[Previous] Taking Children Seriously | Home | [Next] Berkeley Student Protest March

How To Create Knowledge

So you want to participate in the TCS or ARR projects. How does one do that?

The primary thing to do is knowledge creation: learn the ideas and help improve on them. And learning them enables using them to improve one's own life, while improving them allows one to further improve own's own life.

Secondarily, one can promote the ideas, spread the word, hire people to work on it, publish on the topic, and so on. This will help people as well as creating a bigger community with more people to contribute improvements, which can benefit you.

Let's consider knowledge creation in more detail because there are some misconceptions and confusions about how it works, and because some understanding of Karl Popper's philosophy is helpful to doing it better.

How are ideas like TCS or ARR created? Learned? Further refined and improved?

All knowledge is created by *guesses* and *criticism*. It is a process of trial and error, not one of deriving, induction, abduction, justification or empiricism.

Knowledge also addresses problems. Problems aren't necessarily a bad thing but would include any question one has, or anything one wants to get and isn't sure how to get it.

Step one is to identify a problem. Just find anything that one thinks could be better in any way.

In step two, brainstorm ideas which might solve the problem; make guesses. There's no quality standards here, no rules or limitations, anything goes. And don't worry about coming up with enough ideas right away, you'll have unlimited chances to revisit this step later, so just move on as soon as you want.

Step three is to criticize the ideas. If anything at all is wrong with them, that's grounds for criticism. A criticism is an explanation of a flaw in an idea, and we want to be merciless here and find every flaw we can.

Step four is to do a mix of the previous steps in no particular order. You can have a lot of things in progress at once and bounce around between them, or you can be more methodical, either way is fine. Every time you criticize an idea, you have created a new problem: how could that idea be improved not to have the flaw criticized? And you've also created a new idea (the criticism itself) which can be exposed to criticism. So step three naturally feeds back into steps one and three.

So, criticism drives the process. Criticism identifies new problems we can try to solve, must be criticized itself in case it's mistaken, and sends us back to brainstorming as ideas are rejected. Criticism is the reason the initial steps are relaxed and easy: if any mistakes are made, they are supposed to be caught in the criticism step, you don't have to worry about them in the first steps.

The goal is to come up with a single idea which has no criticisms of it which you think will solve the problem best. When you reach that point, you're done. Everything you learned along the way, and this final result, are knowledge you have created.

There's also the possibility of criticizing the problem one is trying to solve, itself. Problems can have flaws too. Maybe there is a better way to frame the issue, or a better goal to try to accomplish instead. If a problem is criticized, one can try to brainstorm better problems or ways to improve it.

Now that we're familiar with the general method of knowledge creation, let's consider some specifics.

This covers how to solve problems, such as trying to improve an idea. How does this apply to learning?

Learning is itself a problem: trying to gain knowledge one didn't have before. This is accomplished by brainstorming what the ideas one is learning are, and how they work, and improving on that with criticism. It is fundamentally the same process. The main difference is that existing material on the topic can provide suggestions for problems to consider, brainstormed ideas to consider, criticisms, and so on. And one can criticize his brainstormed ideas not just by considering if they are good or bad, but also by considering if they are compatible with the existing material on the topic that one is trying to learn.

How much can this be a collaborative process? Or does it work best as an individual process? It works either way. This process is just as valid within one mind as for a group discussion.

People always do some of their thinking in their own mind, even in the midst of a group discussion. That's important and good. And it leads to the question: is collaboration is important too, or can we rely on individual thinking? And does collaboration create too much extra work having to deal with other people?

Collaboration is extremely important and valuable for two reasons. But first let's consider how difficult it is. Actually, people can frequently work together to create knowledge in an efficient and effective way. All they have to do is share what problem(s) they are working on, share any brainstormed ideas for a combined list, and share their criticisms. This is simple to organize since the basic outline of knowledge creation involves two lists associated with each problem, and anyone could add to the list, all they'd need to do is read it first to avoid duplicates.

To collaborate, people also need to explain their ideas clearly enough for others to understand them. This does take some effort but on the other hand clarifying one's ideas is important even if one is doing individual thinking. Making them clear instead of vague improves their quality and addresses the criticism of the vagueness flaw.

The first benefit of collaboration is that if someone else has a good problem, brainstormed idea, or criticism, then I don't have to think of it myself. Instead of having to figure everything out personally, I can benefit from thinking other people do, and they can benefit form thinking I do. This is the same principle as not reinventing the wheel, and learning math from other people instead of trying to figure it all out from scratch by yourself.

A lot of ideas about TCS and ARR have already been figured out, and it's advantageous to learn those instead of trying to think of them all yourself.

The second benefit of collaboration is that we all have weaknesses, blind spots, irrationalities, and hang ups, as well as strengths and areas of expertise. Thus, someone else might be good at what I'm bad at, and vice versa. So that provides an opportunity to help each other.

It's too hard and unrealistic to find all of our own mistakes. We can find a lot, but we're not perfect at everything and will miss some that other people might find. This is one of the reasons that people who don't collaborate enough sometimes get stuck and don't make much progress.

For topics like romance and parenting, virtually everyone has blind spots and irrationalities. Sometimes it's hard enough for one person in a group of a thousand to see an error that everyone else is making without realizing it. Trying to do that all by oneself every time would be folly. Making mistakes where we don't realize anything is wrong is common for everyone, and it's a hard problem, and collaboration helps us better deal with it since only one person has to find a mistake and can share it with thousands of other people to help them get past their blindness or irrationality.

Now we've covered how to create knowledge to help the TCS and ARR projects make progress, and we've seen why collaboration is important. Let's consider a few specifics.

The best medium for collaborative discussion is email lists. That is why TCS and ARR have email lists which you should join here:



Email lists enable people to brainstorm ideas and share them, to explain problems they are interested in solving (such as problems with current TCS or ARR ideas), and to share criticisms. They also enable asking questions people may have.

Some people like to have discussions in person. That's fine, but it's no substitute for an email lists which allows for worldwide collaboration. Worldwide collaboration means that anyone in the world who knows about this stuff can criticize or contribute, and it means that people from different cultures can use the strengths of their differing perspectives to augment weaknesses of other cultures -- the added variety of perspectives is helpful.

Exposing one's ideas to criticism -- from oneself as well as others -- is crucial. Otherwise one's ideas will have lots of mistakes. This includes one's understanding of what TCS or ARR are. It's not just improvement that works by critical discussion but also learning the existing ideas. Learning is not trivial and needs criticism to deal with mistakes.

This means that people who read a lot of stuff, but never write anything which could be criticized by others, almost certainly do not understand the ideas very well (which may be why they never think of anything they want to say). One needs to expose his understanding to criticism to get it to high quality, not just expose his proposals for improvements to criticism.

Besides email lists, online forums are another good tool. Any online collaborative tool can also be valuable such as instant messaging, IRC chat channels, wikis, collaborate document creation websites, and so on. Computers and the internet are amazing technologies which surpass what came before. The various offline options have some use but are more limited.

Elliot Temple on November 6, 2011

Messages (2)

This. Was. Amazing. Thank you so so much for writing this, it really helped in visualizing the production of knowledge. Have a nice day :)

Anonymous at 2:13 PM on November 18, 2018 | #11385 | reply | quote

Update: The TCS and ARR discussion groups have both been moved to the multi-purpose Fallible Ideas group:


curi at 12:33 PM on November 21, 2018 | #11388 | reply | quote

Want to discuss this? Join my forum.

(Due to multi-year, sustained harassment from David Deutsch and his fans, commenting here requires an account. Accounts are not publicly available. Discussion info.)