One-sentence definition. 無因作 (ahetukakriyā, "creation-without-cause") is the claim that something — a person, a deed, an event, a thing — can arise or come into being without any 因緣 (pratītya, conditions) whatsoever; 《中論》 (Mūlamadhyamakakārikā, *Fundamental Verses on the Middle Way*, attributed to 龍樹菩薩 Nāgārjuna, T1564, with Piṅgala's *ṭīkā* rendered by Kumārajīva) treats it as a foundational 邪見 (wrong view) that, if conceded, annihilates 因果 (causality), 業報 (karmic retribution), and 涅槃 (nirvāṇa) in a single stroke.
一、這個概念在說什麼 (What This Concept Is Saying)
1.1 What is actually being denied
The chapter on *The Refutation of the Agent and the Action* (破作作者品) of 《中論》 is in dialogue with four theories of how deeds and events come to be: 自作 (svayaṃkṛta, "self-made"), 他作 (parakṛta, "other-made"), 共作 ("jointly-made"), and 無因作 ("made without any cause"). Nāgārjuna's dialectical move is cumulative: each prior theory is dismantled in turn, and 無因作 is held up last as the most implausible survivor of the four.
The precise claim of 無因作 is bold and unqualified: some things — or all things — come into being with no conditions whatever. Not "we can't trace the conditions," but "there literally are none." The chapter's later verse (corresponding to the chapter summary in chunk:20483) describes such a view as the assertion that 萬物 (the myriad things) "naturally arise without depending on causes." This is the assertion that gets dismantled.
But why bother refuting a claim so clearly counterintuitive? Because Nāgārjuna knows that 無因作 is not just a metaphysical curiosity. It is the structural endpoint of certain ways of thinking about agency. Concede the right premises — that there is a fixed 作者 (kāraka, agent) somewhere, that there is a fixed 作業 (kriyā, action) somewhere — and the conclusion drags itself out by logic whether you wanted it or not.
1.2 The three-term anatomy: 作者, 作業, 作法
Before the five refutations can land, the passage enforces three distinctions that contemporary readers routinely collapse:
- 作者 (kāraka) — the *agent*. The "doer." The one who, in ordinary speech, "did it." In Buddhist analysis, this is a *designated* (*prajñapti-upādāya*) reference, not an inner metaphysical essence. - 作業 (kriyā) — the *action*, the deed as event. Distinct from the doer: a punch and the person throwing it are not the same referent. - 作法 (kṛta) — the *result* of the action, the deed's *fruit* or "thing-done-ness." When the action is harmful, the 作法 is harm's consequence; when skilful, the 作法 is the corresponding ripening.
Most English translations shanghai these three into "doer/deed/done-ness" so quickly that the asymmetry disappears. The 《中論》 keeps them deliberately separate precisely because the 無因作 mistake amounts to smuggling a fixed 作者 into one corner of the picture and a fixed 作業 into another, and then pretending causation still survives across the gap.
1.3 The five layers of refutation
The chapter hands us a tightly structured five-step dismantling. I'll walk each layer with its reasoning intact.
Layer 1 — Without cause, there can be no result. If something has no cause and yet exists, it must be 常 (eternal, unchanging), because everything that begins also ends — that is, what has no beginning has no ending. Nāgārjuna then turns the premise into an ethical absurdity. If no-cause-can-produce-result were true, then: > 布施持戒應墮地獄、十惡五逆應當生天 (Giving and keeping precepts would land you in hell; the five gravest evils and patricide would land you in heaven).
This violates both 世間現量 (direct worldly perception — what any of us can observe about how actions bear fruit) and 聖教量 (reliable teaching — what the Buddha affirmed). The closing verse of this layer (chunk:19956) is clean: 「若離色有因，則是無果因；若言無果因，則無有是處」 — "If you sever a 'cause-without-effect' from form, you have a cause that produces nothing; such a thing has no locus at all."
The metaphor here is implicit in the structure: 無因作 is like trying to drink from a cup that has no bottom — you can point at the cup, you can call it a cup, but the cup-as-cup-to-drink-from never shows up. We'll come back to this concretely in section two.
Layer 2 — "Causeless dharmas" are speech and nothing more. Nāgārjuna's objector tries to escape with a venerable counter-example: "You Buddhists already accept causeless existents! The three 無為 (asaṃskṛta, unconditioned) — 虛空 (ākāśa, space), 擇滅 (pratisaṃkhyā-nirodha, deliberate cessation), and 非擇滅 (apratisaṃkhyā-nirodha, non-deliberate cessation) — have no cause. The non-Buddhist theories of 神 (purusa/ātman), 時 (kāla, time), 方 (diś, direction), 微塵 (paramāṇu, atoms), and their brand of nirvāṇa are also called causeless. So you cannot simply reject 無因作."
Nāgārjuna's reply is brusque: 「此無因法但有言說，思惟分別則皆無」 — these "causeless dharmas" exist only in 言說 (speech/linguistic designation); under analysis, none of them holds up.
He then distinguishes two kinds of 因 that the objector was trying to use: - 作因 (operative cause) — a condition that actually does something, that genuinely produces. - 言說因 (nominal or locutionary cause) — a "cause" that only operates by being spoken of.
The objector retreats to: "Fine — the unconditioned has no operative cause, but it at least has a *nominal* cause, since we can name it." Nāgārjuna rejects this rescue too. 虛空 (space), the most elementary of the candidates, was already dismantled earlier in 《中論》 under what the tradition calls 六種之破 ("the six-fold refutation"); if even an object of ordinary perception cannot be established, then unseeable reifications like atoms and the self certainly cannot. (chunk:19955): 「現事尚皆可破，何況微塵等不可見法」 — "Even visible things can be refuted; how much more so things like atoms that are not visible at all."
The metaphysical lesson: if something is to claim the title "unconditioned," it must first establish that it *exists*, and existence itself is established only through the whole conditioned web. Strip away 因緣 and you strip away the very thing that lets you say "this exists."
Layer 3 — Fix the agent and the action, and you have already fallen into 無因. This is the most penetrating layer and the one most easily missed. The argument runs:
Suppose you posit that there is definitely a 作者 (an agent with fixed nature) and definitely a 作業 (an action with fixed nature). The moment you grant these as fixed existences, the agent is separable from the action and the action is separable from the agent. Two independent existences cannot produce each other. The agent doesn't generate the action; the action doesn't depend on the agent. They are co-present but causally inert.
Verse (chunk:20081): 「若定有作者，亦定有作業，作者及作業，即墮於無因」 — "If you fix the author as existent, and fix the action as existent, then the author-action pair collapses into causelessness."
The commentary (chunk:20083) draws out the wreckage: 「無因無果，無作無作者，無所用作法」 — no cause, no result, no action, no agent, no deed-as-result. Then the devastating continuation: 「罪福等無故，罪福報亦無。若無罪福報，亦無有涅槃，諸可有所作，皆空無有果」 — "Since moral good and bad dissolve, moral retribution dissolves. With no retribution, there is no nirvāṇa. Everything one might do turns out fruitless."
This is the move that turns 無因作 from a strange philosophical position into an existential catastrophe. Once you reify "I did it" as a metaphysical fact instead of a useful conventional designation, the deed empties itself. The very determinateness of the doer you meant to protect is what kills the causal bite of what the doer did.
A subtle but crucial corollary worth flagging: *the argument here applies to you whether or not you call yourself a "no-cause" person.* Many Western readers imagine that 因緣 view is one philosophical option and 無因作 is another, opposite option. Nāgārjuna is showing that the deeper error is reification — pinning down the agent or action as a fixed thing — and that error *structurally* produces 無因作 whether or not you endorse it verbally.
Layer 4 — Neither a determined nor an undetermined author can act. With the third layer's conclusion on the table, Nāgārjuna considers the four ways a thinker might try to save the agent: - 決定 (definitely determined as a fixed self), - 不決定 (definitely not a fixed self), - 亦定亦不定 (both at once), - 非定非不定 (neither).
In every case, 「不能作於業」 (the agent cannot perform the action). The metaphor given here is sharp enough to preserve word for word: 「如化人以虛空為舍」 — "Like an apparition treating empty space as a house." An apparition has no real existence; the "house" it inhabits is also empty; what you see is a scene in a conjurer's show with no actual resident and no actual dwelling. So too for an author who is not solid in nature: deeds with no solid consequences. All that is left is 言說 (speech, nominal designation), and 言說 is precisely what was separated from 作因 (operative cause) in layer 2.
Layer 5 — The cascading damage (殃累) of 無因作. Final tally (chunk:20483): 「即為破因果，作作者作法，亦復壞一切，萬物之生滅」 — "It destroys cause and effect, the doer and the deed and the deed's fruit; it destroys everything, the arising-and-ceasing of the myriad things." Nothing in the Buddhist picture of practice survives if 無因作 is conceded.
The chapter closing (chunk:20183) recapitulates the cumulative structure: 「若此彼苦成，應有共作苦；此彼尚無作，何況無因作」 — "If *this*苦 and *that*苦 are constituted (as self-made and other-made), there should be a jointly-made苦; but even *this*-doing and *that*-doing are not established, how much less causeless-doing."
So the order of operation in 《中論》 is important: refutation of the author-and-action apparatus must precede any serious dismantling of 無因作, because the latter is nothing more than the former pushed to its logical endpoint.
1.4 Where 無因作 fits in the wider map
To place this concept, three coordinates matter:
- Versus 因緣生 (pratītyasamutpāda, dependent origination). 因緣生 is *not* the opposite of 無因作 by being rigidly deterministic. It is the recognition that things arise *from conditions* while *those conditions themselves* are conditioned, interpenetrating, and empty of fixed self-nature. 因緣生 is the middle that escapes both the trap of "self-made fixed agent" (自作) and the trap of "no structure at all" (無因作). - Versus 無為 (the unconditioned). The three 無為 (space, deliberate cessation, non-deliberate cessation) are conventionally spoken of as "unconditioned," but as layer 2 argued, this is 言說 (naming convention), not a structural exception to 因緣. They are called *unconditioned* precisely because conventional analysis finds no *productive* conditions for them — but that is not the same as 無因作's claim that something exists with literally no conditions. - Versus 因果 in the popular sense. 因果 is not fate (命運), not karma-as-destiny (宿命的業), not mechanical law. It is the relational intelligibility of how actions shape their own continuations, which is exactly why reifying the doer-event-fruit triplet into hard metaphysical lumps is what produces 無因作 in the first place.
二、生活裡的走查 (A Walkthrough in Contemporary Life)
The classical definition is rigorous, but 無因作 is rarely argued in the abstract by contemporary people. It sneaks in as the silent assumption beneath certain ways of talking. Here are five scenes in which you can feel it at work, each mapped carefully to the classical categories.
2.1 The promotion that "just happened"
You wanted the role. You applied. The panel chose someone else. A friend texts: "Well, that's just how things go." Notice the wording — "just," "things," "go." Three 言說 (nominations) doing all the causal work.
Mapping to 《中論》: The surface speech-acts "it just happened" function here as the 言說因 from layer 2 — a "cause" that has no 作因 (operative content). The promotion was obviously conditioned: by who knew whom, by how the panel was composed, by your past choices and the timing of your application, by the state of the budget that quarter, by the prior decisions that made certain people visible and others invisible. None of these is hidden from analysis; all are *investible* conditions. The trap: If you accept "just happened" as the final word, the agent-and-action apparatus evaporates. The 作者 (decision-makers) and the 作業 (their choices) become unfixable; the 作法 (where you end up) loses its intelligibility. You cannot even formulate a sensible strategy for next time, because "next time" became causally opaque. The 因緣 fix: Replace "just happened" with "happened *given these conditions*." The event didn't pop out of nothing; it grew out of a web. Naming the conditions doesn't lock the web into a metaphysical eternity — it just lets you see that the next event is being woven right now.
Common mapping error to avoid: This is *not* the same as saying "you're fated never to get the role." 因緣生 is not 命運. Nothing in 《中論》 says the web is fixed. It says only that the web is what is happening, and that denial of the web either gives you a frozen self-causation (自作) or nothing-at-all (無因作).
2.2 The breakup that "just fell apart"
You were together for years. Something shifted — or seemed to. Now neither of you can say what happened. The story that gets told, in private and on social media, runs along the lines: "We just grew apart." "It ran its course." "Sometimes things just don't work out."
Mapping to 《中論》: Three 言說 doing the work of three 因 (causes). "Just grew apart" is 言說 where 作因 should be (layer 2). The story is functionally the breakup-equivalent of "no cause made this happen." Layer 3 is even more clearly in play: a fixed "we" — the relatio nship as a metaphysical thing — dissolves into causelessness. If "we" is a fixed substance, then when it breaks there is no longer anything that did the breaking; it just "happened." But "we" was never a fixed substance; it was a network of habits, attentions, small repairs, and retracted attentions. Those are exactly the conditions that layer 1 says must be investigated if anything is to be intelligible at all. The 因緣 fix: Refuse the 言說 of "just grew apart" and locate at least one actual operative condition that was withdrawn. The condition may be trivial — a habitual conversation, an assumption, a deferred repair. Naming it does not bring the relationship back, but it returns you to a world where your life is made of conditions you can see, instead of a world where your life is made of mysterious voids.
2.3 The viral post and "I got lucky"
A creator posts a piece of work. It travels. A reporter asks how. The creator shrugs: "Right place, right time. Got lucky." The audience reads this as humility and is charmed. But notice the fine print: luck is functioning here as an unnamed-cause.
Mapping to 《中論》: Luck, in this usage, is 言說 without 作因 — a "cause" that is just a placeholder for uninvestigated conditions. The actual conditions are visible to anyone willing to look: years of prior practice, the platform's recommendation dynamics on that day, the network a prior post had built, the topic's cultural temperature, the time of day the post landed. Every one of these is a 因 (condition) and the absence of naming any of them turns "luck" into a virtual 無因. The trap: If this vocabulary travels into how the creator funds a career, decisions become superstitious. "Do more, hope for luck." Operative conditions — target audiences, distribution choices, formats, follow-ups — get starved of attention because they are not in the 言說. The 因緣 fix: The opposite is not "deny luck" but "treat luck as a placeholder for the conditions you haven't investigated yet." This is the practical discipline that corresponds to layer 2's distinction: 言說 (speech) without 作因 (operative content) is, in this domain, precisely the failure mode that 《中論》 is targeting.
2.4 The AI assistant and the seamless reply
Contemporary life presents us with a strange, frequent rehearsal of 無因. You type a sentence into a chat model; an articulate reply appears. The reply looks like it came from nowhere. There is no body, no stance, no biography. Just output.
Mapping to 《中論》: This is a useful *analogy* — and the analogy is not identity — for the way 無因作 feels when it is in operation. The reply has not actually arisen without conditions; it is conditioned by a vast training procedure, by the prior prompts in the conversation, by randomness seeded into inference, by hardware. None of these conditions are transparent at the moment of output. The output gives every appearance of 無因作. Why the analogy is partial (and why the limit matters): A chat model cannot become aware of its conditioning and act differently in response to that awareness. A human practitioner — to take the analogy back to the dharma — *can*. The whole teaching exercise of investigating 作者 (the agent), 作業 (the action), and 作法 (the result) presupposes that a human can do something the model can't: surface the conditions and *vary the response*. So the limit of the analogy is the heart of the practice: we can turn 無因-appearing experiences into 因緣-recognizing ones, because we have access to our own conditioning in a way outputs don't.
Common misreading to flag: Some readers conclude "so 因緣 is a poetic description of how AI works." It isn't. AI is a useful mirror for the felt texture of 無因. But the Buddha's project is concerned with experience that *can be investigated and transformed*, which is precisely the opposite point from the analogy.
2.5 The self-blame spiral: "I did this to myself with no reason"
Something painful happened — a job loss, a relationship ending, a period of depression. The voice in your head tries to make sense of it. It tries blaming "me." It reaches for a fixed, blameworthy 作者 (a "broken me," a "weak me"), and a fixed 作業 ("I just messed up"), and a fixed 作法 ("and now I have to pay"). But the more it tries to fix the self who is at fault, the more it paradoxically cannot find any operative cause for the failing that it condemns.
Mapping to 《中論》: This is layer 3 in miniature. The attempt to nail down a blameworthy self *empties the very causal content of the deed being blamed.* "I did this for no reason" and "I did this for fixed-self reasons" are structurally the same error; they differ only in valence. Both cut the agent away from the network of conditions in which the action actually happened. The 因緣 fix: Re-weave the action back into its conditions. Not to excuse; to enable a useful response. "I was tired. I had been warned. I had missed the early signal. The early signal had been missed by several others in similar situations." Conditions emerge. With them, the deed becomes intelligible, and intelligible deeds are the only kind that can be addressed instead of merely suffered. The most important point for this scene: The 因緣 view is not soft on accountability. It is the *only* view that lets accountability do anything useful, because accountability presupposes intelligibility, and intelligibility presupposes conditions. The bluff of "I did this for no reason" sounds severe, but it is precisely the thought that disables the possibility of repair.
三、當代人為什麼需要它 (Why Contemporary People Need This)
Three pressures make this classical refutation a working diagnostic rather than a museum piece.
3.1 The presentation of 無因 in modern dress
Contemporary culture has produced three particularly common rhetorical figures that smuggle in the structural shape of 無因作 even when nobody is using the technical terminology:
- "It just is." When people say "that's just how I am," "that's just how it works here," "that's just how they are," they are functioning with a kind of frozen essence — a fixed self that produces outcomes for no traceable reasons. The episodes above (promotion, breakup, viral post) all illustrate the same shape. - The "manifestation" frame. A strand of contemporary self-help treats outcomes as arising from inner states without operative conditions: "vibe it into existence," "your frequency is the cause." The vocabulary is metaphysical; the structural commitment is precisely 無因作 — the result arrives without 因 (conditions) traced through 作因 (operative mechanism). - The "I am the cause of everything" frame. At the opposite pole, certain psychological languages push toward a self as exclusive cause: "you create your own reality," "you are the cause of all that happens to you." Reified into a fixed, omnipotent 作者, this is structurally layer 3's error: the agent is fixated and the deed becomes causally unintelligible — no conditions need to be investigated because "I" is held to be the only operative cause. Both the unlimited and the nonexistent author land in the same 無因 trap.
Each of these is not labelled a "philosophical" problem in everyday life. It is experienced as a small pattern: unexplained luck, inexplicable drifts in relationships, persisting self-blame loops, the inability to account for an algorithmic recommendation, the sense that creativity either flows or doesn't. The freedom available in 《中論》 is the freedom to see these patterns as instances of one structural error — and therefore to address them by undoing that error.
3.2 What 因緣生 (dependent origination) actually buys you
The 因緣 view is *not* a refusal to be surprised. It is not the claim that everything is predictable, nor that everything is determined, nor that the world is a closed system. It is the recognition that:
1. Conditions exist. Things don't arise ex nihilo ("from no cause"). When we say "this just happened," we have stopped investigating prematurely. 2. Conditions are investigable. We can name them, sometimes act on them, sometimes change them, sometimes not. Naming them is a competent act, not a metaphysical one. 3. The agent is not the sole cause. Real investigation distributes cause across a network, which means the agent's choices are real, consequential, and embedded — *neither* floating free as in 無因作 *nor* locked in as in 自作-as-fixed-self. 4. The result is intelligible without being predetermined. Operations bear fruit; what fruit is unpredictable in detail, but intelligible in direction. We can say that habit produces more of itself, that heedlessness narrows, that attention opens — without claiming we can predict any particular case.
Some contemporary parallels help here, with their limits kept in view. Cognitive science has documented that perception is conditioned at multiple levels — attentional, affective, hormonal, social, linguistic. Neuroscience has shown the activity of the brain substantially precedes conscious awareness of intending to act (a finding popularized as "your brain decides before you do"). Network science has shown that influence in social systems is distributed, non-local, and path-dependent, not assigned to a single node. All of these are consonant with 因緣生 in the sense that they complicate the picture of a single, fixed doer making a single, fixed deed and producing a single, fixed result. None of them proves 因緣生, because they describe mechanisms at a level of granularity different from the Buddhist analysis; and all of them leave open — indeed, leave more interesting than before — the question of how awareness becomes capable of *investigating* the conditioning rather than merely being pushed around by it.
The practical consequence of holding the 因緣 view is not pessimism. It is the opposite of pessimism. If nothing has conditions, you can do nothing about anything. If everything has conditions, you can investigate, address, and transform.
四、常見誤讀與澄清 (Common Misreadings and Clarifications)
4.1 "Without cause" is not "without fixed cause"
A common contemporary reading of dependent origination treats 因緣生 as the claim that "things have causes but not fixed ones — so they're spontaneous." This is actually very close to 無因作 dressed up in 因緣 clothing. 因緣 is *not* "no fixed self-nature and therefore magic." It is "arisen from conditions, the conditions themselves arisen from conditions, all empty of self-nature." The 中 (middle) in 中觀 (madhyamaka) is precisely the avoidance of two traps: *fixed thing* and *no thing at all*. Spontaneity in the loose sense is fine; spontaneity in the sense of 無因作 is exactly what 《中論》 is dismantling.
4.2 因緣生 is not 命定 (fate or determinism)
A second misreading treats 因緣生 as cosmic determinism: events have causes, all causes are fixed in advance, nothing can be done. But 因緣生 is *conditioned arising*, not *fixed unfolding*. Conditions in Buddhist analysis are themselves conditioned; the web is woven but the threads aren't pre-dyed. Even within the 因 chain there is room for what the tradition variously calls 願 (vow), 努力 (effort), and 修 (cultivation) — practices that change the conditions rather than merely reporting them. Determinism is the other side of the coin that 無因 作 is on: either everything is fixed (自作-as-fixed-self) or nothing is fixed (無因作). The middle says: conditions, investigate-able, transform-able, but no fixed self to be found anywhere in the process.
4.3 無作 (non-action) is not 無因作
Daoist-adjacent readers will recognize "wu wei" (無為) and ask whether 無因作 might be its Buddhist cousin. It is not. 無為 in Chinese contexts and 非造作 in Indic contexts refer to non-coercive, non-clinging responsive activity — *conditioned* but non-appropriative. 無因作 is the structurally opposite claim that something exists or comes to be *without any conditions at all*. The 《中論》 would not even engage 無作 as an object of refutation; 無作 is a way of acting within conditions. 無因作 is a claim about how things appear in the world. Conflating the two reads calmness as nihilism — a mistake that is the opposite of what is being said.
4.4 "The author doesn't exist" is not what is being asserted
Layer 3 sounds to some modern readers as if Nāgārjuna is denying the existence of the agent. He is not. He is denying the *fixed self-nature* of the agent. The 作者 (agent) is a *designation* (假名, prajñapti) that does real linguistic and practical work; it is not a reified metaphysical essence. To act as if there is no author at all is to slide into 無因作 at the level of personal responsibility. To act as if the author is a fixed metaphysical solid is to slide into 自作-as-eternal-self. The path is to use the designation with full pragmatic precision — "I did this; these are the conditions; this is the result being borne" — while refusing to point at any of the three as a *thing-in-itself.*
4.5 If everything is conditioned, where is freedom?
This is the question contemporary readers most often raise. The 因緣 view is sometimes thought to deliver us into a world of trapped conditioning with no room for the practice that practitioners spend decades on. The text we are examining helps answer this directly. Layer 4's metaphor — 「如化人以虛空為舍」 (the apparition in empty space) — shows what is at stake. The "author" that cannot act is the author that has *no conditions* (the apparition) or *the wrong conditions* (empty space). A real author, embedded in real conditions, can in fact act — which is why the entire Buddhist practice from 戒 (śīla) through 禪 (dhyāna) to 慧 (prajñā) presupposes that conditioned things can investigate and transform their conditioning. Freedom in this framework is not "uncaused choice"; it is the capacity to be the kind of conditioned process that knows itself as conditioned and acts on that knowing. That is the opposite of 無因作; it is the practical uptake of 因緣.
4.6 The most subtle trap: skipping the author altogether
If you read 《中論》 and come away thinking "the agent is empty, so authorship is an illusion, so nothing was done, so I need not worry about anything I do," you have — perhaps the most common contemporary error — landed in 無因作 while feeling you have landed in 因緣. The text shows why this is the deepest trap: the moment you drop 作者 (and 作業 and 作法) from your account of actions, you are in the position layer 5 warns about — no 罪福, no 業報, no 涅槃. The Buddhist practice of 因緣 never drops these; it uses them carefully, aware that each is a designation rather than a thing, and that misuse of the designations — either *fixing them* (which produces self-made errors) or *wiping them out entirely* (which produces causeless errors) — is what the 《中論》 is at such painstaking length to refuse.
Closing. 無因作 is the most expensive mistake in the chapter, not because it is the most argued *for*, but because it is the most easily *slipped into.* The fix is not to add more metaphysical weight to the doer, the deed, or the result. It is to *use* these designations with the precision the chapter models — past events investigated as condition-webs, present acts chosen as the conditions of future results, and the entire pattern held lightly enough to leave room for the cessation (涅槃) that the chapter shows depends on its not being frozen into metaphysics.
The 《中論》 at chunk:20483 finishes with a stark promise: 「即為破因果，作作者作法，亦復壞一切，萬物之生滅」 — concede 無因作, and everything goes. The middle path is the one that refuses both sides: investigate the conditions, use the agent as a designation, and don't reach for a foundation that is not there.