There's a particular frustration almost everyone has experienced. You're in a meeting, in a relationship, in a family disagreement, and you know that what you're trying to say is true — you can feel the shape of it. You produce reasons. You cite examples. You gesture at the obvious. And the other person looks at you politely, or irritably, and remains exactly where they were. You have just tried to make an argument, but the argument did not move them. The Buddhist logicians of the classical period — particularly in the tradition known as 因明 (*hetuvidyā*) — spent centuries thinking very precisely about why arguments fail to produce understanding, and what shape a piece of reasoning must have if it is to actually open the unclarity in someone else's mind. That shape has a name: 能立 (sādhana), often translated as "the means of establishing," or "the proof."
This is not a concept about debate tricks. It is not about being clever or winning. It is a quiet, almost engineering-style discipline: how to construct a small argumentative object whose every joint can be examined by both sides, and which, if it is built correctly, must produce understanding in the other. Once you have internalized it, you cannot unsee how rare it actually is in human discourse.
What this concept is actually saying
The core definition
The canonical formulation comes from the 《因明入正理論》 (*Nyāyapraveśa*, attributed to Śaṅkarasvāmin and translated into Chinese by Xuanzang):
「宗等多言，開示諸有問者未了義故，名為能立。」 *Because the synthesis of thesis, reason, and example words "opens up what is not yet understood by the questioner," it is called 能立.*
Three things to notice already. First, 能立 is a *synthesis* — 宗等多言, "many words joined together." It is not a single statement; it is a small composed structure. Second, its purpose is "開示諸有問者未了義" — to *open up, disclose, illuminate* what has not yet been grasped by the other. Third, and this is the quiet radicalism, the goal is 唯悟他 — understanding in another. It is not the *proposition* being established that matters primarily; it is that the other party actually *gets it*. If they assent without understanding, you have not done 能立. You have done something else.
This puts 能立 in sharp contrast to two other cognitive acts in the same system:
- 現量 (pratyakṣa) — direct perception, valid for oneself alone. - 比量 (anumāna) — inference for oneself, where one validly concludes something in one's own mind. - 能立 — inference *for others*, where one explicitly lays out the reasoning so that someone else can perform the same inference and reach the same conclusion.
And opposite to 能立 stands 能破 (dūṣaṇa) — refutation, which points out flaws in someone else's position. Both belong to 唯悟他. But 能立 is constructive; 能破 is critical. In contemporary terms: 能立 is *building the case for P*; 能破 is *showing that Q does not hold*. They are symmetric, but distinct disciplines.
The structure: 宗, 因, 喻
能立 is composed of three members. Let's take the classical example used in the 《因明入正理論》 itself — establishing that sound is impermanent.
宗 (pratijñā) — the thesis. This is what the proponent wants to establish. Its classical formulation is dense: *"極成有法，極成能別差別為性，隨自樂為所成立性。"* Unpacked: the subject of the thesis (有法, the "thing that has the property") must be 極成 — *acknowledged by both parties* as existing. The predicate (能別, the "property to be ascribed") must also be 極成. And the act of *ascribing* that predicate to that subject must be something the proponent genuinely intends to establish. In our example:
"Sound is impermanent."
This is what we want the other person to understand.
Now — and this is a point the 《正理滴論》 insists on very firmly — the 宗 is the thing to be established (所立), and *only* the thing to be established. It is not simultaneously a *means* of establishment. If you reuse some piece of content both as a thesis and as a reason, you have not built a valid 能立; you have confused the layers. The 宗 sits on the ground; the 因 and 喻 reach up to support it.
因 (hetu) — the reason. This is what makes the thesis follow. The classical example:
"Because sound is produced (所作性)."
In other words: produced things are impermanent; sound is produced; therefore sound is impermanent. But 因 is not allowed to be just any supporting statement. It must satisfy three conditions — the 三相 (trairūpya). This is the heart of the system.
The three aspects of a valid reason:
1. 遍是宗法性 — *The reason must be established in the subject of the thesis.* If you say "sound is impermanent because it is produced," the term "produced" must actually apply to sound. If the opponent denies that sound is produced, your 因 collapses at the first joint. The 《正理滴論》 calls this case 兩俱不成 — non-establishment by both sides. A similar failure is 隨一不成 — non-establishment by either one side; for instance, if a Sāṃkhya proponent tries to argue "pleasure and pain are unconscious (無意識), because they have birth (有生)," but they themselves do not accept that what has birth must be unconscious, they have failed on their own side. In real life, this failure is everywhere: we cite a reason whose applicability to the subject the other person does not even grant.
2. 同品定有性 — *The reason must definitely be present in similar instances* — that is, in things that *do* have the property you are trying to establish. In our example: impermanent things (the *similar instances*, 同品) — bottles, pots, lightning, the seasons — are they all produced? Yes. Wherever we find impermanence, we find that the thing was produced. If we found an impermanent thing that was *not* produced, our 因 would fail at the second joint. (The 《因明入正理論》 explicitly names 瓶, *bottles/pots*, as the standard 同品 example.)
3. 異品遍無性 — *The reason must be entirely absent in dissimilar instances* — things that *do not* have the property you are establishing. In our example: things that are *not* impermanent (the *dissimilar instances*, 異品) — space (虛空), perhaps the self in certain philosophical schemes — are they produced? No. If we found a non-impermanent thing that was produced, our 因 would fail at the third joint. (And the 因 is stated in *negated* form for the 異品: 虚空 is "non-produced" — 非所作 — which the text glosses as "無所作," *unproduced*. This little detail matters because it shows that the 因 is being stated not just positively but also by contrast.)
If any one of these three aspects is missing, you do not have a valid 能立. You have what the tradition calls a 似能立 — a proof that looks like one but isn't. 《正理滴論》 puts it bluntly: 「無有因，即不能成立宗。」 — *Without a [valid] reason, the thesis cannot be established.*
喻 (dṛṣṭānta) — the example. The 喻 comes in two mandatory forms.
- 同法喻 — *example of similarity.* It states: wherever the reason is found, the property is found. *"若所作見，彼無常，譬如瓶等"* — "Whatever is observed to be produced is impermanent, like a pot." (瓶 = pot/bottle is the canonical 同法喻 exemplar.) - 異法喻 — *example of dissimilarity.* It states: where the property is not found, the reason is not found. *"若是常，見非所作，如虛空等"* — "Whatever is permanent is observed to be unproduced, like space."
Both forms are required for the full 能立. Skipping the 异法喻 is one of the most common moves in everyday reasoning: someone gives you a clean example of the same-pattern case, but never shows you what happens on the opposite side, leaving the inference half-built.
Beyond the basic shape
The 《正理滴論》 adds further refinements that show how seriously this tradition took the question of when a reason actually has the right kind of relationship to what it is supposed to establish.
Two kinds of relationship between 因 and the thing to be established:
- 自性因 (svabhāva-hetu) — the reason *is* the very nature of what is being established. (For example, using "cognition" itself as the reason to establish "the self is cognizing" — the reason and the established are not distinct things.) - 果因 (kārya-hetu) — the reason is *produced by* what is being established. (For example, using "trembling of the hand" as the reason to establish "fear" — fear produces the trembling.)
If a thing is neither the nature of what is to be established nor its product, it has no 相系屬 (*essential relation*) to it, and cannot serve as a valid 因. *「無有因，即不能成立宗」* — no essential relation, no proof.
未緣到因 (anupalabdhi-hetu) — the non-perception reason. This is a special form with particular contemporary resonance. The structure is: *If this thing existed, it would necessarily be perceived (or cognizable) by us. We have not perceived it. Therefore, it does not exist.* But — and this is the careful part — this only works when the thing, if it existed, would in fact be accessible to perception: not hidden by distance, time, or the inherent nature of the thing itself. The 《正理滴論》 warns that if something is merely hidden from *your* perception (e.g., on the far side of the moon, or in the distant past), you cannot conclude from your non-perception alone that it does not exist. You only get this inference when the thing's existence and your perception of it are tightly bound.
This is the precise move behind a great deal of contemporary argument: "I haven't seen any evidence of X, therefore X is not real." The 因明 tradition does not dismiss such reasoning; it *qualifies* it. It works when X would necessarily have shown up; it fails when X could be there without you having noticed.
The似能立 catalogue
The 《因明入正理論》 provides a strikingly detailed catalogue of failure modes. There are:
- Eight defects of a 似立宗 (defective thesis): 現量相違 (contradicting direct perception), 比量相違 (contradicting valid inference), 自教相違 (contradicting one's own scripture), 世間相違 (contradicting common worldly view), 自語相違 (contradicting one's own prior statement), 能別不極成 (predicate not acknowledged by both), 所別不極成 (subject not acknowledged by both), 俱不極成 (neither acknowledged by both). - Five defects of a 似同法喻: 能立法不成, 所立法不成, 俱不成, 無合, 倒合. - Five defects of a 似異法喻: 所立不遣, 能立不遣, 俱不遣, 不離, 倒離.
The catalogue's existence itself is part of the point: this is a discipline that has thought hard about how argument *goes wrong*, in many specific ways, each with a name. The closing line of the section is: 「如是等似宗、因、喻言，非正能立。」 — *Such defective statements of thesis, reason, and example are not a valid 能立.*
What you should not do is treat the catalogue as a memorizable list. The catalogue is the visible surface of a much deeper commitment: every part of the structure can fail in a specific way, and a person who wants to actually open understanding in another has to know how to check each joint.
Life walk-through
The translation from canonical logic to ordinary situations is best done carefully. Let me take three contemporary scenes and walk them through the structure. I'll deliberately use scenes where someone is *trying* to make a serious point to someone else, not just chatting.
Walk-through A: A team trying to decide whether to rewrite a critical service
The setting. A senior engineer in a small company believes the team's core service needs to be rewritten. The rest of the team is hesitant — they've heard this claim before, from previous engineers, and the rewrite projects always took longer than promised and broke things. The senior engineer has to make the case carefully.
Mapping the structure.
- 宗 (the thesis to be established): *"This service needs to be rewritten within the next two quarters."* (Note: the *what to be established* is the call to action — the rewrite — not some abstract property.) - 因 (the reason offered): *"The current service has no integration tests, no formal API contracts, and three of the last four outages were caused by changes that should have been caught by tests but weren't."* - Checking 三相: - *遍是宗法性:* Does the team acknowledge that the service has these properties? This is where 极成 matters. The senior engineer's claim needs to be a fact both sides agree on, not an accusation. If the team doesn't acknowledge the lack of tests, the 因 fails here. Most failed arguments in real meetings fail right here — one side cites a "reason" the other side doesn't even grant as a description of reality. - *同品定有性:* Are all services with these features ones that genuinely need rewrites? Here's a subtle joint: a service can have no tests and still be small and stable enough not to need a rewrite. So this aspect is not automatic. The senior engineer needs to show that the pattern "no tests + recent outages + brittleness" tracks with "rewrite needed." They might say: *every service we know of that had this profile eventually either was rewritten or caused a major incident* — that's a 喻. - *異品遍無性:* Are all services that don't need rewrites ones that *don't* have these features? In their shop, are there services that are stable, well-tested, and quietly doing their job — i.e., services that have tests, contracts, and no recent outage-causing-change pattern? Yes. So the 异品 has no 因 present. This joint holds. - 喻: - *同法喻:* "Every service we've seen in this category — like the billing service two years ago, like the analytics service last year — that had no integration tests and was changing constantly eventually required either a rewrite or a major rescue effort." (瓶 *pot* here corresponds to "the billing service.") - *异法喻:* "Conversely, services like the auth service, which has strong test coverage and stable contracts, have not needed a rewrite, even though they're old." (虚空 *space* here corresponds to "the auth service.")
Where the standard 似能立 fails here. The senior engineer's most common error would be to assert 因 that the team does not actually 极成 (e.g., "the codebase is a mess" — what counts as a mess is contested). They might also omit the 异法喻 entirely — never naming a service that *is* old and *does not* need a rewrite, which leaves the inference floating. And they might confuse 宗 and 能立: trying to *prove* the rewrite is needed by *saying* it is needed, over and over, instead of building the supporting structure.
The canonical move, translated into a meeting, would be: slow down. State the 宗 once. Pause. State the 因, with the facts both sides already accept. Pause. Walk through the same-pattern examples. Pause. Walk through the counter-examples that *don't* need rewriting. Then stop. If the structure holds, the other team will come to the conclusion themselves. If it doesn't hold, you'll find out exactly which joint is broken — and that is itself a useful result.
Walk-through B: A doctor explaining a diagnosis
The setting. A patient comes in with persistent fatigue, weight loss, and intermittent low-grade fever for six weeks. The doctor believes they likely have a specific condition and wants to make the case to the patient clearly enough that they accept the recommended test.
Mapping.
- 宗: *"Your symptoms are most likely caused by condition X."* - 因: *"You have had persistent fatigue, weight loss, and intermittent low-grade fever for six weeks, with no other obvious explanation."* - 三相: - *遍是宗法性:* Does the patient agree they have these symptoms? Usually yes, because they are experiencing them. 极成 is high here — both sides share the observation. - *同品定有性:* In cases that turn out to be condition X, are these symptoms reliably present? This is exactly the kind of question a careful doctor checks against the clinical literature. If many cases of X present with a different symptom cluster, the 因 may not satisfy this aspect for this particular patient. - *異品遍無性:* In cases that are *not* condition X, are these symptoms reliably *absent*? No — there are many other conditions that produce these same symptoms. So this third aspect is the weak joint. The doctor has to be honest about this: "These symptoms alone are not specific. The reason I'm leaning toward X is also that your lab panel shows Y, which is more characteristic of X than of alternatives."
This is a beautiful example of where real diagnostic reasoning sits: the 因 is not a single fact but a layered one, and the doctor has to keep checking each aspect honestly, including acknowledging where the third aspect fails.
- 喻: The doctor might say: *"Patients with this combination — fatigue, weight loss, low-grade fever, plus your lab findings — in my experience, about two-thirds of them turn out to have condition X."* And implicitly: *"Patients with a different combination rarely turn out to have it."*
The point is not that doctors do formal 因明. It is that good clinical reasoning already uses this structure; the formal vocabulary just makes it visible.
Walk-through C: A partner saying "I feel unheard"
The setting. In a long-term relationship, one partner tells the other: "I don't feel heard." This is a familiar situation and it almost always goes badly. Why? Because both sides end up arguing without a shared structure.
Mapping.
- 宗: *"In our recent conversations about X, I have not felt heard."* - 因: *"When I brought up X, you responded by Y within a few sentences — before I had finished."* - 三相: - *遍是宗法性:* Does the other person acknowledge that this happened? Often the dispute is precisely here. One partner says "I always let you finish"; the other says "you don't." This is a 极成 failure at the very first joint — the *subject* (the behavior) is not agreed upon. - *同品定有性:* In situations where the speaker genuinely is being heard, does this behavior (interrupting-to-solve) reliably *not* appear? Yes — when both are listening fully, the response typically waits. So this aspect can be checked if you can find a recent example where the speaker *was* heard and the pattern differed. - *異品遍無性:* In situations where the speaker is not being heard, does this behavior reliably appear? Not always — there are other ways of not being heard (distraction, dismissal, topic changes). So the 因 is sufficient but not exclusive. - 喻: - *同法喻:* "When your sister talks to her partner about work, and he jumps to solutions before she finishes, she feels exactly like this — unheard." - *异法喻:* "When your sister talks to her friend A — who always waits and reflects back what he heard — she feels heard, even when they disagree."
Where this typically 似能立. The speaker is usually tempted to either skip 极成 (assume the behavior is obvious) or to fuse 宗 and 能立 (say "you don't listen to me" repeatedly, which is the *thesis* being asserted as if it were the *proof*). The structure requires saying: here is the specific behavior; do you agree it happens? Here is the reasoning; here is when this behavior does and does not appear; here are examples outside us where the same pattern holds.
This is not a magic formula that will make the relationship heal. But it does something rare: it puts the disagreement on a structure that both sides can actually examine, instead of leaving it on the level of mutual accusation.
The most common mapping error
Across all three walks, the most common mistake is this: people treat 宗 and 能立 as the same thing. They make a thesis statement ("this needs rewriting"; "you have X"; "I feel unheard") and then re-assert it more loudly or more emotionally as if repetition were proof. But repeating the thesis does not give you a reason, an example, or a counter-example. It just gives you the same thesis again. The structure of 能立 requires *external* support — reasons and examples that both parties can examine. Without that, the argument has no 因三相 at all. 《因明入正理論》 calls this very directly: "如是等似宗、因、喻言，非正能立."
Why contemporary people need this
It is striking that a logic system formalized more than a millennium ago should feel so urgent in our moment. There are at least four reasons.
1. The discipline of three checks
Modern discourse is exhausted by single-criterion arguments. Someone asserts a claim; someone else asserts a counter-claim; nothing resolves because neither side is checking whether their reason actually satisfies three structural conditions. The 三相 give you a small, portable set of questions to ask about any argument you encounter — including your own:
- *Does the reason actually apply to the subject?* (遍是宗法性) - *Is the reason present in all cases that share the predicate?* (同品定有性) - *Is the reason absent in all cases that don't share the predicate?* (異品遍無性)
These are not abstract scholastic questions. They are the difference between a real reason and a *feeling of a reason*. Anyone who has been in a long argument knows the moment when one side produces a "reason" that does not even apply to the thing they are arguing about — and knows how powerless the other side feels in that moment, because they cannot articulate why it doesn't apply. The three aspects are the articulation.
2. The collapse of 极成 in public discourse
Perhaps the most painful contemporary failure is at the level of 极成 — mutual acknowledgment. In any structured debate, both parties must agree on the basic terms (the 有法 and 能別) before the argument can even begin. In contemporary public discourse, this layer is missing. Two sides will use the same word ("freedom," "fairness," "safety," "innovation") to mean different things, and then proceed to argue as if they shared the term. The result is unresolvable disagreement that masquerades as disagreement about a thesis, when it is really disagreement about the subject and predicate themselves.
The 因明 move is unsentimental about this: until both sides acknowledge what they are talking about, no 能立 is even possible. This is not a counsel of despair; it is a counsel of order. Before you try to prove P, ensure the other person agrees on what P is about. Much public conflict is dissolved by this single preliminary move.
3. The diagnostic vocabulary of 似能立
The eight defects of the thesis, the five defects of the same-pattern example, the five defects of the counter-pattern example: these give you a vocabulary for naming things you have always felt were wrong in arguments. You have heard someone argue against their own earlier position — *that is 自語相違*. You have heard someone assert something obviously contradicted by perception — *that is 現量相違*. You have heard someone give an example that does not actually contain the reason — *that is 能立法不成*. You have heard someone argue from cases on one side only, never showing the absence of the pattern on the other side — *that is倒合 or 不離*.
In an age of generated and flooded text, this diagnostic vocabulary is invaluable. A model — whether human or algorithmic — can produce an utterance that *looks like* a proof. Without these categories, you cannot tell whether it actually *is* one. (And here a careful caveat: the parallel between the 因明 structure and the structure of contemporary large language models is suggestive, not identical. The systems have different architectures and different sources of failure. What is genuinely shared is the observation that fluent output can mask structural invalidity. 因明's contribution is the vocabulary for spotting the latter, whatever the former is.)
4. The ethical core: 唯悟他
This is the part easiest to miss. 能立 is not about winning. It is about producing *actual understanding* in the other person. The canonical definition ties 能立 explicitly to "開示諸有問者未了義" — opening up what is not yet understood. If the other person assents but does not understand, you have not done 能立. If they understand but disagree, you have still done 能立; they just have not been persuaded because the *thesis itself* is wrong, which is a different problem.
This subtle separation — between *opening up the unclarity* and *winning agreement* — is one of the most quietly radical moves in the tradition. Most modern discourse conflates these. We assume that an argument has succeeded if the other side agrees. The 因明 tradition says: no, your argument has succeeded only if the other side *understands*. They can understand you and still reject the thesis. That rejection is their prerogative. The proof has done its work; the conclusion is for them.
This is also why 能立 is paired with 能破 — constructive proof with critical refutation. Both belong to 唯悟他. Both serve the other person's clarity, not your victory.
5. The 未缘到因 in our time
Finally, the most contemporary-feeling move in the system is 未缘到因 — the non-perception reason. We live in an age saturated with arguments of the form: "I have never seen evidence of X, therefore X is not real." This move is not inherently invalid; it is *qualified*. The qualification is exact: it works only when the thing's existence would entail its being perceived (which depends on the thing's nature, location in time, and accessibility to inquiry). The 因明 tradition refuses to either embrace this move uncritically (as some forms of naive empiricism do) or dismiss it entirely (as some forms of apologetics do). It specifies the conditions under which non-perception is evidence of non-existence, and the conditions under which it is not. This is the precise middle path through a debate that in our era is usually conducted with blunt instruments.
Common misreadings and clarifications
能立 is not persuasion
The single most common misreading is to translate 能立 as "persuasion" or "rhetoric." It is not. Persuasion is the act of making someone accept a conclusion, possibly without understanding it, possibly through emotional or social pressure. 能立 is the construction of an argumentative object whose structural validity can be examined by both parties. A speech that moves a crowd is not a 能立; a careful chain of reasoning that produces understanding in a single careful interlocutor can be. The measure of 能立 is not assent but understanding.
因 is not just any supporting evidence
In modern usage, "reason" or "evidence" can mean almost anything cited in support of a conclusion — a feeling, an authority, an anecdote. In 因明, 因 is a precisely constrained object: it must be established in the subject, present in all cases that share the predicate, and absent in all cases that don't. If your evidence does not pass these three tests, you do not have an 因; you have a *feeling of an 因*. 《正理滴論》 calls this 似因 and traces it to specific failure modes, including the most common one — citing something as a reason that does not even apply to the subject.
喻 is not the same as analogy
Analogy in modern use is loose: "this is like that, therefore this." 喻 in 因明 is structured. It comes in two required forms — 同法喻 (similar-case example) and 异法喻 (dissimilar-case example). The 同法喻 establishes the pattern "where reason, then predicate." The 异法喻 establishes "where not-predicate, then not-reason." Skipping the 异法喻 is one of the most common moves in contemporary reasoning, and one of the most damaging. If you have only ever seen the pattern on one side, you have not actually checked whether the inference holds.
极成 is not optional
Some modern interpreters treat 极成 — mutual acknowledgment of the subject and predicate — as a formality. It is foundational. Without it, the argument has no foundation to stand on. If you and your interlocutor do not share the meaning of the basic terms, the entire structure is built on sand. The 因明 move is to do the work of establishing 极成 *before* attempting 能立 — or to recognize that no 能立 is possible in the present exchange.
宗 is the thing to be established; 能立 is the means
The distinction between 所立 (what is to be established) and 能立 (the means of establishment) is fundamental. 宗 is the former; 因 and 喻 are parts of the latter. Conflating them — using the thesis as if it were also the proof — is the move the 《因明入正理論》 explicitly warns against. In real life, this conflation is the rule rather than the exception. "X is bad, so we must stop it." "I told you so." "Everyone agrees." These are repeated theses, not proofs. The 宗 does not prove itself.
能立 does not mean the thesis is true
This is a subtle but important point. A valid 能立 produces understanding of the thesis in the other; it does not guarantee that the thesis is true. The structure of 能立 is a *vehicle for understanding*. Whether what is understood is actually the case is a separate question — and one that the system addresses elsewhere (through 現量 and 自悟之 比量, etc.). In modern terms: a logically valid argument for a false conclusion is still a valid 能立 in this technical sense; it has simply successfully produced understanding of something incorrect. The system is not naive about this. The structure checks whether the argument works as an argument, not whether the world agrees with it.
唯悟他 is not "winning"
The whole point of 能立 — and the reason it is grouped with 能破 under 唯悟他 — is that the goal is understanding, not victory. If the other party understands you and rejects the thesis, you have done your job. If the other party agrees with you without understanding, you have not. This is a discipline that pushes against the grain of competitive argumentation. It asks: what is the *purpose* of your argument? If the answer is "to win," you are not doing 能立. If the answer is "to produce understanding," you may be.
A small structure, then: a thesis, a reason with three aspects, two examples (one matching, one contrasting), all built on terms both parties have acknowledged, with the explicit goal of producing understanding in the other. It is austere. It is gentle. It asks almost nothing of you except that you actually mean what you say and are willing to put your reason under examination.
In an age when almost every form of discourse has forgotten these disciplines, *能立* is worth remembering not as a scholastic curiosity but as a quiet instruction in how to speak to another human being in a way that genuinely opens their understanding — or, when it does not, at least shows you exactly where the joint broke.