In school, (cough) years ago, I did a bit of epistemology. It’s the study of knowledge. It asks the questions “why do we know that?” and “how will we find that out?”.

It’s infused through everything we do, we just might not realize it. A couple of practical examples.


One big difference between left and right is trust in markets. The left will say “here is what we need, now how do we get markets (people) to go there?”. The right will say “what we need is what the markets (people) say”.

See the difference in epistemology? The left thinks the answer already exists — a priori. The right thinks the answer must emerge — a posteriori.

Programming and security

Resetting a password is actually an epistemic problem. Before you can reset a person’s password, you need to know who they are, with a reasonable degree of confidence. Yet, to know who the person is, to a reasonable degree of confidence, we usually depend on their submitting a password.

Another security issue: SSL certificates. When you go to, and the little lock icon appears in your browser, you are reasonably sure that the site is actually Amazon. Why?

Because Amazon says it is. But anyone could say they are Amazon. So Amazon buys a certificate from (say) Verisign. So now, Verisign is saying, we verify that this is Amazon.

But why do you believe Verisign? How do you know “Verisign” itself is not an impostor? A security person struggles with these questions every day. They are doing epistemology.


From the NYT today: The premise of the current financial regulatory reform is that the establishment missed the last bubble and, therefore, more power should be vested in the establishment to foresee and prevent the next one.If you take this as your premise, the Democratic bill is fine and reasonable. It would force derivative trading out into the open. It would create a structure so the government could break down failing firms in an orderly manner. But the bill doesn’t solve the basic epistemic problem, which is that members of the establishment herd are always the last to know when something unexpected happens.

The truth is, there is no “ultimate authority” out there — there are only degrees of confidence in what you know. And as long as time moves only in one direction, there will always be the question of discoverability and uncertainty. The smartest people I know are the ones who think this is worthy of exploration.