How Do You Know What Someone Else Is Thinking?
The Mystery of Mindreading
You walk into the kitchen and see your best friend staring at a plate of cookies, frowning. Without a word, you know something’s wrong — maybe there are no chocolate chip ones left. You didn’t read a textbook or follow a series of logical rules. You just felt what she was feeling. Philosophers call this everyday magic mindreading: the ability to guess what other people are thinking, feeling, or wanting. It happens in a flash, but how does your brain pull it off? For more than thirty years, two rival explanations have competed to solve the puzzle.
Two Big Ideas: Rulebooks vs. Replay
The first explanation says mindreading works like a little scientist inside your head. According to Theory-Theory, you carry around a silent theory of mind — a set of rules about how minds work. For example, one rule might be: “If someone wants something and believes a certain action will get it, they’ll do that action.” You use these rules like a detective: you observe the clues (frown, plate), plug them into the rulebook, and deduce the answer (disappointment). Psychologists like Alison Gopnik (b. 1955) and Henry Wellman argue that children build this rulebook the way scientists build theories — by collecting evidence and revising their guesses as they grow up.
In 1986, two philosophers, Robert Gordon and Jane Heal, each proposed a very different idea. According to Simulation Theory, you don’t need a rulebook at all. You reuse your own mental machinery. Instead of following an abstract rule about disappointment, you put yourself in your friend’s shoes: you imagine wanting a chocolate chip cookie and seeing none. That imaginary scenario runs through your own brain’s “emotion engine,” and out pops a feeling of disappointment. You then project that feeling onto your friend. In other words, you simulate her mental state — you create a copy of it inside yourself and use that copy to understand her. The process is like a flight simulator: instead of computing everything from a manual, you rerun a piece of your own system and see what happens.
Mirrors in the Brain
A dramatic piece of evidence for simulation arrived in the 1990s, when neuroscientist Giacomo Rizzolatti and his team in Parma, Italy, discovered mirror neurons in macaque monkeys. They found brain cells that fired both when the monkey grabbed a banana and when the monkey simply watched an experimenter grab one. The same neurons acted as if the monkey itself were performing the action, even though it was just observing. The brain seemed to be mirroring what it saw.
Similar mirror mechanisms turned up in humans. In one experiment, volunteers inhaled foul odors to feel disgusted, and then watched videos of other people making disgusted faces. The same brain area — the left anterior insula — lit up in both cases. As philosopher Alvin Goldman and neuroscientist Vittorio Gallese (1998) argued, these mirror circuits might be the hardware for mental simulation. When you see your friend’s disappointed face, your own disappointment circuit quietly fires, reusing the same machinery that produces your own feelings. This “reuse” version of simulation predicts that simulated states are not quite the real thing: they are off-line copies, pretend emotions that don’t make you cry or run away, but still feel a bit like the real thing.
The Battle for the Truth
If simulation is so powerful, why would anyone need a theory of mind? Theory-theorists fire back with a clever argument. Suppose you try to simulate what a suspension bridge does in a windstorm. That simulation only works if you already know something about physics — a theory of bridges. Maybe when you simulate a friend’s mental life, you secretly rely on a theory of mind to guide your imagination, even if you don’t realize it.
Simulation theorists reply that this collapse only happens when the simulator and the simulated thing are radically different. Human minds, however, are similar enough that you can let your own cognitive engine run on imagined inputs without a manual. You don’t need a theory of disappointment; you just need a disappointment mechanism. And that mechanism generates an honest copy inside you.
A second battlefield is about errors. Simulation Theory predicts egocentric errors: we often assume others think and feel exactly what we would in their situation. The famous “curse of knowledge” shows this perfectly — once you know something, it’s hard to imagine someone not knowing it. If you were using a rulebook, you might still make mistakes, but a pure simulation should be especially vulnerable to projecting your own states onto others. The fact that we do this all the time adds weight to the simulation side.
Yet one of the strongest challenges to a pure simulation account comes from the way children develop mindreading. In a classic experiment with dolls named Sally and Anne, four-year-olds pass a false belief task — they correctly say that Sally will look for her ball in the basket, even though the child knows the ball has been moved. This shows they can represent someone else’s mistaken belief. Three-year-olds typically fail. But in 2005, Kristine Onishi and Renée Baillargeon found that even 15-month-old infants looked longer when a character acted in a way that went against her false belief — suggesting the infants were sensitive to false beliefs long before they could talk about them.
Simulation theorists like Goldman interpret these findings as support for a more primitive, non-verbal form of simulation. The infant doesn’t form a full judgment like “Sally believes X,” but simply imagines the world from Sally’s point of view — re-centering her mental map on Sally. Theory-theorists counter that this still requires a minimal theory of how minds represent the world. The debate remains open.
So Which One Wins? (Spoiler: Maybe Both)
After decades of argument, many philosophers and psychologists now think neither pure Simulation Theory nor pure Theory-Theory can explain everything on its own. Instead, they propose hybrid models. Jane Heal, one of the original simulation theorists, suggested that simulation might be perfect for understanding rational processes — like figuring out what someone will decide when they want sushi and believe calling Yama Sushi will work. But when the link between a mental state and behavior is not rational (like the effect of a drug), we might need a body of knowledge about psychology — a theory.
In these mixed views, simulation often takes the lead, but a theory of mind steps in when simulation fails or when we encounter minds very different from our own. The exact recipe is still unknown. What is clear is that both reusing your own mind and leaning on some kind of psychological rulebook are likely part of the story.
Why This Matters Every Day
The simulation debate isn’t just for philosophers in armchairs. It touches every friendship, every misunderstanding, and every moment of shared laughter. When you wince as your friend stubs her toe, you might be reusing your own pain system. When you catch yourself thinking “I would be furious if that happened to me,” and assume the other person is furious too, you’re living inside the predictions of Simulation Theory — and maybe making an egocentric error. Even the awkward moment when you expect someone to know the punchline before you say it is a simulation glitch.
Understanding how we read minds doesn’t give us a perfect superpower, but it does reveal something surprising: your brain is wired to connect with others by silently replaying their experience inside you. You are, in a literal sense, running a tiny model of your friend in your own head. And when that model gets it right, you don’t have to ask what’s wrong — you just know.
Think about it
- If you could feel exactly what your best friend feels every time she’s sad, would you be a better friend, or would it get overwhelming? Why?
- Think of a recent time when you assumed someone knew something they didn’t. Did that error seem like a “rulebook” mistake or a “simulation” mistake? Can you tell?
- If a computer program perfectly copied your brain’s mirror system and could “feel” your emotions, would you say it understands you? What would be missing, if anything?
