去者

What This Concept Is About

Imagine you are watching someone walk across a room. Before any analysis begins, two things present themselves to ordinary perception: the *act* of walking, and the *person* who walks. We do not experience the walking as floating free of a walker; we do not experience the walker as divorced from the act. They appear to fit together like a hand and a glove. The hand moves, the glove moves with it. The walker walks, the walking belongs to the walker.

The Mūlamadhyamakakārikā (《中论》, "Fundamental Verses on the Middle Way," Treatise T1564), in its famous second chapter — 观去来品 (Guān Qù Lái Pǐn), "Examination of Going and Coming" — accepts this everyday observation as a starting point, then takes it apart. It does not deny that there appears to be a walker, or that there appears to be a walking. What it denies is that either of these, or the pair of them together, can be pinned down as having any self-standing (自性) existence. The chapter's central device is to set up a tiny pair — 去法 (qù fǎ, "the going") and 去者 (qù zhě, "the goer") — and to show that they hold each other up like two mirrors facing each other: each appears to be reflected in the other, but if you look for the original source of the reflection, you cannot find one.

The "goer" is not a Buddhist-fantasy invention. It is the everyday assumption that there is a *subject* who *does* the going, an agent standing behind the action, an "I" or a "self" who walks. The classical argument simply asks: when you look very carefully at this supposed goer, where is it? Try to locate it.

The chapter pursues this question through five interlocking analytical gates, sometimes called the 五门 (wǔ mén, "five gates") of the chapter:

First gate: 三时中求 (sān shí zhōng qiú, "searching across the three times"). Divide time into past, future, and present, and ask where the goer lives. In the past — already gone — there is only a gone. The going is over; if a goer existed there, it would have to be the *already gone*, but "already gone" already implies the going has finished. In the future — not yet gone — there is no going yet at all, so no goer to be found. In the present — the moment of going — the chapter asks the opponent to locate the goer here, and the opponent proposes: "There is going at the place where there is motion; this is where the going-time is; therefore at the time of going, there is going." The reply: this would yield *two* goings and *two* goers in the same moment, but in a single goer you cannot have two goings. The conclusion is delivered as a classical verse:

*"已去无有去，未去亦无去，去时名半去半未去，不离已去未去。是故于三时，定无有去者。"* ("In the already-gone there is no going; in the not-yet-gone there is no going; the moment of going is called half-gone and half-not-gone, not apart from the already-gone and not-yet-gone. Therefore across the three times, definitively, no goer can be found.")

So the first gate shakes the goer loose from time. It cannot be found in any time-slice.

Second gate: 去不去求 (qù bù qù qiú, "searching via going and not-going"). The chapter then asks: is the goer the one who goes, or the one who does not go? The one who goes is not at rest, because the going has not ceased. The one who does not go is not at rest either, because rest (住) only arises when the going has ceased, and without going there is no rest to define. Outside of these two, there is no third category of "goer" waiting to be discovered. The verse runs:

*"去者则不去，不去者不去；离去不去者，无第三去者。"* ("The goer does not go; the non-goer does not go; apart from goer and non-goer, there is no third goer.")

So the goer cannot be located in the dichotomy of "currently going" / "currently not going" either.

Third gate: 一异门 (yī yì mén, "the identity-and-difference gate"). Perhaps the goer and the going are the same thing? But if the agent and the action were identical, then there would be no distinction between the one who does and what is done — every doer would be its own deed, and we would lose the grammar of action entirely. Perhaps they are different? But if the going is one thing and the goer is another, then the going could exist without any goer (a walking with no walker) or a goer could exist with no going (a walker who never walks). Both are absurd. The classical verse:

*"去法即去者，是事则不然；去法异去者，是事亦不然。"* ("That the going should be the goer — this cannot be; that the going should be different from the goer — this also cannot be.")

Fourth gate: 作不作门 (zuò bù zuò mén, "the doing-and-not-doing gate"). Can the goer *use* the going? The verse says no:

*"因去知去者，不能用是去；先无有去法，故无去者去。"* ("From the going one knows the goer, yet the goer cannot use this going; with no going existing prior, there is no goer that goes.")

The point is sharper than it first sounds. To "use" the going, the goer would have to stand over and against it as an already-existing tool. But the goer is constituted *by* the going — we only designate something as a "goer" because there is going. So the going cannot be the goer's instrument, because the goer is not prior to the going. There is no temporal ladder on which a pre-existing goer picks up a going like a carpenter picks up a saw.

Fifth gate: 决定不决定门 (jué dìng bù jué dìng mén, "the determinate-and-indeterminate gate"). Maybe we can rescue the goer by declaring it either inherently existent (决定, "determinate") or only nominally existent (不决定, "indeterminate"). Either way, the chapter says, the goer cannot use the three modes of going (未去/已去/去时, future/past/present). The verse:

*"决定有去者，不能用三去；不决定有去者，亦不用三去。"* ("A determinate goer cannot use the three goings; an indeterminate goer also cannot use the three goings.")

A goer declared inherently existent cannot be dependent on the going to be a goer — but then it would have to be a goer even when there is no going, which contradicts how we identify goers. A goer declared merely nominal cannot ground the three goings because it has no independent reality to anchor them. Neither option works.

The concluding verse ties the whole chapter together, and it is one of the most quoted passages in the Mūlamadhyamakakārikā:

*"不得言定有、不得言定无，是故决定知，三法虚妄，空无所有，但有假名，如幻如化。"* ("It cannot be said to definitively exist; it cannot be said to definitively not exist. Therefore it is decisively known: the three dharmas are illusory, empty of inherent existence, possessing only conventional designations, like an illusion, like a magical transformation.")

Here 三法 (sān fǎ, "three dharmas") refers to 去法 (the going), 去者 (the goer), and 所去处 (suǒ qù chù, "the destination / the place gone to") — three concepts we use every day without thinking, three concepts that turn out to be mutually dependent like the corners of a triangle: remove one, and the other two cannot stand.

Where does this sit in the broader path of liberation? The 观去来品 is one of the most famous early chapters of the Mūlamadhyamakakārikā because it does something very specific: it takes one of the most ordinary, unexamined experiences — bodily movement from one place to another — and shows that even here, the architecture of *agent + action + object* is not solid. If walking cannot yield a findable goer, what about the much more tempting "goer" — the person, the self, the "I" who is said to walk through life? The chapter is a small, clean proof-of-concept that the same kind of analysis, scaled up, applies everywhere. It is a hand-drawn prototype of the larger claim of dependent origination (缘起) and emptiness (空): that *nothing stands alone*. Liberation begins not with denying the appearance of the goer, but with seeing through the assumption that the appearance has a self-standing owner behind it.

Walk-Through in Daily Life

Let's take something almost everyone does daily — a commute to work — and walk through it the way the 观去来品 walks through "going."

Scene: You are standing at the subway platform. The train arrives. You step on, sit down, and forty minutes later step off at your stop and walk into your office.

Step one: The moment-by-moment experience of "I am going."

You say to yourself, or feel, "I am going to work." Now, where is the "I" in that sentence? Try three locations.

*In the past* — last night, when you were brushing your teeth and the going to work had not yet begun. Were you already a "goer-to-work"? Not really. The going had not started; you were just brushing. You are not yet a commuter in the future tense.

*In the future* — tomorrow, when the going-to-work will already be over and you'll be brushing your teeth again. Were you still a "goer-to-work" in tomorrow's retrospect? No, by then the role is finished. You have become a "had-gone-to-work."

*In the present* — the actual moment on the train. The train is moving, you are inside it, and you feel the going is happening *right now*. But "right now" — is that a slice? A second? A millisecond? If you slice finely enough, the moment of going is already made of past-and-future pieces. The chapter's formula 半去半未去 (bàn qù bàn wèi qù, "half-gone, half-not-gone") is uncomfortably accurate the more carefully you look at "now."

So the first gate is doing its work: the commuter-as-goer is not locatable in any of the three time slots, even though the commuting is clearly happening.

Step two: Are you "going" or "not going"?

The chapter says: neither.

When you are physically on the moving train, you are *not* going — your legs are still, you are sitting, your body is not performing the act of walking. (In fact, for most of a commute, you are *not* going in the strict sense of bodily movement.) When you are standing still on the platform waiting, you are *also* not going. There is no third category — "going-while-not-going" — that rescues the goer.

And this matches lived experience. You do not feel like a "goer" for the entire forty-minute commute. You feel like a goer for perhaps the part where you walk down the stairs and across the platform. Most of the rest of the time, you are a "non-goer" — sitting, scrolling, waiting. Yet we casually call the whole trip "going to work," as if the going-ness were a continuous property of "you" across the whole interval. The chapter is showing that this continuous-property assumption doesn't hold up.

Step three: Is the "you" the same thing as the going?

Try an experiment. Tomorrow morning, deliberately do not go to work. Stay in bed. Now: are you still "you"? Yes. Did the "going to work" happen? No.

If the goer *were* the going, then when the going is absent, the goer should be absent too. But you are still there, brushing your teeth, working from home perhaps, being the same person in some sense. So the goer and the going are not the same.

Try the opposite experiment. Imagine a perfect clone of you, identical in every way, who does go to work tomorrow. Does the *going* automatically come along with the clone? Only because there is a walker doing the walking. The going requires a walker, and that walker is *this* walker, not just any body. So the going is not floating free of all walkers either; it is not wholly different from the goer.

The goer and the going are *neither identical nor different*. They fit together, but the fit is not the fit of two halves of the same thing, nor the fit of two independent things. They fit like — to borrow a later Buddhist image, used in Yogācāra and Madhyamaka literature — a reflection fits a face.

Step four: Who is using the commute?

If the goer is constituted by the going (you are a commuter *because* you commute), then the commute cannot be the goer's tool. You did not pick up "commuting" from a shelf and apply it to yourself; "commuting" is the activity through which "commuter" gets defined. In the classical language: 因去知去者 (yīn qù zhī qù zhě, "from the going one knows the goer") — the goer is recognized *through* the going, not the other way around.

This is a strange inversion of how we usually narrate our lives. We say, "I (the real me, the one with preferences, habits, a face, a history) chose to commute." The chapter is gently pointing out: the "I" that chose to commute is itself a figure that crystallizes around the act of commuting. If you've never commuted, never traveled, never moved from one place to another in any sustained way, the word "commuter" is meaningless for you. The goer is what the going produces as its own shadow.

Step five: Where the popular mapping goes wrong.

When contemporary readers first encounter the goer analysis, they often assume Nāgārjuna is doing one of the following:

- *"He's saying nobody really goes anywhere."* No. People do travel. The argument is not that movement is an illusion in the sense that the train is a hallucination. The train is conventional reality (世俗, *世俗 = "the conventional"). What lacks 自性 (zì xìng, "inherent self-nature") is the goer *as an independently existing agent*. - *"He's denying free will."* Not quite. Free will as ordinarily understood (a self that stands behind choices) is precisely what the chapter calls into question. But this is not a libertarian-versus-determinist debate in the modern sense. It is a different kind of inquiry: *when we posit a self that owns actions, can we find this self?* - *"He's saying the self is fake."* Closer, but again imprecise. The Mūlamadhyamakakārikā's move is not to say "there is no goer" but to say "the goer cannot be found when searched for, therefore the goer has no inherent existence, therefore the goer exists only as a conventional designation." The goer is not *fake* — the goer is *empty* (空). This is the difference between saying a reflection is "fake" and saying a reflection has no independent face.

This walk-through can be repeated for almost any movement-involving human activity: cooking dinner, scrolling a phone (the thumb "goes," the eye "goes" across the screen), driving, walking the dog, even typing. Each time, the same five gates apply: the doer is not findable in past/future/present, neither in the doing nor the not-doing, neither identical with nor different from the action, and is not using the action as a pre-existing tool.

Why Contemporary People Need This

There is a particular modern affliction that the 去者 analysis addresses with surgical precision: the experience of being the constant owner of one's actions. The contemporary self is often narrated as an internal executive — a "real me" inside the body who makes decisions, executes them, accumulates the results, and bears the credit or the blame. This executive is rarely examined; it is taken as the default furniture of identity. The chapter's argument is that this executive, examined carefully, cannot be located.

Why does this matter for contemporary life?

1. The weight of self-ownership is enormous.

When someone fails at a task — misses a deadline, loses a relationship, makes a bad investment — the "executive self" model says: *you* did this, *your* mistake, *your* failure, *you* are the kind of person who does this. The model loads the entire outcome onto a single, durable, locatable agent: the goer. Contemporary psychology has begun to notice that this loading contributes to rumination, shame, and what is sometimes called "self-condemnation." The goer analysis does not deny the failure happened; it loosens the assumption that there is a fixed "you" standing behind it as its owner. Failures still happen; mistakes still have consequences. But the goer — the inner executive whose job it is to carry the mistake forever — turns out, on close inspection, not to be a findable thing.

2. The AI-and-algorithm parallel is instructive but should be handled carefully.

When a contemporary person sees an AI assistant execute a task — write an email, plan a route, generate an image — there is a moment of confusing judgment: *who did this?* Was it the AI? Was it the human who prompted it? Was it the company that built the AI? Was it the data the AI was trained on? The same structural puzzle appears. There is a chain of mutually-conditioning factors (prompt, model, weights, server, training data, the human's intention, the company's design choices), and yet we still want to point to one of these factors and say *that* one is "the agent." The chapter's move — to deny a findable singular agent within a network of mutual dependence — is exactly the kind of philosophical move that contemporary discussions of AI authorship, accountability, and credit are slowly circling around. But the analogy is a two-way street. Just as the Madhyamaka analysis can illuminate AI debates, the AI debate can illuminate Madhyamaka: the puzzlement people feel about AI authorship is a high-tech restatement of the very puzzlement the chapter raises about goers. *We should not claim, however, that artificial neural networks "prove" Buddhist philosophy.* What the analogy shows is that the conceptual puzzle about agency is genuinely hard, not that the chapter's solution has been empirically validated by machine learning. The Buddha's analysis stands on its own argumentative ground.

3. The fragmented attention of contemporary life reveals the goer-analysis in action.

A contemporary person rarely "does one thing at a time" in the way the classical model of an agent presupposes. Phones flicker, calendars overlap, a parent is simultaneously texting, cooking, listening for a child, mentally composing a work email. There is no single, coherent "goer" running these activities; the activities are running themselves, and what we call "I" is a label that holds them together after the fact. The chapter's description of the 半去半未去 (half-gone, half-not-gone) present moment is almost a literal description of contemporary multitasking: each micro-activity is partly already started and partly not yet finished, and the "agent" is the provisional name we paste across all of them. Practically, this loosens the pressure to be a perfectly coherent, continuous self at every moment. The "I" that shows up in different contexts — at work, at home, with friends, alone — need not be one single substance to be real.

4. The analysis is liberating precisely because it is not nihilistic.

This is the most important practical point. The chapter's conclusion — *"三法虚妄，空无所有，但有假名，如幻如化"* (the three dharmas are illusory, empty of any self-existence, possessing only conventional designations, like an illusion, like a magical transformation) — could sound like a counsel of despair: *nothing is real, why bother?* But the same verse continues by establishing that the three dharmas are not "definitely non-existent" either. They are 空 (kōng, "empty"), which in Madhyamaka is not "non-existent" but "empty of independent self-existence." The goer, conventionally, works. You can still catch the train. You can still take responsibility for what you do. But you do so without the metaphysical load of positing an unchanging owner behind your actions. This combination — functioning fully in conventional reality while seeing through the pretensions of inherent self-existence — is the practical heart of the path of liberation.

Common Misreadings and Clarifications

Misreading 1: "The chapter denies that anyone ever goes anywhere."

This confuses two levels of discourse. The 胜义谛 (shèng yì dì, "ultimate truth") analysis shows that the goer has no inherent self-existence. The 世俗谛 (shì sú dì, "conventional truth") still includes people walking, trains moving, lives being lived. A later Madhyamaka formulation, found in commentaries by Bhāvaviveka, Candrakīrti, and Tsongkhapa, makes this explicit: things are empty *of* inherent existence, not *as* inherently non-existent. To say "no one goes" in the sense of denying the train ride is to commit the error of 恶取空 (è qǔ kōng, "wrongly grasping emptiness") — taking emptiness as a positive denial of everything, rather than as the absence of self-existence. Candrakīrti's *Madhyamakāvatāra* and *Prasannapadā* devote considerable attention to warning against this collapse.

Misreading 2: "去者 is the same as the self in the five-aggregates teaching."

There is a family resemblance, but the analyses are distinct and operate at different levels. The 五蕴 (wǔ yùn, "five aggregates") teaching — most famously in the 《般若波罗蜜多心经》 (Prajñāpāramitā Hṛdaya Sūtra, "Heart Sūtra") with the line "照见五蕴皆空" (zhào jiàn wǔ yùn jiē kōng, "clearly seeing that the five aggregates are all empty") — analyzes the self as a compound of form (色, *rūpa*), feeling (受, *vedanā*), perception (想, *saṃjñā*), formations (行, *saṃskāra*), and consciousness (识, *vijñāna*). The 去者 analysis is sharper and more localized: it takes a single ordinary experience (going) and unpacks the agent–action–object structure within it. The two teachings are complementary, not redundant. The five-aggregates teaching tells you what the self *is composed of* (and fails to find); the 观去来品 tells you that even in one tiny slice of activity, the *agent* cannot be located.

A second relevant classical expression: the 《金刚经》 (Vajracchedikā Prajñāpāramitā Sūtra, "Diamond Sūtra") contains the famous line "应无所住而生其心" (yīng wú suǒ zhù ér shēng qí xīn, "the mind should be abode-less yet still arise"). This is a different but kin teaching: just as the goer has no fixed abode, the mind that arises should not be anchored to a fixed self or object. The two teachings — one about the agent of going, one about the dwelling of mind — both press toward the same insight: nothing stands alone, nothing is owned, nothing abides in a fixed place.

Misreading 3: "去者 is the same as 行 in dependent origination."

This is a subtle and important distinction. The source material explicitly flags it. In the 十二因缘 (shí èr yīn yuán, "twelve links of dependent origination"), 行 (xíng, "formations" or "volitional activity") appears as the second link — 无明缘行 (wú míng yuán xíng, "from ignorance as condition, formations arise") — and refers to volitional impulses that drive the continuity of saṃsāra. The chapter 观去来品 uses 行 in a different sense: the *continuity* from one moment to the next, like a seed developing into a sprout, stem, leaf. 去 (qù, "going"), by contrast, refers specifically to 身动 (shēn dòng, "bodily movement") — the literal physical act of moving from one place to another. The two concepts are kin (both involve continuity or movement) but distinct: 行 is the continuity of karmic patterning through time; 去 is the spatial displacement of a body. Confusing them is a common error in beginners' readings.

Misreading 4: "The chapter is just abstract logic, not practice."

This is perhaps the most harmful misreading, because it relegates one of the most practically transformative chapters of Madhyamaka to the philosophy classroom. In fact, the 观去来品 sits squarely in the practice of 观 (guān, "contemplation / vipaśyanā") — the deep seeing that is one half of the classical pair 止观 (zhǐ guān, "śamatha and vipaśyanā, calm and insight"). The exercise of looking for the goer in one's own movement — while walking, while reaching for a cup, while typing — is a vipaśyanā practice. The meditator uses the body as a laboratory. Each step becomes a question: *where is the goer in this step?* Each time the search comes up empty, the meditator is not gaining an abstract proposition; they are loosening the habitual grip of agent-ownership on lived experience. This loosening, repeated, is what the tradition calls the maturation of insight.

Misreading 5: "If there is no goer, no one is responsible for their actions."

This is a classical objection — and the Buddhist tradition has a well-developed response. The 观去来品 denies a findable goer, not the functioning of the conventional person. Responsibility operates at the conventional level. The person who commits an action experiences the consequences of that action; the relational, social, and karmic web of cause-and-effect still functions. The emptiness of the goer does not exempt anyone from the consequences of their actions — it exempts no one, because "anyone" is itself part of the conventional framework. As Tsongkhapa's *Lam rim chen mo* and Candrakīrti's *Madhyamakāvatāra* both emphasize, the emptiness of the agent does not collapse moral responsibility; it grounds moral responsibility differently. We are responsible not because there is a permanent self who owns the action, but because actions produce effects within a continuum, and the continuum is itself empty of self-nature. To use the chapter's own example metaphorically: the 三法 (three dharmas — going, goer, destination) are like an illusion — and like an illusion, they still appear, they still function, they still have consequences. The magic show is "empty," but the person who walks onto the stage thinking the magician's assistant is about to be sawn in half is still going to gasp.

Misreading 6: "This is just Buddhist idealist denial of the external world."

Madhyamaka is not Yogācāra (consciousness-only). The Mūlamadhyamakakārikā does not deny external objects; it analyzes how *anything we posit* — external objects, internal subjects, the relation between them — cannot be established as inherently existing. The 所去处 (suǒ qù chù, "destination") is included alongside the goer and the going in the 三法 (three dharmas). The destination, too, is empty of inherent existence — not because places don't exist, but because the place-as-destination is itself constituted through the going. There is no "absolute destination" waiting somewhere outside all movement; the destination is the shape that movement carves out as it moves. The chapter's universe is not idealism, not realism, not nihilism — it is the thoroughgoing analysis of mutual dependence.

A closing reflection

The 观去来品 is, in the end, a quiet chapter. It does not thunder. It does not proclaim metaphysical absolutes. It simply asks, with a patience that has been teaching for two millennia, the same question over and over: *where is the goer?* And each time we look, with however much care, the goer is not there — not as absence, but as something that was never there in the way we assumed. What remains is the going, the goer as conventional name, and the destination as conventional name — 三法虚妄，空无所有，但有假名，如幻如化 — three empty, conventional designations, like an illusion, like a magical transformation, into which we step every morning when we step onto the train and become, for the duration of the commute, *someone going somewhere*.