Toward a Cross-Cultural Multi-Modal Theory of Aesthetics David Howes
This essay opens with a review of recent advances in neuropsychological research on the multisensory organization of the brain. It goes on to critique that research from the standpoint of the anthropology of the senses. Crucially missing from the neuropsychological account of how the senses function is any recognition of the role that culture plays in the modulation of perception, as shown in Part I. Part II begins with an inquiry into the history of the Western concept of "aesthetics." This concept originally stood for the perception of "the unity-in-multiplicity of sensible qualities" (Baumgarten), but the intersensoriality of this initial formulation was quickly lost from view. Part III presents a survey of the role of the senses in the aesthetic practices of a range of non-Western cultures. The traditions covered all show a marked tendency to combine the senses (rather than separate them) for artistic and other purposes, most notably healing. This survey in turn lays the groundwork for the articulation of a cross-cultural multi-modal theory of aesthetics in Part IV, which recuperates some of the elements of the original definition. The essay concludes with a series of suggestions regarding how the disciplines of art history and neuropsychology might reconceptualize their engagement with the sensorium in a manner that is attuned to the historicity and interplay of the senses in art as in everyday life.
Part I: The New Multisensory Psychology of Perception It is commonly assumed that each sense has its proper sphere (e.g. sight is concerned with colour, hearing with sound, and taste with flavour). This modular conception of the sensorium is reflected in the analytic orientation of most current research in the psychology of perception with its sense-by-sense (or "one sensory modality at a time") approach to the study of perceptual processes. In recent years, however, a more interactive, relational approach to the understanding of how the senses function has begun to take shape as a result of the growing body of evidence which points to the “multisensory organization” or “integration” of the brain. As Calvert, Spence and Stein write in their introduction to The Handbook of Multisensory Processes (the most authoritative work in this new field):
even those experiences that at first may appear to be modality-specific are most likely to have been influenced by activity in other sensory modalities, despite our lack of awareness of such interactions ...[To] fully appreciate the processes underlying much of sensory perception, we must understand not only how information from each sensory modality is transduced and decoded along the pathways primarily devoted to that sense, but also how this information is modulated by what is going on in the other sensory pathways” (Calvert, Spence and Stein 2004: xi-xii, emphasis mine).
Examples of such modulation include the well-documented fact that, in noisy surroundings, speakers can be understood more easily if they can be seen as well as heard. This finding is readily explicable in terms of the redundancy hypothesis of classic information theory: two pathways or "information channels" are better than one.
However, the new multisensory psychology of perception probes deeper to explore the relationships among the component parts of a multisensory signal (e.g. Partan 2004). For example, in the case of animal and human communication, redundant multisensory signals can be subclassified into those that produce responses in the receiver equivalent to the response to each unisensory component (equivalence), and those where the overall response is enhanced (superadditive). Multisensory signals may also be made up of stimuli which convey different (i.e. nonredundant) information. In such cases, the relationships between the component parts of the signal are often quite complex, and it takes a great deal of painstaking attention on the part of the researcher to sort them out.1 For example, the relationship may be one of dominance, as in the case of the ventriloquism effect (where the seen lip-movements of the ventriloquist's dummy alter or “capture” the apparent location of the speech sounds), or one of concatenation as in the case of the reproductive behaviour of male oriental fruit moths (such moths "need the visual presence of the female in combination with her pheromones before they will perform their most intricate courtship displays, and they need an additional tactile stimulus of a touch on the abdomen before they will copulate" (Partan 2004: 235)). Emergence, as exemplified by the McGurk effect,2 is a third possibility, and this typology could be extended further.
Many of the studies in the Handbook use modern neuroimaging techniques to reveal the multiple sites of multisensory processing in the brain, including many regions long thought to be modality-specific or "primary sensory" areas as distinct from the so-called higher order, "associative" areas traditionally assumed to be responsible for the formation of unified percepts out of the diversity of inputs. In addition to demonstrating the functional interdependence of the modalities, a number of these studies point to their functional equivalence or adaptability. For example, it is now clear that sensory-specific areas can be "recruited" or "remapped" by other sensory-specific areas in situations of sensory deprivation or intensive perceptual training. Thus, visual cortex in blind individuals has been found to show activation in auditory tasks while auditory cortex in deaf individuals can be activated by visual tasks.
Of note, the quality of sensation associated with activating the visual cortex in congenitally blind individuals, or the auditory cortex in congenitally deaf individuals, appears to derive from the nature of inputs. That is, visual inputs are perceived as visual even when auditory cortex is activated [in the case of the blind, while the reverse holds true in the case of the deaf] .... Furthermore, even in normal, nondeprived humans, there is evidence for extensive multisensory interactions whereby primary sensory areas of the cortex can be activated in a task-specific manner by stimuli of other modalities .... Common to these findings is the principle that inputs recruit pathways, cortical areas, and networks within and between areas that process the information, and the sensoriperceptual modality associated with the input is driven by the nature of the input rather than by the cortical area activated per se (Sur 2004: 690).
Such evidence of adaptive processing, or "cross-modal plasticity," poses a serious challenge to the conventional model of the sensorium as consisting of five structurally and functionally distinct modalities. In light of this challenge, some researchers have proposed that the phenomenon of synaesthesia (i.e. the union or crossing of the senses, e.g. hearing colours, tasting shapes) might provide a more productive model for conceptualizing perceptual processes than the conventional sense-by-sense approach that has dominated research on the senses and sensations to date.
The Model of Synaesthesia The condition of synaesthesia is typically understood to be quite rare. Estimates of its incidence vary from 1 in 200 to 1 in 2,000 to 1 in 20,000 people (Ramachandran et al 2004: 868; Baron-Cohen and Harrison 1997). The most commonly documented form is colour-grapheme synaesthesia in which written words or letters are perceived as having particular colours. To limit synaesthesia to a congenital condition, however, would be myopic.3 Synaesthetic connections can be learned. Take the case of odour-taste synesthesia which, perhaps because it is such a common effect, has failed to attract much popular attention or scientific documentation. Yet the evidence is clear: “the majority of people appear to experience odor-taste synaesthesia. First, sweet is one of the most common descriptors applied to odors ... [Furthermore,] when smelling an odor, most people can more easily recognize a taste-like quality such as sweetness than more specific qualities such as strawberry- or banana-likeness” (Stevenson and Boakes: 2004: 69). This finding raises the question: When we speak of the odour of vanilla or strawberry as "sweet" are we speaking in metaphor rather than reporting an actual olfactory sensation? Not according to Stevenson and Boakes:
The central argument of [their chapter in the Handbook] is that, as a result of eating and drinking, patterns of retronasal odor stimulation co-occur with oral stimulation, notably of the taste receptors, so that a unitary percept is produced by a process of either within-event associative learning or by a simple encoding as one event. Eating sweet vanilla-flavor ice cream will ensure that the retronasal odor of vanilla becomes associated with sweetness; on some later occasion the smell of vanilla will seem sweet, even if no conscious recollecton of eating ice cream comes to mind (Stevenson and Boakes 2004: 81)
The “metaphor explanation” of synaesthetic perception is also rejected by V.S. Ramachandran, E.M. Hubbard and P.A.Butcher in their chapter on "Synesthsia, Cross-Activation, and the Foundations of Neuroepistemology." Their objection rests on methodological grounds: "Since very little is known about the neural basis of metaphor, saying that 'synaesthesia is just metaphor' helps to explain neither synaesthesia nor metaphor" and merely compounds the mystery (Ramachandran et al 2004: 868). The authors go on to tout the experimental procedures they have devised to determine whether an alleged synesthete's experiences are "truly perceptual" or merely conceptual (an important distinction when it comes to selecting test subjects since only those in whom the effect can be shown to be involuntary are considered desirable), and then offer a physiological explanation for the effect having to do with the “cross-activation of brain maps.” Such cross-activation may come about by two different mechanisms, namely: "(1) cross-wiring between adjacent [brain] areas, either through an excess of anatomical connections or defective pruning, or (2) excess activity in back-projections between successive stages in the hierarchy (caused by defective pruning or by disinhibition)” (2004: 872).
In the case of colour-grapheme synaesthesia - Ramachandran et al.'s chosen example - the brain areas corresponding to graphemes and colours are right next to each other in the fusiform gyrus, and the potential for excess cross-activation or “hyperconnectivity” as a result of some genetic mutation in those individuals who naturally experience this effect is therefore strongly indicated. Ramachandran et al conclude that
far from being a mere curiosity, synaesthesia deserves to be brought into mainstream neuroscience and cognitive psychology. Indeed, [precisely because the neural basis of synaesthesia is beginning to be understood] it may provide a crucial insight into some of the most elusive questions about the mind, such as the neural substrate (and evolution) of metaphor, language and thought itself” (2004: 881).
There is much to be said for Ramachandran et al’s “bottom-up” approach to the study of perceptual processes, but I find the physiological reductionism of their position unduly restrictive from my own perspective as a cultural anthropologist, and I think an equally valid case could be made for a "top-down" approach. Such an approach would start by examining the cultural organization of the sensorium and descend via the psychological to the physiological level of brain organization. In point of fact, due to the selective focus of their discipline - neuropsychology - Ramachandran et al never ascend in what they call “the hierarchy” as far as the cultural level.4 This oversight constitutes a serious lacuna, for as cultural psychiatrist Laurence Kirmayer observes concerning the hierarchical systems view of neural organization (which Ramachandran et al presumably share):
Contemporary cognitive neuroscience understands mind and experience as phenomena that emerge from neural networks at a certain level of complexity and organization. There is increasing recognition that this organization is not confined to the brain but also includes loops through the body and the environment, most crucially, through a social world that is culturally constructed. On this view, ‘mind’ is located not in the brain but in the relationship of brain and body to the world (Kirmayer, in press).
Ideally, Kirmayer states, “we want to be able to trace the causal links up and down this hierarchy in a seamless way.”
Following Kirmayer's lead, let us imagine what a "Cross-Cultural Handbook of Multisensory Processes" would look like. Instead of presuming sensory processes to be confined to the brain, it would start with the investigation of the culturally patterned “loops” through the environment - that is, with the study of the relationship of brain and body to the world. Thus, a top-down approach to the study of synaesthetic perception would begin by drawing up an inventory of the range of cultural practices and technologies that generate different sensory combinations across different cultures and historical periods. For example, it is a good empirical question whether the incidence of colour-grapheme synaesthesia would be as high in an aural-oral society as it is in a visual-literate one, such as contemporary Western society.5 In the latter, words and letters are experienced as quiescent marks on paper or a computer screen, which renders them available for colour-coding, whereas in the former words (being experienced aurally) might not tend to be seen so readily as they would be felt or smelled as well as heard. In my own ethnographic research in Papua New Guinea, I found evidence of audio-olfactory synaesthesia. In many Melanesian languages, such as Kilivila (the language of the Trobriand Islands), one speaks of “hearing a smell,” and this association is carried over in Pidgin English, "mi harim smel". The reason for this is that most communication takes place face-to-face (i.e. within olfactory range of the other) and odoriferous substances (e.g. anointing the body with oil, chewing ginger) are used to augment the power of a person’s presence and words (Howes 2003).
It is also common in various African languages, such as Dogon (a language of Mali), to speak of "hearing a smell." According to Dogon conceptions, speech "has material properties that ... despite its invisible nature ... are more than just sound ... [It] has an 'odour'; sound and odour having vibration as their common origin, are so near to one another that the Dogon speak of 'hearing a smell'" (Calame-Griaule 1986: 39, 48 n. 69). The Dogon also classify words by smell: good speech smells "sweet", and bad or impetuous speech smells "rotten" - indeed, "the mouth too ready to speak is likened to the rectum" (Calame-Griaule 1986: 320). This understanding is corporealized in, for example, the steps that are involved in a girl's "education in speaking." The first step occurs at age three with the piercing of a hole and the insertion of a metal ring in her lower lip so as to socialize her speaking. This is followed by the piercing of her ears at age six to socialize her hearing. Should the girl continue to make grammatical mistakes and/or utter impetuous remarks by age twelve, then rings are inserted in the septum and wings of her nose (Calame-Griaule 1986: 308-10). The nose-rings have the superadditive function of modulating both her hearing and her speech by promoting the reception and utterance of "good-smelling words" and the deflection or repression of bad-smelling ones - all in accordance with the intrinsic connection the Dogon postulate between sound and odour.
These findings of audio-olfactory synaesthesia, familiar to scholars of sensory anthropology, would likely come as a surprise to most scholars of neuropsychology. For example, Stevenson and Boakes claim that: "Odors display taste properties but do not elicit auditory or visual sensations" (2004: 73). On the contrary, in Mali as in Melanesia, sounds do elicit olfactory sensations, and vice versa. Thus, what Stevenson and Boakes take to be a physiological given actually rests on certain culturally-contingent (i.e. peculiarly Western) assumptions about the divisions of the sensorium and/or potential for cross-sensory activation. This example underscores the need for more cross-talk between the disciplines of neuropsychology and anthropology if contemporary cognitive neuroscience is to achieve a comprehensive understanding of the multiple possible forms of cross-talk between the senses. To put the matter another way, starting with cross-cultural examples such as these, which are practical (i.e. supported by cultural practices which form part of the “loop” through which all sensations must pass) as well as metaphorical, neuroscientists could well be inspired to discover all sorts of heretofore unsuspected cross-linkages between the senses wherever they may be localized in the brain.6
Part II: Aesthetics - The Science of Sense Perception The concept of "aesthetics" was coined by the philosopher Alexander von Baumgarten in the mid-eighteenth century. It is derived from the Greek aisthesis, meaning sense perception. For Baumgarten (1750), aesthetics had to do with the perfection of perception and only secondarily with the perception of perfection, or beauty. His new "science of sense cognition" was to occupy an intermediary rung, as a "science of the lower cognitive power" (sense perception) in contradistinction to "the higher cognitive power" (reason). By limiting aesthetics to the perception of the "unity-in-multiplicity of sensible qualities," as he put it, Baumgarten hoped to insulate it from being reduced to "arid" intellection. He believed that the intellect or "reason" was the poorer for the fact that it trafficked exclusively in "distinct ideas," as opposed to the "confused and indistinct ideas"’ which were the commerce of the senses. For Baumgarten, therefore, the aesthetic was rooted in the body - specifically, in the disposition to sense acutely - and involved attending to the nature of sensory experience in itself, rather than trying to intellectualize sensation (Gregor 1983: 364-65)
Baumgarten’s new "science" was quickly appropriated and just as quickly subverted by his contemporaries. They replaced his emphasis on the sensuous disposition of the aesthete with a taxonomy of "the five arts" (architecture, sculpture, painting, music and poetry). The scope and criteria of the various arts were delimited in terms of the dualism of vision (epitomized by painting) and hearing (epitomized by either music or poetry). The "dark" or "lower" senses of smell, taste and touch were deemed too base to hold any significance for the fine arts. Theatre and dance were also excluded on account of their hybrid character, since they played to more than one sense at once (see Rée 2000)
Baumgarten’s worst fears concerning the intellectualization of aesthetic perception were realized in Immanuel Kant’s Critique of Judgment (1790). Kant attempted to transcend the dualism of vision and hearing and replace it with a fundamental division between the "arts of space" (e.g. painting) and the "arts of time" (e.g. music), accessible to "outer intuition" and "inner intuition" respectively (Rée 2000: 58-60). It could be said that Kant rarefied aesthetics by divorcing it from perception and substituting intuition. After Kant, aesthetic judgment would be properly neutral, passionless and disinterested (see Turner 1994; Eagleton 1990). This definition of aesthetics guaranteed the autonomy of the enclave now known as "art" but at the expense of sensory plenitude.
The Disincarnation of Aesthetics In its modern incarnation (or more accurately, disincarnation), aesthetics has to do with the appreciation of the formal relations intrinsic to a work of art, and is divorced from that work's content. In one particularly influential characterization of the proper object of aesthetics, Robert Redfield offered the following analogy:
Art ... is like a window with a garden behind it. One may focus on either the garden or the window. The common viewer of a Constable landscape or a statue by St. Gaudens focuses on the garden. Not many people ... 'are capable of adjusting their perceptive apparatus to the windowpane and the transparency that is the work of art'" (Redfield 1971: 46 quoting Ortega y Gasset)
The implication is that it is only the uncommon viewer, capable of training his or her gaze on the windowpane itself, who is capable of enjoying a pure aesthetic experience, and of exercising the proper form of judgment.
Note how this characterization deflects attention from the focus on intersensory relations which was given in Baumgarten's concept of aesthetics as the perception of "the unity in multiplicity of sensible qualities." The idea of a windowpane effectively silences any input from the non-visual senses, which are deemed to constitute so many distractions. Aesthetic perception thus depends on a dual process of sensory demarcation and the elision of non-intrinsic sensations so that the viewer may come to appreciate the "organic unity" of a given work of art, be it a painting or a symphony. By way of illustration, consider the following passage from a Harold Osborne essay on "The Cultivation of Sensibility in Art Education"
In the appreciation of a work of art we concentrate attention exclusively upon a selected region of the presented world. When listening to music we shut out so far as possible the sounds of our neighbours' coughing, the rustle of programmes, even our own bodily sensations. When reading a poem, looking at a film or watching a stage play we tend to be imperceptive and unmindful of sensations from outside. But within the chosen sector we are alert to the intrinsic qualities of the sense-impressions imparted rather than to their practical implications and we are alert to the patterned constructs formed by the relations in which these intrinsic qualities stand to each other ... This is perception for its own sake, and represents the kernel of truth in the traditional formula 'disinterested interest' (Osborne 1984: 32; see further Elkins 2000: xi).
It should be noted that the aesthetic sensibility of which Osborne speaks is not only something that is cultivated by the viewer from within, it is also instilled in the viewer from without by the architectural design and codes of conduct (no touching, no chatting, no perfume, etc.) which obtain within the concert hall or art museum and other such spaces for the production of "single-sense epiphanies" (Kirshenblatt-Gimblett 1998; see further Drobnick 2004 on the sensory ideology behind the art museum as "white cube").
There have been numerous attempts in the history of Western art since the eighteenth century to escape the straitjacketing of the senses that followed the rarefication of aesthetics perpetrated by Kant and his contemporaries. The poet Baudelaire, father of the Symbolist Movement, celebrated the correspondence of the senses, rather than their separation. "Who has not known," Baudelaire proclaimed, "those admirable hours, veritable feasts of the brain, when one’s heightened senses perceive striking sensations ... when sounds ring out as in music, when colours speak, when perfumes tell us of worlds of ideas?" (whence his famous poem, "Correspondences"). Baudelaire's counterpart in the visual arts was Gustave Moreau, whose paintings became Symbolist icons on account of the way they evoked textures, sounds and scents through the representation of intricately patterned surfaces, musical instruments, and exuberant flowers (see Classen 1998: 109-18).
Marinetti and his Futurist confrères also experimented extensively with intersensoriality in their artistic works (see e.g. Marientti 1989, 2004), though the emphasis in Futrurism was more on producing a clash of sensations than the gently concatenating correspondences of the Symbolists (see Classen 1998: 126-31) Both of these movements were, however, eclipsed by the relentless drive toward abstraction in art that unfolded over the course of the twentieth century. Abstract or "non-representational" art is all windowpane, with no garden behind it - the logical culmination of the Kantian vision.
Healing Arts of the Shipibo-Conibo Indians of Peru In many non-Western societies, the aesthetic does not constitute a realm apart, but is rather an aspect of everyday and ritual practice, and the senses are not separated from each other but rather combine in specific ways to achieve specific purposes, such as healing. Consider the geometric designs of the Shipibo-Conibo Indians of Peru (see Figure 1).
Figure 1: Shipibo-Conibo Geometric Design
These designs - which are said to originate in the markings of the cosmic serpent, Ronin - are woven into textiles, incised on pots and houseposts, painted on faces, and even recorded in books (see Ventana). However, their foremost use is in the context of Shipibo-Conibo healing rituals
The Shipibo-Conibo understand medicine to be an art, literally, and their healing practices place a premium on synaesthesia in contrast to contemporary mainstream Western medical practice which is geared to the anaesthetization of the patient.
One important condition of [Shipibo-Conibo] therapy is the aesthetically pleasing [quiquin] environment into which the shaman and the family place the patient. He is carefully surrounded by an ambience designed to appease both the senses and emotions. Visible and invisible geometric designs, melodious singing, and the fragrance from herbs and tobacco smoke pervade the atmosphere, and ritual purity characterizes his food and each person with whom he has contact. The patient is never left alone in the mosquito tent during the critical time of his illness. This setting induces in the patient the necessary emotional disposition for recovery. But how is this indigenous concept of aesthetics [quiquin] to be understood? (Gebhart-Sayer 1985: 161)
The Shipibo-Conibo term quiquin, which means both aesthetic and appropriate, is used to refer to pleasant auditory and olfactory as well as visual sensations. Let us follow how the shaman operates with quiquin-ness on these three sensory levels - visual, auditory and olfactory - and how they are "synaesthetically combined to form a therapy of beauty, cultural relevance, and sophistication" (Gebhart-Sayer 1985: 162)
At the start of a healing session (there will be five such sessions in all), the shaman, under the influence of the ayahuasca hallucinogenic vine, sees the body of the patient "as if with an X-ray machine." A sick person's visual body pattern appears "like a very messy design," or mixed-up pile of garbage, and its pathological aura has a vile stench which is the mark of the attacking spirits (nihue) causing the illness. The healing ritual involves both the restoration of a healthy visual body pattern and the neutralization of the pathogenic aura through life-enhancing fragrance.
The shaman begins by brushing away the "mess" on the patient's body with his painted garment and fanning away the miasma of the attacking spirits with his fragrant herbal bundle, all the while blowing tobacco smoke. He then takes up his rattle and beats a smelling rhythm: the air is now "full of aromatic tobacco smoke and the good scent of herbs." Following this, the shaman, still hallucinating, perceives whole "sheets" of luminescent geometric designs, drawn by the Hummingbird spirit, hovering in the air, which gradually descend to his lips. On reaching his lips the shaman sings the designs into songs. At the moment of coming into contact with the patient, the songs once again turn into designs that penetrate the patient’s body and, ideally, "settle down permanently." However, the whole time the healing design is being sung onto the body of the patient, the nihue will "try to ruin the pattern by singing evil-smelling anti-songs dealing with the odor of gasoline, fish poison, dogs, certain products of the cosmetic industry, menstrual blood, unclean people, and so on" (Gebhart-Sayer 1985: 171), and thereby smudge or contaminate it. This is why it may take up to five sessions for the design to come out "clear, neat, and complete," and the cure to be finished. (If the design does not settle down permanently, the patient is unlikely to recover.)
Another strategy commonly employed by the evil nihue to prevent the cure from taking is to seek out the shaman's medicine vessel which contains all his design songs, and pry the lid off it. This causes the therapeutic power of the songs to escape. "This power is imagined as the fragrance of the design songs or the aromatic gas fizzing from fermenting yucca beer" (Gebhart-Sayer 1985: 172). The design songs thus have an olfactory dimension, in addition to their visual one, as their power is understood to reside in their fragrance.
The synaesthetic interrelationships of the designs, songs and fragrances used in Shipibo-Conibo healing rituals are nicely brought out in the following lines from a shamanic healing song:
The (harmful) spirit pneuma
swirling in your body's ultimate point.
I shall tackle it right now
with my fragrant chanting.
I see brilliant bands of designs,
curved and fragrant ...
(Gebhart-Sayer 1985: 172)
An important point to note here is that, whereas we perceive these designs as visual abstractions, the Shipibo-Conibo perceive them as matrices of intersensory perception, since these geometrical abstractions are at the same time musical scores and perfume recipes. They resonate in each of the senses at once. They are not simply addressed to the eye.