What Does It Mean to Be an Agent?

You’re sitting in class, and you decide to raise your hand. Simple enough, right? You wanted to ask a question, so you moved your arm. But if you stop and think about it, something strange is going on. Your brain sent signals to your muscles, your arm went up—but you did that. Not just your brain, not just your body, but you. What does it mean to be the one who does something, rather than just the place where something happens?

This is the puzzle of agency. Philosophers use that word to talk about what it means to be an agent—someone who acts, who makes things happen, who isn’t just being pushed around like a leaf in the wind. And as you might expect, they don’t exactly agree on the answer.

Acting on Purpose

Most philosophers think that the core of agency is intentional action—doing something on purpose, for a reason. The philosopher G. E. M. Anscombe wrote a famous book about this in the 1950s. She noticed something important: you can describe the same action in different ways. Suppose you turn on a light switch. That’s one thing you’re doing. But you’re also illuminating the room. And you might also be alerting a burglar outside who sees the light go on. Are those three different actions, or one action described three ways?

Anscombe thought it was one action, but what makes it an action at all is that under some description, you did it on purpose. You turned on the light intentionally. That you also alerted the burglar might be unintentional—but it’s still something you did, not something that just happened to you, because it’s connected to your intentional action.

So on this view, agency is basically the capacity to act intentionally. If you can do things on purpose, you’re an agent. Simple, right?

Well, not quite.

When the Agent Disappears

Here’s a problem that has bothered philosophers for decades. The standard explanation of intentional action goes like this: you have a desire (you want to know something) and a belief (raising your hand will get the teacher to call on you), and those mental states cause you to raise your hand. That’s a causal theory of action—your desires and beliefs literally cause your body to move.

But some philosophers say: wait a minute. If your action is just the result of your desires and beliefs causing things to happen, then where are you in this picture? It sounds like the real work is being done by mental states bumping into each other, like billiard balls. The agent—the person—seems to have disappeared. One philosopher, David Velleman, put it this way: the standard theory captures actions that are “defective” or “halfhearted,” where the agent isn’t fully participating. For “human action par excellence”—the kind where you’re really in it—you need something more.

But what’s the “more”? Velleman suggested there might be a special desire—the desire to act in accordance with reasons—that plays the role of the agent. Other philosophers think this is a mistake. Alfred Mele, for example, argues that the standard theory isn’t trying to explain the highest, most excellent kind of human agency. It’s just trying to explain what makes something an action at all. And that’s a different question from what makes someone a fully self-governing person.

The Great Debate: What Kind of Cause?

Now we get into the really tricky stuff. Philosophers have three main ideas about what kind of causation is involved in agency.

Event-causal theories say: actions are caused by events—specifically, events inside your head (like the formation of an intention). This is the most popular view among philosophers today, partly because it fits nicely with science. It says that agency is part of the natural world, where events cause other events, no magic required.

Agent-causal theories say: no, that’s not enough. The agent themselves—the person as a whole—must be the cause. When you raise your hand, it’s not just your intention doing the causing; it’s you causing your hand to rise. This sounds more like our everyday experience, but it raises a difficult question: how does a person (a substance that persists over time) cause something to happen? Usually, when we talk about causation, we talk about one event causing another event. Agent-causation seems mysterious.

Volitionist theories say: the real action happens in the will. There are special mental acts called “volitions”—acts of willing—that are the source of all agency. These volitions are uncaused; they just are actions, by their very nature. This view has the problem that it doesn’t really explain anything—it just says “actions happen” and calls it a day.

Most philosophers go with the event-causal view, but even they admit there’s a serious problem: deviant causal chains.

When Causation Goes Wrong

Imagine a climber holding another person on a rope. The climber intends to loosen his grip and let the other man fall. This intention makes him so nervous that he loses his grip by accident. The intention caused the hand to open, but not in the right way—it made him tremble, not act. So even though the intention caused the movement, it wasn’t an action.

This is the problem of deviant causal chains. Any event-causal theory has to say that the causation must happen “in the right way”—but nobody has been able to give a complete account of what “the right way” is. Some philosophers think this is a fatal problem. Others think it’s just a difficult technical challenge that will eventually be solved.

Do You Need a Mind to Be an Agent?

Here’s another question: what about animals? Or robots? Or simple organisms like bacteria?

If agency requires desires, beliefs, and intentions—mental states that represent things—then bacteria aren’t agents. But some philosophers think that’s too restrictive. They talk about minimal agency: the ability of an organism to regulate its interaction with the environment in a way that keeps itself alive. A bacterium swimming toward food and away from poison is, in this sense, an agent. It’s doing something, on its own behalf, to achieve a goal (survival), even if it doesn’t “think” about it.

This raises the question: is intentional agency just one kind of agency, or is it the only real kind? Most philosophers working on this topic think there are multiple kinds—from bacteria all the way up to humans making complex plans about the future.

The Strange Case of Mental Actions

You might think that thinking is a kind of action. When you decide something, or remember something on purpose, or work through a problem—aren’t you doing something mentally?

But this creates a puzzle. According to the standard theory, for something to be an intentional action, you need to have an intention to do it. But can you really intend to think a particular thought before you think it? That seems weird—you’d have to have the thought in order to intend it, but then you’ve already thought it.

Alfred Mele has a clever solution. When you decide to do something, you don’t necessarily intend to make that specific decision. You intend to settle the question of what to do. The specific decision that results is then something you did, even if you didn’t plan it in advance. It’s like intending to draw a picture without knowing exactly what the picture will look like—you’re still acting.

The Science Question

In recent decades, scientists have started studying agency, and some of their findings have been unsettling.

The most famous experiment comes from Benjamin Libet in the 1980s. People were asked to flex their wrist whenever they felt the urge. Libet measured their brain activity and found that the brain showed signs of preparing to move about half a second before people were consciously aware of deciding to move. Some people took this to mean that conscious intentions are just smoke—they feel like they cause things, but the real work is done by the brain behind the scenes.

Most philosophers think this conclusion is too hasty. The Libet experiment involves a very artificial situation: people are told to act spontaneously, without any reason. In real life, when you have reasons for what you do, the picture looks different. Your reasons genuinely guide your actions, and your conscious intentions play a real causal role.

There’s also research on automaticity—how much of what we do happens without conscious thought. When you drive a car, most of the individual actions (checking the mirror, braking, turning the wheel) happen automatically. But they’re still under the control of your larger goal (getting to school). So automaticity doesn’t mean you’re not an agent; it just means that agency works at different levels.

What About the Feeling of Doing?

When you act, you usually have a sense that you’re the one doing it. This is called the sense of agency. It seems simple, but it can go wrong. People with certain mental disorders sometimes feel like their movements are being controlled by someone else. In other cases, people feel like they’re controlling things when they’re not (like in some magic tricks or video games).

Psychologist Daniel Wegner argued that the sense of agency is just an interpretation we make after the fact—we see that our action matches our intention, and we infer that we caused it. But most researchers now think there’s more to it. Your brain has internal systems that predict what your movements will feel like, and when the prediction matches the actual sensation, you get a basic sense that you’re in control. This happens automatically, below the level of conscious thought.

The Open Questions

Here’s where things stand. Philosophers agree that agency is important—that being an agent, someone who acts rather than just being acted upon, is central to what it means to be a person. But they disagree about almost everything else.

Is agency just about having the right mental states causing the right movements? Or is there something special about the agent themselves that can’t be reduced to mental states bumping into each other? Can simple organisms be agents without having minds? What about computers that make decisions? And what does science tell us about whether we really control our actions?

Nobody has settled answers to these questions. But that’s part of what makes them interesting. The next time you raise your hand in class, you might wonder: who exactly is doing that, and how?

Appendix: Key Terms

Term	What it does in this debate
Agency	The capacity to act, to make things happen, rather than just being acted upon
Intentional action	Doing something on purpose, for a reason
Event-causal theory	The view that actions are caused by mental events (desires, beliefs, intentions)
Agent-causal theory	The view that the agent themselves, as a whole person, causes the action
Volition	A basic act of will that is supposed to be the source of all agency
Deviant causal chain	When a mental state causes a movement, but in the wrong way so it isn’t really an action
Minimal agency	A simple kind of agency that doesn’t require mental representation—just goal-directed behavior
Sense of agency	The feeling that you are the one doing something
Automaticity	When actions happen without conscious thought, like habits or routines

Appendix: Key People

G. E. M. Anscombe (1919–2001): A British philosopher who wrote a hugely influential book on action. She argued that actions are defined by being intentional under some description.
Donald Davidson (1917–2003): An American philosopher who developed the causal theory of action. He argued that reasons for acting are actually causes of actions.
Harry Frankfurt (1929–2023): An American philosopher who argued that what makes someone a person (not just an agent) is having second-order desires—caring about what you care about.
Alfred Mele (born 1951): A contemporary philosopher who defends the causal theory of action, arguing that it can explain mental actions and that the “disappearing agent” objection is overblown.
David Velleman (born 1952): A contemporary philosopher who argues that the standard theory leaves out the agent’s participation and that we need a special mental state to play the agent’s role.

Appendix: Things to Think About

If scientists could show that your brain decides everything before you’re consciously aware of it, would that mean you’re not really an agent? Or could the conscious experience still matter, even if it comes after the decision?
Are you still an agent when you do something out of habit—like brushing your teeth or tying your shoes—without really thinking about it? What about when you act on a whim?
Can a computer program ever be a genuine agent? If it makes decisions based on rules and data, is that the same as what humans do, or is something missing?
Think of a time when you did something and later felt like “I don’t know why I did that.” Did you still do it? Or did it just happen to you?

Appendix: Where This Shows Up

Criminal law depends on the idea of agency: someone is responsible for a crime only if they acted intentionally (or at least could have chosen otherwise).
Addiction raises questions about agency: when someone can’t stop using a substance, are they still acting? Or is their behavior more like something that happens to them?
Artificial intelligence researchers argue about whether AI systems are agents. When a self-driving car “decides” to brake, is that really a decision?
Sports involve intense focus on agency—athletes train to make their bodies do exactly what they intend, but also learn to trust automatic skills.
Everyday excuses (“I didn’t mean to,” “It was an accident”) rely on the distinction between intentional action and mere happening.