Gombrich on Semiotics

I'm going to begin my talk on Peirce next week with Ernst Gombrich's preface to the 2000 edition of Art and Illusion. This short piece fascinated me as much as the contents of the rest of the book when I first encountered it nearly 20 years ago. Art and Illusion is about the relationship between art and nature, and the Greek idea of mimesis. The relationship between art, pictures and signs is clearly an important sub-topic in this, and this is what Gombrich wrote his new preface about - at a time when semiotics was much discussed in the art schools (late 90s, early 2000s). His aim was to correct the current fashion which argued from a constructivist position that all images were signs, that the Greek idea of mimesis was nonsense. Gombrich begins:

[the] 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.

Gombrich argues that this reaction is overstated: the thirst for illusion is unabated - the goal of mimesis captivates the imagination. Gombrich, always ahead of his time, points out the technological advances in pursuit of mimesis:

Simulators were developed for the training of pilots, who put on a helmet through which their eyes were fed the appearance of an environment rushing past, which they were asked to control. More recently, so called "virtual reality" has been perfected, which allows us not only to see and hear an invented reality but even to touch it with specially constructed gloves. I do not know whether this device will, or can become a medium of art; all that matters in the present context is the undeniable evidence that images can be
approximated to the experience of reality

Gombrich talks about the 'mental set' - the field of expectations - through which signs are interpreted. He points out the playfulness in the interpretation of signs, and the shifts in mental set. For example, the puppet theatre which might transfix the child's imagination in a story suddenly disrupts this expectation when the giant puppeteer's hand appears in the scene to move a character.

What Gombrich appears to be talking about are the constraints within which signs are interpreted: that a sign is not a construct of some individual mind, but that it is the result of a game played within multiple contexts (or constraints) of sensory stimuli, life experiences, expectations, education, social situations, and so on. The game of mimesis is played between image, perception and illusion, among many other things.

The contributing factors in the game are additional descriptions. He says:

A string of ovals can also be an ornament purely used for decoration, as in this case: 0000. But add the word "PLUM" underneath  and you transform the mental set: the oval no longer appears to stand on a neutral background, it is surrounded by an infinite halo of space, because we expect plums to be solid, and not only to be edible, but also graspable—an effect we can further enhance by the suggestion of a foreshortened stalk and leaves.

He comes to the crux of the issue, highlighting the importance of the game that is played in recognising a sign:

We come to realize in such cases that the required mental set did not precede the reading, but followed in a rapid feedback process. Where signs and images appear together on the page the feedback works almost instantly—witness the ease with which our youngest read so-called comics, combining pictures with a simple story. 

The difference between images and signs, then, does not lie in the degree of iconicity or conventionality. Images can function as signs as soon as they are recognized. We need only think of the labels on cans to realize that a perfect iconic image can function as a sign.

Gombrich tells a story about Constable whose judgement about early photography is very revealing in both his and Gombrich's attitude to the relationship between the image and nature:

In 1823 Constable visited a sensational display, the diorama constructed by Daguerre, later the inventor of the daguerrotype. "It is in part a transparency," he wrote, "the spectator is in a dark chamber, and it is very pleasing and has great illusion. It is outside the pale of art because its object is deception. The art pleases by reminding, not deceiving."

In reflecting what Constable might have meant by "outside the pale of art", Gombrich says:

Would we go quite wrong in suggesting that, for Constable, art had become something like a game of skill, with its own rules, which must be kept free of labor saving devices? To deceive the eye is to cheat, for the painter must please by reminding, just as the playwright of Shakespeare's Prologue must work on our "imaginary forces." Fidelity to nature has to be achieved within the limits of the medium. Once this compact between the artist and the beholder is destroyed, we are outside the pale of art. Indeed, as soon as Daguerre's and Fox Talbot's mechanical methods entered the field, art had to shift the goalposts, and move the pale elsewhere. 

There's something very profound in what is it to remind rather than deceive. I wrote something about this with regard to music a few years ago: http://www.emeraldinsight.com/doi/abs/10.1108/03684921111160304. Art reminds by overlaying descriptions on top of one another. I think its interplay of multiple descriptions reminds us of the interplay of multiple descriptions in our lived experience. To deceive us of reality is to identify and reproduce as faithfully as possible the descriptions of actual experience. Since the actual experience of one person and another is different, this deception necessarily abstracts from individual experience the principal descriptions which it sees to be universal - some of these abstracted descriptions can be taken as 'signs'. The complexity of the interaction of abstracted descriptions is never the same as the overlaying of multiple descriptions to produce complexity.