The Unity of Philosophy

Scenario

A school board holds a hearing on a proposed vaccine mandate for enrollment. A pediatrician stands and reads from a set of clinical trial numbers: efficacy, adverse-event rates, the size of the sample. She sits down having, in her own mind, settled the question. A parent stands next, not disputing a single number, and says that no rate of adverse events, however low, gives the state the right to put a needle in her child over her objection. The two have not contradicted each other. The pediatrician was answering a question about evidence; the parent was answering a question about whose body it is. The board, and much of the room, treats this as one disagreement getting louder, when it is two different questions being asked past each other, neither speaker noticing the other has changed the subject.

A third speaker then asks a question neither of the first two addressed: even granting the data and granting an obligation between neighbors, does this school board have the standing to compel enrollment on the point, as opposed to requiring a form, offering an exemption, or leaving the matter to a higher authority. The hearing runs another hour without anyone naming the shift.

Definition

Separate what happened at the hearing into its parts before giving the parts names. The pediatrician's claim can be true or false independent of what anyone owes anyone: the trial either shows what she says it shows or it does not, and a second study could overturn her without anyone's obligations changing. The parent's claim survives even if the data is granted in full: her point is not that the numbers are wrong but that a low risk, honestly stated, still does not by itself authorize a bodily intrusion. And the third speaker's question survives even if the first two are settled: granting good evidence and granting a real obligation between neighbors, it remains a separate question whether this body, the school board, may use force, in the form of exclusion from school, to make the reluctant comply, or whether that power belongs elsewhere or nowhere.

Three questions, then, not one, and each survives the other two being answered:

An epistemic question: what does the evidence show, and who is positioned to evaluate it. This is settled by trial design, replication, and expertise, and it can shift on its own, independent of anyone's duties.

An ethical question: given what the evidence shows, what do people owe each other, parent to child, household to neighborhood. This is not read off the evidence; it is argued for, using the evidence as an input.

A political question: given an obligation, if one is established, what may be compelled, by which body, using what means, against a dissenter. This is not read off the obligation; it is argued for separately, using the obligation as an input and asking a further question about legitimate power.

Call these the epistemic, ethical, and political layers. Any public dispute of this shape, not just the hearing, runs on all three at once, and pressure travels in both directions between them: a state that enforces a policy shapes which questions get funded and researched, so the political layer feeds back into the epistemic one over time, not only downward.

Philosophy's subfields divide along this same seam. Epistemology studies the first layer in general, what knowledge is and how it is secured. Ethics studies the second, what is owed. Political philosophy studies the third, what may be enforced and by whom. A dispute like the hearing is not a special hybrid case; it is the ordinary case, and the subfields are what you get when each layer is studied on its own, at length, apart from a specific dispute that forced all three into the open together.

The two seams between the layers are not the same kind of seam. The first seam, epistemic to ethical, is the philosopher David Hume's is/ought gap: a description of what the trial shows (an is) does not by itself entail any statement of what is owed (an ought); the move from evidence to obligation is a genuine inferential gap that has to be argued across, not assumed. This is the ONLY seam the is/ought gap covers. The second seam, ethical to political, is a different kind of question entirely: given an obligation (an ought), whether anyone may be compelled to honor it is not a further is/ought crossing, since both the obligation and the question of permissible force are normative claims already. It is instead a public-justification question, of the kind political philosophy asks under headings like the harm principle: which of the things people owe each other may be extracted by force, and which must be left to persuasion, custom, or conscience. Treating this second seam as "is/ought applied again" collapses two different arguments into one and hides that the second needs its own, separate defense in terms of coercion and legitimacy, not in terms of fact and value.

Limits and boundary conditions

The unity this entry demonstrates is narrower than the title might suggest, and the boundary needs to be stated plainly. What has been shown is that one slice of disputes, public normative disputes that ask what should be believed, what is owed, and what may be enforced, runs on a stacked epistemic-ethical-political structure. This covers a real and recurring class of arguments, but not the whole of philosophy. Metaphysics asks what exists and what a thing's nature is, a question that can be posed with no dispute about obligation or enforcement anywhere near it. Logic asks what follows validly from what, a question that holds for uninhabited possible worlds with no one to owe anyone anything. Aesthetics asks what makes a work good as a work, and while aesthetic judgments can enter public disputes, they do not organize themselves along an is-to-ought-to-may-compel gradient the way the hearing does. None of these three fields sit inside the three-layer stack, and importing the stack into them would misdescribe what they are doing. The claim here is deliberately local: within public normative dispute, three layers; not: within philosophy, three layers.

Common mistakes

Fixing a dispute at one layer and declaring the other two settled or irrelevant is the most common failure, and it comes in a matched pair. The technocrat's version fixes the debate at the epistemic layer: once the trial data is in, the technocrat treats the ethical and political questions as already answered, as though a low adverse-event rate settled what may be done to a dissenter. The libertarian's mirror version fixes the debate at the political layer: any compulsion is ruled out on principle before the epistemic or ethical layers are even examined, as though the evidence were beside the point. Both errors look like confidence; both are actually a refusal to argue one of the layers, dressed as having already won it.

A second mistake is assuming that settling one layer settles the next for free. Strong evidence does not, by itself, yield an obligation; an obligation, once established, does not, by itself, license enforcement. Each step needs its own argument, and the two steps differ in kind, as the Definition section sets out: the first crosses Hume's is/ought gap, the second is a distinct question about permissible coercion. Reaching the ethical layer is not one short step from the political layer; it is the start of a different question.

A third mistake is treating the three layers as a one-way pipeline, evidence feeding obligation feeding policy, and missing the feedback running the other way: what a political body chooses to enforce shapes which studies get funded and which questions get asked, feeding back into what later counts as established at the epistemic layer. Ignoring that return path makes "the evidence" look more independent of politics than it is.

Test yourself

Pick a live public dispute you actually care about and can state in one sentence: housing supply, an AI-safety regulation, a drug-policy change, a school curriculum fight, anything currently being argued in public rather than settled. Write its central question three separate times: once as a pure epistemic question (what would the evidence need to show, and who is positioned to judge it), once as a pure ethical question (granting some state of the evidence, what would be owed), and once as a pure political question (granting some obligation, what, if anything, may be enforced, by whom, and against whom). Then find one real example of the dispute in public, an op-ed, a hearing transcript, a comment thread, and mark which layer its loudest argument is actually operating on.

Success is three genuinely distinct, well-formed questions that are recognizably about the same underlying dispute, not three restatements of one question in different words, plus a one-sentence diagnosis of which layer the loudest public argument occupies and a separate sentence on which layer you think the real disagreement lives in, when the two differ.