A Buddha statue sits in a hall. It was carved once, long ago, and it has not moved since. Whoever walks in finds the same shape waiting — it stores the Buddha's attributes the way a photograph stores a face, fixed and inherited, ready to call them to mind. Now picture a different image: not the statue, but the figure a practitioner builds in the mind during visualization, assembling it stroke by stroke out of their own active involvement, so that without the practitioner's doing there is no image at all. Dukes says these are not the same kind of thing. The first is a rupa. The second is a pratima. And the whole theory of Chuan Fa kata turns on the difference.1
Dukes is untangling how Sanskrit image-words were rendered into Chinese, and the untangling exposes a real distinction.2 Rupa in Sanskrit names an image formed "through the assemblage of a number of previous components into a contemporaneous common characteristic" — which is why the same word describes the fixed, component attributes of human consciousness.3 A rupa is inherited or already-existing. A Buddha statue is a rupa: it "acts as a static mnemonical of the Buddha's attributes."4 The image is there before you arrive; you receive it.
Pratima names something else: a form or image "produced only or mainly by a dynamic interplay between its component factors" — "a shape created by actual involvement of the onlooker in its creation."5 The word shows up in the Chinese and Japanese terms for "imagination" or visualization (Hsiang Hsiang, Japanese sozo) precisely because it means an image the participant brings into being.6 The result may look identical — you can build a religious image physically or mentally — but the process is opposite. Rupa is given; pratima is generated.7
Here is the puzzle Dukes is solving, and it is a good one. Later Chinese translators rendered rupa with the character meaning color — se in Chinese, shiki in Japanese.8 On its face that is bizarre. Why translate "form" or "image" as "color"?
Dukes's answer runs through the kata. In Chuan Fa, he says, a pratima has two aspects. There is the "activity form" — Japanese jozo, the "karma (activity) form" — which is the actual movements themselves that form the pratima.9 And there is the static form — Chinese pu tung hsiang, Japanese fudozo — the total fixed image or mnemonic the pratima embodies, "such as the symbols found in the mandala," also called its Li or Principle.10 The activity form, Dukes writes, makes up the pratima "in much the same way as various colors form a painting."11
That is the hinge. Movements compose the living pratima the way colors compose a painting — each motion a stroke, the whole form emerging only as the strokes are laid down by the practitioner doing them. Dukes suspects this is exactly why the ancient translators reached for the "color" character to approximate rupa: familiarity with the painting-by-colors logic of dynamic form bled the word "color" into the vocabulary of image.12 A strange translation choice becomes legible once you see the kata as a painting made of motion.
Hold the painting metaphor and the two forms snap into focus. The fudozo — the static, principle form — is the finished composition hung on the wall: the mandala's fixed symbol, the Li, the total image you can point to and name. It does not move. It is the rupa-like pole of the pratima, the inherited mnemonic.
The jozo — the activity form — is the act of painting: the brush actually moving, color going down, the image coming into existence under the hand. In the kata, the jozo is the sequence of movements themselves, the living doing that generates the form in real time. A practitioner running a kata is not displaying a stored picture; they are painting one, stroke by stroke, and the form exists only while the painting is happening.
This is why the pratima is the participatory pole and the rupa the inherited pole, even within a single kata. The fudozo is what the kata means — its principle, sitting there like a hung canvas. The jozo is what the kata does — the brush in motion. You inherit the principle; you generate the activity. The difference between rupa and pratima is not two different objects. It is the difference between the painting on the wall and the painting being painted.
This page is the theory of form the whole kata corpus has been waiting for. Pratima: Dynamic Mudra Sequences describes the moving form; this page supplies the precise reason it is a pratima and not a rupa — it is generated by the participant, not received. Kata as Inward Training gets its deepest justification here: a kata works inwardly because the practitioner is making the image, not viewing one, and making recruits the whole person where viewing does not. Mudra: Hand-Gesture Consciousness Encoding treats gestures as encoded meaning; the rupa/pratima split says a mudra has both a static encoded principle (fudozo) and a living generated enactment (jozo), and only the second is alive. And Five Elements Consciousness Manifestations connects because the Okinawan kata are "based upon the four elements" — the elemental content is the palette the activity-form paints with.
The single most load-bearing line is Dukes's aside: the activity form "indicates the actual movements themselves which form the pratima, in much the same way as various colors form a painting."11 Read it slowly and it carries three claims at once.
First, that movements form the pratima — the image is constituted by motion, not merely accompanied by it. Take the motion away and there is no pratima, only the static fudozo principle waiting inertly. Second, the analogy to colors forming a painting tells you the relationship is compositional: each movement is a constituent, not a decoration, the way each color is part of what the painting is. Third — and this is where Dukes earns the whole note — the analogy explains the otherwise inexplicable translation. "I suspect that familiarity with such a distinction was the reason the ancient translators used this unlikely character indicating 'color' to approximate the meaning of rupa."12 He is reverse-engineering a thousand-year-old translation choice from the structure of the practice. If movements form a pratima as colors form a painting, then a translator steeped in that practice would feel "color" and "form" as neighbors, and would reach across. The case study is a piece of linguistic detective work: an odd character becomes evidence that the painting-by-motion model of the kata was alive in the minds of the people doing the translating.
You step to the start of the kata. Before you move, the fudozo hangs in front of you — the finished form, the principle, the meaning of the whole sequence, like a painting you have seen on the wall a hundred times. You could just recite it, trace the remembered shape. Instead you pick up the brush. The first movement goes down: a real stroke, your weight actually shifting, the air actually parting at your hand. That stroke is not a reference to the form — it is the form, the first color on the canvas. The second movement lays the next color. You feel the image coming into being under you, not retrieved but made, and you understand in the body why the old word for this leaked into the word for color. When the last movement lands, the painting is complete and already vanishing — a pratima exists only while it is being painted. You did not show a picture. You were the brush.
The source is Dukes's note 74b in full, a translation-history exposition.1 The whole entry is his own philological argument, and he flags its tentativeness himself: "I suspect that familiarity with such a distinction was the reason the ancient translators used this unlikely character."12 So the central claim — that the kata's painting-by-motion logic explains why se (color) came to render rupa — is offered as informed conjecture, not established etymology.
The terminological field is genuinely tangled, and Dukes says so: the same Chinese characters get combined differently for the same thing across different translations of Sanskrit texts, and "such usage of the character se to represent 'form' can cause confusion."13 He is reconstructing a clean rupa/pratima/jozo/fudozo architecture out of a messy translation record. That is a strength — it makes sense of the mess — but it means the neatness is partly his.
Open: is the "color" rendering of rupa in fact explained by the kata's activity-form model, or does it have a more ordinary Abhidharma source — rupa as the form-skandha, of which visible color (rupa-ayatana) is the prime instance, color simply standing in for the whole category? Dukes's painting explanation is elegant; the skandha explanation is duller and may be the real one. [PLAUSIBLE — needs corroboration] Both can be partly true.
Dukes is doing practitioner-philology — reconstructing a translation history from the inside of a living practice — and the rupa-as-color etymology is his conjecture, marked as such by his own "I suspect."12 Hold it as an illuminating hypothesis rather than a settled finding: a Sanskritist working only from texts might well derive the "color" rendering from the form-skandha's standard analysis, where visible color is the representative case of rupa, without ever invoking the kata. The two explanations are not even exclusive — the skandha sense could be the doctrinal seed and the kata's painting-by-motion the living reinforcement that kept the choice feeling natural. Where Dukes converges with the broader vault: the conviction that doctrine and practice are one fabric, that you cannot fully read the philosophical vocabulary without having done the movement it describes. His rupa/pratima distinction is, at root, the participation thesis — meaning is generated by the doer, not received by the viewer — which is the same instinct running through his whole treatment of kata as self-revealing internal environment.
Enactive Cognition and Generative Perception — Modern cognitive science increasingly argues that perception is not passive reception of a finished world but active generation — the brain constructs the percept, and what you see is in part what you make. The rupa/pratima distinction is that thesis stated in fourteenth-century image-theory. A rupa is the received, inherited form — perception modeled as a stored picture handed to a viewer. A pratima is the generated form — perception modeled as a construction the participant builds through involvement. The insight neither field reaches alone: cognitive science supplies the experimental case that even ordinary seeing is generative, dissolving the rupa-style "image just sits there and I receive it" picture as a folk theory; and Dukes's frame supplies a trained version — a deliberate technology for generating form (the kata as jozo) rather than receiving it, plus a vocabulary that keeps the two modes distinct. Put together: if perception is always somewhat pratima-like, the kata is perception-training, raising the generative element to consciousness so the practitioner feels themselves making the form they had assumed they were merely receiving.
Kata as Inward Training — Kata trains inwardly, this page says why: the practitioner generates the image rather than receiving it, and generation recruits the whole person. The jozo/fudozo split sharpens the inward-training claim into a mechanism. The fudozo — the static principle — is what a kata means; you could memorize it from a book. The jozo — the activity form — is what only doing produces, the painting actually being painted. Inward transformation lives entirely on the jozo side: you are changed by making the form in your body, not by knowing its principle. Neither domain alone gives you this. The martial literature treats kata as a repository of technique (pure fudozo thinking — a stored shape to copy); the image-theory here insists the technique is dead until it is painted, until the activity-form generates it live. The practice the technique-repository reading entirely misses is the only one that does the work.
Sharpest implication: the same image carved in stone and built in motion are two different kinds of thing — and only the one you generate yourself, stroke by living stroke, can change you, because changing requires making, not receiving.
Generative questions: