intersubjectivity.
We are all, of course, Reddy’s toolmakers; we inhabit disparate, mutually inaccessible subjective universes. We send each other concrete messages in the form of signals (public symbols),7 but deciding which to send, and making sense of those received, are complex tasks. Private worlds of thoughts, feelings, and perceptions can be represented only obliquely and must be inferred by others. Our habitual assumption that the exchange of signals equates to the communication of meanings obscures the fact that intersubjectivity is an accomplishment, not a natural consequence of signaling.
Reddy’s own argument is an unintentionally telling, paradoxical piece of evidence for the position he espouses. His main point—that we easily fall prey to the illusion that symbols convey meanings—is underscored by an ambiguity that verges on a contradiction. If (as Reddy asserts) symbols do not convey meanings, then wherein lies the power of the conduit metaphor? By pinning the blame for our confusion on what he calls “our language about language,” does not Reddy fall victim to the very model he renounces?8 For this metalanguage is not reified thought, a symbol system that smuggles warped ideas into our brains, but an objectification to think with. The problem, as I see it, is not that English speakers are coerced by their language, but rather that conduit metaphors articulate closely with an entrenched way of thinking: evocative symbols and evoked thought seem nicely equilibrated, mutually reinforcing. As Naomi Quinn observes,
Metaphorical systems or productive metaphors typically do not structure understandings de novo. Rather, particular metaphors are selected by speakers, and are favored by these speakers, just because they provide satisfactory mappings onto already existing cultural understandings—that is, because elements and relations between elements in the source domain make a good match with elements and relations among them in the cultural model. (1991: 65)
We process such metaphors more or less automatically: they make immediate sense, because their linguistic components link up readily with existing cultural models, that is, complex, broadly shared meaning systems (D’Andrade 1984; Holland and Quinn 1987).
Such “automatic” concordance between language and thought, while common, is hardly inevitable.9 And restructuring thought through language is often difficult. For example, the toolmakers paradigm, an alternative symbolic rendition of communicative events, evokes a way of thinking that vanishes before we can grasp it securely. Focusing now on ideas rather than language, Reddy writes, “I do not claim that we cannot think momentarily in terms of [a nonconduit] model of the communication process. I argue, rather, that that thinking will remain brief, isolated, and fragmentary in the face of an entrenched system of opposing attitudes and assumptions” (1979: 297–98). Reddy’s own seeming inconsistencies (and the many which, I am sure, the reader will find in this chapter) suggest the difficulties of both toolmaker thinking and its objectification in symbols that can sustain the new understanding.10 That is, the interaction between mind and symbol is highly unstable: both understandings and symbols tend to revert to the conduit pattern.
The toolmakers paradigm is, unfortunately, a cumbersome symbolic construction. To facilitate nonconduit thinking about communication, let me offer an alternative, more concise trope: the inkblot. Projective tests like the Rorschach, in which people are asked to comment on blotches of ink, evoke meanings on the basis of highly ambiguous stimuli. Because inkblots cannot easily be thought of as conduits, viewing ordinary symbols as if they were inkblots draws our attention to the interface between physical forms assumed to be meaningful (symbols) and the activity of meaning-making by persons.
It may be objected, however, that the synecdoche {inkblot = symbol} is inappropriate: an inkblot, which does not stand for anything in particular and is indeed so designed, seems quite a deviant symbol. What does an inkblot have in common with, say, the dot-dash of Morse code, a symbol that has a precise referent—a “real” symbol?
Although any figure of speech must be used cautiously, I think the inkblot trope jars us out of a deceptive mode of thought. We regard code units—dot-dash, for example—as better symbols than inkblots because conduit thinking easily accommodates an image of communication as the transmission of code units between cryptographs. But this is a travesty of most real-life communication, where symbols are subject to diverse construals (not one) by biographical persons (not machines). The inkblot trope is a provocation that forwards a narrow but crucial claim about all symbols: neither an inkblot nor dot-dash “has” meaning. It is true that those who know Morse code will make a meaning from dot-dash more automatically than they will from an inkblot.11 But such automaticity is the result of prior learning by discrete persons, not a property of the symbols. The inkblot trope underscores the point that meanings arise in interactions between symbols and human minds, whether those symbols be inkblots, dots and dashes, or ordinary symbols that elicit a combination of conventional and idiosyncratic understandings.12
To summarize: metaphors are not concepts, words are not thoughts, symbols are not culture. I think Reddy’s chief target is not deceptive language but rather a “misleading and dehumanizing” (1979: 308) commonsense cultural model, conduit thinking, which receives powerful reinforcement not only from conduit language, but also from unconscious perceptions. Because we typically experience words and other symbols as suffused with meaning, we sometimes imagine that our ideas and feelings have come, special delivery, from without. Our automatic processing of much everyday symbolic material provides experiential sustenance for conduit thinking: the world seems to speak for itself. Hence the receiver’s meaning-making, essential to any communicative event, gets little attention in our practices of communicating and understanding and, I shall argue, in our theories of culture, which are thereby distorted in two major respects.
Distortions
The First Distortion: Displacing the Natives
For cognitive linguists such as Reddy, symbols are public, concrete objectifications. Meanings, on the other hand, are networks of knowledge accessed or evoked, but not conveyed, by symbols. Such linguists propose a shift from extensional to intensional theories of meaning (D’Andrade 1990: 123). Extensional theories look for reference—for correspondences between words or other symbols and things in the world. If meaning is reference, the mind can be a black box; an account of meaning need not concern itself with cognition. By contrast, intensional theories focus on the sense that symbols have for natives. Intensional semantics insists that what things mean depends fundamentally on what and how people think, not on direct links between signs and referents or on features of language per se.
Because intension mediates between symbol and meaning, communication among toolmakers is a clumsy business; the visiting anthropologist’s job is doubly difficult. It has been tempting for us to collect the messages, which are, after all, readily accessible, discover systematic relationships among them, and present an interpretation. Virtuoso readings of symbols and rituals by informed, sensitive observers well steeped in the local culture—for example, Geertz’s Balinese cockfight (1973a), Victor Turner’s Ndembu milk tree (1967), and David Schneider’s American kinship (1968)—have been highly influential in the elaboration of Anglo-American culture theory. Other readers of cultural “texts” use more formal schemes, drawn from literary criticism, semiotics, structuralism and so on, as keys to unlock the meanings hidden within public symbols. Whether through cultural savvy or formal cryptography, interpretation aims to reveal deeper and more compelling messages hidden beneath an enigmatic or misleading “textual” surface of physical movements, objects, or words.13
The social world in which anthropologists immerse themselves is not, however, a set of boxes within boxes with a treasure (or, perhaps, only more boxes14) at the center. It is people doing things. Communication, rightfully a prime focus of anthropological inquiry, is a social and intrapsychic practice. It cannot be boiled down to a key, a set of meanings “conveyed” or “embodied in” symbols. Indeed, conceiving social and intrapsychic life as a disembodied text rather
