While I’m not a teacher by trade, my website is all about software curriculum and the spread of knowledge. As a result, I figured I’d take the opportunity to touch on a topic that plagues teachers and instructors across the world: the curse of knowledge.
Table of Contents
What Is the Curse of Knowledge?
According to our old friend Wikipedia, the curse of knowledge is a cognitive bias that occurs when an instructor assumes that their students have the proper background to understand the topic at hand. In other words, the instructor is unable to put themselves in the position of the student. Of course, this topic is quite meta because I’m playing the role of the instructor here (though I wouldn’t consider myself an expert on the subject). Regardless, you get the idea.
How Do We Overcome the Curse of Knowledge?
As a society, we trust experts to teach us because they have the knowledge we wish to obtain. However, the curse of knowledge puts experts in a position where they can’t properly give us an education. It would seem then that we’re stuck. How can we possibly learn if we can’t overcome the curse of knowledge?
Fortunately, we know there has to be a way to overcome the curse of knowledge. If not, there would be a bit of a paradox that would arise. How would we get experts in the first place? Would everyone have to discover everything on their own? Obviously that’s not how our world operates, so there must be something else at play.
Let’s Think Like an AI
Hopefully by now you’re used to the little Computer Science interjections. I felt like artificial intelligence would be an appropriate topic to bring up. After all, I am struggling to write Computer Science lessons that everyone can understand. Again, it’s a bit meta, but I think you’ll appreciate the point I’m trying to make.
How do machines learn?
I think this is an important topic because it brings in some fundamental characteristics about learning. Of course, I’m also no expert in cognitive science, but I like to think that the two fields are related. Feel free to call me out on it in the comments.
Machines have several ways of learning, but I think the one that applies most to us right now is inverse reinforcement learning. Now that’s a pretty large term, but basically what it’s describing is exactly what we’ve talked about so far – learning by observing an expert. If you want a better explanation of this type of learning, I’ve embedded the YouTube video that actually inspired this article below:
One of the challenges of inverse reinforcement learning is trust in the expert. By default, the machine is going to offer 100% trust to the expert which is sort of how I imagine students trust their instructors. The problem is that the machine will only learn based on what it’s showed, and it has no way of determining which parts of the demonstration are significant. As a result, a machine may incorrectly interpret a repeated segment of a demonstration as the correct way to complete the process. In reality, the expert is likely just repeating a more difficult segment for clarity.
Applying AI Principles to Human Learning
If we put ourselves in the position of the machine, then it becomes clear what our issue is – there is no dialogue. We have no mechanism of talking out our confusion with the expert. We simply soak in what we experience and attempt to reproduce the process. What we need to be doing is asking questions. That way we can address that lapse in education due to the curse of knowledge.
Don’t believe me? Imagine this. You take a class on the introduction to basic math, and you’ve never learned a single math concept. The instructor decides to teach you multiplication as they assume you already have a background in addition. After a very confusing lecture, the instructor starts listing off the following examples:
5 x 7 = 35 9 x 1 = 9 3 x 3 = 9 4 x 6 = 24
As a bright young individual, you try to observe some sort of pattern in this chaos. After some frustration, you decide you’ll just go back home and study all the combinations up to 9.
Fast forward to next week. You have your first quiz. Naturally, you ace it because you have a nice set of flashcards that taught you all your multiples. At this point, you start to feel good because you think you really understand the subject.
That was until the lesson after the quiz: multiplication of terms greater than 9. Now you realize that there is definitely some background information you’re missing. This whole long multiplication nonsense has you baffled.
That’s when you pop the question – “What do you mean when you say you add the result?” The entire class sighs in relief as if this simple question had relieved all their stress. It’s at that moment the curse of knowledge is lifted, and the education system meets equilibrium.
Hopefully by now it’s clear how we beat the curse of knowledge. As a student, we need to be asking questions. We can’t expect all the information presented to provide us with 100% comprehension. Likewise, as experts, we need to promote an environment in which dialogues are possible. If we do, we’ll ensure that we are providing complete coverage of the subject area.
That’s why I try to offer as many opportunities for you to ask questions and make suggestions on this website. I want you to be able to lift the curse of knowledge, so we can start learning together.
I figure now is as good a time as any. Fire off your suggestions or questions in the comments below. I’ll get back to you as soon as I can! Better yet, head over to your favorite article and leave your comments. If you’re a first time reader, here are a few of my favorites:
Let’s help each other out by making a commitment to lifting the curse of knowledge.
Recently, I was giving a lecture about Java's "common" methods (i.e., all of the methods of Object), and I had epiphany about how Java only has toString() while Python has str() and repr(). So, it...
Magic numbers are numerical constants that have no clear meaning in the code and therefore make code harder to read. Anything that makes code harder to read is something we can use to obfuscate our...