not dissimilar from forms suggested by Levi-Struass.  Other discussions of biblical material, as for example in God’s Place in the World (1998) and We Think What We Eat (2004) extended the analysis to geography and food and other cultural domains. <9>   These analyses suggested that the binary structures found in relation to the narrative material were both pervasive and culturally monolithic. 

The nature of this analysis is best illustrated with reference to the Israelite food rules. <10>   The animal world can be divided into three main groups: air, land, and water.  These three groups should not be considered structurally significant as they are not set in clear opposition to each other.  If Genesis chapter one is considered, all three of these groups are individually created and specifically described by god as ‘good’.  The only relevant structural oppositions in the natural world are the opposition of the creeping/swarming creatures with the other animals and the opposition between the animals, described as ‘good’, and humanity, which is described as ‘very good’.

The three groups, based on the texts in both Leviticus and Deuteronomy can then be structurally analysed on the basis of structurally opposed binaries.  The water animals are the simplest; they are divided on only one level.  The positively valenced half of the opposition is clearly defined as those animals with fins and scales.  Any animal with both of these characteristics is intrinsically in this category.  The negatively valenced category includes all animals that have either only one or none of these characteristics.  In relation to food specifically there is no reason to subdivide this category.  All of the above are equally forbidden.

A similar pattern is evident in relation to the air animals.  They too are divided into two oppositional categories.  The positively valenced category is defined as those animals that fly, hop, (have feathers), <11>  and do not eat meat (other than fish) or carrion.  The negatively valenced category, as in the case of the water animals, includes all those lacking in one or more of the necessary characteristics.  As all of those in the negative category are equally forbidden there is no reason to subdivide this category.  Unlike the water animals, however, there is a sublevel within the permitted category if the cultural domain of sacrifice is added to the discussion.  This sub level recapitulates the structure of the preceding level.  The permitted air animals are divided into two opposing categories: animals that can be sacrificed and those that cannot.  The positively valenced category is defined by birds that are domesticated.  The negatively valenced category includes all other birds.  The key aspect of this structural system is that each level recapitulates precisely the same structural form.

The land animals are categorized in the same way, though due to their significance as sacrificial animals they include further levels of structural recapitulation.  The top level divides land animals into those that are permitted and forbidden to be eaten.  The positively valenced category is defined by chewing cud and having cloven hooves.  The negatively valenced category are all animals lacking in one or more of these elements.  Despite the fact that some scholars (most notably Mary Douglas <12>) consider the pig and other animals, those lacking one element, as forming a mediating category, in terms of the food rules there seems no basis for this distinction.  A bear and a pig are equally forbidden, despite the fact that one is lacking one element and the other two.

In relation to sacrifice the permitted category can be divided in to at least two further sub-levels.  The first is identical to the air animals, that is, an opposition between domesticated and wild animals.  The domesticated category can be sub-divided into firstborn animals in the positively valenced category and other animals in the negatively valenced category.  It is important to note that these non-firstborn animals are only negative in relation to the firstborn animals; it would not be appropriate to set them in relation to higher level categories.

If the structural rule underlying both the food and sacrificial system is abstracted we find a clear binary oppositional structure that can be characterized as A not B.  In each case the categories are clearly and intrinsically defined.  If an animal is born into one category it can never move to the other – the categories are both intrinsic and unbridgeable.  To this point the ethnographic material fully supports both the theoretical expectations of pervasive structure and the binary oppositional form that is found throughout classical structuralism.  This equation (A not B) is also demonstrated in the works cited above to be characteristic of all domains within biblical culture.  The analysis of the food rules, however, also begins the process of dialectically relating ethnography and theory.  Structuralist theory did not suggest the recapitulated levels found, and thus the ethnography has offered new possibilities for the theoretical understanding of underlying structure.

While the first part of this research was an application of a theoretical and methodological approach to a new body of ethnographic material, the second part, that is, the analysis of the diachronic development of the narrative material within rabbinic texts, sought to test the theoretical perspective.  It addressed the issue of transformation within a textual and cultural tradition.  Although arguably all Lévi-Strauss’ work is about transformation, the nature of the transformations he discusses is significantly different from that discussed in this research.  Due to the nature of the cultures he analysed, predominantly in South and North America, none of which had a literary tradition, he focuses on geographic development rather than diachronic development.  Thus, he analysed the processes of transformation as mythic material moved from one culture to the next.  Within the Israelite/rabbinic material, however, there was a strong literary tradition, linking the biblical texts with retellings and expansions of those texts during more than one thousand years of rabbinic tradition.  The research therefore raised the question of the nature of transformation within an arguably continuous narrative tradition, and one that considered itself to be culturally continuous.  It also provided a diachonic basis for analyzing transformation within a continuous cultural context as opposed to the intra-cultural context found in Lévi-Strauss’ work.  The outcome of this analysis suggested that while underlying structure itself was essentially conserved, as would be expected from a structuralist theoretical perspective, transformations did occur, particularly in respect of emphasis or de-emphasis of particular mythemes and the crystallization of structure as the texts were told and retold. 

Interestingly, this process of structural crystallization allowed for the possibility of developing arguments about how structural elements develop through time and the testing of those arguments in relation to an actual textual tradition.  For example, the analysis of the biblical narrative of Genesis 22, the sacrifice of Isaac, suggested that the two young men who accompanied Abraham and Isaac to the sacrifice but did not participate in the actions were structurally in the role of Isaac’s brothers.  This structural identification was borne out in the crystallization of the narrative; by the eighth century C.E. the young men are identified as Ishmael and Eliezar (Isaac’s half and adopted brothers respectively) in almost all rabbinic sources. <13>

While, as indicated, the analysis of the biblical/rabbinic ethnography supported aspects of classical structuralism, the theoretical approach was specifically not a theory about biblical/rabbinic ethnography.  That is to say, it was not intended for a single ethnographic context.  As indicated above, a proper use of theory should extend to any ethnographic context, that is, be abstractable.  The specific structure found in the biblical/rabbinic material is particular, the theory and methodology that provided the basis for determining that structure should not be particular.  On this basis it was important in the diachronic development of my research to extend the analysis to other, non-biblical/rabbinic cultural contexts.

Two new ethnographic contexts, the New Testament and The Book of Mormon, suggested themselves; each of these utilized biblical mythemes and thus would enable a clearer view of the potential transformations in structure while also being (at least arguably) culturally distinct from the ethnographic material studied in the original research. <14>   Although neither of these was analysed to the same extent as the original ethnographic material, they provided a basis for testing structuralist theory, and also, particularly in respect of the LDS material, for challenging the binary emphasis often found in traditional structuralism.

The New Testament material provided a direct comparison with the material analysed in the original project – most of the same mythemes were present, for example, virgin birth and death and rebirth.  These mythemes, however, were placed in a different structural configuration than was found in the Hebrew Bible material.  The analysis suggested that while the structure remained essentially binary it was not strongly oppositional.  The New Testament material was characterized by a mediated binary system with movement possible between the two categories.  This was clearly different from the Hebrew Bible model in which the binary was unmediated and allowed no movement between categories. 

The difference between the two structures is clearly seen in their respective definitions of identity.  The Hebrew Bible definition is composed of two categories each defined on an intrinsic basis.  Israel is defined genealogically (initially having an Israelite father) and thus the rejected non-Israelite category is also intrinsically defined (as lacking the genealogical element). <15>  The New Testament model defines self on a very different, non-intrinsic basis, that is, faith.  An individual starts in the rejected side of the binary, but through faith (and a mediator, Jesus) can move into the privileged category.  The two respective understandings of the divine also highlight the structural difference: the Hebrew Bible model develops a strong unmediated opposition between god and the created world; the New Testament (in a crystallized form) has a mediated relation with Jesus containing both human and divine elements and thus bridging the gap between the two categories.

While the material from the New Testament does not challenge structuralist theory, it provide a basis for both rethinking stereotypical or simplistic understandings of structuralism as well as adding in additional tools for the analysis of structural systems.  Many views of structuralism over emphasized the oppositional nature of structuralist analysis; this, however, was not part of Lévi-Strauss’ own discussion or application of the theory.  The New Testament ethnography presents a binary structure that is not rigidly oppositional and thus suggests that the forms of underlying structural equation are more varied than might have been thought.  The significance of mediation in the New Testament structure also suggests that different forms of mediators may allow for both differences between ‘cultural’ forms and indeed gradation within a particular ‘culture’.  This final point is perhaps the most important; structuralist theory, in line with much contemporary anthropological discussion, tended to present culture as monolithic.  The flexibility and variability of mediation opens the possibility of a wider degree of variation within a single structural complex.  This is further supported by the variations on mediation found within the complex tapestry of the four Gospels.

The material from the Book of Mormon raises an even more significant issue for structuralist theory; it is characterized by triadic rather than binary structure.  Analysis of the Book of Mormon presented both similarities and significant differences from that of the New Testament.  The narrative material in the Book of Mormon was characterized by very similar mythemes to those of the Hebrew Bible and the New Testament.  Thus, for example, many texts developed the death/rebirth mytheme.  The structural relations of these mythemes, however, always included three elements rather than two.  A good example of this is found in the family of Lehi, who is seen as the founding patriarch, moving out from the rejected old Israel, and founding the new Israel.  Lehi had six sons who can structurally be divided into three groupings: the two oldest sons, Laman and Lemuel are in the negative category (and are born prior to leaving Israel), Nephi and Sam  are in a transformed positive category (and are also born prior to leaving Israel), Jacob and Joseph are in an intrinsic positive category (and are born after leaving).  While in some senses the middle category has aspects of mediation, its function in the narrative and in other iterations of the structure suggests that it is a structural category in its own right. 

The structural equation that characterizes the Book of Mormon has elements of both the Israelite (Jewish rabbinic) model and that of the New Testament.  The first part of the equation (the relation between Laman/Lemuel and Nephi/Sam) is characterized by a neutral relation, that is, the categories overlap and there is movement on the basis of, for example, faith between them; this part of the equation is identical to the structure found in the New Testament and can be abstracted as AnB.  The second part of the structure includes a positive relation (this is seen in the structural relation between Nephi/Sam and Jacob/Joseph).  A positive structural relation indicates that the two categories in terms of quality contain the same information – the information is valued differently depending on the other element with which it is set in relation.  This aspect of the structure can be abstracted as B+C.  The significant aspect that is identical to Israelite/rabbinic structure is the relation between C (Jacob/Joseph) and A (Leman Lemuel); these categories are both intrinsic and there is no possible movement between them.

This structure can be further clarified when applied to an understanding of communal identity.  The first category A are the gentiles, this is a negative category.  Gentiles, however, due to the possibility of movement between A and B can be transformed into Mormons through faith.  The interesting aspect of the structure is, however, found in the relation between B and C.  As indicated, although these contain the same content, the content is defined differently in each.  When in B individuals are defined as Mormons by faith; in C they are Mormon’s by birth, that is intrinsically part of the group in the same way that Israelites are defined by genealogy and therefore intrinsically Israel.  The LDS temple ritual of retrospective conversion provides the basis for this change in understanding.  Through conversion of one’s ancestors one becomes intrinsically Mormon rather than merely Mormon by transformation.

The move to new ethnographic contexts allowed for the material to be brought into dialectical relationship with the theoretical perspective.   This dialogue suggested both new aspects that could fit within the preexisting framework and elements that challenged the perspective as it was commonly understood.  Thus, while the role of mediation found in the New Testament material allowed for a more nuanced and tightly focused analysis of ethnographic material, the triadic structure of the LDS material suggests the need for a reassessment of the nature of structure.  This reassessment argues for both an extension of possible structural forms and variations on underlying structure; it also, however, suggests that focusing the analysis on particular structures on the most abstract level may be inappropriate.  It seems that it is necessary to focus that level of analysis on a ‘structuring principle’ rather than a common and fundamental underlying structure.

The product of this dialectical process, however, also allows for a rethinking of the original ethnographic analysis.  Thus, if the concept of mediation (on the structural rather than narrative level) is taken seriously, then there are clear indications of the types of mediation that might be found in the Israelite material.  Given the strongly oppositional structure that is characteristic of all Israelite cultural domains, mediation of the type found in New Testament material would not be possible.   Thus, if any apparent mediators were found in biblical material they would be very problematic for the underlying structural system – these mediators would be strongly negative and need to be defined out of the system; structurally they can not exist.  This allows us to reassess Mary Douglas’ retention of the pig as a mediator, that is, it has one element of the positive category and one element of the negative category.  This understanding of mediation provides an explanation for the extreme negativity of the pig – it needs to be clearly placed in the negative category in order to make it clear that it is not a mediator and that structural mediation is impossible.  It is for the same reason that Esau is a particularly negative character and ancestor of the Edomites and Amalakites – as a twin he appears to mediate between Israel and the nations.

The structural variation within the Gospels also challenges the monolithic view of culture and underlying structure.  This view of internal cultural variation is clearly also important in respect of the biblical material within which biblical scholars have long recognized different schools, authors and traditions.  This reassessment provided  the basis for a reexamination of models of space within the biblical text.  The original analysis had suggested some aspects of variability, particularly a centralized model focusing on a single sacred space, the Temple or Tabernacle, and a multiple model associated with the subdivisions of Israel. <16>  The reanalysis suggested an even greater variability and focused on how ideology relates to the primacy of the centralized model.  An interesting aspect of the reanalysis, however, was that all of the variants were based on a single underlying structural model.  This aspect of the analysis suggests that culture is in many respects an artificial construct with a multiplicity of internal divisions based, for example, on interest or class, and with overlapping and fuzzy edges.  Nonetheless, if there is a perception of commonality and the need for communication (which may be and indeed often is a product of ideological imposition) these multiple entities will be characterized by a fundamentally similar structure.  Variation tends to be found relating to emphasis and de-emphasis and the introduction of subtle modifications in mediation rather than significant transformations in underlying structural equation.

The next ethnographic context provided a very different view of underlying structural processes both because it was a modern setting (of living people) and due to the inherent complexity of the cultural context. <17>   The research focused on the Crypto-Jews of the American Southwest.  This ‘community’ is composed of individuals who are (or at least believe that they are) descended from Jews who were forcibly converted to Catholicism in Spain and Portugal between 1390 and 1497.  Many of these individuals, alongside the perception of either past or present identity, also retain a wide range of practices that they associate with their Jewish origins.  The ethnographic context is complex on many levels.  The forms of identity are variable, ranging from individuals who have a very strong and ongoing connection with Judaism to those that have only a marginal connection.  Added complexity is also found with respect to religious affiliation; some Crypto-Jews retain a strong Christian aspect (with membership today in all forms of Christianity), others are moving into mainstream Jewish communities, while others are secular, seeing their Crypto-Judaism as purely cultural.  A final level of complexity is found in the high degree of fluidity; individuals are constantly redefining themselves in relation to all of these factors.

The Crypto-Jewish ethnography provided an important basis for testing the issue of structural variability suggested by the New Testament material and the reanalysis of sacred space.  Each of the complexity factors was related to subtle variations on underlying structure.  The structural models employed ranges from a strongly oppositional A not B model similar to the biblical form to a mediated model similar to that of the New Testament.  As individuals moved along the different axis of complexity there was a related structural movement.

This ethnographic context suggested a need for a reassessment of the role of agency within underlying structure and a related mechanism for cultural/structural transformation, a factor which had been absent from some more static versions of structuralist theory. <18>   Traditional structuralism, because of its emphasis on the unconscious nature of underlying structure, provided no theoretical basis for agency.  Individuals were bound by the structure and could only unconsciously replicate it.  This theoretical perspective also suggested an obstacle to the understanding of transformation.  If structure could only be replicated (as in a genetic model), then there were only two possibilities for change, outside influence and mutation.  While both of these clearly are material, it seems likely that there are also internal processes that lead to change, at a slower and more imperceptible level.  The Crypto-Jewish ethnography, in which individuals were interviewed and observed over an extended period of time within which many of them underwent changes in self-perception and identity construction, which in turn were mirrored by subtle changes in structure, suggested a possible location for agency.   Agency, on the level of emphasizing and privileging of particular structural elements, was necessary in order to explain the processes observed and the way they were articulated by living subjects. This process was identified as ‘jonglerie’, in which individuals, while not conscious of underlying structure, are able to effect change in underlying structure through the selection and reselection of different elements of their cultural repertoire. This concept includes a very strong notion of fluidity and thus the metaphor of the juggler seems the most appropriate.

A rather simplified but helpful example of this process is found in a Friday night candle lighting ritual used by some Crypto-Jews.  This ritual was composed of three elements: the day of the week, the candles and the Rosary.  The ritual was performed on Friday night, which was associated by many individuals with their Jewish heritage.  It was composed of lighting the candles while reciting the Rosary.  Individuals interviewed over a number of years had various self-identities, some were moving towards normative Judaism while others had a Christian identity with a focus on their Jewish heritage.  Those individuals with the Jewish identity interpreted the ritual as being a strong statement of Judaism, with the Rosary there purely as a defensive mechanism to protect them from the external recognition of the Jewishness of their practice.  The structure articulated strongly emphasized the oppositional aspect, seeing the Christian aspects as negative or meaningless.  Other individuals performing the same ritual, while seeing the time and practice of candle lighting as remnants of Jewish heritage and perceiving them as positive in that respect, saw the Rosary as the operative symbol, reflecting the key aspect of their identity.  These individuals emphasized the Christian aspect of structure and also had a degree of mediation, recognizing the transformational/conversion aspect of their identity as positive.  This example also reflects the role of bricolage in the transformational process; unconsciously bringing together elements from a wide range of contexts provides the material with which the conscious or semi-conscious process of ‘jonglerie’ can work. 
The first section of this paper presented a somewhat brief discussion of phenomenology and reductionism.  We suggested that the phenomenological position was largely self contradictory, arguing both against the external imposition of theory and utilizing an external theoretical stance; even where some scholars in religious studies move outside of phenomenology, their arguments for the unique basis of religion are merely the attempt to create a false disciplinary identity rather than based on convincing arguments of religious uniqueness. With regard to the reductionist stance, with which we are largely sympathetic,  it is argued that due to the basis of the argument in abstraction at times theorists seem to suggest that theory is independent of ethnographic ‘reality’ or evidence.  This stance largely ignores the fact that theoretical positions are themselves culturally and ideologically contextualized and that this contextualization must be disturbed by a process of dialectical engagement.  It is thus suggested that any theoretical stance must be regarded as provisional and set in dialectical relationship with ethnographic research. 
The second half of the paper attempted to situate this position in relation to a specific dialectical development of theory in relation to varying ethnographic contexts.  The discussion illustrates that although the theoretical position with which the discussion culminates is related genetically to the original orthodox structuralist stance, in many respects it has been transformed.  Although some of these transformations are shaped by wider theoretical discussions, it was primarily the issues raised by the ethnography that led to the reassessment of the theoretical position.  The discussion here illustrates a methodology for theoretical/ethnographic practice that includes the following elements: initial conscious articulation of theory; engagement with and analysis of ethnography on the basis of that theoretical stance; reassessment of the theory based on the ethnography; engagement with and analysis of other ethnographic contexts via the revised theoretical stance; reassessment  of the theory; reassessment of the earlier analyses on the basis of the new theoretical position.  A key feature of this development and an essential aspect of the argument developed here, is that in spite of the revaluation and development of the theory, the new position retains the same level of abstraction as did the earlier version. It does not become a particular theory that can only be applied in a particular context – particular theories are not theories at all, they are merely forms of description masquerading as analysis.



<1> This is seen specifically in the work of Eliade. M. Patterns in Comparative Religion (Cleveland: Ohio World, 1965) and van der Leeuw, G. Religion in Essence and Manifestation (Princeton: Princeton University Press, [1933] 1986) and implicitly in the work on the history of religion of Bleeker, C.J. The Sacred Bridge: researches into the nature and structure of religion (Leiden: Brill, 1975) who sees religion as having a single (and therefore separable) history – which demands an origin and basis separate from any particular culture.

<2> Although the view that religion has a single history is not found in many phenomenological analyses, it does provide the only basis for the third element of the phenomenological approach, that is, the eidetic vision, the typological or morphological aspect of the approach.

<3> See for example Radcliffe-Brown, A.R. ‘Religion and Society’ in Journal of the Royal Anthropological Institute vol. 75 (1945), pp. 33-43.

<4> One notable exception that arises from this scholarly tradition is found in the work of Ugo Bianchi - see for example The History of Religion (Brill: Leiden, 1975) - in which he argues against the notion of a single history of religion in favor of histories of religion as well as arguing for a dialectical approach to the defining of the object, which was a key influence on the approach suggested here.  

<5> Segal, R.A. ‘Reductionism in the Study of Religion’ in T.A. Idinopulos and E.A. Yonan (eds) Religion and Reductionism (Leiden: Brill, 1994), pp. 4-14; Strenski, I. ‘Reduction without tears’ in T.A. Idinopulos and E.A. Yonan (eds) Religion and Reductionism (Leiden: Brill, 1994), pp. 95-107.

<6> Structural Anthropology Vol 1 (London: Basic Books, 1963); Structural Anthropology Vol 2 (London Basic Books, 1963); The Savage Mind (London: Weidenfeld & Nicholson, 1966); The Raw and The Cooked(London: Jonathan Cape, 1970); From Honey to Ashes (London: Cape, 1973); The Origin of Table Manners (London: Cape, 1978); The Naked Man (London: Cape, 1981).

<7> The Logic of Incest (Sheffield: Sheffield Academic Press, 1995).

<8> A somewhat abstracted version of this methodology is presented by Lévi-Strauss in ‘The Structuralist Analysis of Myth’ in Structural Anthropology Volume One, pp. 206-31.  An abbreviated version of my own methodology is found in Kunin S. D. (2004) We Think What We Eat(London: T&T Clark, 2004), pp. 27-8.

<9> God’s Place in the World (London: Cassell, 1998); We Think What We Eat.

<10> I present a very brief version of this analysis, a more detailed version that also critiques the work of Mary Douglas is found in Kunin, We Think What We Eat , pp. 29-103.

<11> The biblical text permits the eating of locusts, which clearly do not have feathers,  but it specifically also forbids the eating of bats.

<12> See, for example, Douglas, Mary, Implicit Meanings (London: Routledge & Kegan Paul, 1975), pp. 249-275.

<13> The most detailed and complete discussion of this text is found in Kunin, We Think What We Eat, pp. 104-146.  This also includes a discussion of the use of these mythemes in the Book of Mormon and the New Testament.

<14> A brief discussion of the New Testament material is found in Kunin, We Think What We Eat, pp.133-138.  The LDS material is discussed in Kunin, S.D. ‘The Death/Rebirth Mytheme in the Book of Mormon’, in D. Davies,Mormon Identities in Transition (London:Cassell, 1996), pp. 192-203 and ‘The Allegory of the Olive Tree: A Case Study in (neo) Structuralist Analysis’ in Religion vol. 33, no. 2 (2003), pp. 105-125.  This final article also examines the use of this allegory in the biblical and inter-testamental texts.

<15> Although the biblical/rabbinic system does allow conversion, it has a strong model of rebirth with the convert seen as being a different person (genealogically) after the completion of the process.  For a more detailed discussion of this issue, see Kunin, We Think What We Eat, pp. 234-237.

<16> This reanalysis is found in Kunin, S.D. ‘Neo-Structuralism and the Contestation of Sacred Space in Biblical Israel’ in Temenos vol. 41, no. 2, (2005) pp. 203-224.

<17> Kunin, S.D. (2001) ‘Juggling Identities Among the Crypto-Jews of the American Southwest’, in Religion 31 (2001), pp. 41-61.

<18> The most convincing approach to agency is found in the work of Pierre Bourdieu. See, for example, Outline of a Theory of Practice (Cambridge: Cambridge University Press, 1977).  His discussion of the interrelationship between habitus and improvisation has particular relevance for the development of the role of agency in the theoretical/ ethnographic material outlined here.



Seth Kunin 2006