Judging Goodness

Moral judgment without a god's eye

The journalist Kelsey Piper gives away roughly forty to fifty percent of her income to global health interventions — malaria nets, deworming, direct cash transfers — selected for measurable impact per dollar. Confronted with a case like hers, most people reach for a casual verdict: she is a better person than I am. The verdict feels natural, and it is broken — not because it is too generous, but because it is incomplete. Better by what standard? Until that question is answered, the sentence has not yet said anything.

I have argued that value exists only in relation to agents — that objective value is a myth — and the standard worry about that position is that it kills moral judgment. If there is no god’s-eye standard, the worry runs, then nobody is better than anybody, Piper’s sacrifice is just one lifestyle among others, and the whole vocabulary of admiration and condemnation becomes noise. This chapter is about why the worry is exactly backwards. Judgment survives subjectivism; what dies is only the casual judgment, the verdict issued from nowhere. Three cases will carry the argument: how to praise a person, how to assess a movement, and how to condemn a practice.

Two Ways to Judge

Every moral judgment of a person is made against a standard, and there are exactly two legible places to get one.

You can judge internally, by your own standards. This is genuine judgment, and it is yours: given what I value, is what she is doing good? The answer commits you, because the standard is explicit and it is the one you live by.

Or you can judge externally, by the agent’s own professed standards. Piper has stated what she believes she ought to do, publicly and in numbers, and she does it. Judged externally, her record is exemplary whatever you think of her framework — she chose an ethical identity and lives inside it, her stated values and her concrete actions transparently aligned. External judgment is the measure of integrity, and it is what gives the charge of hypocrisy its teeth: to call someone a hypocrite is to convict them by a standard they cannot disown, because they announced it themselves — a weapon I examine in Honesty and Hypocrisy.

Both judgments are meaningful. Neither is cosmic. And conflating them is where moral discussion goes to die. “She is a better person” typically smuggles two claims into one: that her standards are the true ones, and that she meets them. The second claim is checkable. The first has no owner — it is a naked preference wearing a verdict’s clothes. This is just Conditionalism applied to persons: all truth is conditional, and a bare “X is good” is an unconditioned sentence that evaluates nothing until the standard is bound to it. Bind the standard — mine, or hers — and the judgment becomes evaluable, debatable, and often simply correct. The binding move is agent-binding: attach the moral claim to a specified agent’s values and it turns from a naked preference into a conditionally objective statement.

Note what this framework predicts: two careful people can judge the same act oppositely, both reasoning impeccably, because they are filling in the conditions differently — a case of reasonable disagreement, not a case of one of them being broken. The disagreement is not noise to be suppressed by a universal standard. It is information about where the standards differ, and stating the standards is how you find out whether the difference is one that argument can move.

Even the virtues that seem to float free of frameworks do not. Integrity and authenticity — choosing a framework consciously and living consistently within it — deserve admiration across virtually any ethical system that values coherence and self-awareness. But “across systems that value coherence” is itself a condition. It happens to be a very widely shared one. That distinction, between the unconditional and the nearly universal, is the hinge of this whole chapter, and I will come back to it.

A Movement Judged Charitably

Effective Altruism is the natural worked example, because it attracts exactly the casual verdicts I am warning against — from admirers who call its members better people, and from critics who call them deluded ones. It deserves a more precise judgment than either.

The charitable reading first, because EA has earned it. It is one of the most intellectually rigorous and empirically grounded moral movements of our time. Its logic is elegant: maximize good outcomes; use evidence and reason to compare interventions; measure value with proxies like quality-adjusted life years; fund whatever does the most, regardless of emotional appeal or proximity. The virtues it embodies — rationality, generosity, consistency, a willingness to act on difficult moral math — are widely respected, and rightly so.

But behind the clarity sits an error: the assumption that these values are not just effective but objectively true — that “you should fund malaria nets over guide dogs” is a fact about the universe rather than a consequence of EA’s value function. The aggregation at EA’s core requires comparing utility across persons and summing it over populations, and that requires a vantage to sum from: either a shared agreement on moral goals, or a meta-agent who coherently weighs everyone’s values from above. No such meta-agent exists. Nature is indifferent, God is silent, and reason — powerful as it is — operates on values without generating any. There is no view from nowhere; there are only overlapping and conflicting subjective valuations. So EA’s “should” is true for those who share its value function and simply inapplicable to those who don’t, and no amount of expected-value arithmetic closes that gap. The full attack on the arithmetic — aggregation, demandingness, the erasure of situated agents — belongs to Against Utilitarianism; here I only need the conclusion that the universalist core fails.

The interesting question is what remains when it does. The answer is: everything that actually matters. Read without the realism, EA is a high-agency moral identity — a set of goals, metrics, and heuristics chosen by agents who want to act in certain ways. It is voluntary, consensual, internally coherent, and agent-owned. It is not morally binding on others, not a consequence of reason alone, and not a universal imperative. EA is a LARP — but one that saves lives, which is more than can be said for most moral frameworks.

That is not a demotion. It is a promotion to honesty. Effective Altruists never needed moral realism; they needed shared goals, coordinated action, rigor, and integrity, and they have all four. Let go of the illusion of objectivity and the movement gets stronger: a community of agents acting on deeply held values rather than missionaries enforcing imaginary absolutes. And the two modes of judgment now apply cleanly. Externally, a committed Effective Altruist who gives what she pledged is exemplary — full stop. Internally, you may share enough of the value function to admire the goals as well as the follow-through, or you may not; either way you can acknowledge the clarity of purpose and the measurable impact, which is how respect works across value systems that will never merge.

Conditional judgment keeps operating even inside the identity. Should giving be public or private? Transparency leverages social norms and multiplies impact through example; privacy guards humility and refuses performative signaling. Neither is superior full stop — each reflects underlying priorities, and the choice reveals as much about the chooser as about the act. A movement that understood itself in the terms I am proposing would treat such disputes as what they are: differences in values to be stated, not errors to be corrected from above.

Practices, Not Persons

Praise was the easy case. The real test of judgment without absolutes is condemnation, and it arrives as a tension between two things nearly everyone believes. Bigotry is wrong: hostility toward individuals based on group identity — race, sex, religion, nationality — erases the person by reducing them to a category. And cultural relativism is wrong: the doctrine that moral norms are wholly determined by culture, so that no culture’s practices can be judged from outside, forbids calling foot-binding, caste discrimination, or forced marriage wrong so long as they are locally sanctioned. But the two commitments appear to clash. If I oppose bigotry, how do I condemn a culture’s practice without being prejudiced against that culture? Isn’t the condemnation itself a form of bigotry?

The resolution is a single distinction: target practices, never persons. “Muslims are evil” is bigotry — a sweeping verdict on people, which denies their individual agency and condemns them for membership. “Blasphemy laws are evil” is a judgment of a practice against explicit standards. The first is not even a judgment in the sense this chapter has developed: it binds no standard, neither mine nor the agents’ own, because it refuses to see agents at all. The second is judgment working exactly as it should — and to judge an action is not to condemn the humanity of those who perform it.

But notice the trap waiting in the traditional way of putting this. The anti-relativist wants to say that practices may be measured against universal standards that transcend culture — and if “universal” means unconditional, that is the absolutist’s cure, and I refuse it for the reasons given throughout this volume. What actually does the work is not unconditional standards but widely shared conditions, deployed through agent-binding. “Given agents who value their own bodily integrity, their agency, and their power to refuse — forced marriage is wrong” is not a decree from nowhere. It is a conditional whose conditions are instantiated almost everywhere, because nearly every agent in every culture values not being mutilated, silenced, or disposed of by force. The girl whose feet are being bound shares those conditions. That is what lets the standards transcend culture without pretending to transcend agents: they function as if universal — the way arithmetic and human rights do — while remaining explicit, inspectable, and criticizable rather than oracular.

This inversion also exposes what relativism actually is. It poses as respect, but it treats cultures rather than agents as the units of moral standing — and cultures do not bleed, suffer, or hope; people do. Judging a practice by conditions its own victims share is not imperialism. Deferring to “their culture” over their agency is: it hands the victims to whoever controls the culture’s definition. The relativist and the bigot make the same mistake from opposite directions — both dissolve the individual into the group, one to excuse everything done inside it, the other to condemn everyone born into it. Refusing both is a single move, not a balancing act.

So the worry I opened with gets its answer. Nothing in subjectivism disarms judgment; every tool survives, sharpened. You can praise Piper — internally if you share her values, externally on her integrity either way. You can assess EA — admirable as a chosen identity, wrong about its own universality. You can condemn blasphemy laws and forced marriage — relentlessly, by conditions their victims share, without a flicker of prejudice toward the people trapped under them. What you cannot do is issue verdicts from nowhere, and you never could; the god’s-eye standard was never doing the work. The conditions were. Say them out loud, and judge.