Synthesizing Bonsai

In recent years, there has been a cacophony of calls for synthesis and cross-domain know-how to mend what ails those in knowledge-silos. The thing is, the internet and most platforms do a really good job of helping you find information if you already know what you’re looking for. But what if you don’t?

My own humble note-taking perusals have blossomed into jekyll-bonsai from a full-blown garden of forking paths in the cracks between many silos. I thought I’d share some of the forks so others can wander the wild with more direction (read as: keywords galore!) in the search for prized tips’n’tricks that might help refine their knowledge-handling workflows.

An Unexpected Note-Party 🐝

Everything started when I wanted to do some self-study. I immediately knew I needed to take notes and I was interested in answering questions like, “How do I track my thoughts?”, “What structures lend themselves to effective review practice?”, “How can I seamlessly tie into other functions on my computer, like quick imports from a web browser?” among others.

These curiosities got me started down the road of personal wikis, even though I didn’t know that term until much later. Personal wikis are for note-taking. On the whole they resemble wikipedia and generally provide some form of wikilinking with [[double-square-brackets]] that allow you to quickly link to documents or pages internal to the current context (directory of notes, a website…). I came to see the absolute necessity of wikilinks and continued searching for the optimal workflow – in all these years since I was in school someone must have come up with systematic, efficient flows for learning and knowledge work, right?

Wikilinks apparently sport many identities: “Wikilink”, “bidirectional link”, “wiki text linking”, “internal link”, “humble double bracket internal link”…These terms are all referring to the same syntax, with perhaps tiny differences in usage details. This clash in terminologies landed me in the world of zettelkasten since those folks were trying all kinds of apps with quick’n’smooth linking mechanisms to digitize their respective zettelkastens. Without getting too into the details, zettelkasten is a method of note-taking that breaks notes into atomic bits and links them via random ids (as opposed to keywords in personal wikis) since it was originally developed with pen’n’paper.

But zettelkasten started somewhere in the mid to late 20th century and personal wikis first appeared on the scene in the mid-90’s. The amount of information people accrue in their day-to-day lives has exploded since both of those times. So while their capacity to store information is quite useful, their ability to import is sometimes lacking and you can see people online stitching their workflows together from multiple applications to fill this need.

This is where personal knowledge management (PKM), networked note-taking1, and memexes came in. Apps that sport this description often take some of the core strengths of personal wikis and zettelkasten and even start reaching for futuristic capabilities like assisted Bayesian reasoning2 and adding pretty [[graph]]s to visualize your system. But I found these workflows fell short. Others have pointed out how convoluted your thinking and writing can become after living in networked note-taking apps for long enough and I [[agree whole-heartedly]] with this.3

It’s become evident to me that the workflows I am searching for just don’t really exist yet. So my paths have forked far and wide in search [[first-principle]]-style tidbits to improve my learning and knowledge work workflows.

Out of The Frying Pan and Into The Fire 🐝

The first fork off of note-taking paradigms had to do with learning, education, and pedagogical practices. I wanted a system that would not only store information, but to help me learn material better and improve the functioning of my own mind.

A very popular, but not often well-integrated feature is spaced repetition. There are lots of flashcard apps that supply this review style for you, but finding a solution that integrates seamlessly within a larger context (the rest of your notes) is actually quite difficult. Most existing solutions are a hodgepodge of scripts and open source solutions that may or may not be a great fit for any given person. I think having access to this feature is important, but I also think it should be [[approached with caution]].

Continuing down the path of memory-related skills, there has also been an interest in “[[memory palaces]]” (or “memory castles” in older terminology). This skill is what memory champions hone in order to memorize vast amounts of information in miniscule amounts of time. I personally view this as a fun parlor trick and suspect it won’t improve your understanding of complex topics. Connections themselves are an important feature of understanding anything and memory palaces totally rely on superfluous connections between concepts and physical objects. But I’m still [[mulling it over]]. Considering how effective the practice is, it is probably wise not to shrug it off completely.4

Considering the wild advances in technology, I thought I would end up searching for solutions and features that would completely change how I learn, but probably the most important factor in learning for me has still been simply to write. Write full reviews of books. Verbalize vague gut feelings. Write full letters to friends and family. Do this regularly. You will think much more clearly if you do. So I suspect whatever system I am searching for would need to support simple, traditional writing well and provide some kind of integration with linking that made sense.

All that being said, I’m still pondering over what [[characteristics]] would be fundamental for a note-taking system.

Inside Information 🐝

At some point I stumbled across a blog that mentioned “information architecture” in passing and this one word opened an entire world to me. From there, I was exposed to concepts like “semantic relationships”, search and navigation dynamics, and general principles of information system design.

If it hasn’t become evident yet, I’m pretty deep down this rabbit hole. 🕳

The questions I was asking were now about information inherently: How do you store it? What makes it searchable? What about easily navigable? How do you blend people-focused “API”s with information-focused ones? And then it occurred to me: “You know who’s been thinking about this problem a long time? Librarians” And for some reason the term “library science” was sitting in a dusty corner of my cortex hippocampus and I was off to the rabbit-hole races.

Library science is definitely a thing and that field has kind of turned into “information science” which, ironically enough, is a sort of parent to “information architecture”.

Anyway, the parent term “information science” opened up doorways to “knowledge organization”, which I guess information architecture technically is, but the word usage implies a different sort of use-case. When I think of “information architecture” I imagine a sort of information store like yellow pages, libraries, imdb, goodreads, wikipedia…Stores that are optimized with the information in mind and especially for a certain type of information. “Knowledge organization” on the other hand, focuses more on information stored in the head5 or geared toward some specific goal or project in mind. How is all the information that a person or organization knows, from all of those previously mentioned repositories, managed in one place? In other words, what are the first principles with regard to structuring the things I (need to) know?

The Gathering of the Bots 🐝

This question led me to stumble across the term “knowledge representation”, which also packs a slightly different meaning from either “knowledge organization” or “information architecture”. This phrase is used in the field of artificial intelligence. It refers to how to structure and storew data such that an agent may use that information logically and intelligently.

I found it particularly intriguing that the podcast that pointed me in this direction specifically mentioned that AI researchers haven’t put much time and energy into this sub-field of AI (Though as I go to publish this post it seems there have been extremely recent advances on this front).

Anyway, I do wonder how much of the new-fangled note-taking applications mentioned previously are thinking in terms of AI rather than people for the structure of these notes.

There and Back Again 🐝

All this might feel like needless navel-gazing at similar terms in multiple fields, but there is a method to the minutae-madness: By viewing the same problem through the lenses of varyious contexts (learning, information, libraries, ai, etc.) different aspects of the problem and solutions are revealed to us. This is precisely the kind of wandering one must operate when crossing the boundaries of knowledge silos and searching for novel solutions.

A Thief In The Night 🐝

About those creator and maintainer intentions…I think any software that claims to house your (effectively) “second brain” ought to be approached with extreme care.

I would prefer to use a system that doesn’t peer into any of my data whether manually or with any kind of automation. And data should not be sent over the internet anywhere for any reason (unless you are using services such as sync of your own volition).

I would also much prefer to use a piece of software whose entire codebase is open source. I view it as a human right to be able to see into the innerworkings of the “machinery” that will store (and potentially process) all of my thoughts, feelings, sensitive and personally identifiable information.

All that being said, there are the harsh realities of market competition that must be addressed, especially if opening up one’s codebase. In the end, it will be up to the communities that are consuming these products to choose software that adheres to humane values…And legislation, whenever that happens.

  1. “Networked note-taking”, “networked thought” are terms that might help get your feet wet in this space, but I wouldn’t linger on them for too long for research purposes. 

  2. This sounds freaking cool. I also think that’s a v2.0 feature. v1.0 might look something more like “verify internal consistency and freedom from contradiction”. 

  3. I think trees are likely key component in resolving the convoluted thought problem. 

  4. See [[hippocampus]]

  5. 🎩 Hat tip to [[design-of-everyday-things]]