What Are Concepts, Anyway?

Try this: think of something—anything. A cat. A number. The color red. Justice.

When you think about “cat,” what’s actually happening inside your head? You’re not looking at a cat right now, but you can think about one. You can imagine what it looks like, what sounds it makes, what it feels like to pet one. You can think about cats in general—not any particular cat, but the idea of cats.

Philosophers call these mental things concepts. Concepts are the building blocks of thought. You combine them to make more complicated thoughts: CAT + LIKES + STRING = “Cats like string.” Concepts are what make it possible to think about things that aren’t right in front of you, or things that don’t even exist. You can think about unicorns without ever meeting one, because you have the concepts UNICORN, HORN, and so on.

But here’s the weird part: nobody really agrees on what concepts are. Are they pictures in your mind? Are they abilities you have? Are they something else entirely? This disagreement matters because how you answer it changes how you think about thinking itself.

Are Concepts Like Pictures in Your Head?

One popular idea is that concepts are mental representations—basically, symbols inside your brain that stand for things in the world. Just like the word “cat” on this page stands for cats, your brain has a symbol for CAT that it uses when you’re thinking about cats. This view is called the “mental representation” view, and it’s the most common one in cognitive science.

According to this way of thinking, when you believe something—say, that your friend Alice is taller than Bob—your brain has something like a sentence written in “mentalese,” the language of thought. Your brain puts together the symbols for ALICE, TALLER, and BOB in a particular way, and that combination is what it means to have that belief. Your beliefs and desires and other thoughts are all made of these internal symbols.

This view explains some cool things. For one thing, it explains why you can think an unlimited number of different thoughts. You have a finite number of concepts (maybe thousands), but you can combine them in endless ways, just like you can make endless sentences from a finite set of words. It also helps explain how your brain can run mental processes that are rational—your brain is literally manipulating symbols according to rules, kind of like a computer.

But some philosophers don’t like this picture. A guy named Daniel Dennett pointed out that you probably believe zebras don’t wear overcoats in the wild, even though you’ve never specifically thought about that before. If beliefs are just symbols in your brain, where was that symbol before you read this sentence? It seems like you had the belief without having the symbol. Also, critics worry that if thinking is just manipulating internal symbols, then we need another layer inside our heads to interpret those symbols, and then another layer to interpret that, and so on forever—an infinite regress.

Maybe Concepts Are Just Abilities

Other philosophers think the mental representation view gets things backwards. They say concepts aren’t things inside your head at all. Instead, concepts are abilities—things you can do. Having the concept CAT means you can recognize cats when you see them, you can tell cats from dogs, you can make reasonable inferences about cats, and so on.

This view goes back to Ludwig Wittgenstein, a famous philosopher who was deeply suspicious of the idea that thinking is a kind of inner mental process. He thought that trying to explain language by appealing to a “language of thought” just pushes the problem back a step. If you need mental symbols to explain how you understand language, then you need another system to interpret those symbols, and then another… you never get anywhere.

The abilities view is appealing because it’s grounded in what you can actually observe. We can see that someone can tell cats from dogs; we can’t see their mental representations. But critics point out that this view doesn’t explain much. How exactly do you recognize a cat? Probably by using mental representations. And how do you draw inferences? Probably by manipulating those representations. The abilities view might just be describing what needs explaining, rather than explaining it.

Or Maybe Concepts Exist Outside Your Head

Here’s an even stranger idea: concepts might be abstract objects that exist independently of any human mind—like numbers or shapes. According to this view, the concept CAT isn’t a thing in your brain or an ability you have. It’s a kind of meaning—the thing that the word “cat” and the thought about cats both point to.

This view is inspired by Gottlob Frege, a German logician and philosopher from the late 1800s. Frege noticed that you can think about the same thing in different ways. You might think “George Orwell wrote Animal Farm” and “Eric Blair wrote Animal Farm”—but those are different thoughts, even though George Orwell IS Eric Blair. The difference, Frege said, is that “George Orwell” and “Eric Blair” have different senses—different ways of presenting the same person. Concepts, on this view, are these senses.

One advantage of this view is that it explains how two people can share the same concept even if their brains work differently. The concept is a public thing, not a private mental object. It also makes sense of the fact that there might be concepts nobody has ever thought of yet—concepts that exist as possibilities, waiting to be discovered.

But critics ask: if concepts are abstract objects floating around somewhere, how do we get access to them? How does your brain connect to something that isn’t physical? Philosophers who like this view use the word “grasping”—we “grasp” concepts—but that’s just a metaphor. Nobody really knows how it would work.

Do We Even Need to Pick One?

Some philosophers think maybe all three views are right in their own way, and the debate is mostly about what we want to call a concept. You could call mental representations “concepts₁,” abilities “concepts₂,” and abstract meanings “concepts₃,” and then the argument just disappears.

But most participants in the debate think it’s not just terminological. The different views come with big commitments about how to study the mind. If you think concepts are mental representations, you’re likely to think the best way to understand thinking is through psychology and neuroscience. If you think concepts are abilities, you’re more likely to focus on what people can do and say. If you think concepts are abstract objects, you might think philosophy can discover truths about concepts by thinking from the armchair.

What Are Concepts Made Of?

Let’s assume for a moment that concepts are mental representations. The next question is: what’s inside a concept? Are concepts built out of simpler parts, or are they basic atoms?

Take the concept BACHELOR. For a long time, philosophers thought it had a definition: BACHELOR = UNMARRIED + MAN. According to this “classical theory,” every concept is defined by a set of necessary and sufficient conditions. If something is an unmarried man, it’s a bachelor; if it’s a bachelor, it must be an unmarried man. Simple.

But it turns out that very few concepts actually have neat definitions like this. Philosophers have spent centuries trying to define KNOWLEDGE (as JUSTIFIED TRUE BELIEF, plus something), TRUTH, JUSTICE, and so on—and they still can’t agree. Wittgenstein pointed out that many concepts (like GAME) don’t have a single defining feature. Games involve fun, rules, competition, skill, luck… but none of these are true of all games. They share a “family resemblance” instead of a single definition.

Psychologists have also found evidence that our concepts work differently. When people are asked to categorize things, they’re faster at categorizing typical examples. People say “apple” is more typical of FRUIT than “olive” is, and they’re quicker to agree that an apple is a fruit. This makes no sense if FRUIT has a strict definition. A definition would include or exclude things equally, not have degrees of typicality.

This led to the “prototype theory,” which says concepts are built from typical features rather than necessary conditions. FRUIT might include features like “sweet,” “grows on trees,” “has seeds,” and “you can eat it raw.” Something counts as a fruit if it has enough of these features—not necessarily all of them. This explains typicality effects: apples have more fruit-features than olives, so they’re more typical.

But prototype theory has its own problems. It doesn’t handle complex concepts well. Consider PET FISH. The prototype for PET FISH might include “brightly colored,” but that doesn’t come from the prototype for PET or the prototype for FISH separately. Something weird happens when you combine them. Also, people don’t always use prototypes. If you surgically alter a dog to look like a raccoon, most people (even kids) still say it’s a dog, not a raccoon. They’re using deeper knowledge about what it is, not just surface features.

This deeper knowledge is what the “theory theory” tries to capture. According to this view, your concepts are organized like scientific theories. You have a “folk biology” theory that tells you being a dog is about having dog DNA or dog insides, not looking like one. This theory changes as you grow up—kids’ concepts really do get reorganized, like scientists replacing one theory with another.

But if concepts are embedded in whole theories, then it’s hard to say when two people share the same concept. If your theory about cats is slightly different from mine, do we have the same concept? Probably not, if theory theory is right. That seems too strict—surely we can both have the concept CAT even if we disagree about some cat facts.

Then there’s a radical view called “conceptual atomism.” According to this view, most concepts have NO internal structure at all. They’re like atoms—basic units. A concept like CAT isn’t defined by any features or embedded in any theory. It just points directly to cats in the world. You have the concept because you’ve been in causal contact with cats, not because you know anything about them.

This might sound crazy, but it solves some problems. It explains why definitions are so hard to find—there aren’t any. It explains how people with very different beliefs about cats can still share the same concept. And it’s supported by some famous arguments about how names work. The name “Gödel” doesn’t mean “the discoverer of the incompleteness of arithmetic,” because we could discover that someone else actually discovered it, and we’d still be talking about Gödel. The name just tags the person directly.

Are Concepts Innate?

Here’s another ancient debate: are some concepts built into us from birth? The empiricist tradition (going back to John Locke and David Hume) says no. All our concepts come from experience. We’re born with blank minds, and everything we know is learned through our senses. The nativist tradition (going back to Plato) says yes. Some concepts—maybe the most basic ones—are part of our mental hardware.

This debate has gotten new life from cognitive science. Noam Chomsky famously argued that children learn language too quickly and from too little data to be learning it from scratch. They must have an innate “universal grammar” that guides them. If that’s right, then linguistic concepts (like the concept of a noun or a verb) are innate.

The philosopher Jerry Fodor pushed this to an extreme. He argued that you can’t really learn a concept. Learning a concept involves forming a hypothesis about what it means, but to form that hypothesis, you already need the concept. You can learn complex concepts by assembling them from simpler ones, but the simplest concepts—the “atoms”—must be innate. Fodor concluded that almost all our basic concepts (CAT, DOG, RED, etc.) are built in.

Most philosophers and cognitive scientists find this hard to swallow. It seems obvious that we learn what cats are by encountering them, not by being born knowing. But Fodor’s argument is surprisingly hard to refute. Some philosophers have tried to develop accounts of how you could learn a primitive concept—not by testing hypotheses, but by having your brain physically rewired by experience. But this gets technical fast, and nobody’s fully satisfied.

Can You Have Concepts Without Language?

Finally, there’s the question of whether you need language to have concepts. The philosopher Donald Davidson argued yes. To have a concept—say, the concept of BELIEF—you need to understand that you could be wrong. And understanding that, Davidson thought, requires participating in a linguistic community where truth and falsehood are publicly negotiated. A dog might react to its owner coming home, but it doesn’t have the concept of its owner as such—it doesn’t have a full-blown mental representation of Mr. Smith.

But cognitive scientists have found lots of evidence that animals have sophisticated internal representations. Clark’s nutcrackers (a type of bird) cache thousands of seeds and remember where they are. They also remember which food is perishable and whether other birds were watching when they hid it. If another bird saw them, they’ll go back and move the food later. This requires representing food types, locations, and other birds’ perspectives. It’s hard to explain without saying these birds have concepts.

Whether you’re convinced by the animal evidence probably depends on which view of concepts you already hold. Philosophers who think concepts are mental representations are happy to say birds have concepts. Philosophers who think concepts require language or sophisticated reasoning have to say birds are just sophisticated reflex-machines. The debate is still very much alive.

So What Should You Think?

If you’ve made it this far, you’ve seen that the simple question “What is a concept?” opens up a whole mess of difficult puzzles. Are concepts inside your head? In your behavior? In some abstract realm? Do they have internal structure? Are some of them innate? Can animals have them?

Nobody has settled answers to these questions. Different theories have different strengths and weaknesses, and which one you choose probably depends on what you think philosophy and psychology are supposed to do. Some philosophers think the study of concepts is best done from an armchair, analyzing our intuitions. Others think it’s an empirical science, to be investigated by psychologists and neuroscientists.

The best way to approach these questions is probably to try on different views and see how they feel. When you think about a cat, what does it seem like is happening? Do you have a mental image? A word in your head? An ability to recognize and reason? Probably all three at once. And that’s exactly why the question is so hard.

Appendix: Key Terms

Term	What it does in the debate
Concept	The basic building block of thought; what you use to think about things
Mental representation	An internal symbol in your brain that stands for something in the world
Language of thought	The hypothesis that thinking happens in a mental language, like “mentalese”
Classical theory	The view that concepts have definitional structure (necessary and sufficient conditions)
Prototype theory	The view that concepts are built from typical features rather than definitions
Theory theory	The view that concepts are organized like scientific theories
Conceptual atomism	The view that most concepts have no internal structure at all
Innate	Built into the mind from birth, not learned from experience
Empiricism	The view that all (or almost all) concepts come from experience
Nativism	The view that many concepts are innate
Linguistic determinism	The (controversial) view that your language determines what you can think

Appendix: Key People

Jerry Fodor (1935–2017) — A philosopher and cognitive scientist who championed the language of thought hypothesis and argued (controversially) that almost all basic concepts are innate. He liked radical positions and defended them with clean, forceful arguments.
Gottlob Frege (1848–1925) — A German logician and philosopher who distinguished between the meaning of a word and its reference. His work on sense and reference inspired the view that concepts are abstract objects.
Ludwig Wittgenstein (1889–1951) — An Austrian-British philosopher who argued that many concepts don’t have definitions but work through “family resemblance.” He was deeply skeptical of the idea that thinking is a hidden inner process.
Noam Chomsky (born 1928) — A linguist and political activist who argued that humans have an innate universal grammar, influencing the nativist side of the concepts debate. His work on language acquisition is foundational.
Donald Davidson (1917–2003) — An American philosopher who argued that you can’t have concepts without language. He thought that only creatures who can interpret speech can have beliefs and thoughts.
Eleanor Rosch (born 1938) — A psychologist whose experiments on typicality effects helped overturn the classical theory of concepts and inspired prototype theory. She showed that categories have “fuzzy” boundaries.

Appendix: Things to Think About

The word “game” covers chess, soccer, solitaire, peekaboo, and Call of Duty. Try to find a single feature that all games share. If you can’t, does that mean the concept GAME doesn’t really exist? Or does it mean concepts don’t work by definitions?
Suppose you meet someone who has the same concept of CAT as you, but believes all cats are secretly robots sent by aliens. Do they really have the same concept? If not, how different can two people’s beliefs be before they’re using different concepts?
Do you think a dog has the concept of its owner? What about a newborn baby? What would count as evidence one way or the other?
Fodor argued that you can’t learn a primitive concept—you have to be born with it. But if that’s true, how did humans ever acquire new concepts throughout history? Did people already have the concept of QUARK before quarks were discovered?

Appendix: Where This Shows Up

Artificial intelligence: When programmers try to build a computer that can “understand” things, they run into exactly these questions about what concepts are and how they work. Different AI systems use different theories (some use definitions, others use prototypes).
Psychology: The debate about concepts connects to real experiments on how kids learn, how people categorize things, and whether different cultures think differently.
Law and ethics: Legal systems depend on concepts like INTENT, RESPONSIBILITY, and PERSON. Changes in how we understand these concepts (e.g., whether corporations are “persons”) have real-world consequences.
Education: How we teach concepts to students depends partly on what we think concepts are. If they’re definitions, we teach definitions. If they’re prototypes, we teach with examples.
Everyday arguments: When people argue about whether a tomato is a fruit or a vegetable, they’re arguing about concepts. The question “What does X really mean?” is a question about concepts.