Richard Woodfield, Introduction to Gombrich on Art and Psychology more

Introduction: mapping the ground Richard Woodfield Among the books that Gombrich read while he was still at school was Max Dvo ák's Kunstgeschichte als Geistesgeschichte, soon after its publication. Gombrich later remarked that it had convinced him that `the art of the past offered an immediate and exciting access to the mind of bygone ages.'1 Ten years after reading Dvo ák he took a different view: the idea that art reflected the mind of the past accounted for the `social success of art history today, its receptivity to the art of all times and all peoples'. 2 Gombrich granted that its author had offered a persuasive case for studying art, but that his views on mannerism failed t he test of critical analysis. His change of mind had been brought about by his studentship, under Schlosser, at the University of Vienna's second institute of art history. Schlosser had been sceptical of the Geistesgeschichte approach and insisted on rigorous work in the archives, which could be the only point of entry into history: Art was a separate matter and the `language' of construction of visual imagery, Kunstsprache, something yet again. In preparing his doctoral dissertation for Schlosser on the architecture of Giulio Romano in the archives of Mantua, Gombrich qualified his earlier assessment of his hero's achievement. But it wasn't just a matter of the archives proving that Dvo ák's assessment of mannerism was wrong, it was that Gombrich had entered into a completely different arena for understanding the history of art. This was due to the prevalence of an interest in the methodological problems of art historical explanation, encouraged in Schlosser's special seminar. Explanation, as opposed to chronology, was an integral part of the grand tradition of the Vienna School of Art History, of which Schlosser was the lineal descendent. Perhaps its interest in the psychology of perception has been overemphasised to the exclusion of other matters, but it was psychology which was to gain Gombrich's everlasting fascination: `we all imbibed psychology with the milk of our Viennese Alma Mater.' He has told that particular story himself in his article `Art History and Psychology in Vienna Fifty Years Ago'. 3 Riegl's work had just been republished, Schlosser had an interest in the area, Loewy was testimony to the continuity of a particular kind of psychology and Sedlmayr had just published a book on Borromini's architecture using the new approach of gestalt psychology. Ernst Kris, one of Schlosser's ex-students, was both a curator at the Kunsthistorische Museum and a practising psychoanalyst, a member of Freud's circle. It was more or less inevitable, then, that Gombrich should turn his attentions to the work of Karl Bühler, who held the chair in psychology, follow his lectures and associate with his students. Bühler and Kris together were to have a profound effect on Gombrich's attitude to the history of art. 4 In his Bodonyi Review, written while he was still a student in the summer of 1932, Gombrich argued that any art historical analysis which used psychology should make its assumptions plain for detailed scrutiny; they could not simply be taken for granted. The Bodonyi Review was Gombrich's first published theoretical essay and entrance into methodological debate.5 It was on the use of the golden ground in late antique art, a subject 1. `Focus on Arts and the Humanities', Tributes, p. 14. 2. `Wertprobleme und mittelalterliche Kunst', Kritische Berichte 1937, translated as `Achievement in Medieval Art' in Meditations, p. 84. 3. Art Journal, 1984, pp. 162-4. 4. On Gombrich's work with Bühler see `Art History and Psychology in Vienna...' and Klaus Lepsky's contribution to this volume. On his work with Kris see `The Study of Art and the Study of Man: Reminiscences of Collaboration with Ernst Kris (1908-1975)' in Tributes. 5. "J. Bodonyi, Entstehung und Bedeutung des Goldgrundes in der spätantiken Bildkomposition (Archaeologiai Értesitë, 46, 1932/3)", Kritische Berichte zür Kunstgeschichtlichen Literatur, 5, 1932/33 1 raised by Riegl's Spätrömische Kunst-Industrie (1901). He argued that Riegl's interpretation of late antique art was open to criticism: Riegl treated its imagery as having naturalistic ambitions. The idea that visual imagery resulted from a naturalistic impulse in front of nature, though sweeping from a haptic to an optic pole of perception, failed to recognise the exceptional nature of the Greek Revolution in artistic production. 6 Egyptian and late antique art were fundamentally different in terms of their cultural roles and the demands they made upon the visual responses of their spectators. Gombrich took exception to the view shared by Riegl and Dvo ák that Art was an unchanging absolute. The stimulus to art's production was a historical variable, 7 not a constant Kunstwollen, and the language of visual imagery, its Kunstsprache, was just as historically and culturally specific as verbal language.8 There was a crucial difference between the classical artist's ambition to capture life and the late antique artist's creation of a pictograph: the same modes of visual analysis could not be applied to the two different forms of imagery. 9 Both Riegl and Dvo ák used outdated psychological theories in their analysis of late antique art. Both committed what Gombrich was later to call the physiognomic fallacy, seeing the style of a period as reflecting its Geist.10 He argued that expression was only achievable within a language and that a language could not have an expressive character of its own. At any given moment, the style in which a visual image was produced (that is, its overall characteristics which made it identifiable as being the product of a moment of culture and history) were the mechanics of its actual production. Style was not an abstract mirror to thought. Later, in his work in the Mantuan archives, Gombrich realised that a unified, and abstract, Geist did not lie behind the production of mannerist works of art; history was populated by people, with individual minds, inclinations and habits of work, not by abstractions. 11 Gombrich cited Bühler's work in the Bodonyi Review and Bühler's own concerns with expression had been fed from a number of different sources. Besides having close contacts with the gestalt school of psychology he was also very involved with linguistics. This resulted in, what could be called, a concern with the mind's products rather than with the mind per se. In his fundamental work Sprachtheorie (1934), anticipated by an earlier article `Die Axiomatik der Sprachwissenschaften' (1933), he was concerned to `clarify the capacities of language by glancing at other sign systems': Bühler... starts from the insight that there exists a spectrum, ranging from the extreme fidelity to nature exhibited by a waxwork, which (even so) resembles the model only relatively, to, for instance, a temperature chart, which merely records certain (published in 1935), pp. 65-75. 6. The exceptional nature of Greek art had been pointed out by Heinrich Schäfer, Von ägyptischer Kunst, third edition, Leipzig 1930. 7. He explored this idea further in his subsequent Garger Review, republished in Meditations as `Achievement in Medieval Art'. 8. On this subject see J. von Schlosser `"Stilgeschichte" und "Sprachgeschichte" der bildenden Kunst. Ein Rückblick' Sitzungberichte der phil. -hist. Abteilung der Bayerischen Akademie der Wissenschaften 1935, München 1935, Heft 1. 9. For a further discussion of this article see Richard Woodfield, `Gombrich, Formalism and the Description of Works of Art', British Journal of Aesthetics, 34 (1994). 10. For the fallacy see `Art and Scholarship' and `On Physiognomic Perception' in Meditations. Gombrich revisited the subject in Sense of Order where he says that, ultimately, anything may be made to fit with anything else. A friend's habitual expressions indentify her even when she might change her clothes in an atypical way: they become assimilated into one's thoughts about her identity (pp. 199-200). It should be observed in relation to what used to be called the new art history and is now called the new art criticism, that any argument based simply on appearance, rather than hard historical evidence, is bound to be psychologically compelling at the same time as having null explanatory value. This may be termed the `cicerone syndrome'. 11. `Focus on the Arts and Humanities' in Tributes, p. 14. 2 relationships in a given field. In between we find (if I may simplify and supplement his account a little), for instance, the notes of a musical score, the map, the landscape painting, and the illusionist backdrop of the stage as different but equally valid systems of signs. As you notice, this analysis no longer speaks of the question of the notion of the "conceptual image", because first we have to get our logic right before psychology can come into its own again. What is at stake is the notion of "relational fidelity", which is brilliantly explained in connection with black and white photography. 12 Gombrich was to pick up the strands from Bühler's work after the war, when working on Art and Illusion, but on leaving university he joined forces with Ernst Kris on two projects, first on expression and then on caricature. As Freud's disciple, Kris was keen to write a book on visual humour, an equivalent to Der Witz und seine Beziehung zum Unbewussten (1905). Besides addressing itself to the mechanics of visual satire, drollery, the grotesque, the cartoon and the caricature, the study was concerned with the question of why portrait caricature had emerged at such a late state (c. 1600) in the development of visual art. Kris, accepting Freud's Lamarkian account of the growth of civilisation, argued that it was not possible until humankind was civilised enough to have left the irrational fear of magic. The advent of war prevented the publication of the book 13 and the Holocaust put paid to ideas concerning the growth of human civilisation. But the project left Gombrich with a deeper interest in the mechanics of visual imagery, including the problem of likeness, which Kris had not addressed. How could it be possible that a few lines could bear such a striking resemblance to an individual to a point where one could see that individual in terms of those lines? Before the war, and as a result of Schlosser's insights, Gombrich had been struck by the recurrent use of the formula in artistic creativity: not just the medieval simile but also the schematic image in post-renaissance art. The war also increased his interest in the role of projection in perception, a subject which had already been explored by gestalt psychologists. 14 After the war, he honoured his promise to write The Story of Art. Building on his earlier theoretical work, he argued that different social function demanded different kinds of imagery. The majority of the world's artistic production was dominated by the need to produce magical substitutes for the figure: We remember how the primitive artist used to build up, say, a face out of simple forms rather than copy a real face; we have often looked back to the Egyptians and their method of representing in a picture all they knew rather than all they saw. Greek and Roman art breathed life into these schematic forms; medieval art used them in turn for telling the sacred story, Chinese art for contemplation. Neither was urging the artist to `paint what he saw'. This idea dawned only during the age of the Renaissance. At first all seemed to go well. Scientific perspective, `sfumato', Venetian colours, movement and expression, were added to the artist's means of representing the world around him; but every generation discovered `pockets of resistance', strongholds of conventions which made artists apply forms they had learned rather than paint what they really saw. The nineteenth-century rebels proposed to make a clean sweep of all these conventions; one after another was tackled, till the Impressionists proclaimed that their methods allowed 12. `Art History and Psychology in Vienna Fifty Years Ago', Art Journal, 1984, p. 164. 13. With Ernst Kris, Gombrich published `The Principles of Caricature', British Journal of Medical Psychology, 17 (1938) and Caricature, Harmondsworth 1940. His latest statement on the subject is `Magic, Myth & Metaphor: Reflections on Pictorial Satire', to be republished in my forthcoming collection of his essays. 14. See O. Renier and V. Rubinstein, Assigned to Listen, London 1986 which includes an extract from a memorandum concerning `listening' from Gombrich to his fellow listeners. 3 them to render on canvas the act of vision with `scientific accuracy'. 15 Modern art, by contrast, rejected that ideal. The artist was no longer concerned to produce naturalistic imagery because that function had been usurped by photography; other values came to the fore. It is a mistake to think that the cubists simply extended a line of development inaugurated by the Impressionists, their artistic concerns were entirely different. We need to separate out two phenomena: on the one hand the development of visual imagery, and the other conspicuous success in the use of that imagery. This is analogous to our drawing a distinction between the developments of languages and conspicuous literary success in those languages.16 Our own visual world is dominated by photography and we are aware of developments towards the creation of virtual reality.17 It is not inappropriate to see ourselves as inheriting the classical concern with the production of the naturalistic image, but that is a different matter from Art. It should be remembered that the account of the development of the arts of working with pigment, bronze and marble found in Pliny's Natural History was orientated towards techniques rather than aesthetic values. 18 The two great periods of visual discoveries about the world and its representation, classical antiquity and the Italian Renaissance, formed the focus of Gombrich's post war project which ultimately led to Art and Illusion. Naturalistic imagery was a problem worth addressing, particularly in the light of its rejection by contemporary artists. Gombrich's strong sense of a problem made him ask why its development was such a major human achievement if a child today could draw better than Giotto. In a proposal sent to Walter Neurath on 9th March 1947, Gombrich wrote: Any number of books have been written on pictorial art as a mode of expression, as creative activity or aesthetic experience. Without necessarily questioning the fruitfulness of this approach it may be asked whether the time has not come to investigate the realm of the image as such much in the way modern linguists have studied the functions of speech. The studies of Richards in Cambridge and the development of the Semantic approach, notably in the U.S.A. seem to hold out the promise that ultimately even the analysis of aesthetic values can but benefit from a clarification of these primary matters. It should perhaps be said from the outset that the development of Modern Art facilitates and stimulates such a fresh approach for the various symbolic aspects of art are, of course, much more manifest in Picasso than they were, say, in Manet.19 15. The Story of Art, First Edition, London 1961, pp. 421-22. It was disingeneous of M. A. Hagen to argue `Gombrich explicitly denies that his theory depicts art history as a development towards realism, but his treatment both of that history and of perception leads to that interpretation.' `A new theory of the psychology of representational art' in C. F. Nodine & D. F. Fisher (eds.), Perception and Pictorial Representation, New York 1979. 16. One of the drawbacks of the use of the word `art' is that it can be used as a blanket term to refer to all visual imagery, thus child art, chimpanzee art etc. There isn't a word which stands to art, in that blanket sense, as literature stands to language; children's literature refers, of course, to literature written for children, not by children. 17. Gombrich included photography in the fifteenth edition of The Story of Art, Oxford 1989. When he was asked why he had not included film he replied that he had not earlier included drama. Whether there have been any significant artistic achievements in holography is an interesting issue; there seems to be more potential for virtual reality. 18. I sympathise strongly with Göran Sörbom's point of view in `The theory of imitation is not a theory of art: Some stages in a conceptual tradition', J. Emt & G. Hermerén (eds.), Understanding the Arts: Contemporary Scandanavian Aesthetics, Lund 1992. I would not, however, want to encumber Gombrich with Schlosser's Croceian views on the subject. 19. Manuscript in the possession of Sir Ernst Gombrich, to whom I am indebted for its use. 4 He mentioned Picasso who, of all artists, demonstrated that style was a visual construction. But in The Story of Art Gombrich had also shown himself to be intrigued by Salvador Dali, whose painting: brings it home to us for the last time why it is that modern artists are not satisfied in simply representing `what they see'. They have become too aware of the many problems which are hidden in this demand. ... Dali's way of l etting each form represent several things at the same time may focus our attention on the many possible meanings of each colour and form - much in the way in which a successful pun may make us aware of the function of words and their meaning. 20 Dali's work had also intrigued Freud, who expressed an interest in exploring `the origins of a painting by him analytically', ie. through free association and memories like a dream. 21 But Gombrich chose a different approach: it could be analysed at the level of a joke. Dali's paintings played with visual puns and, like puns which are real linguistic discoveries, they render conspicuous the generative potential of visual configurations. 22 Dali had hit on the phenomenon later explored by Adelbert Ames: the fundamental ambiguity of the monocular view of a spatial configuration. It was with the problems of the visual image in mind that Gombrich wrote his Morris Review of 1949. 23 There he found Morris's theories fundamentally flawed on a number of counts, not least of which was his explanation of iconicity. If the iconic sign is defined as a symbol which looks like its subject then it still needs to be explained how that likeness occurs; it might, after all, be just by habituation. Is it just a matter of convention, for example, that a patch of paint in one of Guardi's paintings comes to look like a gondolier? Is there ever an image which is purely iconic, that just looks like what it is? Guardi depended upon projection for completion of the image: Guardi relies on the beholder's capacity to read `iconicity' into his sign. The contextual, emotional, or formal means by which this interpretation is evoked or facilitated - in other words, the relation between objective `iconicity' and psychological projection would have to form one of the main fields of study of a descriptive semiotic of the image. Perhaps it will show that what has been called the history of `seeing' is really the history of a learning process through which a socially coherent public was trained by the artist to respond in a given manner to certain abbreviated signs. 24 This idea is close to the suggestion, and could have been taken as meaning, that naturalistic imagery could be purely conventional. It is not surprising that some commentators, Nelson Goodman in particular, should have come to understand Art and Illusion in these terms. Two years after the Morris Review, in `Meditations on a Hobby Horse', Gombrich argued that the representational function of the image originated in substitution, not visual similarity, and that the child's hobby horse extended the class of `horse' by offering an object that was treatable as a horse. The hobby horse had two interesting distinct features: it was a viable substitute for a horse because it afforded riding and it only became desirable to ride because of the social prestige of the rider on horseback. The biological substitute of the cat's ball for a mouse can be translated, in more general terms, as the later Gibsonian idea of the object's 20. The Story of Art, p. 443. 21. Quotation and elaboration from E. H. Gombrich, `Verbal wit as a paradigm of art: the aesthetics theories of Sigmund Freud (1856-1939) in Tributes, p. 104. 22. This idea is developed at greater length in "The necessity of tradition: An interpretation of the poetics of I. A. Richards (1893-1979) in Tributes. 23. Reprinted in Reflections. 24. Ibid, p. 248. 5 offering visual information for use. It is both a product of nature, a stimulus to `see as', and culture, a social pressure for use. Given the fact that a hobby horse affords riding, or that a manikin may become the centre of ritual,25 the way is open to the user to increase the information offered by the stick or manikin to enhance its simulative capacities. Where would a horse be without its mane or Athena without her garments? Naturalistic imagery is, needless to say, even more complex than the primitive substitute. `Meditations...' originated the idea that `making comes before matching' and it also considered the way in which conventions, and the perception of pictures as visually constructed objects, come to play a role in picture perception. The extension of a class does not imply that the member of a class refers to a reality beyond itself;26 that is a new step in the representational process. The so-called `illusionist style' took that step; it not only suggested the possibility of creating an extension of the viewer's space but also created the possibility that an element within the picture field must signify a presence. The idea that a picture offers a window on to an imagined nature demands that: we cannot conceive of any spot on the panel which is not `significant', which does not represent something. The empty patch thus easily comes to signify light, air, and atmosphere, and the vague form is interpreted as enveloped by air. It is this confidenc e in the representational context which is given by the very context of the frame, which makes the development of impressionist methods possible.27 Guardi's patch is interpretable as a gondolier as a result of a complicity of understanding between the artist and his public over how his painted imagery actually works. In 1955, Gombrich presented his germinating ideas to the Durham conference of the British Psychological Society.28 He emphasised the nature of depiction as a sorting and classifying process, using the example of the mosaic as a constructed relational model. The idea of prototypes emerged to explain the use of the known to explore the unknown. While Gombrich was deeply interested the Whorf-Sapir hypothesis that language decisively shapes our conception of the world he stressed that the artist's visual vocabulary can also produce mismatches which offer the possibility of correcting drawings. The ultimately correct drawing is one which offers as many different interpretations as its model; such a drawing is possible when it is a drawing of a drawing, such as a Rorschach blot. Even so, Gombrich emphasised the importance of feed-back. The artist's depictions offer models for construing the visual world: a landscape can be seen as a Constable, and a finger print, as in a humorous drawing by Steinberg, can be seen as a Van Gogh. The choice of the title Art and Illusion for his Mellon Lectures `The Visible World and the Language of Art' did Gombrich a profound disservice. It invited two mistaken views: that the book was about art and an exercise in aesthetics, and that illusion was the criterion of successful art. In fact, as Gombrich has often said, visual imagery needs a lingu istics prior to its poetics. Furthermore, anyone with a smattering of knowledge of the history of art criticism will know 25. The allusion, here, is to the account of the emergence of Greek statuary offered by Jane Harrison, Ancient Art and Ritual, London 1913. 26. An idol of a god is not a representation of that god, it is the god; hence the rejection of idols in terms of their material characteristics: `That is not an idol it is only wood', wood not having the powers of the idol. There could be some interesting anthropological reflection on the difference between powers and affordances in this context, along the lines offered by D. Sperber, Rethinking Symbolism, Cambridge 1975. 27. `Meditations...', p. 10. There is an important exception to this: an inscription written across, say, a view of London is not taken to be an pobject hovering over London but simply read as an inscription. This exercise of mental set also conspicuously applies to comics: although the frame indicates spatial action it is also filled with script - two different mental sets come into play. 28. `Art history and the psychology of perception', unpublished manuscript. 6 that `sparrow aesthetics' has never been regarded as the criterion of a successful work of art.29 Gilpin, writing in the eighteenth cen tury, echoed general feelings about illusion when he wrote: But it is not under the idea of deception, that the real artist paints. He does not mean to impose upon us, by making us believe that a picture of a foot long is an extended landscape. All he wishes is, to give such characteristic touches to his picture, as may be able to rouse the imagination of the beholder. The picture is not so much the ultimate end, as it is the medium, through which the ravishing scenes of nature are excited in the imagination.30 Art and Illusion offers a series of theses about the mechanics of the construction of visual imagery rooted in the view that they constitute objective discoveries about the way in which the world's appearance may be simulated. While it may be a convention to want to depict the world naturalistically, and has had a minority appeal in the past until photography swept the face of the earth, operations within that convention are not, of themselves, necessarily conventions.31 It is not a matter of convention that one cannot see around corners, it is a matter of fact. If the cubists decide to represent what can be seen around corners, it is a matter of fact that they are not representing what can be seen from a stationary point of view. Picasso's cubist paintings will never represent the way that we see the world at a momentary point of time. Much of the misunderstanding of Art and Illusion has arisen from a failure to grasp the relationships between natural effects and conventional constraints in the perception of images. The question of whether illusionistic imagery is natural or conventional is a product of a habit of thinking in terms of polarities, of black and white, when it should be recognised that there is a range of grey in between: What we observe is rather a continuum between skills which come naturally to us and skills which may be next to impossible for anyone to acquire. 32 It might be thought, for example, that the outline figure is a matter of convention because outlines do not exist in nature: yet it turns out that the traditional view of the contour as a convention is based on an oversimplification. Things in our environment are indeed separated from their background, at least they so detach themselves as soon as we move. The contour is the equivalent of this experience. ... So important are these boundaries indicating what psychologists call `common fate' in features of the immediate environment that it has been shown that animals too respond to objects in outline as they do to their three dimensional prototypes. The equivalence is so obvious that no special learning appears to be required. 33 Does `no special learning' imply any learning at all? Does it not indicate an inbuilt disposition to respond to such outlines? Are there not inbuil t dispositions, like the disposition of a baby to 29. On the subject of `sparrow aesthetics' see now Goethe `On realism in art' (1798) in Goethe: The collected works, (ed.) J. Gearey, vol. 3, Princeton 1986, pp. 74-8. 30. William Gilpin Observations on the Western Parts of England, London 1798, p. 176 (Gilpin's emphasis). 31. The best philosophical discussion of this matter is Hilary Putnam, `Convention: A theme in philosophy', New Literary History, 13 (1981), pp. 1-14. 32. `Image and Code: scope and limits of conventionalism in pictorial representation', Image and Eye, p. 283. 33. Ibid. 7 recognise a smiling face, or to pick up a language? What are the limits of dispositional behaviour? Gombrich solved his early problem of projection by bringing into his account the new discoveries of ethology: the importance of trigger mechanisms in animal behaviour. Seeing faces in clouds, for example, is not like learning a language but is a product of biological conditioning. On the other hand, seeing the Indian in the white chalk drawing of a typically Indian profile (Philostratos's example), is a product of the cultivation of seeing what one can see in an image, which is a product of culture. 34 The naturalistic artist's use of increasingly sophisticated visual effects depends upon a culture of appreciation. A successful naturalistic image depends upon a range of perceptual triggers to gain the spectator's visual assent. It could be argued that Pliny's Natural History described the development of one particular series that he was able to identify, such as the use of three colours to generate the illusion of relief.35 The ability to depict a knee under a drapery is another such skill: At least from the sixth century onwards Greek art appeals to our imagination, by implying more than it can show. Every genuinely narrative illustration must be thus supplemented by us along several dimensions - by extending that of time we can see the figures move, by supplying that of space we can relate them to each other, and by projecting life into them we give their gestures and expressions meaning. ... ... It was at the same time that the sculptors' statues were seen to `come to life'. We sense the tension of the muscles under the surface, we see the play of the body under the garment, we feel the presence of a mind behind the smile. In discussing the illusions created by art, art historians (including this writer) have concentrated too much on the pictorial inventions of foreshortening, perspective, or light and shade, and failed to analyse the illusion of life that a Greek statue can give. 36 This question of illusion, and the relationship between stimulation and simulation, was one which Gombrich explored further in `Illusion and Art'. Gombrich had already raised the question of whether perceptual triggers were natural or a product of habituation in his Morris Review. His conclusion was that both elements entered into the production of illusionistic images. There are effects which compel visual reaction and there are also those which come into play once one has joined the game, like Guardi's patch. 37 In the context of talking about expectation Gombrich writes: Once more the effect experienced by the trained observer can be most conveniently imitated in the perception of images. It has been found in a well-known experiment that a familiar shape will induce the expected colour; if we cut out the shape of a leaf and of a donkey from identical material and ask observers to match their exact shade from a colour wheel, they will tend to select a greener shade of felt for the leaf and a greyer one for the donkey. We remember that the result of this experiment was anticipated by our ancient author Philostratus: `Even if we drew one of these Indians with white chalk.' Apollonius concludes, `he would seem black, for there would be his flat note and stiff curly locks and prominent jaw... to make the picture black for all who can use their eyes.' He was right. Interpreting, classing a shape affects the way we see its colour. Art and Illusion, pp. 189-90. This as a specialised skill which is the product of a habit of reading images for their potential information content. It is, however, a skill developed from the fact that one looks at a coloured world. I believe it would be true to say that when one misrecognises an object one misperceives its precise colour; recognition results in the correct colour perception - one perceives the colour of the bus as opposed to the colour of the robin. The subject in the psychological test and Apollonius's complicit spectator are both invited to form judgements about aspects of images which may simply not have crossed their minds. 35. Discussed by Gombrich in `The Heritage of Apelles' in Heritage. On modelling and highlights see `Light, Form and Texture in Fifteenth-Century Painting North and South of the Alps' in the same volume. 36. `The art of the Greeks' in Reflections, p. 16. 37. This subject is discussed in chapter seven of Art and Illusion, `Conditions of Illusion'. 34. 8 Triggers may stimulate, though they need not simulate. The phenomenon of imprinting in ducks `shows how far objective likeness can be dispensed with in certain circumstances'.38 The concern of the naturalistic artist is to invent triggers, or keys, to unlock sensory responses: foreshortening for depth, tonal contrast for modelling, highlights for texture `or the clues to expression discovered by humorous art'.39 In `Illusion' Gombrich considered the depiction of eyes. The artist's problem is `not necessarily to fashion a facsimile eye. It is to find a way of stimulating the response to a living gaze'.40 A variety of different conventions of depicting the eye have been invented to achieve this effect, Houdon's probably being the most remarkable. The difficult question concerns what we are actually seeing when we see a successful effect; we are certainly not seeing the object but are we seeing a phantom? This question opens up an enormous area for debate amongst perceptual psychologists in which Gombrich comes out against the empiricists on the one hand and J. J. Gibson on the other. In the course of his career, Gibson had found it growingly intolerable that there could be a distinction between appearance and reality and denied that we could actually experience visual fields. Insofar as we move around the world, our perception of it is self-correcting, enabling us to perceive its invariant structures, but we may not do this with pictures. Confronted by Gombrich's criticisms, Gibson had to admit that appearances may be seen in the world as they may in pictures: witness the Vault of Heaven. These, according to Gombrich, `form the limit of our visual world'. 41 The ambiguities of the world's appearance may be removed by movement but in painting, the ambiguities potentially in an image have to removed by the introduction of reinforcing cues: witness the modifications which are necessary to support a purely geometrical construction of perspective as unambiguously representing a particular natural space. True to Popper's falsificationism, Gombrich believes that there is always the possibility of error in perception, but nevertheless that error may be removed by further tests. The artist's process of making and matching reflects the process of testing and correcting. This `searchlight' theory of perception does not turn Gombrich into a constructivist, however; that represents a massive oversimplification of the problems involved. Monkeys do not make guesses when they leap from tree to tree and humans do not make inferences based upon their experiences: the classical distinction between sensation and perception is too crude to handle what happens. Psychology lacks an adequate concept of feedback, tied, as it is, to an eighteenth century language of experience and causality and, as I understand it, the computer modelling approach, introduced by Marr, is also very primitive, tied as it is to a picture theory of processing: it assumes what it tries to explain. Insofar as pictures are used in experiments to test visual processes and as they necessarily contain less information than objects in the real world, the role of conjecture is bound to play a greater role in their perception. Humans are propelled through a world of objects, not sensations, by a search for meaning which can, itself, condition the experience of pictures and other visual notations. The major discovery of Art and Illusion was that: the human imagination is a powerful thing when expertly manipulated. ... the less information is given, the more what I would call `the beholder's share' comes into play, provided of course the search for meaning is suitably guided. The visual in formation the painter can stimulate may never actually duplicate the information we pick up from solid objects close by. But is it not possible that he can mobilize the system to produce the same phantom sensations which come into play in those processes of search or probing for simplicity precisely in situations of inadequate information? If that is true, our perception of pictures would indeed differ from the perception of the visible world, but the right stimulation from the canvas may still engender a reaction similar to that 38. 39. 40. 41. Illusion, p. 200. Ibid, p. 201. Ibid, p. 205. `The sky is the limit', Image and Eye, p. 167. 9 which we experience in front of nature. A fine landscape or seascape by one of the Dutch masters certainly does not give me the illusion that the museum wall opens into parts of Holland. But I would claim that in getting absorbed in such a painting my search for meaning between and behind its brush -strokes weaves on its surface a rich fabric of uncontradicted sensations. 42 Gombrich's attitude to Gregory's kind of constructivism stems from his aversion to relativistic interpretations of perception.43 If one accepts the idea that all pictures offer insights into visual experience of the world or, as the philosophers put it, criteria of perception, 44 it would have to be accepted that the world was seen differently in the twelfth century fr om the way it is now. The difference in perception is then taken to account for the difference between their pictures and our own. But this is to make the same mistake as Riegl and Dvo ák, remarked on before, which is to assume a unity of pictorial purpose. Once it is recognised that pictures can serve a variety of purposes besides depicting a world a new type of analysis can emerge. Gombrich's work, then, is a form of historical psychology: it asks what may be thought of as being possible at a particular moment of art historical time. The historian has to ask what the devices were that were available to makers of visual imagery, and the range of their possible applications. A fundamental example of this is the potentialities of the painted and drawn mark, explored in Art and Illusion and then developed in `Light and highlights' in Heritage and `Watching Artists at Work: Commitment and Improvisation in the History of Drawing' in Topics. A further example is the use of the frame, used in connection with the p ainted mark in `Meditations' and then later, in connection with the development of narrative, in Means and Ends: Reflections on the History of Fresco Painting.45 Of course, the idea of the device is inseparable from that of the receptive viewer and this is a topic which is still foremost in his mind. `Mental set', or the beholder's share, plays a very important role in the perceptual process. As Gombrich suggested in a recent lecture at the Warburg Institute: `I believe that the difference between signs and images rests in the different mental sets we have to adopt for their understanding'.46 At this point we may return to a remark made by Gombrich in his Durham Lecture of 1955: `One of the things, I believe to have learned, is that it makes very little sense to speak of "seeing the world" or any such generality. We can study perception only through actions or reactions in given situations'.47 In Art and Illusion he reformulated this idea as: `The test of the image is not its lifelikeness but its efficacy within a context of action.' 48 One is invited 42. Ibid, p. 171. 43. This opens a complex argument, as many of Gombrich's comments do. The relativism remark was made in conversation with Richard Gregory on a radio programme `Seeing and knowing' made to celebrate Sir Ernst's 85th birthday. It can be pursued in the difference of opinion aired in Illusion: `I am not quite happy with the suggestion made by Gregory that representations should be classified as "impossible objects" objects, that is which give us contradictory impressions at the same time. ... what we see is not a turning fork of impossible shape but a very possible drawing on paper.' (pp. 236-7) This leads into the question of differences in mental set that are invited by different kinds of drawings, which takes us back to the difference between illusionist and non-illusionist imagery. 44. Wittgenstein's remarks have caused a great deal of debate in this context: What is the criterion of the visual experience? - The criterion? What do you suppose? The representation of `what is seen'. (Philosophical Investigations, Oxford 1963, 198e.) Amongst others, this has been picked up by R. Wollheim, `On drawing an object' in On Art and the Mind, Harvard 1974, and M. Wartofsky, Models: Representation and the Scientific Understanding, Dordrecht 1979. 45. London 1976. 46. `Signs and Images', manuscript of a lecture given at the Warburg Institute on 8.6.94, pp. 10-11, courtesy of Sir Ernst. The interest in mental set runs throughout his work and is to be found in Art and Illusion, Illusion and Image and Eye. 47. Unpublished manuscript, p. 5. 48. Art and Illusion, p. 94. 10 to consider what spectators would have made of the images that were produced for them, what purposes they fulfilled. By recognising the role of `mental set' in picture perception we realise that not all pictures of things are intended to be seen as records of visual experiences of those things. There are important differences between looking at naturalistic pictures and looking at objects in the world but there are dangers in going too far in asserting the extent to which naturalistic pictures are governed by conventions.49 In a very real sense we look into a Leonardo drawing of a face and we look at a typographic symbol for one ( ); we recognise them both as faces, but expect animation from the first and not from the second. 50 How, then, do we look at this symbol ( ) in relation to Shadow Antiqua script? We look into a photograph of a city and at a map;51 a photograph invites a point of view, a map doesn't. We gain an impression, right or wrong, of what a city is like from a photograph but not from a map. The information potentials all vary in character depending on the requirements placed upon the constructed image. This idea of visual processing goes back to Bühler's principle of abstractive relevance:52 features of a configuration become meaningful in terms of their information potential within a given situation. One may wonder what the information potential of a given situation is. In a world of objects, it is their availability for use. In a world of visual imagery, it is again their availability for use. But life gets very difficult when one tries to specify what those uses might be. One looks for different significant features in the sculptures of the founders at Naumburg Cathedral,53 armorial bearings,54 confraternity decorations55 and paintings like Masaccio's Holy Trinity,56 Raphael's Madonna della Sedia57 and Dutch landscapes. 58 Most obvious of all, one is not expected to look at flowers in wall paper as one looks at them in a still-life.59 49. This theme is handled in much greater depth by other contributors to this volume. 50. In this context `The Image Family' in J. -P. Sartre, The Psychology of Imagination, London 1978, pp. 17-62 repays reading. 51. Gombrich's example, Illusion, pp. 225-8. For an extended discussion of the difference between images and maps see `Mirror and Map: Theories of pictorial representation" in Image and Eye. 52. Discussed by Gombrich in `The use of art for the study of symbols', American Psychologist, 20 (1965), pp. 34-50. 53. As Gombrich and Kris discovered in their work on the sculptures at Naumburg Cathedral, it was sufficient that they should be seen as lifelike: their expressive features were actually indeterminate, open to a wide variety of readings. Tributes, pp. 226-7. 54. One does not attribute deep spiritual states to an armorial bearer: The eyes gaze into the distance, they stand in a face that bears the marks of hard experience. This man is no longer a wild adventurer, he is sensitive to the sufferings destiny has laid upon him; it is with sorrow that he awaits the next test, though he is sure that he will win through in the end. `The priority of context over expression' in C. S. Singleton (ed.) Interpretation. Theory and practice, Baltimore 1969, p. 71. 55. See `Tobias and the Angel' in Symbolic Images: not a narrative, or an illustration of a biblical story, but an image of the Archangel Raphael with his attribute Tobias. On the difference between imagini and storie see C. Hope, `Religious Narrative in Renaissance Art', Journal of the Royal Society of Arts, 134 (1986), pp. 804-18. 56. Means and Ends: Reflections on the History of Fresco Painting, London 1976: `what it represents within this highly convincing religious chapel is not an imagined reality but a purely symbolic image, a reminder of a doctrine which cannot be visualised at all. The real presence in art of the First Person of the Trinity in so tangible a form is not easy to reconcile with the Decalogue, but it d oes not seem to have disturbed Masaccio's contemporaries.' (p. 41) 57. `Raphael's "Madonna della Sedia"' in Norm and Form, pp. 66-7: `If we were really meant to relate it to an image of everyday life, we might feel this relationship is false and prettified. It is this feeling which presents the greatest obstacle in our day to the understanding of Raphael's achievement.' (p. 67) 58. See `Mapping and painting in the Netherlands in the seventeenth century' in Reflections, pp. 121-2. 59. Sense of Order, pp. 34-5. The history of mural painting can be seen in terms of the tensions between 11 The question which Gombrich started to address in ` Icones Symbolicae. The visual image in neo-Platonic thought'60 concerned the status of what had hitherto been regarded as symbols in Baroque art: what is the difference between looking at peace personified and the personification of peace? From my own experience, I would suggest that it is the kind of attention that one gives to the various aspects of the image. 61 While Rudolph Arnheim, following Picasso's dealer Daniel Kahnweiler, predicted that the works of Picasso, Braque or Klee would begin to `look exactly like the things they represent',62 the same kind of point can be made again. It is a mistake to think that Picasso intended to offer pictures of reality in his analytic cubist images: they were intended to disrupt normal perceptual processes in front of the object by offering conflicting spatial clues in the depicted object. Synthetic cubist constructions, on the other hand, play ed with the differences between pictorial ground (the `invisible' primed surface), real space (the actual occupied space of a newspaper or piece of veneer) and pictorial space (a depicted guitar, for example). These subtleties of pictorial appeal were appreciated by Picasso's patrons as part of of an avant garde game which was being played by a small number of artists, centred in Paris. 63 If the world actually did look like cubist paintings, we would have enormous difficulty in getting around it and if the world looked as if it was depicted in twelfth century pictures, archers would have had no difficulty in shooting their prey around corners. If naturalistic art is concerned with a moment of vision, it must involve itself with movement as well as space: While the problem of space and its representation in art has occupied the attention of art historians to an almost exaggerated degree, the corresponding problem of time and the representation of movement has been strangely neglected.64 Following Hildebrand's concern with the representation of movement, Wickhoff's interest in narrative and his own early explorations of gesture and expression, Gombrich has turned this into another one of his major areas of exploration. Actions take place through time and a naturalistic visual image captures only a moment. Two obvious questions follow: whether the moment is necessarily a slice of time like a film frame and whether a slice of an expressive gesture forms the foundation of representation of states of mind. To the first question, the answer is that the staged moment, like a film still, could well stray beyond the slice of time to embody a full sense of the action; thus there is something suspect about Lessing's characterisation of visual imagery as simply an art of space.65 To the second question, Gombrich suggests that one might look in the direction of `ritualised gesture' to find a repertoire decoration and illusion; on which, see Means and Ends. 60. Extended in its republication in Symbolic Images. 61. This is another case of `mental set'. A point of entry may be found in considering one's responses to a `documentary' which turns out to be a fiction. A few years ago, on April Fools' Day, British television broadcast a hoax about astronauts' discoveries on the other side of the moon. The film had my absolute attention until I realised that it was a hoax: what was different was the kind of attention which I paid to detail. This is rather different from what Roland Barthes described in `The reality effect', reprinted in T. Todorov (ed.), French Literary Theory Today, Cambridge 1982. Sartre's Psychology of Imagination, might offer some point of entry here as well. More work needs to be done on this; I have started to address the issue in "L'effet du réel: an alternative account" in a special issue of the Polish philosophy journal Dialectics and Humanism, 15 (1988), pp. 63-73. 62. Art and Illusion, p. 22. 63. In this respect their point was completely missed both by R. Arnheim (Art and Visual Perception, London 1956) and M. Hagen (Varieties of Realism, Cambridge 1986) who followed him. 64. "Moment and Movement in Art", in Image and Eye, p. 40. This applies as much to still life as it does to the portrayal of events. The `play' of light is, after all, time based. 65. `Moment and movement in art', `The Mask and the Face: The Perception of Physiognomic Likeness in Life and in Art' and `Action and Expression in Western Art' in Image and Eye. 12 for the expression of mental states.66 The idea of a repertoire highlights a confusion in understanding Gombrich's views on expression: one which he set out to redress in `Four theories of artistic expression'. 67 The idea of a repertoire may characterise either the range of expressions depicted by an artist in an image or by an artist's own expressive range. The idea that it is artists' business to express themselves is a legacy of the Romantic view of art. It is significant that while Gombrich has developed a psychological approach to understanding visual imagery his own approach, despite his work with Ernst Kris, has never been psychologistic.68 He is unconcerned with artists' inner states in giving accounts of intentions69 and is preoccupied by the problem of how visual imagery may be seen to work. This critical stance in relation to psychologism comes out most clearly in his essays on Freud.70 In a letter to Hermann Struck, Freud declared that his essay on Leonardo `does not in the least aim at making the great man's achievements intelligible.'71 Freud's essay on Leonardo72 is well known as an attempt to explore the mentality of the man through his imagery; but he described it himself as `half novelistic fiction ( Romandichtung)'.73 The historian's business is to ask how Leonardo's imagery would have been understood by his contemporaries: at least one had a problem with his St. Anne.74 The first step towards clarification is to determine the representational status of the image: `The traditional group ... [of the Virgin and Child with St. Anne] had never been conceived as a realistic represetation' let alone an expression of Leonardo's state of mind. Leonardo, unlike contemporary artists, wasn't paid to paint his states of mind and no -one had the slightest interest in them. 75 The image is simply a reworking of a traditional motif rendered more complex by Leonardo's interests in naturalistic depiction. While other commentators on Freud have had recourse to his work on dreaming, Gombrich regards his ideas on jokes as more appropriate to an analysis of the workings of visual imagery.76 A joke depends for its effectiveness on the resources of language as such; creativity in image making is, in turn, based upon the resources which emerge out of the traditions of image making. As language is generative and can operate by way of feedback, so too may the visual image. The problem of feedback is a mighty one. Once a work of art has been created it may serve a variety of different functions,77 which may compete with each other in terms of the 66. "Ritualized Gesture and Expression in Art" in Image and Eye. 67. Architectural Quarterly, 12 (1980), pp. 14-19. 68. It must at the same time be said that he has never subscribed to behaviourism, indeed this is one of the issues that he raises in the Morris Review. 69. On this subject Gombrich's remarks are scattered through his writing. See, in particular, `The psycho-analytic approach' in `Aims and limits of iconology' in Symbolic Images, pp. 17-18. 70. Though see also his important Ernest Jones Lecture `Psycho-analysis and the history of art' reprinted in Meditations on a Hobby Horse. 71. Quoted by Gombrich in `Seeking a key to Leonardo' in Reflections, p. 66. 72. See volume 14 of the Pelican Freud Library Art and Literature, Harmondsworth 1985. 73. In the letter to Struck, again, as cited by Gombrich in `Seeking a key to Leonardo'. 74. Fra Pietro da Novellara; see `Aims and limits of iconology' in Symbolic Images, p. 16. 75. On this subject see A Lifelong Interest, pp. 157-9. 76. See `Verbal wit as a paradigm of art: the aesthetic theories of Sigmund Freud' in Tributes. 77. In replying to Peter Burke's suggestion that art history should concern itself with the social functions of images, Gombrich argued: ... one of the things we have learnt from psychoanalysis is that what is successful in society will have many functions at the same time. ... Most things in society ... fulfil many functions. The number of specialised tools is very small, and art certainly belongs to those institutions which meet many demands at the same time. (p. 883) `Ernst Gombrich discusses the concept of cultural history with Peter Burke', The Listener, 27th December 1973, pp. 881-3. 13 appreciation and future development of the work. Nowhere is this problem better illustrated than in Gombrich's hugely complex and ambitious study The Sense of Order. The creation of substitutes is only one dimension of image making activity, the desire to ornament is another. Ornament is generative and plays on the human interest in enriching and extending the visual field. It has quite different characteristics from picturing: `Painting, like speaking, implicitly demands attention whether or not it receives it. Decoration cannot make this demand. It normally depends for its effect on the fluctuating attention we can spare while we scan our surroundings'.78 While we might enjoy examining a Persian carpet, we don't believe that it is necessary to examine its every detail. The mechanism which facilitates this response is our sense of order.79 If a key idea behind Art and Illusion was `making before matching', the one behind this book is `groping comes before grasping or seeking before seeing'. 80 While the picture offers us the delight of recognition, the Persian carpet cultivates the pleasures of `anticipation and memory' which `may largely be subliminal, but they are none the less real.'81 Gombrich had thought of calling his book The Unregarded Art. It is central to his theory of perception, however, that the search for meaning plays a key role. Now meaning is one of those multifacetted phenomena which can spill into imagery in a variety of different ways: one has only to remember gold as a visual metaphor of value to realise how complex the situation might be. Colours and shapes may possess more than formal values. It is for this reason that The Sense of Order is a polyphonic text, including such chapters as `Ornament as Art', `Towards an Analysis of Effects', `The Psychology of Styles', `Designs as Signs' and `The Edge of Chaos', the latter chapter concerned the dissolu tion of figurative images and the play of grotesques and the fantastic. The art of ornament may reach the heights of control, in the Alhambra or Book of Kells, or regress into fantasy. The tendency to regression forms a focus for another of Gombrich's central interests, which is now, after a lifetime's gestation, emerging as a book on the preference for the primitive. Like other members of the Vienna school, Gombrich has been deeply interested in so-called artistic `declines'; this was marked both by the Bodonyi Review and his doctoral dissertation on Giulio Romano's architecture. In the following Garger Review he maintained that a decline of interest in naturalism had: facilitated the emancipation of formal values. ... in as far as the recognizability of symbols is not compromised and the sign remains a sign, primitive predelictions may be allowed free rein. This applies to the pure use of precious colours in medieval illumination as much as to that ornamental elaboration of the whole work which leads to such high decorative achievement.82 In his later essay `Visual Metaphors of Value in Art' he returned to the theme of the golden ground. What had been taken for a valuable addition to t he artist's armoury in late antiquity had become a problem for Alberti, in the fifteenth century. Not only did he see it as an obstacle to a managed depiction of space, disrupting the carefully wrought devices of one point linear perspective, its use posed moral problems as well: it was felt to have a different moral significance. In his Ernest Jones lecture `Psychoanalysis and the History of Art' he argued that contemporary primitivism was a product of forms of sophistication and refinement: ... a compensation, a redistribution of psychological gratifications, must also take place 78. The Sense of Order, p. 116. 79. Gombrich didn't feel that the central thesis of The Sense of Order had been sufficiently appreciated and so he offered a resumé in his preface to the second edition (1992). 80. Ibid, p. 5. 81. Ibid, p. 103. 82. `Wertprobleme un mittelalterliche Kunst', Kritische Berichte, 1937, translated as `Achievement in Medieval Art' in Meditations, p. 74. 14 during the post-Bougereau period. ... Impressionsim succeeded in excluding literary association and in confining the give and take to the reading of the scrambled colour-patches. But in return for this effort of shared activity, it yields a wonderful premium of regressive pleasure. Impressionism stood between the pursuit of appearances and `an openly regressive art, of primitivism.'83 It was Cicero who had remarked: For it is hard to say why exactly it is that the things which most strongly gratify our senses and excite them most vigorously at their first appearance, are the ones from which we are most speedily estranged by a feeling of disgust and satiety.84 Gombrich's reading of ancient rhetorical theory convinced him that there was there a model for the explanation of the development of a primitivising style. 85 It was a theory of literary effects which could be applied with equal validity to the visual arts. He subsequently ma de use of that model in accounting for the distinctive nature of Giulio Romano's deployment of the rustic style of architecture.86 His 1971 lectures `Ideas of Progress and their impact on art' 87 focussed particularly on changes which occurred in the eighteenth century, in the rejection of Rococo. Blake, for example, objected to Reynolds' admiration of works by painterly artists: clean lines can come to have a moral significance superior to decadent conspicuous brush strokes. It is obvious that this was a matter to be explained not by changes in perception but in the values placed upon perceived features of visual imagery: a matter for the social or cultural, as opposed to the perceptual, psychologist. At this point there is an obvious link into his work on `T he logic of Vanity Fair'88 and his account of the historical development of ornament, which in turn, harks back to his fascination with the golden ground and visual metaphors of value in art. Social, or cultural, psychology also found an outlet in his important essay `Botticelli's Mythologies. A study in the neo-platonic symbolism of his circle'.89 But in this book it would be appropriate to turn to the matter of the relationship between image and text as yet another application of the idea of mental set. Earlier interpreters of Botticelli's Primavera read into the appearance of the central figure features corresponding to their anticipations of what the painting depicted 90 and the occasion of its use. The important methodological point is that context dictates the expressive possibilities of the image, not just in terms of the expressions of the individual figures, but the expressive effect of the image as a whole. Thus it is important to grasp the connection between the image and its predecessors in courtly tapestries and celebratory religious images to see its novelty as an image.91 83. `Psychoanalysis and the History of Art', Meditations, p. 41. 84. De Oratore, III, xxv. 98, trans. H. Rackham, London 1948. 85. `The debate on primitivism in ancient rhetoric', Journal of the Warburg and Courtauld Institutes, 29 (1966), pp. 24-37. 86. `Architecture and rhetoric in Giulio Romano's Palazzo del Te', reprinted in New Light. 87. `I. From Classicism and Romanticism', `II. From Romanticism to Modernism' (privately circulated); these have not yet been made publically available in English. 88. Reprinted in Ideals. 89. Reprinted in Symbolic Images. 90. Footnote 23 of the article offers a fascinating microstudy of the varieties of readings, from pensiveness to laughter, with pregnancy and consumption en route. 91. See the discussion of innovation and refinement in `The Necessity of Tradition' in Tributes, pp. 206-7 and the prior discussion of Michelangelo's Moses, pp. 200-2. Is novelty something that one may see or does one simply see that something is novel, or see something novel in some thing? How would a psychologist respond to the question? 15 The theatre-goer who thinks that King Lear is a comedy would rather miss the point of the play; recognition of its genre enables the spectator to appreciate its subtleties. Gombrich argues that genre plays a similar role in imagery: getting the genre wrong is the first step to misunderstanding the image. In his unpublished book on secular iconography Gombrich wrote `H. Wölfflin's carefully balanced formula of historical approach to art `Not everything is possible in every period' applies to iconographical no less than to stylistic questions.' Raphael would not have illustrated Shakespeare and Simone Martini would not have painted a beggar-boy; `To tell a story like the supposed "paternal advice" was simply "not possible" in Terboch's day and could therefore not be the true interpretation of the picture'.92 A book on the history of genres could be written out of Gombrich's published essays. It could start with the legacy of medieval imagery in Raphael's Stanza della Segnatura and Tobias and the Angel by a follower of Verrocchio. It could look at the rise of the portrait, 93 the landscape,94 and the emergence of mythological painting.95 Some serious thought could be given to the emergence of the programme in Renaissance art.96 It could consider the idea of imagery as environment and illustration in the middle ages and also the emergence of anecdotal imagery in the nineteenth century which witnessed a reverse in the relation of image to text: from text to image to image to text.97 It could also consider the way in which a particular theme within traditional art gains its manifold resonances in the same way as different performances of the same piece of music.98 All of this has taken us a very long way from the issues raised in Art and Illusion, but it does raise a very important point for the perceptual psychologist. Traditional psychology worked upon the false assumption that looking at things was like looking at pictures, so pictures could be used unproblematically in experiments on perception. The assumption was compounded by the belief that looking at things in the world was simply a product of processing retinal images, again like looking at pictures. But, as Gombrich pointed out, there a re fundamental differences between looking at objects and looking at pictures of them - and there are also fundamental differences between looking at different kinds of pictures. Psychologists are still exploring the perceptual issues which Gombrich raised. But it seems to me that the area between psychology and linguistics, whose subject is the visual image, needs something better than contemporary popular semiotics to deal with it. Gombrich's critique of Morris still stands and his use of Bühler has resulted in great gains but, as he has frequently said, it is not method which offers a way forward but a sense of the problems which need to be solved.99 In his review of Art and Illusion, Gibson concluded: The discoveries of painters have been far more elaborate than the discoveries of psychologists, if less rational, and Gombrich shows that they are at least potentially investigable. The student of perception is tempted to limit his research to what he can control by the methods he has been taught. This book will widen his horizon and stimulate his 92. Secular Iconography, ms. dated 26.11.41 and held at the Warburg Institute; I am indebted to Sir Ernst for its use. 93. `Giotto's Portrait of Dante?', reprinted in New Light. 94. `The Renaissance Theory of Art and the Rise of Landscape' reprinted in Norm and Form. 95. `Botticelli's Mythologies: A Study in the Neo -Platonic Symbolism of his Circle' republished in Symbolic Images. 96. It would start with the Introduction to Symbolic Images and then consider the issue behind `Botticelli's Mythologies' and `Raphael's Stanza della Segnatura'. 97. The transitions involved are discussed in `Action and Expression in Western Art', reprinted in Image and Eye. 98. This idea was mentioned first in `Achievement in Medieval Art', loc. cit., and then reiterated much later in `Painted Anecdotes', reprinted in Reflections. 99. See, for example, `Approaches to the History of Art: Three Points for Discussion', reprinted in Topics. 16 ambition.100 I think that Gibson's forecast was correct. 100. American Journal of Psychology, 73 (1960), pp. 653-4 (p. 654). 17