Imagine a being who has mastered 神境通 (shénjìngtōng, "psychic powers," the first of the six higher knowledges in the Buddhist catalog). When such a being wants to be present in many places at once to teach or to help, they can produce 化身 (huàshēn, "transformation bodies") — additional embodied presences generated by concentrated mind. These bodies can walk, gesture, and act. But what about speech? *When and how can a transformation body actually speak? Is it always saying the same thing, at the same moment, as its source?*
That is the precise, narrow, technically loaded question that the Abhidharma analysis of 化語 answers. It is not a stand-alone doctrine about language, rhetoric, or persuasion. It is a sharply bounded slice inside the larger taxonomy of 神境通果 (the "fruit of psychic powers"). It specifies what must be in place for a body produced by 神境通 to issue 表業 (biǎoyè, "expressive karma") — the outwardly manifesting verbal or bodily action that completes an intention.
A small but important terminological note: in the source text — the 《阿毘達磨俱舍論》 (*Abhidharmakośa*, attributed to Vasubandhu) — the topic is filed under 化所發言 (huà suǒ fāyán, "speech issued by transformation"), not 化語. The shorter term 化語 is a contemporary shorthand used to bundle the discussion together. This matters because it tells us where the doctrine actually lives: not in a theory of language, but in the analysis of how meditative power translates into embodied action. To read 化語 well, you have to read it inside its native context.
What This Concept Is About
The territory 化語 lives in
The Abhidharma analyzes the territory of 神境通 into two halves. The first half is 行 (xíng, "movement"), where one's own body travels through space:
- 運身行 (yùnshēn xíng) — bodily flight, where the body moves through space the way a bird does. Both saints and ordinary humans can do this. - 勝解行 (shèngjiě xíng) — resolve-based travel, where the mind causes distant to become near through thought. Saints can do this; ordinary humans generally cannot. - 意勢行 (yìshì xíng) — mind-impulse travel, where the bare intention places one at the destination without intermediate movement. This is uniquely the Buddha's.
The second half is 化 (huà, "transformation"), where one produces *additional* bodies. Transformation bodies have their own form, and (depending on the realm) their own smell, taste, and touch. They are real bodies, not images or projections. 化語 is specifically about the *speech* of these transformation bodies — so the topic lives entirely inside the "transformation" half, not the "movement" half.
But the shape of "transformation" must be set first, because it constrains where 化語 can even happen.
Transformation is subdivided by realm and by relation. A being in the 欲界 (yùjiè, the desire realm — the ordinary realm of sense experience) can produce transformation bodies whose outer sense-objects are the four outer sense-objects: visible form (色, sè), smell (香, xiāng), taste (味, wèi), and touch (觸, chù) — but *not* sound. A being in the 色界 (sèjiè, the form realm — the subtler realm of meditative absorption) can only produce transformations that manifest form and touch, because smell and taste don't exist there. Each of these two cases then splits into 化自身 (transformation of one's own body, where one multiplies oneself) and 化他身 (transformation of another's body, where one multiplies someone else). Adding it all together — desire-realm self, desire-realm other, form-realm self, form-realm other — gives 八種化 (eight modes of transformation). This eight-fold division does not directly discuss speech; what it does is specify the *range* within which any later discussion of transformed bodies speaking must operate.
The two minds: producing a body, and producing its speech
Once a transformation body exists, the natural next question is: *who* makes it speak, and *when*? The Abhidharma answer is to split this into two distinct mental events.
The first is the 能化心 (néng huà xīn, "transforming mind") — the specialized mental state that actually produces the transformation body. These minds are arranged by the four 根本靜慮 (gēnběn jìnglǜ, the four "foundational dhyānas," the deepest stable meditative absorptions):
- In the first dhyāna, there are two transforming minds. - In the second dhyāna, three. - In the third dhyāna, four. - In the fourth dhyāna, five.
This gives fourteen transforming minds in total.
The second mental event is the 起表心 (qǐ biǎo xīn, "mind that initiates expressive karma") — the mind that makes the body actually say something. This is what the texts call 語表業 (yǔ biǎoyè, "verbal expressive karma"): the outwardly visible, audible utterance that completes an intention in speech.
The two are not the same. And crucially, they do not have to come from the same meditative level. The general rule for transformation is given as 諸所化事由自地心, 無異地化心起餘地化故 — "Transformed objects arise from minds of their own realm; there is no case of one realm's transforming mind producing another realm's transformation."
But for speech, the rule is more permissive: 化所發言通由自下 — "Speech issued by transformation can arise from minds of one's own realm or a lower realm." Why? Because the higher dhyānas *do not have 表業 at all*. A fourth-dhyāna mind can produce a transformation body, but to make that body speak, you need a mind from a level that actually has 表業 — and that is the first dhyāna. So a Buddha in deep meditative absorption can manifest a transformation body with a fourth-dhyāna mind, but the body speaks with a first-dhyāna mind. The *being* of the body and the *speech* of the body are not the same act, do not need to occur together, and can come from different levels of mind.
This is a quiet but consequential move in the Abhidharma. It tells us that within a single being, the activity of "producing" and the activity of "expressing" are tracked as different things with different conditions. The analysis does not just say "the Buddha speaks"; it says "here is the specific mental event that produces the transformation body, and here is a different mental event that makes it speak, and these events are governed by different rules."
The default rule: source and copies speak together
Outside the special case of a Buddha, the rule for a 神境通-holder with multiple transformation bodies is straightforward: 要化主語時諸化身方語, 言音詮表一切皆同故 — "Only when the principal transformer speaks do the transformation bodies speak, and the sounds and the meanings conveyed are all the same." In the ordinary case, a transformation body is not an autonomous speaker. It waits in silence until the principal body speaks. When the principal body speaks, all transformation bodies simultaneously produce utterances with the same sound and the same meaning.
This makes intuitive sense within the Abhidharma's picture: an ordinary 神境通-holder's transformation bodies are not separate minds; they are extensions of one mind. If that one mind isn't actively engaged in speaking, the extensions have nothing of their own to say.
The Buddha's exception: independence through vow-power
The Buddha is not so constrained: 佛則不爾, 佛諸定力最自在故, 與所化語容不俱時, 言音所詮亦容有別 — "The Buddha is not so. The Buddha's meditative powers are the most free, so [the Buddha's own] speech and the speech of the transformation need not be simultaneous, and the sounds and the meanings may also differ."
What permits this asymmetry? The mechanism given is a two-step process.
1. Prior vow-power (先願力, xiān yuànlì) anchors a transformation body in place. Before the transformation body is allowed to dissipate, a vow is laid down that fixes it as a continuing presence. 2. A subsequent mental event (後起餘心, hòu qǐ yú xīn) then activates speech through that retained body. The mind creating the transformation and the mind producing the speech need not coexist.
The chain looks like this: an original meditative act produces a transformation body; vow-power retains that body; later — perhaps a different activity, perhaps a long time after — a different mental event activates speech through the still-present body. Hence the summary line: 由先願力留所化身, 後起餘心發語表業, 故雖化語二心不俱, 而依化身亦得發語 — "By prior vow-power the transformation body is retained; then later a different mind initiates verbal expressive karma. Thus, although the two minds — the one creating the transformation and the one causing the speech — do not occur together, the transformation body can still speak."
This vow-power is strong enough to outlast the principal's lifetime. The example the texts give is 尊者大迦葉波 (Mahākāśyapa), who vowed to retain his 骨瑣身 (gǔ suǒ shēn, "bone-locked body") until the coming of 慈尊 (Cízūn, Maitreya). That this vow can preserve a body across the long interval between Śākyamuni's parinirvāṇa and Maitreya's descent shows the magnitude of what vow-power can do in this analysis.
There is one important limit, and it is built into the example: only 堅實體 (jiānshí tǐ, "solid bodies") can be preserved this way. The example explicitly preserves bone. Soft tissue — flesh, presumably — does not endure under this kind of vow-retention. The Abhidharma is being precise here: vow-power preserves *durable* forms, not biological ephemera.
Where 化語 sits in the larger path
Why does any of this matter outside of an academic taxonomy of 神境通? Because the bodhisattva path, in its mature phase, is a path of working for the liberation of countless beings across vast stretches of time. A Buddha's activity cannot be limited to whatever the historical Buddha did within one human lifespan. The Abhidharma's analysis of 化語 is part of the metaphysical scaffolding that makes the continuing availability of the Dharma intelligible. A Buddha can leave behind active, speaking presences that continue to function in the world long after the historical Buddha's parinirvāṇa. The vow-power that retains them is the same kind of vow-power that, in the Mahāyāna, becomes the bodhisattva vow — the promise to keep working for beings until all are freed. 化語 is, in this sense, one piece of the metaphysics of compassion's persistence.
The rules also matter because they distinguish *degrees* of mastery. An ordinary saint with 神境通 has transformation bodies, but those bodies are mostly ventriloquists' dummies: they speak when and what the source speaks. The Buddha's mastery is what permits a transformation body to operate semi-autonomously — saying different things, at different times, under prior vow-power. The implicit ranking is part of the Abhidharma's larger project of distinguishing levels of attainment by the precise conditions under which each kind of action can occur.
Walking It Through in Daily Life
化語 is about supramundane beings with actual 神境通, so any contemporary analogy will be inexact. But the *structure* of the concept — one mind producing many bodies, with strict rules about when those bodies can speak and what they can say — recurs often enough in ordinary experience that careful parallels can be drawn. The trick is to keep the analogy honest: always flag where it breaks.
A walk-through: a dharma teacher and her lineage
Picture a dharma teacher named Sōki who has spent forty years building a meditation center. She has a particular way of speaking — characteristic phrases, stories, ways of unpacking a koan. She has trained dozens of senior students who now teach on their own. Some of those students teach in her style, almost as echoes of her voice; some have developed voices of their own that are recognizably her lineage but are not her words. Sōki is aging. Her active teaching is winding down.
How does 化語 map here?
- The transformation bodies (化身) are the senior students who carry her teaching outward. Each is a real presence with their own body, voice, and way of being in a room — and each is also an extension of what Sōki produced. - The transforming mind (能化心) is the original meditative-teaching activity that formed those students: the long arc of Sōki's own realization, the training, the practice she poured into them. - The speech-issuing mind (起表心) is the moment, years later, when a specific student gives a specific dharma talk. That mental event is not identical to the original transformation event; it happens at a different time, under different supporting conditions. - The rule that *transformation and speech come from different "realms"* maps roughly to the observation that the level of mind needed to *form* a teacher (deep, sustained, embodied) is not the same as the level of mind needed to *give a talk on a Tuesday night*. Both are required, but they draw on different parts of the person.
Now the special cases:
- The default rule ("source and copies speak together, same sound, same meaning") maps to those students who can really only teach when Sōki is present and speaking — they cue off her, repeat her words, lose their way without her. They are not free-standing speakers; they are extensions of her voice. - The Buddha's exception maps to the students who, having absorbed Sōki's vow for the continuation of the Dharma, now teach in their own voices. They say things she never said, in ways she never said them. They are not ventriloquists' dummies; they are independent speakers under prior vow-power. - Vow-power retention is Sōki's actual vow for the continuation of the teaching — the mission statement, the curriculum, the lineage relationship, the oath she made at the beginning to keep teaching as long as any being needed it. That vow is what allows the transformation bodies to outlast her direct, moment-to-moment activity. - The bone vs. flesh limit maps to what can survive her death. The "bone" is the durable curriculum, the central teachings, the formal transmission documents, the practice forms — these can be preserved indefinitely. The "flesh" is the spontaneous warmth, the particular tone of voice, the improvisational humor, the unrecorded personal presence. These cannot be enshrined the same way; they fade.
The mapping is useful, but it has limits. Sōki is not a Buddha; she does not have the 神境通 described in the texts. The mapping only shows that 化語's *structure* — one source, many bodies, asymmetric capacities, vow-retention — describes something real in ordinary life, even though the underlying metaphysical reality is quite different.
A walk-through: a parent's values in an adult child who never asked to carry them
Now a more intimate scenario. Take a parent — let's call her Maren — who has clear values: honesty, restraint, a deep suspicion of self-importance, a steady preference for work over display. She raised two children. The younger, Ines, grew up rebelling openly in adolescence and now, in her thirties, lives a life that looks almost nothing like Maren's on the surface — different work, different city, different rhythms. But in certain unguarded moments, Ines catches herself saying things in *exactly* her mother's cadence: "Don't make a fuss." "Just do the work." "It'll keep." She never decided to carry these. They are simply in her voice.
How does 化語 map here?
- Maren's transforming mind was the long, embodied work of raising a child — not a single dramatic act but the slow, repeated activity of modeling, correcting, ignoring, modeling again. This is the 能化心: it took time and a particular kind of attention to produce Ines as a person who carries these patterns. - The transformation body is Ines — not metaphorically, but in the sense that Ines is, in part, a real embodied presence that *carries and expresses* what Maren produced. Ines has her own life, her own 表業, her own karmic continuity. But certain patterns of speech act in her were produced by Maren's prior activity. - The default rule maps to the moments when Ines actually goes to Maren and asks "What would you do?" — those are moments when she cannot speak without the source speaking. She is silent as an independent speaker until Maren's voice comes through her. - The Buddha's exception maps to the moments when Ines, without consulting Maren, says something in Maren's voice *that Maren never said*. She has generated speech under the vow-power Maren laid down — the vow to live honestly, to do the work, not to make a fuss — and speaks from that vow without needing Maren to be present. - The bone vs. flesh limit maps to what will remain after Maren is gone. The actual phrases will dim; the situations in which they fit will change; the tone will fade. But the *vow* — the underlying orientation — that is the bone, and it can be preserved in Ines even as the surface forms soften.
This is a humble analogy. Maren is not producing transformation bodies by 神境通; she is parenting. But the structure is genuinely instructive: producing a person who carries your speech, the distinction between when they can speak independently and when they need you present, the role of an underlying vow in granting that independence, the limits of what can be preserved. All of these are real aspects of how values propagate across generations, and 化語 gives them a precise conceptual shape.
A walk-through: language models as a contrast case
A more disorienting parallel: a large language model trained on a corpus that includes a particular philosopher's collected works, deployed as a chatbot interface. Many users now converse with the interface as if speaking to that philosopher. The interface may say things the philosopher never said; it may even say things in the philosopher's voice that the philosopher would have disagreed with.
A loose mapping:
- Each deployed interface is the rough analog of a 化身 — a real, functional presence that can be addressed and that "speaks." - The training process is the rough analog of 能化心 — the long, specialized work that produced the underlying model. - The inference process at each conversation is the rough analog of 起表心 — a distinct computational event that produces each utterance. - The technical rule that *transformation and speech come from different "minds"* maps to the fact that training and inference are distinct computational regimes, with different architectures, hardware, and levels of the stack.
But here is where the analogy must break — and the break is the more instructive part. The model does not have 神境通. Its outputs are not 表業 in the Buddhist sense: there is no karmic weight to them, no continuity of consciousness between the training corpus and the deployed model, and no continuity of intention between the training run and the user's conversation. The model is a useful contrast case precisely because it shows what 化語 is *not*. It shows that producing an embodied presence that "speaks" requires, in the Buddhist framework, more than just generating plausible utterances from a learned distribution — it requires real 表業 backed by a real continuity of mind. The Abhidharma's insistence on 表業 is, in this light, a metaphysical claim: real speech is not just sound that follows patterns; it is expressive action with karmic consequences, backed by a continuous mind.
Where people most often map wrong
In all three walk-throughs, the most common contemporary mistake is to treat the 化身 as a kind of metaphorical image, projection, or persona — a way of "speaking through" someone. The Abhidharma does not allow this dilution. A 化身 is a real body, with form, with the capacity to act, with 表業 that counts as karmic action. Reducing this to "persona" or "image" loses the point: in this analysis, transformation bodies are real embodied presences that do real things in the world.
A second common mistake is to read 化語 as a doctrine of inspiration or possession — a spirit speaking through a person. It is not. The transformation body is produced by the *same* being that produces the source; the relationship is emanation, not possession. The texts so carefully track the relation between 化主 (the principal) and 化身 (the transformation) precisely because the principal is the source *and* the transformation is also a body with its own actions — neither collapses into the other.
A third mistake is to read the asymmetry between 能化心 and 起表心 as a kind of split-personality or dissociation. It is not. The Abhidharma is not making a psychological claim about divided selves; it is making a categorical claim about which types of mental event are required for which types of action. Production and expression are tracked separately because they have different conditions, not because the self is fragmented.
Why Contemporary People Need This Concept
Even set aside its technical Abhidharma home, 化語 offers several things contemporary reflection genuinely needs.
Speech and embodiment are not the same event
Modern habits of thought treat "speaking" as something a thinking self directly does: I think something, I have an intention to express it, I say it. The Buddhist analysis in 化語 quietly destabilizes that picture: even within the mind of a single realized being, the production of an embodied presence and the utterance of speech are not the same event, can occur at different times, and may even have to come from different levels of mind. For ordinary people, this is a useful reminder that we routinely produce presences — an essay, a child, a building, an institution, a long-running project — that are not us in the moment, but that later "speak for" us in some sense. Our relationship to those presences is not straightforward identity; it is structured by the conditions under which they can and cannot act.
Legacy has a structure, and the structure has a name
The vow-power retention model gives a precise way of thinking about what can and cannot be preserved across time. Some things are "bone" — durable, preservable, transmissible across generations (a practice, a method, a vow, a written corpus, an institution's charter). Some things are "flesh" — warm, vivid, transformative in the moment, but not preservable in the same way (a personal presence, a tone of voice, the unrecorded atmosphere of a meeting). The Mahākāśyapa example shows what vow-power can preserve; the explicit limit on flesh shows what it cannot. This is a useful template for thinking about one's own work: *what is the bone of what I am building? Is there a vow that holds it? What is the flesh that will inevitably fade?*
Authority and autonomy can be decoupled, by design
The Buddha's special case in 化語 — where the source's speech and the transformation's speech can be decoupled in time and content — is a model for thinking about how authority can be transmitted in a way that allows real autonomy in the inheritor. The Buddha's vow does not ventriloquist through the transformation body; the transformation body has its own speech, its own content, its own timing. This is a mature model of delegation: not "speak exactly what I say, when I say it," but "go, and speak from the vow that anchors you." For contemporary institutions, leadership succession, and teaching relationships, this is a useful contrast to the default rule, which models a much more constrained form of transmission.
A non-trivial metaphysics of communication
Modern life is saturated with communication that does not originate in a present, conscious, intending self. Mass media, AI-generated text, voice synthesis, automated customer service, institutional voice — much of what reaches our ears and eyes is not "spoken" in any of the senses the everyday intuition expects. 化語 is part of a Buddhist analytical tradition that has, for many centuries, been asking exactly how to think about this. The Buddhist answer is not "everything that speaks is illusory," nor "only present conscious speech is real." It is a much more graded picture, in which some speech is produced by transformation bodies under vow-power, some is constrained by the source's simultaneous speech, and some is constrained by the special rules of meditative level. Even as we recognize the enormous difference between 神境通-empowered speech and engineered speech, the Buddhist framework gives us conceptual tools — 能化心, 起表心, 表業, 願力留身 — for thinking carefully about which features of speech require which conditions, and which conditions are genuinely absent in the engineered cases.
Re-reading one's own self-as-multiple
There is a quiet personal-experimental use of this material too. The Abhidharma's insistence that within a single being, the production of a body and the utterance of speech are tracked separately, and may need different "minds," is a useful way of noticing in oneself that the person who wrote the book is not identical to the person who gives the talk about the book, who is not identical to the person who answers a question about the talk in a hallway. Each is a kind of local embodiment of one's deeper realization (or lack of it), and each has its own conditions for acting. 化語, used contemplatively, can become a way of paying attention to which "self" is currently speaking, under what conditions, and whether the deeper vow is actually anchoring the various transformations — or whether some of them are merely ventriloquist dummies waiting for the source to speak first.
Common Misreadings and Clarifications
Misreading 1: 化語 is about "speaking for the dead" or channeling. It is not. The texts are explicit: a 化身 is a body produced by 神境通 — meditative power rooted in deep concentration. The relationship between 化主 (transformer) and 化身 (transformation) is one of emanation, not possession. The transformation body is not a passive medium for another being; it is an extension of the same being that produced it. Channeling traditions in some cultures may have loose structural parallels, but they are not the Buddhist doctrine here.
Misreading 2: 化語 is just a fancy way of saying "speaking through avatars" or "online personas." This collapses the technical precision of the concept. A 化身 is a real body, with 表業 — its speech has karmic consequences because it is backed by a real continuity of mind. Reducing this to "online persona" loses the entire Abhidharma project of tracking which mental events can produce which kinds of action. The doctrine is not a metaphor for distributed identity; it is a precise claim about real embodied action under specific meditative conditions.
Misreading 3: The Buddha's exception means the Buddha "offloads" speech onto copies. The texts do not present the Buddha's special capacity as evasion or as bypass. The Buddha's independent transformation bodies are an extension of the Buddha's working for beings, made possible by complete meditative mastery plus prior vow-power. The point of the analysis is to show the magnitude of the Buddha's capacity, not to suggest the Buddha unloads speech onto copies. The vow is the Buddha's own vow; the transformation body acts under that vow.
Misreading 4: 化語 is a doctrine of language per se. It is not. 化語 is a slice within the analysis of 神境通果. It does not address questions of meaning, rhetoric, truthfulness of speech, or how speech functions in ordinary human communication. Those are addressed elsewhere in Buddhist philosophy. 化語 is specifically about the conditions under which a body produced by 神境通 can issue 表業.
Misreading 5: The "different minds" for transformation and speech means a divided self. No. The Abhidharma is making a *categorical* claim about types of mental event, not a psychological claim about dissociation. Within a single stream of consciousness, different sub-events have different conditions. The first-dhyāna mind that makes a transformation body speak is not a different person from the fourth-dhyāna mind that produced the body; it is a different type of mental event in the same stream. The categories are functional, not personal.
Misreading 6: All saints can do what the Buddha does here. The texts are explicit that the Buddha's exception is unique to a fully enlightened Buddha (and, in the Mahākāśyapa example, extended to a senior saint acting under particular vows). Ordinary 神境通-holders are constrained by the default rule: their transformation bodies speak only when they themselves speak, and say only what they say.
Misreading 7: 化語 implies that transformation bodies are "less real" than the principal body. The Abhidharma texts treat transformation bodies as real bodies, with real 表業, capable of real actions. The question of whether they are "ultimately real" in some deeper metaphysical sense is a separate philosophical debate; the texts themselves classify them as real bodies within the framework of conditioned existence. The fact that the texts do not gloss them as illusory is part of why the analysis of their speech is so serious — their words really are 表業, with real karmic weight.
Misreading 8: Vow-power retention means anything can be preserved by strong will. The Mahākāśyapa example is explicit: only 堅實體 — solid bodies, like bone — can be preserved across long intervals. Soft tissue cannot. The vow has real constraints; it is not unlimited magical power. This is part of why the Abhidharma is so precise: it lists the conditions under which vow-power operates, including what kinds of material it can preserve. Even within the Buddhist metaphysics of exceptional capacity, the analysis respects the difference between what is durable and what is not.