What this concept is actually about
憶知 — literally "remembrance-cognition," Sanskrit *smṛti-jñāna* — names something deceptively familiar: the moment when a later instance of knowing *re-knows* something a previous instance once touched. You smell jasmine and instantly recognize it as the jasmine your grandmother grew. You open a notebook, read a sentence you wrote five years ago, and the person who wrote it becomes briefly, strangely present again. That act of re-encountering a past object through the power of *smṛti* (remembrance) is what the *Foxing lun* 佛性論 (T1610, *Treatise on Buddha-Nature*) is interrogating.
The reason the concept deserves slow attention is that it is not the Indian/Chinese Buddhist equivalent of "memory studies" in psychology. Memory research asks *what happens in the brain* when we recall. The *Foxing lun* asks a far sharper metaphysical question: how is re-cognition possible at all, and what must we believe about the self in order to make it intelligible? The treatise doesn't treat *smṛti-jñāna* as a standalone topic; it pulls the term into a debate with a non-Buddhist opponent who insists that any successful act of remembering proves the existence of a permanent, unchanging self (*ātman*, here rendered 自性 *zìxìng*, "self-nature"). The treatise master then uses the very concept the opponent stakes his identity on to dismantle that identity.
So the concept sits at a precise location in the liberation-path's map: it is a *philosophical instrument* used to break the assumption that continuity of experience requires a continuous self. This is important for the *Foxing lun*'s larger project, which is to teach *Buddha-nature* (佛性) without thereby smuggling in a Hindu-style eternal soul. *Smṛti-jñāna* is one of the spots where the text shows its hand: you can have real continuity, real re-cognition, real cultivation across time — *without* needing an unchanging "me" behind it all.
The three conditions for *smṛti-jñāna*
The treatise's opponent claims that if *smṛti-jñāna* is real, then both **perception (證量, *pratyakṣa*, literally "the validating measure") and inference (比量, *anumāna*, literally "the comparable measure")** are validated by it. In plain terms: the opponent is saying we only know the past is real because we remember it, and we only trust memory because there is a real "I" that does the remembering. The treatise master does not first attack this head-on; he *clarifies what genuine memory requires*. Real *smṛti-jñāna* must hold three things simultaneously:
1. Same kind of object (境界同類). The thing being remembered now must be of the same type as what was cognized then. The jasmine I smell now and the jasmine I remember are both jasmines — without sameness-in-kind, recognition could not pick out what it picks out.
2. Same kind of knower (能知同類). There must be continuity of the *knowing* itself, not just sameness in the known. Otherwise an entirely unrelated event could be conflated with the original. A parrot that utters the same words doesn't thereby remember what you said.
3. Moment-to-moment cessation (知念念滅不得一時). Each knowing-moment is impermanent and does not coexist with the others. You don't have one giant mega-moment in which past and present perception happen together. Each flash of knowing arises and extinguishes as it arises.
These three together produce an unsentimental picture. *Smṛti-jñāna* lives in the gap between a thing now-extinguished (the previous knowing) and a thing not-yet-arising (the next knowing), with *sameness-of-kind* as the bridge. Nothing static holds the bridge; the bridge is constantly being rebuilt.
The opponent's trap and the treatise master's reversal
The opponent's move is to claim that these three conditions force us to posit a permanent self: only something truly self-identical could produce *sameness of knower* across vanished moments. The treatise master then performs a dialectical judo. He does two things at once.
First, he shows the opponent's position is internally inconsistent. If you take the self-nature (自性) view seriously, several absurdities follow. You cannot, while currently looking at a vase, simultaneously remember the vase: "When the eye is directly facing its object, it has not yet released it; thus it cannot remember another object. When the mind is taking in color, it cannot also be taking in sound." (This is a careful phenomenological point: attention is sequential, not simultaneous; you can't see and remember-visual-of-the-same-object-at-once.) And if "awareness is only one knowing" without moment-to-moment cessation — as the opponent's self-nature view requires — then within *smṛti-jñāna* there is no *smṛti-jñāna* at all: re-cognition collapses into a single opaque awareness that does no work of re-encountering.
Second, he accepts the opponent's formal framework and turns it. Yes, he concedes, we need *kind-sameness* across past and present. Yes, we need continuity. But — and here is the heart of the matter — continuity does not require permanence. It requires what the seed-and-grain analogy makes vivid: *"The earlier seed can produce the later grain. These two — earlier and later — are neither identical nor different, neither existent nor non-existent. Because the cause has extinguished, it is not 'existent'; because the result has arisen, it is not 'non-existent.' Thus cause-and-effect *smṛti-jñāna* is fully established, and we speak of no self-nature."*
This sentence, more than any other, gives the concept its force. Continuity is real, but it is the continuity of conditioned arising — one conditioned thing (the prior seed) giving rise to the next (the later grain), which itself becomes the cause of still further conditions. Nothing is preserved; everything flows; nothing is lost. The grain on your plate is the rice of the seed — it is *the same stuff of rice* — but it is not the seed. *Smṛti-jñāna* has exactly this structure: the recognition of jasmine now is the same *kind of knowing* as the original encounter, conditioned by it, carrying forward what is relevant — and is yet a different act of cognition.
Impermanence is what makes cultivation possible
A natural objection arises: *"If moments of knowing are impermanent and 'no self-nature,' how can cultivation (修習) be possible? Practice changes something, but if nothing stays, what's there to cultivate?"* The treatise master's answer is paradoxically sharp: precisely because things are impermanent, cultivation is possible.
He illustrates with the *kimpavā* (劫波婆) seed, which when irrigated with *ketaka* (佉陀羅) juice changes its nature — a botanical fact that Indian commentators liked to cite because it showed that a thing can be transformed by repeated conditioning. Applied to mind: a previous mind-moment conditions (熏習, *熏习*) the next mind-moment, not by leaving a static deposit but by shaping what is possible next. Over time, the effects of conditioning ripen — the bitter or sweet fruit appears. "From the no-self-nature mind, the various karmic results arise."
This is unusually counter-intuitive, and it deserves time. In our usual mental model, change seems to require a stable thing that *changes*. A cloud can change shape; the cloud's basic substance or pattern persists. The Buddhist analysis breaks both — there is no substance, and the pattern is not preserved — yet change still happens, because of causal efficacy without self-identity. Seed, plant, fruit; moment of intention, moment of action, karmic result. Impermanence is not an obstacle to practice; it is the *condition* for practice. A permanent self could not be transformed, only re-labeled. An impermanent stream can be redirected.
What the "life-principle" argument refutes
The opponent tries one further flank: *"Well, then the 'life-principle' (壽者) — the eternal lifespan-and-soul — is also a kind of* smṛti-jñāna*."* The treatise master runs three short arguments:
- If *smṛti-jñāna* had self-nature, there would be no forgetting. Yet forgetting is obvious. - If a thing is permanent, repeated cultivation makes no sense. You cannot repeatedly "do" anything to a thing that is already complete. Yet cultivation is real. - If a thing has self-nature, it would guard one object, one matter — and so there could be no sameness of *kind* across different objects, no "remembering many things," no *smṛti-jñāna* of "various matters."
The argument ends up inverting itself again: forgetting, repetition, and the ability to know many things are all evidence *against* self-nature. They are evidence *for* the seed-grain stream structure.
The absurd fallout and the final turn
If you still insist on a self-nature *smṛti-jñāna*, logical consequences degenerate: *every doubt should already be decisive wisdom; theft should be giving; one's wife should be one's mother*. The point is not stylistic — it is structural. A view that fails to discriminate properly between past and present, between known and inferred, between person and relation, has lost the capacity to *know* anything coherently.
The treatise master then moves to the argument that ties *smṛti-jñāna* to the larger Buddha-nature question. He shows that the "I" (我) cannot be established by either perception or inference:
- Perception establishes five-sense objects (visible form, sound, smell, taste, tangible) as the kind of things knowable without needing any further cognition about them — smoke *just appears* through the eye, before any reasoning about fire. - Inference establishes things like fire behind smoke — known *through another cognition* (the reasoning "where there is smoke there is fire"). - The "I" is *like neither*: it is not a sense object directly available (it "transcends the sense-object situation," 過根境故 — the "I" is not the kind of thing the eye sees or the ear hears), and it is not like a sense faculty whose existence we could infer by way of its objects.
So when we think we perceive "me," or infer "me," we are doing neither. *Smṛti-jñāna* — the sense of "I remember, I continue, I am the same one who walked into this room yesterday" — is a powerful intuition, but it is not a perceptually or inferentially established cognition. The text is not denying that you remember; it is denying that what-underlies-your-remembering is the kind of thing perception or inference can get hold of. This is the opening the *Foxing lun* exploits, in the larger work, to introduce Buddha-nature as not the same as *ātman* but as having the qualities the *ātman* claim tries to take credit for.
Life's walk-through
Let's take a normal week and watch where *smṛti-jñāna* is doing all of its work.
Reading an old journal
You open a journal from seven years ago. You read a sentence — *"Today I was afraid and didn't say anything"* — and you recognize the writer. There is a flash of *smṛti-jñāna*. The text on the page is the visible object now; the person who wrote it is the remembered object; and you, the reader, are re-knowing them. The treatise's three conditions fit naturally: the "kind" of person is the same (same handwriting, same concerns, same kind of self-reference); the knower now is of the same kind as the knower then (still you, dealing with the same kind of issues); and yet each moment of reading-and-recognizing is fresh and impermanent — you cannot hold the recognition.
The careful point, though, is what is *not* here. There is no ghostly "you" who existed seven years ago who has been secretly preserved and handed across time to meet today's "you." There is a stream of conditioning that reaches from those years to this sentence on the page: dispositions, vocabulary, scars. The continuity that *smṛti-jñāna* reports is real, but its "substrate" is not a self-nature. To feel the recognition as if it proves the existence of an enduring inner spectator is exactly the move the *Foxing lun* diagnoses — and refuses to authorize.
The song that returns you
A song from a particular room, a particular summer, plays in a café. The recognition is involuntary, total, almost bodily. *Smṛti-jñāna* at high intensity. Notice three things. First, the song is recognized as the same song (object-sameness-of-kind). Second, your listening-self is the same kind of listening-self that heard it then (knower-sameness-of-kind). Third, you did not need to *store* the song and then *retrieve* it; the moment of memory and the moment of hearing interpenetrate in a way that has no permanent container — the same imprecision that neuroscience also reports when it studies episodic memory.
What is *not* in the experience is an independent "I" doing the remembering. What *is* in the experience is conditioned arising — your nervous system and the song and the room. The treatise's point is not to dissolve the experience; it is to dissolve the metaphysical scaffolding we reflexively add to it.
Cultivation: meditation, language learning, therapy
This is where the *Foxing lun*'s most counter-intuitive move matters most: *only because each moment of mind is impermanent can meditation (or any other practice) actually change you.* Imagine trying to teach a word to a permanent, unchanging listener. There is nothing to teach — the listener is already complete. But to a stream whose every moment is fresh, words can land, dispositions can shift, painful patterns can be replaced.
Therapy works, when it does, for the same reason the treatise describes. The mind of last Tuesday's session is *not* the mind of today's session as a permanent thing, but it is *kind-same* and it has been conditioned by what happened in the room. By the time the change in question is visible (the "ripened fruit"), the conditioning may have passed through many impermanent stages. This is why change is sometimes slow, sometimes surprising, sometimes jumpy in ways that feel discontinuous even though the chain is continuous. The chain is real; the self within the chain is not.
Losing memories — and gaining a clue
People with dementia illustrate, painfully, why the *Foxing lun*'s analysis matters. They are *still kind-same knowers* in many domains, but the conditioning thread between yesterday and today is broken or thinned. Their continuity-of-selfhood, which we took for granted, reveals itself to be a process, not a possession. When we love someone through such an illness, we are loving a process too — not a repository that has emptied out.
The treatise's position is not the cold one. It is the *most accurate* one about what is happening.
Where contemporary analogies genuinely help — and where they stop
A wave on the ocean: each wave is not the previous wave (not same), is not unrelated to it (not different), arises from the same kind of conditions, and extinguishes. *Smṛti-jñāna*, the wave-shaped re-recognition of a past wave, is exactly this. Your city where all buildings are gradually replaced but you still say *"my city"* — same kind, not same entity.
The neuroscience analogy is partial and worth its limits. Modern memory research suggests that memories are reconstructed each time they are recalled; there is no fixed storage, no stable trace, only patterns of connectivity that are activated and slightly revised each time. This is a striking convergence with the *Foxing lun*'s seed-grain model — *the same kind of insight*, reached from very different starting points. But the analogy stops where it stops: neuroscience has no concept of *karma*, of *Buddha-nature*, or of the existential liberation the *Foxing lun* is steering toward. Reading neuroscience *as if it proved Buddhism* — or vice versa — is a category error. The convergence on the reconstructive nature of memory is genuinely suggestive; the philosophical consequences differ.
The AI/algorithm analogy is also partial and instructive. Your phone remembers things; you don't have to. Does this change what *smṛti-jñāna* is? In an unsettling way, yes — it changes where the conditioning stream runs. But your phone's "memory" is not *smṛti-jñāna* in the Buddhist sense: it cannot re-know, it cannot suffer over its contents, it has no continuity of conscious kind. The model of memory it embodies is exactly the *opposite* of the *Foxing lun*'s — storage-and-retrieval — and *that is why we need the treatise's vision*.
The place where most people map this wrongly
The most common mapping error goes like this: *"Memory requires a self, because I am the one who remembers."* This is so obvious it feels tautological. The *Foxing lun* is not denying that there is remembering. It is denying the inference. The "one who remembers" is not a separately existing self witnessing memory; it is the re-knowing itself — the conditioned re-arising of a kind-same cognition. The "I" in "I remember" is not a noun standing behind the verb; it is the verb wearing a noun's clothing. (This is exactly the move the treatise makes in its sixth layer, where the "I" is shown to be neither perceptible nor inferable.) Once you see through this, the everyday self-thought loses its evidentiary force. It still appears — it will keep appearing — but you no longer treat its appearance as proof that there is a backstage homunculus pulling its strings.
Why contemporary people need this
We live in an age that is obsessed with the continuity of self. Memory has become a kind of *digital estate* — your photos, your chats, your browsing history. The self is rendered as a *record* to be preserved, restored, curated, mourned when lost. If a person dies, the "loss of memories" feels, in our culture, almost like a *loss of personhood*. Conversely, people who cannot maintain consistent narratives about themselves are diagnosed as impaired.
The *Foxing lun*'s analysis is bracing against this entire cultural frame. It says: a person's continuity is real, and it is the continuity of a conditioned stream. There is no record preserved. The strength of someone's life is not in the preservation of an inner file but in the way each moment has been cultivated to condition the next. The very category of "preserving memory" may, in deep ways, be training us in the wrong metaphysics.
There are practical implications.
- In healing, this view helps. Trauma often feels like a permanent stain on a permanent self. The Buddhist analysis offers a different framing: trauma is conditioning that crosses *kinds* poorly, the way a seed that's been cooked can no longer germinate. Healing is the cultivation of fresh conditioning from moment to moment, and it works precisely because the self is not a permanent disk to be overwritten, but a process that can shift.
- In aging, this view helps. The fear of losing one's memory is, in part, the fear that *the self* will be lost. The treatise is honest: what *will* be lost is some continuity of past-conditioned cognition. What will *not* be lost is the kind-same knower of this moment. The self that is afraid is the false self; the knower that remains, kind-same, is closer to the truth.
- In habit and character change, this view helps. We often say "I'm just not a morning person" or "I have always been this way" as if a permanent entity were constrained by its own nature. The treatise's analysis of cultivation says: there is no fixed entity; there is a stream, and streams are redirectable. This is a metaphysical basis for *becoming* rather than *being*.
- In our relation with the past, this view helps. We tend to load the past onto a permanent self — "I am the one who was hurt, I am the one who achieved, I am the one who failed." The treatise pushes us toward something more exact: the stream that includes those events continues, but it is not a permanent witness accumulating trophies and scars. The past is not stored in a self; the past conditions the present without leaving its marks in any permanent place.
Common misreadings and clarifications
1. Treating 憶知 as just "memory." This is the most reductive error. *Smṛti-jñāna* is the inquiry into *what re-cognition requires of the self*. Reducing it to "memory" — like a cognitive faculty item in a textbook — loses the philosophical stakes entirely. The treatise uses it to *unmask* the assumption of self-nature, not to deliver a chapter on recollection.
2. Reading the text as denying re-cognition. The treatise does not deny that we re-recognize. It affirms re-cognition; it rejects the *metaphysical credit-card* — the imagined permanent self — that the opponent charges it to. To say memory doesn't require an eternal soul is not to say memory fails to happen.
3. Sliding from "no self-nature" into nihilism. The treatise is clear: it is because there is no self-nature that continuity of conditioned arising is possible. A stream has continuous flow without being one identical water-molecule. The empty nature of cognition is the *condition* for cognition's efficacy, not the denial of efficacy. Concluding "if there's no self, nothing really happens" is exactly the position the text has refuted.
4. Treating impermanence as the enemy of cultivation. We covered this above. The treatise's most surprising move is to argue: impermanence is the precondition of cultivation. Nothing permanent can be transformed; only streams can be redirected. People who feel trapped by the apparent solidity of their habits can take heart from this: the apparent solidity is itself part of the trap.
5. Missing that the target is the "I"-thought, not just memory. The final layer of the argument shows that the *real* opponent is the assumed "I" that supposedly does the remembering. *Smṛti-jñāna* is the lever; the lever is being applied to the lever's handle (the ātman-concept). This matters because, in contemplative practice, the work is not to weaken memory itself, but to *see through the silent witness that imagines it owns memory*. Meditation on this point loosens the apparent owner without damaging the actual experience of remembering.
6. Equating the seed-grain analogy with simple continuity. The seed-grain metaphor is doing something precise. The earlier seed is "not existent" (it has extinguished as seed); the later grain is "not non-existent" (it has arisen). This is not a vague "change over time" — it is a specific metaphysical stance, neither eternalism (the cause persists as itself) nor nihilism (the cause disappears entirely). Seed and grain share their kind; they do not share their identity. The same is true of two moments of *smṛti-jñāna*. Read with this precision, the analogy lands; read loosely, it becomes mere poetry about growing.
7. Imagining the concept as a standalone topic. A misreading encouraged by modern reference works. *Smṛti-jñāna* in the *Foxing lun* is a moment in an ongoing debate. Treated as a "topic" in isolation, it looks like a puzzle about memory. Treated as a move in an argument against *ātman*, it becomes legible as one of the sharpest instruments in the Yogācāra-Vijñānavāda tradition's toolbox for distinguishing *Buddha-nature* from *ātman*.
A closing image
You are walking and you pass a place you walked ten years ago. For a moment, the present and the past are continuous in you — same kind of knower, same kind of place, briefly held together. Then the moment passes. You are not the same person who walked ten years ago; you are not a different one either. The walk then conditioned the walk now; the walk now is its own arising; both are impermanent; both belong to the same kind of stream. The recognition is real. The "I" who recognizes is the recognition itself.
The *Foxing lun* would call that recognition 憶知 — *smṛti-jñāna*. And it would also say: do not let that very recognition persuade you that something self-existent stands inside the recognition doing the looking. The wisdom is in honoring the re-knowing without dignifying it with an owner. The freedom of it comes slowly, but it is exactly where the Buddhist path is heading.