Commentary on Stephen E. Palmer (1999). Color, consciousness, and the isomorphism constraint. BBS 22(6):923–989

The authors of this commentary are employed by a government agency and as such this commentary is considered a work of the U.S. government and not subject to copyright within the United States.

1. Although seldom discussed, this stereotyped psychophysical performance underlies the common practice of using only two or three normal observers in most vision experiments (Boynton 1966, p. 277).

2. At first glance, it is rather remarkable that Rubin's anomalous trichromats have such narrow spectral loci distributions. This is likely due to Rubin's exclusion of anomals with large Rayleigh match ranges (poor color discrimination), which restricts the pool of anomalous observers to those with relatively moderate differences in photopigment maxima relative to normals. Rubin's study is also a useful antidote to arguments that color boundaries are taught rather than inherent in biological color coding. Anomalous trichromats – despite indoctrination from the same culture as the color normals – stubbornly insist on shifting unique and balanced hues in the direction predicted by opponent process theory (Pokorny & Smith 1977).

3. Minor differences in cone photopigments (Block 1999) don't seem to have much effect on unique hues of normal subjects; their Rayleigh matches are poorly correlated with unique hue measurements. Similarly, except perhaps for unique green (Mollon & Jordan 1997), minor differences in preretinal absorption have little impact on perception of nearly monochromatic stimuli.

4. Hurvich et al. (1968) felt that the bimodality of unique green (Rubin 1961) is an artifact of chromatic adaptation. However, evidence suggests that this bimodality represents two alternative neural pathways for handling short wavelengths (Ingling 1977; Richards 1967). Bimodal distributions are found only when measurements involve adapting fields, large fields, or bipartite fields. Subjects who have longer wavelength loci for unique green (about 20%) also differ from other observers in additivity of spectral lights, rate of recovery of sensitivity following adaptation, and the chromaticity coordinates for “white” (Hovis & van Arsdel 1997; Ingling 1977; Richards 1967). The trait seems sex linked and rare in females (Cobb 1975; Waaler 1967), but is not related to inherited cone photopigment defects (and indeed is uncorrelated to Rayleigh matches; Mollon & Jordan 1997). For a plausible model based on retinal physiology, see Ingling (1977) and Ingling et al. (1978).

5. There are several reasons why this might be so. Mixed cone surrounds in retinal ganglion cells tend to dilute the influence of excess cones (Billock 1996). Also, it might be possible in a rigid array of units to use fixed ratios of L- and M-cone driven units in the construction of Hering channels. Both ideas are belied by new evidence that the cone mosaic is patchy (Roorda & Williams 1999). If unique hues are viewed as a balancing of cone absorptions (like a Rayleigh match), then unique hues would be expected to be invariant (Miyahara et al. 1998; Mollon 1982; Pokorny & Smith 1977). This would suggest that unique hues ought to covary with measurements like the Rayleigh match or with perturbations in a reference “white,” neither of which seems to occur (Mollon & Jordan 1997; Wallstein 1981). Finally, in some nonlinear dynamic cortical color models, the unique hues are switching points in a winner-take-all competition between cortical-hue-labeled lines; because these mechanisms have rectified responses, differences in cone ratios have little effect on the outcome of the competition (Billock et al. 2001).

6. One of us tried out this argument on a famous philosopher. He countered with an obscure mental condition in which a person becomes convinced that his or her spouse has been replaced with an exact duplicate. We’re not sure whether this proves that the qualia associated with the spouse is different or that philosophers are really slippery.