Concept maps = Text?

c4lvorias

Is there something concept maps cannot capture while linear writing can in terms of knowledge extraction? Right now I am leaning towards using concept maps as my main medium for thinking and creating a Zettelkasten.

Andy

By saying "in terms of knowledge extraction", if you mean knowledge extraction in the technical sense described by Wikipedia, you're implying that what you want to do with your Zettelkasten is limited to modeling information in a highly structured manner.

If that implication is true, then you may not need linear writing (but concept maps may not be the best tool either).

But, in general, linear writing provides access to many features of various modes of discourse that would be difficult to express in concept maps. Linear writing may also be more efficient for representing some kinds of information (that's a guess based on thinking about how complex and hard-to-read concept maps can become when many concepts are involved). So if your Zettelkasten use is not limited to modeling information in the highly structured manner of "knowledge extraction" in the technical sense, it would probably be advisable not to discard the features available to you in linear writing.

tomp

@c4lvorias said:
Is there something concept maps cannot capture while linear writing can in terms of knowledge extraction? Right now I am leaning towards using concept maps as my main medium for thinking and creating a Zettelkasten.

Linear writing can capture detail and complication that can't fit into a concept map. Concept maps can be strong in working out (and reminding you of) the elements and how they relate. I use mind maps a lot, and I put a clickable link to them in some of my notes to make them easy to look at while I'm also looking at or editing a note.

I'm not sure what you are thinking of when you write "extraction". Zettelkasten notes are more effective when they represent the results of your own work on the concepts, etc., that you are studying or thinking about. Concept maps and mind maps help with working out that understanding.

c4lvorias

@tomp @Andy To clarify what I mean, I am basically asking whether linear text provides something more than a proposition. Because with propositions, imho, you can represent fully what a linear text basically says.

My goal is to take literature notes in the form of concept maps. Here's a very basic example:

tomp

@c4lvorias said:
@tomp @Andy To clarify what I mean, I am basically asking whether linear text provides something more than a proposition. Because with propositions, imho, you can represent fully what a linear text basically says.

My goal is to take literature notes in the form of concept maps.

It depends on what kind of texts you want to work with and what you want to get out of them, it seems to me. In general I don't think every sentence or even paragraph tries to prove something, but probably non-fiction comes closer. Sentences can be inserted for rhetorical purposes, for example. A paragraph can have several points that support each other. A paragraph, or even several paragraphs, may be filled with rhetorical text meant to set up and support a complex point in a following paragraph. If you have decided beforehand that all you want to get out of your reading is propositions that I suppose that is what you will notice.

Here is a sentence from a work of non-fiction that doesn't lend itself to the "proposition" formulation. I suppose one could force it into that mold but the fit would be poor, IMO:

"It had never occurred to me that when I took mushrooms to commune with the plants in my garden, I actually might have been doing a kind of science."

• Pollan, Michael, "A World Appears", p. 42.

There is also the matter of scale. If you make a map for every sentence or even paragraph, you may never get finished with even one work, and if you did you would have a large number of fragments that haven't been combined to give a coherent picture. It could be better to make most of the concept maps about the larger-scale ideas in the work.

Some people do better with text, others with pictures and diagrams. It sounds like you tend to the second, and if that's right then by all means emphasize diagrams, including concept maps.

There is a middle ground. Here is your diagram turned into text but not in the form of linear sentences:

Point
    topic sentence.
    what will be proved?
    main claim

Argument
    is a -> reasoning process
    explains
        how evidence proves
        why true

Evidence
    is
        facts
        raw data
    explains
        concrete proof

You could create a concept map, put into something like my example into a note, and also put a link to the map's graphic into the note. Over time, you could rework and flesh out the note as you come to understand it better.

Edmund

@c4lvorias said:
@tomp @Andy To clarify what I mean, I am basically asking whether linear text provides something more than a proposition. Because with propositions, imho, you can represent fully what a linear text basically says.

My goal is to take literature notes in the form of concept maps.

I use 'Concept Maps' as 'Structure Notes' to illustrate the context of my 'Permanent Notes'. These provide context for my observations, questions and propositions. A 'Concept Map' is linked to one or many 'Permanent Notes'.

My 'Literature Notes' are without any 'Concept Maps'. These are NOT my own words, NOT my own ideas, AND NOT my own 'Concept Maps':

c4lvorias

@Edmund I am familiar with your work. I don't know whether you mentioned elsewhere, what software exactly are you using?

Edmund

@c4lvorias said:
@Edmund I am familiar with your work. I don't know whether you mentioned elsewhere, what software exactly are you using?

I’m using Obsidian together with SimpleMind Pro.

zettelsan

@Edmund said:

I’m using Obsidian together with SimpleMind Pro.

Are you using SimpleMind Pro for all your mind maps and diagrams? I was also wondering what you are using to create them.

Edmund

@zettelsan said:

@Edmund said:

I’m using Obsidian together with SimpleMind Pro.

Are you using SimpleMind Pro for all your mind maps and diagrams? I was also wondering what you are using to create them.

Yes, I do. 100%.

harr

@c4lvorias said:
My goal is to take literature notes in the form of concept maps. Here's a very basic example:

I'm curious what you mean by concept map. Are you following a manual?

Are you familiar with the ihmc Cmap site? https://cmap.ihmc.us/docs/learn.php

As far as I can tell, linking words are important in Novak's method. https://cmap.ihmc.us/docs/linkingwords.php

In your example I noticed the linking word "next". What kind of relationship do you want to express with that word? I see a main benefit of such maps in asking that question. Novak's concept maps encourage you to make the relationship between two concepts explicit.

How are point, argument and evidence connected? You could look at the text on the right side of your screenshot. Evidence "supports" the argument, the argument "explains" the point.

BTW, The example looks like a variation of the claim-evidence-reasoning pattern: https://www.wikihow.com/Write-a-CER

c4lvorias

@harr

I'm a newbie to concept maps, I wanted to evaluate its potential first, and they reflect my poor thinking patterns and vocabulary of action verbs. So "next" is BS in terms of theory. What I meant by "next" is the supporting nature of each other consecutively. But you are right, there should be no ambiguity.

I read "Learning, Creating, and Using Knowledge" and "Learning to Learn" by Novak in this 3-day period. The only step-by-step organizational hint was to establish a hierarchy between concepts. Novak heavily cites Ausubel's Assimilation Theory which is a more principle-based approach. Ausubel mentions progressive differentiation (of concepts through time) and subsumption of concepts (a concept dissolving into some other concept after a change in understanding) and integrative reconciliation (linking with the prior knowledge, basically Zettelkasten). However, I've not read it yet.

The reason I started to wrap my mind around concept maps as the main medium is the non-linear nature of my thinking. As I am writing this comment it doesn't progress in a linear fashion. I write a sentence, edit one, change their place etc. So I really believed in concept maps.

I experimented with Tinderbox. In Tinderbox you can drag-and-drop a note onto another note so the second one becomes a map. What a feature! Everything is suddenly a structure note. Plus concept maps. Extra structured structure notes. (Wanted to share my excitement)

Surprisingly, no one in the Tinderbox community uses the Zettelkasten Method. They go directly into synthesis.

PAE is very simple, I took the text from Google's AI summary. It just helps me thinking.

harr

@c4lvorias said:
I read "Learning, Creating, and Using Knowledge" and "Learning to Learn" by Novak in this 3-day period. The only step-by-step organizational hint was to establish a hierarchy between concepts.

I find chapter 2 of Learning how to learn useful. It defines concept maps (emphasis added):

Concept maps are intended to represent meaningful relationships between concepts in the form of propositions. Propositions are two or more concept labels linked by words in a semantic unit. (…) A concept map is a schematic device for representing a set of concept meanings embedded in a framework of propositions.

Later:

Concept mapping is a technique for externalizing concepts and propositions.

I think concept mapping works best, when we use the terms concepts and propositions in the same way as Novak does.

The tables list exercises for students that teach concept words and linking words. For example:

Ask the students to describe what they think of when they hear the word car, dog, etc. Help them recognize that even though we use the same words, each of us may think of something a little different. These mental images we have for words are our concepts; (…)

And:

Now list words such as are, where, the, is, then, with. Ask students what comes to their minds when they hear each of these words. These are not concept words; we call them linking words and we use them in speaking and writing. Linking words are used together with concept words to construct sentences that have meaning.

It's not the step-by-step manual I would hope for, but close enough to get the general idea.

@c4lvorias said:
The reason I started to wrap my mind around concept maps as the main medium is the non-linear nature of my thinking. As I am writing this comment it doesn't progress in a linear fashion. I write a sentence, edit one, change their place etc. So I really believed in concept maps.

I find them useful in addition to linear text, because they force me to think about the linking words. Novak writes in the same chapter:

We also want to emphasize that concept maps become increasingly more useful as students become more proficient at labeling the lines. (…) We now consider careful attention to the words selected to link concepts an essential aspect of instruction in concept mapping.

I find it fun to draw concept maps (and mind maps and other visualizations), but I find them tedious to read.

I experimented with Tinderbox. In Tinderbox you can drag-and-drop a note onto another note so the second one becomes a map. What a feature! Everything is suddenly a structure note. Plus concept maps. Extra structured structure notes. (Wanted to share my excitement)

:-)

@c4lvorias said:
Surprisingly, no one in the Tinderbox community uses the Zettelkasten Method. They go directly into synthesis.

Maybe they perceive Zettelkasten as too narrow and rigid?

@c4lvorias said:
PAE is very simple, I took the text from Google's AI summary. It just helps me thinking.

Does anyone else use PAE or did the AI hallucinate it? When I tried to google PAE, the results indicated that CER is a much more common acronym. I must have overlooked something. ;-)

Edmund

@Edmund said:

@c4lvorias said:
My goal is to take literature notes in the form of concept maps.

I use 'Concept Maps' as 'Structure Notes' to illustrate the context of my 'Permanent Notes'. These provide context for my observations, questions and propositions. A 'Concept Map' is linked to one or many 'Permanent Notes'.
My 'Literature Notes' are without any 'Concept Maps'. These are NOT my own words, NOT my own ideas, AND NOT my own 'Concept Maps':

Here are my four use cases for concept maps:

c4lvorias

I love that it's very action oriented, @Edmund. This looks like a refined output of a thinking process I can imagine.

I see that you mention some templates. Are they actual links? Where do you store these maps, do you export as .png to Obsidian?

And I am super interested in those templates. After my hype in concept mapping, I noticed that I am thinking better with prompts/templates. I am trying to wrap my head around creating ones. I don't understand the rationale of Sascha's flower, he points to the value of knowledge SEP page but text is too advanced for me. And idea compass is not enough in my opinion to explore the depth of an idea. Your approach, as far as I see is more compatible with the way I think, especially since you have a proposition template. More structural.

c4lvorias

This is the latest attempt of mine. It's not perfect, and I cheated with creating circles.

Besides its technicality, main takeaway for me is dyslexic people have problem with top-down modulation of attention especially to visual stimuli.

Concept maps in that sense is a great externalization. You have to separate concepts, link them with linking words and constantly think if something fits to the existing relationships, naturally forcing your attention. As Edmund pointed out, it helps you to create a structure out of the material. I see significant benefits especially in the degree to which I engage with deep processing. It was hard for me to delve into structural relationships of sentences before trying concept maps.

harr

@c4lvorias said:
This is the latest attempt of mine. It's not perfect, and I cheated with creating circles.

Nice!

Your example inspired a few experiments of my own. I learned, again, that I don't like drawing concept maps by hand. (I've tried this before. I have a love-hate-relationship with mapping tools. They help understand things, but I find them unwieldy.) Draw.io is really fast, but it still feels slow.

I prefer to type text, so I also tried Mermaid flowcharts. There was no installation necessary, because Obsidian ships with Mermaid. Here's an example:

This is the Mermaid code:

---
title: field trip to a paper mill
---
flowchart TD
wood --contains--> fibers --hold-together--> paper
wood --comes-in-the-form-of--> logs --are-cut-from--> trees --grow-in--> forest
logs --are-ground-in--> mill
water --is-consumed-by--> mill
mill --produces-soft--> pulp --dries-as--> paper
wasps --build--> wasp_nest --is-made-of--> paper
wasps -.chewing-resembles-.-> mill
paper --is-used-for--> writing
books --are-printed-on --> paper

The example is based on figure 6 of Novak's tutorial. I followed the instructions and started with a list of concepts.

Mermaid has two limitations. First, it lays out the chart as it likes. There's not much opportunity for fine-tuning. Second, an edge can connect only two nodes. If you want to link three or more concepts with a linking word, you'd have to create a node for that linking word. It's possible, but it makes Mermaid's source harder to read. And it feels a bit like cheating.

What worked? Novak's recommendation to start with a list of concepts and then add cross-links worked surprisingly well. In Mermaid I first listed the concepts. Then I combined them into short chains of propositions. I checked with the rendered chart if all concepts were linked and if the wording and connections made sense. Some concepts worked better as linking words. The iterative process really made me think about the concepts in this example.

I like Mermaid because it sits at the intersection of visual map and linear text.

Answering the initial question, I'd say that concept maps capture less detail. I wouldn't use them for extracting detailed knowledge. I find them useful for understanding the relationship between concepts. But when precisions matters, text and tables are my first choice.

GeoEng51

@harr said:

@c4lvorias said:
This is the latest attempt of mine. It's not perfect, and I cheated with creating circles.

Nice!

Your example inspired a few experiments of my own. I learned, again, that I don't like drawing concept maps by hand. (I've tried this before. I have a love-hate-relationship with mapping tools. They help understand things, but I find them unwieldy.) Draw.io is really fast, but it still feels slow.

I've used "mind maps" for many years in my work, mostly as a means of describing all the elements of a particular engineering system and how they are related to one another. For example, I recently completed a mind map for all the elements and steps involved in designing and constructing a dam. This drew on 50 years of experience and accumulated knowledge on the subject, and took me about a week of creation and fine tuning. It had 15 major branches and thousands of individual components.

I find mind mapping (or concept mapping) is an effective way of showing relationships in a system and can, when necessary, also be constructed like a flow chart, showing sequences, and cause and effect.

They are useful in sorting out in our own minds how components, events, tasks, etc. are connected. They are eminently suited when used in this way with a Zettelkasten.

However, by itself, a mind map does not tell a story. For that, you need a narrative. Any good writer can tell you that, as can anyone who reads good fiction or non-fiction, although I find it difficult to describe the essential differences between the two.