Conjectures and Refutations method, by Karl Popper
1. Guessing and criticizing.
This is how people create knowledge.
Somebody makes a guess. Its an idea. And he labels it true to start out. In other words, the idea is innocent until proven guilty. "Proven guilty" is a metaphor for having a flaw/problem/outstanding-criticism.[confusing parts: explain more: only 2 possible truth-values] [2 crits on 1 guess?]
Then we criticize the guess. A criticism is an explanation of a flaw in an idea -- it is itself a guess. So the criticism is innocent until proven guilty too, and so the guess is false. To illustrate this, see the figure below. Notice the logic involved in judging an idea false. It is reduced down to arithmetic of 1 - 1 = 0. So the guess now has a truth-value of 0, which means false.
Then we might criticize the criticism. And that flips the guess back to true.
|Now the guess is true because the criticism that targeted it is now false.|
Then we might criticize that last criticism. And that flips the guess back to false.
So when should we stop criticizing? Also how do we judge which things are flaws and which aren't? The answer to both of these questions is that it depends on the situation.
2. The problem-situation.
It is a principle of rationality that we create ideas in order to solve problems. So if an idea isn't intended to solve a problem, then that's a criticism of it -- so its false.
As for the ideas that do intend to solve problems, they can be criticized and the criticisms are contextual, meaning that they depend on specific details of the problem that the idea is intended to solve. These details are what make up the problem-situation.
Getting back to flaws, something that is a flaw in one problem-situation isn't a flaw in all other problem-situations. A pen that doesn't write well upside-down is flawed for that reason if my problem is that I want to write upside down. But if I want to use the pen to defend myself against an intruder to my home, then the not-being-able-to-write-upside-down thing isn't a flaw here.
So what are the details of a problem? Consider some universal starter questions that can be useful in illuminating the problem-situation.
- What is the problem? In other words, "what are you trying to change?"
- Why is it a problem? In other words, "why do you want to change it?"
- What were the proposed solutions? In other words, what are your "best guesses"? These are the guesses that you or other people did real life testing on. These include existing traditions. Traditions are good to start out with because they have been criticized a lot already by other people and they've been refined a lot with real life testing.
- What are the outstanding criticisms of the proposed solutions? In other words, "why do you think those ideas didn't work?"
- What are your attempted criticisms of the problem? For example, is there another way to get what you want that meets the same goal as this problem?
- What are your attempted criticisms of the idea of working on that problem right now? For example, is there a more pressing problem that needs your attention instead? Another kind of example is where a deadline is reached and we want to put a solution into action now, and revise it later when new flaws are found.
Note that your answers to these questions are guesses. Also, even the questions themselves are guesses that we iteratively refine with criticism.
Here's a visual aid showing these questions as a workflow:
[notes: What are the arrows to the left, what led to noticing a new problem? Why is checking premises arrow back and forth?]
[how about this one?]