Erickson's article raises important semantic issues, some of which echo lucid arguments previously put forward by him about the extent to which findings at one level of analysis are often applied, inappropriately, to a different level of analysis. For example, Erickson (Reference Erickson, Le Magnen and MacLeod1977) cited “Brewster's Fallacy,” whereby Brewster inferred, from studies at the psychophysical level, that at the stimulus level, there were three kinds of light – red, yellow, and blue. Some of the arguments in the present article are complicated by the opposite tendency. That is, implicit in the text is the assumption that if the notion of basic tastes is not appropriate at one level, reference to labeled lines is inappropriate at any level. Whether this is the case or not depends, in large measure, on matters of definition and scope.
Young's insightful model of color vision predated the validating discovery of separate cone photopigments and their associated genes, producing three populations of receptors, each of which, though broadly tuned, have identifiable response characteristics. Such individuality is required to produce the pattern across which a perceptual interpretation is made. The need for these distinct lines of input is evident when one of them is missing in color deficient individuals. Unless the term “labeled lines” is restricted to very narrowly tuned outputs that run uninterrupted from the periphery to consciousness, then identifiable, arguably labeled, lines are required at certain levels within any across-fiber pattern model. To that extent it is probable that both vision and taste involve some short-distance labeled lines. It may be appropriate to consider the three visual photopigments the basis for three short-distance labeled lines, and the ganglion cells the basis for a subsequent layer of opponent-process labeled lines. Notwithstanding, in agreement with Erickson, it would clearly be inappropriate to call red, green, and blue lights basic visual stimuli. In a similar vein, although basic tastes generally imply labeled lines at the neural level, the reverse need not be true. Separately identifiable output lines, if they exist, need not imply a fixed number of basic tastes. In vision direct use is not made of the pattern of responses across the receptors. In principle, as Young concluded, color experience could be explained by the ratio of responding across three broadly tuned color receptors. Despite the economy of this idea, it is clear from both psychophysical data, and the discovery of color opponent output cells, that in practice no such simple pattern directly feeds color experience. Mullen and Kingdom (Reference Mullen and Kingdom2002), for example, have obtained evidence for distinct pathways from the retina through the lateral geniculate nucleus to the cortex for two cone-opponent systems, as well as an achromatic channel. Commonalities among mechanisms across sensory systems are, as Erickson maintains, most likely. In the end, and depending on the particular use made of key terms, most sensory systems are likely to display both across-fiber and labeled-line features. This may be especially true of taste responses, which represent but one component of the overall integrated flavor experience created by complex interactions with other senses (Stillman Reference Stillman2002).
Important differences between visual and chemosensory stimuli limit comparisons between processes at the receptor level in taste and in vision. In color vision the stimulus consists of a continuum of wavelengths. There is not an obvious continuum for chemosensory stimuli, and so the pattern of stimulation is likely to be fed by receptors that differ more than those in color vision, as revealed by studies that have identified both ionic and metabotropic mechanisms at work in taste receptors. Erickson may be correct in maintaining that perceptual basics do not exist, and that in any case the concept of basic tastes is not potentially falsifiable. However, this does not negate the possibility that at the receptor level some basic functional categories, for example, receptors responding to acids, might exist. Conceivably they could, by their mode of operation, contribute to the pervasive nature of basic taste words, inadequate though that vocabulary is for describing taste, let alone flavor, experience.
In recent years the labeled-line versus across-fiber pattern argument has, due in no small measure to the work of Erickson himself, become somewhat of a straw man. Few contemporary researchers or textbooks adopt a narrowly defined labeled-line approach with respect to taste perception. Notwithstanding, the force of Erickson's persuasive and illuminating arguments, namely, that research has become hide-bound by the notion of a set number of perceptual basics, has the potential to broaden the perspective researchers bring to bear on both research design and the interpretation of data.
