Critical Fallibilism, Evolution and Digital Error Correction
Critical Fallibilism (CF) is a philosophy about knowledge, reason and learning. It begins with questions. What is knowledge? How do I learn? Which ideas are true or good? What are ideas for? What makes thinking rational? How can I use thinking to improve my life? And what is the origin of species?
The origin of species is evolution. But why was that a question? Because a wolf appears designed while a rock doesn’t. So people were more interested in the origin of wolves than the origin of rocks. (You might wonder where rocks – or atoms – come from at all, but the shape of a particular rock isn’t notable. With a wolf, you can ask where the atoms came from and ask why they are configured in the form of a wolf with eyes, teeth, paws, etc. It’s the origin of the design in the wolf that we’re asking about, not the origin of its molecules.)
A wolf has fur to keep warm; it hunts; its four legs work together to let it run; its stomach extracts energy from food; its eyes let it see; and it can mate to make more wolves. These traits are not random. It’s as if wolves were created by an intelligent designer who thought about the purpose of each part of a wolf.
Eyes are complicated. Most objects wouldn’t work as eyes. Rocks, dirt, legs, teeth and fur can’t see. Even small changes to an eye would prevent it from seeing clearly or at all.
Rocks aren’t so mysterious. Why does this rock have this particular shape? No special reason. Why does this rock have some iron in it and that one doesn’t? It’s just random. Why is this rock at this location instead of twenty feet to the right? Because it rolled down a hill and ended up there, or the wind blew it, or a dinosaur kicked it … there are many possible explanations and they don’t really matter.
Rocks are purposeless. Purpose is interesting. Humans design purposeful objects, e.g. tools like a spear. So people thought maybe the origin of species is a powerful designer, God, who designed a wolf kind of like how a human designs a spear.
But now we know better. There was no person, or alien, or intelligent being, who designed wolves. Wolves were not designed by intelligence. In the whole universe, we know of two sources of design: evolution and intelligence.
Evolution we understand pretty well. But how does intelligence work? Philosophers have been trying to figure that out, but it’s tough.
Intelligence works by evolution. The reason our ideas seem designed and purposeful is because ideas evolve just like wolves evolved.
Evolution works by replication with variation and selection. There are two known evolutionary replicators: genes (DNA or RNA) and memes (ideas).
The thing evolution creates is the appearance of design. (I say appearance because evolution is not actually a designer.) Design is anything that looks intelligent or purposeful.
Knowledge is the appearance of design. Which is information adapted to a purpose.
Learning is getting knowledge. That’s done by evolution of ideas.
(The parts of Critical Fallibilism above are not original. Thanks to people who worked on understanding evolution, like Darwin, Paley and many scientists. Dawkins and Deutsch were the most directly helpful to me regarding evolution. And thanks to Popper and Deutsch for their work connecting evolution to epistemology.)
Ideas are for purposes. They work in a range of contexts (situations) but don’t work in others. A good idea achieves its purpose. A bad idea, an error, doesn’t work. E.g. a method of starting a fire that won’t start a fire is an error. A wrong answer to a question is also an error.
Rational thinking is thinking which makes progress instead of getting stuck on errors. That requires error correction. Errors will happen and can’t all be avoided. Guarantees against error are unavailable. To make ongoing progress, we need ways to find and fix errors.
Errors are corrected by critical thinking. That means thinking of criticisms – explanations of errors (what the error is, why it’s an error).
Rational thinking requires unbounded critical thinking. There shouldn’t be limits on the types of ideas or criticisms considered. “No questioning religion” is an example of a boundary. “No rude statements” is another boundary (though maybe all criticism can be expressed politely, so it’s debatable whether that boundary makes unlimited progress impossible, less convenient, or more convenient). To make ongoing progress, it’s important not to ban categories of ideas from being considered and analyzed. Put another way, the only reasonable way to shut down an idea is criticism, and criticism is itself always open to potential criticism, doubts, questions and counter-arguments.
Evolution of ideas happens in your mind below the conscious level. But it’s also a good model for conscious thinking. The conscious version is brainstorming ideas and using criticism to reject errors. Do a bunch of brainstorming and criticism and you can improve your ideas. But don’t do all the brainstorming then all the criticism after. For a second generation of ideas (like a second generation of evolving animals), you want to brainstorm (birth) ideas related to the ideas that survive the first generation of criticism. For a third generation, brainstorm more ideas related to those that survive the second generation of criticism. And so on.
Don’t be scared to brainstorm ideas which are refuted by criticism. You can look for ways to fix them. Don’t just entirely avoid stuff once you know one reason you think it’s wrong. You can investigate changing an idea while keeping the good parts and fixing the error. If you don’t give that a try, you don’t know if the error was severe (a reason it’s a dead end) or just needed an adjustment to address.
A key issue is which issues are digital or analog. Digital issues correspond to the counting numbers while analog correspond to the real numbers. Digital issues are often binary – there are exactly two possible answers – but can sometimes have more choices. Analog issues always have infinitely many choices. Error correction requires digital issues.
True or false is a way of judging ideas with only two choices, so it’s binary and digital. Non-refuted or refuted is also binary and digital. But good or bad normally refers to an analog scale, a continuum, a spectrum, a matter of degrees: an idea can have some amount of goodness or badness, a little or a lot.
The rejection of analog issues from epistemology is perhaps my most important original idea. We build digital computers and David Deutsch knew that was necessary for error correction, and I think others knew that too. But no one realized that the concept of stronger and weaker arguments was the use of analog within epistemology where error correction is needed but can’t be done in an analog way.
Rational thinking involves trying to understand what ideas refute what other ideas. A criticism is a reason an idea does not work. It’s a binary distinction: either the criticism says why the idea fails or it doesn’t. These can be called decisive or binary criticisms. All other criticisms have a secondary role at best. Many other criticisms, as well as positive arguments, can be translated to decisive criticisms, but some can’t. They can also be used for helping understand and explain the issues and concepts rather than as arguments.