Objective truth, and what force must answer to

Builds on the Assumptions.¹

I want to begin with the sentence everything I write here hangs on: there is a way things are.² Not a way for you and a way for me — a way, full stop, indifferent to preference, party, or mood. The world was some way before anyone was here to have opinions about it, and it will be some way after.

That is where my philosophy starts. What I am after reaches further, and it is practical: a shared framework. By that I mean standards of evidence and public rules that people on different sides accept. Inside it, claims can be checked, mistakes corrected, and coercive rules challenged. You may disagree with my account of why those practices work. You may come away thinking that framework worth building while rejecting the rest of this project. Even then, this essay will have achieved something I want.

1. What I am claiming, and what I am not

The claim is narrow. There is a way things are says nothing about who knows it, how well anyone can know it, or how confident anyone is entitled to be. Those are separate questions, and I give them the less comfortable answers. Plenty of truths are out of reach, our instruments are partial, and our minds are motivated. Honest inquiry moves toward the facts slowly, closing in on them with no promise of arriving, when it moves at all. None of that touches the claim. A hidden fact is still a fact. The number of fish in the Atlantic at this moment is unknown and unknowable in practice; it is also exactly one number.

Notice what the claim rules out. It has to rule something out: a claim that makes no difference anywhere says nothing, and I hold my own claim to that rule first.³ Mine makes no observation more or less likely; its difference is in what we may conclude. It rules out your truth and my truth as descriptions of the world. It permits, cheerfully, your perspective and mine, your values and mine, your experience and mine. Perspectives differ because we stand in different places, not because there are different worlds to stand in. Two witnesses to a car crash disagree, and we can say something precise about that: if two of their claims contradict each other, at most one is right. The crash happened one way.

A second commitment keeps apart two things I lean on throughout. One is the map: our description of what is there, which the world can correct. The other is the path we choose across it, which I call policy: what we decide, rather than what we find.⁴ Keeping them apart rules out a comfortable retreat: “well, that is true for science, but this domain is different.” A domain where nothing is the case is not a deeper domain; it is empty of facts, and should say so. Most real domains are neither wholly one nor the other: part found, part chosen. My own ethics is mixed in exactly this way. Suffering is bad for the one who suffers, and I take that to be found. Much of the rest of morality I take to be chosen. Wherever there is a fact of the matter, disagreement is a problem with a possible resolution. Wherever there is not, we are no longer describing — we are choosing, and I insist on saying so out loud. I keep the map and the path apart everywhere.

2. The referee

Now watch what the sentence does. Two people disagree about something that matters — a dosage, a bridge load, whether the well water is safe. If there is a way things are, their disagreement has a referee. Not a person: a direction. There is something the dispute is about, sitting outside both of them, that their claims can be checked against. They can be wrong. They can find out. The argument can end — not by one side tiring, but by the world answering.

Every practice we trust works this way. Engineering is argument refereed by collapse. Medicine is argument refereed by mortality. Navigation is argument refereed by the coastline. The referee is often slow, often expensive to consult, sometimes cruel — but it is there, and both parties accept the same ways of consulting it. What makes their disagreement a conversation rather than a contest is those shared tests and standards of evidence, not a shared philosophy of what is real. The loser of a refereed argument has not been defeated by the winner. He has been corrected by the world, and the winner is answerable to the same referee on the next question. That is why refereed losing is survivable, even useful. Nobody submits to a person; both submit to the way things are.

It does not always end that way, and I will not pretend otherwise. Some disagreements run all the way down to where argument no longer tells the two sides apart. Each stands on a last premise, and neither has anything left that the other is obliged to accept. (“Last” marks where these two stopped, not where the reasons end.) There the world does not answer, or does not answer yet.

Even then, the opening claim still matters: if the last premises are contradictory claims about the world, at most one can be true. The deadlock reflects ignorance, not two equally good answers.⁵ Where a last premise is an aim rather than a claim about the world, we have crossed from the map to the path, and there may be no fact to find. That much follows from the opening claim alone, and the claim is a mild one as metaphysics goes. The consequence binds everyone who grants the claim, me included. Against someone who refuses it, the claim is itself my last premise, set against theirs. That limits what this argument proves; it does not show that we cannot cooperate on a practical question.

Will the ignorance lift? I expect sustained inquiry tested against reality to improve our grasp of particular questions over time. What supports the expectation is a requirement: a theory that replaces another has to account for the older theory’s well-supported successes, where they were tested. The expectation is a further, weaker commitment, and it does not promise that any particular ignorance will lift.⁶

Someone who does not accept all of my account of what is real may still accept tests that can correct either of us. Sharing workable standards of evidence and public rules does not require us to agree on every question about what is ultimately real.⁷ Two investigators could disagree about whether a reality no one can ever reach is in some definite state, and still agree on how to test the water. That gives them something useful to do together without settling the deeper question first. So we should find out what the other person actually accepts before treating a philosophical disagreement as a refusal to answer to evidence.

3. The tell

Here is a pattern you have seen this year, whatever year you are reading in. A public number comes out — a rate, a count, a cost. One side embraces it without checking. The other side dislikes what it shows.

There are more honest responses to a number you dislike than the usual fight admits. The measurement is wrong, and here is a better one. The measurement is right, and here is why it means something else. And a third, which I want to protect: the measurement is unreliable, here is why, and I do not yet know the true value. A broken thermometer can be thrown out before you own a working one. Who funded a study, how its sample was chosen, who declined to answer, whether the thing counted is the thing claimed — all of that is evidence about reliability. Asking after it is doing the map’s work, not leaving the map. All three responses accept the referee; all three can lose.

The tell is none of these. It is when the story of where a number came from is offered as a complete substitute for looking at the number. “The number is theirs” is the whole reply. No correction the other side could make, and no re-measurement by anyone, would let the result count. If that is all there is, such a reply says nothing about the world. It does not say the world is other than measured, or that the instrument is broken in some way a better instrument could fix. Allegiance to the claimant, or hostility toward them, has taken the place of an assessment of the claim.⁸

Now go back to the public number, suppose the other side’s reply is just that, and run the test on both moves. The embrace fails quietly, and usually innocently. An unchecked number is most often accepted out of ordinary trust in a familiar source, trust nobody has calibrated. But a number nobody checked answers to the source, not the world; if it turns out right, that was luck.

When it fails, the rejection fails loudly, and at a precise point. The point is not “who produced it?”, which is a fair question with a checkable answer. It is what comes next. The first follow-up is “so is the instrument broken, and how?” Having no answer to it proves little on its own. The speaker may lack the expertise to say, or may be fairly discounting a source whose track record has earned a discount. The question that tells is the second: “what would a measurement you would accept look like?” Suppose the answer is nothing, and no measurement from them could ever count. Then the reply was never “the world is otherwise,” or even “this thermometer is broken.” It was “no number from them can bind us.”

In neither move, as described, does the evidence do the deciding. One rests on trust nobody has checked, the other on a refusal nothing could answer. Investigating bias is legitimate. Refusing to count a result solely because it comes from the other side is something else. Whatever checks or corrections are offered, that refusal blocks the shared practice I am asking for: checking claims and correcting mistakes.⁹

4. What force must answer to

Coercion needs a justification, and the facts are one part of it. Whether the accusation against someone is true is one question; whether the rule under which they would be coerced is justified is another.

Take two public rules: a speed limit, and a rule that people with blue eyes must be imprisoned. Each could have a perfectly checkable trigger, the fact that sets it off: a driver’s speed can be measured, and so can the colour of someone’s eyes. Applying the blue-eyes rule with perfect accuracy would not justify it. What a rule permits has to be defended too, and that defence belongs to policy; it cannot be read straight off a measurement.

Here is the policy I defend. Inside a shared framework, a rule that coerces people must be stated publicly, and the facts that trigger it must be open to scrutiny.¹⁰

Inside the framework, factual scrutiny constrains how a rule is applied. An accusation must answer to evidence about something neither side controls. The accuser can be wrong, and the process must leave room to challenge the accusation. That takes justified standards of evidence and procedures for challenge, not a perfect instrument. Those standards need a defence of their own: an uncertain danger may justify a precaution, while what evidence would justify punishment is a further question.¹¹

That is the link between truth and force, and its limit. A claim about what happened must not become true merely because the more powerful party asserts it. But neither a correct accusation nor a public rule is enough, on its own, to establish that a particular use of force is justified.

What does defend a rule? My answer is conditional. The norms of a shared life bind whoever wants in, and are silent toward those who don’t. Inside it, a coercive rule is defended when the people who want in could accept it as a term of living together, whatever else they believe about the world.¹² This is not a new idea. John Rawls argued that people holding incompatible worldviews can still share the terms of a political life, and the cooperation I hope for is close to his in aim.

“Could accept” includes grudging acceptance. A rule may cut against someone’s convictions, but not so far that they would rather leave the shared life. They stay because staying is worth more to them by their own weighing. That is the threshold for a rule to count as defended. It is lower and less comfortable than Rawls’s, which asks for reasonable endorsement. I think it is closer to how real people actually manage to live together, and it says plainly where justification runs out.¹³ The test I prefer is stricter, and I hold it the better standard to write rules by: if this rule were in place, would I agree to be its subject? It is the old role-reversal form of the Golden Rule. Both are practical standards for real people living together, not a claim to have found the morally right rule for perfectly informed, idealised agents.

The threshold only works for people who can choose. If you can neither reject a rule nor leave, the rule is not a policy you are party to; for you it is part of the terrain, like the weather. It is still a policy for whoever imposes it, and they owe an answer. A rule imposed on people who cannot realistically leave has to pass the stricter test: would I, the one imposing it, agree to be its subject? And it stays bounded by the limits that never depend on anyone’s agreement: established suffering always counts, and the institutional ban on torture stays absolute.¹⁴ Taking someone’s options away when you could leave them open is itself a wrong. That holds even if the person does not know what they are missing and is content: seen with the full picture in view, they are deprived of ways they could have grown.

Now run the two rules from the start of this section through both tests. The speed limit: the driver who likes speed accepts it grudgingly rather than emigrate, so it passes the threshold. If I picture a child near the road, I would agree to be fined under it, so it passes the stricter test too. The blue-eyes rule: a blue-eyed person who could leave would rather go than live under it, so it fails the threshold. For one who cannot leave, the threshold is silent, and the rule meets the stricter test instead. It fails, since nobody would agree to be imprisoned for the colour of their eyes. A checkable trigger could never have told the two rules apart.

Outside a shared framework, the case is harder. In the real world each side faces pressure of time and resources, and negotiating a shared framework is often infeasible and sometimes impossible. There, force is the ultimate conflict resolver. It resolves; it does not justify, and winning confers no legitimacy. The same two limits still hold: the other side’s suffering still counts, and the institutional ban on torture I hold remains absolute. Neither depends on the other side agreeing to my framework.¹⁵

Better information can remove factual disagreements and show where a peaceful agreement is possible. But even perfect information does not guarantee that two sides want compatible things, or a settlement both can accept.¹⁶ Where their aims remain incompatible, this page does not make them compatible. Nor does it supply a shared tribunal where none exists.

5. What I hope we can share

My own commitment remains that there is a way things are, independent of what anyone prefers. What I hope to share with you is a practical commitment: seek workable common standards, let evidence correct either side, and keep public rules and their application open to challenge. You do not have to accept this whole project to join that effort. Agreement does not make every rule just; the rules still need a defence. Uncertainty does not disappear because we cooperate. The arguments that follow remain open to your criticism.

The next essay goes back to the public number, the kind that describes people: what a number about a group can tell us about the person in front of you, and why evidence, treatment, and publication require distinct judgments.

Next: Statistics — the most important tool, and the most abused.

Footnotes

1. This essay builds on A0.1 (throughout), A0.2 (section 1), A0.D2 (sections 1 and 2), A0.7 (section 2), A0.P4 (sections 2 and 4), and A0.P1 and A0.P2 (section 4). ↩

2. A0.1: “There is a way things are, independent of what anyone believes about it.” It commits me to “rejecting a truth for you and a truth for me.” ↩

3. A0.2: “A claim has content by making a specified difference in a declared kind of landscape.” And: “It may change which observations are likely, which models remain possible, which inferences are licensed, or which actions a policy permits.” And: “In the empirical case, content is how far a claim shifts the distribution over possible observations given its background; a claim leaving every observation exactly as likely has none empirically.” Applied to the opening sentence, precisely: this claim makes no observation more or less likely than it was; its difference is in what may be inferred. ↩

4. A0.D2: “Map and policy are different registers and never trade places.” And: “The map is refereed by reality; the policy is refereed by nothing but whether it works, and may contain stances held for their effect — provided they are labelled as such.” And: “My image for it: the mountains are there; what paths people can take is arbitrary.” ↩

5. Precisely: where those last premises are claims about the world, the world is some way on the point the two of them stopped short of: of two premises that contradict each other, at most one is true, and the deadlock is ignorance, not two equally good answers. ↩

6. A0.7: “I expect sustained inquiry tested against reality to improve our grasp of mind-independent structure and constraints.” And: “A replacement theory must account for the well-supported successes of its predecessors within their tested domains, while remaining open to correcting them. This supports a defeasible expectation of convergence on particular questions; it does not guarantee a unique final theory or the resolution of every disagreement.” Applied here, precisely: it does not guarantee that this particular ignorance is temporary or that every deadlock will be resolved. ↩

7. A0.P4: “Sharing workable standards of evidence and public rules need not require agreement on every metaphysical foundation. A disagreement about those foundations does not by itself show that someone rejects factual correction or that cooperation with them reduces to force.” ↩

8. Precisely: there is a practical failure we can identify without diagnosing anyone’s metaphysics: replacing an assessment of the claim with allegiance to, or hostility toward, the claimant. That is the failure in such a reply; it does not establish the speaker’s entire philosophical position. And it is a claim about what would change the speaker’s mind, not about what they happen to say. Silence alone is not the tell. ↩

9. Precisely: investigating bias is legitimate. Making a result inadmissible solely because it comes from the other side, regardless of checks or correction, blocks shared inquiry. It need not prove that force is inevitable to be a practice worth opposing. ↩

10. A0.P4: “Within a shared framework, coercive rules must be publicly stated and their factual conditions open to scrutiny. This is a necessary safeguard, not a sufficient justification: the rule itself also requires a normative defence. Factual uncertainty calls for justified standards of evidence and procedures for challenge; it does not by itself make every use of force permissible.” And: “These commitments set no particular evidence threshold.” ↩

11. Precisely: for parties who share the framework, factual scrutiny constrains the application of a rule: an accusation must answer to evidence about something neither party controls. The accuser can be wrong, and the process must leave room to challenge that accusation. This requires neither an infallible instrument nor indisputable knowledge. An uncertain danger may justify a precautionary response; what evidence would justify punishment is a further question. The standards and procedures need their own defence. The existence of an objective fact does not give us automatic access to it or supply those standards. ↩

12. A0.P4: “What defends a rule inside the frame: the norms of a shared life bind whoever wants in and are silent toward those who don’t, so a coercive rule is defended when the people who want in could accept it as a term of living together, whatever their wider views.” And: ""Could accept” includes grudging acceptance short of wanting to leave, by each person’s own weighing; the stricter test I prefer when writing a rule is whether I would agree to be its subject. Both are practical standards for real people, not claims about the right rule for idealised agents.” ↩

13. A0.P4: “This shares Rawls’s aim of shared terms across worldviews, with a lower bar (grudging acceptance by people who could leave, close to what he calls a modus vivendi) and an exit he leaves out; force used on those who do not want in borrows no legitimacy from the frame.” Rawls is paraphrased here, not quoted. In Political Liberalism (1993) he argued that people holding incompatible worldviews can still share the terms of a political life, and that political power is fully legitimate only when it is exercised under a constitution whose essentials all citizens could reasonably be expected to endorse. He hoped for an overlapping consensus: each worldview endorsing those shared terms from within, for its own moral reasons. On facts we are closer than a quick reading suggests. His limits on public reason cover constitutional essentials and matters of basic justice, not whether a particular accused did it, and in A Theory of Justice (1971) he treats a criminal trial as a procedure that aims at the true verdict and is imperfect because it can miss. That a rule’s trigger must answer to reality is something I share with him, not a correction of him. Where I part from him is the bar, and the door. He distinguishes an overlapping consensus from a mere modus vivendi, a truce that holds only while the balance of interests holds, and ranks the truce below it. My threshold sits close to what he would call a modus vivendi; I accept it as enough for a rule to count as defended among real people, and I keep the stricter role-reversal test as the better standard to write rules by. Rawls sets the bar at reasonable endorsement. And he treats society as closed, entered by birth and left by death, and denies that the bare right to emigrate makes anyone’s acceptance voluntary. I keep exit in the picture, because my threshold is defined by it. Toward those who do not want in, the shared norms are silent: force can still resolve the conflict, but it borrows no legitimacy from the frame. ↩

14. A0.P4: “The threshold applies only to people who can choose: for someone who cannot reject or leave, a rule is terrain, but it is still policy for whoever imposes it, and it must pass the stricter test and the limits that need no agreement. Taking away someone’s options when one could leave them open is a wrong in itself, even if they are content, because it deprives them of ways they could have grown.” The two limits: A0.P1, “Established suffering is bad for its bearer pro tanto: a weight that never reaches zero.”; and A0.P2, “I uphold an absolute institutional prohibition on torture.” ↩

15. A0.P4: “Where no frame is shared, and under time and resource pressure negotiating one is often infeasible and sometimes impossible, force is the ultimate conflict resolver. No rule legitimates it in advance there: it resolves, it does not justify, and winning confers no legitimacy. It stays bounded by the policies that rest on bearers rather than on agreement: the other party’s suffering keeps its pro tanto weight (A0.P1), and the institutional prohibition on torture remains absolute (A0.P2).” And: “Where no workable frame can be negotiated in the time available, force may still resolve the conflict without justifying its outcome.” ↩

16. A0.P4: “Better information can remove factual disagreements and reveal opportunities for peaceful agreement. Perfect information alone does not guarantee compatible aims or a mutually acceptable settlement.” ↩