The Given

If there is no observer standing behind experience, the next question is whether appearance rests on something beneath it: some undesignated layer that is simply there, which concepts then describe. Philosophers have called this the Given. The idea has a long history in the West and an older, more rigorous one in Buddhist epistemology.

Begin with Locke. He divided qualities into primary ones, such as shape and motion, which belong to bodies themselves, and secondary ones, such as colour and warmth, which are only powers in bodies to produce sensations in us. The primary qualities were the bedrock. Wilfrid Sellars later attacked the whole structure from inside empiricism, arguing that the idea of a pure given in experience is a myth: what we call a sensation is already conceptual.

Buddhist philosophy had worked out its own version of the Given centuries earlier, and with greater rigour. Dignāga, in the Pramāṇasamuccaya, opens with a symmetry. There are exactly two means of valid knowledge, perception and inference, because there are exactly two kinds of object. The particular, what characterises itself alone, is grasped by perception. The universal, a general character shared by many things, is grasped by inference and concept. Perception itself is defined as free from conceptual construction.

Dharmakīrti sharpened the scheme. Perception is free from construction and non-erroneous. What makes the particular ultimately real is that it alone can do something: it is capable of causal efficacy. Universals do nothing. The fire that is merely thought of burns no one.

The result is a two-storey epistemology. On the lower floor are momentary, efficacious particulars, touched by perception before any concept arrives. On the upper floor are concepts, which never touch particulars at all. They construct general characters by exclusion, so that "cow" means what is not non-cow, and they are useful precisely because, strictly speaking, they misrepresent what they seem to present.

This position should not be caricatured. Dharmakīrti does not posit a bare substance waiting beneath qualities. He rejects the split between substance and property altogether. His claim is subtler: the real is what functions, and what functions is met by perception before designation. That is why it is the strongest form of the Given. It is a Given without a substrate, built by philosophers who had already accepted momentariness and dependent arising. If any account could keep an undesignated layer while honouring everything else this book has argued, it would be this one.

Candrakīrti's reply

In the first chapter of his commentary on the Mūlamadhyamakakārikā, Candrakīrti answers a view closely resembling Dignāga's. Three of his moves bear directly on the Given.

The first concerns the word own-characteristic itself. A characteristic is the characteristic of something: hardness characterises earth. Is the characteristic different from what it characterises? Then earth, apart from its hardness, has no characteristic, and nothing has been picked out. Is it the same? Then nothing characterises anything, and "own-characteristic" says nothing. This is the analysis Nāgārjuna applied to space. Characteristic and characterised are designated together, and neither is found first. A particular singled out as "this, characterised by itself alone" has already been singled out by contrast, which is the very operation the two-storey scheme reserved for the upper floor.

The second move concerns what perception is of. Candrakīrti does not abolish perception. He accepts it as a means of knowledge, together with inference, testimony and analogy, as the world uses them. But what the world perceives is the pot, not a momentary patch of colour. The object of perception is already a designated whole, met as a pot from the start, in the way a cup on a table is seen as a cup.

The third move comes from Nāgārjuna. Means of knowledge and objects of knowledge, he argues, are not established by themselves. Each is established only in relation to the other. Dignāga's founding symmetry, two means because two objects, is therefore not a foundation. It is co-definition, like long and short.

Tsongkhapa makes the conclusion explicit and makes it a defining mark of the Prāsaṅgika view. Nothing is established by its own characteristic, not even conventionally. Other Buddhist schools refuted the particular's ultimate status and kept it conventionally. Prāsaṅgika refuses it at every level.

What is kept, and what goes

The refusal is narrow. Valid cognition is kept. Perception is still distinguished from inference, a snake in a rope is still corrected by a lamp, and Tsongkhapa's criteria still sort sound conventions from mistaken ones. What goes is only the claim that perception's object is there from its own side, before designation, as the floor beneath it.

Dharmakīrti's deepest reason for that floor was efficacy. Substrateless World answers it. The key's power to open is found only together with the lock, the hand and the door. Causal efficacy is itself a relation. The mark that Dharmakīrti took for the stamp of independent reality turns out to be the clearest mark of dependence.

When the particular is not found, the two-storey building has no ground floor. Perception is not a lower layer, and concept is not a top coat. Vividness and designation are mingled through and through.