How do you know? The Buddha told a brahmin that faith, tradition, and careful reasoning can each turn out false, and told his own monks to examine him. A thousand years later Buddhist logicians allowed only two sources of knowledge, where their rivals allowed four. Why Hume’s two kinds of reasoning come so close, what Kant would object, and what counts as evidence on a path that asks to be tested. 你怎麼知道？佛陀告訴一位婆羅門：信仰、傳統與仔細的推理，每一樣都可能到頭來是假的；又叫他自己的比丘審察他。一千年後，佛教的邏輯學家只承認兩種知識的來源；而他們的對手承認四種。休謨的「兩類推理」為什麼這麼接近？康德會提出什麼反對？在一條要求被檢驗的道路上，什麼才算證據？
D13 Dharma and the West · 佛法與西方哲學Knowledge and Logic · 知識與邏輯Advanced · 高級about 30 min
量
pamāṇa Pāli pramāṇa Sanskrita means of knowledge
The verdict · 判定
Closest: Hume, who also allowed two and only two kinds of knowing, direct experience and reasoning, and gave tradition no independent authority. William James supplies the test of truth that the Buddhist logicians used: a cognition is knowledge if it does not let you down when you act on it. The likeness to Hume ends at the range of experience. Hume’s experience is that of the five senses and ordinary reflection. The Buddhists counted the trained observation of a meditator as perception, and they pursued the theory because the path depends on knowing what is so.
中文翻譯Every paragraph has a Chinese translation. Tap 中譯▾ under a paragraph to open that one, or use Show Chinese · 顯示中譯 at the top to open them all.每一段都有中譯：點段落下方的「中譯」只開那一段，或按上方的「顯示中譯」一次全部打開。
1 · What the Dharma says · 佛法怎麼說
Examine the teacher 審察老師
The early discourses are unusually explicit about grounds for belief. A young brahmin, Kāpaṭika, told the Buddha that the brahmins hold their scriptures to be the only truth. The Buddha asked whether any of them, back through seven generations of teachers, had claimed to know this for himself, and compared the tradition to a file of blind men, each holding the one in front. He then listed five grounds on which people accept things: “Faith, endorsement, oral transmission, reasoned train of thought, and acceptance of a view after deliberation.” Each can go either way. “Even though you have full faith in something, it may be vacuous, hollow, and false. And even if you don’t have full faith in something, it may be true and real.” A person who cares about truth may say “such is my faith,” and may not go on to say “this is the only truth, anything else is futile.”
He applied the standard to himself. To the Kālāmas, who were confused by rival teachers, he said not to go by oral transmission, lineage, canonical authority, logic, or the thought “the ascetic is our respected teacher,” and to rely on what they could know for themselves about what leads to harm and what to welfare. To his own monks he said that an inquirer who cannot read minds “should scrutinize the Realized One to understand whether he is a fully awakened Buddha or not,” and he described how: watch his conduct over a long time, and see whether anything in it comes from greed, hatred, or delusion. Faith has a place in the path as the confidence that makes a person willing to try. It is expected to be replaced by knowledge.
About a thousand years later Buddhist philosophers turned these attitudes into a theory. Dignāga, in the fifth or sixth century, argued that there are exactly two means of knowledge, perception and inference, because there are exactly two kinds of thing to be known. A short manual of his school, translated into Chinese by Xuanzang, puts it in a line: for one’s own understanding, “there are only two means of knowledge, perception and inference.” Perception is “free of conceptual construction,” a direct awareness of a particular before any name or class is applied to it. Inference knows its object “by way of signs,” as one knows fire on a hill from smoke. The rival Nyāya school counted testimony as a third source and comparison as a fourth. The Buddhists held that what we learn from a trustworthy speaker is a case of inference: we reason from his reliability. Dharmakīrti, a century after Dignāga, added the test. A cognition is a means of knowledge if it does not deceive, that is, if the thing it shows is there to do its work when one acts.
The British empiricists held that the mind begins empty and is furnished by experience, and Hume gave the view its sharpest form. “All the objects of human reason or enquiry may naturally be divided into two kinds, to wit, Relations of Ideas, and Matters of Fact.” The first are known by reasoning alone and the second by experience. A book that contains neither kind of reasoning should be committed to the flames, “for it can contain nothing but sophistry and illusion.” The two-part structure, the refusal of any third source, and the consequences for religious authority all match Dignāga’s position. Hume wrote a famous chapter arguing that testimony about miracles can never outweigh experience, which is the Buddhist treatment of testimony applied to a hard case. The difference is in where each ends. Hume found that experience cannot justify even our belief in causes, and concluded with a mitigated skepticism. Dignāga and Dharmakīrti held that perception and inference, properly used, do yield knowledge, enough to walk a path by.
Kant agreed with the empiricists that knowledge begins with experience and denied that experience is something simply received. Two faculties are needed, one that is given objects and one that thinks them. “Thoughts without content are void; intuitions without conceptions, blind.” Raw sensation with no concept applied is not yet knowledge of anything. This is aimed at Locke and Hume, and it lands on Dignāga’s definition of perception as free of conceptual construction. The Buddhist logicians saw the difficulty themselves. They granted that bare perception does not by itself tell us what we have perceived, and that a conceptual judgment has to follow before we can act. They kept the distinction because it did work for them. It marks the point at which the mind adds names, classes, and “mine” to what is there, and the path is concerned with exactly that addition. Kant thought the contribution of concepts could not be removed. The Buddhists thought it could be watched happening.
The pragmatists proposed that an idea be judged by what follows when we act on it. For James a true idea is one we can verify, which means that it leads us successfully through experience to what it promised. This is Dharmakīrti’s criterion almost word for word. A cognition is a means of knowledge if it is non-deceptive, and it is non-deceptive if the object it presents performs its function: the water one saw can be drunk, the fire one inferred can cook. A mirage fails the test at the moment one bends to drink. Both hold that knowledge is to be assessed by its fruit in practice and without first settling what reality is in itself. They used the criterion for different ends. James used it to defend the right to hold beliefs that make life go better where evidence is lacking. Dharmakīrti used it to exclude them. For him a belief that has not yet met the test is not yet knowledge, however much it helps.
Who is closest, and where the likeness ends 誰最接近，相似到哪裡為止
On the conclusion Hume is closest: two sources, experience first, no independent weight for scripture. On the test of truth James is closest. Kant is the critic whose objection the Buddhist tradition also raised against itself. Taken together the three show that the Buddhist theory of knowledge sits squarely within what the West calls empiricism, and that it met the standard objections to empiricism in its own history.
The likeness ends at two points. The first is what counts as experience. The Buddhist logicians listed four kinds of perception: by the five senses, by the mind, the awareness each mental state has of itself, and the perception of a yogi, which is the direct seeing that comes from long training in attention. The claim is that a trained observer perceives what an untrained one must infer or take on trust, as a physician sees in an X-ray what a layman does not. Hume had no place for this, and a Western empiricist will ask how such perception is to be checked by those who lack it. The second is the purpose. The British empiricists wanted to map the limits of the human understanding. The Buddhists wanted to know which cognitions can be relied on by someone trying to end suffering, since acting on error keeps a person bound. Their epistemology is part of the path. It has the same job as a physician’s knowledge of how to tell a true symptom from a false one.
相似之處在兩個地方為止。第一，是什麼才算經驗。佛教的邏輯學家列出了四種現量：五根的現量、意的現量、每一個心理狀態對它自己的覺知（自證），以及瑜伽行者的現量；後者是長期訓練注意力而來的直接看見。這裡的主張是：一位受過訓練的觀察者，知覺得到未受訓練的人只能推論、或者只能憑信任接受的東西；就像醫師在一張 X 光片裡，看得到外行人看不到的東西。休謨沒有為這個留位置；而一位西方的經驗論者會問：那些沒有這種現量的人，要怎麼查核它？第二，是目的。英國的經驗論者想要畫出人類理解力的界限。佛教徒想要知道的是：對一個想要止息苦的人來說，哪些認知是靠得住的？因為照著錯誤去行動，就使人繼續被綁著。他們的知識論是道的一部分。它的工作，就像一位醫師「怎麼分辨真症狀與假症狀」的那種知識。
Perception can be trained, and the theory can be questioned 現量可以訓練；而這套理論也可以被質疑
The first further step has been named: the claim that perception is a skill. The discourses say that the Buddha’s teaching invites one to come and see, and that it is to be known by the wise each for himself. The logicians gave this a place in the theory. A meditator who has stabilized attention and examined experience closely is said to perceive impermanence and the absence of self in the way that others perceive color. This turns the central doctrines from articles of belief into things that are in principle observable, and it places the burden where the Buddha placed it for the Kālāmas, on one’s own knowing. It also creates the tradition’s hardest problem of evidence. A report from trained observation carries weight only with those who have the training or who have independent grounds for trusting the observer. The discourse on examining the teacher is the tradition’s answer: judge the observer by what can be seen of his conduct over time.
The second step is that Buddhism did not leave its own theory of knowledge unexamined. Nāgārjuna, three centuries before Dignāga, had asked what establishes the means of knowledge themselves. If they are established by further means, there is a regress. If they establish themselves, any claim might do the same. If they are established by the objects they reveal, the argument goes in a circle. He concluded that means and objects of knowledge depend on each other and that neither is a foundation. Western philosophy calls the same three-way trap by the name of an ancient skeptic and has used it against foundationalism ever since. So the tradition contains both projects. One builds a careful account of what counts as knowing, and the other shows that the account stands on nothing outside itself. Later Buddhist thinkers mostly held the two together, using the theory of knowledge as conventional truth: reliable in practice, like a map, and without a ground beneath it.
Neither blind faith nor laboratory science 既不是盲信，也不是實驗室的科學
Western listeners come with one of two expectations. Those who think of Buddhism as a religion expect to be asked to believe. Those who have read popular books expect to be told that Buddhism is a science and the Buddha an empiricist ahead of his time. The texts support neither without qualification. Begin with the passage in which the Buddha tells his own students to examine him, which surprises both groups. Then describe the position plainly. Faith is where a person starts, as one trusts a doctor enough to try the treatment. The tradition’s own philosophers held that knowledge comes from direct experience and from reasoning, and that scripture must answer to both. Add the honest limit. Some of what the tradition claims, a trained mind is said to see directly, and until one has that training one is taking a specialist’s word, with reasons for trusting him.
It asks for trust at the start. Say that faith is provisional and is meant to be replaced by knowing.
它在開頭是要求信任的。要說：信是暫時的；用意是由「知」來取代。
“Buddhism is a science.” 「佛教是一門科學。」
It shares the attitude of testing. It has no laboratories, and trained perception is not public in the way a measurement is.
它有「要檢驗」的那種態度。它沒有實驗室；而受過訓練的現量，並不像一項量測那樣是公開的。
“The Buddha said: believe nothing, not even me.” 「佛陀說：什麼都不要信，連我也不要信。」
A paraphrase that drops the second half. He said what to rely on: what you know for yourself to lead to harm or to welfare.
這是丟掉了後半句的轉述。他說了該依靠什麼：你親自知道會導向傷害、或者導向利益的東西。
“Meditation proves it.” 「禪修證明了它。」
To the meditator, perhaps. To others it is a report. Say what the report is and why you trust it.
對禪修者本人，也許是。對別人，那是一份報告。要說出那份報告是什麼，以及你為什麼信任它。
Three languages · 三語對照
The vocabulary, in Chinese, Pāli, Sanskrit, and English 術語：中文、巴利文、梵文、英文
中文
Pāli
Sanskrit
English
量
pamāṇa
pramāṇa
a means of knowledge; a reliable cognition
現量
paccakkha
pratyakṣa
perception; direct awareness
比量
anumāna
anumāna
inference
聖言量
sadda
śabda
testimony; the word of a reliable person
離分別
—
kalpanāpoḍha
free of conceptual construction
自相
salakkhaṇa
svalakṣaṇa
the particular; the unique character of a thing
共相
sāmaññalakkhaṇa
sāmānyalakṣaṇa
the general character; what is known through concepts
不欺誑
avisaṃvāda
avisaṃvādin
non-deceptive: the mark of a means of knowledge
瑜伽現量
—
yogi-pratyakṣa
the perception of a trained meditator
自證
—
svasaṃvedana
the self-awareness of a mental state
信
saddhā
śraddhā
faith, confidence
來見
ehipassika
ehipaśyika
inviting one to come and see
Reading check · 理解測驗
Did it land? 讀懂了嗎
1. What did the Buddha say about faith, oral transmission, and reasoned thought?
2. What did he tell his monks to do about himself?
3. Which two means of knowledge did Dignāga accept?
4. What is Kant’s objection to perception without concepts?
5. What is Dharmakīrti’s test for a means of knowledge?
Write · 你怎麼講？
List three things you believe: one from your own experience, one by reasoning, and one on someone’s word. For the third, write down why you trust the source. About 150 words.
