Part I · The grammar that invents an owner

A sentence can smuggle in a metaphysics

“Who sees?”, “Who chooses?”, and “Who suffers?” sound like requests for missing information. Each has the grammatical form of an open question. Yet each also presupposes that an event must belong to a subject in the way a coat belongs to its wearer. The hidden step is easy to miss because many actions really do require a role-holder. If a payment was approved, a bank needs to know which authorised officer approved it. If a door opened, an engineer may need to know which actuator moved it.

Experience is not automatically organised in the same way. From the fact that seeing occurred, it does not follow that an additional seer-entity caused the seeing. The visual event may already involve an eye, a visible form, illumination, attention, learned discrimination and consciousness conditioned by that encounter. Asking for another object called the seer may duplicate the work already performed by the relation.

This is the force of the redirection in SN 12.12. It is not a ban on pronouns. Early Buddhist texts speak ordinarily of persons, monks, parents, kings and practitioners. The challenge concerns explanatory placement. A conventional subject can organise communication without becoming an extra cause inside the process.

The aggregates are not five replacement souls

The best-known early Buddhist analysis applies anattā, “not-self”, to five khandhas: form, feeling, perception, formations and consciousness. In the Anattalakkhaṇa Sutta, SN 22.59, each is examined as impermanent, liable to affliction and not commandable at will. None warrants the stance “this is mine, this I am, this is my self”.

A common diagram draws a human silhouette divided into five labelled compartments. That is convenient but risky. It makes the aggregates look like a rival substance theory: the old soul has been removed, and five smaller things now constitute the person. Rupert Gethin’s study of the Nikāyas and early Abhidhamma argues for a more dynamic reading. The aggregates are recurring classes of phenomena through which conditioned existence is experienced. They arise and pass in relation to sense activity; as aggregates affected by clinging, they include the grasping as well as what is grasped.⁸

Consider hearing your name in a crowded station. There is bodily form, including the ear and nervous system. There is the tonal feel of the moment, perhaps pleasant or alarming. Perception recognises a familiar sound pattern. Formations include orientation, expectation and the impulse to turn. Auditory consciousness occurs dependent on the relevant encounter. These are analytical cuts through one event. They are not five independent agents taking turns at a control panel.

The aggregates dissolve ownership by distributing the work, not by erasing the event. Feeling still hurts. Perception can still misclassify. Intention can still redirect behaviour. Consciousness still occurs. What disappears is the inference that one permanent item must own all five in order for them to interact.

Not-self is not annihilation

The first predictable reaction is nihilism: if no enduring owner can be found, perhaps nobody exists and nothing matters. Early Buddhist texts resist this binary. The Kaccānagotta Sutta, SN 12.15, warns against reducing the world to the poles of “everything exists” and “everything does not exist”. Its alternative is to observe arising and cessation through conditions.

This matters because “there is a permanent self” and “there is absolutely nobody” share an assumption. Both expect reality to be settled by locating or deleting a thing. A process account changes the unit of analysis. A promise can be kept by a later configuration because it inherits memory, social expectations, bodily capacities and concern from an earlier configuration. A wound can be treated because present conditions were caused by earlier conditions. The word person remains useful as a name for this organised history.

No-self removes a metaphysical privilege, not the causal texture of a life. Pain is not made unreal. Intention is not made causally idle. Conventional identity is not forbidden. What is denied, or at least placed under disciplined suspicion, is the move from coordinated change to an independent, permanent proprietor.

flowchart TD
  A["The monk who asked the wrong<br/>question"] --> B["Part I · The grammar that<br/>invents an owner"]
  B --> C["Part II · Continuity without a<br/>carrier"]
  C --> D["Part III · Agency as<br/>conditioned control"]
  D -. "re-test" .-> A
  classDef first fill:#0a1424,stroke:#D95545,stroke-width:2px,color:#e6f3ff
  classDef step fill:#0a1424,stroke:#3f7da0,stroke-width:1px,color:#e6f3ff
  class A first
  class B,C,D step

Part II · Continuity without a carrier

The carrier fallacy

A flame moves from one candle to another. Nothing material called “the flame” travels intact. Combustion at the first wick creates the conditions for combustion at the second. Yet the second flame is not unrelated to the first. The relation is causal without being the transport of a permanent object.

The analogy is limited, but it exposes a general mistake. We often infer that every continuity must be carried by one continuous item. A company retains obligations while employees, computers and offices change. A melody remains recognisable while no note persists for the duration. A river keeps a course while its water passes. In each case, what continues is an organised pattern of constraints and dependencies.

Early Buddhist accounts place consciousness inside such dependency rather than outside it. In MN 38, the monk Sāti claims that the same consciousness runs and wanders through repeated existence. The correction is emphatic: consciousness arises dependently and is classified by the condition through which it occurs, such as eye-consciousness or ear-consciousness. The text therefore blocks the easiest candidate for a migrating owner.

The Mahānidāna Sutta, DN 15, presents consciousness and name-and-form in a mutually constraining relation. Interpretations of that relation differ, especially across models of rebirth, but the structural lesson is clear: consciousness is not introduced as an isolated lamp that would shine identically without a body, objects, recognition or conditions.

Continuity is better modelled as inheritance of constraint than as transport of essence. A later state carries consequences because earlier events changed the field from which it arises.

The smallest useful formal model

We can compress the intuition without pretending that an equation proves a Buddhist doctrine. Let the state of an episode at time t be:

Every term names something in the situation. Body includes the available sensory and motor condition. Feeling marks pleasant, unpleasant or neutral tone. Recognition classifies what is encountered. Dispositions include intentions and learned tendencies. Awareness-of-object is the relevant episode of consciousness. The function F says only that the next configuration depends on the current configuration and new conditions.

Take a toy numerical case. An unpleasant message produces feeling at -0.7 on a scale from -1 to 1. A learned threat-recognition bias contributes 0.6. A retaliatory disposition contributes 0.5. A pause practice contributes -0.8 to response urgency. Add them in this deliberately simple model: -0.7 + 0.6 + 0.5 - 0.8 = -0.4. The sign still indicates aversion, but the magnitude is below a hypothetical action threshold of 0.6. The reply is delayed.

Remove the pause condition and the magnitude becomes 0.4 in the other direction only if we define signs poorly, which reveals why the model must separate valence from urgency. A better calculation uses absolute unpleasantness as activation:

Without the pause, urgency rises to 1.13 and crosses the threshold. The example is synthetic. Its purpose is not to quantify a mind. It shows what an owner-free model must do: preserve state, explain how conditions combine and produce a counterfactual prediction. If changing a condition changes the action, the model has located leverage without invoking a permanent controller.

The residue of “I am”

Conceptual agreement does not instantly remove the felt centre. The Khemaka Sutta, SN 22.89, distinguishes not identifying any aggregate as self from eliminating the residual conceit “I am”. This is psychologically important. A philosopher may accept process theory while still flinching at insult, defending status and organising memory around a privileged centre.

The distinction prevents a common failure. One may replace “I am a permanent self” with the equally rigid identity “I am someone who has no self”. The proposition changes while appropriation continues. Early Buddhist practice therefore targets the recurrent activity of taking phenomena as “mine”, rather than only a sentence in a belief system.

Part III · Agency as conditioned control

If nobody owns the action, can it still be an action?

The strongest practical objection to no-self is not about memory. It is about agency. A falling stone causes damage, but it does not apologise. A person can notice a reason, inhibit an impulse, form an intention and act differently. Does that capacity require a permanent subject?

Early Buddhism does not solve the problem by denying intention. In the Nibbedhika Sutta, AN 6.63, intention, cetanā, is identified as kamma, ethically consequential action. Body, speech and mind express what has been intended. The same discourse asks for the condition of action and points back to contact. Agency is therefore neither an uncaused command nor mere bodily motion. It is a conditioned process in which intention has a distinctive causal and ethical role.

A modern control term helps, provided we do not project it back as a translation. Call an action agentive when it is counterfactually sensitive to reasons and retained goals. If evidence changes, the intended action can change. If a prohibition is remembered, an impulse can be vetoed. If consequences are anticipated, a plan can be revised. These capacities require state, discrimination and inhibition. They do not logically require an indivisible owner.

Conditioned agency is not the absence of control. It is control whose causes can be inspected, trained and interrupted. The permanent-owner picture can actually obscure this work by attributing a decision to “the self” after it has occurred.

A worked episode: the late-night message

A project lead receives a terse message at 11:47 p.m.: “This is not acceptable. Fix it before morning.” The ordinary summary is “I became angry but decided not to reply.” The aggregate and conditional analyses make the episode more detailed without denying the person.

The unpleasant feeling is not yet the angry email. Perception supplies a learned meaning: “I am being humiliated.” Formations activate a repertoire: counterattack, appease, seek evidence, wait. A retained commitment to avoid late-night escalation enters the field. Intention settles, for now, on saving a draft and sleeping. In the morning, new bodily conditions and additional context alter the appraisal.

What acted? The project lead did, at the conventional level required by work and accountability. At the explanatory level, the action emerged through conditions that included training, memory and foresight. The two descriptions answer different questions. The conventional description identifies the person to whom colleagues can speak. The process description identifies the levers that made restraint possible.

This also clarifies responsibility. If the reply had been abusive, “there is no self” would not erase the causal history or the need for repair. The episode would still be connected to learned habits, choices, ignored warning signs and consequences for others. Responsibility tracks the organised lineage of intention and effect, not possession by a metaphysical substance.

Where intervention has leverage

The chain from contact to feeling, craving and clinging is often drawn as fate. That misses the practical point. A conditional relation is precisely what makes intervention meaningful. Remove or alter a condition and the downstream trajectory can change. Training may change recognition. Ethical precepts can constrain the action set. Attention can detect the unpleasant tone before it becomes a demand for retaliation. Social arrangements can reduce the frequency of provocative contact.

The following chart uses the toy urgency calculation from Part II. Everything except the pause or equanimity term is held constant. It is not a measurement of meditation and does not claim linear human response. It visualises one implication: if action pressure is conditionally produced, a modest gate can move a system across a behavioural threshold.

The counterexample: compulsion is also conditioned

Conditioning alone does not guarantee agency. A reflex, seizure, coerced act and rehearsed compulsion are all caused. The stronger criterion is organised counterfactual control. Could reasons enter? Was an alternative represented? Could inhibition operate? Did the state preserve the relevant commitment? These questions can yield degrees rather than a binary verdict.

This is where a simplistic Buddhist analogy can fail. Saying “all action is conditioned” does not mean all actions are equally voluntary or ethically equivalent. Conditions can destroy control as well as enable it. Severe sleep loss, intoxication, threat and manipulation may narrow the field. Training, supportive institutions and truthful feedback may widen it.

No-self should increase causal discrimination, not flatten it. The more precisely we understand which conditions supported or impaired control, the more intelligently we can assign repair, protection, retraining and accountability.

stateDiagram-v2
  [*] --> Claim
  state "A sentence can smuggle in a<br/>metaphysics" as Claim
  state "The aggregates are not five<br/>replacement souls" as Test
  state "Replace the witness, one<br/>candidate at a time" as Revision
  Claim --> Test: expose mechanism
  Test --> Decision: evidence survives
  Test --> Revision: rival survives
  Revision --> Claim: narrow or rebuild
  Decision --> [*]

Part IV · Intervening where the person is made

A new causal distinction: owner-required versus owner-excess

The phrase “there is no self” invites an argument at the largest possible scale. That is often unproductive because the participants mean different things by self: an immortal soul, pure awareness, a biological organism, a narrative identity, a control model or simply the person named on a passport.

A more disciplined test begins with one target phenomenon. Suppose the target is delayed retaliation after criticism. Build the smallest model that predicts the delay using bodily state, feeling, recognition, dispositions, memory, norms and intention. Then ask whether a further owner-variable changes any counterfactual. If setting the owner to a different value changes nothing unless one of the ordinary process variables also changes, the owner is excess for that explanation.

This yields the owner-excess test:

The test is intentionally local. Passing it for message restraint does not prove that no self exists in any sense. Failing it does not prove a metaphysical self. Failure may expose an omitted biological variable, an unrecorded social condition, a hidden memory state or an inadequate time scale. The test replaces an all-or-nothing metaphysical contest with a sequence of causal obligations.

Apply the test without begging the question

The test has four stages. First, state the phenomenon precisely. “Human agency” is too broad; “whether a remembered rule can inhibit a prepared reply” is tractable. Second, draw the condition graph and include the states that actually persist. Third, intervene on those conditions rather than merely observing correlations. Fourth, add the owner candidate and specify what unique difference it should make.

The last stage is the hardest. “The owner makes it mine” is not yet a causal prediction. Would an identical pain cease to guide avoidance if ownership were absent? Would a commitment fail to influence action? Would the system lose unity, report or ethical concern? Each possibility names a measurable bridge. Without such a bridge, the owner-label redescribes the phenomenon.

What the lab tests

The lab encodes six assumptions. Contact varies across events. Unpleasant feeling is converted into activation through a learned threat gain. Habit adds persistence. Pause strength reduces the transition from activation to craving. Craving accumulates into clinging. Clinging and current activation jointly raise action pressure. Every state is bounded between zero and one so the display remains interpretable.

Run the default configuration, then select “Apply stronger pause”. The intervention changes one variable while preserving the earlier conditions. A lower action trajectory supports a limited claim: in this model, a downstream behavioural difference can be produced by altering a condition rather than by issuing a command from a central owner. It does not show that human minds use these equations, that practice has this dose-response curve or that consciousness is reducible to the model.

function step(state, input, t) {
  const contact = input.contact * (0.9 + 0.1 * Math.sin(t * 1.7));
  const activation = clamp(
    Math.abs(input.valence) * input.threat + 0.35 * contact + state.habit - input.pause,
    0, 1.5
  );
  const craving = sigmoid(4 * (activation - 0.55));
  const clinging = clamp(0.72 * state.clinging + 0.45 * craving, 0, 1);
  const action = sigmoid(3.2 * (activation + clinging - 1.15));
  const habit = clamp(state.habit + 0.06 * action - 0.035 * input.pause, 0, 1.2);
  return { habit, clinging, craving, action };
}

To reuse the artefact, save this page and open it in any modern browser. No server is required. To adapt it for research, replace the hand-set equations with preregistered hypotheses, expose each latent state to independent measurement where possible, and compare it against rival models. A positive result permits only the conclusion specified by the model comparison. A negative result should lead to model revision, not metaphysical triumph.

The metaphysics remains underdetermined

A process explanation does not automatically imply physicalism. A physicalist can say the relevant conditions are wholly physical. A biological naturalist can hold that the organisation depends on distinctive living processes. A functionalist can identify mentality with causal roles that may be realised in more than one substrate. An emergentist can argue that higher-order agency is real although dependent on lower-order processes.

An idealist can treat the changing processes as modulations within awareness rather than matter. Neutral monism can regard mental and physical descriptions as aspects of a more basic reality. Panpsychism can attribute experiential features to fundamental constituents while still denying a permanent personal owner. Dual-aspect theory can hold that one underlying process has mental and physical aspects. Process philosophy can make events and relations metaphysically basic.

Early Buddhist anattā does not adjudicate all these modern positions. It does, however, press the same question against each: where exactly is ownership doing causal or liberative work? No-self is compatible with serious consciousness-primary inquiry so long as conditioned consciousness is not silently equated with an eternal personal witness.