An Enquiry Concerning Human Understanding, decoded


There is a big difference between sensations and ideas. Ideas are much less vivid and forceful. The idea of warmth is much different from actually feeling warm. Talking about love is much different from feeling actual love. While our thoughts are faithful and accurate, they are faint and dull compared to the original perceptions.

As a result, we can divide all perceptions into two kinds, depending on their force or vivacity. The less forceful perceptions are obviously ‘thoughts’, or ‘ideas’. The perceptions are a little harder to name because, I think, we don’t tend to talk about them all as a set, except in philosophy. I will call them ‘impressions’. By the term impression, I mean all the lively perceptions, like when we hear, see, feel, love, hate, desire, or will.

At first, it seems like our minds are free and limitless. We can imagine anything we desire. We can imagine monsters and other universes without any more effort than it takes to imagine the most familiar objects.

But while our thoughts seem free, they are actually quite confined. All we can do is compound, transpose, augment, or diminish the objects of our senses. When we think of a golden mountain, we join two ideas we already had: gold and mountain. Although the complete idea might be new, its components are not. All the objects of thought are given to us. Our minds can only remix them, and can never come up with something entirely fresh.

There are two proofs of this. The first is a challenge: if someone thinks it is possible to invent something new and not composed of impressions that preceded it, she need only produce the idea. If I cannot produce the impressions, she wins.

Second, we can see that people with some disabilities cannot even imagine the corresponding ideas. Blind people cannot imagine colours, and deaf people cannot imagine sounds.

I admit that there is one possible contradiction that I can imagine, and which may prove that there are some fresh ideas that do not come from senses. If a person had never seen one particular shade of blue, and if a rainbow of shades of blue were presented to her, but with that one shade missing, she will be able fill in the blank and imagine the shade she had never seen. Still, this is so small an exception as to be worthless, so I don’t think it changes my general rule: all ideas come originally from sensations.

Therefore, here is a proposition that seems simple yet is quite profound: when we consider a philosophical term or idea, we should enquire what impression it comes from. If taken seriously, this principle would abolish all the metaphysical nonsense people bandy about. Many abstract ideas are faint and obscure, and we become easily confused about them. We assume when we use some philosophical terms that they have a concrete meaning attached, but this is often not the case. Bringing misty ideas into the light will remove all dispute.


Some ideas are consistently followed by other ideas, and all ideas come from somewhere. Our thoughts are like links of a chain, and each thought is connected somehow to the one before it. This is even true in lighthearted conversations; when the thread of a conversation gets broken, the person who broke it can always say, “Oh, x reminded me of y, and that’s why I mentioned z.”

Strangely, no philosopher has ever listed all the principles of connection. There are only three ways to connect ideas: resemblance, contiguity in time or place, and cause and effect.

It is obvious that these relations exist; that they are the only relations is much harder to prove. All we can do is try several examples and see if there are others.


There are two objects of thought: relations of ideas and matters of fact.

Relations of ideas include geometry, algebra, and arithmetic, and all the other things that are intuitively or demonstratively certain. We can be absolutely certain that “three times five is half of thirty” is true. Thinking this statement expresses a certain relationship between the numbers. The truth of this kind of statement can be found through thinking alone and without looking to experience.

Matters of fact are different. They are not found to be true or false through mere thought, because it is logically possible that every matter of fact could be false. Whenever we need to determine whether a fact is true, we must always check our senses and our memories. For instance, the sentence “the sun will not rise tomorrow” makes perfect logical sense and is distinct. To prove it wrong, we must look outside tomorrow. This makes it quite different from logically false statements; we cannot get a clear idea of the statement “9+1=9”. In some way, the statement does not even make sense.

As far as I know, few philosophers have investigated what reasons we have to trust these matters of fact other than our senses and our memories. It seems to me that all our thinking about matters of fact is based on the relation of cause and effect. Cause and effect is what allows us to extrapolate from memory and sense. If you were to find a watch on a desert island, you would conclude that someone had been there before—that they had caused the watch to be there. All thoughts about facts are the same: heat and light are caused by fire, voices in the dark are caused by unseen people, and so on.

But how do we come to know about cause and effect?

Really, we don’t know much about it at all. We do not know cause and effect from a priori reasoning. When Adam and Eve were first shown water, they could never have known that it could suffocate them. If I gave you gunpowder, there is no way you could tell just from contemplating it that it explodes. Nor could you tell that bread nourishes you but does not nourish a lion.

This is hard to admit, but it seems to be true. We believe that we are born into the world able to see cause and effect, able to see that one billiard ball hitting another causes it to move. Yet it is not true. If we look very carefully and think very cautiously, we must admit that the mind never sees an effect in its cause. The effect is totally distinct and cannot be discovered without observation. The motion of the second billiard ball is distinct from the first, and there is nothing in the first that suggests the motion of the second. When I think about a billiard ball moving towards another, I can imagine many things happening: the second ball moves, the second ball does not move, both balls stop, or the second ball causes the first ball to rebound. All of these are consistent, clear, and conceivable. Why would I prefer one to the others?

Every effect is distinct from its cause and cannot be discovered in the cause, and any association with its cause must be entirely arbitrary. We simply cannot look at a thing or phenomenon and intuit through reason alone what effects it will have.

Scientists and philosophers have discovered a few general causes, like ultimate springs and principles. These are the things like elasticity, gravity, attraction, and so on. These are the general principles that govern nature. The trouble, though, is simple: explaining particular phenomena through general principles only postpones the question. How do we explain the general principle? We cannot explain what caused elasticity, gravity, or attraction.

Mathematics, physics, and geometry are of no use. They proceed from premises to conclusions. Geometry says that if x is the case, then y will happen, but it does not say that x is actually true, or even what x is. We need real experience to know whether something is true and what effects it might have.


We still haven’t reached the end of our questioning, though. Every time we answer a question, another one comes up. We ask, how do we reason about matters of fact? The answer is, all of our thoughts are all founded on the relation of cause and effect. Then we ask, how do we know about cause and effect? The answer: Experience. But how do we draw conclusions from experience? This is a very hard question.

Actually, this question is only hard if we try to answer it with a positive answer. It’s easy if we admit our limitations. I will give a negative answer: we cannot draw conclusions from experience.

Nature hides many things from us. For instance, we can see a ball moving, but we cannot see whether it can impart that movement onto another ball. That power is hidden. We eat brown bread, and it seems to nourish us. But how can we know that the next loaf will nourish us in the same way? We cannot. We imagine that we can, but there is nothing in the appearance of the bread that tells us it will. Of course, it always has, and it probably always will, but there is no way that I can see to actually prove it.

Even if you doubt me, you must admit that there is a difference between these two propositions:

  • I have always found this cause to have that effect, and
  • This cause will always have that effect

I know that everyone always infers the second proposition from the first, but I do not know what reasoning leads to the inference. There is no obvious connection between these two propositions.

There is an obvious but deceptive argument: Until now, the future has always resembled the past, so in the future, the future will resemble the past. This is arguing in a circle, though; we need to prove that the future will resemble the past. We can’t just assume it.

There is another, similar argument:

  • Most of our generalizations have been right
  • Therefore, most of our generalizations will be right

To make this argument, we need to make a generalization about our generalizations: that our generalizations are usually right. Clearly that, too, is assuming the very thing we want to prove.

Perhaps we observe the hidden capacities of things after a number of experiments. This seems to have the same difficulties, though. How do we infer what the capacities are? Why do we admit the nourishing capacity of bread after many instances, but not on the first? And how do we determine that bread will always possess the ability in the future? We need to see many instances of cause and effect before we judge the relation to be true. Who would believe that chickens lay eggs the first time they observed it? If the effect were judged by reason, one observation would be enough.

Of course, you may say that I am a bit of a hypocrite: I eat bread and trust that the sun will rise. I look both ways when I cross the street, because I don’t doubt that traffic retains the ability to kill me. As a person, I am quite satisfied that I should continue eating. Yet, as a philosopher, and as a skeptic, I want to know why.

Only the arrogant think that the things they can’t see don’t exist. It might even be that philosophers will look for ages before finally concluding I am wrong. Still, there are some considerations that might remove all doubt.

Infants and animals learn from experience. Even a baby knows learns from one accident that she cannot touch a flame. If a baby learns from reason or argument, I would like you to produce the reasons and the argument. If you can’t, well, the baby is smarter than you. And, of course, if you can, the baby is smarter than me.




[1] Kant will have much more to say on this. He calls it the principle of contradiction, and it’s the starting point for some deep philosophy.

[2] Descartes said this very thing in the Meditations.

[3] I cannot for the life of me imagine why Hume would admit this exception. It begs readers to find similar ones and to expand upon it. What about musical notes, for instance? I think the argument is even stronger. And how big a gap can the mind fill in? Expanding like this, bit by bit, we undermine Hume—perhaps fatally.

[4] Even philosophers realize that philosophers are often full of crap. Every once in a while, one comes along and tries to get us all talking sense. It never works and it never lasts.

[5] This is huge. He is saying that we know much less than we assume.

[6] This is the “problem of induction”. Induction is, roughly, science. We look at the world and draw generalizations about it. How can we do that when the world changes all the time, and we can’t see deep inside its mysterious workings? An example might help.

Alice and Biljana are in a bar. Biljana’s getting a divorce, and she turns to Alice, her lawyer friend, and says “Alice, I’m really stressed. I don’t know how to find a lawyer I can trust. How can you tell?”

Biljana’s eyes brighten. This is something she knows about. “Alice, it’s easy. You look for a lawyer whose shoes match their bag. It’s that simple! All the best lawyers colour coordinate.”

“Wait, wait! Really? That seems crazy.” Alice says. “How do I know I can trust your advice?”

Biljana replies, “Alice, just look at my shoes and bag! They match!”

Scientists are like lawyers. When they’re asked “How do you know that you can draw conclusions from evidence?”, they reply, “Easy—just look at the evidence! Most times when we draw conclusions from evidence, the conclusion’s correct!”

Like Biljana, they’ve missed the point: drawing a conclusion from evidence about drawing conclusions from evidence is arguing in a circle. The question is “why does induction work?” not “what evidence do we have that induction works?” To answer the second question, you use induction to show that induction works, assuming the very point you’re trying to prove.



[7] To philosophers “begging the question” means something entirely different. Question begging is assuming what you want to prove then pretending you proved it.