Upamana (comparison) is a King Charles’s head with me. Like Mr. Dick with the latter (David Copperfield) I am fascinated by it to the extent that I believe that even those who accept it as a pramana aren’t aware of its decisive role in the problem of universals. Nobody understands upamana ! That way lies madness.
When re-reading Matter and Memory I find that Bergson with his sense of the foundational nature of comparison/resemblance takes a similar view:
It would seem, then, that we start neither from the perception of the individual nor from the conception of the genus, but from an intermediate knowledge, from a confused sense of the striking quality or of resemblance: this sense, (pg 206) equally remote from generality fully conceived and from individuality clearly perceived, begets them both by a process of dissociation. Reflective analysis clarifies it into the general idea; discriminative memory solidifies it into a perception of the individual.
Sankara with his theory of vedic words proposes a notion of universals along the same lines as Plato’s. Pramana for Sankara then might be restricted to a way of going from the innately known genus to the unknown particular - from cow to gavaya as per the standard example. Bergson would deny ideas and accept comparison (upamana) which he would term the “striking quality”. Upamana on its own could generate a world of universals and particulars.
Bergson’s statement of the classic problem is succinct:
individual objects. The whiteness of a lily is not the whiteness of a snow-field; they remain, even as isolated from the snow and the lily, snow-white or lily-white. They only forego their individuality if we consider their likeness in order to give them a common name; then, applying this name to an unlimited number of similar objects, we throw
back upon the quality, by a sort of ricochet, the generality which the word went out to seek in its application to things. But, reasoning in this way, do we not return to the point of view of extension, which we just now abandoned? We are then, in truth, revolving in a circle, nominalism leading us to conceptualism, and conceptualism bringing us back to nominalism.
Generalization can only be effected by extracting common qualities; but, that qualities should appear common, they must have already been subjected to a process of generalization.
Moreover what Bergson says about this power of discerning resemblance namely that it is of the nature of a primal force that does not have to be set into motion by a psychological effort, has the true flavour of pramana as irreducible.
But we relegate nothing to the unconscious, for the very simple reason that it is not, in our opinion, an effort of a psychological nature which here disengages similarity; this similarity acts objectively like a force, and provokes reactions that are identical in virtue of the purely physical law which requires that the same general effects should follow the same profound causes.