studies draw upon and perpetuate the conduit model. When interpretive anthropologists look for meaning, they, like ordinary users of language, look for it in symbols. By assigning meanings to symbols, interpretive anthropology imparts a strongly culturalist bias to human studies. The point is not that a concept of culture is irrelevant to human studies—far from it—but that it must be put in its analytical place.
So Where, and What, Is the Culture?
In rejecting the conduit model, I have also rejected the equation of culture with symbol systems. Culture consists of meanings in people, not meanings in tokens. This formulation, however, presents serious conceptual challenges.
Theodore Schwartz (1978) recounts how, as he contemplated the incredible diversity in opinions offered by the inhabitants of Manus, he found himself asking “Where is the culture?”20 He coined the term “idioverse” to designate a person’s subjective reality. Idioverse is a useful heuristic concept. The toolmakers live in disparate idioverses; attaining correspondence between portions of idioverses is, according to Reddy, the object of communication—and, I would suggest, the origin of culture. Culture, then, refers to the overlap of idioverses among members of a given group at a given moment—the temporary set of their intersubjective conceptual networks. As emphasized above, this intersubjective array is not just a determinant, but also an unfolding product, of the public trade in symbols.
Unlike most culture concepts, which stress conformity and continuity, culture thus defined is partial, multiple, and plastic. Some meanings are noncultural; everyone participates in many discrepant “cultures”; and cultural meanings can and do change. A comparison of Reddy’s work with that of a second cognitive linguist, Ronald Langacker, brings some of these points into sharper focus.
Langacker’s innovative, controversial “cognitive grammar” (1986, 1987, 1991), which unfortunately I can do no more than sketch here, offers a new paradigm for the field of linguistics. Denying that “language is a self-contained system amenable to algorithmic characterization, with sufficient autonomy to be studied in essential isolation from broader cognitive concerns,” Langacker insists that “language is neither self-contained nor describable without essential reference to cognitive processing” (1986: 1). Indeed, he equates meaning with conceptualization rather than, as is usual, with bundles of essential features or sets of truth conditions (1986: 2–3).
Thus lexical items are points of entry into “knowledge [conceptual] systems whose scope is essentially open-ended” (1986: 2). Both the entities comprising this system (concepts) and the relationships among them (perceptual and transformative cognitive processes) are postulated to be real features of the mental world rather than formal semiotic units and operations. That is, words “mean” a conventionalized network of concepts interrelated by various cognitive processes. By implication, the network activated (or accessed) by any lexical item is part of a knowledge system of encyclopedic size.
Like Reddy, Langacker distinguishes between symbol and meaning: for him, symbols activate part of an immense knowledge system located in mind. But Langacker emphasizes congruent rather than idiosyncratic zones of idioverses.21 Langacker’s meanings are both cognitive and, because shared, cultural. Such conventionalized meanings—learned, shared meanings that are recognized as shared (Langacker 1987: 62; D’Andrade 1987: 113)—become fixed points of orientation in thought.22 Hence Langacker takes up where Reddy leaves off—after the toolmakers have managed to agree upon meanings. Such consensus meanings need not be accepted as, or felt to be, paramount, but they are inescapable: herein lies culture’s constraining quality.
By contrast, Reddy’s account of toolmaker incomprehension highlights not the symbol’s instantiation of shared conceptual networks but the variability in its construal, the gap of indeterminacy between public symbol and private meaning that reveals culture’s limits. A meaning-system view of culture emphasizes that conventionalized meanings are only part, often a small part, of the story. The clinical literature, of course, is rife with case studies of persons who attach unusual, elaborate, compelling meanings to banal items. Like inkblots, ordinary symbols routinely evoke powerful and diverse personal meanings that lie well outside cultural understandings.23
A comparison of Reddy and Langacker also illuminates issues of cultural temporality. Langacker’s conventionalized symbols are historically given, whereas Reddy’s improvised symbols seem ahistorical. The propagation of language over time is a complex and far from automatic process, but for individual actors at a given moment the words exchanged are recognizable communicative gestures and the meanings evoked by those words are partially intersubjective. The contrived situation of Reddy’s toolmakers, who circulate only slightly conventionalized “inkblots,” differs radically. Historical time does not exist: the focus is on the present communicative event. Intersubjectivity is minimal. Reddy’s premises are intentionally unrealistic; they serve his discussion of the problematics of intersubjectivity, stressing difficulties in real-time communication.
But there is room for a synthesis. Our stock of symbols—words, rituals, physical signs—is, from the perspective of living persons, a social inheritance. Some of the meanings people learn to attach to those symbols— the cultural component of idioverses—are intersubjective. In real time communication, such intersubjective meanings are relatively fixed, that is, reliably elicited as a consequence of common learning. But they are supplemented, at times even overshadowed, by private, biographically salient meanings. Hence, culture-in-the-short-run can seem an immutable, weighty legacy or a drop in a sea of unique and often powerful private meanings. In the medium or long run, intersubjective meanings themselves change, although such changes are hard to identify from available symbolic evidence, a point I explore further in the following chapter. Cultural anthropology does better with microcosmic, short-run analyses because the necessary fieldwork brings us into close contact with living persons: we can explore real-time communication rather than just tokens, the collective representations that those who study macrocosmic and historical events tend to think of as social level conduits of meaning.
Foucault’s sweeping macroethnographies of historical changes in categories of Western thought (e.g., 1977, 1990 [1976]) have inspired a sizable anthropological literature, sometimes called postmodern or poststructuralist, that includes both critiques of the discipline and a growing body of ethnographies. In such studies, the analytic concept “culture,” a staple of both interpretive and cognitive studies, is displaced by “discourse.” This is more than a stylistic move: the notion of discourse “is meant to refuse the distinction between ideas and practices or text and world that the culture concept too readily encourages” (Abu-Lughod 1991: 147). Discourses are not, precisely speaking, conduits, and ethnographies of discourse are notable for their attention to the making of meaning. Nevertheless I will argue that discourse approaches, like the interpretivism they self-consciously reject, rest substantially upon and propagate certain conduit assumptions.
Recovering Lost Minds
Discourse Versus Culture
The discourse perspective has undeniable merits. Such studies examine meaning-making as a historical process. Discourse theorists distinguish themselves from interpretive anthropologists by their attention to diversity, temporality, and practice. Moreover, they have taken special heed of political issues, especially the constraining and oppressive social consequences of the historical production of meaning. Such concern, though not absent, has not always been prominent in either interpretive or psychological anthropology.
Unfortunately, however, I think that discourse analysts, despite their criticism of the reification of culture in interpretive studies, do not themselves break free of reifying assumptions of the conduit model. A review of ethnographies of discourse is obviously beyond the scope of this chapter, but an exemplary study, Richard Handler’s Nationalism and the Politics of Culture in Quebec (1988), offers insights into some strengths and weaknesses of the approach.
