Why It’s So Hard to “Teach Yourself” on the Internet: The Problem of Structure

Formal education offers better sequential organization.

Almost everything taught in college courses can be learned on the Internet for free. But Internet users largely aren’t using it to learn.

Even the most motivated of us are often just left with scattered bits of knowledge that tickled our pre-existing curiosity. Rarely does our “learning” look anything like the rich, well-connected web of knowledge we’d gain from a series of courses in a subject.

(Honestly, who do you trust more to understand physics: someone with a documented A in an undergraduate-level quantum physics course after getting an A in the required intro course, or someone who claims to be “self-taught” in physics topics from online videos?)

It’s not that we can’t reach that level without formal courses. It’s just that in formal courses, you have a clearer roadmap for what to learn — in other words, more appropriate structure.

Fields of knowledge aren’t just bundles of facts. In school, subjects have a level of sequential organization, both within courses and across courses: You need to learn Concepts A and B to understand C, which you’ll eventually need to know in conjunction with X and Y to understand Z.

Internet searches don’t give you that structure. If you want to learn Concept Z, you’ll get sources that just explain Z — either in a simplified way that only draws upon layman-level knowledge, leaving you with only a surface-level grasp, or in a technical way, as in the relevant college course, assuming you know all the other concepts before it.

For example, say you want to learn about neuroplasticity, so you’re reading about “long-term potentiation.” Popular sources will tell you that with repetition, synapses get “stronger” to aid communication. True, but is that all that’s worth knowing?

Here’s the explanation I received in Cellular Neuroscience: when NMDA receptors are persistently activated, Ca²+ concentration rises, activating CaMKII. This phosphorylates AMPA receptors to increase their conductance and helps add more AMPA receptors to the membrane — strengthening the synapse in the short term. The activated protein kinases then upregulate transcription of genes to produce more dendritic spines at this synapse—a slower, longer-lasting change, strengthening the synapse in the long term.

The paragraph above makes sense if you’re familiar with the following concepts: NMDA and AMPA receptors (glutamate-gated protein channels on the membrane of a dendrite that let in positive ions), phosphorylation (addition of a phosphate group to change a protein’s conformation), ion channel conductance, protein kinases, gene transcription, dendrites, and communication across synapses. And if you’d taken prior biology and neuroscience courses, you would be familiar with them, all important concepts with a range of applications.

This isn’t written to be deliberately dense. To a neuroscience student, it reads smoothly and concisely, a fitting answer for a homework question. (In fact, the Wikipedia article’s explanation of these same processes is even more jargon-packed.) A reader equipped only with high school level biology, on the other hand, might be swamped!

Clearly, how you interpret what you read depends on the contents of your long-term memory. While working memory is a tight resource, long-term memory is practically limitless. So just “looking it up as you go” isn’t the best strategy. If you keep stumbling upon words that aren’t in your long-term memory, and you have to interrupt your reading to look them up, you’ll be juggling multiple new concepts freshly crammed in your working memory.

Those who already deeply understand those terms are at an advantage. They simply make sense of what they read when they read it, so they have more mental processing space for the new takeaways.

(What’s even harder to correct is misinterpretation — when novice readers see a word and think they know what it’s referring to, they don’t even recognize the need to look anything up, because they don’t notice the gap in their understanding. For example, if novices read about dopamine, they may connect it to the pop-psych caricature of “dopamine” and come to inappropriate conclusions. This would not be an issue if they’d been introduced to “dopamine” afresh and were explicitly taught the different roles it plays in brain systems.)

Experienced professors understand the big picture of their discipline. They know what “higher” concepts are ultimately unlocked by knowing A, and what “lower” concepts you need to unlock Z. So academic content is structured accordingly. By the time you get to Z, nobody’s going to slow down to explain A.

Novices can’t see that far. To them, Concepts A and Z just look like independent pieces of information, so why does it matter when you learn which? They often search for content haphazardly, in books, articles, and videos that catch their attention. When they can’t see the deep structure of a discipline, their idea of what they should learn is guided by surface features.

The novice who wants to learn about the brain, for example, will be drawn to anything specifically dedicated to the “brain.” This strategy overlooks important sources about heredity, gene regulation, evolution, chemical polarity, protein structure, cellular metabolism, and more that she surely would’ve learned as a student of university-level neuroscience. She can access online neuroscience sources that refer to these, but the neuroscience student will interpret them better, with fewer gaps and misconceptions, and thus learn better from them.

With this in mind, you may not yet know the roadmap of what you want to learn. But strategies like these can help you find it:

Strategies for Structured Learning

• Consult the syllabus for an introductory course in the subject. Seeing what concepts are covered in what order gives you a good place to start.
• Read from a textbook. Here’s an impressively comprehensive list of recommendations for many subjects.
• Familiarize yourself with the different frameworks in the field. “Neuroscience,” for example, includes studies at the molecular, cellular, functional, clinical, and evolutionary levels. Determining what you should know for each subdomain is a more manageable way to break down what you need to know for the domain as a whole.
• Avoid disorganized methods, like opening all the articles alphabetically in a Wikipedia category. Instead, move to new concepts systematically, building upon previous knowledge, such as by clicking on relevant links within the articles you read, and clicking relevant links in them, and so on.
• When you stumble upon an article with lots of unfamiliar terms, take the time to study each of them and then return to it afresh.

If simply putting information on the Internet would revolutionize academic learning, it would’ve done so by now.

Why it hasn’t is no mystery. Without the push of getting graded and progressing toward graduation day, most people have little patience for the slow, ordered accumulation of knowledge they’d need to master a field. Curiosity-driven searches may give us the satisfaction of “learning,” but we can’t even judge how insufficient our mental maps are without a good idea of what we should be learning. That’s a tricky thing for novices to assess on their own, and it’s not fair to expect them to.

After all, the difference between a novice and expert in a discipline isn’t just that the expert “knows more.” The two can put the same article up on their screen, reading the same words, but how well they each understand it depends on what’s already in their long-term memory. The expert can learn more in this discipline, better able to integrate new material.

So learning doesn’t work the way schooling critics seem to portray it, as mere delivery of information. There’s not much value in knowing you have equal “access” to information if you can’t understand it.

With how popular it is to discuss what’s wrong with schools, few have pointed out what’s right with formal instruction. Self-directed Internet exploration has a problem with structure — once we recognize this, we can figure out strategies to overcome it!