There are at least three models of learning
There are at least three models for [way of learning
Mental model of attending lectures in one-way traffic
Self-taught, trial-and-error mental models
Mental models to exchange knowledge and learn from each other
Mental model of attending lectures in one-way traffic.
nishio There are at least three models of learning.
First, there is the mental model of attending a lecture, in which "there is a 'teacher' in the role of teaching, and you are the 'student' in the role of 'being taught', and the teacher gives knowledge to the student in a one-way manner.
知識の流れが一方通行.icon
Many software engineers believe that they will never be able to program that way.
Self-taught, trial-and-error mental model.
nishio In contrast, the second model is a "write your own program, run it, observe its behavior, check whether your understanding is correct, and correct it if it is wrong. If it is wrong, you correct it." This is an approach to learning by oneself through repetition of trial and error.
学びのサイクル.icon
This is called "learning programming on your own".
Because education was slow to catch up with the rapid development of IT, outspoken people in the IT industry across the board have come to believe that they have a large weight of self-taught knowledge, that self-taught learning is inevitable, and that it is important to have a self-taught habit. This is why "self-taught" is now spoken of in a positive light
nishio However, there is an important factor here that is often overlooked.
Programming knowledge can be verified by experimentation to see if the understanding is correct.
Knowledge must be verified.
As anyone who can program has experienced, this process involves being "pointed out" hundreds or thousands of times and learning by correcting > mistakes.
nishio On the other hand, those who "self-study" in such a way that they are not "pointed out" for mistakes are not corrected for wrong understanding. This is the reason why the nature of learning is completely different even though the same term "self-study" is used. The amount of "opportunity to learn from mistakes" is totally different.
Mental model of learning by exchanging knowledge with each other.
Then what do we learn about areas where we don't get "pointing out errors through experimentation"?
nishio So this is where "it is essential to have others who can point out errors.
nishio This "stranger" does not necessarily have to be a "teacher". It does not have to be someone who knows more than you in all areas. If they know more than you in some area, that's fine.
Exchange knowledge and learn from each other
互いに知識を交換して学ぶ.icon
nishio So how do we find such people? Here, "knowledge in organizations is becoming more homogenized," which leads to "the need to belong to multiple organizations," which leads to talk of parallel work. If we call this "side job," some people may interpret it to mean "diversification of sources of income," but the important thing is "diversification of knowledge flow channels.
nishio in my book "The Intellectual Production of Engineers" written 5 years ago. ChatGPT] was released later, and I think that further discussion is needed.
---
Inspiration for writing
Self-study reinforces assumptions
reverberation
@takiuchi: Indeed, programmers are in a special environment where they have a dedicated teacher, a processing system, who can teach them unlimited mistakes. The self-taught group has a huge advantage because they can repeat trial and error, which a human teacher can never reach.
hrjn The problem with the lecture format is twofold: first, the instructor is not teaching everything. In development, of course they don't teach how to use a computer or anything like that, but it's actually necessary. You can't make up the details unless you do hands-on work.
Second, imitation is not the same as application. It is more difficult than one might think to take the case A that one has learned and apply it to B.
hrjn Math, in particular, is one of those things where everyone believes that if they learn it, they will be able to do it, but that's not true. You have to have a certain level of generic know-how to actually be able to solve problems.
It is not possible to do it by following the manual, so the knowledge gained must be generalized.
hrjn I don't know myself how to generalize, but I guess I just have to practice with various cases.
By accumulating experiences of what worked and what didn't, the know-how becomes generalized.
I think that's what you mean, maybe.
@IWAN_NAOK: "Multiple Knowledge Inflow Pathways" for learning
I think the key is that this "knowledge flow channel" is not just an input, but a feedback to your actions/ideas.
I think again that it is a very good growth opportunity for me to interact with Japan/the world both in my company/carling.
I'd love to go to grad school again.
---
This page is auto-translated from /nishio/学び方には少なくとも3つのモデルがある using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.