Hello everyone,

I hope you've been doing well — I certainly have.

My new article is finished. I have now left it with the wonderful Iona Italia for editing. It should be published within a week, you'll receive an email when it is.


Who knows?

"The lunatic who believes he is a poached egg is condemned solely on the ground that he is in a minority..."

The otherwise brilliant Bertrand Russell once remarked on how sad it was that we couldn't really distinguish a scientifically sound theory from the ravings of a lunatic. That when you drill down at the justification of any proposition it dissolves into nothingness.

Can anything truly be justified? Is it uncertainty all-the-way down?

Most people have the sense that our knowledge is justified by repeated observations — that is to say; induction.

Many philosophers including Russell and even David Hume seemed to have taken the step to say that:

  • As we can never justify a statement with repeated observations
  • We cannot have objective knowledge.

I wish to convince you that this is a mistake.

This letter will be yet another that I will use as practice for the upcoming article; How we know what we know.

In a previous letter, I promised I would discuss the four strands of David Deutsch's theory of everything — but only got around to discussing the first. In this one, I will discuss the second: Karl Popper's epistemology.


Deducing that induction is a myth

First I want to explain what deduction and induction are.

  • Deduction is when you arrive at a truth from a general rule.
  • Induction is when you infer a rule from observations

Imagine I knew that all people are good.

I also know that you — dear readers — are people.

Therefore I can deduce that you are all good.

This is deduction: I knew a rule and derived a specific truth.


Induction is to do this in reverse, based on observations one tries to induce a general rule.

By my lights, all of you are wonderful (since you are interested in this, after all).

I also know that you are all people.

Can I then induce that all people are wonderful?

No, there isn't reason to believe that all people are wonderful only because all of you are. It simply isn't justified. The sample size I am working with isn't representative of everyone.

And at the time of writing, there are sixty-seven of you. So n = 67.

Assuming I am right to say that all of you are excellent, can I then induce that everyone is? No.

This is the problem with induction: There is no basis to believe that every future observation will conform to those made in the past.

Even if I personally examine a billion people and verify beyond a shadow of a doubt that they were all very, very good people — it would only take one bad person to throw the rule away.

In fact, even if I checked everyone and knew that every person alive was great — the day a bad person is born, the theory is disproved.

The theory that we acquire knowledge through induction states that we come to truth through repeated observation. However, as we can see no matter how many observations we make (n), the number of possible future observations is infinitely larger.

Therefore induction is a myth. We do not arrive at objective truth merely through repeated observations.


Russell's chicken

I am borrowing this analogy once again, courtesy of David Deutsch.

Imagine you are a chicken.

A farmer comes by and feeds you.

The next day, he feeds you again. And the next day, and the day after that.

This happens again and again, over and over.

You induce a rule: When the farmer comes, I will be fed

One day the farmer comes... and wrings your neck!


No matter how many observations you make to justify your view, there is always the chance that another observation will throw it to the wind. Contrary to the common misconception, induction does not and cannot give rise to knowing what is indeed true in the world.

I hope you can sympathise with the bitter note that the likes of Hume, Russell and many others confessed we that can't really know what is true. Yet this was a premature conclusion.

Indeed, we can't come to truth through induction, but it isn't quite right to say that we can never have objective knowledge...


Enter Karl Popper

I am now reading Karl Popper's excellent work: Objective Knowledge.

I am still not very far in, but I am already amazed.

It is true we do not arrive at knowledge through induction.

It is true no matter how many observations we make, another one could always be made to throw our theory into question.

But is it true that therefore we can never know anything at all? Of course not.


I see a white swan, then another and another. This keeps happening, over and over again. I induce the rule: All swans are white.

One day my knowledge is shattered. I visit Australia and find out that everything I once believed was a lie. A black swan gazes at me, almost gleeful at the lesson in epistemology it is teaching me.

Did I learn anything here? Has knowledge been gained? Was truth found?

Yes, it was!

I now know this: Not all swans are white.

Observations can't fully verify a theory, but they can falsify one.

This is the seed of certainty. This is the grain of truth. This is the kernel of objective knowledge.


This is Popper's contribution to humanity: The philosophy of critical rationalism.

We cannot construct truth from ground up, however, we can and should criticise our theories.

It is through the withstanding of scrutiny can it be said for a theory to hold objective knowledge.

Rather than justifying a theory from the ground up, we do our best to criticise all our theories and see which is left standing. That which remains we take as the truth until better theories inevitably take its place.

Being able to prefer certain descriptions of the world to others is the basis for objective knowledge.


I have made the attempt to explain how we explain in a previous letter, but allow me to recap:

  1. We make observations in the world
  2. We conjecture explanation(s) for them
  3. We test these explanations against reason and reality
  4. Good explanations which survive scrutiny we hold as true until better ones arise

As Popper has said, Newton's physics was in line with everything we could observe for as long as any theory is — only to be replaced with Einstein's general relativity. General relativity is in turn being replaced by quantum theory.

We don't seem to get to a final theory, but we are in a process of developing more and more refined theories. We are getting closer and closer to the truth — and this is the case for objective knowledge.


Objective ≠ Absolute

We need not ever land on a final inscrutable truth to claim objective knowledge. Progressively better theories are enough. In fact to be able to say that some explanations are better than others is enough to establish objectivity.

We all use language differently, but there is one confusion I believe is deeply harming our discourse.

That is the different dichotomies of absolute vs relative and objective vs subjective.

Absolute means final, inscrutable and no longer prone to criticism.

Objective here, simply means it has nothing in principle to do with opinion. That there is discoverable truth beyond what we decide it to be.

That Russell was wrong and we can prefer scientific theory to the ravings of a lunatic and be correct in doing so. That we can know things.


I now have a question for you: Were you able to follow that?

Epistemology can often be difficult to talk about, but I tried my best to explain it simply — did I succeed?

I would love to hear what you thought of that, please consider replying and letting me know — or asking any questions.

Thanks for coming to the end of another email,

Be happy,

Sashin.