Psychophysical Harmony Argument for Theism

The argument from psychophysical harmony (PPH), proposed by Brian Cutter and Dustin Crummett (forthcoming, Oxford Studies in Philosophy of Religion), is a Bayesian argument for theism that proceeds from facts about the relationship between conscious experience and physical/functional brain states. It claims that the alignment between phenomenal states (e.g., pain) and physical states (e.g., tissue-damage-detection + avoidance behavior) is “too lucky” to be expected under naturalistic atheism, but is expected under theism, because God has reason to design psychophysical laws that realize valuable harmonious correlations. ¹

The Two Types of Harmony

Normative Harmony

Normative harmony concerns the match between the normative role of a conscious experience (what it gives us reason to do) and its functional/physical role (what behavior it actually produces). Cutter and Crummett distinguish several sub-types:

• Hedonic harmony: Pain — an intrinsically bad experience that gives reason to avoid its source — is lawfully correlated with physical states that actually produce avoidance behavior. The psychophysical laws could have mapped the pain-role physical state onto pleasure (pleasure/pain inversion), or onto an evaluatively neutral state, resulting in organisms that systematically pursue what they have reason to avoid. That they don’t is claimed to be extraordinarily lucky. ¹

• Epistemic harmony: A visual experience as of a round object — which essentially provides justification for believing there is a round object — is correlated with physical states that produce belief-forming and behavioral responses as if there is a round object. The laws could have mapped the same physical state onto a phenomenal presentation of a triangle, or random static, destroying the harmony between experiential justification and functional role. ¹

• Cognitive harmony (following Goff 2018): If cognitive phenomenal states (occurrent phenomenal beliefs/desires) are distinct from functional states, then the fact that they tend to be matched in rationally appropriate ways — phenomenal beliefs that rationalize behavior, phenomenal beliefs that are appropriate to sensory evidence — calls for explanation. This sub-type requires the controversial premise of cognitive phenomenalism. ¹

Semantic Harmony

Semantic harmony concerns the match between our verbal reports/judgments about consciousness and the actual phenomenal states that make those reports true. Under dualism + causal completeness, there is a complete physical explanation of why we say “I’m in pain” or “consciousness is non-physical” that makes no reference to actual pain or non-physical consciousness. It would be a fortunate coincidence if the physical mechanism output the sentence “I am experiencing consciousness” and that sentence happened to be true. Cutter and Crummett illustrate this with the “Text Machine” analogy: a machine whose text output (“HIGH,” “LOUD,” “RISING”) semantically corresponds to its sound output (high-pitched, loud, rising) cries out for explanation beyond a mere common cause. ¹

The Bayesian Structure

The argument has the form of a likelihood comparison:

P(harmony | theism) ≫ P(harmony | atheism)

The key claims:

1. The vast majority of conceivable psychophysical mappings would be disharmonious (including the simplest mappings, e.g., mapping all states onto buzzing noise).
2. On theism, harmonious laws are not terribly unlikely, because harmony is valuable (enables rational agency, knowledge, meaningful action).
3. On naturalistic atheism, there is no reason to expect harmonious over disharmonious mappings.
4. The prior probability of theism is not “fantastically low” (it posits a mind with a few natural properties to a maximal degree — simplicity confers intrinsic probability).
5. Therefore, conditionalizing on psychophysical harmony significantly raises the posterior probability of theism. ¹

Cutter and Crummett provide an illustrative calculation: if harmony is 50× more likely on theism than atheism, and the prior ratio P(theism)/P(atheism) = 1/10, then after conditioning, P(theism)/P(atheism) = 5. ¹

Robustness Claims

Cutter and Crummett argue the argument survives dropping its initial dualist assumptions:

• Interactionist dualism doesn’t help: the phenomenal-to-physical causal laws could have pushed in any direction; harmony requires a very special form of these laws.
• Physicalism/functionalism doesn’t help: functionalism says pain = some functional property, but doesn’t predict it equals the harmonious one specifically — the a priori epistemic probability of harmonious identity is no higher than disharmonious alternatives.
• Russellian monism faces the same “experiential lottery” problem: why do we have the right combination of quiddities?
• Idealism is even worse off without God (the order of experience requires explanation).
• Multiverse doesn’t help: psychophysical harmony is not subject to an observation selection effect — it would be surprising to find ourselves in a harmonious universe even given many universes. ¹

Critical Analysis: Multiple Angles

1. The Evolutionary Response

Summary: Evolution by natural selection shapes the mapping between mental states and physical behaviors because organisms with adaptive sensory-to-experience correlations survive and reproduce. Those that felt pleasure from harm or pain from food died out.

Cutter and Crummett anticipate this: they argue evolution presupposes but does not explain normative harmony. The physical functional state (tissue damage detection → avoidance behavior) has a straightforward evolutionary explanation. But the question is why the psychophysical laws map that functional state onto an experience (pain) whose normative role (badness, reason to avoid) harmonizes with the functional role. Since “evolutionary forces cannot affect the psychophysical laws,” evolution cannot explain the harmonious correspondence — it merely presupposes it. ¹

Critical evaluation of the evolutionary response:

This exchange hinges on a crucial ambiguity in what “evolution explains.” The Cutter-Crummett framing treats psychophysical laws as fundamental background constants that predate biological evolution — like gravity, they specify “if physical configuration X occurs, conscious experience Y manifests.” Evolution merely assembles the physical configuration; the background law determines the experience. On this framing, evolution cannot touch the laws, only work within them.

But this framing begs the question against physicalism and functionalism. If consciousness is a structural/functional property of complex matter (as physicalists hold), then there are no independent psychophysical laws to be “lucky” about. Pain just is the brain’s integrated aversive system for registering tissue damage and commanding avoidance — they are the same thing, and “inverting” them is metaphysically impossible, not merely improbable. On this view, evolution didn’t “find” a pre-existing law; it built the system whose functioning constitutes the experience. The co-evolution of behavioral complexity and experiential complexity was not a lottery but a gradual, continuous process: simple chemical sensitivities in single cells → basic nociception → integrated pain systems, each step selected for its adaptive value. ²

The deeper evolutionary objection is that the argument misrepresents biology by treating the mind-body link as a single improbable lottery pick on raw nerve fibers, ignoring how hierarchical neural networks gradually evolved over millions of years to process and integrate sensory data. The “combinatorial explosion” of possible mappings that depresses P(harmony | atheism) is an artifact of treating each possible brain-state–experience pairing as an independent random variable — a presupposition that is only plausible under dualism, which is precisely what physicalists deny. ²

2. The Bayesian Critique

Summary: The Bayesian formula P(harmony | theism) ≫ P(harmony | atheism) collapses if the prior probability assignment is fundamentally flawed. Three hidden assumptions drive the artificially low P(harmony | atheism):

• Combinatorial Explosion Assumption: Every physical state can be paired with any conscious state randomly.
• Principle of Indifference: Every conceivable chaotic mapping is equally likely a priori.
• Logical Independence Assumption: Physical laws and psychophysical laws have no structural relation.

²

Critical evaluation:

(a) Illegal reification of laws. The argument assumes psychophysical laws exist as independent, top-down cosmic constants dictating how a brain behaves from the outside. If a physicalist or functionalist view is correct, consciousness is a structural property of complex matter — not a separate substance pinned to it by a law. Under constitutive identity, the “prior probability” of mind-body divergence is not low but zero: a specific brain state couldn’t map to a chaotic experience any more than H₂O could “map” to behaving like gold. The combinatorial explosion is calculated over a space of possibilities that is metaphysically empty. Cutter and Crummett’s response to this (§3.2) is that functionalism subdivides into many specific identity hypotheses (pain = F₁, pain = F₂, …), and there’s no a priori reason to prefer the harmonious one. But this response itself presupposes that the “harmony” of the identity is a contingent fact requiring explanation — which is precisely what constitutive physicalists deny. An identity is not a correlation; it does not “cry out for explanation” in the way a nomological coincidence does. ^{1 2}

(b) Begging the question against non-theistic order. The argument assumes that a random, chaotic mapping is the “default” state of a non-theistic universe — that without God, we should expect disharmony. But nature exhibits intrinsic mathematical structure at every level (fractals, quantum symmetries, thermodynamic efficiency, self-organizing systems) without needing a conscious director. Assuming that non-theistic psychophysical laws must be a chaotic lottery, while theistic ones are neat and tidy, is not an objective probabilistic calculation — it is, as the SEP’s general assessment of ontological-arguments notes, “dressing up a theological preference in mathematical notation.” ²

(c) The reference class problem. The argument’s force depends entirely on what counts as the “reference class” of possible psychophysical laws. If the reference class includes only laws compatible with physicalism (where experience is constituted by function), then P(harmony | atheism) = 1 and the argument collapses. If the reference class includes all conceivable dualist mappings (where experience is independent of function), then P(harmony | atheism) is vanishingly small. The argument’s power derives entirely from which reference class you assume — and that assumption is the very point at issue. ¹

3. The Lance Bush / Type-A Physicalism Response

Summary: Lance Bush (2023) argues that the entire explanandum dissolves if we adopt type-A physicalism — the view that physicalism is true and there is not even an epistemic gap between physical and phenomenal truths. If there are no phenomenal states (in the robust, irreducible sense the argument presupposes), then there is no psychophysical harmony to explain. ³

Cutter and Crummett preempt this by explicitly assuming “the falsity of so-called ‘a priori’ (or ‘type-A’) physicalism.” But as Bush argues, this is question-begging: the quality of an argument is only as good as its assumptions. If type-B physicalism (which accepts an epistemic gap) fails to explain psychophysical harmony, the rational response is not automatically “therefore theism” — it could equally be “therefore type-A physicalism.” Problems for (3) type-B physicalism increase the plausibility of both (1) theism and (2) type-A physicalism. ³

Bush invokes Williamson’s principle: “one philosopher’s modus ponens is another’s modus tollens.” If PPH leads to theism, and theism is false, we can reject PPH (and the phenomenal states it presupposes) rather than accept theism. Type-A physicalism offers a clean denial of the explanandum — which is, Bush notes, a virtue: a theory that doesn’t suffer from a problem is in a better position than one that does. ³

Critical evaluation:

Bush’s response is powerful against the argument as Cutter and Crummett present it, but it faces the standard objection to type-A physicalism: most philosophers (and most people) find it self-evident that there is something it is like to be in pain, and that this phenomenal character is not deducible from the physical/functional description. Type-A physicalism is widely regarded as requiring the denial of evident facts about consciousness. However, Bush argues that the typical objections to type-A physicalism consist of “little more than claiming that it’s obviously wrong” — which is not a substantive philosophical argument. The chinese-room-argument and the turing-test debate show that the boundary between “obvious” and “argued” is itself contested in philosophy of mind.

4. The Constitutive Identity / Functionalism Response

Summary: Physicalists and functionalists bypass the entire dilemma by rejecting the premise that psychophysical laws exist as independent background constants. On this view:

• Constitutive identity (no separate laws): A conscious experience is the physical functioning of the evolved organ. Pain is just the brain’s integrated system for registering tissue damage and commanding avoidance. They cannot be separated or inverted because they are the same thing.
• Co-evolution of mind and matter: The mind-body mapping did not wait for a fully formed organ to suddenly connect to a pre-existing cosmic law. Basic cells developed simple chemical sensitivities; as organs evolved greater behavioral complexity, internal experiential complexity co-evolved as an inescapable structural property of that architecture.
• Rejecting external norms: Pain doesn’t feel bad because of an independent moral law; it feels bad because evolution sculpted the experience to be inherently aversive. The “normative role” of pain is not a metaphysical fact independent of its functional role — it supervenes on it. ²

Cutter and Crummett’s response (§3.2) is subtle: functionalism says pain = some functional property, but the general thesis of functionalism doesn’t predict that pain = the specific harmonious functional property F rather than some disharmonious F*. Among all the a priori possible psycho-functional identity hypotheses, there’s no reason to prefer the harmonious one. So P(harmony | atheism & functionalism) ≈ P(harmony | atheism & dualism) ≈ very low.

But this response has a structural flaw: it treats the identity “pain = F” as if it were a discovery that could have gone either way — like discovering that Hesperus = Phosphorus rather than Hesperus = some other planet. For the physicalist, the identity is not a contingent fact waiting to be discovered; it is a metaphysical necessity grounded in the nature of the properties involved. If pain just is aversive functional integration (and this is constitutive, not merely nomological), then there was never any “lottery” — the harmonious identity obtains in every possible world where the functional state exists. The “other possible identities” (pain = F₂, pain = buzzing noise) are not metaphysically possible scenarios that were “rolled against” — they are a priori epistemic possibilities that don’t correspond to anything. ¹

5. The Sceptical / Meta-Level Response

At a meta-level, the argument faces the same structural vulnerabilities as other Bayesian arguments for theism:

• subjectivity-of-priors: The prior probability of theism is itself deeply contested. Cutter and Crummett argue it’s not “fantastically low” because theism is simple (a mind with a few natural properties to a maximal degree). But this is the same Swinburne-style reasoning that has been extensively criticized (see dwindling-probabilities, agrippas-trilemma).
• dwindling-probabilities: Even if P(harmony | theism) is higher than P(harmony | atheism), the full theistic hypothesis is a chain — God exists → God is good → God wants conscious beings → God designs harmonious laws — and each link introduces its own probability penalty.
• The problem of evil: If God designed psychophysical laws for harmony, why is there so much disharmony — phantom limb pain, migraine, chronic pain without tissue damage, depression? Cutter and Crummett don’t address this, though it bears directly on P(harmony | theism).

6. Stack Exchange Critiques: Identity, Evolution, and Explanatory Underdetermination

Two answers to the Stack Exchange discussion sharpen the preceding physicalist and evolutionary objections. They are useful as criticisms of the argument’s framing, although neither answer by itself settles the metaphysics of consciousness.

NotThatGuy: the alleged lottery may be an artefact of the hypothesis space

NotThatGuy argues that the paper is primarily effective against a dualist naturalist picture in which phenomenal states are distinct from physical states and are attached to them by contingent laws. On reductive materialism or a constitutive form of functionalism, the relevant physical state is not merely correlated with pain: it is (or realizes) pain. The alignment between nociceptive processing, aversive experience, and avoidance behavior is therefore not a lucky match between independent items. It is the operation of one evolved system described at different levels. NotThatGuy’s answer makes the related evolutionary point that selection favors neural circuits linking sensory input to adaptive action; under this model, systematic psychophysical alignment is less surprising than its absence. ⁴

This directly challenges the paper’s treatment of functionalism. Cutter and Crummett grant that, conditional on functionalism, harmony may have probability close to 1, but argue that functionalism is only a small region of the atheist probability space. NotThatGuy’s reply is that this gets the mixture weights backwards: evolutionary evidence is a reason to assign substantial prior weight to physicalist/functional explanations, not a reason to dismiss them because many other naturalistic hypotheses fail to predict harmony. The car-versus-magic analogy in the answer illustrates the error: the fact that most candidate explanations are poor does not make the one explanation that independently fits the causal evidence improbable merely because it is specific. ⁴

The force of this criticism depends on what kind of “functionalism” is under discussion. If it is a constitutive identity claim, the alleged alternative mappings (for example, the same functional organization accompanied by an arbitrary pain/pleasure inversion) may be metaphysically unavailable rather than low-probability outcomes. If it is only a contingent psychophysical correlation or a non-reductive realization relation, Cutter and Crummett’s epistemic-probability objection remains more intelligible. Thus the dispute is not resolved by Bayes’ theorem alone; it concerns the admissible hypothesis space and the prior weighting of physicalist theories.

NotThatGuy also objects that “we cannot explain this without God” is not yet a positive case for theism. An unexplained coincidence can motivate a search for an explanation, but it does not by itself discriminate between theism, a non-theistic fundamental psychophysical law, or other sufficiently specified alternatives. To establish a theistic likelihood advantage, the argument would need an independently justified account of why a good personal creator should make harmony more probable, together with a principled prior comparison against rival explanations—not merely the observation that harmony seems difficult to understand. ⁴

variableization: biological systems can produce real psychophysical variation

variableization presses the evolutionary objection at the level of actual nervous systems. Congenital insensitivity to pain can result from loss-of-function variants in SCN9A, which disrupt NaV1.7 sodium channels in nociceptors and prevent injury signals from reaching the brain (MedlinePlus Genetics). The condition illustrates that the presence, absence, and adaptive efficacy of pain experience depend on an evolved and genetically variable biological implementation. It is not necessary to imagine a whole universe whose experiences have been randomly remapped: changes to the organism’s sensory machinery can generate substantial psychophysical mismatch, including failure to avoid injury. ⁴

Synesthesia supplies a different kind of example. In synesthesia, a stimulus or concept can automatically elicit an additional phenomenal experience without a corresponding external sensory input; mappings are often individual-specific and relatively stable (van Leeuwen, Singer, & Nikolić 2015). This supports the more modest claim that neural connectivity and processing architecture can alter the mapping between stimuli and experience, rather than consciousness being a fixed, biologically inert accompaniment to an otherwise complete physical story. ⁴

These cases should not be overstated. They demonstrate variation in the biological mechanisms that mediate nociception and perception; they do not by themselves establish that evolution changes fundamental psychophysical laws in the strong dualist sense used by Cutter and Crummett. A dualist can reply that the same laws apply to different physical configurations, or that a damaged system fails to instantiate the relevant physical state. Nevertheless, the cases weaken a sharp premise that evolution is wholly irrelevant to psychophysical harmony and make it less obvious that naturalistic harmony should be assigned an extremely small likelihood. They also show that “harmony” is graded: ordinary pain/avoidance alignment coexists with congenital analgesia, maladaptive pain, cross-sensory associations, and other departures from ideal fit.

Assessment

The argument is philosophically sophisticated and genuinely novel — it identifies a real explanandum (the apparent non-accidental fit between experiential and functional roles) that is distinct from traditional consciousness-to-God arguments. Its claim that the argument survives under multiple metaphysical frameworks (dualism, physicalism, idealism, Russellian monism) is ambitious. The multiverse immunity claim — that psychophysical harmony, unlike cosmological fine-tuning, is not subject to an observation selection effect — is a genuine structural advantage over the fine-tuning argument. ¹

However, the argument’s force is highly sensitive to background metaphysical assumptions — particularly the treatment of psychophysical laws as independent, brute facts whose harmony is a contingent matter requiring explanation. Under constitutive physicalism, the “harmony” is not contingent but necessary (identity is not correlation), and the argument’s probability space is empty. Under type-A physicalism (Bush), the explanandum itself is denied. The Bayesian priors are driven by the combinatorial explosion assumption and the principle of indifference, both of which are question-begging against physicalist views on which the “alternatives” are not metaphysically possible.

The argument is therefore best understood as a conditional argument: if you accept (a) an epistemic gap between physical and phenomenal truths, and (b) that psychophysical laws are contingent brute facts rather than necessities grounded in the nature of the properties involved, then psychophysical harmony is significant evidence for theism. But both premises are precisely what physicalists deny, and the argument does not provide independent reasons to accept them. ^{1 3 2}

Relationship to This Wiki

• chinese-room-argument — Searle’s argument for the insufficiency of syntax for semantics; the PPH argument builds on a similar intuition that physical mechanism alone can’t explain why our reports about consciousness are true
• turing-test — The behavioral criterion for intelligence; PPH challenges the functionalist identification of behavior with understanding by arguing the match between report and reality is “lucky” unless designed
• what-is-intelligence-book — Agüera y Arcas’s functionalism (“a function is what it does”); PPH argues functionalism can’t explain which functional identity obtains
• bayesian-apologetics — PPH is a Bayesian argument in the same family as Swinburne’s, subject to the same subjectivity-of-priors and dwindling-probabilities objections
• ontological-arguments — A different (modal) a priori argument for theism; PPH is a posteriori (empirical) rather than a priori
• agrippas-trilemma — The regress problem in justification; PPH’s Bayesian structure presupposes stopping somewhere (priors), which the trilemma explains is unavoidable
• active-inference — The free-energy principle views organisms as self-organizing systems with intrinsic teleology; this provides a non-theistic framework where mind-body “harmony” is structural rather than lucky
• john-searle — His distinction between original and derived intentionality bears on whether consciousness can be “constituted” by function or requires something more
• emergence — Strong emergence is one naturalistic reply to the harmony puzzle (novel psychophysical powers as emergent), but inherits the brute-contingency worry the argument exploits; organizational emergence leaves the alignment question open

See Also

• chinese-room-argument — Syntax vs. semantics; can symbol manipulation produce understanding?
• emergence — Weak vs. strong emergence; whether consciousness needs novel causal powers
• turing-test — Behavioral criteria for intelligence
• what-is-intelligence-book — Functionalism extended to biology and AI
• bayesian-apologetics — Bayesian arguments for theism
• subjectivity-of-priors — The prior probability problem
• dwindling-probabilities — Chained hypotheses in probabilistic arguments
• ontological-arguments — A priori arguments for God’s existence
• agrippas-trilemma — The justification regress
• active-inference — Self-organizing systems with intrinsic teleology
• john-searle — Original vs. derived intentionality
• fine-tuning-argument — The older design argument from physics; its multiverse debate is the comparison point for PPH’s “multiverse immunity” claim

References

• Cutter, B. & Crummett, D. (forthcoming). “Psychophysical Harmony: A New Argument for Theism.” Oxford Studies in Philosophy of Religion. PhilArchive. ¹
• Bush, L. S. (2023). “There is no such thing as psychophysical harmony (Part 1).” Lance Independent. ³
• Pautz, A. (2020). “Consciousness and Coincidence: Comments on Chalmers.” Journal of Consciousness Studies, 5–6, 143–155.
• Goff, P. (2018). “Conscious Thought and the Cognitive Fine-Tuning Problem.” Philosophical Quarterly, 68(270), 98–122.
• Chalmers, D. (2018). “The Meta-Problem of Consciousness.” Journal of Consciousness Studies, 25(9–10), 6–61.
• Williamson, J. (2022). “One philosopher’s modus ponens is another’s modus tollens: Pantomemes and nisowir.” Metaphilosophy, 53(2-3), 284–304. ²
• NotThatGuy (2024). “Does psychophysical harmony strongly point toward theism?” Philosophy Stack Exchange, answer 1. ⁴
• variableization (2024). “Does psychophysical harmony strongly point toward theism?” Philosophy Stack Exchange, answer 7. ⁴
• MedlinePlus Genetics. “Channelopathy-associated congenital insensitivity to pain”.
• van Leeuwen, T. M., Singer, W., & Nikolić, D. (2015). “The Merit of Synesthesia for Consciousness Research.” Frontiers in Psychology, 6, 1850.

Footnotes

1. raw/articles/cutter-crummett-psychophysical-harmony.md ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵

2. raw/articles/google-ai-mode-pph-critique.md ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

3. raw/articles/bush-2023-type-a-physicalism-pph.md ↩ ↩² ↩³ ↩⁴ ↩⁵

4. raw/articles/stackexchange-pph-answers-2024.md ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷