Proving It to Someone Else

Two different skills, easy to confuse

You can privately work through a piece of reasoning until you're completely convinced — that's svārthānumāna, "inference for yourself," Session 4's whole subject. Then you try to explain it to a friend, out loud, and watch it land flat. Nothing about your private reasoning changed. What's missing is a second, separate skill: parārthānumāna, "inference for others" — putting a valid inference into words that actually work on someone whose mind doesn't already contain your reasoning.

This is the last of the four chapters this pilot set out to cover, and it closes a loop that's been running since the very first color-and-shape debate: everything so far has been about getting the logic right inside your own head. This chapter is about what changes when another person is the one who has to be convinced.

What a proof statement actually is

Here's the chapter's central, almost deflating point: a spoken proof statement is not itself a third kind of valid cognition, alongside perception and inference. It's just words. What it does is trigger a valid inferential cognition in the listener's own mind — and it only works if the listener can independently verify, or already grants, the pervasion involved. Say a perfect proof statement to someone who doesn't accept "wherever there's smoke, there's fire," and nothing happens. The words didn't do the logical work; they pointed at logical work the listener has to complete themselves.

A leaner proof than you might expect

Older Indian logic (the Nyāya school) used a five-part proof: thesis, reason, example, application, conclusion. Dignāga and Dharmakīrti trimmed this down hard. Their claim: if the reason is stated correctly, a competent listener already has everything they need — the example is a teaching aid, useful for a beginner, but not a separate logical requirement. In its leanest form, a proof statement needs only two working parts:

When a reason actively backfires

Session 4 gave you two ways a reason can fail: it can be false (not established), or true-but-toothless (no pervasion). This chapter names a third, sharper failure: a reason that doesn't just fail to prove the thesis, but actually proves the opposite of it.

Classic case: "Sound is permanent, because it is a product." "Product" is a real, working reason — you met it as far back as Session 4 of Collected Topics — but it's the textbook proof of impermanence, not permanence. The reason doesn't fail to connect; it connects to exactly the wrong side.

1. Have you ever been privately certain of something and then genuinely failed to explain it to someone else? Looking back, was the problem your reasoning, or your proof statement?
2. Dharmakīrti treats the example as a teaching aid, not a logical requirement. Does that match your own experience of when examples actually help you understand an argument, versus when they're just decoration?
3. Try to recall a real argument you've heard (politics, a group chat, anywhere) that used a contradictory reason — one that actually supported the opposite conclusion from the one being argued.
4. If a proof statement only works on someone who already grants the pervasion, is logical "proof" ever really neutral — or does it always depend on what the listener walked in already accepting?