Saturday morning, February 21st. A news segment shows a single child, named, photographed, her story told in careful detail. Donations pour in within hours. The same day, a report estimates that 10,000 children in the same region will die from preventable disease over the next year—a number derived from careful epidemiological work. The report is noted, filed, forgotten. Nobody sends money. The math is grotesque, but the psychology is completely consistent.

Empathy—the experience of feeling what others feel—is widely treated as the foundation of morality. It isn't. It's a bias. It reliably directs your concern toward whoever is most emotionally salient, most narratively compelling, most visually present—not toward whoever most needs help. The people who feel the most are not the most moral. They are the most manipulable. Genuine moral seriousness requires something closer to the opposite: caring about outcomes without requiring emotional resonance to motivate you.

What Empathy Actually Is

It's worth being precise about what we mean by empathy, because the word covers several different things.

Emotional empathy is what people usually mean: the capacity to feel what someone else feels, to have their distress activate your distress. See someone in pain, feel pain. This is what makes humans instinctively helpful, generous, kind—and it's also the source of the problems I'm about to describe.

Cognitive empathy is different: the ability to accurately model another person's perspective and mental state without necessarily feeling it emotionally. This is closer to what a skilled therapist, negotiator, or detective uses. It's genuinely useful and does not have the same failure modes.

The moral praise we give to "empathetic people" is typically praise for emotional empathy—the feeling kind. But emotional empathy has a structural feature that should immediately concern you: it is not proportional to need. It is proportional to salience.

The Scope Insensitivity Problem

The single most damning evidence against empathy as a moral guide is scope insensitivity: the well-documented finding that human concern for suffering does not scale with the number of people affected.

In one classic study, people were asked how much they'd donate to save 2,000 birds from an oil spill, then 20,000, then 200,000. The amounts didn't differ meaningfully. The numbers changed by a factor of 100; the donations barely moved.

The reason is not callousness. It's that emotional empathy runs on vivid mental imagery—on the ability to picture a specific thing suffering. You can picture one bird, coated in oil. You cannot picture 200,000 birds in a meaningfully different way than you can picture 2,000. The image is essentially the same size. So the response is essentially the same.

This produces a systematic distortion in where concern ends up. One identifiable victim—named, photographed, story told—reliably generates more concern than one thousand statistical victims presented as aggregate data. Psychologists call this the "identifiable victim effect." It is not a bug in human psychology that can be fixed by pointing it out. It persists even after people are told about it.

The implication is ugly: a society that allocates concern according to emotional empathy will systematically neglect large-scale, diffuse, statistical suffering in favor of vivid, individual, proximate suffering. Not because people are bad. Because emotional empathy is not a scale.

Empathy Gets Hijacked

The other problem with empathy-as-moral-guide is that it is extremely easy to manipulate.

Political campaigns know this. They don't lead with statistics about policy impact. They lead with a constituent whose life was affected, shown in close-up, telling their story in their own words. The emotional activation is immediate and bypasses critical evaluation. You feel before you think. By the time you're thinking, the feeling has already formed a prior.

Charities know this. The iconic image of a starving child—named, aged, with their story in two sentences—raises orders of magnitude more money than rigorous analysis of which interventions save the most lives per dollar. People who give based on emotional resonance often fund less effective charities over more effective ones, because the more effective ones are fighting diseases or poverty in ways that don't produce moving individual stories.

Media knows this. Proximity, visual vividness, narrative arc, and individual identification all increase emotional activation. This means coverage—and subsequent concern—is radically disproportionate to the actual scale of suffering. A single attack in a wealthy country generates months of anguished coverage; a famine killing thousands in a poor country gets a paragraph.

Empathy makes you a lever. Anyone who can produce an emotionally salient stimulus gets to point your moral concern wherever they want.

The Difference Between Empathy and Compassion

The alternative to empathy is not indifference. It's what the psychologist Paul Bloom calls "rational compassion"—caring about the wellbeing of others, wanting to improve it, without requiring that you feel their pain in order to act.

A surgeon who feels the patient's pain performs worse than one who does not. The feeling is not what produces good outcomes; it's what produces hesitation, distress, and clouded judgment. Emotional empathy can be exhausting and depleting—a well-documented problem in healthcare workers, where "compassion fatigue" is a real burnout pathway. The doctors and nurses who remain most effective over time tend to be the ones who care about outcomes without being overwhelmed by emotional resonance with every case.

The same applies to moral decision-making generally. If you want to actually help people, the relevant question is not "who do I feel worst for?" but "where can I do the most good?" These questions have very different answers. The first is answered by your emotional system; the second requires thinking.

Peter Singer's famous argument about global poverty follows this logic: if you can prevent something very bad from happening without sacrificing anything of comparable moral importance, you ought to do it. The argument works whether or not you can emotionally simulate the lives of the people affected. It works at scale. Emotional empathy doesn't work at scale—it collapses as the numbers get larger.

What To Do Instead

None of this means becoming cold. Warmth, connection, and responsiveness to people in your immediate life are genuinely valuable—both for them and for the quality of human relationships. The problem is specifically with using emotional resonance as a guide to the allocation of moral concern and practical action.

Distrust the story that moves you most. When a particular narrative activates strong empathy, that's a signal to think harder, not to act faster. Ask whether the intensity of your feeling corresponds to the actual need—or to how well the situation has been packaged for emotional activation.

Give based on impact, not resonance. The charity that makes you cry in the commercial is probably not the one that saves the most lives per dollar. Organizations like GiveWell exist specifically to evaluate effectiveness rather than emotional appeal. Giving to the most effective charities rather than the most moving ones is the single highest-leverage thing most people can do to convert moral concern into actual improvement of lives.

Watch for scope insensitivity in yourself. When you notice that your concern for 1,000 people is basically the same as your concern for 10,000, that's a signal that you're operating on emotional empathy rather than rational compassion. Deliberately recalibrate: the larger number matters more, even if it doesn't feel like more.

Separate feeling from acting. Feeling sad about something is not a plan. It's not even a commitment. You can feel a great deal about a great many things without this translating into anything useful. The question is what you're going to do, which requires calculation rather than feeling.

The Takeaway

The moral intuition that feeling more makes you better is wrong in an important way. Emotional empathy is a tool shaped by evolution for small-group social life—for tracking the wellbeing of people you can see, who can reciprocate, who are part of your tribe. Applied to a world of seven billion people, statistical risk, and global problems, it fails systematically and predictably.

The people who do the most good are not usually the ones who feel the most. They're the ones who asked where their concern could accomplish the most and followed the answer—even when it pointed away from the vivid, the nearby, and the emotionally compelling.

Feeling is cheap. Choosing where to care, deliberately, at scale—that's the harder and more important work. The empathy that moves you easily is the empathy that is easiest to weaponize. The compassion that survives contact with scale and evidence is something colder, more reliable, and ultimately more useful.

Care about people. Just don't let the feeling do all the deciding.

Today's Sketch

February 21, 2026