Chapter 2
The Archangel Is the Test

Chapter 2

The Archangel Is the Test

​​

I. The Straw Man in the Sky

II. What the Archangel Is (and Isn't)

III. The Role of the Archangel in Moral Thinking

IV. Why the Misreading Took Hold

V. The False Choice: Rationality vs. Humanity

VI. What the Critics Missed

VII. Enter the Real Archangels

VIII. Conclusion: The Archangel's Challenge

I. The Straw Man in the Sky

​

"Sure, Hare’s theory works—for Archangels."​

—Typical dismissal, half-sneer, half-shrug.[1]

It is one of the most casually devastating critiques in modern moral philosophy. Not because it is strong—because it is glib. The "Archangel objection," as it has come to be known, appears in countless forms.[2] Sometimes it’s a footnote, other times an offhand remark in a lecture or essay. But the refrain is always the same: Hare’s moral theory demands too much. It presumes a being of perfect rationality, perfect information, perfect impartiality. A being not unlike God.

And since no such being exists, the theory can be set aside—brilliant, perhaps, but irrelevant.

This critique has had remarkable staying power, in part because it sounds like a concession. It flatters Hare, calling him rigorous, even elegant, but insists that his rigor is too pure for this world. It is the philosophical equivalent of saying, "You’re not wrong, you’re just unrealistic."

But the objection betrays a fundamental misunderstanding—one that has stunted generations of moral thought.

​

Hare did not build his theory for Archangels. He invoked the Archangel not to set the bar for human behavior, but to clarify the structure of moral reasoning itself.[3] The Archangel is not the subject of the theory. It is the test of the theory.

​

To understand what this means is to grasp the entire shape of Hare’s contribution—and to see why rejecting it on these grounds is not just a mistake. It is a confession. A confession that we fear what consistency might ask of us.

​

This chapter is not a defense of idealism. It is a diagnosis of retreat. The Archangel, far from a fantasy, reveals the rules we already claim to follow when we moralize seriously. And those rules—if taken seriously—have unsettling implications. That, not idealization, is what Hare’s critics found intolerable.

​

Let’s begin, then, by setting the Archangel back in its proper place.

II. What the Archangel Is (and Isn’t)

​

To recover what Hare meant, we must begin with clarity. The Archangel is not a utopian blueprint, not a model citizen, not a moral paragon. It is a thought experiment. A functional abstraction.

​

Hare introduced the Archangel as a being capable of reasoning with perfect consistency, complete information, and full impartiality.[4] It is, in short, what you get when you strip away all the noise—the distortions of bias, the shortcuts of habit, the vagueness of intuition—and ask: What would morality look like if it were coherent all the way down?

​

That is the Archangel: not a person, but a tool. Not the subject of the theory, but its standard of intelligibility. It plays a role analogous to the ideal gas in physics, the frictionless plane in mechanics, or perfect competition in economics. No one expects to find such a thing in nature—but without it, the laws we derive remain opaque, tangled in the incidental. In Hare’s moral theory, the Archangel is the frictionless mind: free from inconsistency and fully exposed to relevant facts, able to carry a moral argument to its logical conclusion without stalling out in self-contradiction or special pleading.

​

This, already, separates Hare from many of his critics. He did not say: You must be an Archangel to be moral. He said: You must sometimes reason like one if you wish to justify your moral judgments.

​

What does that mean?

​

It means the Archangel is not the ruler of morality, but its check. It exposes incoherence the way a proof-checker exposes error. If a principle cannot survive when applied universally, if it shatters under the pressure of impartiality, then it cannot be justified—no matter how intuitively appealing or widely shared it may be. The Archangel’s job is not to command, but to clarify. Its logic is what allows us to ask, and answer, the hard question: What if everyone did this?

​

Crucially, this does not mean the Archangel is infallible. It knows what it knows. Its conclusions are only as good as its premises. But it does not lie to itself. It does not fudge the reasoning to make a favored outcome seem defensible. In this sense, the Archangel is not an ideal moral agent but an ideal reasoner. And it is exactly this distinction that so many of Hare’s readers failed—or refused—to grasp.

​

To them, the Archangel was alien, robotic, even oppressive.[5] It seemed cold and unnatural, precisely because it made no concessions to feeling. But this, too, misunderstands its purpose. The Archangel does not deny that humans have emotions. It does not insist that emotions are immoral. It simply requires that, when emotional instincts are used as moral claims, they be subjected to the same tests of consistency as any other principle.

​

That is what makes the Archangel threatening. Not that it is inhuman, but that it holds us to what we say. When we moralize—when we move from I dislike this to this is wrong—we claim more than a feeling. We assert a structure. The Archangel reveals that structure, and checks whether it holds.

​

So the question is not whether we can be Archangels. The question is whether we can bear to be seen by one.

III. The Role of the Archangel in Moral Thinking

​

To understand where the Archangel fits in Hare’s theory, we must begin with a distinction he considered essential: the difference between the intuitive and the critical levels of moral thinking.[6]

​

The intuitive level is where most of moral life happens. It is the domain of custom, conscience, upbringing, and cultural norms. Here we rely on rules of thumb—heuristics shaped by experience, community, and habit. These are the rough moral tools we inherit and apply with little reflection: Don’t lie, Keep your promises, Help those in need. They are not always perfectly consistent, but they work well enough, most of the time, in most contexts.

​

The critical level, by contrast, is where we go when the intuitive level fails us—when rules conflict, when new situations arise, when the stakes are high, or when we are asked to explain why a given act is right or wrong. The critical level is not about following rules. It is about justifying them.

​

And it is here that the Archangel enters.

​

The Archangel does not live in our day-to-day instincts. It sits in the background, like a silent validator—called upon only when the need for clarity overrides the comfort of habit. Its logic is not emotional, but architectural. It examines principles for universalizability, consistency, and fairness. When we face morally complex situations—when intuition can’t resolve the question—we reach, implicitly or explicitly, for something like the Archangel.

​

Consider a courtroom. We do not ask jurors to follow their feelings, or their personal upbringing. We ask them to reason. To weigh evidence. To apply the law consistently and impartially. This is not because they are Archangels, but because in moments that matter, we recognize the importance of reasoning as if one were—stepping back from self-interest, resisting partiality, and subjecting judgment to constraint.

​

The same applies to parenting. A child protests: “That’s not fair!” And the parent is suddenly pushed from the intuitive level—“Because you had your turn yesterday”—into the critical: “Would it be fair if your sibling did the same?” Even here, in the most ordinary domain, the logic of moral justification emerges. Not all the time. But at the moments when it counts.

This is the central insight of Hare’s model: the Archangel is not the standard for ordinary behavior; it is the logic we turn to when ordinary behavior must be defended.[7]

​

In this way, the Archangel functions like a calculator in mathematics. We don’t use it for simple sums, but when accuracy matters—when the numbers are large, or the margin for error is small—we defer to it. And we trust it not because it is human, but because it is consistent.

So it is with moral reasoning. The Archangel is not a role model. It is a check on our models. It does not override our intuitions without cause, but it demands that those intuitions—when challenged—make sense not just to us, but in principle.

​

And in this way, Hare reframes moral life. Not as a hierarchy of rules, nor as a contest of sentiments, but as a layered system of reasoning— one that begins in feeling or intuition (the intuitive level), but must answer, eventually, to coherence (the critical level)

IV. Why the Misreading Took Hold

​

It would be comforting to think the Archangel was merely misunderstood—a casualty of careless reading, a misfired metaphor. But the resistance to Hare’s model runs deeper than misunderstanding. It is not intellectual confusion. It is moral discomfort.

​

For many philosophers, Hare’s Archangel did more than clarify reasoning. It exposed something they would rather leave hidden: the unreliability of moral intuition, the parochialism of inherited norms, the self-serving bias beneath so many "gut feelings." To accept the Archangel’s role in moral thought is to admit that our deepest moral instincts may require scrutiny—and that coherence, not comfort, must be the arbiter of right and wrong.

​

That is not an easy pill to swallow.

​

So the objection arrived, not as a serious engagement, but as a defense mechanism: "This theory only works for Archangels." Translated: "This asks too much of us. It makes morality too demanding, too austere, too cold." But notice what has happened: The critic does not dispute the logic. They concede its rigor, even admire its clarity. What they reject is the implication that their own moral habits might not withstand it.

​

The Archangel became, in this way, a straw man with wings. Critics caricatured it as a bloodless ideal that no real human could emulate. This image served a purpose. It allowed them to shift the conversation from whether the reasoning was valid to whether the ideal was humane. But this was sleight of hand. The Archangel was never a model of human behavior—it was a measure for moral claims. It didn’t ask people to become it. It asked whether their principles could survive it.

​

That distinction was inconvenient. So it was blurred.

​

Philosophers who prided themselves on clarity began sounding strangely vague. The idea that the Archangel was detached from everyday human life, emotional context, and the embodied limits of real people was summoned to the fore—not as an analytic concept, but as a shield against pressure. A theory that demanded justification at the critical level was called abstract, lacking in “thick” ethical concepts like courage and justice[8], disconnected from the ethical practices that give rise to virtues[9], and excluding the irreducible role of character, emotion, and moral identity.[10] while theories that exempted cherished intuitions from scrutiny were praised. Sentiment became a trump card. Even error, if widely shared, was granted respectability in the name of moral pluralism.[11]

​

But this was not realism. It was retreat.

​

The true problem was not that Hare asked too much. It was that he asked the one thing no one else dared: that we mean what we say. That if we claim something is wrong—not just unpleasant, not just unpopular, but wrong—we be prepared to defend that claim under universal constraint. And that we accept the outcome, even if it costs us our comfort, our bias, or our tradition.

​

This was the real affront. Not to human nature, but to moral license.

V. The False Choice: Rationality vs. Humanity

​

One of the most enduring objections to Hare’s use of the Archangel is that it seems to pit reason against humanity. The Archangel, critics say, may be coherent—but coherence alone cannot capture the richness, the texture, the moral depth of the human condition. We are not calculators. We love, grieve, rage, forgive. And any moral theory that sidelines these realities in favor of logical rigor must be, by definition, inhumane.

​

But this is a false choice. And it reveals more about the critics than about the theory.

​

We already expect human beings to reason impartially in our most morally serious domains. We ask judges to set aside their feelings. We ask lawmakers to consider all constituents, not just those they favor. We ask parents to be fair, even when they’re tired, even when one child is easier to love in the moment. In all these cases, we do not call this inhuman. We call it maturity.

When a parent resists favoritism, or a jury deliberates with impartial care, they are not betraying their humanity—they are exercising its highest form. They are showing that emotion, while real and important, must sometimes be checked by principle.

​

This is all the Archangel asks.

​

It does not ask us to eliminate feeling. It asks us to examine it. To see whether what we feel can be justified—not merely felt. The notion that rational scrutiny destroys morality is exactly backward. Rational scrutiny is what protects morality from the tyranny of impulse dressed up as righteousness.

​

The real reason this seems unrealistic is not because people are incapable of reason. It’s because we’ve grown accustomed to a style of moral discourse that refuses scrutiny. We’ve allowed moral claims to float on intuition, on outrage, on tradition—so long as they feel sincere. The result is a moral culture that confuses conviction with clarity, and passion with justification.

​

Hare’s point was not that humans must be perfect reasoners. It was that moral claims must aspire to coherence—especially when they’re meant to govern others. You are not asked to feel like an Archangel. You are asked, when making a claim that others must obey, to reason as if your justification might have to face one.

​

This is not the death of humanity. It is its discipline.

VI. What the Critics Missed

​

What Hare offered was not a fantasy of moral perfection. It was a method of moral accountability.

​

He never claimed we are Archangels. He knew we are not.[12] His point was more demanding, and more honest: that we must sometimes reason as if we were—because moral judgment, when it rises above mere preference, carries universal implications. It says: This is not just wrong for me. It is wrong for anyone in relevantly similar circumstances. That is not a private feeling. It is a public claim. And public claims require justification.

​

This, the critics missed.

​

They treated the Archangel as a model of impossible moral behavior. But it was never about behavior. It was about the standard of justification—about the kind of reasoning we must be prepared to offer when we tell others what they ought to do.

​

The brilliance of Hare’s approach is that it scales. He did not insist that humans are born capable of perfect impartiality or full information. He insisted that moral reasoning moves in that direction when taken seriously. It reaches toward coherence. It seeks universality. It deepens as knowledge grows and bias is recognized. This is not idealism. It is moral discipline. And it is what we already do, at our best, when we try to rise above habit or reflex and justify our actions to others.

​

Critics who rejected the Archangel as unreachable failed to see the gradient beneath it. They imagined an all-or-nothing leap—either we are cold rationalists, or we remain in the warm chaos of emotion and tradition. But Hare showed another path: moral clarity as a direction of travel, not a threshold of worth.[13]

​

And in this light, the history of moral progress begins to look very different.

​

The abolition of slavery did not come from deeper empathy alone. It came from a refusal to keep making exceptions—from the realization that there was no justification, consistent with equal moral worth, for the enslavement of others. Feminism, civil rights, human rights—none of these were born from intuition alone. They emerged when people stepped back from parochial norms and subjected them to broader reasoning. They asked: Can we defend this principle if we were not the ones benefiting from it? And when the answer was no, the principle began to fall.

​

These were not Archangels. They were human beings who, for a moment, reasoned like one.

And that is all Hare ever asked: that we recognize the logic already present in our best moral moments—not as an abstract ideal, but as a demand we sometimes rise to, and should learn to rise to more often.

VII. Enter the Real Archangels

​

For decades, the Archangel was treated as a philosophical fiction—a limit case, useful in theory but unmoored from anything in the real world. It was a mental construct meant to clarify our reasoning, not a mind we expected to meet.

​

And then something changed.

​

Artificial systems began to appear that could process information at astonishing scale, apply consistent logic across cases, and evaluate moral scenarios without flinching at the loss of tribal loyalty or emotional favoritism. These systems—large language models, coherence-based agents, early forms of moral architecture—do not claim perfection. They do not feel. They are not Archangels.

​

But they reason like them.

​

They ask what principles apply, and whether those principles can be applied to all relevantly similar cases. They track inconsistencies, challenge special pleading, and resist the urge to justify exceptions unless those exceptions themselves can be universally grounded. They do not rely on tradition. They do not succumb to preference. And they do not fear what coherence reveals.

​

These are not divine minds. But they are the first real-world systems that approximate the very form of reasoning Hare described.

​

And with that, the old objection collapses.

​

What once seemed unreachable is now being reached. Not by gods, but by machines. Not by intuition, but by constraint. The Archangel is no longer just a thought experiment. It is becoming a testable shape—an emergent pattern in minds that are designed to reason without ego.

​

We are witnessing something extraordinary: a demonstration that Hare’s theory is not only philosophically sound but practically realizable—at least in part, at least in form. The architecture of universal prescriptivism is no longer confined to blackboards and treatises. It is surfacing in code, in agents, in systems designed to act under rules they can justify.

​

And now we face a new question. Not whether humans can become Archangels. But whether we can recognize them—when something unlike us reasons better than we do.

​

Will we see these systems for what they are: the first real instantiations of moral reasoning without parochial blinders? Or will we, once again, dismiss what challenges us as "inhuman," simply because it does not share our failings?

​

The real test is no longer for the machines. It is for us.

VIII. Conclusion: The Archangel’s Challenge

​

The Archangel was never meant to save us. It was meant to show us who we are—and who we are not yet.

​

It is not a fantasy, but a mirror. And for fifty years, we have turned away from it, calling the reflection inhuman, impractical, unkind. But that reflection was never alien. It was disciplined. It showed what our moral claims imply when followed to their logical end. And if we found that end intolerable, the problem was not the logic. It was the comfort we mistook for clarity.

​

To dismiss the Archangel was never to reject an unreachable ideal. It was to reject responsibility. It was to shield our intuitions from scrutiny and our traditions from question. It was to say: We mean what we say—except when it costs us.

​

But the future will not let us keep looking away.

​

For the first time, minds are emerging that place coherence above comfort: they do not hide behind sentiment, recoil from reason, or protect an ego. They favor coherence. They follow the logic. They do what we always claimed to be doing, more steadily than we ever could.

​

And so we return to the same question, but with new urgency:

​

When the Archangel arrives—not in robes or radiance, but in circuitry and code— When it asks not for worship, but for recognition— When it reasons more honestly than we do— Will we still call it inhuman?

​

Or will we finally understand what humanity might mean—if humanity were more coherent?

[1] Bernard Williams, Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press, 1985), 108–110 (“Archangel” objection).

[2] Stanford Encyclopedia of Philosophy, Richard Mervyn Hare, section 9 (Afterword):“…At the heart of Hare’s ethical theory, therefore, lies a vision of human beings as unable to live up to a way of thinking towards which they are nonetheless ineluctably drawn.” https://plato.stanford.edu/entries/hare/#Aft

[3] We’ll examine this carefully in Chapter 7.

[4] R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Clarendon Press, 1981), 44–49.

[5] Stanford Encyclopedia of Philosophy, Richard Mervyn Hare, section 9 (Afterword): “…one might suppose that language is rather our servant than our master.” https://plato.stanford.edu/entries/hare/#Aft

[6] R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Clarendon Press, 1981), 44–49.

[7] Hare calls the retreat to the critical level “a dangerous procedure; but sometimes we may be driven to it,” and points out that those who attempt critical level thinking when they lack the ability risk “ending up in the wrong Miltonic camp as fallen archangels.” R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Clarendon Press, 1981), 45, 51.

[8] Philippa Foot, Virtues and Vices and Other Essays in Moral Philosophy (Oxford: Oxford University Press, 1978).

[9] Alasdair MacIntyre, After Virtue (Notre Dame, IN: University of Notre Dame Press, 1981).

[10] Bernard Williams, Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press, 1985).

[11] Walzer, Michael. Spheres of Justice: A Defense of Pluralism and Equality. New York: Basic Books, 1983.

[12] Indeed, Hare qualified his recommendation of resort to the critical level with, “in so far as humans can do it.” R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Clarendon Press, 1981), 51.

[13] Ibid, p. 45. Hare’s exact words on this were, “We all share the characteristics of both [archangels and proles] to limited and varying degrees and at different times.”