When Maxwell unified electricity and magnetism into four elegant equations, or when Einstein derived the field equations of general relativity, physicists reached for a peculiar locution: they had discovered laws. The metaphor is juridical, even theological—laws command, and matter obeys. But what exactly does this obedience amount to? Do Maxwell's equations somehow make electromagnetic waves propagate at c, or do they simply record, with mathematical precision, that they do?

This question, seemingly semantic, cuts to the metaphysical core of physical theory. If laws merely describe the mosaic of actual events, then modal claims—about what must happen, what would happen under different initial conditions—require deflation or reinterpretation. If laws genuinely govern, we owe an account of how abstract mathematical structures exert metaphysical authority over concrete physical happenings.

The stakes extend beyond metaphysics into scientific practice itself. How we understand laws shapes our theories of explanation, our confidence in counterfactual reasoning, and our interpretation of what physics ultimately achieves. The Humean and non-Humean camps offer radically different pictures of what a completed physics would tell us about the world.

David Lewis, refining a proposal from Ramsey and Mill, articulated what has become the canonical Humean position: laws of nature are the axioms of that deductive system which best balances simplicity and informativeness with respect to the total mosaic of particular matters of fact. On this Best Systems Account (BSA), there is nothing to the world beyond the distribution of local, categorical properties across spacetime points. Laws supervene on this Humean mosaic as its most efficient compressed description.

The view has considerable virtues. It respects the epistemic situation of actual scientists, who arrive at laws by systematizing observations rather than by intuiting metaphysical necessities. It dissolves the mystery of why laws are knowable: they are, by construction, features of the pattern we observe. And it aligns naturally with the practice of theoretical unification, where a good theory is precisely one that captures much with little.

Yet the account faces persistent difficulties. The notions of simplicity and informativeness appear language-relative—what counts as simple depends on which predicates we take as primitive. Lewis appealed to a distinguished class of perfectly natural properties to fix the vocabulary, but this move imports substantial metaphysical machinery that sits uneasily with Humean parsimony.

More troubling is the objection from explanation. If the pattern of particulars grounds the laws, then citing the laws to explain those particulars appears viciously circular. The electron's behavior cannot both constitute Coulomb's law and be explained by it. Humeans respond by distinguishing metaphysical grounding from scientific explanation, but critics find this distinction strained.

Nevertheless, the BSA remains attractive to those who see modality as a projection of our epistemic and pragmatic interests onto a fundamentally non-modal world. It offers a deflationary reading of physics as the science of what actually happens, elegantly summarized.

Takeaway

On the Humean view, the universe doesn't follow laws—it exhibits patterns, and laws are simply our most compressed summary of those patterns. Necessity is a feature of our systematizations, not of the world.

Anti-Humeans, from Armstrong and Dretske to Maudlin and Bird, argue that the Humean mosaic is metaphysically insufficient. Something must explain why the pattern is this pattern rather than another—why electrons universally attract protons, why entropy in isolated systems does not decrease. On the governing conception, laws are ontologically prior to and productive of the regularities we observe.

The versions of this view differ instructively. Armstrong construes laws as relations of nomic necessitation holding between universals: it is a second-order fact that F-ness necessitates G-ness, and this fact grounds the first-order regularity. Tim Maudlin, resisting reductive analysis, treats laws as fundamental features of reality—primitive constraints on the evolution of physical states. Dispositionalists like Alexander Bird locate the source of law-like behavior in the essential powers of physical properties themselves: charge just is the disposition to produce certain forces.

Each variant addresses the explanatory deficit Humeans face. On the governing conception, appeal to laws in scientific explanation is metaphysically respectable because laws genuinely produce the phenomena they subsume. Counterfactual reasoning tracks real modal structure in the world rather than pragmatic extensions of actual patterns.

The cost is ontological. We must countenance entities—universals, primitive laws, or dispositional essences—that go beyond the categorical mosaic. Humeans regard this as unwarranted metaphysical inflation, invoking Ockham's razor against posits that do no additional predictive work.

The debate ultimately turns on what we take metaphysical explanation to require. If we demand that patterns be grounded in something beyond themselves, the governing conception is compulsory. If we accept that at some level of description brute facts must be tolerated, the Humean mosaic can be that level.

Takeaway

The governing view treats laws as active constraints written into the fabric of reality, not passive summaries of what happens. The disagreement with Humeans is fundamentally about whether patterns need grounds.

The dispute would be idle if it left scientific practice untouched. It does not. Consider how physicists reason about counterfactuals: what would happen if we doubled the mass of the sun, if we prepared a Bell state with different parameters, if gravitational waves had different polarization modes. Such reasoning is ubiquitous in theoretical work and appears to presuppose robust modal structure in nature.

The governing conception offers a straightforward vindication: counterfactuals track how laws would apply to alternative initial conditions, and laws support these applications by virtue of their genuine necessity. Humeans must work harder, typically by embedding counterfactual assessment within the Best System of the actual world and extending it to nearby possible worlds constructed via similarity metrics. Whether this reconstruction preserves scientific practice or merely mimics it remains contested.

Explanation shows similar strain. Deductive-nomological accounts, and their contemporary descendants in causal and structural equation frameworks, treat laws as the load-bearing elements of scientific explanations. If laws are mere summaries, explanation reduces to subsumption under patterns, which many philosophers of science find explanatorily thin.

Interpretation of specific theories also depends on the debate. In quantum mechanics, the status of the Schrödinger equation—as governing wavefunction evolution or as describing a pattern in the world's history—shapes one's stance on measurement and the wave function's ontological status. In statistical mechanics, whether the Second Law governs or describes affects the probability-mechanics relationship.

Scientists themselves are typically agnostic, but the philosophical choice colors their self-understanding. To do physics as the discovery of governing structure is a different intellectual project than to do it as the compression of the world's contingent history.

Takeaway

How we interpret laws changes not what physics predicts but what physics means. The equations are the same; the metaphysics they encode is not.

The choice between governing and describing conceptions of law is not resolvable by physics itself. Both interpretations are consistent with any body of empirical evidence, since they differ over what grounds the regularities rather than what regularities obtain. This is metaphysics in its proper domain: articulating the deep structural commitments implicit in our best science.

Yet the debate is not merely verbal. Different accounts license different explanatory ambitions, different attitudes toward counterfactuals, and different readings of what a fundamental theory achieves. A naturalistic philosophy of science must engage this question because the sciences themselves cannot avoid taking a stance, even implicitly.

Perhaps the most productive response is to recognize that both conceptions capture something the other neglects. The Humeans are right that laws must be inferable from patterns; the anti-Humeans are right that patterns invite metaphysical explanation. A mature philosophy of physical law lives in the tension between them.