Turrell and Visual Science:

Perception as Unconscious Thought


Shinsuke Shimojo
Published in supplement to James Turrell:
Where does the light in our dreams come from ?
The official catalog book for the exhibition tour in Japan, 1997.


1. An eye-opening view of visual science in works by James Turrell The motivation for analyzing the works by James Turrell from scientific viewpoints, such as those of physiological optics, visual neurophysiology and psychology of visual perception, is twofold. Using these analyses, one can better understand the medium and techniques employed by Turrell. If his style is "to deal with lights just as other painters deal with paints" and he achieves it based on a knowledge of visual science, then we need to take a closer look at this basis to further appreciate his works. Thus, to understand his materials and methods is the first motivation.

However, a more intrinsic relationship seems to exist between Turrell’s art and visual science. He regards the aim of his works as ’seeing the act of seeing’ and ‘realizing the process of perception itself.’ Turrell attempts to accomplish this by manipulating light and space, not the actual forms and surfaces themselves. Visual perception is brought about by light, and the solid image perceived is merely the result. Therefore, only by manipulating and showing lights, not images, he can reveal the process of visual sensation.

It is not easy, however, to demonstrate the actual process of seeing, mainly because the major part of visual perception is automatic and subconscious. For example, when we see a rectangle which is opaque at a distance, though still within our reach, we perceive its color, form and distance. Yet, we could not consciously experience the optical process in the eyes and neurophysiological process in the brain which have actually caused this mental event of perception. It is an unconscious process which takes place despite our conscious awareness of the percept. Although introspection allows us to describe the quality of perception phenomenally, we usually cannot describe the underlying neurophysiological process. In the extreme cases where we do not have the slightest knowledge about the underlying mechanism, it still brings our perceptual experience to us correctly. (An observer who does not know the opponent color theory could still experience an aftereffect whose color is opponent to the adapting color, and those young children who would not understand the principle of binocular stereopsis still could perceive depth in stereograms.) Turrell seems to use his experimental-psychological methods to bring these implicit processes into our awareness.

2. The art of additive color mixture Art has mainly concentrated on the media of painting in the past, and even today this is still true. (The increasingly popular technique of computer graphics can also be considered as a variation of painting in a broad sense.) Although the skills and techniques have changed, the basic concept is the one which has gradually developed over the long history of painting.

The traditional mindset in painting becomes an obstacle to overcome when Turrell tries to extend his methods to show the very formation of perceptual space by manipulating lights themselves. Over the years in traditional styles of painting, a wide variety of materials and new types of brushes and canvases have offered an increasing number of technical options to artists. However, the technology of generating lights in order to achieve Turrell’s creative purpose still remains primitive. This does not stem from the lack of technology but because the need for this art type has never fully developed. Furthermore, this lack of development may have resulted from basic conceptual differences. The perceptual theory of lights advocated by Turrell differs substantially from the fundamental theory of painting.

We can observe the difference by comparing the ways of mixing colors. (1) Traditionally the subtractive color method is used, where mixing red, green and blue produces black. Turrell uses the other method, that of additive color mixture, where the same color mixture produces white. Turrell recently recalled "I went to art school but it did not help me in my work as it mainly taught the theory of subtractive color mixture."

As will be mentioned later on, these differences may be further emphasized. While color technology in art depends upon limited reflective optics, visual perception is actually the opposite process (inverse optics). Reflective optics is over when it describes the wavelength, amount and ratio of the lights reflected from surfaces. However, this is exactly where the theory of inverse optics (= visual perception) starts. Its main goal is to explain how these lights are absorbed by the retina and lead to the quality of perceptual surfaces and space. These complicated and little known neural functions and mental capacities are hidden behind our experience of perception. And Turrell realizes this clearly.

3. Surface color, film color and volume color Another great achievement made by Turrell, using lights instead of paints, can be understood in terms of ’surface color,’ ‘film color’ and ‘volume color.’ (2) In another way, these color types may be defined as reflection, transparency, and volume. This terminology, though not widely known, becomes a powerful tool when we attempt to describe his works from a phenomenological viewpoint.

It is Turrell who has been aware of this distinction and very successful in introducing the experimental techniques of film and volume colors, while throughout the history of art, technology had been mainly limited to surface color. One of Turrell’s best examples can be seen in the idea used for ‘Aperture.’

Those who enter the ‘Viewing Space’ (the first compartment of ‘Aperture’) first see a rectangle painted in color and hung on the front wall. On approaching, one realizes that there is something out of the ordinary about it. The degree of sky-like transparency seems to have suddenly increased. Then they see a square window. What has started out as a flat 2-dimensional image on the wall has turned out to be light filling the ‘Sensing space’(the second compartment). The initial impression is that of volume color, specifically ‘aperture color’ and ‘film color.’ But then, by making free use of techniques, he made it into ’surface color.’ This space uses many visual techniques and phenomena, such as dark adaptation, night myopia, empty myopia, the elimination of cues for form and distance (Turrell calls it neutralization of form), interpolation and formation of surfaces by contour, propagation and leveling of color, and figure-ground segregation in relation to closing of contour. As a psychologist of visual perception, these are issues that are very intriguing to me.

In his creation known as ‘Wedge Work,’ he projects light onto a space, takes away the form and depth cues almost completely, and captures the remnants using a contour which is projected through a slit. This is all he provides, and the viewer’s visual system automatically takes care of all the rest of surface formation processing based on the contour and color. The result is an opaque surface that appears out of nowhere. An expert on visual perception would describe it as deprivation of natural perceptual cues and re-combination of them. Moreover, the visual system is left at unstable levels of dark adaptation. No wonder the impressions that people experience vary from person to person, ranging from unrealistic to super realistic, purified, sharp, uneasy, shimmering, or expanding in depth. This type of input to the visual perception system at a lower level forces emotional and cognitive completion at a higher level. Perceptual interpretations received initially do not differ greatly within individuals, yet variations occur at the higher processing levels.

Two other works by Turrell, ‘Gas Works’ and ‘Telephone Booth’, feature ‘volume color’ employing the Ganzfeld technique which is well known in the psychology of visual perception. (3) In the Ganzfeld, there is no contour, no surface and no form, thus no object. Color and brightness are homogeneous and no particular depth or movement is indicated. Naturally enough the Ganzfeld has become one of the favorite techniques of Turrell, who advocates ‘neutralization,’ that is the emphasis on light and space by minimizing image or surface. Also it seems to be a matter of course that Turrell has no interest in so called technology art because it tries to build up rather than eliminate forms.

4. Cone and rod systems: physiological optics of twilight With the full use of physiological optical devices around the exhibition room, Turrell throws observers into a carefully adjusted semidarkness to get them ready to experience his work.

It has long been known in visual neurophysiology that the human retina has two major information processing systems corresponding to the two photoreceptor types. These systems, the cone system and the rod system are supposed to have different visual functions, to see in brightness and darkness, respectively. (4) In the process of adapting to darkness, visual processing quickly shifts from the cone system to the rod system. Also, which of these systems will be used depends on the degree of brightness (or darkness). Given an appropriate time for adaptation to darkness, the pupils of the eye open wide to compensate, resulting in a more efficient absorption of light into the retina, which then creates a very clear image. The rod system thus gradually becomes dominant, and a space filled with light and without definite form and depth, emerges. As the retina becomes more sensitive to short wavelengths and less sensitive to long wavelengths, only blue becomes distinct (the Purkinje phenomenon). It is in this way that an unusual color sensation is experienced. Turrell seems to have found a huge space for some freely floating ‘thoughts without words’ in the twilight zone where photopic and scotopic visions are pushing against each other.

Not only in the physiological space on the retina but also on the axis of physiological time, Turrell demonstrates his experimental ability. In ‘Gas Works’ and ‘Telephone Booth’ mentioned above, a strobe light is turned on and off evenly on the visual field, with subtly adjusted frequency, and volume color by Ganzfeld as its background. This strobe light is imprinting an afterimage, (5) as well as synchronizing the activities of neurons in the retina, or causing the multiple interactions such as interference and lateral inhibition. (6) An illusion of dynamic texture known as the Purkinje image thus occurs.

It seems to me that Turrell must have repeated research trials and errors in his parameter space, where data is scarce and therefore even a visual neuro-physiologist could not predict perceptual results clearly.

5. Ambiguity and multi-stability of visual perception It is an important function of visual perception to provide the observer with an interpretation of the retinal input that relates to the real world. Thus for instance, the observer can perceive overlapping parallelograms and squares not as such but rather as a wireframed cube. If the formation of a retinal image is considered an optic process, visual sensation could be the process of ‘inverse optics’ which calculates the process backwards. To make the matter more complicated, however, there are many possible real world interpretations, given a particular retinal input; there is ambiguity in this particular sense. It is often pointed out that the essence of perception lies in the process of solving this ambiguity. (7)

Suppose we see a blue rectangular area which is clearly outlined on a surface such as a wall. Immediately, there are already several possibilities. One is that a blue board is hung on the wall, and another is that it is part of the wall which is made of different material. There is yet another possibility, as in the work ‘Aperture’ which I mentioned earlier: the blue field may be a window cut into the wall and the blue is the light filling in space outside the window. Finally, there is even another remote possibility that the rectangular shape may be floating in the air in front of the wall.

This last interpretation may not happen so easily because visual perception has a strong tendency to localize or ‘attribute’ the image onto a visible surface of unambiguous depth. In fact, an afterimage imprinted by a strobe light, looks as though it is on the surface at which one is gazing whether it is your palm or a wall. Nonetheless, Turrell created an unlikely perception in his work ‘Blind Sight.’ He again forced observers to lose orientation in the darkness, and removed reliable spatial cues such as the frame and texture of the wall. As a result, observers perceive the colored object and the halo around its edges at a much closer distance within their reach. Furthermore, a ‘crawling effect’ can also take place where the image begins to move around in the air (probably due to minute eye movements and fluctuation of lens accommodation).

6. Perception as a fact Some critics consider Turrell’s methods unrelated to illusion demos and assume that he has no interest in illusions. I cannot agree with this opinion. I feel that the relationship between his work and visual illusions is much more complicated. It may be true that he is not interested in which percept is the ‘correct’ one. But this is also the case for psychologists of visual perception studying illusions.

Both psychologists and Turrell have an interest in the normal functions of the visual perception system that will bring about both illusions and ‘correct perception.’ To probe into them, they pay special attention to stimulus situations that lead to perceptual ambiguity and optimize the stimulus parameters. I have already described the essence of visual perception as a process of ambiguity solving. Many of Turrell’s works deal with visual effects similar to illusions studied in experimental psychology. These effects in general reflect the real characteristics of unconscious neural information processing.

With regard to this, Turrell has emphasized that his psychological background is not that of clinical psychology such as Freud and Jung, but that of experimental psychology. Although some psychology schools tend to associate themselves with him due to superficial similarities in phrases such as ‘background of psychology’ and ‘ light and space, not image or focus,’ they are merely wrong.

For example, we cannot regard him as a Gibsonian, because his way of removing cues for form and depth is highly experimental and reductionistic, quite the opposite to the Gibsonian philosophy. Moreover, as it has been mentioned already, what is emphasized in his case is ambiguity and multi- stability of perception (Gibsonians would either not admit the existence of these aspects of perception at all, or deny their importance at least.)

By the same token, it is also quite inappropriate to position his work as a ‘counterattack against minimalism’ or even ‘agreeing with minimalism’. The impact that his works carry is paradoxical with their rich, emotional and spiritual outcome. It is paradoxical in a sense that critical cues are minimized to define the contents of ’seeing’ at a lower level of visual information processing, whereby cognitive and emotional systems at a higher level are left with more processing or required to process more.

Turrell is not interested in anything symbolic. He seems to be standing on the horizon where only facts of perceptual experience matter. Even young children who know nothing about trends of thought in art and psychology can still enjoy his works. I imagine that most of his works would give the same perceptual experience even to monkeys, which is something that could be tested experimentally (although the emotional and cognitive responses would of course differ).

7. Thinking without symbols My review has revolved around visual sciences, particularly the psychology of visual perception. My intention was not to comment on Turrell’s artistic motif as I am no expert. However, perhaps it is not meaningful to totally isolate the visual scientific parts from the artistic ones in his works. These are tied together, as it is his aim to reveal the formation process of perception in the act of seeing (and indeed, this is the ultimate purpose for most visual scientists).

Despite this, it may be said that Turrell chose to be an artist and not a visual scientist for a reason. There should be differences between the fields. One difference lies in the purpose. A visual scientist views perceptual phenomena as stimulus variables, where the phenomena are merely dependent variables which could be predicted from physical stimuli that are independent variables. Only a limited set of the dependent variables in the experience of perception are of concern and all the others are discarded. In studying the effects of binocular disparities on depth perception, for instance, all elements other than the mere impression of depth will be abandoned as noise, including the sharply defined clear impression of space, an extraordinary feeling of relaxation, feelings of familiarity and nostalgia, together with an impression of a glimpse of the cosmos, etc. Sometimes even more fundamental properties such as color, brightness, movement and texture are all ignored as being unrelated. The perceptual attribute selected for the experiment is also constrained by theory and hypothesis. The discipline of visual science limits its scope for itself.

Accordingly, the experimental manipulation which repeats subtle adjustment of stimulus variables in order to experience the very primitive moment when the perception occurs is not allowed, regardless of the scientific and experimental nature of the method. Turrell’s intention to produce a medium through which we could be ‘thinking without language,’ or ‘not no thinking, but not thinking with language’ would not be welcomed in formal scientific circles. For it to be accepted, everything must be re-written into explicit scientific terms.

Perhaps this has something to do with the reason why Turrell did not choose to be a visual scientist. People are often saying that ’subjectivity, emotion and inspiration – all these are not relevant to science, but are cherished in art.’ I hate to hear this. Scientists are often excited about their discoveries, and this emotional outpouring often occurs. Intuition and subjective impressions are often the key to another important discovery. Scientists can be amazed at, or even afraid of, the very existence of the universe and self, something which is yet unexplainable in modern science. It is not an exaggeration to say that this very feeling of awe drives them to inquire into their issues deeper and farther. The only difference between art and science is the means of communication. Any scientific discovery, however impressive and exciting it may be, must be supported by the language of numerical formulas and symbols. Unlike the scientists, Turrell seems to enjoy expressing the experience in purely visual language.

The relationship between Turrell and visual science is thus intimate and unique in a complicated way.

8. Towards the Crater It is probably more appropriate to say that the era we live in has inspired Turrell, rather than that he has merely become successful in the contemporary art scene. The relationship of science and modern times has brought about changes to art and society, thus providing an opportunity for him to create something different. This is well proven in his ‘Roden Crater’ project.

This project, which took 20 years since the idea was born to complete the initial groundwork alone, is about to plunge into the first stage of construction. The total construction cost of the first stage is predicted at $85 million and over $30 million will be used for the purchase of further land in the near future. A number of astronomers and architects joined together for the sake of an artist with a dreamy grandiose idea. Such a phenomenon can be only interpreted in the context of the global needs of the times.

Turrell tells us that he simply desired to experience the nature of perception in natural surroundings. From such a rather naive motivation, he searched around to discover, and become fascinated with, craters. At present, the construction of Roden Crater has just completed only the initial groundwork, in which the form and height of the crater rim has been perfectly modified. Already, this creation raises many challenging questions to psychologists and visual scientists. (See my ‘Notes on Roden Crater’.)

Turrell will continue to inspire us to think over our own experiences without words.


(1) In painting, the principle of subtractive ‘color mixture’ is applicable, in which the colors become darker by mixing. For example, the primary color red changes to dark yellow or brown when mixed with green, and if blue is added to this, black would result. This is due to the simple fact that the pigment color depends upon the reflected light from the surface layer of particles. The surface which reflects light with a particular wavelength or combination of wavelengths appears as a certain color, and looks bright or dark according to the ratio of reflection versus absorption. The principle of additive color mixture, on the other hand, applies to color mixture by light. For example, when red and green spot lights overlap on a white wall, the result is bright yellow approaching white. The long and the middle wavelengths of light will be simultaneously absorbed into the retina to yield the impression of yellow, and the more the reflected light increases, the brighter it is perceived. In addition, if a blue spot light is added to this, it offsets the yellow, equally stimulating the three color channels, that is, the three cone types, thus resulting in the formation of a bright white spot light.

(2) The categorization of perceptual color qualities into ’surface color,’ ‘film color’ and ‘volume color’ was originally proposed by a Gestalt psychologist, David Katz. Surface color refers to the color of an object’s surface, and needless to say, it depends on the reflective characteristics of that surface. It is interesting, however, that even though the features of the illuminating light change, we can still see the original ’surface color’. (This is called color constancy.) The illuminating light varies in wavelength and color, but we can perceive it as not having a changing surface color. Film color refers to color which is flat but not explicitly localized in depth. ‘Aperture color’ is similar to this, an example of which would be the sky cut into frames by a window. Aperture color is the dominating element in the ‘Roden Crater.’ (See ‘Note on Roden Crater’.) Film color occurs when depth cues are eliminated and color is generated only by the penetration of light through space, yet the framework is clearly localized in depth. Finally, Katz refers to color which is not even film color because of the lack of a framework, and therefore appears to be vaguely and homogeneously filling in the space, as ‘volume color.’ It is something like the state of a dense fog.

(3) Ganzfeld refers to a visual field with perfectly even brightness which excludes all cues for form, surface, shading and distance, where only a vague amount of light with uncertain depth just like fog, appears to fill in. That is, only volume color can be perceived. This phenomenon was studied systematically for the first time by a German psychologist, Wolfgang Metzger. Although this can be simulated easily by applying half cut ping-pong balls to the eyes, Turrell avoids this method as he probably thinks that the constricted application of balls to the skin will prevent the perceptual extension of space.

(4) The ‘cone system’ in the human retina consists of groups of photo receptors, each of which responds to one of the three primary colors. These receptors enable excellent color vision and spatial resolution due to their high density in the fovea. The ‘rod system’ on the other hand, has a high sensitivity to light so that it is powerful in the dark, and is more dominant in the peripheral visual field. This system has an excellent temporal resolution so that it can detect movement and flicker well. Owing to this system, we can see slight movements of people around us in the darkness of a movie theater as our eyes become adapted to the darkness. Similarly, it is easier for us to see in the peripheral visual field when we look for a star on a dark night.

(5) As is well known in retina anatomy, light penetrates through blood vessels and other neuronal layers and is absorbed by a group of photoreceptors in the back of the retina. Neural signal processing advances from here to higher level layers, thus moving backward in the direction it comes from.

(6) All neurons send messages known as impulses through their ‘axons’, and information is encoded by the timing and frequency of these impulses. A neuron does not necessarily represent an object or an attribute, rather it is considered probable that populations of neurons which are overlapping each other encode multiple attributes all together. Information transmitted through the axons is then converted to chemical substances in the synapse (the nerve junction), causing the impulse to leap to the next neuron. Parallel interactions, such as facilitation and inhibition, arise among many synapses and enable the brain to perform complicated information processing and to learn from experience.

(7) Shimojo, Shinsuke. Adventure of Vision: From Illusions to Cognitive Science, Sangyo Tosho, Tokyo, 1995.