A reader's note before we begin. In everyday Buddhist speech, 法燈 (fǎdēng, dharma-pradīpa) is a familiar image: the teaching as a lamp that dispels the darkness of ignorance, the *avidyā* (無明) that obscures how things are. You find this image in hymns, in dhāraṇī, in the standard closing line of a *sūtra* recitation: *"願以此功德，莊嚴佛淨土"* — and the implicit metaphor that the dharma itself is a lamp lit in the mind. The picture is warm, settled, devotional.
What the source material you've supplied gives us is almost the opposite of that picture. In the two foundational Mādhyamaka texts cited — the 《中論》 (*Mūlamadhyamakakārikā*, MMK, attributed to Nāgārjuna, translated by Kumārajīva) and the 《十二門論》 (*Dvādaśadvāraśāstra*, "Treatise of Twelve Gates," likewise attributed to Nāgārjuna) — the compound 法燈 does not appear at all. No *dharma-lamp*, no *wisdom-lamp* (慧燈), no *bright lamp* (明燈). What appears is the bare word 燈 (lamp), and it does not shine reassuringly for the reader. It is brought into the courtroom by an opponent, used as a piece of evidence, and dismantled.
The dismantling is the teaching. To honor the source material, I will spend most of this piece on what these two texts actually do with the lamp — namely, treat it as a contested simile in a refutation of svabhāva (自性, "own-being" / "intrinsic nature") — and only afterward indicate where the more familiar positive image of 法燈 does its work in the wider Buddhist tradition, so that readers do not collapse the two senses together.
1. What this concept is actually about
The two Mādhyamaka texts situate 燈 inside one of the most consequential arguments in Buddhist philosophy: the analysis of 生 (ś *生* / *utpāda*, arising). The opponent in MMK's "examination of the three characteristics" (觀三相) chapter wants to establish that arising is real — that things genuinely come into being, with their own nature — and offers the following analogy:
"如燈能自照，亦能照於彼；生法亦如是，自生亦生彼。" *Just as a lamp can illuminate itself and illuminate others, so too an arising dharma can self-arise and give rise to others.*
The grammar of this analogy matters. The opponent is *not* saying the lamp is dharma. The opponent is saying: *the lamp has an obvious, intuitive, self-evident efficacy — it self-shines and shines on what is near it. Arising dharmas must work the same way, with the same obvious self-evident efficacy. That efficacy is what we mean by svabhāva.* The lamp is offered as a felt guarantee that "arising" is not an empty word.
Nāgārjuna's response is to take up that guarantee and refuse to cash it. He does not say "lamps don't work." He says: let's see what "working" consists of in this case. And then he walks through four moves. I will preserve each one with its classical language, because compressing them loses the structure that makes the argument a *prasaṅga* (歸謬, *reductio ad absurdum*) rather than a denial of ordinary experience.
Layer 1 — The definition of 照 (illumination) collapses
"燈中自無闇，住處亦無闇，破闇乃名照，無闇則無照。" *Within the lamp itself there is no darkness; in the place where the lamp is, there is also no darkness. To break darkness is what we call illumination; with no darkness, there is no illumination.*
Key technical terms, kept intact:
- 照 (*prabhā* / "illumination") — here strictly defined as 破闇, the *breaking of darkness*. - 明闇相違 (*prabhā-tamasor virodha*) — light and darkness are mutually exclusive; they cannot occupy the same locus at the same time.
The force of the move: if *照* is a verb, it has an object. The only legitimate object, given the definition, is darkness. But the lamp, by its very nature, excludes darkness from itself and from where it is — that is precisely what it means to be a lamp. So the lamp has nothing to illuminate. You can still call it a lamp; you cannot call it *self-illuminating* on this definition. The "self-evident efficacy" of the lamp, when pinned down, vanishes.
The 《十二門論》 makes the same point more bluntly: *"燈為何所照？"* — "what does the lamp illuminate, exactly?". The question is rhetorical. The answer is: nothing it can name without first locating some darkness to break.
Layer 2 — The retreat to the moment of arising
The opponent is not stupid; they shift ground. "We do not claim that an unlit lamp illuminates, or a fully lit lamp continues to illuminate. We claim the lamp *illuminates at the moment it is arising*, when it is half-formed, half-unformed." That is the 生時 (the moment of origination) retreat.
Nāgārjuna closes that retreat in two steps:
- The "moment of arising" is *半生半未生* — half arisen, half not yet arisen. At that point the lamp-body is not yet complete; how can something incomplete do the work of breaking darkness? - Even granting that moment, the lamp has not yet *reached* the darkness. The 《十二門論》 states this most clearly: *"燈初然時不能及於闇"* — at the moment of first lighting, the lamp cannot yet reach the darkness. The classical simile, from MMK, is *"如人得賊乃名為破"* — as one catches a thief, one is said to "break" the thief; reaching is the precondition for breaking. Translation: you cannot break what you have not reached.
Key term preserved: 到 (*prāpta*, "reached / arrived at") as the necessary condition for 破 (*bhaṅga*, "breaking").
Layer 3 — Reductio to absurdity: the lamp that breaks everything
If you grant, against the previous layer, that the lamp somehow breaks darkness *without reaching it*, then the consequence is universal:
"燈若未及闇，而能破闇者，燈在於此間，則破一切闇。" *If a lamp, not having reached the darkness, can still break the darkness, then a lamp here would break darkness everywhere.*
The justification, kept in the source's own wording: 俱不及故 — "because it equally does not reach". By the opponent's own premise, the lamp is not in contact with any darkness; therefore, on their own terms, it should work on all of them equally. Empirically, of course, it doesn't. The lamp on your desk does not break the darkness in another room, let alone in another country. 《十二門論》 underscores this with the empirical observation: 此間之燈實不能破一切世間闇 — "a lamp here in fact cannot break all the world's darkness".
This is the prasaṅga move in its cleanest form: take the opponent's premise, run it outward, and let it collide with what anyone can observe.
Layer 4 — Symmetric reductio: the darkness that illuminates itself
The final move turns the analogy against itself:
"若燈能自照，亦能照於彼；闇亦應自闇，亦能闇於彼。" *If a lamp can illuminate itself and illuminate others because it stands opposed to darkness, then darkness too, standing opposed to the lamp, should darken itself and darken others.*
The 《十二門論》 phrases this with the parallel verbs 自照／照彼 ↔ 自蔽／蔽彼. The conclusion in both texts is the same, with slightly different wording: 是故燈喻非也 / 汝喻非也 — "therefore the lamp-simile is invalid" / "your simile is invalid".
Key term preserved: 相違 (*virodha*, opposition) — opposition to an opposite is not, by itself, proof that one has efficacy. If you grant efficacy to one side of an opposition, symmetry demands you grant it to the other; if you cannot grant it to the other, you cannot grant it to the first. The lamp's felt obviousness does not survive the symmetry test.
What is being proved, and what is not
It is important to register what has *not* been said. Nāgārjuna has not denied that lamps light rooms. He has not denied that arising happens. He has shown that *neither the felt obviousness of the lamp's illuminating power nor the felt obviousness of arising's "self-and-other-producing" power can do the philosophical work the opponent wanted them to do.* The lamp was offered as the intuition-pump that would secure *svabhāva* for arising. The intuition-pump fails. So does what it was pumping for.
This is the meaning of the careful phrase in the source material: 燈在此純為論敵之喻例 — "the lamp here is purely a simile used by the opponent." It is *所破* (that-which-is-refuted), not *所立* (that-which-is-established). If a reader carries away the warm image of 法燈 as dharma-light and applies it to these texts, they have reversed the direction of the argument.
2. Walking the concept through contemporary life
The four-layer structure is not an ivory-tower exercise. Each layer maps onto a familiar modern situation, and the mapping is exact — not metaphor. Let us walk through one.
Scenario: You have just downloaded a meditation app. The marketing says, *"This app will teach you to focus. Studies show that ten minutes a day rewires your brain."* You are now the opponent in the lamp analogy. The app is your lamp. The darkness is your scattered, anxious mind.
Layer 1 — Where is the darkness? You sit down. You follow the breath for thirty seconds. You notice, already, that your mind is calmer. But — Layer 1 — *the calm was already arriving*. The act of sitting, of stopping, of closing the eyes, had begun to do its work before the app's voice even spoke. By the strict definition the MMK is using — illumination as "breaking darkness" — the app had no darkness left to break; the quiet had already arrived through other causes. *Where, exactly, did the app illuminate?* This is the 燈為何所照 question.
A more honest account: the app *accompanied* a process that was already happening. Saying it "did" the calming, in the sense of *being the agent that broke the darkness*, is to give it a self-nature it does not strictly possess.
Layer 2 — At what moment did it "teach you to focus"? The app's claim is that it *teaches*. Teaching, like the lamp's illumination, requires reaching. But when did it reach you? Was it the moment you opened it (too early — the lamp-body not yet complete)? Was it the moment you closed it (too late — the teaching already done)? Was it the in-between — the *half-formed half-not-formed* moment of starting? Try to point to it. The finger cannot land. And yet the felt sense that "the app taught me" persists. That felt sense is what the lamp analogy is structurally about.
Layer 3 — If it teaches without reaching, it teaches everyone. Take the opponent's premise seriously. The app "teaches focus" by a mechanism that does not require actual contact with your specific mind. If so, by the same premise, it should teach focus to anyone who owns it — every subscriber, every download, every person who has ever heard its name. It does not. Its efficacy is empirically uneven, contingent on conditions (your prior state, your willingness, the day's cortisol level, whether you remembered to use it). The 俱不及故 "equally-not-reaching" condition is satisfied on the opponent's own terms, and the absurd consequence follows: the app works for everyone or for no one in the strict sense.
Layer 4 — The symmetric reductio. The app "teaches focus" because it stands opposed to unfocus. By the symmetric logic, anything that stands opposed to unfocus should also "teach focus": a piece of paper with the word *focus* on it; a friend's casual remark; a near-miss car accident that makes you sharply present for ten seconds. If none of these *strictly* teach focus in the same way the app does, then the app's claim to teach focus — in the strong, svabhāva sense — is also compromised. What is left is a softer, truer claim: the app is one of many conditions that sometimes contribute to focus, none of which possesses the kind of standalone efficacy the marketing implies.
The reader should notice: I have *not* said "the app does nothing." I have said: the felt obviousness of "the app teaches focus" cannot bear the philosophical weight the marketing wants it to bear. The MMK is not anti-lamp; it is anti-the-strong-claim-about-the-lamp.
The mapping works for other scenarios too:
- Medication: "This pill cures depression." Layer 1: the darkness (depression) was already being modulated by sleep, food, social contact. Layer 2: at what millisecond does it "cure"? Layer 3: if it cures without reaching specific neural pathways, it cures everyone who takes it. Layer 4: by symmetry, anything opposed to depression (a phone call from a friend, sunlight, exercise) should "cure" in the same strong sense — and the fact that we do not grant that shows the strong sense was always an overreach. - Therapy / coaching: "This method will change your life." Same four layers. Especially Layer 2 — the moment of change is unlocatable, half-formed, half-unformed, exactly as Nāgārjuna describes. - A relationship that "saved" you. Layer 1: you were already saving yourself in ways you cannot fully name. Layer 4: by symmetry, the things that hurt you (the losses, the failures) were also "opposed" to your flourishing and did not strictly save you — so the saving cannot be assigned to the relationship in the strong sense either.
Where people most often mis-map
The single most common mis-mapping is to take Layer 1 as a denial: *"See, the lamp doesn't really work, just like the app doesn't really work."* That is not the move. The move is: the lamp's working *cannot be cashed out* in the way the opponent needs it to be cashed out in order to secure the conclusion (arising has svabhāva). The lamp still lights the room; the app still sometimes helps. What is lost is the strong metaphysical claim of *intrinsic, self-standing efficacy* — not the everyday observation of function. Conflating these two is the dominant contemporary misreading of this whole line of argument.
A close second: skipping Layer 4. Many readers find Layers 1–3 clever but resist Layer 4 — *"That's not fair, darkness and light aren't really symmetric."* The MMK's claim is precisely that *as far as the argument for svabhāva goes, they are symmetric.* If you let one side of an opposition carry the weight, symmetry forces you to let the other carry it too; if the other cannot, neither can the first. The intuition that "of course light and dark aren't symmetric" smuggles in precisely the everyday familiarity that the argument is putting in question.
3. Why contemporary people need this analysis
The cultural air we breathe is dense with implicit svabhāva-claims. Every advertisement, every productivity guru, every "this changed my life" testimony, every algorithm that promises to *know you* — all rest on a structure of the form: *X works. X works self-evidently. X works because of X's own nature. Therefore X is worth your money, your time, your trust.* This is the lamp analogy generalized. The MMK is not the first text to notice the structure, but it is the first to give it a careful four-layer diagnosis, and the diagnosis has not been improved upon.
Three contemporary reasons to engage it seriously:
First, because the felt obviousness of efficacy is being weaponized. Modern attention economies depend on the intuition that *this thing works* — this supplement, this method, this course, this candidate. The MMK is not saying these things fail; it is saying that the *felt obviousness* with which we attribute efficacy to them is structurally the same felt obviousness we have about the lamp. And that felt obviousness, when examined, cannot deliver what the strong claim wants it to deliver.
Second, because the symmetric reductio (Layer 4) is a tool against the moral form of the same error. We routinely say: *"This practice is good because it opposes bad."* Meditation is good because it opposes agitation. Discipline is good because it opposes distraction. Productivity is good because it opposes waste. By Layer 4's logic, agitation, distraction, waste — whatever opposes the good thing — should also possess the same strong "good-making" efficacy by virtue of opposition. They don't. So the strong claim — that goodness-inheres-in-things-that-oppose-badness — is not available. We are left with a softer, more accurate account: *these are conditions that, in context, sometimes contribute*, and so are their opposites.
Third, because the unlocatability of the moment of change (Layer 2) bears directly on how we evaluate self-transformation. Modern therapeutic culture is full of *before-and-after* stories with a clean break-point: "I started this practice, and three weeks later, I was a different person." The MMK's *half-formed, half-unformed* critique says: the moment of change is unlocatable; the conditions were already in motion; the felt cause is doing less work than the felt sense of causation claims. This is not discouraging. It is *realigning*. It lowers the metaphysical stake of the "thing that changed me," which in turn lowers the metaphysical stake of losing it, and lowers the credit we assign to whoever sold it to us.
A note of methodological honesty: I am not claiming neuroscience "proves" Mādhyamika analysis. The structural resemblance between the lamp-simile analysis and what contemporary cognitive science says about *causal attribution*, *causal overestimation*, and *the illusion of explanatory depth* is real — but resemblance is not proof. The MMK was written long before these empirical findings, by methods that are not empirical. What is fair to say is that the MMK's analysis is *structurally congruent* with what modern research has independently discovered about how humans misattribute efficacy. Whether this congruence reflects a deep feature of how things are, or a deep feature of how minds tend to mis-see how things are, is itself a question Mādhyamika philosophy refuses to settle in advance.
4. Common misreadings and clarifications
Misreading 1 — "法燈 means the dharma is a lamp that lights the mind of ignorance." This is the dominant popular sense, and it is well-attested elsewhere in the Buddhist tradition — in hymns, in *Avataṃsaka*-influenced literature, in the familiar image of 慧燈 (*prajñā-pradīpa*, wisdom-lamp) or 法燈 (*dharma-pradīpa*, dharma-lamp). It is *not* the sense operative in MMK and the *Dvādaśadvāraśāstra*. In those two texts, the compound does not appear, and the bare 燈 is the *opponent's* piece of evidence, not the *论主's* symbol for dharma. When reading or citing these two texts, do not import the popular sense; you will reverse the direction of every argument.
Misreading 2 — "Nāgārjuna is denying that lamps work." No. He is analyzing the conditions under which the claim *the lamp illuminates itself and illuminates others* can be cashed out. The everyday observation that lamps light rooms is untouched; what is undermined is the strong claim that this observation proves the lamp has svabhāva-like efficacy, and that such efficacy in the lamp grounds svabhāva in arising.
Misreading 3 — "The four arguments are independent." They are not. They form a cascading closure. Layer 1 closes off the fully-lit and fully-unlit cases. Layer 2 closes off the *moment-of-arising* retreat. Layer 3 shows that the only escape from Layer 2 produces universal (and empirically false) efficacy. Layer 4 closes off the structural move (opposition-implies-efficacy) that the entire simile depends on. By the time Layer 4 lands, the opponent has no remaining foothold. Treat them as one move with four hinges, not as four separate points.
Misreading 4 — "Layer 4 is just wordplay." The symmetric reductio is doing serious philosophical work. It targets the inference from *x is opposed to y, therefore x acts on itself and on y* — an inference we make constantly. Symmetry is a constraint on inferences about oppositions: if the inference form licenses x-efficacy, it licenses y-efficacy under the same conditions. That the form licenses both but reality (we think) licenses only one is a sign that the inference form was never delivering what it appeared to deliver.
Misreading 5 — "If the lamp-simile fails, all similes fail." No. Nāgārjuna is *not* mounting a general attack on analogy. He is showing that this specific analogy fails on its own terms, and that its failure removes the felt-evidence backing for svabhāva in arising. Other analogies (or the same analogy used carefully in other contexts, as it is elsewhere in Buddhist literature as a positive image of dharma) may do different work in different arguments. The lesson is local: do not let felt obviousness do metaphysical labor it cannot bear.
Coda: where the lamp does light
Because I have spent this entire piece on the *negative* use of the lamp in MMK and the *Dvādaśadvāraśāstra*, it would be misleading to end without acknowledging that the *positive* image — dharma as lamp, wisdom as lamp, the teaching as that which dispels the darkness of ignorance — is genuinely important in other parts of the Buddhist corpus. It is not present in these two texts, but it is present elsewhere, and readers should hold both senses in mind:
- The lamp as *contested simile*, an intuition-pump that fails when analyzed — the sense of the source material you provided. - The lamp as *positive symbol* of dharma/wisdom dispelling ignorance — the sense embedded in much devotional and commentarial literature, and the sense most readers will encounter first.
Both senses are part of the Buddhist tradition's long, careful, and at times self-critical reflection on what it means for a teaching to *illuminate*. The Mādhyamika contribution, in the texts you have asked me to read, is to refuse to let the image do its work unexamined — to insist that even the most obvious-seeming lamp, when held up to the light of analysis, shows that its light cannot be cashed out the way the obvious-seeming claim wants it cashed out. That refusal is, in its own way, also a kind of lamp: not the lamp of the lamp, but the lamp that sees the lamp.