On Color Constancy….and Arthur Koestler’s Treatise “The Ghost in the Machine”

I am reading for perhaps the fifth time Arthur Koestler’s “Ghost in the Machine” (which I heartily recommend!). On page 78 Koestler, in a discussion of a hierarchy of physiological actions, reaches the topic of constancy – of size, shape and finally, color in the human vision process. This reminded me of how a logical explanation for the mechanism underlying the color constancy of vision is defined for the first time in this work. To wit:

The retinal surface contains a geometrically-defined area at 7-8 degrees that is spatially correlated with light refraction in the eye and located at the point where mid-band, i.e., 550 nm, wavelengths interact. Physically, this is the location where rod density is first sufficient to completely surround each cone. It is this spatial arrangement that geometrically defines mid-band. Light interaction at this location provides a wavelength-defining reference point from which all other wavelengths can be inferred/calculated to produce the color signal.

I would add that this geometrically-defined effect is in the nanometer (or, sub-optical wavelength) spatial domain. In fact, the entire hypothesis based on similarly dimensioned “optical wavelength antennas”. I will shortly write more about nanometer technology and its impact on the eye and the vision process. As anyone who reads the popular press knows this technology is under intense development, for example, in the field of semiconductor research. In this context one might even go back to 1993 and read “Optics of Nanostructures” in Physics Today (June 1993). With the development of “quantum dots and wires” one can see the connection with vision – retinal receptors are quantum wires! An example that I have cited – it has recently been found (references cited elsewhere in this work) that when light guides (read “retinal receptors”) are dimensioned in the sub-optical wavelength range light travels outside of the guide. All of the work that has been done modeling retinal receptors as traditional light guides must now be re-thought.

But…back to Koestler (p.78) on size constancy:

“The next stage in processing is very striking – once one starts thinking about it. If you hold the index finger of the right hand ten inches – the same finger of the left hand twenty inches, in front of your eyes, you see them as being of equal size, although the image on the retina of one is twice as large as the other.” (italics in the original)

K. obviously believes, and could not have believed otherwise, that the retina forms the image plane of the eye as depicted in those ubiquitous “inverted arrow” diagrams that populate vision texts…larger objects equal larger images on the retina.. But….on the Fourier plane retina an overall outline sketch of the perceived scene is detected on the all-cone fovea solely at the long wavelength limit of the visual band. Relative image size…? It would seem to me that image size is always the “proper size” for the image perceived .and any constancy is a mental construction and not related to the image formation process……”I expect all objects that I am in the habit of observing to be of a certain size no matter how they appear in my visual scene…and I will not take notice or be surprised”.

I will go on with this later but other things intrude…..

GCH