1989 - Seeing Voices

by Oliver Sacks

  Thus there develops in signers a new and extraordinarily sophisticated way of representing space; a new sort of space, a formal space, which has no analogue in those of us who do not sign. 104

  104. There may be other ways of establishing such a formal space, as well as a great enhancement of visual-cognitive function generally. Thus with the spread of personal computers in the past decade, it has become possible to organize and move logical information in (computer) ‘space,’ to make (and rotate, or otherwise transform) the most complex three-dimensional models or figures. This has led to the development of a new sort of expertise—a power of visual imagery (especially imagery of topological transforms) and visual-logical thinking which was, in the pre-computer age, distinctly rare. Virtually anyone can become a visual ‘adept’ in this way—at least, anyone under the age of fourteen. It is much more difficult to achieve visual-computational fluency after this age, as it is much more difficult to achieve fluent language. Parents find again and again that their children can become computer whizzes where they cannot—another example, perhaps, of ‘critical age.’ It seems probable that such enhancement of visual-cognitive and visual-logical functions requires an early shift to a left hemisphere predominance.

  This reflects a wholly novel neurological development. It is as if the left hemisphere in signers ‘takes over’ a realm of visual-spatial perception, modifies it, sharpens it, in an unprecedented way, giving it anew, highly analytical and abstract character, making a visual language and visual conception possible. 105

  105. Novel—yet potentially universal. For as in Martha’s Vineyard, entire populations, hearing and deaf alike, can become fluent native signers. Thus the capacity—the neuronal apparatus—to acquire spatial language (and all the nonlinguistic spatial capacities that go with this) is clearly present, potentially, in everyone.

  There must be countless neuronal potentials that we are born with which can develop or deteriorate according to demand. The development of the nervous system, and especially of the cerebral cortex is, within its genetic constraints, guided and molded, sculpted, by early experience. Thus the capacity to discriminate phonemes has a huge range in the first six months of life, but then becomes restricted by the actual speech to which infants are exposed, so that Japanese infants become unable, for example, to discriminate anymore between an ‘I’ or an ‘r,’ and American infants, similarly, between various Japanese phonemes. Nor are we short on neurons; there is no danger that developing one potential will ‘use up’ a limited supply of neurons and prevent the development of other potentials. There is every reason to have the richest possible environment, linguistically as well as in every other way, during the critical early period of brain plasticity and growth.

  CORRECT SPATIAL LAYOUT

  DISTORTED SIGNED SPATIAL LAYOUT

  CORRECT SIGNED SYNTAX

  Figure 3. A massive lesion in the right cerebral hemisphere destroys Brenda I.’s ability to ‘map’ on the left side, but not her ability to use syntax. Figure (a) shows the actual layout of Brenda’s room, as it would be correctly signed. Figure (b): In describing her room, Brenda leaves the left side bare, and (mentally) piles all the furniture in the right side of the room. She can no longer even imagine ‘leftness.’ Figure (c): But in her actual signing, Brenda uses a full space, including the left side, to represent syntactic relations. (Reprinted by permission from What the Hands Reveal About the Brain, H. Poizner, E.S. Klima, & U. Bellugi. The MIT Press⁄Bradford Books, 1987.)

  One must wonder whether this linguistic-spatial power is the only special development in signers. Do they develop other, nonlinguistic, visual-spatial powers? Does a new form of visual intelligence become possible? This question has led Bellugi and her colleagues to launch a fascinating study of visual cognition in deaf signers. 106

  106. Bellugi et al., 1989.

  They compared the performance of deaf, natively signing children with that of hearing, non-signing children on a battery of visual-spatial tests. In tests of spatial construction, the deaf children did much better than the hearing ones, and indeed, much better than ‘normal.’ There were similar findings with tests of spatial organization—the ability to perceive a whole from disorganized parts, the ability to perceive (or conceive) an object. Here again, deaf four-year-olds did extraordinarily well, getting scores that some hearing high school students could not match. With a test of facial recognition—the Benton test, which measures both facial recognition and spatial transformation—the deaf children were again markedly ahead of the hearing children, and far in advance of their chronological norms.

  Perhaps the most dramatic test results have come from deaf and hearing children in Hong Kong, where Bellugi has investigated their ability to perceive and remember meaningless Chinese ‘pseudo-characters’ presented as swift patterns of light.

  Here the deaf, signing children did startlingly well—and the hearing children were almost unable to do the task at all (see fig. 4). The deaf children, it seems, were able to ‘parse’ these pseudo-characters, to achieve a very complex spatial analysis, and this enormously facilitated their powers of visual perception, enabling them to ‘see’ the pseudo-characters at a glance. Even when the experiment was repeated with deaf and hearing American adults who had no knowledge of Chinese characters, the deaf signers did notably better.

  These tests, in which signing children perform far above normal levels (a superiority that is especially marked in the first few years of life), all emphasize the special visual skills learned in acquiring Sign. As Bellugi notes, the spatial organization test involves not only the recognition and naming of objects, but also mental rotation, form perception, and spatial organization, all of which are relevant to the spatial underpinnings of Sign syntax. The ability to discriminate faces, and recognize subtle variations of facial expression, also carries extreme importance to the signer, since facial expression plays an important role in ASL grammar, 107

  107. This linguistic use of the face is peculiar to signers, is quite different from the normal, affective use of the face, and, indeed, has a different neural basis. This has been shown very recently in experimental studies by David Corina. Pictures of faces, with expressions that could be interpreted as ‘affective’ or ‘linguistic’ were presented, tachistoscopically, to the right and left visual fields of deaf and hearing subjects. Hearing subjects, it was apparent, processed these in the right hemisphere, but deaf subjects showed predominance of the left hemisphere in ‘decoding’ linguistic facial expressions (Corina, 1989).

  The few cases studied of the effects of brain lesions in deaf signers upon facial recognition show a similar dissociation between the perception of affective and linguistic facial expressions. Thus, with left hemisphere lesions in signing subjects, the linguistic ‘propositions’ of the face may become unintelligible (as part and parcel of an overall Sign aphasia), but its expressiveness, in the ordinary sense, is fully preserved. With right hemisphere lesions, conversely, there may be an inability to recognize faces or their ordinary expressions (a so-called prosopagnosia), even though they are still perceived as ‘propositionizing,’ fluently, in Sign.

  This dissociation between affective and linguistic facial expressions may also extend to their production: thus one patient with a right hemisphere lesion studied by Bellugi’s group was able to produce linguistic facial expressions where required, but lacked ordinary affective facial expressions.

  TARGET STRUCTURE

  POINTING LIGHT MOTION

  DEAF CHINESE CHILDREN

  HEARING CHINESE CHILDREN

  Figure 4. Asked to reproduce a Chinese pseudo-character (presented as a moving point-light display), deaf Chinese children do extremely well, and hearing Chinese children extremely badly. (Reprinted by permission from ‘Dyslexia: Perspectives from Sign and Script,’ U. Bellugi, Q. Tzeng, E.S. Klima, and A. Fok. In A. Galahurda, ed. From Reading to Neurons. The MIT Press⁄Bradford Press, 1989.)

  The ability to separate discrete configurations, or ‘fr
ames,’ from a continuous stream of movement (as was done with the Chinese pseudo-characters) brings out another important ability of signers—their enhanced power of ‘movement parsing.’ This is seen as being analogous to the ability to break down and analyze speech from a continuous and ever-changing pattern of sound waves. All of us have this ability in the auditory sphere—but only signers have it so dramatically in the visual sphere. And this too, of course, is essential to the comprehension of a visual language, which is extended in time as well as in space.

  Is it possible to detect a cerebral basis for such enhancements of spatial cognition? Neville has studied the physiological correlates for such perceptual changes, by measuring changes in the brain’s electrical responses (evoked potentials) to visual stimuli in particular, movements in the peripheral visual field. 108 (Enhanced perception of such stimuli is crucial in Sign communication, for the signer’s eyes are generally fixed on the other’s face, and signing movements of the hands therefore lie in the periphery of the visual field.)

  108. For an overview of Neville’s work, see Neville, 1988, and Neville, 1989.

  She has compared these responses in three groups of subjects: deaf native signers, hearing non-signers, and hearing native signers (usually born of deaf parents).

  Deaf signers show greater speed of reaction to these stimuli—and this goes with an increase of evoked potentials in the occipital lobes of the brain, the primary reception areas for vision. Such increases of speed and occipital potentials were not observed in any of the hearing subjects, and seem to reflect a compensatory phenomenon—the enhancement of one sense in place of another (greater auditory sensitivities, similarly, may occur in the blind). 109

  109. The ancient insight that the loss of hearing may cause a ‘compensation’ of sight cannot be ascribed simply to the use of Sign. All deaf—even the postlingually deaf, who stay in the world of speech—achieve some heightening of visual sensibility, and a move toward a more visual orientation in the world, as David Wright describes:

  I do not notice more but notice differently. What I do notice, and notice acutely because I have to, because for me it makes up almost the whole of the data necessary for the interpretation and diagnosis of events, is movement where objects are concerned; and in the case of animals and human beings, stance, expression, walk, and gesture…For example, as somebody waiting impatiently for a friend to finish a telephone conversation with another knows when it is about to end by the words said and the intonation of the voice, so does a deaf man—like a person queuing outside a glass-panelled call-box—judge the moment when the good-byes are being said or the intention formed to replace the receiver. He notices a shift of the hand cradling the instrument, a change of stance, the head drawing a fraction of a millimeter from the earphone, a slight shuffling of the feet, and that alteration of expression which signals a decision taken. Cut off from auditory clues he learns to read the faintest visual evidence (Wright, 1969, p. 112).

  A similar acuity may also occur, and persist, in the hearing children of deaf parents. Thus in the case described by Arlow (1976):

  The patient would look intently at his parents’ faces from early childhood on…[He] became extremely sensitive to intentions and meanings which can be communicated through expressions on the face…Like his [deaf] father, he was particularly sensitive to people’s faces and could make good judgements about the intentions and sincerity of those with whom he was engaged in business…[he] felt that in ordinary business negotiations he had a serious advantage over his opposite numbers.

  But there were also enhancements at higher levels: the deaf subjects showed greater accuracy in detecting the direction of motion, especially when the movement was in the right visual field, and coincident with this was an increase in evoked potentials in the parietal regions of the left hemisphere. These enhancements were also observed in the hearing children of deaf parents and have therefore to be seen not as an effect of deafness, but as an effect of the early acquisition of Sign (which demands very superior perception of visual stimuli). It is not only detection of motion in the peripheral field that is shifted, in signers, from being a right hemisphere to a left hemisphere function. Neville and Bellugi obtained evidence—indeed, quite early on—for a similar left hemisphere specialization (and shift from the ‘normal’ right hemisphere specialization) in deaf signers for picture identification, dot localization, and the recognition of faces. 110

  110. Neville and Bellugi, 1978. It should not be supposed that all visual-cognitive processing in deaf signers is transferred to the left hemisphere. The disturbing (even devastating) effects of right hemisphere lesions on signing make it clear that this hemisphere is equally crucial for some of the visual-cognitive abilities underlying the capacity to sign. S.M. Kosslyn has recently suggested that the left hemisphere may be better at image generation, and the right hemisphere at image manipulation and transformation (Kosslyn, 1987); if this is so, lesions in opposite hemispheres may differentially affect various components of the mental imagery, and mental representations of space, in Sign. Bellugi and Neville are planning further studies to see if such differential effects (both in simple perceptual tasks and in complex forms of imagery) may indeed be found in signers with damage to one or the other hemisphere.

  But the very greatest enhancements were observed in deaf signers—and in these, intriguingly, the enhancement of evoked potentials spread forward into the left temporal lobe, which is normally regarded as purely auditory in function. This is a very remarkable and, one suspects, fundamental finding, for it suggests that what are normally auditory areas are being reallocated, in deaf signers, for visual processing. It constitutes one of the most astonishing demonstrations of the plasticity of the nervous system, and the extent of its adaptability to a different sensory mode. 111

  111. Although Neville, thus far, has obtained only electrophysiological evidence for such reallocation (neuro-imaging, PET scan studies, are planned) striking anatomical evidence for this has recently been obtained. Thus if newborn ferrets are centrally deafened (by cutting fibers to the chief auditory nuclei), many normally auditory pathways and centers are modified, and become exclusively visual in morphology and function (Sur et al., 1988).

  Such a finding also raises fundamental questions as to the extent to which the nervous system, or at least the cerebral cortex, is fixed by inborn genetic constraints (with fixed centers and fixed localization—areas ‘hardwired,’ ‘preprogrammed,’ or ‘pre-dedicated’ for specific functions) and to what extent it is plastic and may be modified by the particularities of sensory experience. The famous experiments of Hubel and Wiesel have shown how greatly the visual cortex may be modified by visual stimuli, but leave unclear how much input merely kindles built-in potentials, and how much it actually shapes and molds them. The experiments of Neville suggest a shaping of function to experience—for it can hardly be supposed that the auditory cortex has been ‘waiting’ for deafness, or visual stimulation, to become visual and change its character. It is very difficult to explain such findings except by a radically different sort of theory, one that does not see the nervous system as a universal machine, hardwired and preprogrammed for (potentially) everything, but sees it as becoming different, as free to take on completely different forms, within the constraints of what is genetically possible.

  To comprehend the significance of these findings one also needs a different way of looking at the cerebral hemispheres and their differences and their dynamic roles in dealing with cognitive tasks. Such a way has been provided by Elkhonon Goldberg and his colleagues in a series of experimental and theoretical papers. 112

  112. These include Goldberg, Vaughan, and Gerstman, 1978; and Goldberg and Costa, 1981. See also Goldberg, 1989.

  Classically the two cerebral hemispheres are seen as having fixed (or ‘committed’) and mutually exclusive functions: linguistic⁄nonlinguistic, sequential⁄simultaneous, and analytic⁄gestalt are among the dichotomies suggested. This view runs into obvious diff
iculties when one confronts a visuo-spatial language.

  Goldberg would first enlarge the domain of ‘language’ to one of ‘descriptive systems’ in general. Such descriptive systems, in his formulation, constitute superstructures imposed on elementary ‘feature detection’ systems (for example, those of the visual cortex), a variety of such systems (or ‘codes’) being operative in normal cognition. One such system is, of course, natural language; but there may be many others—such as formal mathematical languages, musical notation, games, etc. (insofar as these are encoded by special notations). It is characteristic of all of these that they are first approached in a tentative, groping way but later acquire an automatic perfection. Thus there may be with these, and with all cognitive tasks, two ways of approach, two cerebral ‘strategies,’ and a shift (with the acquisition of skill) from one to the other. The right hemisphere’s role, as thus conceived, is critical for dealing with novel situations, for which there does not yet exist any established descriptive system or code—and it is also seen as playing a part in assembling such codes. Once such a code has been assembled, or emerged, there is a transfer of function from right to left hemisphere, for the latter controls all processes that are organized in terms of such grammars or codes. (Thus a novel linguistic task, even though it is linguistic, will initially be processed predominantly by the right hemisphere, and only subsequently become routinized as a left hemisphere function. And a visuo-spatial task, conversely, even though it is visuo-spatial, will, if it can be embedded in a notation or code, come to show a left hemisphere superiority.) 113