What Holds the Reasons Up

Ask somebody why they believe something. They give a reason.

Ask why they believe the reason. They give another.

Keep going. This is not a rhetorical trick; it is a genuine feature of justification, and it cannot continue indefinitely, because you are a finite person with a finite number of beliefs. So the chain has to end, and there are exactly three ways it can.

It can stop. There are beliefs that need no further support, and the chain terminates on them. This is foundationalism, and the beliefs at the bottom are usually said to be things like what you are experiencing right now: you are not inferring that you are in pain from anything.

It can circle. Beliefs support one another, and the structure is a web rather than a tower. This is coherentism, and the circularity is meant to be acceptable if the web is large enough and tight enough that no belief is simply propping itself up.

Or it can run on without end. This is infinitism, it has serious defenders, and the obvious objection is that you do not have infinitely many beliefs and cannot have them.

Each exit costs something, and the costs are easy to state.

Foundations have to justify themselves somehow, and if they can do that, they are not really the bottom. If they cannot, calling them basic looks like stopping the question rather than answering it.

Webs can be perfectly coherent and entirely false. A well-constructed delusion is internally consistent; that is what makes it a delusion rather than a muddle. Coherence alone cannot distinguish a good web from a beautifully organised mistake.

And infinite chains are not something a finite person has.

None of these has won, and people carry on knowing things, which is the fact any account has to accommodate.

Here is a way of holding it that I find more useful than picking a side.

The regress assumes a picture in which you built your beliefs upward from certainties, and that is not how anybody's beliefs came to exist. You arrived somewhere in the middle of a going concern, inherited most of what you think from people who did the same, and repaired it as you went. You have never once, in your life, justified a belief all the way down.

On that reading, the regress is a problem for reconstructing justification after the fact rather than a problem for having it. It is a difficulty about how to present your reasons, not about whether they are any good.

That is not a refutation of the problem and it does change what the problem is about. It also suggests something practical: since most real arguments end when somebody gets tired rather than when a foundation is reached, it is worth noticing which of the three exits you actually took.

Did the chain stop somewhere you are prepared to defend? Did it quietly circle back? Or did it just run out of road?