Team Members

Principal Investigator:
Anthony Synnott, Department of Sociology and Anthropology

Michael Bross, Department of Psychology
Charles Davis, Department of Religion
David Howes, Department of Sociology and Anthropology

Background to the research

The first cross-cultural study of sensory perception was conducted by the British Torres Strait Expedition (Haddon 1901). Haddon and company were primarily interested in testing the acuity of the senses of the natives they encountered. This concern with measurement was gradually supplanted by an interest in the structuring of sensory experience, particularly tactile experience, which flowed out of the American Culture and Personality School's interest in child-rearing practices (e.g. Williams 1966). As well, various figures on the fringes of anthropology, such as Leenhardt (1947), Carpenter (1955) and Hall (1969), not to mention Simmel (1921) in sociology, began to speculate on the meaning of sensory experience in different cultures.

The publication of Berlin and Kay's (1970) Basic Color Terms represented a major turning point in the anthropology of the senses. The concern with measurement and that with meaning were brought together in a single study, and judging from the results, it appeared that natural languages could be arranged on an evolutionary scale consisting of five stages. The discovery that there exist certain universal focal colours, and that there is an evolutionary sequence to the discrimination of colour terms, led many anthropologists to abandon the so-called "Whorf-Sapir hypothesis", that is, the thesis that culture or language "determines" perception (Whorf 1956; Cole and Scribner 1974).

However, in defence of the relativist position, it should be observed that Berlin and Kay's methodology was seriously flawed. They chose to treat the sense of sight in isolation, whereas people perceive the world through all their senses simultaneously. What is more, as the work of such cultural historians as McLuhan (1964) and Ong (1967) had already shown, cultures vary in the intensity with which they attend to a given sensory field. Thus, to take Berlin and Kay's example of the Dani of New Guinea, the reason the Dani colour vocabulary consists of only two terms may be because vision is not for them (as it is for us) a field of "productive specialization" (Ong 1967: 6). It could be that their vocabulary of sound or smell is highly elaborate, but this would not show up on Berlin and Kay's evolutionary scale.

Recent research in the anthropology of the senses has called attention to this blindspot. Kuipers, for example, has stressed that a useful direction for comparative research would be to focus on "variation in the relationships within and between culturally defined sensory domains" (1984: 97). One of the most interesting case studies in point is Seeger's (1975) work on body decoration among the Gê-speaking peoples of Mato Grosso, Brazil. His work shows that variation in the ornamentation of a body part (eye, ear, nose) reflects variation in the emphasis on the related faculty. Recent research is also characterized by a concern with documenting "how people think they perceive" (Seeger 1981: 81), that is, with describing indigenous theories of perception. The current situation contrasts markedly with the experience of the Torres Strait Expedition: "Attempts were made, but with very little success, to find out what was actually passing in the minds of the natives while making these observations [i.e. doing the tests]" (Haddon 1901: 27).

2. Objectives, Conceptual Framework and Scholarly Significance

The theoretical approach which we adopt is in basic accord with the comparative, holistic and emic orientation of Kuipers and Seeger's work. However, we feel that their perspective is lacking in historical and psychological depth. It is for this reason that a specialist in psychology, Michael Bross, and a specialist in cultural and religious history, Charles Davis, are included in our research team. The latter disciplines are more advanced than anthropology insofar as the development of models for the study of alterations in sense ratios is concerned. It is in the light of such models, and the work of certain current anthropologists, that we propose to analyze (or re-analyze) the casual observations of earlier ethnographers, and arrive at a comprehensive picture of how the organization of the sensorium may vary cross-culturally.

The objectives of our research are as follows:

a) To construct a typology of cultures based on variations in the organization of the sensorium. It would seem that the senses are normally ordered hierarchically, rarely equally. An example of the latter possibility would be the culture of sixteenth century France as described by Febvre (1942; Mandrou 1975). Examples of the former include Leenhardt (1947) on the tactile underpinnings of Canaque collective thought; Seeger (1981) on the role of olfaction among the Suya Indians of Brazil; Feld (1982) on the primacy ascribed to sound symbolism among the Kaluli of New Guinea; Kuipers (1984) on the social significance the Weyéwa of Sumba attribute to taste; and Tyler (1984) on our own seeing-as-knowing complex, or Babb (1981) on the visualism of Hinduism. As these works attest, different cultures display different perceptual orientations, and can be classified accordingly. Of course, it is not enough merely to identify which sensory mode is predominant within a given culture: the sensorium must always be analyzed as a whole.

It is in connection with this problem of devising suitable models for the study of inter-sensory relationships that anthropology has the most to learn from psychology. Michael Bross (1978, 1980 and 1985) has done extensive research into how the application of sensory restriction techniques to one sense organ alters sensitivity in other sensory channels. His research will form the basis on which we construct our typology (or repertoire) of possible sense ratios, the premise being that cultures are sensory restriction techniques writ large. With the aid of this typology, we shall proceed to study the ethnographic literature. Cultures will be sorted in terms of which of the possible sense ratios they actualize.

Our approach may be said to be inspired by the doctrine that "human societies, like individuals, never create absolutely, but merely choose certain combinations from an ideal repertoire that it should be possible to define" (Lévi-Strauss 1975: 178 and 1985: 157-58; Howes 1987c). In other words, our method is consistent with the first principles of "structural analysis". As for the scholarly significance of our way of proceeding, its potential is best summed up by Ong: "Given sufficient knowledge of the sensorium exploited within a specific culture, one could probably define the culture as a whole in virtually all its aspects" (1967: 6).

b) To assemble an encyclopedia of native theories of perception. Our concern in this section is with the meaning given to different sense organs. For example, in many non-literate societies, the eye is regarded as the seat of witchcraft, hence an anti-social faculty (Spooner 1970; Seeger 1981: 86-89), while the ear is thought to be a social faculty, as well as the most informative of the senses (Stoller 1984a and 1984b; Guss 1986). The meaning or value attached to a given sense organ will obviously affect the degree to which it is cultivated.
Native theories of perception are frequently implicit as opposed to explicit, and the problem of extracting them from a given ethnography is further exacerbated by the fact that the ethnographer may not be sufficiently alert to follow up on the hints that such theories exist. For example, Malinowski records that in the Trobriands "the sense of smell is the most important factor in the laying of spells on people; magic ... must enter through the nose" (1929: 44), but in all of his subsequent discussions of magical spells he focusses exclusively on their verbal (oral-aural) aspects (Malinowski 1923: 322-23 and 1961: 428-63). It could be that the Trobriand theory of "the power of words in magic" was, in fact, a theory of smells.

Thus, one must always be attentive to the possibility of an ethnographer placing a false interpretation on "the facts". It seems safe to presume that English-speaking ethnographers will normally err in the direction of the visual (Tyler 1984), but not exclusively (e.g. Malinowski's ear). The best way to correct for such distortions is to follow Ihde in asking: "In what way is each sense unique? And what implications does each area of sense have for ... thought?" (1973: 25), that is, to perform a sort of phenomenological reduction. At the same time, we must be wary not to over-intellectualize the sensory experience of other cultures (Davis 1976 and 1986), which is one of the paradoxical dangers of too "pure" a reduction.

c) To make an inventory of the rules and practices (ritual, child-rearing, instructional, artistic, etc.) which have the effect of altering or maintaining particular sense ratios. Innis (1951), McLuhan (1964), Ong (1977) and Goody (1977) have all made important contributions to our understanding of the effects of the practice of literacy on the organization of the sensorium. Literate cultures tend to place much emphasis on the eye, non-literate cultures on the ear. However, even in the most advanced literate society, "recessive traditions" of perception may occasionally come to the fore (Ihde 1973: 43). An example would be post-Revolutionary France, where olfaction enjoyed a brief golden age, seemingly at the expense of sight (Corbin 1986).

In making our inventory, one point to which we attach special importance is avoiding the kind of reductionism displayed by Montagu when he attributes the fact that the Aivilik adult views space "not as a static enclosure but as a [constantly fluctuating] direction in operation ...[to] the twistings and turnings the [adult as] infant experienced while being carried on the mother's back" in the hood of her parka (1978: 239). Child-rearing techniques are important, and show remarkable variation in time and space (Synnott 1983), but artistic conventions and architectural norms may be just as influential vis-à-vis the inculcation of a particular way of perceiving the world (e.g. Leenhardt 1979: 11-23; Tuan 1974).

In our research, we shall also be concerned to investigate ritual practices, such as the visualization techniques of the Hindu tradition (Teskey n.d.), or the use of smell symbolism is passage ritual (Howes 1987a). Rituals may also be designed to induce inter-modal transfers, or "synesthesia", a phenomenon which cultural historians (Chidester 1981; Camporesi 1986), linguists (Williams 1976), psychologists and literary scholars (Marks 1978) have investigated at length.

d) To study the cognitive (or epistemological) and affective implications of variations in the ratio or balance between the senses. How people think is influenced by what they perceive in two ways: first, the senses provide us with data; second, they provide us with metaphors (Chidester 1981). Thus, it is common for English-speakers to say "I see", as if knowing were a kind of seeing. Other cultures use other sensory imagery or metaphors to think about thought with, and as Boman (1960) has shown, this may have a profound effect on their conception of "truth", or what counts as "real".

English-speakers are convinced that one cannot know the way another person feels from how they smell, so convinced that a phrase like "You smell angry" is regarded as symptomatic of schizophrenia (Winter 1978: 122-23). In the Arabic-speaking world, where the link between smell and disposition is recognized, a prospective bride may be rejected if she gives off a "smell or anger or discontent" (Hall 1969: 49). In this case, one culture's sensibility is another culture's psychopathology. Our research will undoubtedly turn up further examples of alternative regimes of sensory values. It is important to understand how different sensory regimes give rise to different sets of assumptions; even notions of mental health are affected.

e) To determine the relationship of divergent modes of perception to social structure. Finally, our intention is to adapt "grid-and-group analysis" (Douglas 1973 and 1978) to the study of the organization of the sensorium. A plausible hypothesis would be that societies which display "weak control by grid and group" will privilege the proximity senses (touch, taste, smell), while societies characterized by "high classification" and "strong control" will be oriented towards the distance senses (sight and hearing) (Douglas 1973: 103-4). We freely acknowledge that the theory of "grid-and-group" creates as many problems as it sets out to explain, and therefore intend to overhaul this theory before any attempt is made to apply it.

This research was generously supported by a grant from the Social Sciences and Humanities Research Council of Canada.

The results of the study are reported in the following publications:

Constance Classen, Worlds of Sense: Exploring the Senses in History and Across Cultures. London: Routledge, 1993

Anthony Synnott, The Body Social: Symbolism, Self and Society. London: Routledge, 1993

David Howes (ed.), The Varieties of Sensory Experience: A Sourcebook in the Anthropology of the Senses. Toronto: University of Toronto Press, 1991.

Journal of Religion and Culture (1990) vol. 4(2), special issue

Anthropologie et Sociétés (1990), vol. 14(2), special issue

Anthropologica (1990), vol. 32(1), special issue