At the turn of the millennium, Ernst Gombrich published a new preface to Art and Illusion. Unlike its predecessors it did not simply update his argument with new research but instead made a defence of his entire project. He argued that there was a fundamental difference between reading a text and responding to a figurative image and it lay in the different mental sets required for the two activities. Furthermore, he argued that in an age of new visual technologies it was absurd to argue that the illusionist image was a myth and there were not ways in which it could become instrumentally better. The difference between the early photograph and virtual reality lay in objective discoveries in the area of visual simulation. This line of argument was prompted by a body of criticism of his work that had emerged in England and America in the 80s, sparked offby Norman Bryson's book The Vision and the Gaze:
What the ancient Greeks called mimesis (the imitation of nature) has proved a difficult task: it took the artists of the ancient world some 250 years of systematic research to achieve this end, and artists of the Renaissance took the same time before they were able to eliminate what Albrecht Dürer called "falseness" in pictures.
...this commonsense interpretation of the history of Western art has recently been attacked on the ground that the whole idea of mimesis, truth to nature, is a will-o'-the-wisp, a vulgar error. There never was an image that looked like nature; all images are based on conventions, no more and no less than is language or the characters of our scripts. All images are signs, and the discipline that must investigate them is not the psychology of perception - as I had believed - but semiotics, the science of signs.1
To explain the difference between reading a text and responding to a figurative image, Gombrich invoked the notion of mental set. Now, while the Anglophone reader might assume that 'mental set' was a straightforwardly psychological concept, many European readers would see it as central to a semiotic tradition of thought. As Ernst was fond of saying, European scholars regarded him as a semiotician; in which case they would regard the preface as a defence of semiotics. 2 Perhaps the problem is compounded by the fact that while the subtitle of his book is A Study in the Psychology of Pictorial Representation, the Mellon Lectures, on which the book was based, were called 'The Visible World and the Language of Art'. In the main body of its text, Gombrich wrote:
Just as the study of poetry remains incomplete without an awareness of the language of prose so the study of art will increasingly supplement an inquiry into the linguistics of the image. Already we see the outlines of iconology, which investigates the function of images in allegory and symbolism and their reference to what may be called the 'invisible world of ideas.' 3
In his earlier review of Charles Morris's book Signs, Language and Behavior he had written:
The distinction between poetry and language has always been accepted as natural; the distinction between art and imagery is only gradually becoming familiar. Mr Morris himself stresses the need for more descriptive studies on visual signs.... It is all the more a pity that he does not seem to have taken cognizance of the emerging discipline of iconology which must ultimately do for the image what linguistics has done for the word. ... The work of such pioneers in the study of the symbolic aspects of the image as A. Warburg and E. Panofsky are absent.4
Thus it might be felt that Gombrich would have saved everyone a great deal of confusion if he had written a book called Iconology and the Linguistics of the Image, in which he had combined material from both Art and Illusion and Symbolic Images. However, if he had done that, he would probably not have had a best-seller on his hands. That book would have been a product of the confluence of two traditions: Viennese sematology and warburgian iconology.
Gombrich was always somewhat reticent about his connections with sematology, which in its Viennese incarnation originated with the work of Heinrich Gomperz and culminated in that of Karl Bühler. Neither was mentioned in Art and Illusion, though they did emerge in an essay 'Art History and Psychology in Vienna Fifty Years Ago'.5 He also acknowledged Bühler's influence in a letter published in Leonardo in which he tried to rebut David Carrier's claim that his 'illusionistic' theory could be contrasted to Bryson's semiotic approach:
... it really will not do for Professor Carrier to construe a convenient contrast between my alleged "notion of art as illusion" and Bryson's reliance on semiotic theories. In a recent number of the Art Journal (Summer 1984, p.164), I acknowledged the early influence which Karl Buehler's Sprachtheorie (1934) had on my development; witness also my extensive review of Charles Morris, Signs, Language and Behavior in the Art Bulletin of 1949 (pp. 68-75). If the reader cares to look at p.120 of Art and Illusion he will find me making use of the linguistic notion of `distinctive features' in a discussion of board games, which "allow us to study articulation, the creation of distinctions without the intrusion of the problem of likeness or representation". I have continued these trends of thought in various papers of my volume The Image and the Eye (1982)...6
No one seems to have read that letter and Carrier did not pay it any serious attention. However, Gombrich's acknowledgment of Bühler's influence was only partial. Any reading of Art and Illusion that ignores its relation to Viennese sematology and the example set by Gombrich's acknowledged mentors is bound to be partial, if not blind.
Viennese sematology
Gombrich's problem was born in 1905 with Heinrich Gomperz's article 'Ueber einige psychologische Voraussetzungen der naturalistischen Kunst', published in the Munich newspaper Beilage der Allgemeinen Zeitung.7 The article was prompted by the exhibitions in Germany of new contemporary art, in particular French impressionism. It addressed the question of how a naturalistic painting came to be taken as the picture of its subject. It was a long and interesting discussion that started with the Christian complaint against pagan worship of images, dwelt on the different mentalities of iconoclasts and iconodules and the role of imagination, beliefs and feelings in the development of the figurative image from its aniconic predecessor. It discussed the transition from image magic to the magical image and the aniconic image to the naturalistic image, using the hobby horse to explain the origins of art. He argued that with the development of naturalism, the image lost its magical efficacy, and the centrality of the subject of art was replaced by the importance of the artist. He formulated the problem of representation, how a material thing comes to be regarded as an image of something, that was picked up and answered by Karl Bühler. Gomperz concluded his essay by announcing a forthcoming publication, 'an entire system of theoretical philosophy', that would study 'the relationship of ideas to objects and of speech to meaning'. It was duly published as Weltanschauungslehre between 1905 and 1908.
Gomperz did not, however, attempt to explain how figuration itself developed. The archaeologist Emanuel Löwy had undertaken that task in his book Die Naturwiedergabe in der älteren griechischen Kunst in 1900. It is highly likely that both men knew of each other's work through mutual acquaintance with Sigmund Freud and Julius von Schlosser, who was himself interested in psychological issues.8
The next major development came in 1933, when Karl Bühler published 'Die Axiomatik der Sprachwissenschaften',9 acknowledging the significance of Gomperz's Weltanschauungslehre and focussing on his treatment of the problem of representation. How can the actor Mr. Bassermann be Wallenstein?:
'he is something and yet he is not that.' ... The spectator takes the costume and gestures, the words and deeds of Bassermann the individual as something that enables him to experience the Wallenstein created by the writer. ... Bassermann puts at the disposal of the Wallenstein of the writer the previously mentioned elements so that the figure created by the writer can come to appearance.
... If a canvas covered with spots of color awakens in the viewer the charming appearance of a landscape, this state of affairs is, according to Gomperz, also describable with the aid of the same formula: the canvas lends its accidents, so to speak, to the ... imaginally presented thing.10
The business of sematology would be to investigate how this happened. It will build upon the results of sociology, psychology and linguistics but its insights 'must derived constructively by reliance on a model, just as the propositions of mathematics are'.11 That model would be the Organonmodell der Sprache, which was intended to 'limit the dominance of the representational function of language'.12 Bühler described that model more extensively in Sprachtheorie: die Darstellungsfunktion der Sprache (Jena, 1934),13 the first of three projected books. The other two would have encompassed expression (Ausdruck) and appeal (Appell).
The Organonmodell der Sprache consists of a concrete sound phenomenon with three dimensions: Darstellung (representation - having reference to what is spoken about, which may be objects or states of affairs), Ausdruck (expression - having reference to the speaker) and Appell (appeal - having reference to the people spoken to). The language sign is 'a symbol by virtue of its being co-ordinated to objects and states of affairs; a symptom ... by virtue of its dependence on the sender, whose inner states it expresses; and a signal by virtue of its appeal to the hearer, whose inner or outer behaviour it directs as do other communicative signs.'14 Although he did not dispute the significance of the representative function of language, he argued that its expressive and appellative functions should be given due weight as well. It is unfortunate that Bühler did not get around to writing the other two books in the project as both expression and appeal both deserved more intensive treatment than he gave them in Sprachtheorie.15 Although the book was primarily focussed on verbal language, he devoted a section to 'Symbolic Fields in Non-Linguistic Representative Implements' where he discussed the image in terms of its relational fidelity. Bühler discussed the problem of such things as paintings, photographs, sculptures, maps, music scores, the theatrical performance and graphic representations replacing the concrete sound phenomenon by visual configurations, analyzable as relational models governed by the principle of abstractive relevance.
Gombrich applied Bühler's model to the visual image in Art and Illusion but in an unsystematic way.16 Part one explored the representational dimensions of the image using the notions of the relational model and the 'Synsemantics of pictorial values in the painting'. 17 Part three explored the appellative dimension (the beholder's share) and part four, the expressive dimension.18 Part two explored the issues originally raised by Heinrich Gomperz, combined with further exploration of the issues raised by Löwy. In contrast to Bühler's tome, it was all done eloquently and with an abundant use of images, nevertheless it made extensive use of his ideas.
Gombrich's mentors
Before he entered university, Gombrich had already developed an interest in art history but by the time he left, he had no interest in becoming a conventional art historian. His conversion was largely down to admiration for Emanuel Löwy, inspiration of Julius von Schlosser and his friendship with Ernst Kris. They subsequently became the three dedicatees of Art and Illusion.
Gombrich had met Löwy first and was immediately struck by his lovable personality. He was even more taken by his psychological approach to the study of art. He
never failed in his lectures and seminars to refer to the psychological laws that dominate the rendering of nature in ancient Greece. Not only in ancient Greece, however; "wherever we are able to trace the development of art on a large scale" - we read in his early book - "it moves, to speak morphologically, from the psychological towards the physiological, the image on the retina, the objectively recorded slice on nature ... True, that last final consummation is never attainable, were art ever to arrive there, it would destroy itself" (p 17), a remark that incidentally throws a flood of light on the artistic situation at the turn of the century.19
Whilst Löwy was a friend of Freud's he did not subscribe to his psychology, preferring instead rather dated ideas of sensation, perception and the memory image. Nevertheless, Gombrich regarded his work as having the scientific ambitions demanded by the Vienna School of art history.
In his second year, Gombrich met Schlosser, who shared an interest in psychological issues with Löwy:
While Löwy was of course mainly concerned with the evolution of Greek art, it is well known that in Schlosser's writings similar psychological ideas come into play in his analysis of medieval styles, where he repeatedly concentrated on the psychological problem of the formula, the "simile."
If Löwy was interested in the progressive articulation of the schematic image in antique art, Schlosser was interested in the general problem of Kunstsprache: the individual artist's use of a visual language to create a style. Schlosser was adamant that the history of art was not to be described as the history of style but as the history of development of artistic language.20 It was only the masters who could be described as having a style. As Gombrich remarked, that placed him in a quandary:
He had been much impressed by the demand of his friend Benedetto Croce to treat every work of art as an incommensurable individual expression; but how could the historian get hold of that elusive insularity of the work of art if he found it embodies in the artistic idiom?21
Gombrich attempted to address that problem in his work with Josef Bodonyi in 1933 on the phenomenon of the gold ground in late antique art:
The history of art in late antiquity has repeatedly been a favoured field for methodological experiments. In tandem with the dissolution of an artistic tradition that had always been highly regarded, the basis was laid for a new artistic language (in the sense of the term used by Julius von Schlosser).22
That new artistic language was the consequence of a change of function of the figurative image from the imaginative recreation of a scene, as it might have been perceived by an eyewitness, to a pictographic script operating within a culture that had an aversion to naturalistic imagery. Of course late antique art was anonymous, thus the question of individual style did not emerge, but Gombrich would return to the problem later in his essay on Botticelli.
Working in Schlosser's Seminar, in his third year (1931-2), he prepared a paper on Der Sachsenspiegl in which he became interested in gesture as an expression and the difference between symbolic and spontaneous expression. This led him to attend Bühler's lectures on expression, which would be published as Ausdruckstheorie in 1933. He also met the museum curator and Freudian analyst Ernst Kris, who persuaded Gombrich to abandon a career in academic art history in favour of a more scientific approach to the subject.23 It was also through Kris that Gombrich encountered the work of Heinrich Gomperz, in their discussions, with Otto Kurz, of Die Legende vom Künstler.24
After leaving university, Gombrich became Kris's research assistant and they investigated the expressive features of the Naumburg sculptures, becoming even more deeply involved with Bühler's students. Then they became collaborators on a project on caricature, which was interrupted by his move to the Warburg Institute. He had learned a number of very important lessons in that brief interval, the first being the importance of tradition in the continuity of artistic motifs and the second that there was no language of form. Artistic form did not, of itself, speak.
If we now look at Art and Illusion we can see:
1. Schlosser's presence in a concern with the language of art and the use of the simile;
2. Löwy's presence in his account of the development towards naturalism out of the schematic image;
3. Kris's presence in the discussion of caricature;
4. Gomperz's presence in terms of the origins of art in substitution and the problem of representation as re-presentation, and
5. Bühler's presence throughout the book.
Schlosser had prompted the idea of a study of Kunstsprache and Bühler provided the methodological framework for contributing to a linguistics of the image.
Transition to iconography
In the mid 30s the situation in Vienna was becoming increasingly dangerous for Jews and Kris used his influence with Fritz Saxl to obtain a post for Gombrich at the newly arrived Warburg Institute in London. His job was to act as an amanuensis for Gertrud Bing and prepare Aby Warburg's papers for publication. The Warburg Institute's mission was to facilitate the study of cultural history and it had been founded on the basis of its famous library, the Kulturwissenschaftliche Bibliothek Warburg, using art history as a documentary resource. Thus it complemented the University of London's Courtauld Institute, whose purpose was the academic study of art history. Faced with the problem of making its work accessible to an English audience, Saxl was keen to make its resources usable and it was in this context that he arranged for Ernst Gombrich and his friend Otto Kurz to teach iconography at the Courtauld Institute in 1938. As part of that assignment they prepared a textbook that would contain introductions to the subject -Gombrich on secular iconography and Kurz on religious iconography - in combination with introductory bibliographies. However, the war intervened. Erwin Panofsky's Studies in Iconology, published in 1939, became the text that established the discipline in the Anglophone community and Gombrich finally published Symbolic Images in 1972, using material dating from the decade following the end of the war supplemented by an introduction written for the occasion. Symbolic Images was essentially a set of case studies with a reflective preface. It was not a systematic handbook, as the original text would have been. The shape of the proposed text on Secular Iconography can be seen from an undated plan that was produced some time in the early war years (Figure 2).The relationship between image and text would have played a central role in the book's argument. It was a warburgian preoccupation, somewhat transformed in Panofsky's hands. Gombrich had already become interested in the topic in his work with Kris on the Naumburg figures. They had come to the conclusion that a verbal setting for an image would determine its subsequent interpretation. Gombrich's work on renaissance material deepened his interest in the problem of the relationship between painted images and literary texts. He took the view that there was a spectrum of possible connections. At one end, an image could be an illustration of a specific text; at the other, an image could require no text at all. All images would require elucidation in terms of their social use or function. The first chapter would have clarified the different dimensions of image analysis - representation, signification and expression - and the impact of traditions of imagery on the creation of relationships between form and meaning. Subsequent chapters would have pursued a systematic analysis of various kinds of imagery and the book would have concluded with case studies of particular works. It was intended to be a handbook.
Panofsky's publication of Studies in Iconology had rather taken the wind out of Gombrich's sails. Panofsky was the senior scholar with an established reputation and Gombrich's junior position at the Warburg Institute meant that he was hardly in a position to compete with him. He was, in any case, busy with writing The Story of Art and thinking about other books that would be fit for popular consumption. He continued to work on the various case studies that would have been included in Secular Iconography but each text was published separately. 25 His profile of publications in the area of iconography was consistent with his new role as a lecturer in Italian Renaissance culture and later Professor of the Classical Tradition at the Institute. However, he maintained his work on imagery within the context of a new project 'On the Realm and Range of the Image' that was intended to appeal to the same audience that read the books of Herbert Read and Susanne Langer. The popularity of The Story of Art and his subsequent invitation to give guest lectures and accept visiting professorships gave added impetus to the production of such a book.
'The Realm and Range of the Image' and Art and Illusion
As may be seen from its initial plan, the proposed book would have proceeded systematically and would have emerged in a form that could be used as a textbook. However, history intervened. The Story of Art became an overnight success and in consequence Gombrich was flooded with invitations to give lectures. With the exception of his biography of Warburg, all of his later publications originated as lectures. A number emerged, reorganised as monographs, but the majority were published as essay collections. This had a profound impact on his presentational style and as he later observed:
I have become increasingly sceptical of the value of that survival of epideictic oratory called "the formal lecture". I was all the more happy to see, that this scepticism is shared by the supreme master of the genre, Lord Clark, who really said all there is to be said on this topic in his splendid self-portrait. "Historical truth," he says, "is usually complex and frequently dull, and anyone with a sense of style or a love of language is tempted to take short-cuts and omit the qualifications that would make a statement less telling". "The lecture form" he remarks, "encouraged all the evasions and half-truths that I had learned to practice in my weekly essays at Oxford". With this damning accusation in mind I suggest that this Society should spawn a Society for the Reform of Lecturing with a radical wing agitating for its total abolition. I would at least plead for the abolition of lectures which are subsequently to be published. I believe that what is most suitable for discussion in front of an audience are precisely those half baked ideas and arguments which are not yet ripe to be printed.26
It is a central characteristic of a lecture and lectures in a lecture series that they are best presented as autonomous, self-standing, arguments. Collected together they are not written as continuous expositor arguments and, especially in Gombrich's hands, they lose out in scholarly rigor what they gain in rhetorical effect. It is remarkable that while Gombrich was consistently praised for his clarity and ease of access, very few scholars were capable of giving a coherent exposition of his ideas. Interpretations of what he said differed wildly. The finally published text of Art and Illusion was much more restricted in scope than the originally planned book (Fig. 3) and effectively consisted on meditations on a variety of topics, strung together in a loose narrative, rather than a rigorous pursuit of an argument. While 'Light into Paint' takes as its starting point Constable's Wyvenhoe Park, for example, the 'relational model', which is so central to his argument is not systematically clarified. Instead the reader is offered a mosaic by way of illustration. 'Truth and the Stereotype', because it is focussed on truth is obviously focussed to the expense of the latter. As the book is focussed on the mechanics of the naturalistic image, that is treated as an ideal type against which paintings become contingent examples of imagery. I would not want to argue that there is anything inherently wrong with such an approach because it all leads to greater readability and appeals to a popular audience.
Gombrich, in later life, commented that he saw his activity as that of a commentator on the history of art:
My ambition - and it was rather a lofty ambition - was to be a kind of commentator on the history of art. I wanted to write a commentary on what actually happened in the development of art. I sometimes see it as representation in the centre with symbolism on the one hand and decoration on the other. One can reflect about all these things and say something in more general terms. It was my ambition to do precisely this.27
In this context, the project 'The Realm and Range of the Image' would lead to a turn away from 'symbolism' to focus on 'representation', which, from a systematic point of view, was unfortunate because to talk fully about representation would indeed involve a discussion of symbolism.
The semantic approach
In a proposal sent to Walter Neurath at Thames and Hudson in 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 that modern linguists have studied the functions of speech. The studies of Richards in Cambridge and the development of the semantic approach, notably in U.S.A. seem to hold out the promise that ultimately even the analysis of aesthetic values cannot 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 Monet.28
It was in the course of preparing that book that he published a review of Charles Morris's Signs, Language and Behavior, where he invoked notions of the linguistics of the image and a general iconology. He rejected the peircean notion of the iconic sign on the basis that the very idea of iconicity is up to question. The iconic image is built up of non-iconic elements and there is:
hardly an image that is purely iconic. We are always confronted with an admixture of 'post language symbols'. 29The analysis of primitive art and of child art has made us familiar with the role which the 'conceptual sign'30 plays in these styles. Its link with the linguistic faculty (drawing is enumeration of characteristics) is known. Does the decrease of the conceptual (and conventional) image in art mark an increase in 'iconicity'? Is the primitive manikin, complete with ten fingers and toes, less iconic than the patch of colour which may denote a gondolier in one of Guardi's paintings? Here again many difficulties may be avoided by concentrating less on a morphological classification of signs than on an analysis of their interpretation. Guardi relies on the beholder's capacity to read 'iconicity' into his sign. The contextual, emotional or formal means by which this type of 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 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.
Once the beholder's attitude rather than the objective structure of the sign is moved into the focus of attention, even more problems relating to the concept of 'iconicity' will probably arise. Is God the Father in an altar painting, or Justice in a courtroom, to be regarded as an iconic sign (of God or Justice) or as a post-language symbol, a mere substitute for a written word? The answer clearly depends on the attitude and beliefs of the beholder. It even seems that certain classes of religious and allegorical imagery derive their appeal from the fact that their 'degree of iconicity' remains indeterminate. They are neither regarded as mere ideograms of rational concepts not as iconic representations of a visible reality but rather as the visual embodiments of a suprasensible entity.31
This passage concluded with a footnote cross-referring the reader to his article 'Icones Symbolicae' that had been published in the Journal of the Warburg and Courtauld Institutes in the previous year.
From this brief passage the reader is alerted to a number of problems: the inadequacy of the notion of the iconic sign; a notion of interpretation linked to projection; a notion of a visually trained public and the role of that public's attitudes and beliefs towards the ontological status of the depiction. But what is also clear from the review is that Gombrich did not see himself as a potential creator of a linguistics of the image so much as a commentator on the problems that such a project would have to face. He alerted theorists such as Morris to importance of an awareness of iconological research as practised by art historians and he alerted art historians to the problems that they would face in pursuing a 'scientific' approach to their study. Morris's ambition had been to create a scientific approach to the humanities but, Gombrich argued, his scientificity was an illusion: his empirical science of semiotic was not scientific at all as it did not rest on what he called 'the practical semiotic of the past' which was a product of the work of iconologically sensitive art historians: 'The work of such pioneers in the study of the symbolic aspects of the image as A. Warburg and E. Panofsky are absent.'32
The roots of artistic form
Gombrich's essay 'Meditations on a Hobby Horse or the Roots of Artistic Form' was concerned, he said, 'with the status and origin of the "iconic sign"' and developed the argument of the Morris review, mentioning in passing 'Icones Symbolicae'.33 In this essay he discussed the notion of representation as being grounded in substitution, thereby picking up Gomperz's earlier discussion from 1905, giving it a modern twist by incorporating a critique of the notion of image as abstraction. Looking back on the paper at the end of his life, he complained that it had failed to generate any debate34 and he also regretted that he had not made sufficient distinction between two and three dimensional imagery.35 Despite its readability, its argument is very complex and deserves to be unwoven. One passage deserves particular scrutiny:
Is it quite correct to say that [the hobby horse] consists of features which make up the 'concept' of a horse or that it reflects the memory image of horses seen? No - because this formulation omits one factor: the stick. If we keep in mind that representation is originally the creation of substitutes out of a given material we may reach safer ground. The greater the wish to ride, the fewer may be the features that will do for a horse. But at a certain stage it must have eyes - for how else could it see? At the most primitive level, then, the conceptual image might be identified with the minimum image - the minimum, that is, which will make it fit into a psychological lock.36
The problem with this passage is that Gombrich was still operating with the old language of the conceptual image, which effectively returned him back to Gomperz's analysis of naturalism. To move forward he would need to make use of Bühler. The hobby-horse is a three-dimensional object, while the picture of the hobby-horse is a two-dimensional image inviting recognition of a three-dimensional object. The picture invites the recognition because it offers the hobby-horse's distinctive features. But the stick became a hobby-horse not because of its visual features, which could later be supplemented to look more like a horse, but because it was an object that afforded riding. A bare, unadorned stick will do the job. 37 We should, in this context, take note of Bühler's observations concerning both the index field and the behavioural context of speech utterances, the deictic field (Ziegfeld), which is constituted by deixis.38 The stick does not picture a horse but becomes one because of its user's behaviour. The child imagines herself as riding a horse. The index field is behavioural and the deictic element is the child prancing around on the horse. The stick by itself is not a horse. It needs the child's riding behaviour to become one.39
When one turns to prehistoric cave art, seemingly a case of two-dimensional picture making, another problem presents itself in connection with the pictorial field. With Gombrich's rapid switches between three-dimensional objects and two-dimensional surfaces, he overlooks the very real and transitional uneven and messy use that prehistoric artists made of the interior of caves. The photographs of such things, as for instance in The Story of Art, are no substitute for the original things: they offer a two-dimensional image of something that was not two-dimensional and they seemingly offer configurations in a noise of interference. While impressionist paintings demand to be viewed in natural light as opposed to electric light bulbs, cave paintings demand to be viewed in the flickering light of torches as opposed to the permanent glare of studio lighting. The problem of setting, or physical context, was posed by The Story of Art itself. The first chapter showed a painting of bison found in a cave in Altamira in Spain. The original paintings were covered by a range of markings that, up until recently, have been counted as visual noise: hand prints, dotted patterns and engraved marks. In a book review, Gombrich referred to 'mysterious scratching and paintings in remote caves' as if the scratching and painting were separate phenomena.40 Following Bühler, one asks 'what is the field?' and 'what is the notation?' Are we looking at imagery 'defaced' by graffiti? What are we looking at? What would have been the spectators' and producers' expectations and anticipations?
'the emerging discipline of iconology'
Although Symbolic Images was published in 1972, the bulk of its material was written much earlier, in the decade following the Second World War. This was important because it was not until the late 50s that Gombrich adopted Bühler's relational model and extended his range of interests beyond the work that he had earlier accomplished with Kris. It was not as systematic as the earlier proposed text on Secular Iconography. The dynamic of writing individual articles for publication undermined the project of producing a unified manual as the case studies became the tail that wagged the dog. The introduction, 'Aims and Limits of Iconology', consisted of a set of loosely connected reflections and lacked the programmatic intent of Panofsky's introduction to Studies in Iconology. Gombrich had a deeply rooted aversion to theory construction but in this case it undermined the possibility of the presentation of a coherent set of ideas.
The Preface opened with a critical examination of Panofsky's first, pre-iconographical, level of analysis and attacked the idea of 'representational meaning':
It looks quite plausible to speak of various 'levels of meaning' and to say, for instance, that Gilbert's figure has a representational meaning - a winged youth - that this representation can be referred to a particular youth, i.e. the God Eros, which turns it into the illustration of a myth, and that Eros is here used as a symbol of Charity. But on closer inspection this approximation to meaning breaks down on all levels. As soon as we start to ask awkward questions the apparent triviality of representational meaning disappears and we feel tempted to question invariably the need invariably to refer the artist's form to some imagined significance. Some of these forms, of course, can be named or classified as a foot, a wing, or a bow, but others elude this network of classification. The ornamental monsters around the base no doubt are meant to represent marine creatures, but where in such a composition does the meaning end and the decorative pattern begin?41
The immediate problem created by this example is that Gombrich used a monument, the Picadilly fountain, rather than a statue or painting to address the topic of 'the elusiveness of meaning'.42 Given the notion that one may approach visual imagery in terms of its 'distance from a text' - from images illustrating a specific text to images illustrating no text at all - how should one to approach a sculpture placed in a decorative setting? If one thinks instead of a picture within a frame, it is not obvious that one is obliged to extend the meaning of the picture out to the frame. If Gombrich truly wanted to offer a critique of Panofsky's notion of representational meaning he should have offered a better example. He should have followed Gomperz's example and started with the Iconoclastic controversy, which centred on the problem of calling an image of God an image of a man. What precisely was the status of an image of Christ?
Gombrich attempted to settle the question of meaning by using D.E. Hirsch's Validity in Interpretation, particularly his arguments that the meaning of a text is the artist's intended meaning43 but this fitted uncomfortably with his observation that 'Images apparently occupy a curious position somewhere between the statements of language, which are intended to convey a meaning, and the things of nature, to which we can only give a meaning.'44 He would have done better to refer back to Bühler's Organonmodell der Sprache to explore further the notion of 'appeal' and apply it to the case of Veronese's Feast in the House of Simon.45 When the artist was interrogated about its figures it became clear that many simply had an ornamental purpose and to talk about their symbolic significance was beside the point:
Q. Did anyone commission you to paint Germans, buffoons, and similar things in that picture?
A. No, milords, but I received the commission to decorate the picture as I saw fit. It is large and, it seemed to me, it could hold many figures.46
Rather than violate the decorum proper to the Last Supper, he simply changed the painting's title to Feast at the House of Levi. The moral of this story is that it is not the case that every figure in a picture necessarily had any meaning beyond its compositional significance.
The idea of the brief for the Piccadilly fountain led naturally enough into the larger question of the difference between iconography and iconology. Panofsky had described iconology as the study of 'Intrinsic meaning or content, constituting the world of 'symbolical' values' controlled by 'History of cultural symptoms or 'symbols' in general (insight into the manner in which, under varying historical conditions, essential tendencies of the human mind were expressed by specific themes and concepts.'47 Ignoring that, Gombrich declared 'by and large we mean by iconology, since the pioneer studies of Panofsky, the reconstruction of a programme rather than the identification of a particular text.' 48 Thus, in one fell swoop, Gombrich demolished the relevance of Cassirer's notion of the symbol, which Panofsky had adopted in his scheme. The problem is plain. Gombrich was aware of the difference between two different senses of the word 'symbol':
I believe that the very pleasures of the symbol hunt have tended to obscure the fact that the vast majority of monuments in our churches and museums do not present [a] riddle. We all recognize the Statue of Liberty without benefit of iconological study, just as we know that if she held a balance instead of a torch, we would call her Justice.
The symbol here would be the mark of identification, not different in principle from a label or a tag. This is indeed the sense of the term we often use when we speak of mathematical symbols or of the symbols used on road signs. But if iconology were only concerned with the identification of labels, its psychological interest would be slight. But is it? Not if we follow a different usage that insists that a symbol is more than a sign, whatever 'more' may mean in this context.49 This clash over terminology that has much bedevilled discussions is partly due to a divergence of philosophical traditions in Western thought.50 The extension of the term 'symbol' to cover any kind of sign can be traced in the Anglo-Saxon tradition from Hobbes to Peirce, and has led to such coinages as symbolic logic. Against this expansionist tendency, the German Romantics and their French successors stressed the religious connotations of the term and wanted it restricted to those special kinds of signs which stand for something untranslatable and ineffable. It is not my intention to adjudicate between these usages. But I think it might be worthwhile to examine the restrictionists' case in the light of psychology. One of the advantages of the study of art is precisely that it should provide material for such an examination. For it is in the study of art and its history that even the rationalist and agnostic historian has to grapple with the symbol that is called 'more than a sign' because it is felt to be profoundly fitting.51
Given the importance that Panofsky had attached to iconological analysis in the 'deep' sense and the fact that Gombrich had chosen to cite Bühler in an explanatory footnote it is very disappointing that he had not had more to say on the subject at this point. Bühler had argued:
Different resources are used to render the strength, resoluteness and beauty of the goddess of justice than are used for the attribute of justice itself. It seems to me that this is an approach towards a proper definition of the concept of symbol within the framework of sematology.52
This is an observation that Gombrich applied in his essay on Raphael's Stanza della Segnature ninety-one pages further into the volume:
Take another of Raphael's personifications, Poetry, or better still his ravishing drawing for that figure. From one point of view it is a pictorial sign with the obvious and enumerable attributes of the wings, the lyre and the book. But this vision not only signifies numine afflatur, it also displays or expresses it. The sign limits are blurred. The upturned gaze may still be a conventional sign for inspiration, carried by the tradition of art from ancient times, but the tense beauty of the figure is Raphael's own, and not even he could quite transfer and repeat it, for it may well be that the finished image is a little less convincing as an embodiment of the Divine afflatus though it has an added laurel wreath.53
If we now refer back to Bühler's original argument we can see that the two notions of the 'symbol', one logical and the other Romantic, are grounded in a difference between their utilitarian values:
Let us take note of the following point about the modern history of the concept of symbol: the romantics loved the concept of symbol and lavished a plenitude of meaning upon it that comes quite close to the wealth of meaning contained in the idea of an "image or likeness" of things, whereas the logicians (one could almost say for professional reasons) advocated emaciation and formalization of the content of the concept so that in the end nothing was leftother than the arbitrarily agreed coordination of something or other as a sign to something or other as the signified.
In addition to the mention of these two easily understood motives, we need only add a word about the broad circle of applications of the concept, and then we have extracted everything from the history of the concept of symbol that concerns us. In addition to the signs that are called symbolic, which have a representational value, are there not also symbolical actions to be encountered everywhere, and are not certain unique things such as kings' insignia (for example, the Crown of St. Stephen, the Imperial Orb) also "symbols", whether of the rights and dignity of sovereignty itself, or of their bestowal and possession? Of course that is the case, and the list of applications is by no means exhausted here. One observation that can be made here is almost amusing, namely how the difference of taste between romantics and non-romantics also becomes evident in this domain. The one party regards an action as symbolic inasmuch as it is removed from real purposiveness and cut offfrom any clearly physical effect, precisely because it is no longer genuinely effective, but rather a "merely symbolic" gesture: another will call the same action symbolic because after its emancipation from lower (say animal) purposiveness it has taken on a higher human function and now stands before us as an allegory, or because the legally binding character of an act or some other import it has depends on its "symbolic character".
It would be a pure waste of energy to write an apologia for the one or the other motive of definition. There will always be romantics and non-romantics; they must simply try to understand each other in science. It seems to me that it is not now possible to revoke the concession that there are two concepts of symbol. If it were possible to revoke it, the same difference of mentality would come to the fore somewhere else on some other topic.54
On this line of analysis Raphael's drawing of poetry took on a 'higher human function' than mere signification and it would have been appropriate for Gombrich to analyse what that higher human function might have been. As we shall see, to say that the drawing is expressive of its content while the painting fails in this respect is to misinterpret Bühler's Organonmodell der Sprache. And this betrays a weakness of the whole volume: it draws together material written before and after Art and Illusion thus failing to reconcile an early model of representation with a late one.
The discussion of philosophies of symbolism in the preface introduced the reader to the Renaissance Platonic and Aristotelian approaches, which of course would be developed in the essay 'Icones Symbolicae' but it ended on a note of caution:
Nobody who has looked into medieval and Renaissance texts concerned with symbolism can fail to be both impressed and depressed by the learning and ingenuity expended on applying the techniques of exegesis to a vast range of texts, images or events. The temptation is indeed great for the iconologist to emulate this technique and apply it in his turn to the works of art of the past. But before we yield to this temptation we should at least pause and ask ourselves to what extent it would be appropriate to the task of interpreting the pictures or images of the past.55
Apart from the obvious use of symbols, to be found where they are appropriately located, Gombrich warned that there is no evidence that 'applies this doctrine to pictorial works of art' or rather that they 'were commissioned to be painted'. One must distinguish between possible habits of thought amongst spectators and the brief the artist was given. And, one might add, one must give due weight to artists adopting traditional modes of depiction. Indeed, it would be quite inappropriate to regard every Annunciation as freshly born out of an artist's mind, concerned with the latest theological controversy.
With Gombrich's methodological warnings in mind, the contemporary reader must approach the opening essay 'Botticelli's Mythologies' with some degree of caution. It was written early and was intended to demonstrate his commitment to warburgian work and at the same time offer a show piece of iconological analysis. Although new evidence subsequently came to light concerning the location of the paintings the essay still raised useful and important questions - what is the relationship between image and text? What is the significance of stylistic affinities between the paintings under consideration and other paintings? How does style affect meaning? Not least, how does a magical approach to the use of visual imagery affect its users' expectations? If Botticelli's paintings are to be treated as symbols, in the deeper humanly significant sense, then the point of the paper was to demonstrate the multiplicity of meanings made possible by neo-Platonic habits of exegesis. The Primavera was shown to be a symbol with a particular expressive significance56 and an equally significant appeal: the image functioned in a highly specific constellation of patrons' interests and artistic skill. In hindsight this was a perfect example of a sematological analysis exploring the linguistics of an image. Beyond linguistics and into poetics, it also established the artistic values explicit in the painting by drawing connections with other related artworks of the day through visual comparison and an appeal to nuance. Commentators seem to have completely missed its point. 57
From an explicitly semiological/sematological point of view the next essay in the volume, on Raphael's Stanza della Segnatura, makes more interesting reading. Written and revised much later than the immediately post-war publication, Gombrich made an explicit reference to Bühler's principle of abstractive relevance, connecting linguistics to art history:
No sign or symbol can refer to itself and tell us how much it is intended to signify. All signs have a characteristic which Karl Buehler called 'abstractive relevance '. The letters of the alphabet signify through certain distinctive features but in normal contexts their meaning is not affected by their size, colour or font. The same is true of the images which interest the iconographer, be they coats of arms, hieroglyphs, emblems or personifications traditionally marked by certain 'attributes'. In every one of these cases there are any number of features which are strictly speaking without a translatable meaning, and only a few which we are intended to read and translate. But while we can easily identify those in the case of codified signs, the limit of significance, what Buehler called the 'sign limit', is much more open in the case of symbolic images. Take the personification of Causarum Cognitio on the ceiling of the Stanza (Fig. 4). Care has been taken for us to know the meaning of the books she holds, for they are inscribed. Nor need we doubt that the decoration of the throne was intended to signify an aspect of Philosophy-the many-breasted Diana stands traditionally for Nature. Now Vasari also tells us in some detail that the colours of her garment, from the neck downward, are those of fire, air, earth and water, and are therefore symbolic. He may be right, but what of the garments of the other personifications? He does not tell us, but even if he did, we could always ask further questions. Are the configurations of the folds significant? Are the positions of the fingers?58
Gombrich's point about Vasari is disappointing as careful observation of Raphael's painting of Causarum Cognitio would reveal that Raphael himself had encouraged the observer to give a symbolic reading to the colours of the lady's cloths as they were also inscribed with stars, marine creatures and plant forms. These features would have provided the motivation for a symbolic reading, which was evidently not simply Vasari's bright idea. While Jones and Penny drew attention to Theology's green, white and red garments, which they say were the colours of the Theological Virtues59 Vasari did not feel similarly inclined. There was certainly nothing in his copious writing to suggest that he would attach any symbolic significance to folds and fingers. As Bühler's model demonstrated, significance is directed: it is an appeal. In this instance Gombrich's use of Bühler's principle of abstractive relevance was almost casual: he ignored the contribution that the stars etc made to the pictorial field, rather in the same way that he missed the possible significance of the 'mysterious scratching and paintings in remote caves' connected with pre-historic art.
Having made the connection to Bühler, Gombrich threw the problem of interpretation open again as the information imparted by an image is inherently different from that imparted by a text. As we know from the popular expression 'a picture is worth a thousand words':
But though it is obvious that the limits of signification are invisible when looked upon from outside, as it were, when we are confronted with any particular image, the situation changes dramatically when we look at the problem from the other side, when we consider a text which is to be translated into an image. The reason is simple. Language operates with universals, and the particular will always slip through its net, however fine we may make its meshes. It is because language is discrete and painting continuous that the painted allegory cannot but have an infinity of characteristics which are from this point of view quite devoid of meaning.
Of course what is allegorically meaningless is not, thereby, without significance of some other kind. If it were, the image would be a pictogram and not a work of art.60
Take another of Raphael's personifications, Poetry (Fig. 5), or better still his ravishing drawing for that figure (Fig. 6). From one point of view it is a pictorial sign with the obvious and enumerable attributes of the wings, the lyre and the book. But this vision not only signifies numine afflatur, it also displays or expresses it. The sign limits are blurred. The upturned gaze may still be a conventional sign for inspiration, carried by the tradition of art from ancient times, but the tense beauty of the figure is Raphael's own, and not even he could quite transfer and repeat it, for it may well be that the finished image is a little less convincing as an embodiment of the Divine afflatus though it has an added laurel wreath.
Elsewhere in this volume the history of the distinction has been traced between the symbol and the allegory. The allegory was felt to be translatable into conceptual language once its conventions were known, the symbol to exhibit that plenitude of meaning that approaches the ineffable.61
But the same study has also suggested that historically speaking the distinction is falsely drawn. Personifications were not simply viewed as pictographs denoting a concept by image rather than by label. They were rather true representations or embodiments of those superior entities which were conceived as the denizens of the intelligible world. It is here, again, that the intellectual tradition embodied in the Stanza met the artist half-way. Indeed the orator who selected these universals as objects of praise could appeal to the visual imagination to clothe them with flesh and blood.62
If we turn our minds back to Bühler's model we must probe Gombrich's argument more closely before the fog sets in. Symbols have representational, expressive and appellative dimensions. The sign limits were not blurred. Raphael's drawing simply had a different expressive force from his painting. Following Bühler one may say that an image has an expressive dimension but that is within the context of its production: both the artist and the spectators would have deemed the use of the drawing's characteristics in the painting inappropriate to its setting. The painting had a majesty that the drawing lacked. Gombrich's artistic sensitivities were those of a twentieth-century observer. In this context Gomperz's original observation becomes pertinent: in modernity the centrality of the subject of art was replaced by the importance of the artist.
Gombrich's paper on Tobias and the Angel demonstrated the importance of genre in establishing the meaning of a work of art. Renaissance artists themselves distinguished between imagini, which were images of the saints and suchlike, and storie, which were narrative episodes.63 These were different genres of painting and they occupied different discursive fields. Tobias and the Angel might seem to be a storie, but it was in fact an imagini. This has fundamental consequences for the modern spectator's recovery of the image's historical appeal.64 A close scrutiny of the image will reveal the peculiarity of the physical contact between Tobias and the Angel: the two figures hold each other's hands very strangely. It is not as if Tobias holds the angel's hand so much as he uses his hand to indicate that he is linked to it. With a warning that 'the study of meanings cannot be separated from the appreciation of forms' Gombrich drew attention to the competing demands of symbolic clarity and realistic representation. In representing the imagini of the Archangel Raphael:
The painter had to display him before us in full view like any Saint represented to receive the prayers of the faithful. Here we can still discern the hieratic formula of the symbolic art of the Middle Ages. Tobias, on the other hand, was from a different world. The painter was not restricted by the requirements of devotional art. Tobias' part in the story was to stride along confiding himself to his helpmate. In his person the new realism could be given free play. The problem was how to fit the two together. ... It is important to remember that in the mind of those who ordered and painted the picture this framework of references was not a matter of rational thought. To them meaning and effect were still inextricably entwined. Raphael's image does not merely 'represent' the Biblical symbol. It partakes of its meaning, and hence of its power; it helps the donor to enter into communion with those forces of which it is the visible token.65
There are echoes in this text of Gomperz's formulation of the problem of representation:
The picture as image of something ('Abgebildetes') and the picture as the source or vehicle of an image ('Abbildendes') represent two distinct conceptions; they must correspond to two completely distinct complexes of psychological elements. As far as the facts of the visual perception and association, the sensual qualities of the picture, there is no difference in what occurs in either of these: its colours, surfaces etc. are completely identical for the iconoclast and the worshipper. The difference is also not possibly due to the fact that the worshipper consciously associates other ideas ('Vorstellungen') with the subject matter of the image in a sort of amalgam of associations ('"Assoziation von Vorstellungen"'). We have already stressed more than once: that the iconoclast might equally well be reminded of the deity itself by its mere image and yet this does not make him an image worshipper. When I see a boy playing on a toy horse, I might after all myself be reminded of a real horse: this does not mean that the stick has become a reality to me to the extent to which it has for the child. The function of the image does not involve a tangency of parallels, but rather a co-mingling of the image and its subject matter ('ein Ineinander von Abbildendem und Abgebildeten'). Such an interpenetration cannot be the result of a mere addition of the conception of the visual subject to that of the visual object ('dem Abbildenden von dem Abgebildeten'). If this difference of psychological elements cannot be explained as a difference in conceptions, then it can only be due to a difference in psychological elements not attached to the subject of sensual impressions and yet still essentially properties of our conception of the objects.
To put it more simply, the modern spectator, who to all intents and purposes adopts an iconoclastic position, sees a representation of the angel Raphael but the contemporary Christian spectator sees the angel Raphael in the figure of Azarias, in a way analogous to seeing an actor play the role of Julius Caesar. Bühler would have described it in terms of a difference in mental set and we will return to this shortly.
The last essay in the book, 'Icones Symbolicae' approached the illusionist image via the doctrines of neo-Platonism. Nevertheless, it still sat comfortably in the problems discussed by Gomperz and Bühler and the warburgian fascination with Denkraumverlust, the 'tendency of the human mind to confuse the sign with the thing signified'.66 Its interest, for this paper, comes from the schema that it adopted to discuss the functioning of the image:
[The] three ordinary functions of images may be present in one concrete image - a motif in a painting by Hieronymus Bosch may represent a broken vessel, symbolize the sin of gluttony and express an unconscious sexual fantasy on the part of the artist but to us the three levels of meaning remain quite distinct.
As soon, however, as we leave the ground of rational analysis we find that these neat distinctions no longer hold. We know that in magical practices the image not only represents the enemy but may take his place (the very word re-present still has this dual meaning). We know that the 'fetish' not only 'symbolizes' fertility but has it. In short our attitude to the image is inextricably bound up with our whole idea about the universe.67
This schema, of representation, symbolization and expression recurred frequently in Gombrich's notes for his project on the realm and range of the image but it disappeared from view by the time he published Art and Illusion. When we look at his later paper 'The Visual Image: Its place in Communication', published in Scientific American in 1972 we notice that he openly used Bühler's model but gave it a twist:
Students of language have been at work for a long time analysing the various functions of the prime instrument of human communication. Without going into details we can accept for our purpose the divisions of language proposed by Karl Bühler, who distinguished between the functions of expression, arousal and description. (We may also call them symptom, signal and symbol).68
Translation is always a tricky business and it is very important to capture the sense of the original. Bühler use the term Darstellung (representation) as having reference to what is spoken about, which may be objects or states of affairs. It is the symbolization achieved by an utterance and is not a description. The utterance 'Help!', for example, is not a description. It is precisely because it is not a description that it is transferrable to other non-linguistic modes of representation: it is a symbol that is an object of thought or feeling. Second, it is overly simplifying the expressive function to refer to states of mind such as anger or amusement. The expressive function is intended to indicate the utterer's position in relation to the objects or states of affairs referred to: it is an adopted stance. Finally, arousal is not merely intended to refer to a psychological state but on a call to behaviour that could be theorised as interpellation in an almost Althusserian sense, where the individual is engaged by the address to participate in an imagined realm.69 As Robert Innis correctly pointed out:
Bühler, by reason of his general anti-cartesianism, constantly focussed on the social matrix in the genesis of sense and by no means allowed the essential determinant of sense or meaning to be confided to a private mental act. ... Bühler insisted on the primacy of the social and hence publicly observable data for laying the foundations of language theory, combating in this way Husserl's tendency to explain all differences in meaning by differences in acts of meaning, themselves accessible to some sort of introspective process.
Throughout all Bühler's work, the central and ever-recurring thesis ... is the irreducible social matrix of meaning in both the human and nonhuman spheres. ... Social life is characterised first and foremost by a mutual steering or guiding of its members, involving the 'set' (Einstellung) of one individual or group of individuals toward another.70
The upshot of this is that in his earlier 'Icones Symbolicae' paper, Gombrich adopted an approach to the artist's communicative endeavour that harked back to his work with Ernst Kris. In Art and Illusion he dropped that to emphasise the ecology of perception, the causality of the environment spelled out by Bühler's student Egon Brunswik.71 In Symbolic Images he still devoted a section of his introduction to psychoanalysis, though denying its utility to an analysis of symbolic meanings. In 'The Visual Image' he misread Bühler to suggest 'the visual image is supreme in its capacity for arousal, that its use for expressive purposes is problematic, and that unaided it altogether lacks the possibility of matching the statement function of language'.72 Gombrich confused speech with language and the fundamental unit of speech is the sentence, or utterance, not a statement. Consequently all of the very interesting things he had to say about the relationship between the image, the code, the caption and the context, while useful, are misplaced. Captions are typically given in museums and art history books. One does not find them underneath pictures in use in churches, public buildings or private houses unless they have been placed there for the benefit of tourists.
Although Gombrich used Bühler, in respect of mental set, the relational model and the principle of abstractive relevance, he misunderstood the meaning of the Organonmodell der Sprache.73 Bühler's model raises some very interesting questions indeed when used with art historical insight.
At this point we come full circle back to Gombrich and the fundamental question of the nature of representation: more precisely, the symbol as a point of interaction between the sign, its producer and its spectators. If we remember Bühler's emphasis on limiting 'the dominance of the representational function of language', we can see why Gombrich should have stressed the importance of mental set: the attitudes and expectations of creators and spectators and, what he called, 'the beholder's share'.
The creators were, of course, the makers but one shouldn't forget the circumstances of production. Our own mental set insists on the work of art as the production of an individual and identifiable person but, as Gomperz had pointed out, this is a relatively modern idea and an older idea would see it as a crafted object produced for a patron. One is then entitled to ask how the artist was intended to satisfy the patron's expressive needs. Unlike a verbal statement uttered by an individual to achieve an effect with reference back to that expressive individual, a crafted object was produced by an exercise of skill in response to a demand. The artist was an agent for the patron.74 More recently, when the notion emerged of the artist as distinctive creator in his own right, the artist's workshop could become an agent for the artist himself.75 In contemporary artistic practice, the artist becomes an agent for the curator. Within this context the question arises concerning the exercise of power and wealth and how these were concretely mobilised.76
Conclusion
This paper is rough and schematic and just a fragment of a much larger project, which is to understand the roots of Gombrich's thought and the potentiality for its development. As I claimed at the outset, the Anglophone tradition of semiotics in the barthesian strain is different from the Viennese tradition termed sematology. There are differences in the constellation of problems, critical concepts and proposed solutions. The lecture format enabled Gombrich to be unsystematic in forging links between iconology and the linguistics of the image but that was never, as he saw it, his primary task, which was simply to be a commentator on the practice of art history when he was not teaching Renaissance cultural history at the Warburg Institute. Nevertheless Symbolic Images will provide a rich resource for the further study of the 'linguistics of the image'.
* This is a reworked version of a paper presented to the Gombrich centennial colloquium Venice 2009.
1 E.H. Gombrich, Art and Illusion, Princeton 2000, xxvi and note 1. Note 1 reads:
See Nelson Goodman, Languages of Art: An Approach to a Theory of Symbols, Indianapolis, 1968 (I may here mention that the late author allowed me to quote a letter that he wrote to me dissociating his own views from those of the extremists: see The Image and the Eye, Oxford, 1982, p.284); also Norman Bryson, Vision and Painting, the Logic of the Gaze, Macmillan (London), 1983. For the history of this approach, see my article Voir la Nature, Voir les Peintres in Les Cahiers du Musée national d'art moderne, vol.24, Été 1988, Art de Voir, Art de décrire II, pp. 21-43.
2 The irony of this situation is plain as Gombrich regarded himself as a European: 'I don't feel I am English; I feel precisely what I am - a Central European working in England. Thanks to my work with the BBC I learned English thoroughly.' Adi Wimmer (ed.), Strangers at Home and Abroad: Recollections of Austrian Jews Who Escaped Hitler, Jefferson, North Carolina and London: McFarland & Company 2000, 135.
3 Art and Illusion, 8-9.
4 'Signs, language and behaviour', originally published in Art Bulletin, 31, 1949, and reprinted in E.H. Gombrich, Reflections on the history of art: Views and reviews ed. Richard Woodfield, Oxford 1987, 246.
5 Art Journal, Vol. 44, 1984, 162-4.
6 Reply to David Carrier, Leonardo, Vol. 18, No. 2, 1985, 126.
7 'Ueber einige psychologische Voraussetzungen der naturalistischen Kunst', Beilage der Allgemeinen Zeitung, Jahrgang 1905, Nummer 160, München Freitag 14. Juli, pp. 89-93, Nummer 161, Samstag 15. Juli, pp. 98-101, translated by Karl Johns in this issue of the journal. I will publish a full analysis of this paper in the future.
8 I will have to discuss the difference between my version of events and Gombrich's in another place. Suffice it to say that he barely touches the significance of Gomperz's essay for Schlosser.
9 Karl Bühler, "Die Axiomatik der Sprachwissenschaften," Kant-Studien, Band 38, 1933, 24-90. Translated as 'The Axiomatization of the Language Sciences' in Robert E. Innis, Karl Bühler Semiotic Foundations of Language Theory, New York and London: Plenum Press 1982.
10 'Axiomatization', 95.
11 'Axiomatization', 102.
12 'Axiomatization', 153 (Bühler's emphasis).
13 Karl Bühler, Theory of Language: The representational Function of Language, translated by Donald Fraser Goodwin, Amsterdam/Philadelphia: John Benjamins Publishing Company 1990.
14 Theory of Language, 35 (Bühler's emphasis), referring to his earlier terminology, which he now replaced by expression, appeal and representation.
15 It would be useful to incorporate Sir Alan Gardiner's observations in The Theory of Speech and Language (1932) as Bühler felt that it complemented his own work. Theory of Language, 32 n.1. Gardiner started his analysis by focussing on a concrete speech situation, which he analysed in detail, looking and communicative intent and reception.
16 Bühler's work was also fundamental to Art and Illusion through his students Egon Brunswick and Karl Popper: the former for his ideas on the causal texture of the environment and the latter for his searchlight theory of the mind. Gombrich would have admired J.J. Gibson's work in this light. The later dispute with Gibson centres on the latter's rejection of a feedback element in perception, which Gombrich learned more about from Colin Cherry. The concept of feedback sits well in both Brunswick and Popper's theories.
17 Theory of Language, 186-7 based on his work in colour first published as Karl Bühler, Die Erscheinungsweisen der Farben, Jena: Fischer, 1922. Gombrich certainly knew this book as he referred to it in his Bodonyi Review: 'J. Bodonyi, Entstehung und Bedeutung des Goldgrundes in der spätantiken Bildkomposition (Archaeologiai Értesitë, 46, 1932/3)', Kritische Berichte zur Kunstgeschichtlichen Literatur, 5, 1932/33 (published in 1935), 65-75, 66.
18 For a discussion of this see Karl Lepsky, 'Art and Language: Ernst H. Gombrich and Karl Bühler's theory of language' in Richard Woodfield (ed.), Gombrich on Art and Psychology, Manchester: Manchester University Press 1996, 27-41.
19 E. H. Gombrich, 'Art History and Psychology in Vienna Fifty Years Ago', Art Journal, Vol. 44, 1984, 162.
20 His interest in the language of art followed Riegl's, although Riegl was not the inspiration. See Alois Riegl, Historical Grammar of the Visual Arts, translated by Jaqueline E. Jung, New York: Zone Books 2004.
21 E. H. Gombrich, 'Art and Scholarship', Meditations on a Hobby Horse, London: Phaidon 1963, 112.
22 E. H. Gombrich, 'J. Bodonyi, Entstehung und Bedeutung des Goldgrundes in der spätantiken Bildkomposition (Archaeologiai Értesitë, 46, 1932/3)', Kritische Berichte zur Kunstgeschichtlichen Literatur, 5, 1932/33 (published in 1935), 65-75 dated Vienna 1935. Translation by Karl Johns (Los Angeles).
23 'The Study of Art and the Study of Man: Reminiscences of Collaboration with Ernst Kris (1900-1957)' in Tributes, Oxford: Phaidon 1984.
24 Ernst Kris and Otto Kurz, Die Legende vom Künstler; ein Geschichtlicher Versuch, Vienna: Krystall-Verlag 1934.
25 Many of the book's topics surfaced in his later publications but the projected textbook never emerged. Instead, he wrote up and published two of the case studies for The Burlington Magazine: 'Reynolds' Theory and Practice of Imitation', which appeared in 1942 and 'The Subject of Poussin's Orion', which appeared in 1944. On returning to the Warburg Institute after the war he wrote and published 'Botticelli's Mythologies', in the Journal of the Warburg and Courtauld Institutes in 1945 and 'Icones Symbolicae: The Visual Image in NeoPlatonic Thought' in 1948 and 'The Renaissance Theory of Art and the rise of landscape' was published in Gazette des Beaux-Arts, 6e Période, XLI, 1953, 335-60.
26 Ernst Gombrich, 'Topos and Topicality', Annual Lecture of the Society for Renaissance Studies, delivered at University College, London, 10 January, 1975. Gombrich Archive: www.gombrich.co.uk.
27 'An autobiographical sketch', reprinted in The Essential Gombrich ed. Richard Woodfield, London: Phaidon 1996, 34.
28 Literary Estate of E. H. Gombrich.
29 A term borrowed from Bühler (RW).
30 Löwy's expression.
31 'Signs, language and behaviour', 248. I have quoted this passage at length because few commentators seem to have read the paper or they have ignored it.
32 'Signs, language and behaviour', 242.
33 Meditations, footnote 2, 162.
34 Perhaps the reason for this is that while the essay is very rich in suggestions it does not develop its argument in a systematic way but floats from one topic to another by a process of association. To take my analysis any further than I have done would require an extended paper in its own right.
35 E. H. Gombrich, The Preference for the Primitive, London: Phaidon 2002, 273.
36 Meditations, 8.
37 I cannot agree with Umberto Eco's account of the hobby-horse in terms of a theory of pseudo-iconicity in virtue of the stick's linearity: linear objects come in different lengths, weights and structures (string is linear when it is held vertically). The child needs a stable object that he can ride. More importantly, it is the surrounding activity which defines the assigned use. Umberto Eco, A Theory of Semiotics, Bloomington: Indiana University Press 1979, pp. 208-9.
38 See particularly Section II of Theory of Language, 'The deictic field of language and deictic words'.
39 Brunswick and Tolman applied the idea of abstractive relevance to features of the environment that would be biologically important to the organism.
40 E.H. Gombrich, 'The Miracle at Chauvet', The New York Review of Books, 14th November 1996, 8-12.
41 'Aims and Limits of Iconology',E. H. Gombrich, Symbolic Images, London: Phaidon 1972, 2-3.
42 One might add that the Piccadilly fountain was not a Renaissance work of art and therefore out of place in the volume.
43 Gombrich's use of Hirsch caused some irritation amongst delegates at the Venice conference. He would have actually further if he had pursued Bühler's interest in the work of Gardiner's The Theory of Speech and Language. This would have led him directly into the idea of the performative dimension of utterances and, by analogy, works of art. One creates a work of art to do a particular job and the analysis should start by asking for the nature of the job to be done. This fits with Gombrich's last collection of essays The uses of Images: Studies in the Social Function of Art and Visual Communication, London: Phaidon 1999.
44 'Aims and Limits', 2.
45 Elizabeth G. Holt, A Documentary History of Art, II, New York: Doubleday Anchor 1958, 65-70.
46 Holt, 69.
47 Erwin Panofsky, Studies in Iconology, New York: Harper Torchbook 1967, 14-5.
48 'Aims and Limits', 6.
49 Gombrich's footnote: E. Castelli (ed.), Umanesimo e Simbolismo (Padua, 1958) and L. C. Knights and B. Cottle (eds.), Metaphor and Symbol (London, 1960).
50 Gombrich's footnote: K. Bühler, Sprachtheorie (Jena, 1934), pp 185-6.
51 E. H. Gombrich 'The Use of Art for the Study of Symbols', reprinted in The Essential Gombrich, 438.
52 Theory of Language, 208 (Sprachtheorie, 185).
53 'Raphael's Stanza della Segnatura', Symbolic Images, 96.
54 Bühler, Theory of Speech, 210 (Sprachtheorie, 186).
55 'Aims and Limits', 15.
56 There is a little gem of a footnote that illustrates to perfection Gombrich and Kris's earlier work on the interpretation of a figure's expressive meaning in the absence of textual evidence. Symbolic Images, 204-5.
57 See, for example, Kenneth Clark's review of Symbolic Images, 'Stories of Art', The New York Review of Books, XXIV, 19, November 24, 1977, 36-38.
58 'Raphael's Stanza della Segnatura', Symbolic Images, 95-6, referring to Karl Bühler, Die Darstellungsfunktion der Sprache (Jena, 1934). The correct reference should have been to Sprachtheorie : die Darstellungsfunktion der Sprache (Jena, 1934).
59 Roger Jones and Nicholas Penny, Raphael, New Hampshire and London: Yale University Press 1983, 56.
60 This is an enigmatic paragraph that Gombrich sought to clarify in the next but it played on the ambiguity of the expression 'work of art'. In his Bodonyi review, Gombrich characterised late antique art as pictographic. He thought that it was devoid of artistic skill but nevertheless it was an art form that satisfied its audience.
61 I find this paragraph puzzling. The paintings in the Stanza della Segnatura were symbolic images but it would not seem appropriate to call them allegorical paintings as they had no narrative content.
62 'Raphael's Stanza della Segnatura', Symbolic Images, 96.
63 On this topic see C. Hope, 'Religious narrative in Renaissance art', Journal of the Royal Society of the Arts, 134 (1986), 804-18.
64 'Tobias and the Angel' republished in Symbolic Images, 26-30.
65 'Tobias and the Angel', Symbolic Images, 29.
66 'Icones Symbolicae', Symbolic Images, 125.
67 'Icones Symbolicae', Symbolic Images, 124-5.
68 E. H. Gombrich, 'The Visual Image: Its Place in Communication' in The Image and the Eye, Oxford: Phaidon, 1982, 138.
69 See 'Althusser, Ideology and Interpellation' at http://csmt.uchicago.edu/glossary2004/interpellation.htm accessed 30.09.2011.
70 Robert E. Innis, Semiotics: An Introductory Anthology, Bloomington: Indiana University Press, 67-8.
71 The enumeration of books representing different schools of psychology will arouse, in the mind of the specialist, the suspicion that my approach must be fundamentally eclectic. Up to a point this suspicion would be justified, but my selection was not without a bias of its own. If any student of the subject should wish to know at this stage what direction this bias took, I would refer him to the famous joint paper by E. C. Tolman and E. Brunswik, `The Organism and the Causal Texture of Environment', Psychological Review, 1935, which stresses the hypothetical character of all perceptual processes.
Art and Illusion, xxxiv.
72 'The Visual Image', 138.
73 It was a misunderstanding that he shared with Karl Popper. See Karl Popper, Unended Quest, London: Fontana 1976, 74.
74 On which see Michael Baxandall, 'Rudolph Agricola on Art and on Patrons', Words for Pictures, New Haven and London: Yale University Press 2003, 78: 'The Renaissance sense of who was responsible for works and their quality is nimble in that observers can slip to and fro between a sense of the patron as author and a sense of the artist as author in a way that is hard to follow.'
75 See E.H. Gombrich, 'Raphael: A Quincentennial Address', New Light on Old Masters, Oxford: Phaidon 1986 and 'The mastery of Rubens', Reflections on the History of Art.
76 See E.H. Gombrich, 'Patrons and painters in Baroque Italy', Reflections on the History of Art.
Richard Woodfield is Editor of the Journal of Art Historiography. He has edited Gombrich on Art and Psychology (1986), Gombrich's Reflections on the History of Art (1987), The Essential Gombrich (1996), Ernst Gombrich, Dal mio tempo: Città, maestri, incontri (1999) and constructed the Gombrich Archive www.gombrich.co.uk.
Department of History of Art
The Barber Institute of Fine Arts
The University of Birmingham
Edgbaston
Birmingham B15 2TS
United Kingdom
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Copyright Journal of Art Historiography Dec 2011