The process by which the KJ method was created
... I tried to solve the classification problem in a crooked and practical way, and as a result, I ran into unexpected difficulties.
That is, no matter how neatly we tried to classify the materials, the result was that the materials did not automatically fit together well.
Originally, classification was based on "homogeneity," or the grouping of materials in the same category. (p.52)
The collection of things with certain common properties in one place is Classification.. But...
In some cases, "summarize" requires more than just a homogeneous summary and analysis. What is it? In other words, it is the process of collecting materials that are completely different in nature and cannot be compared, and "putting them together" in the sense of what meaning can be found in the combination of these materials.
For "summary" in this sense, the procedure of classification alone is far from sufficient in the first place. (p.53)
In other words, grouping with an eye on "homogeneity" does not mean "be collected". In some cases, it is necessary to detect new associations from combinations of heterogeneous data How to integrate heterogeneous data.
Data Integration Methods
That much I figured out by stumbling upon it, but then I had to figure out how exactly to do it. As I was twisting through the field materials, I realized that there was a certain way of doing things. I don't know why, but I began to understand by experience that if I did it this way, it would come together better than before. Such a way of doing things was very vague at first. However, as I gradually refined it, I came to realize that it could be made into a systematic method. The following method of thinking was born out of this need. I had an interesting experience here. This idea was born before its implementation. The logic of why it would work if it were carried out that way was only a reflection of what came later. (p.54)
In other words, after trial and error of various methods, I found a method that was easy for me to do.
The "method" was vague at first, but gradually organized into a system.
How to do it" and "Why it works" are afterthoughts.
How can we "bring together" the chaos and heterogeneity of data in the field sciences? The final bottleneck is thus how to integrate the data. And that leads to the method of conception. (p.55)
Bringing together heterogeneous data
... Information is not just listed, it must be assembled. It must be assembled into something that has some kind of structure. The KJ method, which I will discuss later, is used to find integration in this assembly. This is structure building. (p.61)
Assemble information into something with structure.
If you just do a lot of information gathering, the boxes just line up flat and don't stack up (p. 3).
https://gyazo.com/40bb5c7183431ffabd19ee3b907dbb7b
The boxes piled up on top are not limited to those collected. We often create new boxes by looking at and thinking about the boxes we have collected. Creating new boxes, or abstract knowledge, is "modelling (e.g. a system, etc.)". (p.4) https://gyazo.com/7cecbddf9318a0c723b91b779929f5d3
OLD TITLE: Failure before the KJ method was created
---
This page is auto-translated from /nishio/KJ法が生まれたプロセス 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.