Most Buddhist concepts point us inward or upward. *Lùn fǎ* — sometimes translated loosely as "the dharma of discussion" or "the principles of discourse" — is unusual because it points at a medium we usually forget we are swimming in: *speech itself*, in all its forms. The *Xián Yáng Shèng Jiào Lùn* (顯揚聖教論, *Exposing the Sacred Doctrine Treatise*, T1602, attributed to Asanga, with commentary by Vasubandhu, translated into Chinese by Xuanzang) treats *lùn fǎ* not as a single technique but as an entire taxonomy of how we talk, what we talk about, where we talk, what makes our talk fail, and what makes it free. In an age when every pocket carries a discourse factory, the old taxonomy has surprising bite.
This reading is grounded in T1602 itself. Where the source only lists a name without defining it, I will say so. Where I offer a contemporary analogy, I will mark it as analogy, not proof.
What This Concept Is Saying
Two layers woven together
The *Xián Yáng Shèng Jiào Lùn* sets up a two-layered architecture for *lùn fǎ*:
- Seven aspects (七門) — the dimensions one must consider about any instance of discourse. - Six kinds (六種論體性) — the six basic *natures* a piece of discourse can take, set down under the first of those seven aspects.
The seven aspects are listed twice in slightly different wording, and the difference is worth seeing because the tradition itself noticed the slippage. The verse-summary (嗢柁南, *gāthā*) version is compact: *"論體、論處所、論據、論莊嚴、論負、論出離、論多所作法"* — *"the discourse-body, the place, the basis, the refinement, the loss, the escape, and the manifold functions."* The prose exposition stretches it: *論體性* (the nature/essence of discourse), *論處所* (the place/locus), *論所依* (what it relies on), *論莊嚴* (its ornament or refinement), *論墮負* (its falling-into-loss), *論出離* (its liberation/exit), *論多所作法* (its many functions).
The shifts are meaningful: - *論體* becomes *論體性* — adding the character 性 (*nature/essence*), so we are being asked about the *kind of thing* discourse is. - *論據* becomes *論所依* — what discourse *rests on* (texts, authorities, evidence, premises), not just "data." - *論負* becomes *論墮負* — adding 墮 (*falling into*), so loss is not just losing but *falling into* loss. Discourse can sink.
Only the first of these seven aspects — 論體性 — gets substantial treatment in the source. The other six are listed but not unpacked in the passages we have. I will respect that limit below and not invent the rest.
The six kinds of discourse
Within 論體性, the *Xián Yáng Shèng Jiào Lùn* names six kinds, in a definite order. They are not ranked from best to worst; some are wider, some narrower, some are pathological mirrors of each other. Walking through them in the canonical order matters, because the text cares about the distinctions.
**1. 言論 (*yán lùn*) — speech as such.** This is the broadest category: any utterance, any vocalization, any sentence. T1602 describes it as "all speaking, all vocal sound, all words." When you narrate to yourself in the car, sigh, mutter at the weather app, or whisper "I am so tired" — that is *yán lùn*. It is the universal container.
**2. 尚論 (*shàng lùn*) — favored speech.** This is a technical term in this text — the source is explicit that it should not be paraphrased into "popular discourse" or "preaching" by default. It means: *"discourse that people in the world enjoy hearing."* The criterion is the world's pleasure, not the world's truth. A motivational podcast that makes the commute feel inspiring, a viral essay that wittily skewers a public figure, a celebrity interview whose rhythm you can't stop scrolling — these are *shàng lùn*. They may or may not be *true*. Their defining feature is that the audience wants more.
**3. 諍論 (*zhēng lùn*) — disputation.** This is where the text gets specific. *Zhēng lùn* is discourse driven by conflict, and the source identifies its three trigger-zones clearly:
- **依諸欲 (*yī zhū yù*) — arising from desire.** The fuel here is wanting: things one loves, things one wants to possess, partners, audiences, status, comfort. When two coworkers compete for the same promotion, when partners argue about who has less leisure, when two creators feud over a niche — the *zhēng lùn* is *yī zhū yù*. The source emphasizes that those who have not left desire (未離欲) and are stained by the greed of the desire-realm (欲界貪染), through clinging, binding, indulging, and craving, *manifest their force in anger* (發憤現威) and create *different, oppositional, harmful disputes* (異諍、乖諍、違害諍論). - **依惡行 (*yī è xíng*) — arising from wrong conduct.** Here the fuel is the mutual exposure of mistakes — bodily and verbal wrongdoings. Modern moral pile-ons, in which people drag out each other's lapses, are textbook *yī è xíng* — heavy greed, hatred, and delusion have those people in their grip. - **依諸見 (*yī zhū jiàn*) — arising from views.** This is the most intellectual *zhēng lùn*. The source enumerates: *identity view* (身見), *annihilationist-eternalist views* (斷常), *non-causality views* (無因), *bad-cause views* (惡因), *and other heterodox views including the Ājīvika-style "rain-clouds ascetics"* (雨眾外道) "and the rest of the unwholesome views." Someone may *zhēng* either to *overturn* what the other holds, or to make them *abandon their old view and accept a new one* — sometimes one's own, which itself is unexamined.
These three are not separate at the roots — the source describes them as three *causal grounds* (所緣) of the same disputatious energy, all driven by the three unwholesome roots (貪瞋癡, the three poisons) showing up in different dress.
**4. 毀謗論 (*huǐ bàng lùn*) — defamatory discourse.** This is the easiest category to misread because it is *wider* than it sounds. The source describes two layers:
- The obvious layer: people, in mutual rage, with defiled minds, *manifest their power and defame each other* (以染污心振發威勢而互相謗毀). The speech involved can be coarse words (麁語), lies (妄語), or frivolous/ornate speech (綺語). - The less obvious layer: *systematically teaching within an ill-said dharma/law tradition* (在「惡說法律」中宣說其法，乃至加以銓量刊定、教授教誡). That is, if the doctrinal source is judged a "bad dharma tradition," then *even polished, well-organized, articulate teaching* of its contents counts as *huǐ bàng lùn*. It is not just name-calling.
The justification the text gives is the diagnostic: such speech "negates the actual nature of things, leads toward evil destinies, and merely wastes effort without benefit" (非撥實相、引向惡道，徒費功勞而無義利). Notice the three-part claim — it is *wrong about reality*, it *misdirects the speaker and listeners*, and it is a *waste of time*. All three hold at once.
**5. 順正論 (*shùn zhèng lùn*) — discourse in accord with correct dharma.** This is the structural mirror of *huǐ bàng lùn*. Same activity — *systematic exposition, definition, and instruction* (宣說其法、銓量刊定、教授教誡) of a doctrinal tradition — but performed *within* a "well-said dharma/law" tradition (善說法律). The three explicit purposes are: (a) *removing the doubts that arise in sentient beings* (斷除有情所起之疑), (b) *enabling them to penetrate the deep meaning* (善能通達甚深義), and (c) *bringing their wisdom-insight to ultimate purity* (令智見達到畢竟清淨).
This is the standard against which *huǐ bàng lùn* is judged. Same speech acts, same structure of teaching — what changes is the *doctrinal context* (善說 vs 惡說) and the *fruit* (義利 vs 無義利).
**6. 教導論 (*jiào dǎo lùn*) — instructional discourse.** The source lists this name only. It does not continue to define it. I will not invent a definition either; in walking through contemporary life I will mark the gap rather than fill it with a guess.
Where this sits on the path
*Lùn fǎ* is one slice in a much larger dharma-categorization project in T1602 — the *Exposing the Sacred Doctrine Treatise* spends much of its work cataloguing what the Buddha taught, and "the dharma of discussion" is one slice among many. Theologically, the three poisons (貪瞋癡) hover behind the entire taxonomy: *yī zhū yù* in *zhēng lùn* is greed's signature, *yī è xíng* is hatred's, *yī zhū jiàn* is delusion's. *Huǐ bàng lùn* and *shùn zhèng lùn* are framed not as neutral speech types but as *karmically consequential speech acts* — one leads toward evil destinies, the other moves toward the purification of view. The framework, then, is not sociological ("interesting ways people talk"). It is soteriological: *your speech is one of the places liberation or its opposite is being worked out, right now, in this conversation.*
Walking It Through Contemporary Life
Let's run the matrix over a single ordinary day.
6:40 a.m., kitchen table, before anyone else is up. You mutter at the coffee machine, narrate a frustration about work, whisper something tender to the cat. This is *yán lùn* — speech as such. The matrix notices: yes, you are doing the most basic thing a human can do with a voice. The matrix is not yet judging, only classifying.
7:30 a.m., bus to work, earbuds in. A podcast host tells a story about productivity with perfectly timed jokes and confident metaphors. The host is clearly fluent. You enjoy it; you almost want to share the episode. This is *shàng lùn*. The defining feature is not that it is wrong. It might even have insights. The defining feature is that *the pleasure of the form* — the rhythm, the wit, the polished cadence — is doing the work, and your wanting-to-listen-more is the criterion the text uses. The text does not say this is bad; it says this is *shàng lùn*, and that it is not the same thing as *shùn zhèng lùn*. A common mapping error is to assume something deeply enjoyed must therefore be deeply true — the matrix says they are independent axes.
10:00 a.m., office. A meeting becomes a dispute. You and a colleague both want the same budget line. Each of you begins *zhēng lùn yī zhū yù* — disputation from desire. Notice the texture the source describes: *clinging (堅執), binding (縛著), indulging (耽嗜), craving (貪愛), manifesting force in anger (發憤現威)*. By 10:45 the dispute has migrated — it is no longer about the budget line; it is about being right. Now it is *zhēng lùn yī è xíng*: each of you is starting to drag out the other's recent mistakes. By 11:15 you are no longer arguing about money but about a software philosophy: which of you has the better framework. Now it is *zhēng lùn yī zhū jiàn*. The same original impulse has rotated through all three grounds, and a competent onlooker can watch the unwholesome roots burning through the register shift. The matrix's gift here is to make it *legible to you* that one dispute is not three; one is three faces of the same fire.
1:00 p.m., lunch scroll. A well-produced channel with a charismatic host argues, calmly and with good production values, that a particular group of people is subhuman. The argument is fluent; the citations are studiously chosen; the moral framing is intricate; the production is high-budget. *This is the easy form of* huǐ bàng lùn *to miss*. The textbook image of defamation as road-rage insults — *bǐ nǐ*, an angry slur — is layer one. Layer two is *the polished exposition*, with definitions and citations, of a "bad dharma tradition." In a Buddhist diagnostic frame, what matters is not the volume or the vocabulary; what matters is the doctrinal home of the teaching and where it leads. A friend watching with you and saying "it's so well argued" is, by this taxonomy, being asked to evaluate on the wrong axis.
6:00 p.m., phone call with an old teacher. You describe the office fight, the podcast, your confusion about what to believe about yourself. They listen a long time without rushing you to a conclusion. They ask two questions that crack open the question. They offer, plainly, a way of seeing that has been tested. They are not entertaining you; you do not enjoy every moment. You leave less confused and more capable of seeing. This is the texture of *shùn zhèng lùn*: *doubt removed* (no longer paralyzed), *deep meaning penetrated* (something clicks open), *wisdom-insight moving toward purity* (you can look at what you are doing and call it by name without being flooded). The pleasure criterion that defines *shàng lùn* is conspicuously absent — *shùn zhèng lùn* is often not the most enjoyable conversation in the room.
9:00 p.m., your partner asks why you were so quiet. You try to explain the day. You are, by this point, just trying to *teach them what happened to you*. Whether this becomes *shùn zhèng lùn* or merely *shàng lùn* depends on whether what you are handing over is actually fitted to free them — or is just narrative you find pleasant to repeat. *Huǐ bàng lùn* is also in the room: if you shift into "and today at work, this *particular colleague*" with the slow whisper of contempt, you have entered layer one of defamation. The line between *yi è xíng zhēng lùn* (carrying yesterday's slight) and *huǐ bàng lùn* (defamation in intimacy) is thinner than people think.
The seven aspects woven through the same day. The day also passes through the seven *aspects* in a less explicit way: the *place* (論處所) shifts — kitchen, bus, meeting room, phone; the *basis* (論所依) shifts — anecdote, scripture, lived memory, doctrine; the *refinement* (論莊嚴) shifts — from mumbled self-talk to polished production; the *falling-into-loss* (論墮負) shows up as the moment when you realize you have been arguing for twenty minutes and have nothing to show; the *escape* (論出離) shows up when someone finally says "let's not do this today," and you feel your chest loosen; the *manifold functions* (論多所作法) is the most visible of all — the day's speech has informed, soothed, alienated, defamed, freed, and wasted your energy, all without changing its surface activity.
Why Contemporary People Need This
We are not short of speech. We are short of *classification*. Several pressures on contemporary discourse make the *lùn fǎ* map unusually useful.
Discourse saturation. Every smartphone is a discourse factory — message threads, podcasts, captions, comment sections, AI chat windows. We are exposed to more speech in a day than pre-modern people heard in a year. Without a way to *type* what we are hearing, we cannot choose what to do with it. *Yán lùn* and *shàng lùn* and *zhēng lùn* and *huǐ bàng lùn* and *shùn zhèng lùn* are flowing through us simultaneously, and a person who has only one word for "speech" cannot navigate that stream.
**The collapse of *shàng lùn* into *shùn zhèng lùn*.** Algorithmic feeds reward what we enjoy hearing; this is, by the source's definition, the criterion for *shàng lùn*. But listeners, primed by years of conflating "well-produced" with "well-argued," often treat *shàng lùn* as if it were *shùn zhèng lùn*. The distinction is not that the favored speech is shallow; some of it is deep. The distinction is that the *criterion of acceptance* is audience enjoyment, not removal of doubt and purification of insight. When we forget this, we mistake charisma for liberation.
**The polish of *huǐ bàng lùn*.** The second layer of defamation — *organized exposition within an unwholesome doctrinal tradition* — has become more common, not less. Hate-ideology now has professional voice actors, slides, citations, even peer-reviewed-looking papers. The matrix says: the level of polish is not what matters; the doctrinal home and the fruit are what matter. This is uncomfortably practical advice for evaluating online content.
**The three causes of *zhēng lùn* as a self-knowledge tool.** Most contemporary arguments are diagnosed as "the issue" when in fact they are running on *yī zhū yù* / *yī è xíng* / *yī zhū jiàn* — and sometimes all three. The source gives a remarkably clean diagnostic for noticing, in the middle of one's own heat, *which* fuel is feeding the fire. You are arguing about money but you can't let go — that is *yī zhū yù*. You are dragging out their old mistake — that is *yī è xíng*. You are defending your worldview and cannot let it bend — that is *yī zhū jiàn*. They are not three separate fights; they are three registers of one dispute.
Where this touches AI-era experience, with limits. As a *loose analogy*: large language models are trained to predict the next pleasing token — which is, in form, structurally close to *shàng lùn* machinery optimized (an analogy, not a metaphysical claim). Conversational AI is, in the *lùn fǎ* sense, a *shàng lùn* engine with no inbuilt way to move toward *shùn zhèng lùn*. A user who wants more than favored speech has to remember this. A user who wants liberation has to remember this. The analogy is useful; it is not a doctrinal claim about AI, and it does not authorize any inference about the inner life of these systems (which is, scientifically, an open question with no settled answer).
Common Misreadings and Clarifications
Misreading 1: "論法 just means debate." No. The *Xián Yáng Shèng Jiào Lùn* explicitly says it is a complete classification covering the *nature, place, basis, refinement, loss, escape, and manifold functions* of discourse. Even the six sub-types include *yán lùn* and *shàng lùn* and *jiào dǎo lùn* — none of which are debate. Treating *lùn fǎ* as a debate manual would be like treating the Periodic Table as a recipe book.
Misreading 2: "尚論 is bad speech." Not at all. *Shàng lùn* is, by definition, *speech people enjoy hearing*. The text does not say it is always wrong or always corrupting. It says it is *not the same thing as* *shùn zhèng lùn*, because the criterion for acceptance is pleasure, not liberation. A podcast can be *shàng lùn* and still helpful. It is just not what it sounds like if you mistake the pleasure-of-form for the penetration-of-meaning.
Misreading 3: "毀謗論 is mainly insults." Half true. The text clearly *includes* angry insult (the coarse-words / lies / ornate-speech layer), but it is equally explicit that *exposition within an ill-said dharma tradition* also counts. This is the half most readers under-apply. The structural mirror — *shùn zhèng lùn* — is defined by *the same activity of exposition*, just within a well-said dharma tradition and toward liberating fruit. Therefore: any system that allows you to recognize polished defamation *as defamation* needs to know what counts as an "ill-said dharma tradition" and how that judgment is made. The source takes that judgment seriously; it is not something a casual reader can perform after one sitting.
Misreading 4: "諍論 is the same as having a strong view." No. *Zhēng lùn* is discourse *driven by conflict*, with the three poisons behind it. Holding a position firmly is not, on its own, *zhēng lùn*. You can state your position cleanly, listen to another, change your mind, or hold your ground — and none of these, on their own, are *zhēng lùn* in the source's sense. *Zhēng lùn* arises when *clinging, binding, indulging, craving, and anger-into-power* (堅執、縛著、耽嗜、貪愛、發憤現威) are running.
Misreading 5: "教導論 means teaching in general." The source, in our passage, lists the term *教導論* only, without defining it. We should not, on the basis of the name alone, fill in a definition. It may overlap with parts of *shùn zhèng lùn*, or be its own distinct category — the text we have does not tell us. Honesty about the gap is the right move.
Misreading 6: "The seven aspects are all fully defined in the text." Only 論體性 gets substantial treatment in our passage. The other six (論處所, 論所依, 論莊嚴, 論墮負, 論出離, 論多所作法) are named but not unpacked in the source we have. We can sketch what they might point at based on the *names* — the text says *ya*, but does not, in our passage, say *what*. Anyone offering a "full account" of these six is reconstructing, not quoting.
Misreading 7: "順正論 means 'a Buddhist arguing.'" Not quite. *Shùn zhèng lùn* is defined by the *dharma tradition the discourse lives within* and the *fruit it bears for the listener*. A Buddhist teacher whose words are incoherent, or who performs within a tradition the text would classify as 惡說法律, is not, by this taxonomy, automatically *shùn zhèng lùn* — and a non-Buddhist teacher whose words genuinely *remove doubt, penetrate deep meaning, and move insight toward purity* may, by the *fruit criterion*, count. The matrix judges speech by its results, not by its badge.
A Closing Frame
The *lùn fǎ* of T1602 is not telling you to be quieter. It is not telling you to argue better. It is telling you something more austere: that *the kind of speech you make is one of the places your liberation — or your entanglement — is being worked out, in real time, in every conversation.* The six kinds are not labels to paste on others; they are mirrors. In a day dense with speech, the matrix asks each of us, again and again, what kind we are making right now, and whether we can tell.