The bridge hidden in an ordinary act

A strict identity theorist can say the painful episode just is a brain process, discovered empirically in the way a familiar phenomenon can turn out to be something described by science. Place and Smart defended versions of this strategy without claiming that “pain” means the same as a neuroscientific phrase. The proposal is an identity of what is reported, not a dictionary translation [2] [3].

A functionalist changes the emphasis. What makes the state pain is its place in a causal organisation: it tends to arise from damage, interact with beliefs and attention, motivate relief and alter behaviour. The same role might be realised in different biological or artificial materials. Putnam’s multiple-realisability argument made this flexibility central [4]. A non-reductive physicalist may similarly hold that the mental depends on the physical while remaining an indispensable higher-level description.

A dualist can preserve the distinctness of experience and thought, then face the old question posed sharply by Princess Elisabeth to Descartes: how can an unextended mind move an extended body? The objection is not that interaction is unimaginable. It is that the theory needs a relation that pairs this mind with this body and explains how influence enters a physical sequence [1].

Other positions reverse or dissolve the starting picture. Idealism takes the mental, or something mind-like, as metaphysically basic and treats physical order as derivative structure. Neutral monism and dual-aspect views place both descriptions over a basis that is neither simply mental nor simply physical. Panpsychism treats experiential features as fundamental and must explain how many small subjects or proto-experiential elements form one human subject. Process philosophy replaces enduring substances with events and relations, then owes an account of why a stable person appears across the flux.

These are more than different names for the same answer. Each changes which link requires explanation. Yet they share a discipline: every position must tell us what the bridge preserves and what it gives up. Without that discipline, “everything is physical” and “consciousness is fundamental” can both function as promissory notes.

Correlation is the beginning, not the bridge

Imagine that every reported pain episode in a study co-occurs with neural pattern N. That finding matters. It can support dependence, prediction and intervention design. It does not yet tell us whether pain is identical with N, realised by N, caused by N, grounded in N, merely correlated with N, or one aspect of a neutral process. The same observational table is compatible with several relations.

This is why more accurate brain imaging, by itself, does not dissolve the metaphysical question. Better measurement can narrow candidate mechanisms and expose false theories. It cannot decide a relation that no theory has stated. Chalmers called attention to the unanswered bridging question between functional availability and experience, while Levine named an explanatory gap even for those who remain physicalists [9] [8].

flowchart LR
  A["The hand that rose twice"] --> B["The bridge hidden in an<br/>ordinary act"]
  B --> C["Three tests every position<br/>must survive"]
  C --> D["Four controlled changes"]
  D -. "re-test" .-> A
  classDef first fill:#0a1424,stroke:#7B4CD8,stroke-width:2px,color:#e6f3ff
  classDef step fill:#0a1424,stroke:#3f7da0,stroke-width:1px,color:#e6f3ff
  class A first
  class B,C,D step

Three tests every position must survive

The phrase “mind-body theory” suggests one question. In practice it contains at least three. Keeping them separate prevents a solution to one from masquerading as a solution to all.

Test one: redescription

Take one episode: a sharp pain after touching a hot pan. Describe it from the first person, as a felt sting with urgency. Describe it from the third person, as neural, bodily and behavioural organisation. Now ask what makes these descriptions concern the same episode.

An identity view says the descriptions pick out one event under different concepts. Functionalism says the mental description abstracts a causal role that this physical organisation realises. Emergentism says the higher-level property depends on the lower level but may introduce explanatory regularities not visible at the microlevel. Dual-aspect and neutral monist views say both descriptions reveal aspects of a common basis. Idealism makes the physical description a stable structural abstraction within experience. Each supplies a different relation.

The test is not whether the relation sounds plausible. It is whether the theory states what must remain invariant across redescription. If the mental state is a role, specify its inputs, internal relations and outputs. If it is an identity, specify the level and grain at which identity is claimed. If it is an aspect, specify the common bearer and why the aspects co-vary. If experience is primary, derive why public physical descriptions are stable and shared rather than private or arbitrary.

A bridge that only says “both occur” is a correlation report. A bridge that says “these are one because relation R preserves properties A, B and C” becomes testable and criticisable.

Test two: intervention

Return to the raised hand. If a decision causes movement, what would differ had the decision differed? Causal language asks for a counterfactual. Yet the intervention must be coherent. If the decision is identical with a neural state, changing the decision while holding that very neural state fixed is like changing the morning star while holding Venus fixed. The proposed intervention violates the theory it is meant to test.

The classic causal-exclusion problem appears when a mental property M and its physical realiser P are treated as separate sufficient causes of the same bodily effect Y. If P is already sufficient, what causal work remains for M? If both independently cause Y on every occasion, the effect seems systematically overdetermined. If M does no work, it becomes epiphenomenal. Kim made this pressure central to the fate of non-reductive physicalism [11].

Interventionist accounts offer a different route. A higher-level variable can be causally informative when changing it through admissible manipulations changes an outcome across variations in its lower-level realisation. Temperature can cause pressure without competing with the motions of molecules as a second force. Likewise, an appraisal state may causally explain behaviour if interventions that alter appraisal, through varying neural routes, change action reliably. This does not prove a non-physical cause. It shows that the higher-level variable supports stable difference-making [12].

A more defensible higher-level test changes an appraisal process through different physical routes and asks whether a stable macro-variable predicts withdrawal or report across those routes. A positive result supports the causal usefulness of the macro description. It does not show that the macro cause floats free of its physical implementation.

Test three: reconfiguration

The first two tests can succeed while the subject remains undefined. Suppose a functional role is preserved across a neural prosthesis, remote memory store or duplicate. Which processes now constitute one experiencer? When does a replacement continue the same subject rather than create a successor? Which branch owns an earlier promise after a fork?

Subject individuation is not an optional ethical appendix. It determines the unit to which experience, responsibility and welfare are attributed. A panpsychist must explain composition. An idealist must distinguish local subjects within a mental reality. A biological naturalist must state which living organisation is necessary. A functionalist must say whether continuity of causal organisation is enough across duplication. A process view must explain how a temporally extended stream is one stream.

A theory that maps states and predicts effects can still fail to identify whose state or effect it is. This is why the bridge has three clauses rather than two.

What the strongest rival positions actually buy

It is easy to caricature the landscape as science on one side and mystery on the other. The strongest versions are harder to dismiss. They accept much of the same dependence evidence and disagree about what that evidence licenses. The bridge tests therefore compare commitments rather than levels of scientific seriousness.

Physicalism is broader than type identity. A physicalist may claim that every mental event is physical, that mental properties are realised by physical properties, or that the mental is grounded in the physical without being reducible to a single neural type. Its advantage is continuity with the causal success of physical science. Its pressure point is not the existence of neural mechanisms. It is whether those mechanisms explain subjective character and whether higher-level mental explanation remains genuinely causal rather than convenient shorthand. Kim’s exclusion argument targets this middle position, while interventionists reply that a macro-variable can be causally legitimate without becoming an extra force [11] [12].

Functionalism and biological naturalism disagree about what replacement must preserve. Functionalism treats the right causal organisation as constitutive, which gives it a principled route to multiple realisation. Its best form is not behaviourism. It includes internal state, recurrence, learning, counterfactual sensitivity and control. Biological naturalism accepts consciousness as a real, causally effective feature of the natural world while tying its production to biological processes. Its strength is refusing to infer consciousness from abstract computation alone. Its debt is to identify which biological powers matter, rather than using “biology” as an unanalyzed boundary [10].

Interactionist dualism takes experience and agency literally. It does not need to explain how non-experiential matter acquires experience because mind is already fundamental. Its hard problem is coordination: why this mind is paired with this body, how a mental difference changes a physical trajectory, and how the relation fits conservation and closure claims. Epiphenomenal property dualism avoids reverse influence by allowing the physical to produce experience without experience producing physical effects. That preserves irreducible phenomenality at the cost of making deliberation, pain and intention causally idle in the physical story.

Idealism reverses the direction of derivation. Instead of asking how matter produces experience, it asks how stable physical order arises within or from mind. This is not the claim that one person imagines the universe. Strong idealist views posit a mind-like ground or an experience-structured reality that is not reducible to any individual subject. Their bridge debt concerns public regularity, mathematical structure, the apparent autonomy of nature and the individuation of many finite perspectives [19].

Neutral monism and dual-aspect theory seek one bearer beneath the contrast. Physics may describe relational structure and dynamics while experience reveals intrinsic or first-person character. Russellian monism develops this possibility by proposing underlying properties relevant to both physical dispositions and consciousness. The move can avoid two independently interacting substances, but it must still specify how the underlying properties compose, cause and form a subject [17].

Panpsychism and process philosophy reject radical emergence in different ways. Panpsychism assigns experiential or proto-experiential character at a fundamental level, then faces the combination problem: why many small experiential bearers constitute one unified subject rather than a crowd. Process philosophy makes events, relations and becoming primary, so the static split between mental substance and material substance loses its grip. It must then recover enduring personal identity, precise intervention variables and the quantitative regularities described by neuroscience and physics [18] [20].

No option escapes by changing vocabulary. Each can be coherent, and each can be developed with scientific discipline. The common demand is a bridge specific enough to incur risk. A position advances when it names an observation, intervention or reconfiguration that would force revision, not when it merely redescribes every possible result as expected.

Four controlled changes

Thought experiments become useful when they move one causal property and hold the rest as steady as possible. They do not deliver verdicts by intuition. They reveal which clause a theory needs.

Replacement: change substrate, preserve organisation

The first intuition is continuity. Nothing abrupt happens between 49 and 50 per cent. Functionalism predicts preservation if the relevant organisation truly remains. A strong biological naturalism may predict failure when a specifically biological causal feature is lost, but it must identify that feature and a measurable break point. Type identity tied to a particular neural kind faces pressure from gradual replacement. Dual-aspect or neutral monist views can allow continuity if the replacement preserves the relevant underlying structure, but that cannot be assumed.

The hidden variable is the phrase “preserves function”. Does it include only visible behaviour, or recurrent dynamics, global availability, homeostatic regulation, developmental history and embodied coupling? A crude silicon lookup table can match a narrow input-output map while destroying the organisation that generated it. Substrate replacement tests a theory only after functional equivalence has been graded.

Counterfactual generation: preserve the report, change the mechanism

The shared sentence once looked like evidence of a shared state. The interventions reveal different generators. System A adapts coherently because the content participates in a wider organisation. System B fails outside its trigger. This does not prove that A experiences red. It defeats the inference from a report string to a mental mechanism.

The case also disciplines functionalism. A causal role is not one output. It is a network of counterfactual dispositions. The more of that network we preserve, the stronger the functional equivalence. Still, Nagel’s first-person challenge remains: an objective structural account can be complete in its own terms while leaving open why that organisation has a particular subjective character [6].

Boundary movement: preserve computation, move state outside

The obvious answer may track location, but location is not the only causal property. Tight, automatic, exclusive and continuously available coupling makes an external component function differently from a library consulted occasionally. The boundary may follow control, bandwidth, ownership, privacy, error correction or bodily regulation. A functionalist can extend the system boundary when the role is preserved. A biological view may keep the subject within the living organism even while acknowledging external assistance. A process or enactive view may treat the coupled loop as constitutive only while reciprocal dependence remains.

Now notice what the experiment does not establish. It can reveal the boundary of a task-performing system. It does not automatically reveal the boundary of a subject of experience. Cognitive extension and phenomenal extension are separate claims. A theory must supply the extra bridge.

Fork and reset: preserve history, duplicate the claimant

Before the fork, continuity is simple. After it, numerical identity cannot remain one-to-one with both branches under ordinary logic. Psychological continuity can be one-to-many, but responsibility and welfare then need lineage rules. A substance dualist might assign one non-physical subject to one branch, but must explain the pairing. A functionalist may treat both as legitimate continuers without claiming they are one individual. A process view can describe branching streams, yet still needs a norm for inherited commitments.

The reset adds another pressure. Restoring informational state does not undo the fact that branch B underwent an interval, unless the theory says only current configuration matters. If experience occurred, deletion may remove evidence without removing the past harm. The case therefore links metaphysics to governance: subject criteria determine whether copying, pausing and rollback are mere operations or interventions on moral patients.

stateDiagram-v2
  [*] --> Claim
  state "Correlation is the beginning,<br/>not the bridge" as Claim
  state "Test one, redescription" as Test
  state "Test two, intervention" as Revision
  Claim --> Test: expose mechanism
  Test --> Decision: evidence survives
  Test --> Revision: rival survives
  Revision --> Claim: narrow or rebuild
  Decision --> [*]

From metaphysics to a research contract

Consider a realistic configured system: a person uses a closed-loop neuroprosthesis that decodes intended movement, proposes a motor command, stimulates peripheral nerves and verifies the resulting motion. The system contains biological intention formation, learned decoding, software control, bodily feedback and conscious report. Asking “did the mind or the machine cause the movement?” is too coarse.

A useful analysis types the variables. Intent is the participant’s task-level aim. Neural activity provides a physical signal. The decoder proposes an action. Authority rules determine whether stimulation is permitted. Effect receipts record the resulting movement. The mental description may explain selection, the neural description implementation, and the control description error correction. These explanations need not compete if they answer different contrasts.

The mind-body claim begins only when we ask what constitutes the intention, whether it is causally efficacious as mental, and where the subject ends. A successful decoder shows that a neural signal contains usable information about intention. It does not prove that intention is identical with that signal, non-physical, or extended into the decoder. A lesion or stimulation effect can establish dependence and causal contribution at a physical level. It still needs a bridge to support an ontological conclusion.

The bridge contract

The following instrument turns a metaphysical slogan into a research object. It asks for six fields: a candidate boundary, the mental and physical variables, the relation between them, a coherent intervention, a subject-persistence rule and a rival prediction. The fields are deliberately asymmetric. A theory can pass the causal test and fail the subject test. It can offer a rich ontology and no discriminating observation. This discipline converts metaphysical disagreement into explicit experimental and design obligations.

{}

A positive result means the theory is articulated well enough to compare with a rival, design an intervention or make an architecture decision. It does not establish consciousness, identity or metaphysical truth. A negative result identifies a missing or contradictory clause. It does not refute the entire ontology. The instrument is a gate against premature inference, not a consciousness meter.

const relationDebt = {
  identity: "Explain the identity grain and why first-person character follows.",
  realisation: "Show that causal organisation, not one output, is preserved.",
  emergence: "Specify macro difference-making without double causation.",
  biological: "Name the necessary biological feature and replacement break point.",
  interaction: "Name a mediator or an explicit exception to physical closure.",
  epiphenomenal: "Accept that experience adds no physical causal difference.",
  idealism: "Derive public regularity and the individuation of finite subjects.",
  neutral: "Specify the neutral basis and how both aspects are derived.",
  panpsychism: "Specify combination and subject-summing conditions.",
  process: "Recover stable subjects and measurable intervention relations."
};

function evaluateBridge(contract) {
  const required = [
    "mapping", "intervention", "closure",
    "boundary", "persistence", "rival"
  ];
  const completed = required.filter(key => contract.clauses[key]);
  const missing = required.filter(key => !contract.clauses[key]);

  let status = "label only";
  if (completed.length >= 3) status = "descriptive bridge";
  if (completed.length >= 5) status = "causally articulated";
  if (completed.length === 6 && contract.candidate.trim()) {
    status = "test-ready bridge";
  }

  return {
    ...contract,
    status,
    completeness: completed.length / required.length,
    missing,
    relation_debt: relationDebt[contract.relation],
    permitted_claim:
      status === "test-ready bridge"
        ? "May compare preregistered rival predictions."
        : "May not infer a solved mind-body relation.",
    prohibited_claim:
      "This record cannot establish subjective experience or metaphysical identity."
  };
}

A minimal adversarial study

Apply two contracts to the same neuroprosthesis before collecting more data. An organisation-sensitive functionalist specifies the recurrent control relations, error correction, learning and counterfactual responsiveness that must be preserved. A substrate-sensitive biological naturalist specifies the biological power whose removal should break or alter consciousness, rather than merely performance. Both may predict intact movement while the device assists an existing pathway. Their disagreement begins only when the device substitutes for a relation one view treats as constitutive and the other treats as insufficient.

Build a replacement ladder rather than attempt an imaginary all-or-nothing transplant. At the first rung, vary decoding while biological planning remains intact. At the second, route feedback through the device and measure whether adaptation generalises when the mapping changes. At the third, use a validated computational twin to bypass progressively more of the biological control loop while preserving its input-output and recurrent profile. At each rung, collect task performance, error correction, ownership reports and confidence, but keep these measures separate. A report can change because expectation, latency or interface friction changed, even when the proposed constitutive relation did not.

Pre-register a prediction delta. Functionalism must name a preserved organisation whose disruption should matter even when gross output survives. Biological naturalism must name a biological property whose loss should matter even when that organisation survives. Include sham changes, blinded mappings and a lookup controller that reproduces familiar outputs without the relevant counterfactual depth. These controls test the rival explanation that apparent continuity is produced by demand characteristics or shallow behavioural matching.

The null policy is as important as the intervention. If every accessible measure agrees across the ladder, record the theories as empirically tied at that rung. Do not declare substrate independence, because the decisive biological feature may not yet have been removed. Do not declare biological necessity, because no break was observed. Spend on the next rung only when the contracts state a new prediction difference. This protocol strengthens the bridge-friction hypothesis if explicit clauses reveal divergence before the technically impossible endpoint. It weakens the hypothesis if repeated, well-specified ladders leave every mature rival observationally and interventionally equivalent.

What the contract cannot settle

The contract does not privilege physicalism. Consciousness may be fundamental, emergent, biologically constituted, processual or one aspect of a neutral reality. The instrument asks each view for the same clarity: source structure, target structure, preserved relation, break condition and discriminating observation.

It also does not convert first-person evidence into a third-person variable by decree. Reports are evidence of what a person says and, under ordinary human conditions, strong evidence of experience. Their interpretation depends on trust, training, elicitation and mechanism. In artificial or altered systems, the link must be rebuilt rather than borrowed.

Finally, the contract cannot guarantee that a decisive intervention is possible. If mental and physical variables are identical or linked by supervenience, independent manipulation may be incoherent. If rival theories predict the same accessible data, the honest result is underdetermination. An explicit null is progress when it blocks an invalid victory.

sequenceDiagram
  participant Q as Question
  participant M as Mechanism
  participant E as Evidence
  participant D as Decision
  Q->>M: The hand that rose twice
  M->>E: Correlation is the beginning, not the bridge
  E-->>M: compare rival explanations
  E->>D: Believe the theory only as far as its bridge holds
  D-->>Q: revise scope or act

Believe the theory only as far as its bridge holds

Stop asking a system, a brain scan or a philosophical slogan to answer “Is mind physical?” in one move. That question compresses description, causation and subjecthood into a single word and then hides the bridge that would make an answer meaningful.

For any consequential claim, require a one-page bridge contract. Name the mental phenomenon, the physical or neutral process, the relation proposed, the intervention the relation permits, the subject boundary it assumes and the rival result that would weaken it. Reject tests that independently manipulate variables the theory identifies. Reject output matching as evidence of preserved organisation. Reject a causal explanation that silently becomes an identity claim.

This rule changes experimental design, AI evaluation, neurotechnology governance and philosophical argument. It allows plural explanations when they answer different contrasts, while preventing them from double counting one effect. It lets consciousness-primary views enter the discussion without granting them empirical victory by assertion. It lets physicalist views use dependence evidence without treating dependence as identity.