I think of teaching, in the general sense, as simply the process of explaining things to others, a form of communication and knowledge transfer. It seems that teaching should be a straight forward information transmission task, but I think it is actually one of the hardest things to do well.

The first thing to realize is that a lot of the knowledge in our brain is very unstructured. The different pieces of information that we know about a certain topic are “just there”, vaguely linked, without much inherent organization. Of course, some knowledge feel more structured in our minds than others. These are usually things that we learned in a more ordered manner (e.g. in school), and the structure we perceive is simply the way we learned it ourselves. Most knowledge though, is not acquired this way, but rather acquired through various life experiences, and as a result is very disorganized in our mind.

This presents a real challenge when we are trying to teach or explain something, because the process of articulation (which takes place over time) is a sequential process. To teach something, we have to bring an explicit ordering to the unstructured pieces of the knowledge in our head. It’s very much like formulating the knowledge into a story¹, where what comes first must prepare the audience for what is yet to come. Very often I find that the act of trying to explain something to others helps me understand it better, and I think this forced ordering process is the main reason for that.

Then there’s the question of what pieces of information to include in the teaching process. I feel that this aspect of teaching is very similar to the process of lossy compression. The teacher often knows a lot more details than what is necessary to understand something. Good teachers are like super efficient lossy encodings, and are able to convey a large amount of information (almost the same quality as the original in their heads) extremely concisely. The key here I think is figuring out what information is (relatively) superficial to the topic and hence can be discarded. A lot of domain expertise, experience, and critical thinking is required to do this well. MP3, the popular lossy audio encoding standard, does this very well. In variable bit rate encodings of MP3, the quieter, simpler (i.e. information-sparse) sections of songs are encoded using much lower bit rates than the louder, more complex parts. The encoded music still sounds good to our ears but the file size is a lot smaller than the original.

In my opinion, this process of selectively discarding information when teaching is what separates a lot of good teachers from great ones. I’ve always admired the ability of Richard Feynman to explain very complex subjects in a very clear and concise manner. Besides his natural charisma, I think it has a lot to do with his ability to identify the critical pieces of information and ignore the rest.

To teach well, it’s also crucial to know your audience. All the techniques involved in teaching must adapt based on the audience, specifically the size and shape of the knowledge gap between the teacher and learner. For example, what information can be glossed over and what needs to be explained in greater detail can vary greatly depending on the learner’s existing knowledge on the topic. Even the choice of words (e.g. what level of jargon to use) must be considered carefully. It sounds very obvious, but very often I see people forgetting this and just assume that the audience have the same background knowledge as themselves². This audience-based variability makes explaining things on the spot (i.e. without preparation) very difficult, and I have great respect for the people who can do it well.

The ability to teach and explain effectively is one of the most important skills to have in life, as so much of our world relies on knowledge transfers³. But because of how difficult it is, not many people can do it well. I personally spend a lot of my spare time looking for better ways to explain certain things to others, as a way to improve. Here are some of the things that I’ve learned over time that haven’t been mentioned above.

• When explaining a high level concept, don’t get bogged down by the specifics. Think of it this way: a compressed, lower resolution version of the whole image is almost always better than randomly selected, detailed partial bits of the image⁴
• Analogies can be extremely helpful in teaching something effectively, since they are based on a common understanding (it’s like using a bridge elsewhere to cross the current knowledge gap). However they are difficult to get right, because they require expertise in at least two different topics, one of which must be shared between the teacher and learner
• Whenever possible, trace back to when you first understood something (i.e. the eureka moment), and think about what was the key to your own understanding. Then simply try to invoke the same feeling when teaching it to others