Since Reid's day, two challenges to the identity theory of secondary qualities object on purely philosophical grounds. The first, promoted by Howard Robinson, criticises the causal connection between physical and mental, external property and subjective phenomenon. Second, Bertrand Russell complains that differing knowledge conditions for secondary qualities and physical properties prevent the identity theory from meriting serious consideration. Consider them in that order.

Robinson's A Priori Sufficiency

As in the preceding case, the next two objections confuse sensations with perceivable properties. Part II offered an account of secondary qualities whereupon they count as physical without becoming causally superfluous. How? Secondary qualities are identifiable with certain physical, scientific properties. Howard Robinson, who featured in Part I, says that such a proposal is impossible because no candidate for identification is a priori sufficient to explain ‘phenomenal’ secondary qualities. Perhaps his criticism communicates best in light of some observations about scientific explanations in general.

Consider a paradigm case of scientific identification. The phases of matter – solid, liquid and gas – may be specified as microphysical properties relating to chemical bonding. Firm bonds create the rigid structures of solids, whereas weak or non-existent bonds cause substances to behave as liquids or gases. One reason that such an explanation works is that its truth entails the phases of matter. That is, the explanans necessitates the explanandum. It is impossible for the molecules of a substance to bind rigidly without the substance becoming a solid, and likewise for other bonding strengths and phases. As Robinson puts it, the explanans of chemical bonding strength is ‘a priori sufficient’ for the explanantia of the material phases. He explains,

A solid has certain dispositional properties: it does not flow, and it has a certain elasticity, plasticity, tensile strength, shear strength and so on.