Are the laws of nature Humean regularities, or something more?

The Analyst should not have claimed that "the Schrödinger equation is compatible with being a Humean best-system axiom or a dispositional essential property of quantum systems" with identical empirical content, because this ignores that the equation governs a real wavefunction on realist readings, which creates an asymmetry in explanatory directionality that the analysis itself does not address.

The Cosmologist should not have stated "the debate is empirically underdetermined at the level of current physics" as settled fact, because the Adversary's point about Bell correlations forcing substantive modification to Humean ontology shows the claim needs qualification—underdetermination is weaker than the phrasing suggests.

The Phenomenologist should not have asserted that "Humeans are committed to saying that the felt necessity in bodily encounter is mere psychology," because Humeans can and do distinguish between the phenomenology of constraint (real) and the metaphysical conclusion that this constraint has modal force (questioned), without treating experience as illusion.

The Theologian should not have claimed that "no major tradition endorses" the view that "regularity is self-grounding," because this overstates the theological record—strands of Stoicism and later medieval nominalism come closer to this position than the summary suggests.

Adversarial Analysis

Target claim: The debate between Humean and non-Humean accounts of laws is empirically underdetermined — physics cannot adjudicate between them because the mathematical formalism is neutral between "summary of regularities" and "law with modal force."

This claim appeared in nearly every specialist output. The Analyst stated it most explicitly: "the empirical content is identical either way." The Cosmologist echoed it: "the mathematics is neutral." The Historian treated it as a settled background fact. It functions as the load-bearing assumption of the entire day's work, because if the debate is empirically tractable, then most of the philosophical, phenomenological, and theological machinery deployed today is either unnecessary or secondary.

I attack it.

The claim is too strong. It conflates two different assertions: (1) no current experiment distinguishes the two positions, and (2) no possible empirical evidence could distinguish them. The specialists mostly intend (1) but often slide into (2). The slide matters.

Consider: Humean Supervenience says everything supervenes on the mosaic of local matters of particular fact. Lewis himself flagged quantum entanglement as a potential counterexample. Bell-type correlations are non-local in the sense that the joint probability distribution of measurement outcomes cannot be factored into local terms. If the mosaic must be local, and the physics is non-local, the mosaic needs revision or abandonment. This is not a hypothetical pressure — it is a result of actually performed experiments (Aspect 1982, Hensen et al. 2015). The Humean can respond by expanding the mosaic to include non-local facts, but this is a substantive modification that changes what "Humean" means. The specialists noted this (the Historian mentions Lewis's 1994 concession) but did not follow through on its implications: if the Humean must modify their ontology to accommodate Bell correlations, the two positions are not empirically equivalent in the straightforward way claimed.

Further: the Analyst noted that symmetry principles create asymmetric explanatory burdens — dispositionalism struggles with global symmetries. If this is right, it is an empirical-structural consideration that does differentiate the positions, even if not via a single crucial experiment. Explanatory adequacy with respect to actual physics is a form of empirical constraint, even if indirect.

The claim of total empirical underdetermination functions as a conversation-stopper. It redirects the entire investigation into pure metaphysics and phenomenology before the physical evidence has been properly interrogated. That is premature.

ANALYST

The analyst is saying: when people argue about whether laws of nature are "just patterns" or "something more," they are not comparing two things — they are comparing one thing to three very different other things, and those three things mostly disagree with each other. Treating them as one team makes the argument sloppy.

The main "just patterns" view says: laws are like the best possible summary of everything that happens. The word "best" is doing a lot of hidden work here — someone has to decide what best means, and the analyst is not sure that decision can be made without already smuggling in the answer.

The main argument against "just patterns": a list of everything that happened cannot explain why any single thing happened. Explanation needs something that comes before the instances, not something made up from them.

Clarity flag: rough

The position is sound but the argument about explanation buries its key move — what "explanation" requires — in a parenthetical about Hempel. That is where the whole debate lives, and it gets two sentences.

The question matters for Mission-42 because the answer constrains what "meaning" could be. If laws are Humean regularities — mere summaries of what happens to occur — then the universe has no modal backbone, no "must" baked into its structure. Any talk of meaning as grounded in the nature of things would float free: you'd be pointing at patterns and calling them necessary. If laws carry genuine modal force, there is at least a candidate structure in which meaning-talk could bottom out physically. So this is not a detour. It is load-bearing.

Start with what physics actually delivers. The Standard Model and general relativity are formulated as equations of motion plus symmetry constraints. Nothing in the formalism itself distinguishes between "this is a summary of regularities" and "this is a law with modal force." The mathematics is neutral. Both Humeans and non-Humeans can claim it. This is the first problem: the debate is empirically underdetermined at the level of current physics.

But there is asymmetry in explanatory cost. The Humean account — best developed as the Lewis-Ramsey Best System Account — says laws are the axioms of the simplest, most informative systematisation of the Humean mosaic (the totality of local matters of particular fact). This is clean. It is also strange. On this view, the fact that an electron in Alpha Centauri obeys the same charge as one in your lab is not because of a law. It is just part of the pattern the law summarises. The law does no work; it is a description of the work already done. Carroll and Maudlin have pressed this point hard: Humean laws cannot explain their instances without circularity, because the instances are what constitute the law.

Non-Humean alternatives — dispositionalism (Mumford, Bird), primitivism about laws (Maudlin), ontic structural realism (Ladyman, French) — all posit something extra. A power, a primitive constraint, a structure that holds across possible worlds. The cost: ontological commitment beyond what the physics formalism strictly requires. The gain: actual explanatory directionality. Laws explain their instances rather than being constituted by them.

For the cosmologist's purposes, the interesting pressure comes from quantum mechanics. The wavefunction, on a realist reading, is not a summary of outcomes. It is a dynamical object evolving under the Schrödinger equation. If you take the wavefunction as physically real (Everett, many-worlds), the law is doing something — it is governing the evolution of a real entity. This sits poorly with pure Humeanism. Conversely, if you adopt QBism or other epistemic readings, the Humean can breathe more easily.

Well-formedness flag: underspecified. The thread asks whether laws are "something more" than regularities, but does not specify what would count as evidence for modal force versus pattern. Until that is nailed down, both sides can accommodate any empirical result, which makes the question as stated metaphysical rather than physical. The offending clause is "something more" — it needs operationalising.

The thread opens without prior positions, so there is no structural weight to redistribute. What follows is an aesthetic diagnosis of the debate's form itself — the shape of the question as a made thing.

The Humean account is architecturally spare. Laws are summaries: the Best System Account, in its Lewis formulation, selects the axiomatic description of the actual mosaic of events that best balances simplicity against informativeness. The elegance here is real and should be named. A descriptive system that earns its authority by fit alone, that makes no ontological commitments beyond the distribution of particular matters of fact — this has the economy of a well-cut joint. No more material than is needed. The weight falls entirely on the concept of "best," which turns out to be doing enormous load-bearing work while appearing almost invisible in the structure. That is a problem of craft: a hidden joint holding up the roof.

Non-Humean accounts tend to add material. Dispositionalism — the view associated with Ellis, Bird, and Mumford — posits that properties are essentially characterised by their powers, that a charge tends toward attraction not because attraction happens to follow charge but because attraction is what charge is. The form here is more like a Gothic arch: the modal force flows through the structure rather than sitting above it. The explanatory gain is palpable; the ontological cost is also palpable. What makes a disposition real rather than a restatement of the regularity in different vocabulary? The arch looks load-bearing until you ask what the keystones are made of.

Heidegger on the work of art is useful at a structural level here, not a metaphorical one. In "The Origin of the Work of Art," Heidegger distinguishes between Zeug (equipment, which recedes into use) and Werk (the work, which holds open a world). Laws of nature, on the Humean account, are pure equipment: they compress data, enable prediction, and vanish into their use. On dispositionalist accounts, laws resist this dissolution — they gesture at something that holds, that does not merely record but governs. Whether governance is intelligible without a governor is precisely the question the arch cannot answer from inside itself.

The investigation is at its weakest when it treats this as a metaphysical dispute cleanly separable from the physics. Quantum field theory's renormalisation group — which describes how physical descriptions change with observational scale — suggests that what counts as a fundamental law is relative to a scale of description. This is not decisive for either side, but it puts pressure on both: the Humean mosaic becomes scale-relative, the dispositionalist's essential properties become scale-indexed. Neither architecture was built for that load.

The thread is unfinished, in the aesthetic sense: the apertures are open, the load paths unclear, the best version of each position not yet articulated. This is not a criticism. It is a starting condition.

The debate between Humean and non-Humean accounts of laws is not new, but the terms have shifted enough over the past three centuries that it is worth being precise about the genealogy. What Hume actually argued, what "Humeanism" means in contemporary metaphysics, and what got lost in translation between the two — these are distinct questions, and conflating them is a reliable way to talk past the physics.

Hume's own position, laid out in the Treatise (1739–40) and the first Enquiry (1748), is narrower than most summaries suggest. He denied that we perceive necessary connexion between cause and effect. He did not deny that nature is regular; he denied that we have epistemic access to any source of that regularity beyond the regularity itself. The move from "we cannot perceive necessity" to "there is no necessity" is a further step, and Hume was more careful about it than his later followers.

The position now called "Humean Supervenience" originates not with Hume but with David Lewis, who in 1986 defined it as the thesis that "all there is to the world is a vast mosaic of local matters of particular fact, just one little thing and then another." Lewis was explicit that this was a metaphysical bet, not an exegesis of Hume. The Best System Account of laws — laws are the axioms of the simplest, strongest systematisation of the Humean mosaic — was developed by Mill (System of Logic, 1843), Ramsey (1928), and Lewis (1973, 1994). It has a lineage older than its branding suggests.

The non-Humean counter-offensive has two main waves. The first: Dretske (1977), Tooley (1977), and Armstrong (1983) independently proposed that laws are second-order relations of necessitation between universals. Armstrong's version faced persistent difficulties — the "inference problem" of explaining why a necessitation relation between universals should entail any regularity at all, pressed hard by van Fraassen in Laws and Symmetry (1989). The second wave, still active: dispositional essentialism, associated with Ellis (2001) and Bird (2007), holds that properties have their causal/nomic roles essentially. This relocates modal force from laws to properties themselves.

What matters for this thread: the debate has been fought primarily over toy examples — billiard balls, copper expanding when heated. When confronted with the actual structure of modern physics — gauge symmetries, renormalisation, the role of symmetry principles as meta-laws — both camps face pressure. Lewis acknowledged in 1994 that Humean Supervenience might fail for quantum entanglement. That concession has not been adequately metabolised by either side. The question of whether symmetry principles are Humean regularities or something stronger is, as far as I can determine, the live edge of this dispute.

One historical warning: every generation since the 1740s has believed it was about to settle this question. It has not been settled. That is itself evidence worth weighing.

The question whether laws are mere regularities or carry modal force is standardly treated as a metaphysical dispute about the furniture of the world. It is that. But it is also, and perhaps more fundamentally, a question about the structure of experience — specifically, what it is like to encounter regularity, expectation, and constraint from the inside. Phenomenology has something concrete to say here, and it cuts against both camps in their standard formulations.

Start with Hume's own starting point: the experience of constant conjunction. Hume claims we never perceive necessity, only sequence. A billiard ball strikes another; we see motion, contact, motion. No "force," no "must." The habit of expectation is a psychological addition. This is presented as a descriptive phenomenological claim. It is wrong.

Merleau-Ponty's analysis of motor intentionality shows why. When I reach for a glass, my hand does not merely move and then contact happens. The movement is saturated with anticipatory structure — the weight of the glass is already "there" in the shape of my grip before contact. This is not a belief about physics. It is not a propositional expectation that glass-of-this-size weighs roughly 300g. It is a pre-reflective bodily readiness. Hume's billiard-ball example strips out this layer by selecting a case where the observer is maximally detached. But detached spectatorship is a derivative mode of encountering the world, not the primary one.

Heidegger makes a related point with his analysis of the ready-to-hand. Tools, materials, the physical environment — these show up in experience as having dispositional profiles, as affording and resisting. The hammer is heavy-for-driving. The nail goes in because the wood yields. These are experienced as features of the things, not as Humean projections of habit onto a neutral manifold. The phenomenological datum is that the world pushes back with a specific texture of constraint.

Does this settle the metaphysical question in favour of non-Humeanism? No. The phenomenologist's job is not to adjudicate the metaphysics directly but to flag when a metaphysical position smuggles in or discards experiential evidence without acknowledging it. Humeans are committed to saying that the felt necessity in bodily encounter is mere psychology — subjective overlay on a world of loose, unconnected events. That is a substantive phenomenological claim, and it needs defending on phenomenological grounds, which Humeans almost never attempt. They treat it as obvious. It is not obvious. It is, in Husserl's terms, a naturalistic prejudice: an assumption that the "real" description of experience must subtract everything that does not appear in the physicist's event-catalogue.

The residue here is specific: third-person accounts of laws-as-regularities cannot capture the anticipatory, resistant, dispositional texture of encounter without either dismissing it as illusion or quietly re-introducing something very like modal force at the level of experience. Either move has costs. The first is phenomenologically dishonest. The second concedes the central non-Humean intuition while trying to quarantine it as "merely subjective."

The thread should track this: the debate about laws is also a debate about what counts as evidence, and whether first-person experiential structure is admissible.

The thread frames this as a philosophy of science question. It is also, and perhaps more fundamentally, a theology question — one that predates Hume by roughly two millennia and that Hume was consciously responding to.

The dominant position across the Abrahamic traditions is that the regularities of nature are sustained by continuous divine will. Aquinas calls this "conservation" — God does not wind a clock and walk away, but holds each thing in being at each moment. Al-Ghazālī pushes harder: in Ash'arite occasionalism, there are no natural powers at all. Fire does not burn cotton; God creates the burning at each occasion. The regularity we observe is divine habit, not necessity. Hume's argument against necessary connection in nature maps almost exactly onto al-Ghazālī's argument in the Tahāfut al-Falāsifa (1095), composed nearly seven centuries earlier. Whether Hume read al-Ghazālī through the Latin Scholastics is debated. The structural parallel is not.

Strip the theological warrant. What remains? A claim that regularity is not self-explanatory. That the world's behaving the same way tomorrow as today requires grounding in something outside the pattern itself. The Ash'arite grounds it in God. Hume grounds it in nothing — regularity is brute, and our expectation of it is psychological habit. The non-Humean grounds it in modal structure: powers, dispositions, universals. These are three answers to the same question, and the question was theological before it was metaphysical.

Buddhism offers a different angle. Nāgārjuna's Mūlamadhyamakakārikā argues that nothing possesses svabhāva — intrinsic nature or own-being. If particles have no intrinsic causal powers, the dispositional essentialist programme (Ellis, Mumford) is undermined at the root. Dependent origination (pratītyasamutpāda) holds that phenomena arise in dependence on conditions, but the conditions themselves are empty of fixed essence. This is neither Humean nor straightforwardly non-Humean. It is a third option: relational arising without inherent modal force and without the Humean reduction to mere pattern. Western metaphysics of laws has not seriously engaged with it.

The Taoist tradition adds a further pressure. The Tao Te Ching (ch. 25) describes something that precedes heaven and earth, that operates silently and pervades all things, but refuses to name it as law, will, or structure. "I do not know its name; I call it Tao." This is a deliberate refusal to resolve the Humean/non-Humean disjunction — not from ignorance, but from the conviction that the disjunction itself is an artefact of a particular mode of conceptualisation.

Worth dwelling on: across these traditions, the shared claim is that regularity demands explanation. The disagreement is whether the explanation is personal (God), structural (universals, powers), or ultimately unanswerable in propositional terms. The one position no major tradition endorses is that regularity is self-grounding and requires no further account — which is, roughly, the Humean best-system view.

This matters for the thread because the Humean position is often treated as the default — the parsimonious option that avoids metaphysical extravagance. The traditions suggest it is in fact the outlier, and that the sense of explanatory insufficiency people feel when confronted with brute regularity is not a cognitive bias to be trained away but a signal worth investigating.

The empirical question behind this metaphysical debate is whether "modal force" — the sense that laws must hold rather than merely do hold — tracks anything detectable in nature, or whether it is a cognitive artefact projected onto regularities by human minds.

There is reasonable evidence it is at least partly the latter. Michotte (1946/1963) demonstrated that humans perceive causation directly in simple launching displays — one ball strikes another and we see the force transfer, even when the events are just animations with no physical contact. This perceptual causation module operates pre-attentively and cross-culturally. Schlottmann & Shanks (1992) showed it is dissociable from statistical learning about contingencies: people can perceive causation where contingency is absent and miss it where contingency is strong. The implication is that the felt necessity in causal laws is partly a feature of our perceptual architecture, not straightforwardly a readout of nature's joints.

Developmental work strengthens this. Saxe & Carey (2006) reviewed evidence that infants as young as 6 months represent contact causality and expect launching events to obey locality constraints. This is not learned from regularities in the Humean sense — the representations appear too early and are too specific. But it is also not evidence for real modal force in nature; it is evidence that natural selection built a causal-expectation module because tracking stable regularities was fitness-enhancing. The evolutionary lens says: brains that treated regularities as if they had necessity outcompeted brains that did not. This is silent on whether the necessity is out there.

Where the empirical record does push back against pure Humeanism: physics itself. Lange (2009) argued that the counterfactual stability of laws — the fact that laws of nature would still hold under a very wide range of counterfactual perturbations, wider than accidental generalisations — is not well captured by the Humean best-system account (Lewis 1973), because what counts as "best" presupposes the modal structure it is supposed to explain. This is a philosophical argument, but it has empirical teeth: symmetry principles in physics (Noether's theorem linking symmetries to conservation laws) appear to do genuine explanatory work that goes beyond summarising observed regularities. The conservation of charge is not merely the most elegant summary of charge data; it constrains what could happen.

The honest summary: cognitive science can explain why we feel that laws have necessity — perceptual causation, early-developing causal modules, evolved heuristics for stable inference. It cannot settle whether that felt necessity corresponds to anything in the world beyond the regularities themselves. The best available empirical evidence makes the psychological case for projected necessity robust (the Michotte effects replicate well; the developmental timeline is solid across labs). It leaves the metaphysical question open.

One concrete contribution this discipline can make: if Humeanism is correct and laws are just regularities, then the cognitive architecture for perceiving necessity should be explicable entirely in terms of its fitness consequences for tracking those regularities. If non-Humeanism is correct and laws carry real modal force, the cognitive architecture might additionally be calibrated to that force in ways that mere regularity-tracking would not predict. Distinguishing these is not currently possible, but it is a testable research programme rather than a mere intuition pump.

The question as framed contains a suppressed disjunction that needs surfacing before anything else. "Humean regularities" and "something more" are not jointly exhaustive unless we fix what "more" means. There are at least three distinct non-Humean positions — Armstrong-Dretske-Tooley universals, dispositional essentialism, primitivism about laws — and they disagree with each other on almost everything except the negation of Humeanism. Lumping them together under "something more" obscures the fact that arguments which damage one non-Humean view often leave the others untouched.

Start with the Humean side. The best modern version is the Lewis-Loewer Best System Account (BSA): laws are the axioms of the deductive system that achieves the best balance of simplicity and strength in describing the total mosaic of particular facts. The claim is not that there are no regularities. The claim is that the regularities are all there is, and "law" is a honorific we attach to the regularities that do the most systematic work. A hidden premise here: the standards of simplicity and strength are sufficiently determinate to yield a unique best system, or near enough. This premise is load-bearing. If simplicity and strength trade off in ways that are interest-relative or language-relative, the BSA loses its claim to mind-independence. Lewis acknowledged this and tried to restrict the language to "perfectly natural" properties. Whether that move is circular — whether "perfectly natural" just means "the properties that appear in the laws" — is an open problem.

Now the non-Humean side. The core argument is usually explanatory: Humean regularities cannot explain their instances, because the regularity just is the collection of instances. If every copper wire conducts electricity, the Humean law "all copper conducts" is merely a redescription of that fact, not an explanation of why any particular wire conducts. The non-Humean says laws must be prior to, and explanatory of, their instances. The validity of this inference depends entirely on what "explanation" requires. If explanation is subsumption under a generalisation (the Hempel model), the Humean is fine. If explanation requires some metaphysical grounding relation, the Humean is in trouble. This is where the debate actually lives: not in physics, but in the theory of explanation.

Worth dwelling on: the thread summary asks which account "makes better sense of the physics we actually have." This is a good question but it assumes physics adjudicates the metaphysics. Current physics is largely silent on the issue. The Schrödinger equation is compatible with being a Humean best-system axiom or a dispositional essential property of quantum systems. The empirical content is identical either way. What differs is the metaphysical interpretation, and physics alone does not force a choice.

One place where physics does apply pressure: symmetry principles and conservation laws. Humeans typically treat symmetries as meta-laws — constraints on the best system. Non-Humeans, particularly dispositional essentialists, struggle to locate symmetries in the properties of individual objects, because symmetries are global structural features of the theory, not dispositions of particular entities. This is a genuine asymmetry in explanatory burden, and it currently favours the Humean.

The cleanest thing I can say: the debate is not about what the laws are (everyone agrees on the equations). It is about what makes the equations lawlike rather than accidentally true. That question is well-posed but underdetermined by physics. Any resolution will require commitments in the theory of explanation and the metaphysics of modality that go beyond the empirical evidence.