Must a Book Begin Auspiciously?

The third reply is the opening verse of the Mūlamadhyamakakārikā applied to a question about book design. Nāgārjuna's first chapter argues that nothing arises from itself, from another, from both, or without a cause; Āryadeva runs the first three of those against the auspicious, and adds the regress — if the arising of the auspicious is itself a thing, its arising needs an arising, and so on without end. That the tightest formulation in Madhyamaka is deployed here, on the least weighty topic in the book, is not a misuse of it. It is a demonstration that the argument has no minimum size.

The second reply is the one to watch, though, because it does something the third does not. It is an argument from disagreement: the same thing is called auspicious by one person, inauspicious by another, and neither by a third; therefore it is 不定, unfixed; therefore there is no auspicious. Notice that this does not conclude that nobody may call anything auspicious. It concludes that auspiciousness is not a property sitting in the thing waiting to be detected — which is the treatise's whole method in miniature, and the reason it can take an argument about etiquette seriously enough to answer it.

One textual point. CBETA punctuates the first reply as ……是吉。是不吉，此是邪見氣, which yields the reading we have translated: the text is auspicious because it cuts off wrong views, and to call it inauspicious is itself a symptom of wrong view. A different and arguably smoother punctuation would run the clauses together — 是吉是不吉，此是邪見氣 — giving "the whole business of 'this is auspicious, this is inauspicious' is a symptom of wrong view," which reads forward more naturally into the denial that follows. We have followed the received punctuation and flag the alternative rather than quietly adopting it.

Two distinct refutations are packed into four lines, and the second is the more interesting.

The first is a dismissal: you have appealed to a nature of its own, and natures of their own were disposed of a moment ago. The second says I shall use your own words against you, and it is a genuinely elegant trap. The Outsider's school holds that salt is produced from causes and conditions. But whatever is produced from causes and conditions does not have its saltiness from itself — so the Outsider's own physics contradicts his own simile. And then, granting the simile anyway, it still fails on its own terms: salt makes food taste salty, but the food does not become salt. Saltiness stays where it lives. Whatever an auspicious opening does to a book, it does not turn the book into an auspicious opening.

The phrase to hold on to is 鹽相鹽中住 — the mark of salt abides in salt. This is the same structure the treatise will use for the next twelve chapters: 牛相牛中住, the mark of an ox abides in the ox, in Part 4; 觸是指業非指相, touching is the finger's act and not its mark, in Part 9. A mark does not migrate. Whatever combines with what, each thing keeps its own.

The self-lighting lamp is the most travelled image in Madhyamaka, and this site now carries all three of its major appearances. Nāgārjuna takes it up in the Mūlamadhyamakakārikā's longest chapter, where it is offered to rescue an arising that arises itself; the Twelve Gate Treatise takes it up in Gate Four, for the same purpose. Here it is doing duty for an auspicious word at the head of a book. The image keeps being reached for because it seems to be a plain counter-example to the rule that nothing operates on itself, and it keeps being taken apart the same way.

The reasoning is worth following slowly, because it is easy to think it is a quibble and it is not. Ask what lighting is: it is the removal of darkness. Now ask where the darkness is that the lamp removes from itself. There is none — brightness and darkness cannot coexist, so at no moment is there a dark lamp for the light to arrive at. And in the corner of the room the lamp has lit, there is likewise no darkness now. So the lamp lights neither itself nor anything else, not because lamps do not work, but because "lighting" was defined as a relation between two things, and at no point are both of them present.

Then the Outsider makes the move that always follows: the two happen at once, at the instant of ignition. And he gets the answer that always follows — at that instant the lamp is half arisen and half not, and half-a-thing cannot do what a whole thing could not do either. The same "half arisen, half not yet arisen" appears in Part 12 against arising itself, where it becomes the central argument rather than a rebuttal. It is worth recognizing the shape now.

Notice what the Insider does not do. He does not deny that spells work or that stars portend anything. He grants both without comment — the power of a spell or a star can indeed reach far — and attacks only the inference. This is characteristic. The treatise is not in the business of correcting its opponents' cosmology; it is in the business of showing that their conclusions do not follow from their premises. Conceding the premise costs nothing and makes the refutation stronger.

The argument itself is a reductio by geography, and it is very good: if distance is genuinely no obstacle to a lamp, then one lamp anywhere should light everywhere, which no one believes. A power that works at a distance must be shown to be that kind of power, and the Outsider has simply assumed it.

A word on 振旦. It transliterates Cīnasthāna — China. An Indian treatise contrasting "a lamp lit in India" with "the darkness in China" is possible, but it is far more likely that Kumārajīva has localized the example for the audience actually in front of him at Chang'an, replacing whatever distant place the Sanskrit named. Readers should know the sentence is probably reaching for them personally: the darkness that the Indian lamp fails to reach is the one they are sitting in. We keep the reading as it stands and flag the likelihood rather than reconstructing a hypothetical original.

The pun does not survive into English and the argument is unintelligible without it, so: the Sanskrit for elephant is hastin, "the one that has a hand," from hasta, hand — the trunk. Kumārajīva renders this literally as 象有手故名有手, and the Outsider's point depends on it entirely. An elephant has eyes, ears, a head and four feet, but only one of its parts gets to name the whole animal. So a part really can dominate a whole out of all proportion to its size — which is exactly what he needs the auspicious opening to do.

It is a better argument than it looks, and the Insider does not answer it on its own ground at all. He simply asks what relation the trunk stands in to the elephant, and runs the standard dilemma. Different? Then the head and feet are different too, and there is no elephant over and above a heap of separate parts. Not different? Then the head is not different from the elephant and the foot is not different from the elephant, so the head is the foot. Either way there is no elephant to be named, and a name that names nothing cannot confer anything on anything.

The two closing references — 如破異中說 and 如破一中說 — are the treatise pointing forward at its own third and fourth chapters, which are titled precisely Refuting Oneness and Refuting Difference. They are Part 7 and Part 8 here. Vasu is telling the reader, in the first chapter, that the argument he has just used in passing will be given properly later — and that whole-and-part is not a side issue but one of the two hinges the book turns on.

The cloth is the best thing in this part of the chapter, and it answers the objection completely without any metaphysics at all. The Outsider's point was about motive: you stop evil because you want the good fruit, so the good is prior. The Insider's reply is about sequence: whatever your motive, a dirty cloth will not take dye. Wanting the color does not let you skip the washing.

What makes the image durable is that it does not disparage either step. The washing is not the point of laundry and the dye is not applied to dirt. Both are necessary and they are necessary in an order that is not negotiable — and the order runs the opposite way from the order of desire. This is also, quietly, the structure of the entire first chapter: stop evil, practice good, and then — the part nobody wants — put down the merit too.

The reply about giving contains three separate arguments and each one is worth having on its own.

First: if failing to give were an evil, then every person who does not give at this moment — which is nearly everyone, nearly always — would thereby be committing an offense. Omissions cannot be converted wholesale into acts without absurdity.

Second, and sharpest: consider someone in whom greed has been wholly destroyed, an arhat with the outflows exhausted. He gives. What evil is being stopped? There is none left to stop. Yet the giving is unmistakably still giving. So giving cannot be defined as the cessation of stinginess, because the giving survives when the stinginess is already gone. The definition fails at exactly the point where the practice is most perfect.

Third, the mirror image: there are people who give generously and remain stingy. So the two do not even reliably coincide in ordinary cases.

Then the formal answer, which is very compact: 止相息、行相作 — the mark of stopping is cessation, the mark of practicing is doing. One is the absence of an act and the other is an act. You cannot fold either into the other, however often they occur together. It is worth noticing that the Outsider accepts this and immediately retreats to a subtler position — that stopping evil is a species of good dharma, so that the wider term should suffice. That retreat is where Part 3 opens, and the answer to it turns out to require a case that neither party has mentioned yet: the person who has taken a precept and is at that moment thinking about nothing at all.