What This Concept Is Saying
There is a moment almost everyone has experienced: you look into a polished surface and see "yourself," yet the figure in the glass is not you. It has your shape, but no bones, no breath, no tomorrow. It is a *pratibimba* — a "reflected image." The Buddhist tradition takes this ordinary phenomenon and elevates it into one of the most precise tools it possesses: a tool for understanding both how the meditated mind works, and how all phenomena — including you, the screen, the relationship, the career — can appear so vividly while lacking any underlying self-nature (*svabhāva*).
The classical definition is sharp and worth quoting directly: *"非實所緣自相，唯是內心所現彼相"* — "not the real self-characteristic of the object [of cognition], but only the appearance of that object as manifested by the inner mind" (cf. *Yogācārabhūmi* via *Bodhicaryāvatāra-ṭīkā*'s received commentarial tradition). This single sentence contains the whole teaching. Whatever appears to consciousness is not the thing-in-itself; it is the *image* of the thing, conjured by the conditions of cognition. Yet — and here is where the tradition refuses to slide into either idealism or nihilism — the image is not nothing. It can be seen, it can function, it can be the object of an eye-consciousness, it can even be classified under the category of "form" (*rūpa*). What it lacks is not existence but *self-existing essence*.
The concept lives a double life in the Buddhist canon, and both lives must be honored.
Life One: Image as the Object of Meditation
In the *Yogācārabhūmi-śāstra* (《瑜伽師地論》) and in Tsongkhapa's *Lam rim chen mo* (《菩提道次第廣論》), 影像 is the technical name for what appears in a stabilized mind when you hold an object of contemplation. The object need not be physically present. When you sit and cultivate *bhūmi*-by-*bhūmi* meditation on the sixteen impurities of the body, on loving-kindness, on dependent origination, on the aggregates and sense-bases, on the four truths — what appears before the meditative mind is an "image." The full technical name is *jñeya-sabhāgya-pratibimba* (所知事同分影像), "an image commensurate with the thing to be known."
Two kinds of such images are distinguished:
- 有分別影像 (sa-vikalpa-pratibimba / image-with-analysis). This is the image that arises *while you are actively analyzing*. You are doing *vipaśyanā* (毘缽舍那, insight): discriminating, examining, parsing the image into its parts. The sutra explicitly says: *"若說有分別影像，即是此中毘缽舍那品"* — "wherever 'image with analysis' is spoken of, that is the vipaśyanā side." - 無分別影像 (a-vikalpa-pratibimba / image-without-analysis). Once you have taken hold of the image-as-analyzed, you stop dissecting. You simply rest the mind upon it — through the ninefold *śamatha* (奢摩他, calm-abiding) progression: *inner abiding* (內住), *equable abiding* (等住), *steadied abiding* (安住), *close abiding* (近住), *taming* (調伏), *pacifying* (寂靜), *thoroughly pacifying* (最極寂靜), *one-pointed abiding* (一趣), and *equipoise* (等持). The mind settles on the image without further discrimination. *"若說無分別影像，即是此中奢摩他品"* — the *śamatha* side.
This pairing is the technical engine of every Buddhist meditation that proceeds through analysis-then-rest. The meditator first *constructs* a discriminated image by examining teachings received from a teacher, lets the right view arise *with* that image as its object, then drops the analysis and lets the same image serve as the resting ground of stable absorption. The image is the hinge between investigation and absorption.
A subtle but important elaboration from the *Yogācārabhūmi*: these images belong to a *world-realm* (界). If a meditator in *dhyāna* is contemplating form-realm phenomena, the images that arise are themselves said to be "made of the great elements of the form realm" (色界大種所造). Contemplating desire-realm phenomena yields desire-realm-element-made images. This is the tradition refusing any naive notion of "mental images are unreal." They participate in the ontological fabric of the realm being contemplated.
The *Lam rim chen mo* adds a practical note: images have 粗 (gross) and 細 (subtle) levels. You begin with the gross — a clearly defined, easily held image (for example, a clearly visualized form of the Buddha) — and only when that is solid do you proceed to the subtle. The text also warns that switching objects too often in early meditation is the chief obstacle to stable *śamatha*. You commit to one image.
The image also accumulates many alternative names, and the names themselves are teaching-tools. In the *Yogācārabhūmi*, the no-analysis image is called: *samādhi-lakṣaṇa* (三摩地相, "characteristic of samādhi"), *samādhi-gocara* (三摩地所行境界, "field of samādhi"), *samādhi-mukha* (三摩地口, "mouth of samādhi"), *samādhi-dvāra* (三摩地門, "gate of samādhi"), *manasi-kāra-sthāna* (作意處, "place of attention"), *antaḥ-pratītya-samutpanna-vijñapti* (內分別體, "inner discriminative body"), and *chāyā* (光影, "light-shadow"). The *Lam rim* lists mostly the same names with a small variation: *samādhi-upāya* (三摩地方便, "means of samādhi") instead of *samādhi-mukha*. Together they teach that the image is not an exotic mental event — it is the *form* samādhi takes whenever it has any object at all.
Life Two: Image as the Master Analogy for Empty Yet Causally Efficacious Phenomena
The second life of 影像 lives in Madhyamaka philosophy, especially as transmitted through Kamalaśīla's *Madhyamakāvatāra-bhāṣya* (《入中論善顯密意疏》). Here 影像 is the central simile used to dissolve the apparent contradiction between emptiness (*śūnyatā*) and the legitimate operation of cause and effect.
The cornerstone verse: 「如影像等法本空，觀待緣合非不有」 — "Phenomena such as the reflected image are *empty by nature*, yet they are not 'not-existent' when the conditions for their arising are assembled." This verse refuses both horns of the familiar dilemma:
- It is not *permanent* — the image is empty of the essence it appears to have. There is no face inside the mirror. - It is not *non-existent* — when light, mirror, original object, and observer are all present, an image undeniably appears. It can be photographed. It can be examined. It is classified as a *rūpa* (色, form) because the eye-consciousness can take it as its object-condition (所緣緣).
The text then asks: what makes the image "false" (*虛妄*)? The answer is precise. "False" here does *not* mean "fabricated by imagination" in a way that disqualifies it from reality. "False" means: *it appears in the mode of being the original, yet it is not the original*. The mode of appearance is false; the appearance itself is not non-existent. Kamalaśīla is fiercely insistent on this distinction because the entire Buddhist case for karma, rebirth, and liberation runs through it.
From there comes the analytical refinement that anyone who thinks seriously about Buddhist philosophy must hold:
| Treated correctly | Treated wrongly (and so collapses the teaching) | |---|---| | A phenomenon *exists* (有) | Mistaking "exists" for "exists with self-nature" (自性有) — leads to eternalism | | A phenomenon *arises from causes* (生) | Mistaking "arising" for "arising with self-nature" (自性生) — collapses cause-and-effect | | A phenomenon is *empty of essence* (無自性) | Mistaking "empty" for "utterly non-existent" (畢竟無) — leads to nihilism |
The image illustrates each of these distinctions. The reflected face is *there* — it is not a barren woman's child (石女兒, an example of something wholly non-existent and never even theoretically producible). It is not the original either. It arose because conditions gathered. When conditions disperse, it ceases. It is therefore a perfect exemplar of *dependently originated, empty-of-essence, yet functioning*.
The same logical move is captured in the *Lam rim chen mo*'s syllogism (比量): "芽無自性，依自因緣所生起故，譬如影像" — "A sprout has no self-nature, because it arises from its own causes and conditions, like an image." The image is the *example member* of an inference: if even something so obviously appearance-only as a mirror-image can arise, function, and cease according to conditions, then of course the sprout can do so without needing to possess a self-nature.
It is worth pausing on a famous classical verse that gathers many of these similes into a single breath. The *Diamond Sūtra* (《金剛經》) says: 「一切有為法，如夢幻泡影，如露亦如電，應作如是觀」 — "All conditioned phenomena are like a dream, a magical illusion, a bubble, a shadow (影), like dew, and like lightning; thus should they be contemplated." The character 影 here is in the same family of images as 影像. This is the most widely known canonical articulation of the *eight-simile* (八喻) tradition to which pratibimba belongs: dream (夢), magical illusion (幻), bubble (泡), shadow/image (影), dew (露), lightning (電), mirror-reflection (影像), echo (谷響), mirage (陽燄), and the city of gandharvas (尋香城). Each simile highlights a slightly different feature of the same truth: the moon's reflection in water teaches that the appearance is locationless; the dream teaches that vivid experience can occur without an external referent; the magical illusion teaches that what is conjured can still affect us; the shadow teaches that what follows the form is not the form. 影像 stands out because it preserves the *causal lineage*: the image is *generated* from conditions (mirror + light + original), not conjured arbitrarily. It is therefore the bridge simile — the one that allows the Madhyamaka to say "empty *and* dependently originated."
The Epistemological Architecture: Three Terms That Must Not Be Confused
The classical commentators are meticulous about keeping three terms distinct:
- 本質 (the "original" / the referent). The face before the mirror. The thing the image *resembles*. - 影像 (the image itself). What appears in the mirror. Treated as genuinely different from the inner consciousness, classified as an external object (外境), and treated as a valid *ālambana-pratyaya* (object-condition) of eye-consciousness, hence placed under *rūpa-dhātu* (色處). - 行相 (the aspect / the cognitive mode). The specific shape the consciousness takes when apprehending the image. This belongs to the *consciousness*, not to the image. The image is the *object*; the aspect is the *act of apprehending it in a particular way*.
Why does this matter? Because in a culture saturated with hand-wavy "it's all in your mind" talk, the Buddhist tradition refuses the collapse. The image is *not* "in your head." It is treated as genuinely other than the eye-consciousness. It can serve as a *rūpa* (form). When you see the moon in water, what your eye sees is a *rūpa*-classified image; what your consciousness *does with* that image is a *行相*. Kamalaśīla uses the famous case of the moon's reflection to show that *what is seen* is real-form, even though *the mode of its appearing* (as a moon in water) is not the actual moon.
A particularly striking distinction in the *Madhyamakāvatāra-bhāṣya*: hair-like visual phenomena seen by the eye are classified under *rūpa-dhātu* (色處, the form-sense-field), because they are seen by the *eye-consciousness* with an image-aspect. But a *bone-lock* (骨鏁) visualized by a meditator — internally visualized but with such clarity that the meditator "sees" it — is *not* classified under *rūpa-dhātu*; it is placed under 法處之徧計所起色, "form arisen through conceptual imputation within the dharma-sense-field." This is the Buddhist tradition refusing to flatten all appearances into one category. Some images are *visually perceived* (and so count as *rūpa*); others are *merely imaged by the mind* (and so belong to a subtler classification).
Walking It Through in Contemporary Life
Let us take this classical apparatus into a few settings that almost everyone knows.
Scenario 1: The Phone Feed at 11 p.m.
You are lying in bed. Your thumb scrolls. An ex-colleague's post appears: a beach photo, a caption about a promotion, a comment from a mutual friend reading "so happy for you." A feeling moves through you — envy, deflation, perhaps a small story: *"they are doing better than I am."*
What, in classical terms, has just happened?
- The 本質 (original) is the actual life of your ex-colleague — which you have not seen, which is invisible to you, which includes their loneliness at 2 a.m., their tax worries, the half of the day the post did not capture. - The 影像 (image) is the visible, scrollable, eye-consciousness-encountered form on the screen. This 影像 is not imaginary; it can be enlarged, screenshotted, sent to a friend. Kamalaśīla would say: classify it as *rūpa-dhātu*, it is a real visual form. - The 行相 (aspect) is the *mode* in which your consciousness apprehends it — as evidence of their flourishing, as a comparison point, as a stab. - And the entire scene is governed by the eight-simile insight: it is *bubble-like* (gone tomorrow), *dream-like* (assembled in your mind from the inputs), *shadow-like* (it follows the actual events but is not the actual events), *image-like* (it appears in the mode of the original but is not the original).
The crucial move is not to conclude "the post is fake, therefore I should feel better." That would be the *nihilistic* misreading. The post *is* something; your ex-colleague *did* go to the beach. But the *mode of appearance* — that this little square of pixels tells you anything about how their life is going, let alone how it compares to yours — that mode is false in exactly the sense the Madhyamaka is naming.
Scenario 2: The Voice Message from Your Partner
They left you a voice message. Three sentences. A pause before the second one. They said "fine" but the pause was half a second too long.
You now *have* an image of their emotional state in your mind. It has the force of perception. You begin to compose a response that addresses the pause rather than the words.
Classical analysis: your *eye*-consciousness cannot perceive their emotional state; you have not *seen* their face. What you have is an *aural* form (the actual audio waveform — classified as *rūpa*) plus a *conceptually constructed image* of their inner state. The second item is 法處之徧計所起色, a conceptually imputed form arising in the dharma-sense-field — exactly the bone-lock case. It is image, not perception.
This is precisely where the 有分別 vs. 無分別影像 distinction becomes useful as a practice-tool. The disciplined move:
1. 有分別影像 (image-with-analysis): let your *vipaśyanā* engage. Investigate: what is the actual evidence? What am I adding? Have they historically said "fine" while being fine? Could the pause just be them putting the phone down? Is my own anxiety filling in the gap? This is the discriminated image, the *vipaśyanā* phase. The classical material insists: do not skip this. Do not collapse it into either suppression or rumination. *Analyze the image.* 2. 無分別影像 (image-without-analysis): having analyzed, release the conclusion and simply rest the mind in openness before sending a reply. *Śamatha* phase. Not to suppress what came up — but to let the mind settle without piling further conceptual imputation on top of the already-imputed image.
The image (the partner's emotional state as you picture it) is the object of meditation in this moment. The practice is to hold it without either believing in it as essence or annihilating it as nothing.
Scenario 3: The 8 a.m. Email from the Boss
A short email. Five sentences. Sentence three is one word longer than necessary. You read it four times.
Here 影像 is almost entirely conceptual imputation — you are constructing an image of the boss's attitude (furious? disappointed? testing?) from extremely thin textual traces. This is *again* 法處之徧計所起色, an image in the dharma-field. It has *zero* visual form to undergird it. It is overwhelmingly the 行相 (aspect) of *your own mind's apprehension*, dressed up as if it were *影像* of theirs.
The classical instruction is humbling: the more the "image" is purely mind-constructed (lacking the eye-consciousness-encountered *rūpa*-classified base), the more honest it is to call it *your* image, not *theirs*. Treating it as if it had 本質 status — as if it were an actual perception of their mind — is precisely the "appears-as-essence-but-has-none" structure that 影像 names.
Scenario 4: The Image of the "Good Parent" You Carry
Perhaps the most existentially loaded 影像 of contemporary life is the culturally circulated image of the *good parent*. It appears in your mind's eye with startling clarity. You compare yourself to it constantly. You feel deficient.
This image has no original in the sense the Madhyamaka means. There is no archetype parent whose self-nature this image captures. It is assembled from podcasts, books, friend's Instagram posts, perhaps your own childhood. It is a textbook pratibimba: *appears in the mode of being a standard, yet is empty of any original self-nature*.
It is not "fake" in the dismissive sense — it has real effects on your behavior. (Recall: 影像 can still function, still produce actions, still be classified as *rūpa* when encountered visually in a magazine.) But to *believe* in its essence — to *behold* it as if you were perceiving an external standard — is precisely the suffering it produces. The practice: discriminate the image (what is it made of? whose voices constructed it? what specific behaviors does it demand?), then rest without either obeying it or hating it. See its nature as image.
Common Mappings to Avoid
In walking this through, several mis-mappings are easy and worth naming in advance:
- Don't map 影像 to "subjective." The classical material explicitly treats it as *external* (外境) and *other than consciousness*. The image is not "just your opinion." - Don't map 影像 to "fake." The image is a valid object of perception. The feedback in the mirror is not a hallucination. The boss's email is not fiction. - Don't map 影像 to "brain illusion" or "Matrix simulation." The classical framework is more subtle than "we live in a simulation." It is not denying the existence of the appearance; it is denying that the appearance has the kind of essence it *appears* to have. - Don't map 影像 to "nothing matters." This is the nihilistic trap. Kamalaśīla devotes significant energy to distinguishing 影像 (empty but not nothing) from 石女兒 (a barren woman's child — wholly non-existent). If you collapse the two, the entire structure of dependent origination and liberation collapses with it. - Don't map 影像 to a single sense. Eye-based images are the most discussed (mirror, reflection, moon-in-water), but conceptually constructed images (法處之徧計所起色) are equally 影像.
Why Contemporary People Need This Concept
The eight-simile tradition was articulated in a world where the most common illusion was the moon-in-water. Contemporary life has produced a media ecology in which we are *surrounded* by 影像 — visible, audible, interactively manipulable, increasingly indistinguishable from the original. The conceptual apparatus the Buddhist tradition developed for one moon-reflection now needs to be applied to a saturated feed.
Three concrete reasons stand out:
1. The information environment is image-dense. Smartphones, social platforms, AI-generated media, the algorithmic curation of what we see — every one of these is a technology that constructs 影像 at industrial scale. They are not "fake" in the crude sense. They are *visible, functional, classified-as-form* representations. But they appear in the mode of being a window onto reality, when they are in fact generated from conditions (data, prior clicks, business model, attention economy) that have no necessary connection to the original. The classical distinction between 本質 / 影像 / 行相 becomes urgent. Without it, we either naïvely believe the image (and feel inadequate, frightened, envious) or we dismiss it as nothing (and lose the capacity to be moved by anything we see).
2. The mental-health landscape is shaped by image-confusion. Much of what contemporary therapy calls *rumination*, *catastrophizing*, *mind-reading*, *personalization* in cognitive-behavioral terms is, in classical Buddhist language, the habit of treating mentally constructed 影像 (the partner's anger, the boss's disappointment, the child's future failure) as if they were directly perceived forms with 本質 status. The Buddhist prescription is not the modern CBT prescription — but the analysis is strikingly congruent: discriminate the image, identify its components, observe it without fusing with it, do not let it become the absolute.
3. The philosophical stakes of AI demand the distinction. When a contemporary generative model produces a synthetic image that is visually indistinguishable from a photograph, what is the ontological status of what we are looking at? Contemporary discourse oscillates between "it's fake" (a nihilistic misreading) and "it's still a real image" (correct but incomplete). The Buddhist framework offers a third position: it is a 影像 — a representation that arises from conditions (training data, model architecture, prompt), that can function (inform, persuade, deceive), but that lacks 本質 status (it does not *show* reality in the mode its appearance suggests). The image *is* something; it is not *the thing it appears to be*.
It is worth being honest about the limits of any cross-tradition comparison. Modern neuroscience's "predictive processing" — the view that perception is the brain's best-guess inference rather than a passive recording — is *analogous* to the image-analysis but is not identical. Predictive processing still aims at accurate *representation* of an external world; the Buddhist image-analysis is making a stronger claim about the ontological status of that representation. Cognitive psychology's distinction between *perception* and *interpretation* overlaps with the *rūpa*-image / *dharma*-conceptualized-image distinction, but the Buddhist tradition has a richer taxonomy and a different purpose: not to clean up perception but to release grasping at any perceived essence. Drawing the analogy is helpful for orientation; collapsing the two traditions into one another would mislead both.
Common Misreadings and Clarifications
Misreading 1: "影像 means 'illusion,' so everything I see is unreal."
This is the *nihilistic* reading. It collapses 影像 with 石女兒 (barren woman's child). The classical material explicitly refuses this: the image *appears* (有), *functions* (能作所緣), *is classified as form* (色處), and *arises from conditions* (緣起). The teaching is not "nothing is real." The teaching is "what appears does not appear in the mode it appears to appear in." This is much harder, and much kinder, than either realism or nihilism.
Misreading 2: "影像 = subjective projection, so other people don't exist."
This is the *idealist* reading. The classical material explicitly refuses this too. Kamalaśīla treats the image as 體異 from inner consciousness (substantially different from the inner consciousness), as 外境 (external object), and as 所緣緣 (object-condition). The Buddhist framework is not Berkeley. Other people are not your image of them. The point is that the *image* of them you carry is not them, while they themselves are not denied.
Misreading 3: "影像 is just a meditation technical term."
It is *also* a meditation technical term, with a precise function in the *Yogācārabhūmi* framework. But reducing it to that would lose the second life — its role as the bridge analogy that allows Madhyamaka to say "empty yet causally efficacious" without contradiction. The image works for meditation precisely because it is a real ontological example, not a meditator's whim.
Misreading 4: "影像 = 幻覺 (hallucination)."
A hallucination is a perception *without an external referent* of the appropriate kind. 影像 is the opposite: it is a perception *of a real appearance* — the reflection in the mirror is a real visible form. The "image" in 影像 is not the absence of an object; it is the presence of an object that *appears in the mode of being something it is not*. A hallucination of a person who is not there is a different phenomenon. The classical framework distinguishes them by whether there is a *rūpa*-classified form encountered by the eye-consciousness.
Misreading 5: "The point is to see through images and be done with them."
The *Diamond Sūtra*'s instruction (應作如是觀, "thus should they be contemplated") is the operative word. The contemplative task is not to annihilate images but to *see* them *as they are* — as images. This is a transformation of *seeing*, not a deletion of the seen. The mirror reflection does not need to be destroyed. The mind that mistakes the reflection for the face needs to be examined.
Misreading 6: "影像, 本質, and 行相 are the same thing."
Three distinct terms, three distinct functions. 本質 is the original; 影像 is what appears resembling the original; 行相 is how the consciousness takes up the image. The image is the object of cognition; the aspect is the cognitive act's shape. Collapsing them — particularly collapsing 影像 into 行相 — produces the "everything is in the mind" trap. Keeping them distinct preserves the Buddhist refusal to dissolve appearance into mind *or* mind into appearance.
Misreading 7: "In an age of AI and VR, the Buddhist similes are obsolete."
Quite the opposite. They may be more needed than ever. When a generative model produces a portrait, when a VR system produces a simulated conversation, when an algorithm assembles a feed that *resembles* your interests but is optimized for engagement — these are 影像 by every classical criterion: they appear in the mode of the original (a real face, a real conversation, a real interest), they arise from conditions (training data, prompt, attention economy), they function (inform, move, persuade), and they lack the self-nature their appearance suggests. The classical similes were developed for a world with one mirror; the contemporary world has multiplied the mirrors to infinity.
A Final Orientation
If one sentence could carry the teaching, it would be Kamalaśīla's: 「如影像等法本空，觀待緣合非不有」. Phenomena such as the reflected image are empty by nature, yet they are not "not-existent" when their conditions are assembled. The teaching of 影像 is, finally, a teaching about *how to hold what appears*. Not to cling to it as if it had an essence. Not to dismiss it as if it were nothing. Not to project it onto the world as if it were a window. Not to lock it inside the mind as if it were a dream. To hold the appearance as an appearance — present, functioning, causally connected, yet empty of the essence it wears. This is the image's gift to anyone willing to look at it carefully, in the mirror and beyond it.