What this concept is saying
自續 (Svātantrika, “the Autonomist approach”) names a way of practicing Madhyamaka reasoning—and, in the Tibetan doxographical tradition, one of the two principal divisions of Madhyamaka, paired with 應成 (Prāsaṅgika, “the Consequentialist approach”). The distinction is not simply that Svātantrikas use logic while Prāsaṅgikas do not. Both reason, refute error, and accept that reasoning can function conventionally. The dispute concerns what must already be established, and in what manner, for a Madhyamaka inference to work.
The term “autonomous” can therefore mislead a contemporary reader. It does not mean intellectual independence, personal autonomy, or a self-existing consciousness. According to the classical materials here, 《辨了不了義善說藏論》 treats “自在” and “自續” as equivalent translations in this context: the thesis is established through an inference whose subject, reason, example, and intended conclusion are secured by valid cognition rather than merely borrowed from the opponent’s commitments.
A 自續比量 (svatantra inference, “autonomous syllogism”) has several indispensable parts:
- 前陳有法, the subject or basis under discussion: that about which something is to be proved—an eye faculty, a sprout, or another phenomenon. - 所立法, the property to be proved: for example, being without intrinsic production. - 因相, the reason or logical mark: that by which the conclusion is established—such as arising dependently. - 喻有法, the exemplifying case: an agreed example that displays the relevant relation. - 遍 or 周遍, the entailment: the reason must pervade the property to be proved in the required way.
In the Buddhist logical idiom summarized by the classical entry, the reason must satisfy the relevant 三相 (three characteristics of a valid reason): it must apply to the subject; it must occur in appropriate similar cases; and it must be absent from relevant dissimilar cases. Crucially, the proponent and opponent must establish these elements through valid cognition in a jointly acceptable way. As 《菩提道次第廣論》 states in the supplied passage: “若以立、敵共成之因，成立所立，名自續因；若不以彼，唯由敵者所許三相，成立所立，名為應成。” In other words, establishing a conclusion by a reason validly established for both parties is called an autonomous reason; deriving the conclusion through the three characteristics accepted by the opponent alone is called a consequence.
Imagine two people debating whether a building is unsafe. A Svātantrika-style argument tries first to secure common evidential ground: both parties recognize this particular building; both accept the observed structural condition; both accept the engineering relation between that condition and instability; and both accept an appropriate comparison. On that shared platform, the proponent establishes a conclusion. A consequential argument may instead provisionally accept the opponent’s own account and show that, if consistently followed, it produces a result the opponent cannot accept.
That analogy illuminates the form, but it does not yet reach the doctrinal depth of the dispute. In Madhyamaka, the very status of the “building,” the evidence, the knower, and the conclusion is at issue. Can the two parties establish a common subject while one sees it as possessing an objectively identifiable character and the other denies precisely that mode of existence? This is where the disagreement becomes more than a preference about debating technique.
Within the Geluk presentations supplied here, the autonomous inference presupposes that its subject, reason, and probandum can be established through 自相 (svalakṣaṇa, an “own-character” or objectively identifiable character) at least conventionally. This does not mean that Svātantrikas affirm ultimate, independent, or truly established existence. Thinkers such as Bhāviveka are still presented as Madhyamikas because they use reasoning to reject true existence. The finer issue is whether conventional phenomena possess a basis of establishment that is not merely posited through names and conceptual imputation.
The materials draw the distinction carefully. Svātantrikas reject a nature existing as it appears wholly apart from the mind’s presentation and superimposition. Yet they are said still to accept a nature that is not established *merely* through nominal imputation. Thus “mind-established,” “false,” and “illusory” cannot be assumed to mean exactly the same thing in every Madhyamaka system. Two schools may employ similar vocabulary while locating the boundary between valid convention and reification differently.
This explains the Svātantrika analysis of phenomena described in 《入中論善顯密意疏》. Knowable things are examined as conditioned or unconditioned; conditioned things are further considered as material or nonmaterial. A material thing is examined through its spatial parts—east, west, and so forth. A mental event is examined through temporal parts—earlier and later moments. Whatever is conditioned is therefore something with parts.
One then investigates the relation between the parts and the whole. If they were utterly different, no intelligible relation could join them. If they were strictly one, the apparent distinction between whole and parts could not be explained. What appears as a unitary thing therefore cannot withstand analysis as a self-grounded entity. It is empty “like an illusion”: it appears and functions, but it is not established in the substantial manner projected upon it. The same mode of analysis is extended to unconditioned phenomena such as space and the nature of phenomena. In the account attributed to Śāntarakṣita and his followers, all knowable things are thereby determined to be without real existence.
“Like an illusion” does not mean that nothing happens. An illusion can appear, affect attention, prompt fear, and condition action while lacking the reality attributed to it. Likewise, emptiness does not erase dependent arising. It corrects the imagined way in which dependently arisen things seem to stand by themselves.
The Prāsaṅgika criticism, as articulated in the supplied works of Tsongkhapa’s tradition, presses on the proposed common foundation. If the opponent apprehends the subject as established by its own character while the Madhyamika does not accept that manner of establishment, what exactly has been jointly established? One cannot simply strip away the disagreement and announce a neutral “general subject,” because the debate concerns how that very subject is established. If the subject is not established in the required common manner, the reason resting upon it cannot be an autonomous reason either.
This is why rejecting autonomous inference is not equivalent to rejecting inference, valid cognition, subjects, reasons, or conclusions altogether. 《辨了不了義善說藏論》 explicitly preserves their conventional operation after rejecting establishment by own-character. A Prāsaṅgika can use “dependent arising” as a reason and “reflection” as an example to remove a mistaken belief about a sprout. What is denied is that these elements must first be independently secured through self-character as a common object of the two parties’ cognitions.
Nor does “having no thesis” mean refusing to say anything. In this context it means not putting forward a thesis affirming intrinsic or self-characterized existence under ultimate analysis. Conventionally, one still distinguishes saṃsāra from nirvāṇa, harmful conduct from beneficial conduct, and sound reasoning from confusion. The supplied 《菩提道次第廣論》 material emphatically rejects the claim that Madhyamikas accept nothing even conventionally. Emptiness is not intellectual muteness and not nihilism.
The distinction therefore concerns three connected levels:
1. The logical level: Must the proponent establish an inference’s subject, reason, example, and entailment through valid cognition common to both sides, or may one expose contradiction from within the opponent’s own commitments? 2. The conventional-ontological level: Are phenomena conventionally established through an objectively identifiable own-character, or only through dependent designation without such a character? 3. The liberative level: What kind of reasoning most completely removes the subtle reification that makes persons and things seem to possess self-grounding cores?
The third level explains the concept’s place on the path to liberation. Svātantrika is not chiefly a badge in the history of philosophy. Its reasoning is a medicine for grasping. By dividing, examining, and testing the relation between whole and parts, it weakens the felt certainty that “this thing is simply there from its own side.” That loosening matters because attachment, hostility, fear, and pride feed upon apparently solid objects: *my* fixed self, *your* intrinsic defect, *this* independently existing success, *that* inherently unbearable loss.
From the Prāsaṅgika perspective represented in these materials, however, the medicine must go further. If conventional things are still thought to possess even a subtle objectively identifiable character, some residue of the disease remains. Tsongkhapa therefore follows Buddhapālita and Candrakīrti in rejecting Bhāviveka’s autonomous inference. That is a sectarian judgment within a particular interpretive lineage, not a neutral definition accepted in identical form by every historian or every Buddhist tradition. “Svātantrika” itself is especially prominent as a later Tibetan classification, although the supplied 《菩提道次第廣論》 argues that the distinction was grounded in Candrakīrti’s 《明顯句論》 rather than arbitrarily invented.
A walkthrough in everyday life
Suppose you send an important message to your partner. Several hours pass. You see that they have been active online, but they have not replied. The thought arrives: “They do not care about me.” This small episode lets us walk through the Svātantrika issue without turning it into a loose metaphor.
First identify the subject, the 前陳有法. It is not yet “an uncaring partner.” Properly speaking, it is the concrete situation under examination: this person’s non-reply during this interval. If you build contempt into the subject from the start, you have smuggled the conclusion into the premise.
Next identify the property to be proved, the 所立法. Your mind may be trying to prove: “This non-reply establishes lack of care.” Alternatively, Madhyamaka analysis may investigate a different property: “This episode is not established as rejection from its own side.” These are quite different theses.
Then identify the reason, the 因相. The emotionally charged inference may be: “Because they were online and did not answer.” But that mark must be tested. Is it genuinely present in the subject? Probably. Does it invariably accompany lack of care in relevant similar cases? No. Someone may open an app while exhausted, distracted, working, caring for a child, or unable to formulate a thoughtful reply. Is it absent whenever care is present? Again, no. The inference fails because the alleged reason does not pervade the conclusion.
This is the first benefit of Buddhist logical discipline: it interrupts the speed at which a feeling becomes evidence and evidence becomes an essence. The fact “no reply yet,” the painful feeling it conditions, the interpretation “rejection,” and the judgment “uncaring person” are not one indivisible object.
Now examine the episode through parts and dependence. Spatially and materially, there are devices, bodies, rooms, signals, and visible interface elements. Temporally, there was the sending, the waiting, moments of checking, the partner’s separate sequence of activities, and your later interpretation. Conceptually, “being ignored” depends upon expectations, relationship history, norms about response time, and what counts as an adequate reply. Causally, it depends upon many conditions neither person controls completely.
Search for “the rejection itself,” independent of all those parts, causes, temporal stages, and acts of designation. Is it identical with the unread message? Then every unread message would already be rejection. Is it wholly different from the events? Then it would have no intelligible connection to them. Is it located in the screen, the delay, your bodily unease, your partner’s intention, or the story linking them? No single candidate is the self-grounded whole your distress initially presented.
This does not prove that neglect never occurs. That would be another error. Patterns of avoidance can be conventionally real, ethically important, and worthy of a clear conversation. Emptiness does not require remaining in a harmful relationship. It means that even a harmful pattern exists dependently—through repeated events, intentions, effects, interpretations, and conditions—not as an immutable substance inside a person.
Where does *autonomous* reasoning enter? You and your partner might try to establish a shared evidential basis: which messages were sent, what response expectations had been discussed, what happened on similar occasions, and what consequences followed. From commonly established evidence and an agreed relation, you draw a conclusion. That resembles the practical aspiration of a svatantra inference.
But the Prāsaṅgika objection asks whether the supposedly shared object is truly shared in the required manner. Perhaps you experience “four hours without reply” as intrinsically dismissive, while your partner experiences it simply as unfinished communication. The words are common, but the object as apprehended is not. If your argument quietly depends on “dismissiveness” already being present in the event from its own side, the common ground is counterfeit.
A consequence-style intervention might therefore begin inside your own position: “If being online without replying necessarily proves lack of care, then whenever I open an app before answering someone I must lack care for them. But I do not accept that. So my rule cannot stand as stated.” Nothing has been rendered unreal. A reified rule has been made to collapse under its own implications.
The common mapping error is to say that the unread message is “empty” because it is digital or because it is “all in your head.” That is not the analysis. The material device and external event are not reduced to private thought. Nor does mind-only fabrication explain everything. The point is that the event, its meaning, and the persons involved are not established as the self-contained entities they appear to be; they arise through causes, parts, relations, cognition, and designation.
The same walk-through works while scrolling social media. The subject may be “this person’s successful life as presented in this post.” The proposed reason may be “because the image shows wealth, beauty, and applause.” The conclusion may be “their life is intrinsically better than mine.” But the reason does not entail the conclusion. The post is a selected frame, temporally dependent on events before and after it, materially dependent on a device and platform, socially dependent on conventions of prestige, and algorithmically dependent on ranking systems. None of this proves that the person is secretly miserable; it only blocks the unjustified inference from a curated appearance to an inherent life-value.
AI provides another useful but limited comparison. A model’s fluent answer can appear to contain a single, self-grounded “view,” although the output depends on training data, architecture, prompts, sampling, system constraints, and the reader’s interpretation. This can help us notice how unity and authorship are constructed from conditions. But an AI system is not a scientific proof of emptiness, and unsettled questions about cognition or consciousness cannot be solved by importing a Madhyamaka term. The comparison is pedagogical, not an identity claim.
Why contemporary people need it
Contemporary life repeatedly converts dependent patterns into apparent essences. Recommendation systems turn prior attention into “what you truly want.” Workplace metrics turn a limited measurement into “your value.” Diagnostic labels can become “who I am.” A partner’s lapse becomes “who they are.” A child’s difficult week becomes “the kind of child I have.” Political and social identities become supposedly self-existing natures.
Svātantrika analysis trains a pause between appearance and assent. It asks:
- What exactly is the subject under discussion? - What property am I attributing to it? - What evidence serves as my reason? - Does that reason genuinely entail the conclusion? - What causes, parts, relations, and acts of designation make the apparent whole possible? - Am I treating something dependently established as though it supplied its own foundation?
That pause is not merely academic. It can reduce the compulsion to defend a fixed self. If “failure” cannot be found as an intrinsic core—only particular outcomes, capacities, expectations, causes, and interpretations—then responsibility remains possible without metaphysical self-condemnation. One can repair an error without becoming “an error.” One can oppose another person’s conduct without needing to imagine an inherently corrupt person beneath every action.
This reasoning also guards against credulity. The requirement that a reason genuinely support its conclusion is valuable amid algorithmically amplified certainty. Repetition is not pervasion; virality is not valid cognition; emotional force is not a logical mark. Yet Madhyamaka adds something ordinary critical thinking may not: it examines not only whether a claim is well supported, but also whether the subject and evidence are being treated as possessing a solidity they do not have.
On the liberation path, this discipline serves insight rather than debate for its own sake. Ethical conduct steadies life; meditative concentration makes the mind capable of sustained examination; wisdom investigates the self and phenomena. Reasoning first produces an inferential understanding of emptiness. Through repeated cultivation, that understanding is meant to transform how things are experienced, weakening grasping at persons and phenomena.
The tradition represented here also connects the dispute with the depth of obscuration. It reports a Svātantrika account according to which practitioners of the two lesser vehicles may temporarily suppress manifest grasping at a self of phenomena, while the seeds of two forms of such grasping require immeasurable accumulations and appropriate supports to be eradicated; it also distinguishes afflictive obstruction from the cognitive obstruction associated with habits of dualistic appearance. Whatever later scholastic disagreements surround that mapping, its practical point is important: understanding an argument is not yet the exhaustion of deeply habituated appearance. A person may sincerely understand dependence and still reflexively experience blame, status, body, and identity as solid.
Thus Svātantrika occupies a middle and pedagogically powerful position: it rejects ultimate reality and reveals phenomena as illusory and dependent, while—according to its Prāsaṅgika critics—retaining a subtle conventional foundation that must itself be relinquished. To understand this position sympathetically is to see why autonomous reasoning can be a profound solvent of naïve realism. To understand its criticism is to see why Madhyamaka refuses to turn even the conceptual proof of emptiness into a final ground.
Common misunderstandings and clarifications
“Svātantrika means believing in a truly existent self.” No. “Sva” in the name does not refer to an autonomous person or ātman. It concerns an inference established from the proponent’s own side through premises taken to be validly secured for both parties. Svātantrika Madhyamikas reject true existence; the later dispute concerns the subtler conventional status of phenomena and logical subjects.
“Svātantrika is simply anyone who uses syllogisms.” No. The supplied 《入中論善顯密意疏》 explicitly warns against this collapse. Accepting a reason, a probandum, or an inference that refutes production from the four extremes does not by itself make one a Svātantrika. If every refutation that indirectly establishes an alternative counted as autonomous, the distinction between autonomous inference and consequence would disappear.
“A consequence contains no assertion of any kind.” Too crude. A consequence shows what follows from the opponent’s commitments without first establishing the disputed elements through an autonomous common measure. But Prāsaṅgikas still accept conventional reasoning, valid cognition, dependent arising, and communicable distinctions. “No thesis” is restricted to refusing a thesis grounded in intrinsic nature or own-character under ultimate analysis.
“Prāsaṅgikas think language and logic are useless.” No. Their objection is not to the conventional function of reasons and conclusions. It is to the claim that these must be established through self-character as independently available common objects. Logic remains useful precisely because dependent conventions have regularity and consequences.
“Svātantrika and Prāsaṅgika differ because only the former qualifies negations with ‘ultimately.’” The supplied 《菩提道次第廣論》 rejects this as a defining distinction: “說於所破加勝義簡別，唯是中觀自續派者，極不應理。” Both approaches must avoid negating conventional existence when denying intrinsic existence. Saying “a sprout does not exist ultimately” can be appropriate; what matters is how the subject, reason, and negation are established, not the mere presence of the word “ultimately.”
“If things are like illusions, pain and ethics are imaginary.” No. “Illusory” describes the mismatch between appearance and mode of existence. Pain arises dependently and can be devastating; actions have consequences; compassion is necessary. Indeed, dependence is what makes change, responsibility, treatment, learning, and liberation possible. An intrinsically fixed condition could neither arise nor cease.
“Established by mind” means that my private thoughts create the external world.” No. The phrase belongs to a technical dispute about imputation and the basis on which phenomena are established. It should not be reduced to solipsism. The classical material itself warns that Svātantrikas and Prāsaṅgikas can both speak of things as false or mentally established while meaning significantly different things.
“If the whole cannot be found apart from its parts, the whole does not exist at all.” That confuses absence of intrinsic establishment with total nonexistence. A car transports people, a team completes work, and a relationship can nurture or injure. These wholes function conventionally because of their parts, arrangements, causes, purposes, and designation—not because an independent “whole-essence” is found in addition to them.
“Refuting one side automatically makes every argument autonomous.” No. Ordinary contradiction often means that excluding one alternative establishes another. The supplied 《入中論善顯密意疏》 insists that this general structure—sometimes discussed as establishing something through the reverse implication of a refutation—must not be confused with a self-standing autonomous inference. Otherwise “consequence” and “autonomy” would become indistinguishable.
“The Prāsaṅgika criticism proves Svātantrika was foolish or spiritually useless.” No. The debate is subtle because both sides are attempting to eliminate reification while safeguarding dependent convention. Svātantrika reasoning can expose the impossibility of self-grounded wholes and reveal all knowable things as illusion-like. Its critics argue that its account of common evidential establishment leaves a finer objectifying residue. One should understand the achievement before evaluating the critique.
The deepest lesson is not that one must acquire the correct scholastic label. It is that even a proof of emptiness can become something to grasp. Autonomous reasoning teaches us to examine our grounds carefully; its Prāsaṅgika criticism asks whether those very grounds have been made too solid. Together they turn logic inward. The question ceases to be merely “Can I defeat the other position?” and becomes: “What hidden mode of existence must I assume before my certainty feels compulsory?”
When that assumption is seen, reasoning has begun to do liberative work. It no longer supplies the ego with a more sophisticated opinion. It loosens the imagined foundation on which fear, craving, hostility, and self-importance try to stand—while leaving intact the dependent world in which choices matter and compassion is urgently needed.