What this concept is about
Let us begin with the most easily misunderstood point: the "利他" (parārtha, "for others") in 利他比量 (parārthānumāna, "inference for others") is not the usual sense of charity, helping others, or the altruistic activities of the bodhisattva path. Here "利他" is a term from 因明 (hetuvidyā, Buddhist logic), meaning "to set forth a statement for the sake of making others understand."
When a person, on the basis of reliable reasons, draws a conclusion in his own mind, this is called 自利比量 (svārthānumāna, "inference for oneself"); when this understanding is organized into a linguistic formulation that others can examine, so that they too may come to know it through the reasons, this is 利他比量 (parārthānumāna, also translated as "inference for others" or "inference on behalf of another's meaning"). 《正理滴论》 (Nyāyabindu) summarizes it: "利他比量，说为三相因者，谓于因安立果故。" *("Inference for others is the statement of a reason possessing the three characteristics; that is, establishing the conclusion on the basis of the reason.")*
It is not about voicing one's convictions loudly, nor about overwhelming others with rhetoric. What it truly does is this: it turns "why I judge in this way" into a path of cognition that the other person can inspect and follow.
For example, you see smoke on a distant mountain and judge that there is fire there. This is 自利比量 (svārthānumāna, inference for oneself), occurring in your own mind. If you wish a companion also to understand, you say:
此山有火，因为有烟；凡有烟处皆有火，如灶等。
*("This mountain has fire, because it has smoke; wherever there is smoke, there is fire, as in a kitchen hearth.")*
Only at this point does it enter 利他比量 (parārthānumāna, inference for others). The difference between inference for oneself and inference for others does not lie in the standard of truth of the reasons: both must rely on a 正因 (saddhetu, valid reason) endowed with the full three characteristics. The difference is only that one is an inner act of inference, the other is a linguistic establishment offered to others.
The so-called 三相 (trirūpa, "three characteristics") cannot be vaguely rendered as merely "having three conditions." With respect to the 宗 (pratijñā, thesis) to be established, the 因 (hetu, reason) must satisfy:
1. 遍是宗法 (paksadharmatā, "universally a property of the subject"): the reason must actually exist in the object under discussion. To say "this mountain has fire, because it has smoke," one must first confirm that the mountain indeed has smoke—one cannot mistake mist for smoke. 2. 同品定有 (sapakṣe sattvam, "present in similar instances"): in cases already recognized as possessing the property to be established, the reason must be observable. A hearth has fire and also has smoke, thus providing a positive relation of concomitance. 3. 异品遍无 (vipakṣe 'sattvam, "absent from dissimilar instances"): in cases lacking the property to be established, the reason should not be present. If places without fire could equally have smoke, then "smoke" could not decisively establish "fire."
The three characteristics jointly prevent three kinds of slippage: the reason not existing in the subject at all; the reason and conclusion not being stably concomitant; and the existence of counterexamples sufficient to overturn the inference. 利他比量 (parārthānumāna) is not a fixed sentence pattern but the expression of such a determinable relation.
《正理滴论》 (Nyāyabindu) divides the forms of inference into two types according to the mode of expression:
- 具同品法，又称随因后遍 (anvaya-vyāpti, "positive concomitance"): expressed through the positive conjunction of the reason with similar instances. The classical stock example is "凡有，皆是无常，犹如瓶等" *("whatever is produced is impermanent, like a pot")*. It states positively: whatever possesses the reason is accompanied by the property to be established. - 具异品法，又称随因遣遍 (vyatireka-vyāpti, "negative concomitance"): expressed through the reverse relation of the reason being absent from dissimilar instances—that is, if the property to be established is excluded, the reason must be excluded along with it. The typical form of an effect-reason is: "若无火，则不出烟；此处有烟。" *("If there were no fire, there would be no smoke; here there is smoke.")*
These two are not two unrelated sets of truth. As 《正理滴论》 (Nyāyabindu) says: "无论立何种量式，皆具此二义，无有差别。" *("Whichever form of inference one employs, both aspects are present; there is no difference.")* The positive formulation emphasizes concomitance with similar instances; the negative formulation emphasizes exclusion from dissimilar instances. A reason that truly possesses the three characteristics can withstand examination from both directions.
The classical sources also preserve a more difficult example of the negative form: "若是青色等支分，而于此处缘虑之所明了者为瓶，以未如实缘到瓶。" *("If [the object] were a pot with blue and other features, it would be clearly apprehended here by the mind; yet the pot is not actually apprehended.")* The point lies not in whether the Chinese reads smoothly, but in establishing an exclusion through the reasoning: "if the object really exists and is in a condition accessible to cognition, it ought to be apprehended; now it has not been truly apprehended." This involves the 未缘到因 (anupalabdhi-hetu, reason of non-apprehension), and cannot be simplified to "I did not see it, therefore it does not exist." Non-apprehension has probative force only when, if the object existed, it would necessarily be reliably cognizable, and the conditions of observation have been fulfilled. Not seeing someone in the courtyard at night does not directly establish that no one is there; only after searching thoroughly under adequate light with an unobstructed view and still finding nothing might a corresponding reason of non-apprehension be constituted.
The three kinds of valid reasons spoken of in 《正理滴论》 (Nyāyabindu) each involve different relations of determination:
- 自性因 (svabhāva-hetu, nature-reason): inferring from a property inherent in the thing itself that determines its classification, as in "此是树木，是沉香树故" *("this is a tree, because it is an aloe-wood tree")*. It is not that merely attaching any category-name to something allows one to infer a higher-order property; rather, "aloe-wood tree" is itself subsumed under "tree." - 果因，又译果性因 (kārya-hetu, effect-reason): inferring the cause from its result, as inferring fire from smoke. The "result" must have a reliable relation to the cause being inferred; damp ground might be caused by rain, by sprinkling, or by a leak, so damp ground alone cannot without qualification establish that it has rained. - 未缘到因，又译不可得因 (anupalabdhi-hetu, reason of non-apprehension): establishing non-existence through not apprehending something where it should be apprehendable. The key is that the condition "if it existed, it would be obtainable" has been fulfilled, rather than treating personal ignorance as proof of non-existence.
利他比量 (parārthānumāna) also involves the rigorous arrangement of 宗、因、喻 (pratijñā, hetu, dṛṣṭānta—thesis, reason, example). The 宗 (pratijñā, thesis) is the proposition to be established; the 因 (hetu, reason) is the ground for establishing it; the 喻 (dṛṣṭānta, example) helps display the relation of pervasion between the reason and the property to be established. But Dharmakīrti did not regard the example as a decorative element working independently of the reason. 《正理滴论》 (Nyāyabindu) says: "喻虽为能立之支分，但离之别无有体，故未另立其相。" *("Though the example is a branch of the proof, apart from this [relation] it has no separate essence; therefore no separate characteristic is set up for it.")* The meaning is not to abolish examples, but to say: the probative force of an example comes from the relation between reason and property that it displays; apart from that relation, however many examples one cites, they are merely stories.
Therefore, "I know three people who started successful businesses and all woke up early, so waking up early must make people successful" is not a valid 利他比量 (parārthānumāna). Examples can only help display pervasion; they cannot replace it. Three striking samples certainly cannot automatically rule out a large body of counterexamples.
The 宗 (pratijñā, thesis) itself is not merely any sentence awaiting debate. 《正理滴论》 (Nyāyabindu) says that a thesis free of fault must possess the four aspects: "自之自性、自己、意乐、非被排除" *("its own nature, itself, the intention [of the proponent], and not being excluded")*—it must be the proponent's own claim genuinely intended to be established, put forward as what is to be established, and not already excluded by valid cognition or common ground. If a claim collides at the outset with reliable knowledge, it becomes a 似宗 (pratijñābhāsa, "apparent thesis"). The treatise lists four kinds of contradiction:
- 现量相违 (pratyakṣa-viruddha, contradicted by direct perception), such as "声非所闻" *("sound is not audible")*; - 比量相违 (anumāna-viruddha, contradicted by inference), such as "瓶是常住" *("the pot is permanent")*; - 世间相违 (loka-viruddha, contradicted by common usage), such as "怀兔非月" *("the hare-in-the-moon is not the moon")*; - 自语相违 (svavacana-viruddha, contradicted by one's own words), such as "比量非量" *("inference is not a valid cognition")*.
A reason may also fail to hold at all. For example, establishing "声是无常" ("sound is impermanent") by means of "眼所见" ("being seen by the eye")—since sound is not seen by the eye, this reason does not hold with respect to the object under discussion; it is a case of 两俱不成 (ubhayāsiddha, "not established for either party"). Even if the conclusion "sound is impermanent" has a sound basis elsewhere, this cannot rescue the erroneous proof. 因明 (hetuvidyā, Buddhist logic) examines whether this particular argument holds, not merely whether the conclusion happens to be right.
《菩提道次第广论》 (Lamrim Chenmo, *The Great Treatise on the Stages of the Path to Enlightenment*) advances this method into the specific context of Madhyamaka debate. The text mainly speaks of 他比量 (parārthānumāna in the opponent-directed sense), 他许量 ("inference from what the opponent accepts"), or 他已成比量 ("inference already established by the opponent"), and does not directly use the name 利他比量. Here it is connected with the general sense of "setting forth an inference for others," but it particularly emphasizes: drawing on premises the opponent already accepts, to expose the internal contradiction within his own position.
"他" (the other) is relative to a standpoint. For the opponent himself, those premises are "自许" (what he himself accepts); from the standpoint of the Madhyamaka proponent, they are "他许" (what the other accepts). It is not that there exists a mysterious set of "truths belonging to the other"; rather, the proponent does not for the moment require the opponent to accept his entire theory, but asks only: given the conditions you yourself already acknowledge, can your conclusion still stand?
The formulation cited in 《广论》 is:
"凡不能见自者，亦不能见他，如瓶；眼不能见自，故眼不能见他。"
*("Whatever cannot see itself also cannot see other things, like a pot; the eye cannot see itself, therefore the eye cannot see other things.")*
The intention here is to use conditions the opponent accepts—such as "the eye cannot see itself"—to expose the internal difficulty faced by his claim that the eye can see other things. It does not first construct a separate theory of "self-nature established" commonly accepted by both parties and then prove the Madhyamaka position from it; rather, it is "以他已成比量而为破除" *("refuting by means of an inference already established by the other")*.
But "他许" (what the other accepts) by no means equals "anything the opponent has said may be used to infer whatever one likes." In an ordinary inference, the 有法 (dharmin, subject) that serves as the basis of discussion cannot carry qualifications that only one side—the proponent or the opponent—accepts while the other does not; both parties must at least be able to point to what is being discussed, and the property to be established, the reason, and the example must each have a corresponding basis. 《广论》 (Lamrim Chenmo) especially emphasizes that the subject, the property, the reason, and the example must be tenable within conventional designation (名言). What Madhyamaka rejects is the existence of a 量 (pramāṇa, valid cognition) commonly accepted by both parties that could prove that phenomena are established by intrinsic nature; it does not, by refuting intrinsic nature, destroy the harmless cognitions of conventional language.
This also explains the crucial difference between 他比量 (parārthānumāna in the opponent-directed sense) and 自续比量 (svatantrānumāna, autonomous inference). In the Madhyamaka debate as set forth in 《广论》, autonomous inference requires that proponent and opponent first jointly acknowledge a 量 (pramāṇa) capable of establishing things as possessing intrinsic characteristics, then jointly establish the three characteristics of the reason, and finally draw the conclusion. The Prāsaṅgika does not accept this starting point, and therefore uses inference-from-the-other's-acceptance to reveal the consequences of what the opponent himself acknowledges. This is by no means "the Prāsaṅgika school does not use logic" or "it only asks counter-questions"; it is a refusal to smuggle in an ontological premise not yet commonly accepted, while still rigorously employing the form of inference.
Placing it back in the entire path of liberation, the position of 利他比量 (parārthānumāna) becomes clear. In Buddhism, 闻、思、修 (śruta, cintā, bhāvanā—hearing, reflecting, cultivating) do not replace meditation with logic, nor do they require one to stop reasoning and believe directly. Hearing brings one into contact with correct views; reflection reduces doubt and distortion through observation, analysis, and inference; cultivation turns the view that has been examined and decided into a stable force that directly affects body and mind. 利他比量 (parārthānumāna) is located mainly at the juncture of "hearing" and "reflection": a spiritual teacher or commentator lays out the reasons he has understood, so that the listener is not merely having his doubts temporarily suppressed by authority, but can personally come to know through the reasons.
It is not itself liberation. A person can be proficient in argumentation while afflictions blaze, and can use logic for the sake of victory. But in the right direction, it removes the obstacles of wrong views before the path: distinguishing what is merely habitual assertion from what is established cognition; seeing how the attachments to permanence, pleasure, self, and purity depend on unexamined premises; and thereby providing an unshakable basis of discernment for the contemplative cultivation of impermanence, selflessness, dependent arising, and emptiness. It serves wisdom, and must itself be constrained by dispassion, compassion, and honesty.
A walkthrough in daily life
Consider a very contemporary scenario: after work, you send a few messages to your intimate partner, and there is no reply for two hours. A conclusion quickly forms in your mind: "He no longer cares about me." Then comes hurt, anger, even dredging up old grievances.
Let us first unpack the 自利比量 (svārthānumāna, inference for oneself) occurring in the mind:
- 有法 (dharmin, the subject, i.e., the object being judged): this instance of two hours without a reply. - 所立法 (sādhya-dharma, the property to be established): the other person does not care about me. - 因 (hetu, the reason): the other person did not reply to my message in a timely manner. - The hidden pervasion relation: whoever does not reply to me in a timely manner does not care about me. - A memory serving as a 同喻 (sapakṣa-dṛṣṭānta, similar example): last time someone drifted away from me, they also went a long time without replying.
The most likely point of mis-projection here is treating "a phenomenon that accompanied estrangement" as "a valid reason sufficient to determine estrangement." An example is not pervasion; correlation is not determination.
Now walk through it step by step according to the three characteristics.
First, does the reason hold with respect to the subject? "No reply yet" probably holds, but "deliberately not replying after seeing the message" does not. The phone may be on silent; the other person may be in a meeting, driving, caring for a child, or caught up in an urgent matter. Your rewriting of "no reply received" into "deliberate neglect" has already smuggled the conclusion into the reason.
Second, is the reason present in similar instances? There are indeed people who do not care about a relationship and habitually put off replying, so "not replying" and "not caring" do sometimes co-occur. But this only shows possible correlation; it cannot by itself complete the proof.
Third, is the reason absent from dissimilar instances? Are people who still care about their partner never delayed in replying? Clearly not. As long as there exists a reliable counterexample—someone who cares deeply but fails to reply promptly due to work, exhaustion, or an unexpected event—the original pervasion cannot stand.
Therefore, this reason is not a valid reason with three characteristics that can decisively establish "not caring." The more careful conclusion can only be: "No reply has been received at present; the cause is not yet certain." This is not forcing oneself to be optimistic; it is refusing to let the evidence bear a conclusion it cannot support.
If you need to communicate this understanding to yourself or to your partner, you might form a more honest "statement for the other":
我看到消息两小时没有回复，这是已经发生的事实；但仅凭延迟回复，不能决定你是否在乎我，因为在乎关系的人也可能因许多事务暂时不回。我的不安是真的，但我对原因的判断还需要更多信息。
*("I see that two hours have passed without a reply to my message; that is an established fact. But the delay alone cannot determine whether you care about me, because even people who care about a relationship may temporarily be unable to reply due to many affairs. My unease is real, but my judgment about the cause still needs more information.")*
This passage may not be a standard syllogism in a traditional debate arena, yet it embodies the spirit of 利他比量 (parārthānumāna): separating the subject, the claim, the reason, and the pervasion; not passing off feelings as proof; and not denying feelings because the proof is insufficient.
Now consider the 未缘到因 (anupalabdhi-hetu, reason of non-apprehension). You open the team document, do not see the report your colleague was supposed to submit, and say: "He simply didn't do it." There are at least two layers to check: are you looking in the correct folder? Has the synchronization completed? Did the other person name and submit the file as agreed? Only when the condition "if it had been submitted, it would necessarily be displayed here" reliably holds can "not seen here" possibly establish "not yet submitted." Otherwise, it is merely that you have not apprehended it—not that the object does not exist.
Pseudo-inferences also frequently appear while scrolling through one's phone: after several similar short videos in a row, one infers "everyone thinks this way." What the algorithm displays is a sample filtered by a recommendation mechanism, not the distribution of similar and dissimilar instances across society as a whole. One may use one's experience with algorithms to understand that "examples are not pervasion," but one should not equate Buddhist logic with modern statistics: 因明 (hetuvidyā) is concerned with how a reason determinately establishes a thesis, whereas statistical inference deals with sampling, probability, and error; the two can remind each other, but their theoretical aims and methods are not the same.
In an overtime-at-work scenario, a supervisor says: "Excellent employees all take the initiative to work overtime; Xiao Lin often works overtime, so Xiao Lin is excellent." Even leaving aside the question of how "excellent" is defined, this argument may conflate causation and classification: working overtime might stem from low efficiency, understaffing, unreasonable scheduling, or an inability to refuse. If there are people who do not work overtime yet complete work of high quality, the dissimilar-instance relation does not hold. Using a visible behavior saturated with institutional factors to infer a sweeping evaluation of character is particularly in need of caution.
In parenting, when a child fails one exam, the parents say: "You simply don't work hard." Here "declining grades" is the effect, yet the parents directly designate a single cause. Sleep, anxiety, changes in question types, study methods, and family stress may all play a part. Effect-reason inference requires reliable qualification; the existence of a result does not mean that the cause that first comes to your mind necessarily exists. A better question is: "What conditions might have produced this result? What evidence do we have to rule out other explanations?"
If put to use in real conversation, 利他比量 (parārthānumāna) also carries an ethical check: can the other person understand and freely examine it? If you use jargon, status, or speech speed to leave the other person no room to respond, then even if the sentence form is complete, you have departed from the core aim of "bringing the other to understanding." The completion of 利他比量 lies not in my having finished speaking, but in the reasons having been placed where both parties can examine them together.
Why contemporary people need it
Contemporary life does not lack conclusions; it lacks a visible path between conclusions and reasons. News headlines, short videos, advertisements, workplace rhetoric, and even generative AI can all rapidly produce sentences that sound complete; but fluency of language does not guarantee that the reason holds, still less that there is a relation of pervasion between the reason and the conclusion.
利他比量 (parārthānumāna) offers an extremely practical mental discipline:
- First ask whether both parties are discussing the same object; do not talk past each other over the 有法 (dharmin, subject); - Then ask whether the reason really exists in this object; - Examine positive examples, but also actively seek counterexamples; - Distinguish correlation, consequence, definitional relations, and non-apprehension; do not mix different kinds of reasons; - Do not forgive a bad argument merely because its conclusion suits your position; - Do not require the other party first to swallow your entire worldview; try instead to begin from common acknowledgments both sides already share.
This is especially important for public discussion. Many disputes are not about value conflicts as such, but about each side injecting qualifications valid only within its own theory into a shared object of discussion. On the surface everyone speaks of "freedom," "harm," "fairness," or "tradition," but they are actually referring to different things. 因明 (hetuvidyā) requires first establishing a subject that can be jointly examined; it will not automatically eliminate disagreement, but it can make genuine disagreement visible.
"他已成比量" ("inference already established by the other") also carries a precious wisdom for dialogue: when criticizing the other, do not rush to proclaim your entire position; instead, faithfully understand what the other acknowledges, and then ask whether these commitments are compatible with one another. This is more rigorous than seizing a sentence out of context, and more powerful than affixing labels. It demands a nearly compassionate accuracy—only by truly understanding the other can one point out his contradiction without distortion.
The same applies to one's own practice. We often say, "I must be recognized to have worth," "if things get out of control, everything will be ruined," "this feeling will never end." These sentences torment us not because they have been proven, but because they have been repeated so long that they are experienced as facts requiring no proof. A hetuvidyā-style walkthrough lets one gently ask: What is the reason? Does it hold with respect to the current object? Are there counterexamples? Does the conclusion exceed the evidence?
This is not expelling emotions with logic. Pain can be a real experience, but it does not follow that every explanation accompanying the pain is true. Distinguishing "I am afraid" from "danger is certain to occur" is precisely a small step from being carried along by thoughts toward observing dependent arising as it is.
Modern cognitive science also studies confirmation bias, selective attention, and sample bias, and can help explain why people tend to look only at similar instances and ignore dissimilar ones. But this can only serve as a comparison; it cannot be said that science has "proven the Dharma." Cognitive science has different internal models, and experimental results are constrained by tasks, samples, and interpretive frameworks; Buddhist 因明 (hetuvidyā) sits within an overall context of epistemology, debate, and the practice of liberation. Where the two meet is in their shared reminder: the feeling of certainty is not itself sufficient proof of reliability.
On the path of liberation, there is a deeper layer: attachment often clings not only to conclusions but also to "my view must necessarily be so." 利他比量 (parārthānumāna) trains one to make reasons public and to let cognition be examined; the Madhyamaka 他许量 ("inference from what the other accepts") further shows that certain views of intrinsic nature cannot even withstand the consequences entailed by their own premises. Its purpose is not to build a harder fortress of concepts, but gradually to loosen one's grasping at mistaken modes of existence.
Thus 利他比量 (parārthānumāna) has both an aspect of wisdom and an aspect of compassion. Wisdom prevents the argument from descending into confusion; compassion prevents the argument from becoming humiliation. A debate truly oriented toward liberation is not "I won, so you lost," but rather: letting erroneous reasons lose their power to delude, so that both parties come closer to seeing things as they are.
Common misreadings and clarifications
Misreading one: 利他比量 means reasoning in order to help others.
"利他" here is in the first place a technical term meaning "statement for the sake of others," not a moral sense of selflessness. A person may well use the form of 利他比量 (parārthānumāna) to win a debate. Whether it becomes a skillful means on the path of liberation depends on the motivation, the content, and the manner of use.
Misreading two: 自利比量 is selfish, 利他比量 is selfless.
They are not two ethical ranks. 自利比量 (svārthānumāna) is inference within one's own mind; 利他比量 (parārthānumāna) is stating the same relation of cognition to another. The three characteristics on which both depend are no different; the difference lies in the form of expression.
Misreading three: as long as one utters "宗、因、喻," it is a valid proof.
A sentence pattern does not produce probative force. The reason must hold with respect to the subject, be concomitant with similar instances, and be excluded from dissimilar instances; the thesis must not directly conflict with direct perception, valid inference, common convention, or one's own words. A pseudo-argument with a neat form is still a pseudo-argument.
Misreading four: the 喻 (dṛṣṭānta, example) means citing several examples in support of oneself.
The function of the example is to display the relation of pervasion between the reason and the property to be established, not to manufacture momentum through the number of stories. Dharmakīrti's statement that the example has no independent essence apart from the probative relation is precisely meant to prevent treating examples as proof itself. A single counterexample can sometimes pierce through the false universal judgment behind ten similar stories.
Misreading five: 具同品法 and 具异品法 are two unrelated kinds of inference.
They express the valid reason respectively through positive concomitance and negative exclusion. A qualified valid reason cannot merely show "when it is present, the conclusion is often seen" while evading the question "when the conclusion is absent, is the reason still present?" 《正理滴论》 (Nyāyabindu) explicitly says that the two forms of inference in fact possess both aspects and have no substantial difference.
Misreading six: the 不可得因 (anupalabdhi-hetu, reason of non-apprehension) means "no evidence, therefore non-existence."
For non-apprehension to establish "non-existence," one must first satisfy: if the object existed, it would be cognizable under the complete conditions at hand. Not having searched for and found extraterrestrial life does not directly prove that there is no extraterrestrial life in the universe; having examined a transparent empty cup and seen no water can, within the corresponding scope, establish that there is no water in the cup. What one personally has not seen, what a database has not recorded, and what a search engine has not returned are all not naturally equivalent to non-existence.
Misreading seven: inferring cause from effect means, upon seeing a result, designating a cause.
An effect-reason must rely on a reliable relation and exclude ambiguity. Smoke can infer fire under appropriate qualification, but damp ground cannot unconditionally infer rain. Real-world events are usually multi-causal and multi-conditional; assigning a person a fixed character on the basis of a single result is often not an effect-reason at all, but prejudice wearing the clothing of inference.
Misreading eight: the "他比量" of 《广论》 is simply a synonym for the general 利他比量, with no special context.
They are related, but cannot be flattened into one. 《正理滴论》 (Nyāyabindu) defines 利他比量 (parārthānumāna) mainly as "expressing an inner inference to others in language"; 《广论》 (Lamrim Chenmo) in the context of Madhyamaka debate highlights "relying on what the opponent already accepts to show his contradiction." The latter is a crucial application of inference-for-others, and involves deeper questions such as Prāsaṅgika versus Svatantrika, common establishment, and whether the position is established by a valid cognition.
Misreading nine: the 他许量 ("inference from what the other accepts") means accepting all of the opponent's premises as true.
The proponent may provisionally rely on the opponent's commitments to display their consequences; this does not mean incorporating all those commitments into his own system. "If you insist on A, you must face B" and "I acknowledge that A is really established" are two different things.
Misreading ten: "唯他成" ("established only by the other") means that the subject, reason, and example exist only in the opponent's imagination.
《广论》 (Lamrim Chenmo) does not mean this. The eye, not seeing itself, a pot, and the related property to be established can all be accepted by one's own system within conventional designation; what is genuinely not commonly established is a 量 (pramāṇa) capable of proving these items to be established by intrinsic nature. What Madhyamaka refutes is establishment by intrinsic nature, not the erasure of everything cognized by harmless conventional awareness.
Misreading eleven: Prāsaṅgikas do not use inferential forms; they only point out contradictions.
The meaning of 《广论》 (Lamrim Chenmo) is not to abolish argumentation, but to refuse the requirement that a self-contained proof must first be established through a 量 (pramāṇa) of intrinsic nature commonly accepted by both parties. Prāsaṅgikas still set up inferential forms and rigorously draw out consequences according to what the opponent already accepts. Refusing one starting point of proof is not refusing rational examination.
Misreading twelve: logical correctness is sufficient to lead to liberation.
因明 (hetuvidyā) can clear away wrong views, hesitation, and confusion in argument, but it cannot alone replace the complete practice of 戒、定、慧 (śīla, samādhi, prajñā—ethical discipline, concentration, wisdom). Being able to prove selflessness is not the same as no longer grasping at a self when insulted; being able to speak of dependent arising is not the same as body and mind ceasing to crave. The determinative understanding achieved through inference must still enter into repeated contemplative cultivation and be actualized in action that is less greedy, less angry, less deluded, and more honest and compassionate.
The deepest significance of 利他比量 (parārthānumāna), therefore, is not to teach one to "argue better," but to teach one to be responsible for one's cognition: what I assert, on what grounds do I assert it; which parts are facts and which are inferences; whether the other person can reach the same conclusion from the reasons I give; and if not, whether the problem lies with the other person, or with my reason not having been established.
When this sense of responsibility enters practice, language ceases to be merely a weapon defending the self. It can become a road shared by both parties on the way to clarity: reducing misapprehension through testable reasons, reducing hostility by not distorting the other, supporting contemplative cultivation with a view arrived at through discernment, and finally allowing the subtler attachment of "I must be right" to enter into the scrutiny of dependent arising and emptiness.