Criteriological Arguments in Perception

Abstract

I wish to consider here arguments by D. W. Hamlyn and Sydney Shoemaker that there is a necessary or conceptual connection between the qualities (e.g. colours) which objects actually have and the ways that people perceive them in normal conditions. I will maintain against these arguments that no thesis concerning the relation of the secondary properties of independent physical objects to the ways they appear or are perceived could be necessarily true. The central claim of criteriological arguments in perception, put negatively, is that it is nonsense or unintelligible to deny that objects have certain qualities which they appear to have in, normal conditions. This claim rests mainly on the Wittgensteinian notions of criteria and of understanding terms in a language, according to which understanding a term involves knowing the criteria for its proper application. Understanding the concept 'red', for example, implies being able to apply it correctly. Furthermore, that a thing is red implies that it meets the criteria for the proper application of the concept or term 'red'. But proper application in this case according to Hamlyn is application to things which normally appear red-appearing red in normal conditions constitutes a criterion for the proper ascription of the property. Using the criteriological jargon of Scriven and Aune, Hamlyn would hold (against Alan White) that such normal appearance constitutes a 'normic' rather than an 'analytic' criterion, since the concept of red is claimed to be semantically more basic than that of appearing redly, but the necessary connection is still held to exist. Thus it is implied that a person is simply not applying 'red' correctly if he refuses to apply it to objects which appear red in normal conditions-to refuse to follow the rules of language is to speak nonsense. There is also a subsidiary argument by Hamlyn that if all those objects which we normally count as red were not, then the concept 'red' would have no application, and we would not understand what redness is. Thus it is claimed again that if we do understand the concept, then there must be more than a mere contingent connection between how we see things and how they are.' If this argument is valid, then it seems that the sceptic with respect to perceptual knowledge can only maintain his position by denying that we know what redness is, for example, that we have a concept of the colour at all, a highly implausible position to take. But if the argument is to have any plausibility itself, it must be shown how 'red' differs from other concepts the understanding of which is ontologically neutral. We have numerous concepts (e.g. valid criteriological arguments in perception, unicorns) without there being existent instances. What Hamlyn is claiming is that if there were no red objects, then we could not have built up the concent: not only would there be no actual instances, but no conceivable