Science news at the speed of thought

Machines Like Us interviews: Steven Lehar

July 17, 2007

Steven Lehar is an independent researcher who has made a number of radical proposals on theories of philosophy, psychology, biological vision, and consciousness. His most radical theory is that the solid spatial world that we see around us in visual experience is not the world itself, but merely a miniature replica of that world in an internal representation.

This is known variously as the theory of Indirect Perception, Indirect Realism, Epistemological Dualism, and Representationalism. Although this idea is not new -- having been first proposed by Immanuel Kant and promoted by Bertrand Russell, Wolfgang Köhler, and the Gestaltists of the Berlin School -- the idea has never taken hold to become generally accepted, and remains to this day a minority view. In his paper on Gestalt Isomorphism and his book The World In Your Head, Lehar refutes the most common objection to Representationalism, which is the homunculus fallacy. He further argues for the indirect nature of perception by pointing out the curvature of perceived space, or phenomenal perspective, suggesting that it is not a property of the external world.

Interview conducted by Norm Nason, who spoke to Steve from his home in Manchester, Massachusetts.

MLU: Thank you for joining us, Steve. You are a vigorous supporter of the theory of Indirect Perception. Can you summarize this view for our readers who may not be familiar with it, and tell us how you became convinced that this view of perception is the correct one?

SL: The problem dates back to Descartes, and his neurophysiological discovery of the causal chain of vision: Light from the world enters your eye where it produces an image on your retina, which in turn is projected up the optic nerve to the brain. The flow of information is uni-directional: from the world, through the eye, to the brain.

How then does it come to be that we can see the world out there, beyond the sensory surface, as if following the light rays back to their source? How can our experience of the world escape the confines of our head to appear back out in the world? Is experience projected out from the brain like the light from a lantern? This problem has wracked the world of philosophy for centuries, and remains unresolved to this day, although few today are even aware that there remains a problem. Most see it as a pseudoproblem that had been resolved centuries ago.

Descartes' own solution invoked the mystery of an immaterial soul, who indeed views the picture in the brain sent up from the eyes, like a man viewing a television monitor of a remote scene. But the soul supposedly does not see the picture on the monitor; instead, it sees the world itself directly, out where it lies, seeing transparently backwards through the optic nerve, retina, and lens as if they didn't exist! This explanation is an unhappy compromise between the irrefutable neurophysiological facts and our personal experience of vision. For we do indeed seem to experience the world out where it lies, not as a picture inside our head. Many modern theorists have equally confused views on this issue.

But the problem with direct perception becomes perfectly clear as soon as we set out to replicate that principle in a robot intelligence. How on earth do you get a robot to "see" beyond its sensory surface and to experience the world itself directly, out where it lies? The whole concept becomes incoherent when you try to implement it in insensate physical matter. Like Descartes, these people still cling to some magical mystical concept of Mind which can never be replaced by mere machines. They are the modern equivalents of vitalism, the theory that there is some vital essence inherent in living conscious beings, which can never be reduced to plain physical terms.

But as soon as we acknowledge that mind is nothing other than the electrochemical functioning of the physical brain, then conscious experience reduces to an information theoretic problem. Visual experience has an information content, and information cannot exist without some kind of carrier to carry that information. Experience cannot exist in the abstract, it must be encoded in some information processing system, and our experience is encoded as a physical state of the physical brain. No aspect of the external world can possibly appear in our experience except by being represented explicitly in the brain.

MLU: How do you believe that experience is represented in the brain, then?

SL: Well, this has been the great stumbling block that has led to the current impasse. Everything we know about neurophysiology seems to suggest that experience is encoded in the brain as electrical activity in networks of neurons. What is prominently absent from our neurophysiological understanding of the brain is the question: Where are the moving colored pictures in the brain, that we know from our experience to be in there?

The consensus in modern neuroscience seems to be that there are no pictures in the brain, because we have not found them, and modern philosophy has attempted to conform to this neurophysiologcal reality by claiming that there are no pictures in our experience either, so don't even bother looking for them in the brain! But our experience is clearly spatially structured, and not just our perceptual experience, but also our experience of dreams and hallucinations, which are obviously constructs of our brain. So my first point is that however they might be encoded in the brain, there are in fact explicit moving colored pictures in the brain that correspond to our experience of the world, and if modern neuroscience cannot find them, that is a problem for modern neuroscience, not for the notion of pictures in the brain.

My own proposal is that the brain encodes spatial images by way of standing waves in a Fourier type representation. This is the more speculative component of my theories, and all the details are not yet fully worked out. However, there is solid evidence that supports some aspects of the harmonic resonance theory of brain function. For example, a number of perceptual grouping phenomena are best explained by a harmonic resonance model. But there is abundant evidence of a more indirect sort, in the worlds of music, art, ornament, architecture, rhythm and dance. One of the greatest unsolved mysteries of human nature is the question of why the resonant vibrations of musical instruments promote the most intense emotional and spiritual experiences in the human soul.

What is it about patterned vibrations in the air that move us so? And why do we feel compelled decorate our clothing, wallpaper, pots, carpets, and just about any item that we use, with periodic and symmetrical ornamental patterns, especially those objects of particular religious or ceremonial signficance in which we place the most value? And why do we humans tend to respond to auditory rhythms by spontaneously oscillating our bodies in the periodic and symmetrical motions of dance? Why does it give us pleasure to oscillate our bodies in this manner? All these phenomena are actually evidence for a harmonic resonance mind, a mind that organizes its experience in terms of patterns of symmetry and periodicity across space and time. The harmonic resonance theory promises to finally link the worlds of science and art through the scientific phenomenon of harmonic resonance in the brain, a physical phenomenon with inherent aesthetic appeal.

MLU: When you say that pictures in the brain correspond to our experience, do you mean that they are spatially analogous?

SL: We don't know much about how images are actually represented in the brain. But one thing we can determine for certain by inspection is that the representational strategy used by the brain is an analogical one; that is, the objects and surfaces in our experience are not represented by some kind of abstract symbolic code, but by explicit visuospatial analogs or effigies of the external objects they represent, because that is how they appear to us. This follows logically as soon as we recognize that the objects in our experienc are separate and distinct from the actual real-world objects that they represent.

MLU: Some of the most intriguing evidence you site for Indirect Perception is the phenomenon of visual perspective. How does this support your thesis?

SL: Yes, that was one of my most surprising discoveries. I spent months pondering whether the sides of a road, or the walls and ceiling of a long hallway, appear to converge toward a vanishing point in the distance, or whether they appear straight and parallel and equidistant throughout their length, as we know them to actually be. In the end I concluded that the answer is both. They appear to converge to a point, and at the same time, they appear straight and parallel and equidistant. Our experience captures both realities in a single representation. That is, it is the same road (in our experience) that is simultaneously straight and parallel, and also converging toward a point.

This is irrefutable proof that the world of our experience is separate and distinct from the world that it is an experience of. It demonstrates conclusively that our experience has a variable representational scale. Things in the distance are both smaller by perspective, and at the same time undiminished in size (perceptually), just as in a museum diorama, or a theater set. What was amazing to me with this discovery was that this fact is so plainly obvious for all the world to see, even to young infants, and yet as we grow to adulthood we learn to ignore it, until it becomes so invisible to us, that many theorists, like Eric Schwitzgebel, adamantly deny that they see any convergence at all in their experience.

MLU: Johnjoe McFadden, Professor of Molecular Genetics at the University of Surrey, has proposed a harmonic resonance theory that is similar to the one you describe. He suggests that the brain's electromagnetic information field is the physical substrate of conscious awareness. Do you agree?

SL: Yes, Johnjoe McFadden's CEMI theory of electromagnetic fields in the brain is one of the very few proposals out there to address the issue of explicit spatial representations to account for the explicitly spatial nature of experience. In fact, my own harmonic resonance theory can be seen as an extenstion to a static electromagnetic field theory. But any system of electrostatic fields, whether in the brain or elsewhere, also automatically includes dynamic oscillatory behavior with no additional hardware required, unless it is actively suppressed. Feedback systems tend to oscillate, and I believe that instead of suppressing the oscillations, nature makes use of the oscillations to generate standing waves of oscillating electrostatic fields, because a standing wave system can encode a lot more rich spatial information than a simple static field. In fact, there is compelling evidence from a wide variety of sources, in support of some kind of harmonic resonances in the brain. So I believe that McFadden's theory is on the right track, but does not go far enough.

MLU: What is the practical difference between static electric fields and oscillatory fields?

SL: A static electric field, like the field between the plates of a capacitor, performs a kind of interpolative function, filling in the empty space between the two plates with a continuous electrostatic gradient from the positive plate to the negative. But the pattern itself is established by the plates, based on their electrical charge, shape, and spacing.

An oscillatory electric field, on the other hand, is like the field that appears between capacitor plates on which the voltage is constantly oscillating -- for example with a sinusoidal alternation. This will set up a pattern of standing waves between the plates which will be in constant alternating flux, but the alternation will be fixed about a pattern of nodal points between the plates, that is, points, lines, or surfaces that separate regions of opposite electrical potential. This kind of system has a great deal more potential for defining spatial patterns which are not determined so much by the geometry and configuration of the plates, but by the frequency and waveform of the electrical oscillation. A single system of plates can thus create a whole repetoire of geometrically related spatial patterns, like the standing waves of acoustical vibration that form within a bugle as the bugler blows different notes.

The patterns in a linear resonance, like the tube of a bugle, define simple alternating patterns along the length of the tube. But standing waves in a two-dimensional resonator create a far more complex repetoire of two-dimensional patterns, as demonstrated by the Chladni figures, produced by standing waves in a flat steel plate. And the standing waves in a three dimensional volume, like the space between capacitor plates, create an even more complex repetoire of volumetric spatial patterns that subdivide the volumetric space into periodic alternating patterns in three dimensions, as seen in the acoustical standing waves in a three-dimensional box shown in Paul Falstead's most excellent Math and Physics Applets. Here we have a simple homogeneous system that exhibits the most exquisite pattern formation capacity, and it is this capacity which, I propose, is exploited by nature as a spatial representation in the brain.

MLU: It sounds as if an analogy to music is not incidental.

SL: That is right. The harmonic resonance theory finally offers a solution to the age-old question of what music is, and why it triggers such emotional response in human listeners. According to harmonic resonance theory, music is the language of the brain, and the powerful emotional response evoked by these inanimate resonances reflects an unconscious recognition of the essential affinity between mind and music. I propose that the musical instrument represents man's first modest success at replicating the physical principle behind biological computation.

MLU: Fascinating. And art then may be man's attempt to do the same using the visual sense.

SL: Yes, thats right. We see this most clearly in visual ornament, the decorative patterns used by all cultures throughout history to decorate our clothing, pots, wallpaper, carpets, vases, lamps, and such, in fact, virtually every artifact we use, most especially those items of symbolic and ceremonial use in which we place the most value. Although there is considerable variation in the exact styles used by different cultures, the principles of symmetry and periodicity in both simple and compound hierarchical forms are common to the ornament of all cultures. It is no accident that symmetry and periodicity are also characteristic properties of harmonic resonance standing waves.

We also see symmetry and periodicity in virtually every dimension of aesthetic activity, including music, rhythm, poetry, dance, and architecture. Even the repeating patterns in animal bodies, the periodicity of ribs and vertibrae, fingers and toes, the bilateral symmetry of face and body, the pentagonal symmetry of the starfish, the symmetry and periodicity of plants and flowers, and the markings on animal hides, such as zebra stripes and leopard spots, have all been explained by a chemical harmonic resonance known as reaction-diffusion. Harmonic resonance theory finally unites the worlds of science and aesthetics by providing a physical mechanism to explain our aesthetic appreciation for symmetry and periodicity in the world.

MLU: If the harmonic resonance theory is responsible for such deep-seated, intense emotional experiences, could it also be the basis of religion?

SL: Not beyond an explanation for the significance of music, dance, and chanting -- which form a prominent part of many religious traditions -- and as an explanation for the elaborate level of ornament seen in churches, mosques, and cathedrals of all faiths.

But the other aspect of my theory of perception -- that is, the philosophy of indirect perception, or representationalism -- does indeed account for one aspect of spirituality observed across the vast majority of religious beliefs: the belief in an alternate reality of the spirit world, which is almost universally recognized as immaterial, usually invisible, insubstantial, and not subject to the laws of the material world. Although superimposed on the physical world, the so-called spirit world is thought to be in a separate plane of existence, which can occasionally make contact with the physical world, often by way of a human medium.

I propose that the dualism inherent in the notion of separate physical and spiritual realms is an unconscious recognition of the ultimate unreality of the world of perception. This is suggested by the existence of dreams and hallucinations, that clearly reveal that there is not a direct relation between reality and experience. This in turn casts doubt on the objective reality of all experience.

Once we recognize the world around us as an internal representation which is distinct from the world it represents -- and once we recognize the ability of the brain to construct vivid spatial entities in that representation -- the phenomena of dreams, hallucinations, and ghostly spirits become much less mysterious; they are revealed to be merely spatial thoughts in the brain, superimposed on perceptual structures, also in the brain.

MLU: Moving from human minds to artificial ones: Searle, Penrose and others have argued that constructing strong AI is not feasible under any circumstances. In his famous "Chinese room" thought experiment, Searle argues that no matter how advanced a computer is made, it can at best be only a simulation or model of a mind -- not an actual mind. How do you feel about this?

SL: I find that to be a very peculiar point of view, because it seems like a return to vitalism, the idea that mind is something other than the mere functioning of a material brain. No, problem with strong AI is not that mind can never be replicated in a physical mechanism, but rather, that strong AI is the wrong paradigm for expressing mind. In fact, the "Chinese Room" thought experiment demonstrates this most clearly -- where the translation to and from Chinese is performed in a robotic rule-based manner without any understanding of the overall context; the kind of process that would translate the biblical phrase "The spirit is strong, but the flesh is weak" as "The vodka is good, but the meat is rotten" -- a classic of machine translation errors.

Instead of denying that mind can ever be replicated in a physical machine, we should focus our effort on the question of what it is in the nature of mind that gives it the kind of overall contextual understanding that seems so profoundly lacking in digital computers and other rule-based systems. I think the Gestaltists were on the right track, identifying an emergent holistic field-like computational process as the root of creativity and understanding in mind, a kind of process that cannot be easily reduced to simple step-by-step rules.

MLU: Finally, Steve: why is it important to understand how the brain functions?

SL: Mankind finds itself at a crucial crossroads: We have the technological ability to destroy our planet and exterminate ourselves, but we have not yet developed the wisdom to know what to do with such frightening power. We are endlessly embroiled in acrimonious partisan debates over what path we should take next, with no reliable methadology for determining which is the right course. I sincerely hope that we will discover the essential principles behind how our mind works, in order to manufacture artificial minds that can be tested and tuned for optimal wisdom and judgment, and then copied and replicated in numbers in a form that does not disappear irretrievably with every new generation. This will complement our great storehouses of global knowledge, represented by the accumulated literature of the world, with a great new storehouse of global wisdom as implemented in an artificial mind, without the volatility and inconstancy of a frail biologically based intelligence. I just hope we can achieve this objective before it is too late.

Read more Machines Like Us interviews here.