LEFT, RIGHT, AND ART
Gravity provides humans with an unambiguous basis for distinguishing up and down. Although Wu’s cobalt-60 experiments revealed standards for labeling left and right, left and right are still confusing mirror images in physical space. Are differences in pictorial space confusing in the same way? Is there a left and right in pictorial art, and, if so, are they related to left and right in the brain?
Aestheticians have frequently asserted that left and right in a picture are absolutes.24 Wölfflin and Gaffron have both emphasized, for example, that mirror-reversing a painting often drastically alters its meaning; thus they claimed that many of Rembrandt’s etchings can be understood only by looking at their mirror image. For at least some paintings, this reversal effect is indeed striking. For example, mirror-reversing Janssens’s Reading Woman (figure 6.7) clearly alters the composition.25 Similarly, the effect of reversing Brueghel’s Parable of the Blind is to change the impression of the painting so that “instead of tumbling into the ditch after their leader, [the blind men] seem to be pressing upon him.”26
If, as aestheticians say, mirror reversal so changes the meaning of a painting, why have so many artists, from Raphael and Rembrandt to Munch, remained apparently indifferent to the reversal of their originals when reproduced as prints or tapestries?27 And why, conversely, did a few, such as Dürer and Van Gogh, take great care to etch originals in their mirror image?
In fact, objective studies involving a number of observers and different paintings have lent little support to the generality of the claims of art historians that mirror-reversing paintings consistently changes the content or tone of the original. For example, when children and young adults were asked to tell whether they preferred the original or the mirror-image views of a series of classical (e.g., Raphael, Rubens, and Rembrandt) or modern paintings (e.g., Mondrian, Picasso, and Degas), preference for the original view was slight or nonexistent.28 Similarly, in two other studies subjects could not accurately remember the left-right orientation of a set of representational pictures.29
Figure 6.7
Pieter Janssesn Elinga, Die lesende Frau. Bayerische Staatsgemäldesammlungen, Alte Pinakothek Munich. Left: Veridical reproduction. Right: Mirror-image reproduction. The viewer’s position relative to the perspective of the floor boards seems to change in these two versions. In the original, the viewer stands naturally at the woman’s back looking over her shoulder; in the mirror-reversed version the viewer must change position to effect such an identification. In the veridical version the woman, in the left pictorial space, is important; in the mirror-reversal her slippers assume salience out of proportion, i.e., they are nearer and more conspicuous. Indeed pictorial depth is greater upon mirror-reversal. Gaffron (1950) suggested that these and other phenomenal differences arising from mirror-reversal evidence the existence of an unconscious central visual process or glance curve in pictorial perception. Thus, the slippers assume great prominence and appear closer because they stand at the head and in the path of the glance curve.
A possible explanation for the failure of experimental psychologists to find the perceptual differences between paintings and their mirror reverses claimed by aestheticians might be that the psychological experiments involved collections of both symmetrically and asymmetrically organized compositions; by contrast, aestheticians exemplify their point with highly asymmetric paintings. Highly asymmetrical paintings, with marked perspective and lighting differences between the two sides (such as figure 6.7), clearly do alter on reversal. However, for naive adults viewing most paintings, as for children learning to read or for octopuses discriminating obliques, mirror images tend to be seen as equivalent. In all these cases the confusion or equivalence of mirror-reversed images represents a fundamental mode of visual analysis originating in organic evolution.
Independent of the difficulties people usually experience in telling left from right in nature or detecting mirror-image reversals of works of art, there appear to be verifiable differences associated with the left and right halves of physical space and pictorial space. Many aestheticians have suggested that the right visual field or objects located there tend to be perceived as heavier, larger, more distant, brighter, and more conspicuous, but less textured and less clear, while the left pictorial field or objects located there tend to have the opposite attributes.30 Several of these claims have been verified experimentally. Specifically, under controlled conditions with a number of subjects, objects on the left have been found to appear closer and clearer and content on the right to appear heavier.31
BRAIN MECHANISMS OR ARTISTIC CONVENTIONS?
Asymmetries of pictorial space could arise from asymmetries of the brain or from cultural conventions. We suggest that both contribute to the anisotropy of art but in different ways.
As discussed above, the left hemisphere of the brain is principally concerned with the perception and production of linguistic material, and the right cerebral hemisphere is principally concerned with the perception and discrimination of nonverbal and spatial material. When visual input is strictly lateralized, and the eyes are not free to roam, the right half of visual space is processed primarily by the left hemisphere, and the left half of visual space is processed primarily by the right hemisphere (figure 6.5). Strict asymmetric input is, of course, not the case in artistic experience, since one does not just fixate the center of a painting. However, both Kinsbourne and Gur have demonstrated that visuospatial thinking and verbal thinking (or anticipation of these types of thinking) activate processes in the right and left hemispheres respectively and, at the same time, excite correspondingly lateralized eye movements.32 Apparently, differential processing of information between the hemispheres is so basic as to maintain lateral biases even in the absence of strictly lateralized input, hence the importance of cerebral lateralization to art.
Anisotropies of visual space that are universal in humans probably reflect the influence of lateralized brain functions rather than cultural convention. One artistic asymmetry that appears to be universal in this way is profile orientation. Portraits are rarely full-face; one study found that of 1,474 painted portraits produced in Western Europe between 1500 and the present, a majority face leftward.33 Similarly, right-handed children and adults of both sexes have a strong tendency (74% of 9,874) to draw profiles facing leftward.34 This was found in the United States and Norway (where reading is from left to right), Egypt (where reading is right to left), and Japan (where reading is from top to bottom and right to left). By contrast, left-handed children were equally likely to orient their profiles in either direction, presumably reflecting the heterogeneous nature of left-handers and demonstrating that profile orientation is not a simple function of how the hand holds a pencil. Leonardo da Vinci, perhaps the most famous left-handed artist, preferred to draw right-facing profiles. An examination of the profiles in the Dover edition of Leonardo da Vinci’s Notebooks shows that most of the profiles therein face right (figure 6.8).35
Figure 6.8
Leonardo da Vinci’s Heads of Girls and Men ca. 1480 from the Royal Library at Windsor (No. 12276 v).
Profile orientation appears to be a function of laterality, not direction of reading, age, or sex. When a face is fixated centrally, the half of the face in the left visual field is processed by the right hemisphere (figure 6.5). As noted above, face recognition is largely a right-hemisphere function, and, when right-handed people look at the two halves of a front view of a face, the half of the face in the left visual field looks much more “like the person” than the other half.36 Thus the tendency for portraits to locate profiles in the left visual field presumably reflects the fact that facial information there would be perceived more readily and accurately by the majority of people (i.e., right-handers). Similarly, as shown in figure 6.9, it is the expression on the half of the face in the left visual field that usually determines the right-handed viewer’s impression of it.
In contrast to profile orientation, other aspects of visual a
nisotropy appear to reflect cultural conventions. Wölfflin, and others, have suggested that individuals typically enter a picture at the left foreground and proceed along a specified path or “glance curve” into the depth of the picture and over to its right-hand side.37 They point out how this directional scan lends an aesthetic dimension of movement in graphic art. Movement from left to right in a painting is easier and faster, while movement from right to left is slower and perceived as having to overcome resistance. The former signals attack or approach; the latter withdrawal: in addition, the / diagonal is often associated with ascent and triumph, while is associated with descent and defeat. Poussin’s The Rape of the Sabine Women (figure 6.10) epitomizes the use Western artists have made of these associations. The painting is a single dramatic frame out of an episode of the Roman legend. Figure 6.11 indicates the diagonals at work in the painting. The strong diagonal (figure 6.11, bottom) begins at the left with the entrance of the Roman fasces-bearer below (and Romulus above) and is continued in the Roman rout of the Sabines, which proceeds off the canvas to the right. The weak diagonal (figure 6.11, top), mainly focused in the father’s futile attempt to save his daughter’s honor in the lower right, is overwhelmed. These crossed diagonals represent opposing forces and express the theme of conflict and counterconflict in a classical manner (the theme originally derives from the metopes in the Parthenon); they inwardly organize a scene otherwise emotionally explosive and chaotic. Wölfflin believed that the left-to-right glance curve represented a fundamental aesthetic vector. However, the glance curve in Asian art appears to be in the opposite direction. Compare the directional selectivity of the Poussin with that evident in a detail from a Chinese handscroll—opened and read from right to left (figure 6.12). Here the direction of viewing, first to the lower right and then toward the upper left, is compelling.
Figure 6.9
Stare at the nose of each face. Which looks happier? Jaynes (1976) found that most righthanders choose the bottom face, with the smile in their left visual field, presumably because the smiling side is processed by the right hemisphere on central fixation. Copyright © 1976, 1990 by Julian Jaynes. Reprinted by permission of Houghton Mifflin Harcourt Publishing Company.
Similar aesthetic vectors seem to operate in stage direction and audience expectation in the theater. According to Dean, the right stage (audience left) in Western theater is strong and elicits audience attention, so, as the curtain rises at the beginning of an act, the audience can be seen to look to the left front.38 In Chinese theater, contrariwise, the important positions are to the audience right. Thus the direction of the glance curve in both painting and theater appears to be a cultural convention, presumably related to the direction of reading, left to right in the West and right to left in the East. As Gaffron says explicitly, “we ‘read’ a picture in a certain way just as we read a page of a book.”39
The term “glance curve” may be a misnomer, however, since studies of eye movements across both Eastern and Western pictures do not reveal glance curves in either direction.40 Rather such studies suggest that the eye roams over a picture in an arbitrary manner, only stopping to rest on salient features. The glance curve may be some kind of covert cognitive scanning with its direction set by reading habits. Or, alternatively, it may reflect a cultural organizing principle implicit in graphic art. Whether in observers or in pictures, conventions like the glance curve prevent representational art from approaching excessive abstraction with its multiple and ambiguous points of view.41 If there were no such conventions about left and right to show the way, artists and observers, like children, might be confused as to how to view a picture.
Figure 6.10
The Rape of the Sabine Women by N. Poussin (1639). Metropolitan Museum of Art, New York.
Figure 6.11
Weak and strong diagonal in Poussin’s The Rape of the Sabine Women. Top: the “weak” diagonal (lower right to upper left). Bottom: The “strong” diagonal (lower left to upper right).
Figure 6.12
Twin Pines against a Flat Vista, by Chao Meng-Fu (1254–1322). Metropolitan Museum of Art, New York. Detail of a Chinese handscroll. Such handscrolls typically show dominant diagonals that follow the direction they are opened and read: right to left.
POSTSCRIPT
In this 1978 paper we suggested that the well-known phenomenon of confusion of lateral mirror images (e.g., b and d ) should not be conceived as a perceptual limitation but as an adaptive mode of processing: since in the natural world lateral mirror images are almost always views of the same object it would be adaptive to treat them as equivalent rather than distinguish them. That is, mirror-image “confusion” is not a confusion but an efficient perceptual constancy (unless you are a human learning to read).
This speculation has received considerable support since we first proposed it. First, visually responsive single neurons in inferior temporal cortex (crucial for visual pattern recognition) respond similarly to members of a lateral image pair,42 an equivalence that may provide the basis of lateral mirror image confusion. Second, monkeys with lesions of inferior temporal cortex have great difficulty in distinguishing different patterns but they can distinguish members of a lateral mirror image pair as well as normals. This may be because lateral mirror images are easier to distinguish if mechanisms that specify their equivalence are lost. Since the animals with inferior temporal lesions have lost the mechanisms for equating lateral mirror images, they can learn to discriminate them. By contrast, normal animals treat mirror images as equivalent and therefore have difficulty in learning to tell them apart.43 Third, in functional magnetic resonance images of human brains, lateral mirror images are treated as identical stimuli.44 Similarly, in studies of visual search lateral mirror images are treated as equivalent.45
Another suggestion we made was that “dyslexics should be superior to normal children in those perceptual and perhaps artistic skills that do not involve language.” In spite of the long list of great artists such as Leonardo and Picasso who were supposedly dyslexics,46 there were actually very little adequate data on the artistic ability of dyslexics for the reasons we listed. Since then there has been at least one careful study of this question that did not have the weaknesses of previous studies. This involved the comparison of students at an elite university art school with comparative students not studying art.47 There was a much higher incidence of dyslexics among the art students, confirming our prediction.
Since this article was written there have been several good popular treatments of right and left in “brain, bodies, atoms and cultures”48 and of left-handedness.49
NOTES
This chapter consists of an article published in the MIT journal Leonardo (11: 29–38 [1978], “Left and Right in Science and Art”), with minor edits and a new postscript. The co-author, Marc Bornstein, is now Senior Investigator and Head of the Section on Child and Family Research at the National Institute of Child Health and Human Development.
1. Leibnitz’s critique of Newton occurred in his running controversy with Samuel Clarke, an English philosopher and divine acting as Newton’s spokesman. See particularly Leibnitz’s third letter to Clarke in Morris, 1951.
2. Handyside, 1929. For a discussion of the relation between mirror images and Kant’s ideas of space see Remnant, 1963.
3. William James (1890) held a view somewhat similar to Kant, but he preferred the term “sensations” to “intuitions.” “We can only point,” he said, “and say here is right and there is left, just as we should say this is red and that is blue.” However, James thought Kant incorrect in invoking the structure of space as a referent and thought the structure of the body to be sufficient.
4. This has been termed the “Ozma problem” by Martin Gardner, after Project Ozma, the first major attempt to develop a method of extragalactic communication. Ozma is the queen of the mythical kingdom in L. F. Baum’s The Wizard of Oz. For discussion of this and many other problems of right and left in science, see Gardner, 1969; Adams, 1969
; and Whyte, 1975.
5. For a nontechnical account of the fall of parity and its significance, see Wisner, 1965. There is a “meta-Ozma” problem: there may be pockets of antimatter in the universe and in these the results of the cobalt-60 experiment would be reversed. Thus, to communicate extragalactically about right and left it may also be necessary to know if our communicants are made of matter or antimatter. On the other hand, it may be possible to send a circularly polarized radio or light signal that would communicate the meaning of “right” and “left” even to an antimatter world without describing Wu’s experiment.
6. Mach, 1914.
7. The classic animal and child studies are Sutherland, 1960, and Rudel and Teuber, 1963. Comprehensive reviews include Tee and Riesen, 1974, and Corballis and Beale, 1976. See also Bornstein, Gross, and Wolf, 1978.
8. Orton, 1937.
9. Corballis and Beale, 1976.
10. Critchley, 1953.
11. The following provide reviews and extensive bibliographies on hemispheric asymmetries: Schmitt and Worden, 1974—see particularly articles by Milner, Sperry, Berlucchi and Teuber; Kinsbourne, 1976, particularly articles by Levy and Trevarthen; Dimond and Beaumont, 1974; and Harnard et al., 1977.
12. See, e.g., Allman and Kaas, 1971; Gross, 1998a, figure 5.6.
13. Allman and Kaas, 1971; see, e.g., Brooks and Jung, 1973; Zeki and Sandeman, 1976.
A Hole in the Head Page 12