Recently, I introduced the concept of ultralearning—deep, intense self-education. This kind of learning is characterized by grappling with deep concepts and hard practice.
My bigger learning projects have used this approach out of necessity. If you’re trying to learn something like differential equations or Chinese characters in a short period of time, those constraints make other strategies unworkable. Necessity is the mother of invention, so I believe highly time-constrained projects might be a good place to start looking for effective learning methods.
However, I don’t believe ultralearning is necessarily about doing large, all-consuming, full-time projects. The principles work, even if you’re only spending a couple hours per week. In this article, I’d like to talk about a different kind of MIT Challenge I’ve been pursuing privately in the last year.
Mastering Cognitive Science
Cognitive science is an interesting emerging field. Cross-disciplinary, most universities don’t have a cognitive science department. Instead, you find ideas about the science of how we think scattered in philosophy, linguistics, computer science, neurobiology and psychology.
This field directly intersects my writing, and I got a taste for the ideas as some of the MIT Challenge brushed against the same topics. After I returned from my language learning project, I got interested in the idea of doing something similar for cognitive science the way I did the MIT Challenge for computer science.
However, I didn’t want to pursue the project full-time. Big, full-time projects are fun. But they’re also hard to juggle with running a business. Documenting the projects is also a non-trivial amount of effort, and especially after recording efforts of my last project, I wanted to learn without needing to make a spectacle.
Quite unlike my two previous big projects, this one had no time constraints, no public blogging and, hopefully will take place over several years, instead of several months.
Despite these differences, the learning approach is quite similar to the MIT Challenge. Going deep first, seeking hard practice and strategically using materials are still the main focus. I also think this provides a more accessible example of ultralearning than some of my previous projects.
Step One: Finding a Curriculum
My first approach to trying to efficiently learn a broad topic was to get a map of the territory. This is one of the tricky parts with learning broad domains of knowledge or skill. It can often be overwhelming where to start.
One approach is just to learn anything and go from there. And while this approach has merits, I’ve found a big drawback is that it tends to push you away from ultralearning. Ultralearning is hard, so by just picking things randomly, you tend to navigate away from harder topics and sit on easier ideas.
Just picking popular books on cognitive science would likely do this approach. I may learn some ideas, watered down in popular science books. But I probably won’t be getting at understanding the hard, university-level concepts that I desire.
The way I got around this with the MIT Challenge was choosing MIT’s undergrad curriculum as a road map. That allowed me to take the vast field of computer science, and break it down into something I could reasonably accomplish, without sacrificing too much depth or breadth.
For this project, I found UC San Diego’s cognitive science reading list. This is a list of 40+ textbooks they recommend as an introduction to their new PhD students. Although I may not stick to this list as rigidly as with the MIT Challenge, it provides a good starting point for covering the field.
Step Two: Exploring and Deepening
After locating the books, I purchased around a dozen of them to start off. Reading, on it’s own, isn’t a terribly effective learning strategy. Good ultralearning should eventually consist of some kind of practice.
However, it wasn’t possible to dive right into problem sets. So the best I can do is to read the books and then look for opportunities to selectively deepen that base of knowledge. I just read the books normally—no speed reading or crazy highlighting system. The goal here is to get a good gist of the ideas in the books, but probably not to know them to the level of rigor that I would need to pass a hard class.
After reading several of the books, I could gauge which topics deserved more depth. The first textbook on neurobiology, for instance, was difficult because I lacked a lot of the background molecular biology and chemistry to make sense of it. This led me to seek out a related course, Duke’s Medical Neuroscience on Coursera, which could bring me up to speed before going deeper.
Other books I wouldn’t read all the way through. George Lakoff’s Women, Fire and Dangerous Things is a long book to suggest a single idea about cognition. After accepting and understanding the gist of the idea, it didn’t feel necessary to read a few hundred more pages of examples and evidence.
Step Three: Hard Practice
For now, most of my effort has still been set on reading the main books and ideas. However, as I get a better sense of the landscape of this emerging field, I’ll shift more of my time into using my knowledge instead of just consuming it. That might mean doing problem sets, toy projects or writing exams.
So far, the only facet of this project that has reached this stage has been taking Medical Neuroscience. Fortunately, they have difficult exam questions that aren’t too watered down for the MOOC audience.
Taken individually, a single source of practice has biases. Learning a language and testing yourself only via vocabulary memorized, conversational success or even a proficiency test, can all be misleading. But added together, it becomes much harder to fool yourself into believing you’ve learned something you haven’t.
Continuing to Evolve the Project
I hesitated to write about this project at first. Part of that is that publicly stating things makes it harder to make changes later. Since this project is still at a fairly fluid stage, without a clear end point or stable curriculum, it means that exactly how far and where I go with it are still undecided.
However, I think there is also value in showing these less formal examples of ultralearning as well. It shows that the ideas of learning hard things deeply is something anyone can do in their spare time. In future articles, I hope to show that these ideas don’t just apply to lofty pursuits of knowledge, but practical learning that can advance your career and passions.