Selfish describes a person, action, or motive that prioritizes benefit to oneself while disregarding the needs, desires, or expectations of others. That core meaning holds across everyday conversation, psychology textbooks, and moral philosophy, but the word leads a double life: biologists use “selfish” to describe genes and behaviors that have nothing to do with conscious intent, economists build models around agents assumed to be perfectly self-interested, and psychologists treat selfishness as something more specific than just looking out for yourself. Unpacking how the same word operates in these different arenas clears up a surprising number of misconceptions about human nature.
What Psychologists Mean by Selfishness
In everyday speech, calling someone selfish is a moral judgment. Psychology tries to get more precise. A recent framework defines selfishness as a perceived situation-specific desire to benefit the self that disregards others’ desires and the prevailing social expectations for that situation.1PubMed. Psychological Selfishness That definition has three moving parts: you want something for yourself, you ignore what other people want, and you violate the norms that apply in the moment. All three have to be present for an act to register as selfish rather than merely self-interested.
This distinction matters more than it sounds. Buying yourself a nice dinner is self-interested, but nobody calls it selfish unless doing so meant breaking a promise or ignoring someone who needed help. The social expectation piece is what separates healthy self-care from behavior that other people recognize as selfish. It also explains why the same act can feel selfish in one culture or situation and perfectly fine in another: the norms shift, so the judgment shifts with them.
The psychological definition is also about perception, not just behavior. People detect selfishness in others by reading context, and they can detect it in themselves too. That means selfishness isn’t a fixed personality label so much as a pattern that emerges when desire, disregard, and norm violation line up in the same moment.
Selfishness vs. Self-Interest
One of the most persistent confusions in discussions of selfishness is treating it as identical to self-interest. They are not the same thing. Self-interest simply means acting in ways that benefit you; selfishness means doing so at a cost to others or in violation of shared expectations. Every selfish act is self-interested, but the vast majority of self-interested acts are not selfish. Eating breakfast, applying for a job, saving money for retirement: all self-interested, none selfish under normal circumstances.
This confusion has deep roots in economics and philosophy. Classical economic models often assume that agents are rational self-interest maximizers, and that assumption sometimes gets casually restated as “economics assumes people are selfish.” But researchers have shown that the willingness to rely on other people’s self-interested behavior is itself a separate question from whether people actually behave that way.2Management Science. Do People Rely on the Self-Interested Maximization of Others? An Experimental Test In practice, people cooperate, share, and sacrifice far more often than a purely self-maximizing model would predict. The economic “self-interested agent” is a simplifying assumption, not a description of what people actually do.
In ethics, the philosophical position that people should act purely in their own self-interest is called egoism. Debates about whether egoism can serve as a coherent moral framework have been running for centuries, and the arguments against it come from several directions, including the observation that a world of pure egoists would be socially unworkable. Distinguishing self-interest from selfishness matters practically: you can advocate for your own needs without being selfish, and understanding where the line falls helps in relationships, workplaces, and everyday moral reasoning.
The Biological Meaning of “Selfish”
When biologists call something selfish, they don’t mean it has feelings or intentions. They mean it behaves in a way that increases its own success at the expense of something else. The most famous use of the word in science is probably the “selfish gene” concept, which proposes that it is more correct to view adaptation as occurring at the level of the gene rather than the organism.3Journal of Evolutionary Biology. A formal theory of the selfish gene Under this view, organisms are essentially vehicles that genes “use” to get themselves copied into the next generation. A gene that makes its carrier slightly more likely to reproduce spreads through the population, regardless of whether that outcome is good or bad for the species as a whole.
The selfish gene concept doesn’t mean organisms are selfish. It means that natural selection favors genes whose effects promote their own replication. This can actually produce cooperation, altruism, and self-sacrifice at the organism level. A gene that makes you help your close relatives survive can spread, because those relatives carry copies of the same gene. The “selfishness” is entirely at the genetic level; the behavior it produces can look generous.
Beyond the metaphorical selfishness of ordinary genes, there are genetic elements that are literally selfish in a more direct sense. Selfish genetic elements are segments of DNA that enhance their own transmission at the expense of other genes in the genome, even if this has no benefit or a negative effect on the organism carrying them.4PubMed Central. Selfish genetic elements Meiotic drivers, for example, are selfish genetic elements that bias their own transmission during cell division so they end up in more than their fair share of offspring.5PubMed Central. A selfish supergene causes meiotic drive through both sexes in Drosophila These elements can spread through a population even when they reduce the health or fertility of the organisms carrying them. They are genetic parasites, and they are found throughout the tree of life.
Are Humans Naturally Selfish?
The question of whether humans are fundamentally selfish or fundamentally good is one of those debates that has been going on since ancient philosophy and still shows no sign of being settled. Modern neuroscience and behavioral genetics suggest that the answer is neither, or rather both. Research on the neurobiology of the “selfish-selfless spectrum” indicates that instead of rigid assignment of human nature as universally selfish or universally good, both characterizations are partial truths that describe different segments of a behavioral range.6PubMed Central. Psychopathy to Altruism: Neurobiology of the Selfish-Selfless Spectrum People vary, and the same person can shift along that spectrum depending on circumstances, training, and experience.
This spectrum view is supported by what we see in very young children. In one study of 15-month-old infants given the chance to share toys, roughly a third shared their preferred toy (an altruistic choice), about a third shared a less-preferred toy (keeping the better one for themselves), and the remaining third didn’t share at all.7PLOS ONE. Fairness Expectations and Altruistic Sharing in 15-Month-Old Human Infants Even before children can talk in full sentences, they already show a spread of selfish and generous behavior. This variation isn’t random noise; the infants who showed stronger expectations of fair distribution in a separate task were also more likely to share their preferred toy, suggesting that even early selfishness and fairness are linked to something meaningful about how a child understands social exchange.
The neuroscience research also shows that where someone sits on the spectrum isn’t fixed for life. Cognitive therapy, social experience, cultural context, and practices like mindfulness training can all shift people toward more prosocial behavior. The brain circuits underlying selfish and selfless behavior are plastic, not hardwired into a single setting. So framing the question as “are humans selfish?” misses the point. Humans are capable of both, and the balance depends heavily on context and experience.
Personality Traits That Predict Selfish Behavior
While most people move along the selfish-selfless spectrum depending on the situation, some personality profiles consistently predict more selfish choices. The cluster of traits psychologists call the Dark Triad, which includes psychopathy, narcissism, and Machiavellianism, is the best-studied predictor. People who score high on Dark Triad measures are more likely to behave selfishly in experimental games that test cooperation and competition.8PubMed Central. The Dark Triad and framing effects predict selfish behavior in a one-shot Prisoner’s Dilemma
Each of the three traits connects to selfish behavior in a somewhat different way. In financial decision-making experiments, psychopathy predicted gambling recklessly with someone else’s money but not with one’s own, a pattern that fits the trait’s association with callous disregard for others. Narcissism was associated with losing more of another person’s money, consistent with overconfident risk-taking at others’ expense. And Machiavellianism predicted strategic selfish behavior, the kind of calculated self-serving choice that weighs costs and benefits rather than acting on impulse.9Journal of Research in Personality. What’s mine is mine and what’s yours is mine: The Dark Triad and gambling with your neighbor’s money These aren’t three names for the same thing. They represent distinct flavors of selfishness: impulsive, grandiose, and calculating.
It is worth noting that most people score low on all three traits. The Dark Triad research is about the tails of the distribution, the people who are consistently and reliably more selfish across different situations. For the average person, whether you act selfishly in a given moment depends much more on circumstances, mood, social pressure, and the stakes involved than on deep personality structure.
How Societies Keep Selfishness in Check
If selfishness is part of the human behavioral repertoire, what stops it from taking over? The short answer is punishment and reciprocity, and these mechanisms work in remarkably similar ways across human societies and animal groups.
In cooperative animal societies, one of the most robust explanations for sustained cooperation among unrelated individuals is reciprocity: individuals help partners who have previously helped them and stop helping those who don’t reciprocate. Models of repeated social interactions show that strategies involving copying a partner’s previous behavior, cooperating when they cooperate and defecting when they defect, tend to produce stable cooperation, especially when cooperators occasionally forgive a defection.10Philosophical Transactions of the Royal Society B. Cheating and punishment in cooperative animal societies This “tit-for-tat with occasional forgiveness” approach doesn’t require anyone to be noble. It works precisely because it makes selfishness costly.
Punishment is the other major force. In social insect colonies, for instance, workers actively police selfish individuals. In one striking example involving desert ants, a single chemical compound on the body surface allows colony members to identify workers who are reproducing selfishly, triggering aggressive policing behavior.11PubMed. Social evolution: the smell of cheating The selfishness-policing dynamic can escalate into something resembling an arms race, where increased policing pressure selects for more extreme selfish strategies, which in turn selects for stronger policing.12Evolution. Arms Race Between Selfishness and Policing
In humans, punishment takes on an additional dimension because of social norms. Research on how punishment shapes behavior in observers found that when a third party, someone not directly harmed, punishes selfish behavior, it communicates stronger social norms than when the injured party does the punishing. Observers who witness third-party punishment develop more positive beliefs about how much others share and how much sharing is socially expected, and those beliefs in turn reduce their own selfish behavior.13PLoS ONE. The source of punishment matters: Third-party punishment restrains observers from selfish behaviors better than does second-party punishment by shaping norm perceptions This is how norms propagate: not just through direct punishment of the selfish individual, but through the signal that punishment sends to everyone watching.
Evolutionary models add a counterintuitive twist. Punishment can more readily invade a selfish population when the punishers themselves start out as selfish rather than altruistic.14PubMed Central. When hawks give rise to doves: the evolution and transition of enforcement strategies In other words, cooperation doesn’t always emerge from good intentions. Sometimes it emerges because even selfish individuals find it useful to enforce rules, and once the rules are in place, more genuinely cooperative behavior can follow.
The Costs of Being Selfish
Research on social motivation has examined the consequences of selfishness and its opposite, sometimes called “otherishness,” for psychological well-being, physical health, and relationships.15PubMed. Social Motivation: Costs and Benefits of Selfishness and Otherishness The general direction of the evidence is that chronically selfish behavior tends to erode relationships and undermine well-being over time, while prosocial behavior tends to strengthen both. This isn’t because selfishness feels bad in the moment; it often feels great. The costs are longer-term and social: people avoid you, trust erodes, and the network of mutual support that humans depend on starts to thin out.
That said, the relationship between selfishness and well-being is not perfectly linear. Some degree of self-interest is necessary for survival and mental health. People who chronically suppress their own needs in favor of others, sometimes called “pathological altruists,” can burn out, develop resentment, or find themselves in exploitative relationships. The research on the selfish-selfless spectrum mentioned earlier suggests that both extremes carry costs. The healthiest position appears to be somewhere in the middle: attentive to your own needs but not at the expense of your social bonds.
Selfish Agents in Artificial Intelligence
The concept of selfishness has found a second life in computer science, where researchers build simulated agents that are programmed to maximize their own reward and then study what happens when those agents interact. In multi-agent systems trained with reinforcement learning, agents given a purely selfish objective function, meaning they are told only to maximize their own score, can still learn to cooperate if the system allows them to choose their partners. When agents can select who they interact with, they naturally gravitate toward other cooperators and avoid defectors, producing a prosocial society despite each individual agent being motivated entirely by self-interest.16Proceedings of the AAAI Conference on Artificial Intelligence. Partner Selection for the Emergence of Cooperation in Multi-Agent Systems Using Reinforcement Learning
The parallel with biological and human selfishness is striking. Just as reciprocity and partner choice keep selfishness in check among animals and people, giving artificial agents the ability to choose who they work with produces cooperation from purely selfish motivations. The agents learn a strategy that essentially mirrors the tit-for-tat approach seen in animal cooperation: cooperate with cooperators, retaliate against defectors. No agent needs to be programmed with fairness or altruism. The structure of the environment, specifically the freedom to walk away from bad partners, does the work.
This result is a useful reminder that selfishness as a starting condition doesn’t have to produce selfish outcomes. Whether in genetics, human psychology, or artificial intelligence, the architecture of interaction often matters more than the motivations of the individual actors. Systems that allow punishment, reputation, and partner choice tend to channel selfish motivation toward cooperative behavior. Systems that don’t, where you’re stuck with whoever you’re paired with and there’s no cost to defecting, tend to let selfishness run unchecked.
Why the Word Itself Causes Confusion
Part of the difficulty with “selfish” is that it carries strong moral weight in everyday language but gets used as a neutral technical term in biology. When Richard Dawkins titled his famous book The Selfish Gene, he wasn’t arguing that organisms are morally selfish or that selfishness is natural and therefore good. He was using “selfish” as a metaphor for a process that has no moral content at all. But the metaphor stuck, and it regularly gets misread as a claim about human nature. People hear “genes are selfish” and conclude “so science says we’re all selfish,” which is roughly the opposite of what the theory actually implies. Selfish genes frequently produce cooperative and altruistic organisms, because cooperation often serves the gene’s interest in getting itself replicated.
The same slippage happens in economics. When economic models assume self-interested agents, people sometimes interpret this as economists endorsing selfishness or claiming that selfishness is rational. But the models are simplifications designed to generate predictions, not descriptions of how people should behave. And empirical tests consistently show that real people deviate from pure self-interest in ways that the models struggle to capture.
In psychology, “selfish” is somewhere between the everyday moral label and the technical biological term. Researchers try to be precise about what they mean, defining it in terms of disregarding others’ desires and social expectations, as described earlier. But in clinical and popular contexts, the word still carries its full emotional charge. Being called selfish stings in a way that “self-interested” doesn’t, even when the behaviors described might be identical. That emotional weight is useful, because it signals a genuine social violation, but it also makes clear-headed discussion harder. Recognizing which definition of “selfish” is in play in any given conversation, the moral, the biological, or the psychological, goes a long way toward avoiding arguments that are really just about word choice.