John J. McGraw.com-MA Thesis

Introduction

Cognitive anthropology remains a young field. Though it grew out of trends in anthropology present since the 19^th century, the specific questions and methodologies of cognitive anthropology did not distinguish themselves until the late 1950s. This new turn in anthropology grew out of a larger movement, the so-called “cognitive revolution” that spawned the multidisciplinary field known as cognitive science. Cognitive science is a blend of cognitive psychology, computer science, linguistics, and the neurosciences. The central focus of these overlapping fields is the investigation of cognition. Cognition relates to all things mental: perception, reasoning, inference, memory, attention, and the organization of knowledge. In other words, the processes by which the human nervous system interacts with and makes sense of itself within its social and ecological environment. Cognitive science has made important contributions to economics (North 2005), linguistics (Lakoff & Johnson 1999; Fauconnier & Turner 2002), mathematics (Lakoff & Nuñez 2001), and anthropology (D’Andrade 1995). Many of the topics and approaches of cognitive science have become central to cognitive anthropology.

One of the richest veins of thought within cognitive anthropology is that of schema theory. Borrowed from cognitive psychology, anthropologists developed sophisticated notions of schemas. Schemas can be thought of as compressions of the salient characteristics of experience. As will be discussed at length later on, schemas are the underlying language of the mind which enable complex patterns of representations like culture to emerge. Many cognitive anthropologists collaborated to work out how schemas might be thought of vis-à-vis anthropology and how larger representations known as cultural models could yield powerful new tools with which to understand the transmission and retention of cultural knowledge. Based on these ideas, cognitive anthropologists wrote about cultural models in relation to gender types (Holland & Skinner 1987), problem solving (White 1987), notions of marriage (Quinn 1987), emotions (Lakoff & Kövecses 1987), and parenthood (Harkness, Super & Keefer 1992), to name some prominent analyses. But for unclear reasons religion largely escaped the attentions of cognitive anthropologists until the last few years.

Like virtually every other aspect of human existence, we cannot have religion without cognition. Better understanding the cognitive foundations of religion can provide anthropologists with important new insights into the nature of religion and culture. Unfortunately, the noteworthy gap in the schema and cultural model literature left some to wonder how these tools might be used to better understand religion. Since the early 1990s a handful of cognitive anthropologists have developed a field entitled the cognitive science of religion (Scott Atran; Pascal Boyer; Dan Sperber; Harvey Whitehouse; et al.). To a man, these theorists have focused upon the selective transmission of religious representations. In the same notable way that schema theory overlooked religion, these cognitive scientists of religion have overlooked schema theory. By focusing too much on the transmission of religious representations, these thinkers have yet to develop sophisticated theories about the organization and retention of religion in individual minds, dimensions which schema theory helps to clarify. By ignoring the role of schema theory theorists they oversimplified ideas about transmission. For transmission, by its very nature, means nothing without reception. What has been lacking most in their framework is an appreciation for ecological approaches to cognition, the dynamic interplay of transmission and reception that occurs within a “socio-cognitive” system (Malley 1995; Malley 1997; Sørensen 2004). By better understanding the ecology of ideas, that unique constellation of beliefs and other representations by individuals who develop within particular historical and social environments, the cognitive science of religion will be able to proceed with a more sophisticated notion of mind. Schema theory provides a well-developed structure for understanding this important process by highlighting the particular structure of representations. This work will attempt to understand how schema theory and the cognitive science of religion might be brought together for a more comprehensive study of religion by developing an ecological approach to mental representations.

Representationalism

The ideas that follow rely on a framework of representationalism. Representationalism derives from the philosophy that mental phenomena can be largely accounted for by the nature of representations (Clark 2001:175). In short, the content of phenomenal experience cannot be disentangled from its representations. Pain, pleasure, emotions: no matter how inchoate the feelings, they gain much (but probably not all) of their meaningful experience from the various mental representations held about them; to a great extent these feelings are representations. This framework necessarily makes the individual an eminently cultural being because the representations utilized to make sense of experience come from the individual’s enculturation and his dynamic interaction with public and private cultural representations. Though much of what follows seeks a better understanding of the properties and constraints of internal representations, in fact, representationalism involves external representations, material anchors, and more generally ideas about distributed cognition outlined by Edwin Hutchins (1995, 2005). In this sense, representationalism thoroughly conjoins the inside world and the outside world, the mind and the body, the subject and object, the individual and society; all are unified in the traffic of representations (cf. Peirce 1982), by the “ecology of mind” (Bateson 1972). While the unit of analysis for psychological approaches must be the mind, a new definition of the mind is in order. The idea of a self-contained mind—a brain in a vat—can no longer be taken seriously when trying to account for the stochastic and dynamic development of the human nervous system which always occurs within natural and social ecologies. The individual brain, the mind, and the larger world in which these develop must be analyzed as intermeshed processes.

An ambitious attempt to reframe the mind within a larger system of representations was undertaken by Dan Sperber. In Explaining Culture (1996), Sperber employed the notion of an epidemiology of representations as a naturalistic approach to interpreting culture (1985, 1996). Sperber provides an interesting definition of culture as “ecological patterns of psychological phenomena” (1985:76). In the same way that epidemiology is used to study the spread of disease, an epidemiology of representations aims to understand the distribution of cultural representations. In Sperber’s words, “to explain culture is to answer the following question: why are some representations more successful in a human population, more contagious, more ‘catching’ than others?” (Sperber 1985:74). The idea of culture as a pattern of representations eschews simpler ideas about culture as monolithic and grounds culture in a naturalistic framework. According to Sperber, understanding the process of the differential transmission of cultural representations is to understand culture itself.

Epidemiology provides a useful metaphor for Sperber’s ideas but, as he is quick to point out, it can lead to confusions as well. Foremost, representations are not like viruses in any simple sense (1985:75). Viruses replicate in a reliable fashion with rare mutations. Cultural representations, in contrast, are necessarily transformed each time they are taken up by an individual subject through the process of semantic construction. Sperber does not assert a simple dualism in which some objective idea is transmitted to a vulnerable subject in the way, for instance, that a virus invades a host. Rather, he proposes a susceptibility to certain ‘catching’ representations based upon their inferential salience and hence their greater intuitive appeal (1985:82-83). Because people everywhere possess similar neural architecture, the kinds of representations that one person finds salient usually overlaps with what another person finds salient, though the degree of salience relates to the representational domain and its susceptibility to cultural shaping. Sperber talks about the variety of representations available:

…each of us knows by personal experience that some representations, say Gödel’s proof, are very hard to comprehend, however much we would like to; some representations, say a figure of twenty digits, though not hard to comprehend, are hard to remember; some deeply personal representations are hard or even impossible to convey without loss and distortion; on the other hand there are some representations, say the story of Little Red Riding Hood or a popular tune, which we cannot help remembering, even though we might wish to forget them. (Sperber 1985:80)

While it seems obvious to us that remembering a story is much easier than remembering a sequence of numbers, a story usually contains far more information in it than a sequence of numbers. But something about our evolved neural architecture leads us to remember stories more easily than strings of numbers. Part of this phenomenon derives from the fact that representations are underdetermined and underspecified. Human understanding comes from a process in which a maximum amount of meaning is inferred from a minimal amount of information. In making sense of a narrative, we employ a great deal of “backstage cognition” (Fauconnier 2002). In other words, we employ a number of schemas, scripts, and inferences in order to understand a narrative. We never approach a story as a blank slate. In understanding even a short narrative, we utilize massive amounts of prior information. All of this background information serves to constrain our interpretations. In this sense, all our past representations go into the construction of meaning in the present. By interpreting new narratives with the schemas that we are already familiar with, we become capable of constructing new meaning with great rapidity.

The most critical representations for cultural knowledge are schemas. Sperber believes that schemas relate to the domains of thought that we naturally employ:

I assume that we have an innate disposition to develop concepts according to certain schemas. We have different schemas for different domains: our concepts of living kinds tend to be taxonomic; our concepts of artifacts tend to be characterized in terms of function; our concepts of color tend to be centered on focal hues; etc. Concepts which conform to these schema are easily internalized and remembered. Let us call them basic concepts. (Sperber 1985:82)

Based on this preliminary treatment, we can see that to properly understand an epidemiology of cultural representations, we need to clarify what schemas are, how they allow us to infer the maximum amount of meaning with the minimal amount of information, and what processes lead us to prefer some schemas over others and thus to more easily internalize these preferred schemas. By making sense of these issues, we can begin to understand religious cognition.

Part I. Cognitive Anthropology: Where Mind Meets Culture

To understand why someone acts the way they do it is not enough to know the discourses, objects, and events to which they have been exposed; we need to know the psychic structures that assimilate those things and render them a basis for meaningful action.

—D’Andrade & Strauss, Human Motives and Cultural Models, p. 7

In the 1969 volume entitled Cognitive Anthropology, Stephen Tyler defined this new discipline as a study of the mental organization of culture. In Tyler’s words: “Cultures…are not material phenomena; they are cognitive organizations of material phenomena” (3). The cognitivist approach to anthropology set out by Tyler and others responds to the (economic) materialist slant as typified by Marvin Harris (1968). Anthropologists interested in language and ideas began to assemble a cognitivist account of culture. In the same way that the “black box” approach to the mind had been perpetrated by psychological behaviorism, mid-20^th century anthropology had sidelined a number of psychological components of culture in favor of political and economic approaches to culture. As Sherry Ortner noted in her comprehensive review essay, “Theory in Anthropology Since the Sixties:” “Unfortunately, anthropologists have generally found that actors with too much psychological plumbing are hard to handle methodologically” (1984:151). But a field like anthropology could not persist long without a thoughtful approach towards the human mind. In a later survey of the field, the cognitive anthropologist Roy D’Andrade rendered a definition for the discipline: “Cognitive anthropology investigates cultural knowledge, knowledge which is embedded in words, in stories, in artifacts, and which is learned from and shared with other humans” (1995:xiv). Though the early development of the field revolved around linguistic methods and concerns, a fact that Tyler’s volume makes clear, cognitive anthropology soon expanded into broader psychological questions, which D’Andrade notes in “The Cultural Part of Cognition.” In this 1981 article, D’Andrade wrote that “the cultural part of cognitive anthropology is in its concern with the way in which cultural content ‘interfaces’ with psychological processes” (182). Anthropology deserves multiple organizing frameworks for its layered subject. While cognitivists study mental structures and semantic concerns, materialists seek clarification of the political and economic infrastructure; both fields typically affirm the importance of the other’s perspective even if they embrace their own preferred methods for studying culture. Cognitive anthropology, then, while not denying the importance of the particular materialist foundations for culture, aims to understand how the human mind apprehends the culturally-mediated world.

Early Cognitive Anthropology

After the first generations of anthropologists completed their groundbreaking work in the early twentieth century, a number of later anthropologists returned to their field sites to highlight other aspects of the cultures in question. It became clear that no matter how excellent an ethnographer may be, his focus on certain aspects of culture—to the exclusion of others—led to a plethora of potential ethnographies; in short, myriad potential “cultures.” Repeated challenges to the founders and their ethnographies led to deep questions about methodology (Colby 1996:211). Appealing to decidedly more scientific methods, anthropology began a new campaign.

Linguistics, which had developed a highly specific terminology and refined set of conceptual tools, served as an appealing model for those social scientists who wanted to transform anthropology into a more scientific study of culture (Colby 1996). The ascent of linguistics during the 1950s tied in very closely with anthropology’s pursuits. In Cognitive Anthropology, Tyler emphasizes the complex semantic systems implicit in the use of language (1969:11). Linguistic anthropologists (and subsequently, cognitive anthropologists) thus sought to understand through the structures of a native’s language what meaning systems lie behind the use of words. By charting various taxonomies, paradigms, and branches (related though divergent methods of classifying the meanings inherent in language use), anthropologists began to delineate semantic structures (ibid). Floyd Lounsbury and Ward Goodenough’s systematic approach to kin term systems led to componential analysis, later to be known as feature analysis (D’Andrade 1995:21). By charting and classifying term relationships, Lounsbury and Goodenough were able to identify and analyze the “units of ideas” that coalesced into cultural understandings, at least insofar as kin relations were concerned. An emphasis on ethnographic semantics dominated the field for some time but other aspects of cognition—especially perception, memory, and reasoning—led to different approaches to the study of culture.

Cognitive anthropologists sought to better understand the way in which different cultures organized their specialized knowledge (e.g. botany, biology, and other technical fields). They coined this new field “ethnoscience.” Though an interest in the organization of mental life originated in Boasian anthropology, it was during the late 1950s and the 1960s that the concerns of ethnoscience took precedence. As early as 1954, Harold Conklin wrote about the ethnobotanical classifications of the Hanunóo, a group of people in the Philippines, who utilized an especially sophisticated system of classification for plants (Conklin 1954). He continued his research with a study of color categories among the Hanunóo (Conklin 1955). Following upon Conklin’s study of color categories, Brent Berlin and Paul Kay began an ambitious multicultural study of basic color terms (1969). Using an array of 320 color chips, Berlin and Kay sought to discover the differences (if any) in the perception and subsequent organization of color among various cultures. The findings were quite complex but generally supported the thesis that the elaboration of basic color terms followed a predictable sequence that depended on the most salient colors in a culture’s environmental setting (Boster 1988:71). Vision appeared to be a developmentally shaped domain of experience that possessed fairly tight physiological constraints.

By the mid-1970s ethnoscientific research had yielded enough data for anthropologists to be able to determine relatively universal features and parameters of such organization. For instance, Cecil Brown et al. noted that: “Non-biological taxonomies, like biological ones, rarely, if ever, exceed a maximum hierarchic depth of five levels” (1976:83). While knowledge of individual artifacts may become quite complex, it is generally true—no doubt because of the limitations of memory and perception—that taxonomies of any sort tend to be about five levels deep. Such assessments helped to illustrate the interaction of the human nervous system with its environment in the process of cultural understanding.

One of the pioneers of ethnoscience, Eleanor Rosch, aimed to better understand human categorization in her fieldwork with the Dugum Dani of New Guinea (MacLaury 1991). Inspired by Berlin and Kay’s discussion of basic color terms, Rosch developed an interest in generics, those things most likely to show up at the basic level of any category not only as terms but as real objects. Defined by their most recognizable types, categories could be indicated without reference to an amorphous, floating collection of terms and attributes. In looking for the most readily identifiable examples for qualities, colors, and objects, she came up with the notion of a prototype: “By prototypes of categories we have generally meant the clearest cases of membership defined operationally by people’s judgments of goodness of membership in the category” (Rosch 1978:35). From prototypes we learn that meaning often derives from the particulars of a representative object. The prototype theory brings to mind Plato’s notion of Ideal Forms, but whereas Plato had to assert a transcendental realm of Ideas, Rosch and like-minded ethnoscientists knew that their ideal forms—in prototypes—were aspects of a human nervous system shaped by its environment. The mind continually engages in a process of pattern-matching, comparing the new to the exemplary cases in mind in a traffic of salience based on cues between memory and perception. Rosch reasoned that the nervous system sought to acquire the maximum amount of information with the least amount of effort (1978:28). It is far more efficient for the mind to typify a category with the best example of the type: reasoning from the part to the whole, as it were. The exceptions, because they are less likely to be encountered and rendered useful or salient, lie at the edge of categories.

In a study of the perception of birds in the San Francisco area, James Boster discovered that the prototype for bird represented an average of the birds seen in the environment, weighted especially around those most numerous within the local habitat (1988). The prototype for the category of bird, then, came from the most commonly perceived birds in one’s environment. To be concrete, robins and sparrows embodied “birdiness” for these people. One can easily imagine, though, how these prototypes for Bay Area citizens would seem like poor imitations to Costa Ricans whose notion of size, color, and overall birdy splendor would make a Bay Area sparrow a very poor avian substitute. With her elucidation of prototypes, Rosch and colleagues helped to uncover the mental elements and cognitive structures for cultural meaning. An outsider to the discipline might well look at all these findings from ethnoscience, scratch his head, and offer up a “so what?” What end does this research serve?

Semantic analysis and ethnoscience indicated that structures existed in the human mind and in culture that helped to organize information. In its essence, these inquiries led to a new epistemological quest within anthropology and cognitive science. D’Andrade hints at the most basic assumption of cognitive science: “It is as if the human cognitive system were a structure seeking device” (1995:120). The study of cultural classification systems led to a greater desire to understand how cognitive structuring organized all cultural knowledge and, reciprocally, how culture structured cognition. One thing Rosch and colleagues realized is that at the highest level of classification, taxonomies formed around the most salient objects in one’s environment. Prototypical objects help define categories and provide people with basic kinds of things, anchoring ontologies. The basic objects, defined by their most relevant and distinctive attributes, indicate a cultural synthesis of the philosophical debate between a priori knowledge and a posteriori knowledge. While the processes through which the human nervous system finds structure in experience emerges through a priori aspects of cognitive architecture, their ultimate products are definitively shaped by the a posteriori encounters between the individual and her physical and cultural environment. In short, ethnoscience helped to answer previous questions about the universal properties of the human mind as well as highlight the cultural components that shape category and object perception.

By the mid-1970s, as typified in the work of Rosch and colleagues, anthropological interest in the way the mind perceives and organizes experience had shifted from the relatively low-level and specific to the much higher-level and generic. The linguist Charles Fillmore illustrated how even single words (e.g. “bachelor”) rely on complex cultural knowledge that went well beyond simple prototypes. His idea of folk models sought to uncover the cognitive scaffolding behind language (1975). While prototypes seemed a fruitful subject for understanding some units of cultural understanding, a quest for more comprehensive knowledge structures commenced. The quest for this wider purview led cognitive anthropology to refine schema theory.

Schemas and Models

Cultural knowledge is best thought of as a distributed system of models.

—Shore, Culture in Mind, p. 312

In schemas, the staggering complexities of a procedure, an understanding, or a behavior become simplified, caricatured, or otherwise altered until only the most salient aspects remain. David Rumelhart referred to schemas as the “building blocks of cognition” (1980). Schematization is the manner by which the limited capabilities of the human nervous system may obtain a functional representation of reality for the purposes of meaning-making and efficient action.

In his review of schema theory, Ronald Casson defines a schema as a conceptual abstraction that mediates between sensory stimuli and behavioral responses (1983:430). Cognitive psychologist George Mandler notes that: “Schemas are built up in the course of interaction with the environment…schemas are abstract representations of environmental regularities” (1984:55-56). While this description might evoke the concept of the prototype described above, in fact D’Andrade notes crucial differences: schemas are sufficiently open and flexible to allow lots of possible organizations of experience (1995:123). For instance, when one thinks of hunting, the schematic representation of hunting might include a number of ‘slots’ waiting to be filled: the type of men stalking, the types of animals pursued, the varieties of weapons used, the methods for attracting the prey, the rituals for slaughtering any captured prey, and more. In spite of these slots being open, though, the gestalt of these different elements combining into the complex we refer to as “hunting” becomes clear. One might imagine Yanomamo hunting monkeys with blowguns, Virginians hunting deer with rifles, or Japanese hunting whales with explosive harpoons; but with any of these particular scenarios we can casually use the word “hunting” and have a good sense of what that means and entails. In contrast to the schema of hunting, the prototype of hunting may be quite specified and carefully elaborated according to the experience of the subject. The prototype of hunting for a North American likely involves scoped rifles and alternately camouflaged, fluorescent orange bedecked stalkers in possession of the latest devices and paraphernalia. Ronald Langacker defines the prototype as a highly typical instantiation (1987). The schema, in contrast, emerges as a vast network of linked potentials, potentials to be clarified and “filled out” as the data amasses.

Cognitive psychologists Nancy Stein and Thomas Trabasso clarify some basic notions about schemas (var. schemata):

Schemata are composed of generic or abstract knowledge; used to guide encoding, organization, and retrieval of information; 2) Schemata reflect prototypical properties of experiences encountered by an individual, integrated over many instances; 3) A schema may be formed and used without the individual's conscious awareness; 4) Although schemata are assumed to reflect an individual's experience, they are also assumed to be shared across individuals; 5) Once formed, schemata are thought to be relatively stable over time; 6) We know more about how schemata are used than we do about how they are acquired. (1982:212)

Schema theory developed over a long period of time and has been innovated and revised numerous times by philosophers, cognitive psychologists, learning theorists, and anthropologists. Though the idea of a schema is relatively simple—namely a mediating representation between perception and action—no single comprehensive theory has been put forth. In the conclusion to his review of the literature, Casson calls schema theory (as of 1983) “overly powerful and too general” (455). While this may be the case for more simplistic schema theories, increasingly refined ones have been proffered by cognitive anthropologists in the 1980s and 1990s.

By the mid-1980s cognitive anthropologists had begun to systematically refine schema theory for their own purposes. One of the first transformations involved a transition from schema theory to the study of cultural schemas and cultural models. Cultural schemas are those schemas shared by numerous members of a culture that allow them some conventional understanding of a given phenomenon (Quinn 1997a:140). The notion of a cultural schema being shared distinguishes it from personal schemas which may be primarily idiosyncratic. For instance, the idea of Neil Armstrong setting foot on the moon and reciting his famous lines would certainly qualify as a cultural schema whereas Proust’s memories about madeleines were idiosyncratic schemas, at least until he published Swann’s Way. Naomi Quinn and Dorothy Holland expand such notions into a cultural model: “Cultural models are presupposed, taken-for-granted models of the world that are widely shared (although not necessarily to the exclusion of other, alternative models) by the members of a society and that play an enormous role in their understanding of the world and their behavior in it” (1987:4). D’Andrade holds that many cultural models are not schemas even though they may be composed of multiple schemas. If anthropologists are to accept Mandler’s stipulation that a schema is “a bounded, distinct, and unitary representation” (1984:55-56), then by necessity many cultural models cannot be classified as schemas. Complex cultural models tax the capacities of short term memory and thus cannot fulfill Mandler’s criterion of a schema (D’Andrade 1995:152). But as the definitions of schema, cultural schema, and cultural model begin to make clear, the concept of how a schema may be shared and how it becomes sufficiently complex to lose its identity as a “bounded, distinct, and unitary representation” breeds confusion. Strauss and Quinn develop this point:

Cultural models are in people’s heads, but a given cultural model need not be in everyone’s head. Since subgroups such as households, genders, regions, ethnic minorities, and historical cohorts crosscut one another, an individual may share schemas with as many different groups of other individuals as he or she shares a history of membership—at the same time sharing all of his or her schemas with no one else in any of these groups. This way of thinking about culture as differentially distributed cultural understandings, or schemas, seems to us to account for that cultural sharing which does occur without reifying culture as a bounded entity. (1994:293)

The idea of a cultural model might still remain vague from these theoretical definitions. In dealing with such abstractions, the presentation of some persuasive examples may clarify things.

A good ethnographic example of a cultural model appears in Quinn’s study of American marriage (1987). Quinn used extensive interview material from husbands and wives in 11 marriages. In a careful discourse analysis, Quinn studied the metaphors and key words to discover a series of propositions that characterize the notion of marriage. Among the proposition-schemas that supported the cultural model of marriage were: 1) marriage is enduring; 2) marriage is mutually beneficial; 3) marriage is unknown at the outset; 4) marriage is difficult; 5) marriage is effortful; and 6) marriage is risky (179). What Quinn’s depiction of the model of marriage elucidates is that, in contrast to a simpler psychological analysis, her model not only posits desire but illustrates, through these proposition-schemas, how desire gets organized cognitively and affectively. Even more interesting, these proposition-schemas show how desire may be generated by collective beliefs, values, and goals. A desire for marriage does not erupt from a unitary source but originates in culturally mediated understandings of numerous schemas. In the same way that Fillmore’s analysis of the folk model behind the word “bachelor” illustrated a world of associations and understandings, any cultural model entails elaborate schematic structures which may or may not be internalized by the individual.

With these ideas of cultural models come a host of cultural consequences and properties. In A Cognitive Theory of Cultural Meaning, Strauss and Quinn discuss properties of culture as envisaged in their cognitive theory. They distinguish between those forces that lead to variation and change in culture (centrifugal) and those that lead to sharing and reproduction of cultural forms (centripetal) (1997:118-199). Two centripetal properties of schemas that they develop are the historical durability of schemas and their thematicity (1997:111-120). Historical durability refers simply to the properties many schemas have for reproduction from one generation to the next. Thematicity refers to the property some schemas possess for spreading across numerous domains of experience and in a wide variety of contexts.

While it is clear from the work of these cognitive anthropologists that schema theory has come a long way since its early days in other social sciences, a “ failure to launch” unfortunately persisted. No matter how these ingenious researchers integrated schema theory into the study of culture, the field remained limited to a small pool of scholars and did not get taken up onto the front stage of anthropology. A quest to refine and elaborate the theory ensued in hopes that a more elegant set of ideas might result in parsimonious analyses attractive to a greater number of anthropologists.

Culture in Mind

Culture, in our formulation…is not some free-floating abstract entity; rather, it consists of regular occurrences in the humanly created world, in the schemas people share as a result of these, and in the interactions between these schemas and this world.

-Strauss & Quinn, A Cognitive Theory of Cultural Meaning, p. 7

Nearly every anthropologist who has worked with schema theory has noted that, as currently formulated, the theory is too general (e.g. Casson 1983:455; Shore 1996:45). In adapting schema theory to account for cultural understandings, Casson attempted to distinguish various types of these structures. Of schemas, he wrote, “some are universal, some idiosyncratic, and some cultural” (1983:440-441). Casson also suggested that schemas worked hierarchically, “in which schemata at the higher levels represent the most general concepts, and schemata at successively lower levels represent more and more specific concepts” (1983:437). He went on to postulate three distinct kinds of schemas: object schemas, orientation schemas, and event schemas. Object schemas relate to the taxonomic classifications of objects discussed earlier in the section on ethnoscience. Behind any given object is a whole series of implicit relationships to like objects and to unlike objects. Orientation schemas relate to the layout of buildings, cities, and other architecture; they are sometimes referred to as cognitive maps. Event schemas relate to a whole series of actions and events such as behavior in a bus, in a restaurant, with one’s girlfriend’s parents, etc. A linked set of schemas sometimes goes by the name ‘script.’ Casson’s development of schema theory aimed to create sensible distinctions among schemas and convey them in a practical way so that anthropologists might take schema theory on as yet another tool with which to study culture. Unfortunately, Casson’s elegant organization of schema theory did not receive much of an audience within the anthropological community. It would take many subsequent iterations of schema theories for these ideas to gain momentum.

D’Andrade was the next to convey an organized approach for schema theory. He stipulated three hierarchical levels of “goal” schemas: the top level (also called master motives), the middle level, and the lower level (1995:232). Master motives are the unifying, directive themes of many behaviors. Examples of master motives include such things as love, fear, and success. These are obviously quite expansive categories that cut across numerous domains of behavior. Middle level schemas include one’s spouse, one’s children, one’s career, one’s hobbies and avocations. While these might seem to be the highest level of priorities, in fact they derive their significance from the underlying master motives aforementioned. The lower level schemas include all the important, day-to-day behaviors and understandings that allow one to support higher level schemas. This level would include such things as taking the children to school, writing company reports, and buying gifts for one’s spouse. Again, while all such schemas are important, they derive their motivational force from the higher level schemas. While D’Andrade’s formulation of schemas and their goal hierarchies aimed to illustrate the link between schemas and motivation, it failed to distinguish between the numerous kinds of schemas.

The most elaborate of the cultural schema theories was put forward by Bradd Shore in Culture in Mind (1996). While he applauds the attempts of earlier cognitive anthropologists to make use of schema theory, he feels that their undifferentiated conceptions retarded further development: “So long as the notion of cultural models remains undertheorized and vague…its potential to provide a long overdue bridge between anthropology and cognitive science will be seriously hampered” (Shore 1996:45). Shore stipulates a modular account of schemas and models which expresses both the private and public sides of cultural knowledge.

Shore distinguishes between personal mental models, conventional mental models, instituted models, and foundational schemas. As can be seen, Shore refers to the more personal schemas as models and the most public ones as schemas. He does this in an attempt to clarify the shifting definitions of schemas and models. Shore also distinguishes between mental models as relatively private and instituted models as public: “Instituted models are social institutions” (1996:51). As one moves up the ladder from Shore’s personal mental models to the foundational schemas, conservatism (or to use Strauss and Quinn’s term, centripetalism for historical durability and thematicity) prevails. Within the scale of personal mental models the individual possesses great freedom in the creation and alteration of models, but by the time she interacts with foundational schemas she has practically no ability to alter or revise the schema. Greater and greater normativity, social pressure, and tradition compel the individual to pay heed to public models and schemas (Shore 1996:48).

Mental models originate in the person’s robust capacity for schema creation. As Shore puts it: “At the personal level, each of us is adept at constructing idiosyncratic models of experience on the fly, as a basic meaning-making strategy” (1996:46). The human mind abhors disorder and bestows orderliness to experience from instinct, it seems. Whereas other animals rely on fixed or relatively closed behavioral patterns, humans are great innovators in possession of remarkably open and flexible behavioral routines. Meaning-making serves as a profoundly functional tool in the symbolic environment of human culture. Ultimately, this symbolic environment interacts with the natural environment in a co-evolutionary system in which the symbolic world and the natural one shape each other (Deacon 1997).

However prolific personal mental models are, they exist only insofar as conventional mental models—the particular models conveyed in a time and place to the individual—involve the person with her social environment. Shore proffers an increase in complexity as the individual goes between personal mental models and conventional ones: “Their creation is more complex than that of personal models, since they have been externalized as shared institutions as well as internalized by individuals” (1996:47). As Shore moves to larger scales of analysis, the models he describes become more entrenched and overarching. The models and schemas exist as modules, nesting beneath and within increasingly public conventions.

Shore describes instituted models as “conventional, patterned public forms such as greetings, calendars, discourse genres, houses, public spaces, chants, conventional body postures, and even deliberately orchestrated aromas” (1996:51). This type of model enables the individual and her local culture to link up with an increasingly shared social sphere: “Instituted models are the external or public aspect of culture and represent common source domains by which individuals schematize conventional mental models” (1996:312). As can be seen here, as “source domains,” instituted models play a significant part in setting the parameters for the smaller scale mental models that each individual generates personally and interpersonally. So while individual freedom cannot be doubted, the framework within which the individual produces her mental models emerges from the jungle of conventional and instituted models. Instituted models provide foundations for social existence. As Shore puts it: “Instituted models always lead a double life, as part of an external social world and as products of intentional behavior” (1996:51). Instituted models seem to be the essence of a culture, the particular ways that it organizes and expresses its distinct way of life in word and behavior. The instituted models get taken up by the individual as her own conventional mental models.

At the highest level of cultural knowledge, Shore introduces the notion of foundational schemas: “A foundational schema is a high-level model of great generality and abstractness. Generally, foundational schemas are not dedicated to a single domain of social life but organize and underlie a large number of specific cultural models whose forms are roughly analogous to one another” (1996:117). In not being dedicated to a single domain, Shore’s foundational schemas evince the kind of thematicity that Strauss and Quinn postulated (1997:118-120). A “cultural value” may be a more colloquial translation of these large-scale schemas. For a number of Americans, then, (especially those in Western parts of the U.S.) “self-reliance” would be a foundational schema. In learning the numerous skills, perspectives, and versatility required of an “open territory,” the pioneering type that spread through the West needed to embody the value of self-reliance. The countless smaller level models necessary for life in a rugged environment derived from this foundational schema of self-reliance. Though discerning foundational schemas within some cultural context can be challenging (like considering macro-level processes in any discipline), they tend to capture the essence of a culture.

The cowboy as the symbol of self-reliant masculinity, for instance, would certainly express something crucial about American life in the Western United States. Thus, Americans continue to organize a great many things according to the mystique of this antiquated schema: they give their sports teams the unlikely name “Cowboys,” they drive impractical “Wranglers,” “Mustangs,” or “Broncos,” they name their streets “Rodeo Drive,” they wear long out-of-date “Cowboy hats” and “Cowboy boots” practical for horseback riding but not for SUV navigation, and they refer to their living accommodations as “homesteads” or “ranch houses.” Once this ecology has become prolific such that it includes a vast collection of symbols, values, artifacts, speech communities, and beliefs, then it has transformed into a larger system of representations that exist both within an individual nervous system and outside of it to form a network of meanings and conventions distributed across these various media. A great irony of this situation is that many contemporary Americans have come to embody the social conventions of an archaic figure—a 19^th century ranch hand—as an affirmation of their distinctiveness and self-reliance, never sensing the contradiction because they have come to internalize, and then externalize and reproduce, the convention.

Internalization of Models

A central concern of Culture in Mind involves demonstrating this irony: the individual’s ability to generate her own models at the same time that her creativity does little more than subtly rearrange and adjust the stock models of her culture. Shore refers to this process as the “twice-born” nature of meaning construction:

Instituted models are the public life of culture, empirically observable social institutions that are available as resources for a community. Mental models, by contrast, are cognitive representations of these instituted models but are not simply direct mental mappings of social institutions. The complex relationship between mental and instituted models defines what I have termed the ‘twice-born character’ of cultural forms. (1996:68)

In a dualist’s interpretation, what would be seen here is an artful plagiarism. The individual adopts and integrates the models of her community and quickly forgets they were ever anything but her own, there are “no new ideas under the sun,” to paraphrase Ecclesiastes. But from a more ecological point of view—one that eschews such simple and traditional dualisms—the neotenous birth of the human organism into culture (half-baked as it were) and then the long, slow simmering of the organism into the stew of culture makes such a plagiarism impossible. From the beginning, the person and all her models engage in an endless back-and-forth entanglement of social models and their personal reception and alteration.

Any synchronic “snapshot” of a culture would freeze in place a nearly infinite number of models and schemas. At such a discrete moment in time, the unit of culture—the individual—flashes into existence with the fusion of sperm and egg, a zygote results subsequently developing into an embryo which eventually erupts from the female body as a few pounds of human potential. During the prolonged postnatal development within the matrix of cultural influences, the human being finds itself adrift among a sea of complex representations which eventually help to shape its very consciousness. If such a creature may be called an “information processor,” then its processing is that of schematization. As Shore writes in “Twice Born, Once Conceived”: “the most basic work of culture would be to provide individuals in a community with a stock of common orientational models for constructing experience. Such models make possible a significant degree of ‘coordinated schematization’ and shared meaning construction of a common community” (1991:9). This pool of coordinated schemas allows for the distinct intersubjective understanding of individuals within a culture. Now, of course, such intersubjective understanding never proceeds without error and misunderstanding but it usually remains a more reliable process among members of the same culture than between representatives of different cultures. This higher fidelity derives from the shared stock of conventional and institutional models. Enculturation is the process of developing as an individual amid such models.

Shore’s subtle distinction of the “twice-born” nature of meaning construction answers those critics who parody schematization as a “fax model” of socialization (Strauss 1992:9). The fax model that Strauss characterizes would claim that culture exists as a set of schemas, models, or “programs” (to extrapolate from the misguided characterization) and that humans are automata which do nothing more than serve as blank slates for cultural inscription. The fax is transmitted and—voilà—one is programmed as a member of the culture. Creativity, change, and complexity in the fax model come from accident alone. But Strauss and Shore both see this fax model as a bastardization that critics of cognitive anthropology concoct as a straw man for their arguments. Understanding this entire system as an ecology of representations in which both the individual and his socio-cognitive environment engage in numerous levels of exchange breaks down simple notions of transmission and reception. The organic metaphor implied by an ecological approach serves to highlight the complex interaction between individual organism and environment. Even if the fax model parody does not stand up to the ecological and systemic approaches in which representations truly exchange, genuine challenges to the internalization of schemas remain.

Melford Spiro, in his paper “Collective Representations and Mental Representations in Religious Symbol Systems,” asks: “why are some cultural beliefs no more than clichés, while others are held with strong conviction and emotional intensity?” (1987). Related to this question, the study of culture must ask why people are moved and motivated in different ways by the very same models and schemas. No doubt the particulars of enculturation, the idiosyncratic natures of one’s parents, peers, and mentors, as well as the contingent social and historical conditions play a large role in this process. Spiro identifies five levels of cognitive salience related to internalization of cultural beliefs:

1) The actors learn about the doctrines; 2) They understand the doctrines, at least so far as traditional meanings conveyed by authoritative texts or experts are concerned; 3) They believe that the cultural beliefs are true, correct, or right; 4) The cultural beliefs are accepted by the actor and help guide behavior—they cease to be clichés; and 5) They not only guide but serve to instigate action by arousing strong affect. (1987:164)

According to Spiro’s model, each member of a culture registers all known cultural beliefs (schemas) in one of these five levels of cognitive salience. Charles Lindholm sees the differential internalization of varied models and schemas as the very nature of intracultural diversity: “within the cognitivist theoretical framework, variations in the degree to which cultural schemas are internalized and shared (for whatever reason) account for differences between people” (2001:259). Again, if one believes in the caricatured “fax model” of socialization, then the variation among people within a culture remains a great mystery. But the nuanced version that Shore and others put forth illustrates the developmental and social complexity of enculturation. Because we interact with a common stock of models during varied stages of development people may differentially interpret the importance of conventional models. Related to this, the sequence through which an individual gains exposure to different models alters her later receptivity to other models. In short, the contingencies of development and exposure provide some basic explanations for intracultural variability and the idiosyncratic adoption of cultural models. But perhaps more than these two factors, conflict, desire, and emotion play a substantial role in the internalization of cultural knowledge.

Psychodynamic Processes

No matter how the cognitive theorists affirm the complexity and variability within schema theory, it remains somewhat wooden and counterintuitive. One can’t help but see these symbolic systems as fit for automata rather than eating, breathing, copulating, excreting, dying beings. Cognition seems “cold.” What is lacking in these modular nests of models and schemas is that most elusive prey of psychological anthropology: the person. The richness of psychodynamic theories of culture, such as those put forth by Spiro and others, convey a struggling and conflicted agent within culture, someone who may be at odds with the very culture that spawned her. How can cognitive anthropologists “heat up” their accounts to elucidate the nature of emotion, desire, and conflict? Anthropologists have been in search of ways to marry the best psychodynamic concepts of the person with cognitive accounts (Westen 1991; Quinn 1997b). But an important step in this important project is the removal of entrenched obstacles.

A prejudice transmitted through 2500 years of Platonic philosophy would have thinkers believe that emotions exist opposite of reason. That which is cognitive, mental, and reasoned stands in contrast to the embodied, irrational, unpredictable nature of emotion. It is because of these very prejudices that cognition seems “cold.” It is not typically associated with the value of living, with the artful negotiation of personhood that each human being engages in on a daily basis. This idea about “dumb” emotions and “cold” reasons could not be more wrong. Even a good number of ancient Greeks knew that this Platonic miscarriage bore false witness. Epictetus addressed this in the aphorism: “Men are disturbed not by the things which happen, but by the opinions about the things” (14). According to Stoic principles, then, judgment shapes emotion. And judgment can only proceed within the nests of cultural models that the person has internalized.

In Upheavals of Thought, Martha Nussbaum finds a middle ground between those who believe emotions to be cognitive appraisals and those who believe emotions to be physiological upsurges mainly independent of culture and context. She postulates an account of emotions that marries the physiology of emotional arousal to the varied cognitive judgments and appraisals that elicit and then shape these upsurges. A long debate in emotions research, one originally set forth by Darwin, holds that some emotions are universal (Ekman 1993). While it would be sensible to imagine some universality in emotions in the same species, there’s no question that individual and cultural differences abound. Basing emotions in the joint interaction of physiological arousal and cognitive judgment supports the universal basis of emotion in the body according to the individual and cultural particularities shaped by one’s cultural models:

…taking up a cognitive/evaluative view makes it easy to see how society could affect the emotional repertory of its members. If we hold that beliefs about what is important and valuable play a central role in emotions, we can readily see how those beliefs can be powerfully shaped by social norms as well as by an individual history; and we can also see how changing social norms can change emotional life. (Nussbaum 2001:142)

Just as it would be odd to imagine a human being without emotions, so would it be odd to imagine human beings as all possessing the same emotional lives. One’s natural predispositions, enculturation, and personal experience all interact to shape billions of possible emotional beings. Such an account of emotion fits quite well with schema theory and bridges schema theory with psychodynamic accounts of the self. As Keith McNeal notes in “Affecting Experience: Toward a Biocultural Model of Human Emotions,” “To analyze emotions we therefore need to place changes in feeling-states within their dynamic environmental, developmental contexts, taking into consideration the culturally mediated cognitive processes of evaluation, labeling, and interpretation involved” (1999:244). The person cannot be envisioned as a simple biological creature, in possession of universal affective responses, but neither can she be thought of as a cognitive agent, coolly calculating through various schemas like a computer.

In fact, emotions and motivation are the crucial links between a cognitive account of culture—as expressed in schema theory—and a psychodynamic account of culture. For what psychodynamic accounts stress, the shaping of the individual by culture and developmental dynamics, answers the underemphasized account of the person in schema theory. McNeal expands upon this more complex, variegated notion of the person:

Variation in human emotionality can be seen as stemming (a) from the ontogenetic history which has shaped the individual (or group of individuals) to relate, interpret, and react to socioecological interactions in various ways, and (b) from the culturally mediated knowledge schemas which influence perception and cognition in particular ways. (1999:244-245)

Both accounts, the cognitive and the psychodynamic, ultimately mesh well because of the evaluative aspects of emotion. The little discussed but crucial aspect of a cognitive account of emotions comes from its appreciation for developmental processes. Nussbaum shows great appreciation for the psychology of object relations which helps to explain the seemingly unconscious aspects of emotional life. Important sources of the unconscious include those layers of one’s emotional life that were shaped in early stages of development (Nussbaum 2001:174-175). Such developmental stages of the emotions deeply influence one’s later emotional responses. Certainly part of the “heat” of one’s emotional existence derives from these inchoate, early experiences and associations. Through psychodynamic approaches to emotion—and to the self—the seemingly “cold” aspects of schema theory may gain the dynamism and capricious emotionality that imbues all aspects of human existence—both personal and social. Nussbaum writes, “in a deep sense all human emotions are in part about the past, and bear the traces of a history that is at once commonly human, socially constructed, and idiosyncratic” (Nussbaum 177).

Once this messier sense of the person comes through—as a psychodynamically shaped, developmentally formed human being—it becomes easier to understand the conflict and ambiguity that must go along with schema theory. Human beings are not automata busily sorting through a catalog of programmatic models but are psychodynamic creatures utilizing information selectively according to its relevance for their essential projects. To complicate the situation even more, the models help to inform that psychodynamically shaped person who she is so that the agent may be alternately understood according to which models she emphasizes at a given moment. The agent at church, who conceives herself as a spiritual being who originated and will return to spiritual realms, acts quite differently from the agent at the marketplace who competes for the scarce resources necessary to her family’s flourishing. Time, context, and various emotional stimuli help to elicit the variety of cognitively mediated emotional responses that people demonstrate.

Another aspect of this richer combination between psychodynamic understandings and cognitive ones derives from its ability to account for intrapsychic conflict, an aspect of the cultural models literature that some anthropologists believe has been undeveloped (Nuckolls 1996; 1998). Cognitive anthropologists have attempted to illustrate the way in which schema theory may lead to such internal conflicts. In “The Motivational Force of Self-Understanding,” Quinn illustrates how conflicting schemas of wifely duties, career advancement, and fairness can lead to uneasy balances in the psyches of American women. About one of her interviewees, Quinn writes: “Still she feels uncomfortable demanding that her professional advancement be taken as seriously as his [her husband’s]. As she tells it, this persistent ambivalence is founded not only on her understanding of her role as a traditional wife but also on her understanding of herself as a woman…With this claim about gender difference, Patsy reinforces an appeal to what is right with an appeal to what is natural” (119). Because potentially conflicting schemas have been internalized to different degrees, Patsy gains motivation to act a certain way even if it contravenes some of her other understandings and values. And in “Behind the Make Up,” McNeal shows how the cultural models available for American gay men necessarily lead to gender ambivalence and subsequent intrapsychic conflict, a conflict deriving from the classic “double-bind” discussed by Bateson (1972). So if critics lodge a complaint about the lack of intrapsychic conflict possible in cognitive theories, they have probably not accounted for differences in schema internalization nor looked at those fault zones where cultural models necessarily create double-binds and other painful dilemmas.

Criticisms and Responses

In reviewing Shore’s elegant, dynamic model of schema theory, Lindholm feels that: “It remains as yet unproved whether this proliferation of categories will prove useful, or whether, like the epicycles used to try to save the principles of Ptolemaic astronomy, they simply reflect a fundamental incoherence at the core of the theory, that is, the failure to postulate any integrating element—a “me” that is in charge of the ramifying schemas” (Lindholm 2001:258). Lindholm refers to the purely cognitivist approaches that conceive of people as information processors shuffling through their hierarchically organized schemas and models. In the “cold” cognitive model, Lindholm’s critique might cause some worry but in the newer models that fuse the best of psychodynamic theories of the person and cognitivist theories such criticism does not carry weight.

A more difficult problem for schema theory relates to what Hutchins has entitled the “transparency” of schemas (1980:12). He notes that once internalized, schemas become “what one sees with, but seldom what one sees” (ibid). As Strauss and Quinn write: “It is the typical transparency of cultural models that gives them the ability to constitute our reality—including our language” (1997:154). While this makes perfectly good sense and is considered by most theorists to be an expected property of internalized schemas, it also results in two rather difficult issues: that of methods for the discovery of schemas and that of the verifiability of inferred schemas. The examples from Quinn’s work demonstrates how careful linguistic analysis can yield a fairly reliable set of schemas from cultural informants. Nevertheless, this hardly resembles the hard-and-fast scientific methodologies that the cognitivists originally sought. But perhaps such linguistic analysis, prone to interpretation as it must be, still yields a more systematic approach to cultural knowledge than the employment of no system at all.

Another way that theorists attempt to account for the transparency of schemas relates to ideas developed by Rumelhart and colleagues concerning Parallel Distributed Processing, also known as connectionism (1986). According to this theory of cognition, a network of decision making nodes interact by dynamically adjusting the strength of their connections between one another resulting in an evolved cognitive system that lacks a central processing unit. Of course, in such a purely cognitive system, Lindholm’s critique would be severe. But while cognitive theorists such as Strauss, Quinn, D’Andrade, and Hutchins assert that connectionist networks likely operate in human cognition they do not claim that these are the only such cognitive structures at work in humans. D’Andrade firmly believes that both serial processing (that is, rule-based behavior) and connectionist networks can be found in the study of cultural transmission (1995:145). Strauss and Quinn clarify the notion of connectionism for cultural models: “If the elements of a cultural model are like the units of a connectionist network, they need not be in consciousness. Indeed, given the limitation on what can be held in short-term memory, the many parts of such large schemas could not all be in consciousness at the same time” (1997:154). Since only parts of these networks can be held at any given time due to memory constraints and bottlenecks in attention and processing, the informant can only give the ethnographer tantalizing parts leaving her to infer the larger whole. But the parallel activation of large schemas could well explain their ability to be transparent, used but not seen. Nevertheless, the property of transparency is as much of a bane as it is a boon, depending on one’s scientific philosophy.

A more anthropological critique of schema theory comes from Holland. In “The Woman Who Climbed Up the House: Some Limitations of Schema Theory,” Holland describes limitations she feels to be quite serious: including “the messy situation” (1992). By “the messy situation,” Holland means that when one is faced with a novel situation, without easy recourse to past solutions—personal or shared—or when one is faced with conflicting schemas for a resolution then how does one account for the behavior that follows? In a simple schema theory, one cannot account for this situation. According to Holland, schema theory has a vulnerability to imperfect fit and multiple interpretations. In relation to novelty, as Shore stated, “each of us is adept at constructing idiosyncratic models of experience on the fly, as a basic meaning-making strategy” (1996:46). Humans are not faxes receiving and transmitting invariant schemas. We are schema makers. Even when we internalize foundational schemas we retain the ability to alter and interpret our individual experience, the occasional conversion experience indicates an ability to reformulate deeply internalized understandings. Additionally, Shore stipulated that analogical reasoning lies behind most human cognition and allows us to construct novel schemas (1996:354). The tremendous power of analogy and metaphor that humans find so easy to manipulate still remains a vast challenge for digital computers. The ability to pattern-match and to selectively borrow and blend different patterns provides us with an infinite source of creativity within our culturally mediated world.

Concerning Holland’s critique of multiple interpretations; this doesn’t really create much trouble in a complex schema theory that takes into account such things as Spiro’s levels of cognitive salience and resultant internalization. Additionally, a framework like Shore’s clearly illustrates how a number of levels of models interact for any given decision. It is all but impossible to have two such complex models end up with exact equivalence. In spite of its timeless quality as a play, Hamlet—with its meditation upon an interminable dilemma—fails to be a problem in the world of nonfiction. People act. They may employ careful deliberation but action cannot be too long postponed in real life, away from the stage. In spite of these answers to Holland’s critiques, there’s no doubt that a number of vivid ethnographies that make use of schema theory, in even the messiest of situations, need to be undertaken before the theory gains the credence of mainstream anthropology.

The work of Pierre Bourdieu and his ideas about practice have been forwarded as criticism of schema theory. Shore discusses Bourdieu’s critique of structuralism which has been brought against schema theory (1996:266). Bourdieu feels that structural accounts make strong claims about objectivist models which he feels ring false to native experience. Objectivist accounts read like the writings of foreigners who do not understand the native’s point of view and thus use abstract tools to make some sense of the alien environment (Bourdieu 1977:2). And while Shore agrees with Bourdieu’s critique that abstract general models are not sufficient representations of an agent’s practical knowledge, he feels it is equally untrue to presume—as Bourdieu seems to—that all knowledge is exclusively egocentric in orientation. A competent theory of cultural transmission must account for both sociocentric and egocentric orientations (1996:266). Similarly, Strauss and Quinn appreciate aspects of Bourdieu’s practice theory while rejecting some of his stances (1997:44-47). For instance, they appreciate Bourdieu’s discussion of the way in which an individual’s habitus is structured by his or her experience and largely without the learning of explicit propositions. But while their belief in connectionist networks can answer this aspect of Bourdieu’s practice theory, his does not support the demonstrable fact that a great deal of learning and behavior is rule-based. Bourdieu’s anthropological work has been performed in nonliterate societies without formal schooling (1977:89) and his theories seem to fit the contours of such societies quite well. But in a great many societies, especially literate ones that emphasize extensive schooling, rule-based learning is quite common, if not the norm (Strauss and Quinn 1997:46). So while Bourdieu’s practice theory receives some support from cognitive theorists who feel that it resembles aspects of schema theory, in general they see it as a biased and incomplete theory of culture. And the critiques that Bourdieu submits against objectivist accounts only apply to parts of schema theory.

The last set of criticisms to be reviewed comes from Roger Keesing, whose essay “Models, ‘folk’ and ‘cultural’: Paradigms Regained?” ends the seminal volume Cultural Models in Language and Thought (Holland and Quinn 1987). Keesing asks what is gained from such attempts at cognitive anthropology. Are these unwieldy theories just too complex? Are there not more parsimonious ways of accomplishing as much as these theories do but with less effort? Keesing wonders: “Do we gain from our cognitive assumptions and commitments? Or, do we achieve, perhaps more pretentiously and less gracefully, what is achieved through the word-magic of a Geertz depicting personhood in Bali (1966) or a Schutz (1962; 1967) or Natanson (1967; 1970) depicting the phenomenology of self and role?” (1987:376). Indeed, the “cost” of cognitive anthropology is great. These are not easily wielded tools nor do they make ethnographies especially amenable to narrative structures. But the cognitivist approach, because it seeks to understand the organization of representations, may provide a more ecological portrayal of the individual within society. According to an ecological approach, the smallest unit of analysis is not “bounded by the skin or the skull” (Hutchins 1995:292). An ecological approach to culture includes the internal and external representations that serve to intersect the individual and the cultural. We have seen, through a brief review of schema theory, how culture and cognition reciprocally shape one another in a dynamic process. Should anthropology desire to capture both the psychological as an internal process and as an interdependent process, then it will need to get comfortable with the relatively complex set of tools offered by cognitive anthropology. If anything, the sophistication of one’s tools should serve as a rough gauge of progress. If anthropology used the same set of tools, with only minor modifications, it would have consigned itself to an Acheulian stasis. But the evolution of anthropology will force its students to continually shape, refine, and discard successively sophisticated toolkits in order to gain some understanding of the complexity that is culture.

Part II. The Cognitive Science of Religion

The cognitive science of religion (CSR) utilizes cognitive approaches borrowed from other fields to understand religion. CSR is a very young discipline, its critical texts emerging only since the early 1990s. CSR draws heavily from cognitive psychology; its prominent theories involve attention mechanisms and memory. Strangely, it does not derive much from ethnoscience, linguistics, or schema theory—the major concerns of cognitive anthropology from its inception in the mid-1950 through the mid-1990s. Schemas are mentioned from time to time by different CSR theorists but without any earnest attempt at reconciliation. Nevertheless, ‘traditional’ cognitive anthropology and CSR do share important assumptions, namely that the latest ideas from neuroscience and cognitive psychology ought to be used to investigate culture. Both fields share the belief that a better understanding of the cognitive constraints of the human mind will shed light on important aspects of culture.

In his article “Towards a Cognitive Science of Religion,” philosopher E. Thomas Lawson writes that “A cognitive science of religion would be possible if it could be shown that despite the obvious variability of religion across cultures and throughout history there lay a similar specifiable commonality” (2000:340). The search for commonality and for the widespread recurrence of religious representations has led many cognitive scientists of religion to the conclusion that for all its apparent diversity, religions everywhere possess a surprising number of commonalities (Boyer 1994). To some this bears no surprise: always and everywhere human beings have eaten basic foods, spoken translatable languages, enjoyed the pleasures of sex, and created art. No matter the era or locale, all human beings participate in the universality of the body, including its evolved neural architecture. Nevertheless, taking a domain as open-ended and free as religion—a domain that allows the greatest indulgence in imagination and flights of fancy—one might expect endless creativity and divergence. In fact, though, ideas about souls, witches, spiritual beings, and the magical efficacy of ritual show up again and again around the world in uncannily similar forms. These striking patterns support the idea that certain religious representations are, for lack of a better term, ‘natural’ for human beings. This does not imply that religion is innate in a Platonic or Jungian sense but rather that, because of their peculiar properties as representations, certain ideas possess an overwhelming intuitive appeal. As background, CSR subscribes to cultural evolution of the sort outlined by Sperber and described in the introduction. By understanding the selective transmission of representations, cognitive scientists of religion can better grasp why some religious representations hold such appeal and generate so many salient inferences for people everywhere. In the same way that prototypes and schemas possess constraining characteristics, so do aspects of selective transmission originate in the cognitive constraints upon the possibilities of human meaning.

Cognitive Optimality

In Modes of Religiosity (2004), Harvey Whitehouse discusses why certain types of religious representations achieve “cognitively optimality:” “people everywhere seem to acquire certain similar kinds of information about supernatural agents, rituals, and myths. Such concepts are clustered around a cognitive optimum position in the domain of religion” (29). By cognitive optimum, Whitehouse means that certain concepts possess more intuitive appeal for human minds than others—they are more attention-grabbing, easier to recall, and subsequently highly transmissible, epidemiologically ‘catchy.’ Stories about witches, for instance, can be found everywhere and routinely grab our attention whereas the axioms and corollaries of Euclidean geometry require careful processing and elaborate memorization before their meaningfulness can be ascertained. Both of these representations are quite abstract and can only be understood with relevant details and discussion yet witches win out over geometry most of the time.

Whitehouse considers cognitive optimality to be “essentially a universal attractor position, around which many cultural concepts, including religious ones, will be liable to congregate in the absence of countervailing tendencies” (2004:29). Representations that draw attention and encourage memorability by subtly violating our expectations approach a cognitive optimum. Cognitively optimal representations resemble Bastian’s “elementary ideas.” Boas helps to elucidate this idea:

They may be indigenous, they may be imported, they may have arisen from a variety of sources, but they are there. The human mind is so formed that it invents them spontaneously or accepts them whenever they are offered to it. This is the much misunderstood elementary idea of Bastian. (Boas 1940:272)

Bastian, Boas, Whitehouse, and others believe highly recurrent patterns in representations exist because of the structure of the human mind. By approaching a cognitive optimum, certain representations either lend themselves to independent invention or exceptional transmissibility across time and space.

Hypersensitive Agency-Detection Device

Another aspect of cognitively optimal religions derives from Stewart Guthrie’s ideas about religion as “systematic anthropomorphism” (1993). Like Boyer, Guthrie adopts some important assumptions from evolutionary psychology. Most significantly, Guthrie believes that humans evolved great sensitivity to agency detection or, to use Barrett’s interpretation, a hypersensitive agency-detection device (HADD) (Barrett 2004:32). Those that did inherit such hair-trigger for potential agency in the environment were more successful in surviving and reproducing than humans that did not possess hypersensitive agency-detection. One need only remember the last time that a late night noise disrupted sleep to realize how strong the instinct is to perceive threat in an ambiguous environment. The fear and awe that accompanies such a moment is not easily shaken, even if the noise can be attributed to an icemaker or to tired lumber.

By imagining agency everywhere—behind each rustle of leaves and every strange noise—ancient humans alert to the possibility of predators escaped death. But by projecting agency where it often is not, human beings acquired a tendency to worry about invisible agents—in other words, spirits. For Guthrie, religion amounts to systematized anthropomorphism on this basis. The world at large has taken on the characteristics of human agency due to this survival enabling HADD. In relation to cognitive optimality, those representations that highlight agency receive greater attention than others. And even more poignantly, agency that may present danger or threat receives our greatest attentions. Whatever representations contain elements of threatening agency achieve maximal attention and transmissibility.

Minimally Counterintuitive Representations

The idea that mysteries and curiosities of various sorts draw attention was something that Sperber noted originally (1985). But, as he writes: “There are infinitely many mysteries competing for mental space, and hence for cultural space. What advantage do the winning mysteries possess? They are, I want to suggest, more evocative, and as a result, more memorable” (85). What does it mean to say that some mysteries are more “evocative” and thus selectively win out in an evolutionary selection of representations? According to Sperber, the most evocative representations share commonalities with lots of other representations but cannot be given any final interpretation because they violate some expectations. He refers to these culturally selected representations as relevant mysteries (Sperber 1985:85). Their relevance originates in the fact that they relate to many other representations yet offer up a mystery that is open for interpretation and that stimulates wonder.

A few years later, Scott Atran and Sperber (1991) further clarified these “relevant mysteries”: “They are generally inconsistent with commonsense knowledge, but not at random: rather they dramatically contradict basic commonsense assumptions” (91). These representations violate category distinctions and intuitive ontologies. So, for instance, the category (or semantic domain) of PERSON includes a schematized understanding of the many vital properties that characterize a person. One can minimally violate this category by selectively blending it with that of a bird. From such a blend, we might arrive at the notion of a person that flies. This is a serious rupture, a true break from normal ontological notions about PERSON and yet the break is minimal enough that it is comprehensible, eminently memorable, and certainly attention-arresting.

Atran and Sperber sketched out the basics about counterintuitive ideas and their contribution to the recurrence of religious representations. But it was Pascal Boyer who developed a sophisticated set of theories about the importance of these minimally counterintuitive representations (MCIs) (1994, 1996, 1997, 2000). Of critical importance to Boyer’s theories about MCIs is the assumption of an evolved intuitive ontology.

Adopting Tooby and Cosmides’ ideas about the emergence of human cognitive capacities during the Pleistocence (1992), Boyer asserts a species-wide affinity for representations of ontological domains (1996). Intuitive ontology promotes a set of schematic representations concerning such categories as PERSON, ANIMAL, PLANT, and ARTIFACT. Were human beings required to learn virtually everything from scratch, generating sophisticated distinctions between all the sorts of basic kinds that are to be found in a terrestrial environment, the process of learning might well be endless. According to Boyer and others (Mandler & Bauer 1989; Mandler et al. 1991), the existence of numerous ontological categories can be demonstrated in children as young as 18 months old. Such evolved intuitive ontologies elicit the schematization of properties that can broadly be attributed to all PERSONS, all ANIMALS, all ARTIFACTS, and so on. Additionally, these and other domains receive significant cultural enrichment and ramification. Through intuition and learning, we recognize that all ANIMALS require food in order to live. ANIMALS cannot survive in the way that PLANTS do by remaining stationary and, apparently, not eating. Such default expectations make us susceptible to the sort of counterintuitive surprises that form the basis for Boyer’s ideas about MCIs.

Importantly, there are limits to counterintuitive representations. Cognitive psychologist Justin Barrett demonstrated that counterintuitions are better recalled than intuitions in general, but that at least some counterintuitions achieve greater recall than merely bizarre representations (1998). Greater weirdness does not equate to greater recall. Rather, specific patterns in countering human intuition approach the cognitive optimum. Boyer and Ramble (2001) performed a set of cross-cultural experiments in which they found that minimally counterintuitive violations (e.g. a flying person) were more highly memorable and compelling than maximally counterintuitive violations that were simply obscure (e.g. a person who lived under a carrot). Specifying how MCIs improve recall, while it seems a relatively arcane point of cognitive psychology, may lead us to a fundamental insight about religion. In his 1990 work, The Naturalness of Religious Ideas, Boyer writes: “Once we understand how mental representations can be transmitted, once we see that not all ideas can be transmitted equally easily, then we have the first elements of an explanation of religious representations” (ix). Of course, a notion of selective transmissibility applies just as much to religious ideology as to any other set of representations. But religion possesses special properties in that category violations and fantastical concepts which would not easily be tolerated in other domains remain preferable.

An as yet undeveloped finding about MCIs is their direct mapping to Fauconnier and Turner’s ideas about conceptual integration and blending theory (2002). These ideas can lend greater specificity to the process of countering intuitions and can illustrate how human beings can derive endless forms of representations from a basic ontology. Fauconnier writes: “Conceptual integration consists in setting up networks of mental spaces which map onto each other and blend into new spaces in various ways” (24). In the sense used here, a mental space can be directly related to a schema. And, using our example, the schema of a BIRD can be projected into a new schema which preserves an aspect of BIRD (namely flight) with the schema of PERSON. This type of cognitive blend is called a double-scope integration because it receives structural elements projected from two different inputs to create a new schema with its own emergent properties. For instance, the notion of a flying person may or may not have wings, may or may not lay eggs, may or may not eat worms, etc. What properties it has can be selected according to how we integrate the two inputs. Because we are left a significant role in the construction of this blend, it elicits a great deal of attention. Does our flying person live in a nest? Does he hunt like a raptor or peck like a chicken? Does he pass for a normal PERSON when on the ground? Our curiosity and imagination attract us to such blends. Because a blend like this receives input from two highly meaningful domains, it is a tremendous attractor for cognition. In contrast, a blend that receives projections from PERSON and projections from ROCK leads us further from a cognitive optimum because these two ontological categories veer from one another quite radically. Such a conceptual integration poses greater challenges: do many such ROCK-PERSONS compose a mountain? Does this new representation just stand in solemn fixity? While such mappings are possible, their wide ontological divergence makes them less likely to draw attention and achieve memorability.

What’s most important about borrowing from Fauconnier and Turner’s ideas about conceptual integration is a better understanding of the process by which religious representations come about (as ontologically relevant double-scope blends) and how they evolve. The nature of such evolution is that each new blend can serve as a source domain for a subsequent blend. Imagine that our BIRD-PERSON, who has all the attributes of a PERSON and can fly like a BIRD, then becomes a source domain and gets blended with a fearsome CAT. Now the flying person possesses cat eyes, fangs, and claws. Through any number of such blends, we create chimerical monsters which serve as the basis for such things as witches, vampires, and many other mythological creatures. Furthermore, they trip up our HADDs so that we really concern ourselves with them even though our better judgment convinces us that they are imaginary. The blended features of such chimeras exhibit tremendous complexity were one to consider how many ontological categories became selectively blended yet the cognitive processes at our disposal make this a relatively straight-forward process. Furthermore, to discuss such a fearsome creature, say for the first time around a campfire, poses no great challenges for the listener. Through the “backstage cognition” of conceptual integration, the neophyte quickly assesses the wondrous characteristics of this MCI which is, after all, just a blend of PERSON and ANIMAL.