Skip to content
Mind Only

Index

T1586 · VERSE 1 OF 30 · THESIS

Self and dharmas are designations, not entities

Through provisional designation of “self” and “dharmas,” appearances of every kind proceed. These depend on what consciousness transforms — and that which transforms is only three.

ātmadharmopacāro hi vividho yaḥ pravartate | vijñānapariṇāme 'sau pariṇāmaḥ sa ca tridhā || 1 ||

由假說我法　有種種相轉 彼依識所變　此能變唯三

Triṃśikā, verse 1 · from Xuanzang's Chinese · tr. Claude (Anthropic), 2026

The opening move is structural and linguistic, not metaphysical. Vasubandhu does not say self and dharmas do not appear; he says they appear as provisional designations (upacāra): labels the appearances have to carry, not entities they have to refer to. Both terms matter. “Self” is the obvious target, but “dharmas”, the basic ontological elements of Abhidharma analysis, is the more radical inclusion. Even the units you would reach for in order to ground “self” are themselves upacāra. All that is left is the transformations of consciousness from which the designating arises. The remaining 29 verses unpack what those transformations look like, layer by layer; this verse tells you how many there are (three) and what they are not (independent things).

LAYER 3 · PRAVṚTTI-VIJÑĀNA The Six Sense Consciousnesses eye · ear · nose · tongue · body · thinking mind stateless consumers LAYER 2 · MANAS Self-Consciousness always-on reader · reifies the substrate as "I" stateful cache LAYER 1 · ĀLAYA-VIJÑĀNA Store Consciousness all karmic seeds · karmically neutral · "like a river flowing" append-only event log replays seeds conditions experience deposits impressions writes new seeds
The three transformations of consciousness as a layered information system. Verse 2 names the layers; the rest of the text unpacks how impressions flow between them.

Apparatus — five witnesses

DSE

The Distributed Systems Engineer

Treats alaya as an append-only log, manas as a hot cache, the six senses as consumers.

Read this as a claim about identifier provenance. “Self” and “dharmas” are not entities the system points at; they are labels produced inside the system. The verse refuses the naïve architecture where a self sits at the keyboard and a world full of typed objects sits on the other side of a serial port. Both endpoints are records emitted during the same continuously running transformation.

The closest engineering analog is the move from a client/server mental model to an event-sourced one. The “client” and “server” are not physically prior to the events that mention them; they are reconstructed by replaying events. Vasubandhu is opening with the strongest version of this: not only is the self derived from its own records, so is every type it would reach for to describe what it is encountering. The pipeline preview, only three, is also DSE-friendly: storage, indexing/cache, and the user-facing read layer. The next 19 verses are this preview spelled out.

Draft not yet reviewed
CYN

The Cynefin Practitioner

Maps each verse to Clear, Complicated, Complex, Chaotic, or Confused.

This is a domain claim made up front and unambiguously. Vasubandhu is asserting that the phenomenology of subject and object, self and the typed ontology you would reach for to talk about self, is Complex, not Complicated. There is no decomposition into a self-component plus a dharma-catalogue that, wired correctly, yields experience. The pieces are constituted by the wiring. You only see what they mean in retrospect, after the transformation has run.

Mistaking the Complex for the Complicated (treating experience as if it had separable parts you could swap out) is the mistake Cynefin warns gets you killed. Vasubandhu, in verse 1, is preempting it. He also preempts the Clear-domain reduction “self is just X” (brain, soul, function, role) by naming both terms, self and dharmas, as upacāra in the same breath. You cannot escape the Complex domain into a Clear one by pointing at a smaller unit; the smaller units are conventional too.

Draft not yet reviewed
COG

The Cognitive Scientist

Reads through predictive processing, Bayesian inference, and self-model theory.

Predictive processing makes the same move when it treats the self/world boundary as a prediction the brain generates, not a feature of the input. The cortex models “external world causing sensations” because that model minimizes free energy, not because it has independent access to an external world. Anil Seth’s “controlled hallucination” phrasing is verse 1 in a different idiom; Metzinger’s Phenomenal Self-Model is the same move applied specifically to “self.”

What the new wording sharpens is the linguistic angle. “Self” and “dharmas” are designations: how the model labels its outputs, not metaphysical commitments the model makes. The brain doesn’t decode pre-existing categories of self and dharma; it constructs them as the posterior of a generative model whose priors are themselves shaped by previous self/dharma inferences. Recursive, with no ground floor. That recursion is what pariṇāma names.

Draft not yet reviewed
PRO

The Process Philosopher

Whitehead, Heraclitus, Bergson — reality as flow, not substance.

Whitehead’s actual occasions of experience are events, not substances; a “subject” is a society of such occasions, not a thing they happen to. Verse 1 is process-philosophical to its core: it asserts that the apparent furniture of the world, me here, dharmas there, is dependent on a more fundamental flow.

What the new translation makes more explicit is the language of furniture. “Self” and “dharmas” are names the flow gives to its own patterns. In Whitehead’s vocabulary they are abstractions from concrescence, not the concrescence itself. In Bergson’s they are the spatialized representations the intellect superimposes on durée. In Heraclitus, the river is the standing example; the people who name it “the same river” twice are themselves the larger river the metaphor is reaching for. You do not step into the same self twice.

Draft not yet reviewed
SKP

The Skeptic

Mandatory on every verse. Names where the metaphors break.

Notice what the verse does not say. It does not say self and dharmas are illusions. It does not say there is no world. It does not say everything is mental in some idealist sense that the Cartesian could recognize. It says they are upacāra, provisional designations, inside a transformation. The technical force of upacāra is “metaphorical, derived, secondary use”, not “untrue.”

Three traps:

  1. The systems-engineer reading is useful but leans on a metaphor (event sourcing) with a physical substrate. Vasubandhu isn’t telling you what the substrate is. Resist filling in “the brain” or “physics.”
  2. The “controlled hallucination” framing in cognitive science still posits a brain doing the hallucinating. Verse 1 doesn’t.
  3. “Provisional designation” gets compressed into “useful fiction” by Western readers. Upacāra is sharper: the appearances are real, but the names you reach for to describe them carry implications the appearances don’t carry. The verse is a claim about labels, not a claim against the world.
Draft not yet reviewed