Chapter 10 Dependent origination¶

This chapter is in the first person and stands outside the proof.

Following the earlier example, begin with a declaration. This chapter records my experience, not an argument. Whether this essay's theorem holds has nothing to do with how I came to understand it. Remove this entire chapter and not a word of the first nine needs to change. Essay 13's final chapter recorded part of my encounter with the early teaching and left one sentence unsaid at the end: Madhyamaka shook me once more and corrected me once more, a story left for the Madhyamaka essays. This chapter supplies that sentence.

The title's two Chinese characters, 缘起, make a pun. They name dependent origination, the matter this essay has addressed throughout. They also mean "how something came about," the origins of a book or a piece of writing. This chapter takes the latter meaning.

Like most readers of Chinese, I grew up in the territory of the northern Buddhist tradition. At first, Buddhism was hardly a choice for me. It was the air: the look of temples, the names of bodhisattvas, a few familiar verses. They were there before I knew what I was taking in.

I had always felt a little unease about Mahāyāna that I could not quite name. It was directed at no particular school. It had the shape of a question: a religion that takes saving people from suffering as its duty ultimately judges whether someone has awakened by how that person speaks, chooses words, and presents a view thoroughly and rigorously. Judging awakening or its absence thus becomes a language examination.

Another side of the same question was those two verses. Shenxiu says to polish diligently at all times; Huineng says that originally there is not a single thing. The latter was judged higher, on the grounds that it was more thoroughgoing. Yet polishing is something to do, and "not a single thing" is something to say. Why should the statement rank above the deed? The verdict did not explain. The textual layers of that verse contest are discussed in Part I, Essay 10, Chapter 8. Huineng's verse in the Dunhuang version says "Buddha-nature is always pure," not the received version's "Originally there is not a single thing." The contrast between the two verses is less sharp there than it is today. I learned this philological point only later. What troubled me then was the received version's sentence.

The third difficulty was hardest to answer. If originally there is not a single thing, then there is nothing. Yet whenever someone pursued the literal sense and asked, "Then does anything exist or not?", the skilled debater immediately turned back: it does not mean there is nothing. What, then, does it mean? A statement is taken one way and it means this; taken the opposite way, it still means this. The speaker decides both readings. My judgement at the time was that this was rhetoric, not insight.

With these three questions, I came to Madhyamaka.

When I first read the Mūlamadhyamakakārikā, I did not understand what it meant to say. I had only a rough judgement: something here is wrong. That judgement later proved right. I had simply been wrong about where the error lay.

So I tried to refute it with a clear target: find one self-contradiction in emptiness. If everything is empty, is this sentence empty too? If there is no intrinsic nature, is this "no" itself also a kind of nature? These challenges did not originate with me. I simply did not know that then and reached them through my own questions.

For the next three months, discussion, questions and answers, attacks and defences, disputes followed round after round. I did not succeed. The precise character of that failure must be clear: I found no self-contradiction in the Mūlamadhyamakakārikā. Every objection I thought fatal turned out, once raised, to have a distinction of syntactic levels already waiting for it. Rejecting does not mean asserting the negation; negating does not mean positing the contrary thesis; "emptiness too is empty" is no flaw but a clause already written. These points later found their respective places in this essay, Lemma 1 in Chapter 3 and the level clauses in Chapter 5. My objections failed. The other side was not evading them.

After losing the debate, my unease did not fade. It became more specific. A system had withstood three months of attacks without a visible flaw. I should reasonably have been convinced. I was not, and at the time I could not say where that unease pointed.

So I tried another method: give each argument in the Mūlamadhyamakakārikā a modern proof, write its definitions as definitions and its derivations as derivations, and list every premise used at every step to see whether it withstood questioning. The result is the essay now in the reader's hands.

The conclusion was that it withstood questioning, and more firmly than I had expected. The definitions came from criteria the opponents themselves used; the derivations used only logic. Buddhist material performed only two jobs throughout the essay: sources and historical instances. Once the work was done, I had no reservations about the theorem's logical rigour.

It was also only after finishing that work that I located my unease. It was not in the Mūlamadhyamakakārikā's conclusion. The conclusion was right. Something else troubled me.

For a time afterward, I committed the very misuse to which the Mūlamadhyamakakārikā itself most readily leads.

With such a tool, it is hard not to use it. In any debate, whatever the other side posited, I could dissolve it. Was the thing posited established in dependence on another or self-existent? If dependent, it was not what they wanted; if self-existent, it cancelled identity and difference and could not be identified. This way of questioning worked in any debate. What arose in me when the other side could not answer, I knew very well.

After using it many times, I looked back at what I was doing. I saw that I did not care what remained after each dissolution. I cared about dissolving.

This is functional attachment. It clings to no particular view. Quite the reverse: what it clings to is the method of cancelling views. The thing to which it clings is called "no thesis." It is harder to detect than attachment to a view because it appears empty. It talks about emptiness, speaks emptiness, and dissolves the question. It does not show on the outside, but it is there all the same.

This has its formal place in the book. Chapter 7 in the latter part of Volume 2's Madhyamaka essay is devoted to functional attachment and ultimate truth. Chapters 1 and 2 of the same essay, on the truths of suffering and its origin, give its mechanism. Madhyamaka shifts the field of work from prereflective craving to views at the reflective level. I first saw this shift in myself. My craving had not diminished at all. It had only changed its object, from "I want to be right" to "I want to be able to take you apart."

I dare to make this judgement because I have gone through the derivation afresh from outside Buddhism. Chapter 8 tests the theorem in a setting without Buddhist premises. What it tests is not only the theorem's scope of application but its character. A tool that remains effective where there are no Buddhist premises at all shows precisely that it belongs to logic, not liberation. It is a philosophy, a constraint whose function is to mark what statements cannot be made and what combinations cannot be posited. It does not supply what accounts within Buddhism promise for it. It is no path that leads out of suffering if followed to the end. This is no disparagement. The law of noncontradiction is likewise extremely useful, and no one expects to leave saṃsāra through it.

With this understanding, rereading the Diamond Sūtra and the Heart Sūtra made formerly obscure passages clear. "What is called X is not X; it is named X" is no mystical wording but a distinction of levels: reject the reading of X as really existent, preserve the force of X in language and names. The Heart Sūtra's string of "no" is likewise a list of rejections, not a declaration of annihilation. That skilled debater's claims on both sides had been half right. It really did not mean there was nothing. He simply could not explain the level at which the other half was spoken, so it sounded like rhetoric. The Mūlamadhyamakakārikā wrote that level out.

Precisely because it wrote it out, I could see what that level can and cannot do.

The result of this essay's whole argument is that my confidence in the Mūlamadhyamakakārikā's logical core is greater than it was then, not less. It is a true theorem, not a bound collection of reductios. Yet I put my name to the closing sentence of Appendix 2 in Volume 2: the more perfect a philosophical form, the more respectably, and the further, it stands apart from the original question, "How does suffering cease?" Madhyamaka is far from a theory of liberation, so far that a whole volume is needed to explain the distance.

For a certain kind of person, it may be no help and may well be an obstacle. I mean people like me: full of intellectual craving, skilled in debate, delighted by thinking something through. Essay 13's final chapter records that I still feel intense craving for logic and the intellect themselves. Give such a person the Mūlamadhyamakakārikā, and the likeliest result is no cessation but an upgrade. Attachment rises from content to method, from views to the technique of dissolving views. After the upgrade it is harder to notice, because it looks no different from letting go.

So this chapter's verdict is one sentence. The Mūlamadhyamakakārikā remains an extremely effective tool, and in this essay I have given it the strongest defence I can. But it should remain in the place of a tool, the place of a constraint.

Chapter 9 ended with the raft analogy: crossing a river requires a raft, but on reaching shore one does not carry it onward. When writing that passage, I was discussing the theorem's limitation of itself. Only in writing this chapter do I admit that the sentence was addressed first to me. For a long time I carried the raft and took pride in doing so.

The first nine chapters are my proof of this theorem. This chapter is the theorem's proof, carried out on me.

References¶

This chapter is a personal record in the first person and cites no new material. For the theorem, lemmas, clauses, and verdicts concerned, see the first nine chapters of this essay and Appendix A. For the textual layers of the two verses, see Part I, Essay 10, Chapter 8. For functional attachment and the shift in the field of work, see Chapters 1, 2, and 7 in the latter part of Volume 2's Madhyamaka essay. The statement beginning "the more perfect a philosophical form" and ending "the more respectably, and the further, it stands apart" comes from that volume's Appendix 2, final paragraph.