Kaivalya Pada 4.21 — If Mind Were Known by Mind, an Endless Regress and a Confusion of Memories

If the mind were seen by another mind, there would be an endless regress of cognition cognizing cognition, and a confusion of memories.

Unpacking the Sanskrit

The sūtra is compact, and each of its terms carries weight. Citta-antara joins citta (mind, the field of cognition) with antara (other, another, a further instance) — so cittāntara is "another mind," a second cognition distinct from the first. The compound cittāntara-dṛśye adds dṛśya (from the root dṛś, to see) in the locative, "in the case of the mind's being seen by another mind." This is the hypothesis the sūtra entertains and then destroys.

The first consequence is buddhi-buddheḥ, literally "of the cognition of cognition" — buddhi (from budh, to wake, to know) doubled, naming the stacking of one act of knowing upon another. To it is attached atiprasaṅga: ati (beyond, excessive) plus prasaṅga (attachment, entailment, the dragging-in of a consequence) — an unwanted over-extension that, once admitted, cannot be stopped. The second consequence is smṛti-saṅkara: smṛti (memory, from smṛ, to remember) and saṅkara (mixing, commingling, confusion) — a jumbling of memories. The connective ca ("and") binds the two ruinous results together as a single twofold refutation.

What the sutra asserts

Patañjali is closing off an escape route. The preceding sūtra denied that the mind can know itself within one and the same act, since one thing cannot be simultaneously the agent and the object of a single operation. A defender of the mind-as-knower might then retreat to a subtler position: grant that no single cognition grasps itself, but let each cognition be grasped by a further, succeeding cognition. The mind would still do all the knowing, only across a relay of acts rather than within one. This sūtra shows that the relay collapses.

It collapses for two reasons. First, the relay never ends. If cognition A can be known only by a later cognition B, then B — being itself a cognition, itself something that must be known if it is to function as a knower for us — requires C, and C requires D, and so on without terminus. The phrase buddhi-buddheḥ names exactly this spiral: cognition of cognition of cognition, an atiprasaṅga that explains nothing because it never arrives anywhere. An account that demands an infinite series before any single knowing is secured has, in effect, declined to give an account at all.

Second, the memories would run together. If every cognition were registered and retained by a separate succeeding cognition, the impressions belonging to this swarming, endless series would commingle — this knowing's memory contaminated by that knowing's memory, with no clean thread of "this experience, then its trace." Yet our remembered life is in fact orderly and single-lined: I recall my experiences as mine, in sequence, without the chaos the theory predicts. The evident coherence of memory is therefore the empirical refutation of the relay. Smṛti-saṅkara is the absurdity that would follow if the theory were true, and since it does not follow, the theory is false.

The place in the pada's argument

This sūtra is the second blow of a tightly built sequence near the heart of the Kaivalya Pāda's epistemology. Patañjali has been establishing, against any view that would make the mind self-sufficient, that the mind is something known rather than the ultimate knower. The line of argument runs: the mind is not self-luminous (it does not light itself); it cannot be both seer and seen in one act; and now — here — it cannot be rescued by appeal to a succession of further minds, because that succession is both endless and memory-wrecking.

With both horns blocked — neither self-knowing within one act nor knowing across an infinite relay — only one possibility remains. The witnessing of the mind must come from something that is not itself a mind awaiting further witnessing: something whose knowing is not one more cognition in the queue but a self-established, self-luminous ground. That is the puruṣa, the changeless seer, and the very next sūtra (4.22) names how, through it, consciousness is known without the seer ever moving. This sūtra is thus the negative clearing that makes the positive doctrine of 4.22 necessary. It does its work by reductio: assume the alternative, and watch it ruin both inference and memory.

The commentary tradition

Vyāsa, in the foundational Yoga-Bhāṣya, reads the sūtra as the decisive refutation of the Buddhist position that a cognition can be cognized by a subsequent cognition rather than by an abiding witness. He presses the regress vividly: if the second cognition needs a third to be known, and the third a fourth, there is no stopping, and meanwhile the orderly succession of memory becomes impossible. For Vyāsa the lived fact of coherent recollection is the hammer that breaks the theory — we plainly do remember in clean order, which the relay forbids.

Vācaspati Miśra, in the Tattva-vaiśāradī, sharpens the dialectical setting, locating the opponent precisely as the doctrine of momentary cognitions each illumined by the next. He underscores that the regress is not merely long but genuinely unending (anavasthā), and that an unending requirement is no requirement that can ever be met; the would-be explanation defeats itself. Vijñānabhikṣu, with his theistic and Vedānta-inflected reading, frames the conclusion in terms of the necessity of an unchanging conscious principle — the puruṣa — that grounds awareness without entering the series, treating the sūtra as a step toward affirming a witness beyond all the moving cognitions. Bhoja, in the Rājamārtaṇḍa, gives the compact gloss: were mind known by mind, you would have endlessness on one side and the ruin of memory on the other, so neither limb stands. Across these voices the shared verdict is the same — the regress and the confusion of memory together force the recognition of a knower outside the chain.

The Samkhya frame and the interpretive crux

Underlying the argument is the Sāṃkhya partition that Yoga inherits. Everything that changes, combines, and serves a function belongs to prakṛti, unconscious nature; only puruṣa, pure consciousness, is the seer that does not change. The mind (citta) is the subtlest evolute of prakṛti — luminous-seeming but not itself conscious in the originary sense. Within this frame the regress argument is not an abstract puzzle but a metaphysical necessity: since every citta-act is a modification of nature, no stacking of such modifications could ever generate the self-grounding awareness that the witness alone supplies.

The interpretive crux worth marking is what the sūtra does and does not prove. It does not, by itself, demonstrate that the puruṣa exists; it shows that the mind cannot be the final knower, which leaves the door open for the witness that the system independently affirms. Read carelessly, the sūtra can seem to assume its conclusion; read carefully, it is a disjunctive elimination — self-knowing within one act (ruled out earlier) or knowing by another mind (ruled out here) — whose remainder is the changeless seer. The force of the argument is in the closing of options, not in a direct sighting of puruṣa; that sighting belongs to the discriminative discernment toward which the pāda is moving.

One may also ask why the meter and compression of the sūtra matter to its meaning. Patañjali's aphoristic form forces every term to bear maximal weight, and here the doubling of buddhi in buddhi-buddheḥ is itself a kind of icon of the regress it names — the word repeats as the cognitions would, and the ear hears the spiral before the mind parses it. The sūtra thus performs its argument as much as it states it. This marriage of form and sense is part of why the text rewards slow study: the very economy that makes it terse is what makes each surviving word a deliberate signal, and the reader who lingers on atiprasaṅga and saṅkara finds the whole refutation folded into two well-chosen nouns.

The move to halt a regress

The regress argument here is a textbook instance of a strategy found across the world's philosophy: to stop an explanatory chain, posit a term that does not itself need the explanation it gives. Aristotle's argument in the Metaphysics for an unmoved mover — a source of motion that is not itself moved — has the identical shape, halting the series of movers by a first that does not require a prior. Patañjali's seer is the unmoved knower behind the chain of knowings: the term that grounds the series precisely by standing outside it.

The homunculus and the witness

The danger of "cognition of cognition of cognition" without end is exactly the homunculus problem that later haunts theories of mind. If understanding a perception requires an inner perceiver who perceives it, that inner perceiver needs another within, and so on forever — the little man inside the little man. Patañjali names the trap many centuries before it was rediscovered in this form, and resolves it the only way it can be resolved: by a knower that is not one more inner perceiver in the queue, but a self-luminous awareness that is not itself an object awaiting another's gaze.

Testing theory against lived memory

The appeal to the confusion of memories is a distinctively careful touch, turning to ordinary experience to break a theoretical regress. Because our memory is in fact orderly and single-lined, no mechanism that would jumble it can be the truth of how the mind is known. This kind of argument — measuring a metaphysical proposal against the evident coherence of mental life — shows the empirical sobriety running beneath the Yoga system's metaphysics, a sobriety it shares with the wider Indian debate over self-awareness (svasaṃvedana) in which Buddhist and Brahmanical thinkers pressed one another on precisely this point. The same instinct animates Western philosophers who test theories of mind against the lived unity of experience: any account that predicts a chaos we plainly do not live is, by that prediction, refuted. Patañjali's confidence in the orderliness of memory as a fixed point against which to measure theory is exactly this move, made many centuries earlier.

Some explanations defer the real question instead of answering it — they push it back a step, and then another, forever. To watch for the endless regress is a discipline of clear thinking: when an account only relocates a puzzle rather than resolving it, the account is incomplete. Real understanding has somewhere to rest, a place where the chain of "and what explains that?" finally comes to ground. An honest mind learns to feel the difference between an answer and a deferral.

Applied inwardly, the sūtra cautions against trying to secure self-knowledge by piling thought upon thought — examining the examiner, doubting the doubter, watching the watcher. That ladder has no top rung. The coherence of one's own remembered life already points past the regress to a single, steady awareness that holds it all together. It is there, not in the endless stacking of mind upon mind, that the self is found — not as the last in a series but as the ground the series never reaches. To stop climbing is not to give up the question but to discover that the questioner was the answer all along.

The layered-monitor impasse

The infinite regress this sūtra exposes is the impasse that any account of awareness as "a process watching a process" eventually meets. If every level of monitoring needs a higher monitor to make it conscious, the explanation never bottoms out — a difficulty still very much alive in contemporary efforts to describe awareness as layered information-handling, where each tier of representation seems to require a further tier to be experienced rather than merely processed. Patañjali's response is that some ground must do the witnessing without itself being one more thing witnessed.

The spiral of rumination

On the personal side, the warning against the self-watching-the-self speaks to the spiral of rumination — thinking about one's thinking about one's worry, then worrying about the thinking, with no resting point. The loop feels like effort toward clarity but only deepens the entanglement. The way out is not a further, cleverer layer of analysis but a step off the ladder altogether, into the simple awareness in which all the layers appear.

Where attention can settle

The coherence of one's life is held in that simple awareness, not in the regress. Recognizing this reframes the aim of inner work: not to add another tier of self-observation, but to rest as the one in whom every tier is already seen. The settling is downward and inward, toward the ground, rather than upward toward a final, never-arriving overseer.