1. What this concept is saying
1.1 The three-fold structure: 見、可見、見者
When Nāgārjuna, in the *Mūlamadhyamakakārikā* (《中論》, T1564), examines perception, he doesn't begin with abstract questions like "what is consciousness?" or "what is the self?" He begins with a structural observation: every act of cognition implicitly involves three terms —
- 見 (the seeing-function, and by the commentary's gloss, the eye-faculty on which that function depends): the actual operation of seeing. - 可見 (the visible): what is seen — color, form, the visual object. - 見者 (draṣṭṛ, "the seer"): the assumed subject of the seeing act.
This three-fold template is not unique to vision. The text extends the same pattern to hearing (聞、可聞、聞者), to feeling and appropriation (染、可染、染者), and — by implication — to all cognition and action. 見者 functions here as a paradigm term (範型術語): an index for the entire pattern of *agent* (能作), *action* (所作), and *object* (作用) that quietly structures our everyday framing of experience.
The crucial point is that this is not a phenomenological description but a structural diagnosis. The moment we say "I see," "I hear," "I feel," we have already *posited* a seer — a unified subject that possesses and deploys these powers. The analysis in this section is dedicated to showing that this posited seer cannot actually be located anywhere in the structure we have set up.
1.2 What kind of "self" is 見者?
The commentary on the *Mūlamadhyamakakārikā* is unusually direct: 「見是眼根、可見是色塵、見者是我」 — "seeing is the eye-faculty, the visible is the form-object, the seer is the self."
But this "self" is carefully circumscribed. It is *not* the Brahmanical ātman — the eternal, omnitemporal Self that survives death and migrates. The text identifies this 見者 with a cluster of interchangeable labels: 我 (self), 本住 (the "originally dwelling one"), 神 (the "spirit"). All three denote the same structural assumption: a unified subject that pre-exists the sense-faculties and uses them. If such a subject exists, it must:
- pre-exist the sense organs (it must be "originally dwelling" before the eye, ear, nose come online); - remain unified across all senses (the same "spirit" sees, hears, feels); - be capable of deploying any sense-faculty at will (since it stands behind them all).
This is the precise target. Not "self" in some diffuse spiritual sense, but the very specific philosophical assumption of a unified, pre-existing subject that *operates* the sense-faculties.
1.3 The mutual-conditioning template (去者與去法)
Before pressing the case against 見者, Nāgārjuna lays a structural pattern using 「the goer and the going」 (去者與去法):
去者因去法成，去法亦因去者成。
The goer exists only insofar as there is a going; the going exists only insofar as there is a goer. You cannot first establish a self-standing goer who *then* performs a going; nor can you first establish a self-standing going that *then* has a goer attached. They are mutually conditioning (互相因待).
This is not a paradox or wordplay. It identifies a real feature of certain concepts: they are co-arising, co-dependent. The agent is constituted by the action; the action is constituted by its being *someone's* action. The very same structure is applied to seeing and the seer: the seer exists only insofar as there is a seeing; the seeing exists only insofar as there is a seer. Therefore the seer cannot be the metaphysical *ground* of seeing — it is a conceptual postulation that gives seeing its grammatical form ("someone is seeing"), nothing more.
1.4 Four lines of argument against the seer
(a) Seeing cannot see itself.
見不能有見，非見亦不見。
If seeing were already a self-standing thing, it ought to be able to see by itself — but it cannot. The function of seeing is to see *something other than itself*; it has no self-luminosity. A non-seeing (a stone, a closed eye) certainly doesn't see. So neither seeing-as-already-established nor non-seeing can serve as the seer. If the seeing function cannot establish itself as the seer, the seer as a separate subject has nothing to attach to.
(b) Whether identified with seeing or separate from it, the seer is unfindable.
離見不離見，見者不可得。
This is the central verse. Two scenarios:
- *If the seer is identical to the seeing*, then the seer just *is* the seeing function. But the seeing function has no "seer-character" — it is a perceptual process, not a person. There is no extra face to it that constitutes "the one who sees." - *If the seer is separate from the seeing*, then a person without the seeing function (e.g., someone born without eyes, or with the visual cortex damaged) should still possess a seer, and could perhaps see by other means. But this contradicts observable fact.
Either way the seer is not findable. And since the seer is not findable, the seeing function loses its "owner," and the visible object loses its "owner-relative-to-whom-it-is-visible."
(c) The unity / multiplicity dilemma.
If we assume one unified seer stands behind all senses, this single seer must also be the hearer, the feeler, the thinker. The text presses the consequence: if it is truly one, the senses should be interchangeable — the eye should be able to hear, the nose should be able to see. They are not.
Conversely, if each sense has its own independent seer, then while you see, a separate "hearer-seer" should also be operating; while you hear, a separate "see-seer" should also be operating. But experience does not present itself as multiple parallel subjectivities; nor do we ever encounter several "spirits" inside one chest.
Both the unified-subject and the many-subjects assumptions collapse under their own consequences.
(d) The three things never meet — and knowledge has no time.
Now the analysis becomes almost geometric. Suppose:
- 見 = the eye, located at A; - 可見 = the visible form, located at B; - 見者 = the self, located at C — perhaps behind the eyes, perhaps nowhere physical.
*If they are in different locations*, they never meet, and no seeing ever occurs.
*If they meet*, then in every location where there is a visible form (B) there must also be an eye (A) and a self (C). This contradicts fact: there are colors in the room where I am not.
*If they do not meet*, then a seer in location C should be able to see a form in location B without the eye in A being involved — through walls, across distances. This also contradicts fact.
Finally the text checks when "knowing" (知) occurs relative to "seeing" (見) — before, after, or simultaneously:
- Knowing-before-seeing: one would know what one has not seen. - Knowing-after-seeing: the seen object would have to persist as a known object after the seeing. - Knowing-simultaneous: requires temporal coincidence of two events whose separate existence has not been shown.
Each case fails. The conclusion: 知無故 — since "knowing" has no fixed locus, 見、可見、見者同樣無有定相可得 — seeing, visible, and seer alike have no fixed character that can be pinned down.
1.5 Where this sits in the path of liberation
The *Mūlamadhyamakakārikā* is not denying that there are visual experiences, that eyes function, or that people act and perceive. It is undoing the structural assumption that there is a subject-owner behind these processes. This undoing is not theoretical entertainment; it is liberative.
Because every form of suffering in Buddhist analysis — craving, clinging, the sense of "me and mine," pride, jealousy, fear of death — is built on the assumption: *there is a "me" that owns, that loses, that is threatened, that needs protecting*. If the structure that posits this "me" is exposed as having no fixed referent — if the seer cannot be located in the analysis — then the foundations of *me-and-mine* begin to loosen. This is not annihilation. It is the discovery that what we took to be a solid owner was always already a conceptual postulation.
2. Life examination: walking through the concept in contemporary scenes
Let me map the three-fold structure onto familiar situations — and flag where the default mappings tend to go wrong.
Scene A — Scrolling at midnight
You're in bed, scrolling through images on a phone. Three things are present:
- 可見: the actual visual data — faces, products, text, color gradients, animations streaming across the screen. - 見: the visual processing — light hitting the retina, signals moving through the optic nerve, the visual cortex assembling edges, motion, recognition. - 見者: the *you* that feels like it's "behind" the eyes, watching, evaluating, being moved.
The default mapping. Most people identify 見者 with the brain, or with "consciousness," or with "the mind." This is already a reification. The classical analysis is not asking "where in the brain is the seer located?" It is asking whether the *structural assumption* of a seer is even coherent.
Walking the argument through. - "The seer is identical to the seeing" → in this scene, that would mean the visual processing itself *is* "you." But visual processing has no character of "me-ness"; it is a process. You don't feel the retinal firing as "you" — you feel the *content* of what you see. - "The seer is separate from the seeing" → that would mean there is a "you" who can see *without* visual processing — perhaps by reading the meaning behind the screen directly. But when your eyes are closed, you can't see the screen; when the visual cortex is damaged, the screen becomes invisible. - "The three never meet" → the data is in the phone (B); the processing is in your brain (A); the "you" feels like it's somewhere in the middle of your head — or perhaps nowhere physical at all. Where, exactly, is the seer?
Where the walk leads. The sense that "I am watching this" is *real as an experience*, but the assumption that there is *a thing* doing the watching is what the analysis is unpacking. "I am watching" is closer to a grammatical construction ("it is raining") than to a report about an entity.
Scene B — The job interview
You walk into a room. Three things again:
- 可見: the interviewer's face, body language, the office furniture, the light through the window. - 見: your visual processing, your reading of expressions, your monitoring of the room. - 見者: the assumed "I" — the one being perceived, the one who must perform well, the one who must "show up."
The default mapping. In anxious moments, people often feel the 見者 as a kind of *inner homunculus* — a small version of themselves standing behind the eyes, observing and being observed.
Walking the argument through. - The "one seer" assumption: in the interview, you feel there is one "you" that has to be coherent across multiple channels — you have to *look* confident, *sound* confident, *feel* confident. This is the 本住 (originally-dwelling) assumption in action: a unified subject that must be present consistently across all senses. But notice: your visual self, your auditory self, your proprioceptive self are not coordinated by a single "you" — they are coordinated by habit, training, and contingencies. The sense of a unified "you" is a *narrative overlay*, not a structural fact. - The "many seers" assumption: at any moment, multiple sub-personalities are running — the part that wants to impress, the part that wants to flee, the part that is calculating word choice. These are not separate "spirits" in the metaphysical sense, but the assumption that they must be unified under one owner creates pressure and stress.
Where the walk leads. The demand for *self-coherence* in performance situations comes from the 見者 assumption. If there is no fixed owner of these processes, the pressure to be a "unified self" begins to relax. You can show up *as a process* rather than as a defended owner of a process.
Scene C — Grief after a loss
When someone close to you dies, the experience of "I" undergoes a strange transformation. The classical analysis helps here in an unexpected way.
- 可見: the absence — the empty chair, the unsent message, the photograph. - 見: the perception itself — eyes still functioning, still registering the world. - 見者: but now the "seer" feels fragmented. "Who am I now that they are gone?" The 見者 that was partly defined in relation to the lost person seems to have lost its coherence.
The default mapping. People in grief often experience this as the *destruction* of the self — "I'm not myself anymore." This is felt as catastrophic.
Walking the argument through. The classical analysis suggests that the 見者 was *never* a unified, self-standing thing to begin with. It was always constituted by its relations — to the people around it, to the roles it played, to the body that housed it. Grief doesn't *destroy* a self; it reveals that the self was always *relationally constituted*. This is not an additional blow on top of the loss — it can also be a profound release from the assumption that you must always be the same "you" you were before.
Where the walk leads. The 見者 analysis is not just an abstract exercise. It is a way of meeting moments when the assumed self-structure becomes unstable — and discovering that instability is not annihilation.
Scene D — The "I" that watches thoughts in meditation
In meditation, many practitioners encounter the sense that "there is an I that is watching my thoughts." This is often mistaken *for* mindfulness, but it is actually the 見者 assumption appearing in subtle form.
The default mapping. "I am the observer of my thoughts. This is the witnessing self. This is liberation." This is the 見者 in its most refined form — a seer behind the seeing of thoughts.
Walking the argument through. Where is this "witnessing self"? Is it a thought? If so, it is not witnessing thoughts — it *is* a thought. Is it separate from thoughts? Then it must always be present, even in deep sleep, even in unconsciousness — but in those states there is no witnessing. Is it identical to awareness? Then "awareness" is doing the witnessing — but awareness itself is a process, and processes do not have owners.
Where the walk leads. What is often celebrated in some meditation circles as "the witness" is precisely what the classical analysis is pointing at as a structural assumption that needs to be examined, not affirmed. Genuine release is not finding a deeper seer; it is the exhaustion of the seer-assumption.
Common mapping errors in contemporary language
- "The seer is consciousness." Consciousness is itself a process; the analysis applies to consciousness-as-seer just as it does to eye-as-seer. There is no consciousness that stands behind consciousness. - "No seer = no person = no moral responsibility." The analysis denies the *metaphysical assumption* of a fixed seer; it does not deny the functional, relational, karmic continuity that allows moral discourse. A wave is not a separate "thing" from the ocean — but it can still damage the shore, and we still respond to it as a wave. - "The seer is the brain." Neuroscience can describe visual processing; it cannot locate a seer in the processing. The contemporary binding problem — how distributed neural events become unified experience — is the modern echo of the classical 見者 problem: how do these processes come to feel like *mine*? The classical answer is: they don't — there is no unifier to be found, only the assumption that there must be one. - "No seer = I don't exist." This is nihilism, and it is the *opposite* of what the analysis is doing. The analysis is not denying the experience of "I"; it is examining the *assumption* that this experience points to a self-standing thing. The experience remains; the metaphysical assumption is what's being investigated.
3. Why contemporary people need this
3.1 The exhaustion of selfhood
Contemporary life, especially in digitally saturated environments, asks the self to be many things at once: a personal brand, a professional role, a partner, a parent, a consumer, a citizen, a self-care practitioner, an inner-child, a future-projected identity. Each of these demands *seer-like coherence*: a "you" that owns them all, that decides among them, that maintains continuity across contexts.
This is exhausting in part because the underlying assumption — that there is such a "you" to be coherent — was never a stable foundation to begin with. The classical analysis of 見者 is not asking people to abandon their identities; it is offering a way of *relaxing the grip* on the assumption that any of these identities point to a real, unified owner.
3.2 The mental-health connection (with caveats)
Contemporary psychology and psychiatry often work with models of the self — narrative self, minimal self, autobiographical self, self-as-process. Some of these models are remarkably close to the classical analysis; others implicitly reify a "core self" that the classical analysis would question.
Where the parallels illuminate: when therapeutic work helps someone notice that the "I" who is anxious is not the same as the "I" who is observing the anxiety, this is *already* a movement in the direction of the 見者 analysis. The classical analysis goes further: it asks whether there is any "I" at all — even the observing one.
Where the parallels must be handled with care: neuroscience cannot prove or disprove the classical analysis. The fact that the brain shows "default mode network" activity during self-referential thought is interesting, but it does not adjudicate the philosophical question of whether the self being referred to is a real entity. The classical analysis is a logical and phenomenological investigation, not a scientific hypothesis. Studies of neural correlates *describe* the conditions under which the seer-assumption arises; they do not *establish* that there is a seer.
3.3 The "homunculus" in philosophy of mind
Philosophers of mind have long recognized the homunculus problem: if perception requires an inner observer, who observes the inner observer? Who observes *them*? The regress never terminates. Contemporary theorists like Daniel Dennett have argued that there is no central observer — only "multiple drafts" of experience, no unified witness.
The classical analysis of 見者 is doing something structurally similar, but with a different goal: not just describing how the brain works, but dissolving the assumption that there must be a witness. The philosophical parallel is striking, but the classical project is *liberative* in a way that philosophy of mind typically is not. The classical analysis is not trying to describe the mind better; it is trying to undo a specific structural assumption that is producing suffering.
3.4 The AI parallel — and its limits
Modern AI systems often present a kind of "user-facing self": an interface that presents itself as a unified, coherent agent even though the underlying computation is distributed and fragmentary. Users interact with "the AI" as if it were a single thing, even though what they are interacting with is a complex web of processes.
The classical analysis of 見者 applies a similar suspicion to the human "interface": the sense that there is a unified "I" behind the eyes may be a comparable user-facing construction, not a metaphysical fact.
But this parallel must not be pushed too far. Humans and AI systems are not the same kind of thing. AI systems don't (currently) suffer from the assumption of being a unified self — they don't crave, cling, or fear the loss of selfhood. The parallel illuminates the *structure* of the 見者 assumption; it does not collapse human experience into machine processing, nor does it let us claim that "AI research has confirmed Buddhist philosophy."
4. Common misreadings and clarifications
4.1 "Nāgārjuna says there is no self"
True but misleading. Nāgārjuna is not denying that there is a person, that there is continuity, that there is functional behavior. He is denying a very specific assumption: that there is a *self-standing, unified, pre-existing subject* that owns the senses and the actions. The person-functions remain; the metaphysical owner is what's being investigated.
A useful analogy: when physicists say there is no "solid matter" at the subatomic level, they don't mean your chair disappears. The chair's functional reality remains; what is rejected is the *metaphysical assumption* of solid, atom-like substance underneath. Similarly, the person-functions remain; what is rejected is the *metaphysical assumption* of a seer underneath.
4.2 "見者 = ātman, and Buddhism rejects ātman"
Yes and no. The 見者 is functionally equivalent to what other traditions call ātman, puruṣa, or jīva — a unified subject. And Buddhism does reject this. But the rejection is *not* of personal existence or moral responsibility. It is the rejection of a specific structural assumption. Many popular explanations of "no-self" (無我) flatten this into either nihilism ("nothing exists") or essentialism ("there is a true self, just not the false one"). The 見者 analysis avoids both.
4.3 "No seer means I am not responsible for my actions"
This is a common misunderstanding, with a long history in Indian philosophical debates. The classical reply: moral responsibility is a *functional* matter, not a *metaphysical* one. The wave damages the shore; we respond to the wave. The fact that there is no fixed "owner" of the wave doesn't mean the wave has no causal efficacy. Similarly, the fact that there is no fixed "owner" of actions doesn't mean the actions have no consequences. The analysis cuts the metaphysical ground from under *the owner*; it does not cut the causal ground from under *the behavior*.
4.4 "This is just a clever logical trick"
The arguments in the *Mūlamadhyamakakārikā* can *look* like mere puzzles. But they are designed to expose a deeply embedded structural assumption. The assumption of a seer is not something most people have ever explicitly formulated — it operates pre-reflectively, in the grammar of every sentence that begins "I…". The arguments are tools for bringing this pre-reflective assumption into reflective awareness, so that it can be examined rather than simply acted on. The "trick," if there is one, is that the assumption is so structural that once it is seen, the world looks different without it having changed at all.
4.5 "The 'real self' is the seer"
Some contemplative traditions — including some that claim Buddhist lineage — hold that beneath the constructed self there is a "true self" or "witnessing awareness" that is the seer. The classical analysis of 見者 explicitly rejects this move. There is no deeper seer to be found; the seer-assumption *is* the thing being investigated, and it does not have a hidden referent. To install a "true witness" behind the analysis is to bring back exactly the kind of unified, pre-existing subject that the *Mūlamadhyamakakārikā* has spent this section dismantling.
4.6 "This is epistemology, not liberation"
It is easy to read the *Mūlamadhyamakakārikā* as a work of epistemology or philosophy of language. But its project is soteriological — aimed at undoing the structural conditions of suffering. The 見者 analysis is part of this because the assumption of a self-owner is what makes craving and clinging feel like *my* cravings and *my* clingings. When the seer-assumption is loosened, the affective grip of "me and mine" begins to relax. This is not a theoretical side-effect; it is the whole point.
4.7 The wider Madhyamaka frame
The 見者 analysis is one instance of a structural pattern that runs through the entire *Mūlamadhyamakakārikā*: every posited entity — seer, time, motion, causation, self — is examined via mutual conditioning. The pattern applies across the whole text, to motion, to the self, to causation, to nirvāṇa itself. The pattern is consistent: don't accept the seer as self-standing; don't accept it as nonexistent; don't accept it as both; don't accept it as neither. Instead, examine the structure that generates the assumption, and let the structure come to rest.
Closing reflection
The 見者 analysis is not a demolition; it is a decompression. It takes a structure we have been unconsciously running on — *there is a me who is seeing, hearing, feeling, choosing, owning* — and makes it available for inspection. Once we see the structure, we don't have to keep running on it.
The visual processing continues. The visible world continues to appear. The functional person continues to act and respond. What changes is the grip: the assumption that there is a central owner of all of this begins to relax, and with it, a particular kind of suffering — the suffering of having to be a unified self at all costs — begins to ease.
This is not a small thing. It is, in fact, a very large thing — but it happens in very small ways: in the moment you notice that the "you" who is anxious about the future is not the same process as the "you" who is breathing right now; in the moment grief reveals that the self was always constituted by its relations; in the moment meditation shows that the "witness" is itself a thought. These are not losses. They are relaxations of an assumption that was producing suffering disproportionate to its reality.
That relaxation — the slow, ordinary, ongoing one — is what the *Mūlamadhyamakakārikā* is after, in this section, under the name of 見者.