Why Experts Make Bad Teachers | Hacker News Reader

Why Experts Make Bad Teachers

medium.com

98 pointsbaptou1266 comments

Saying that all experts are bad teachers is like saying all <insert ethnic group> are <insert stereotype>.

Good teachers are good teachers. If the good teacher is an expert, then they are better than the good teacher who is not an expert, especially if they are trying to teach a nuanced topic.

Does every teacher need to be an expert? Of course not. But plenty of non-expert teachers will actually lead students astray by teaching them the wrong model.

I've had plenty of "bad teachers" who were experts. I've also had plenty of "bad teachers" who were not experts. I've had very few good teachers, period. Bad teachers in lower-level topics are not noticed very much, but they are the bane of undergraduates who are subjected to them often. Survivorship bias might lead me to conclude that most experts are bad teachers, but the better explanation is simply that teaching and being an expert in field X are distinct abilities/skills, and that it is rare to find people who have mastered both.

I aim to be an expert in my fields (currently 3 disparate ones in which I hold degrees). I also aim to be a good teacher, which means that I take time to figure out the best approach and experiences for teaching a particular topic. I consider my expert status to be an advantage that complements my teaching ability. Being an expert does not make me a bad teacher. Being a bad teacher makes me a bad teacher.

I don't think you can be a great teacher without being an expert or close to it. It's not so much content but knowing what to emphasize.

Habit X: will kill you, will hurt, will cost time/money, will make you look like an idiot, will save time.

Maybe it should have been titled "why some experts think teaching is easy, and why they're wrong". As in, an expert might naively feel like they can teach a subject well simply because they know it well. However, the article shows why that intuition doesn't actually work. If, on the other hand, you want to put significant time into learning how to teach, then I imagine anyone can be a good teacher.

I found the article itself fascinating, because I've gone through that experience myself. It may be obvious to some, but for me it was neat to hear that the realization I came to in my teaching is a common one amongst people who have spent lots of time learning and not a lot teaching.

Warning: anecdota.

Having taught music and dancing at various points in my life, I've noticed that -- at least in these sorts of fields that require some sort of body mastery -- the best teachers often weren't the most naturally gifted. They often had to struggle and reason their way through issues in ways that the most gifted/natural musicians/dancers didn't. And because of that, they were often vastly superior teachers simply because they had learned how to think about and articulate a given subject. They could translate the wealth of techniques they had developed to learn the subject to others.

And perhaps, having struggled themselves, they would be more patient.

True, that's all ad hoc theorizing on my part, so I wouldn't promote that as some sort of universal truth, but it's something that I've heard from lots of musicians.

(All of the above is tangential to the OP, btw, which I don't necessarily find that compelling.)

That's what Carl Sagan said about his ability to communicate science information to his audiences. He had difficulty with the material, struggled through it, and found that be was more conscious of each step of the process, and could better explain it than his peers who got it naturally.

>> "Saying that all experts are bad teachers is like saying all <insert ethnic group> are <insert stereotype>."

To be fair, the linked article makes exactly 0 claims about "all experts." Instead, the claims pertain to "experts," which could be fairly interpreted as "some experts" or even "a preponderance of experts."

If the article meant "all experts" it surely would have said so. Neither the string "all experts" nor "every expert" appears in the linked article.

I know my objection seems pedantic but I see this happen all the time, where someone makes a claim about e.g. "bicyclists do X" and then someone else gleefully and sanctimoniously jumps in with an objection like "How can you stereotype all cyclists like that, you are a terrible person!" When in fact the original claim explicitly did not pertain to "all cyclists" but rather merely to plural "cyclists." But now the claimer is super vilified as a cyclist-hater.

I don't think "some experts" nor "plural cyclists" is the most reasonable interpretation.

If I write "Black lives matter", it is not understood to mean that at least 2 lives matter (even though that's a perfectly valid, literal reading of the phrase).

You are arguing from extremes on both sides. A blog post is not a math PhD defense. Nor is it a formal systems proof. The words do not have one and only one possible interpretation.

It is extremely common writing style to dedicate an article to "X does Y" when trying to talk about an interesting tendency or correlation. Unfortunately, it is also extremely common for people who work in formal systems to get an allergic reaction and feel compelled to point out that "X does Y" is not formally, literally, universally correct under all conceivable circumstances. This leads to distracting arguments that waste huge amounts of time that could have been spent discussing the interesting tendency/correlation and what to do about it.

The proposed alternative is usually that the author should have peppered the article with caveats, exceptions, disclaimers and preferably citations of formally verified statistics. Again, this is not the Journal of Science. This is an opinion piece. Watering down every statement with disclaimers would be a distraction from the central point and would only serve to satisfy the allergies of a few highly formal readers.

This feels like every discussion on the internet.

We give no quarter when it comes to generalizations.

Wow, how's that for some generalizations?

>Now imagine an expert who understands this model and tries to teach it to you. Would they take you out in the world and let you encounter Binkles? Probably not.

So it's not "experts" that are the problem, but "bad teachers". Trust me, not taking the time to teach properly is not something unique to experts. Plenty of non-experts fail to teach well, too. And plenty of experts teach well.

The best teachers are people who know the subject and know how to teach. Eliminating them from teaching simply because they know the subject will eliminate the best teachers.

Yes this article is like "Experts have built all the abstractions which makes them completely unaware of these abstractions, which is likely false. And even more so if the expert has an advanced degree where they nearly always must teach some classes.

One factor, is that it really depends on how that person arrived at mastery and their level of metacognition in the process. If anyone fails to grasp the mindset of the person they are trying to explain a concept to, they will do a poor job.

What do a I know? What concept am I trying to convey? What context does the other person have? Based on their questioning what gaps do they have? How can I fill in the links in that concept graph so that making a connection to that new knowledge will fit in.

How does Feynman fit into the expert/teacher spectrum?

I agree. I call what I believe the critical factor in being a good teacher, "sympathetic imagination"-- the ability to imagine what the other person perceives.

I understand the article, but the best teacher I ever had was an expert in his field, and taught by the Socratic method. Only teacher I've ever had that had enough knowledge of a topic that he could both truly answer any question we asked via the method, _and_ answer in such a way as to direct the lesson. Miss that guy.

Everyone who thinks they understand a topic should try to conduct a lecture with this method. It's so hard to not fall back on "because that's the way it is" or make declarations of truth. The effort involved in getting your audience to think though a subject often reveals how little careful thought we've put in ourselves.

I have often fallen into a trap where

  1. I really like a subject, I jive on it for awhile
  2. I try and teach it to people
  3. They immediately reach the bounds of my knowledge and I cannot answer their questions

To adequately teach a subject, the gap between the what you are teaching the student and your own knowledge needs to be adequately large because concepts will have "stringers" that exit that beginner sphere, loop around and re-enter. You can't teach to the limits of your own knowledge.

I've done this - but I also find for a limited group of subjects it's fine!

I know my music theory and I know how to play piano [0]. I gifted a close friend a keyboard and taught her everything I knew about piano and music theory. Within 3 months she was a better pianist than I was - and two years later she was playing for the school choir and talent show. But that is a subject in which "You learn the basics - and the rest comes with hours and hours and hours of practice."

Another example is language (can be iffy if your pronunciation/grammar isn't already near-fluent, even if your vocabulary or niche knowledge (eg: idioms) isn't fluent). You can teach someone the pronunciation, some basic grammar/vocabulary, and guide them towards becoming fluent themselves. A large part is being able to guide people towards resources so that after you've taught them all you know - they can teach themselves from there.

[0] Not very well, but just well enough to play this with only a few mistakes: https://www.youtube.com/watch?v=UNajh166DBc

I think what you are describing is boot-strapping your peers. This is often the case when there is a mix of advanced-beginners and total neophytes. Like one student explaining IDE configuration, or how to debug a problem by looking at log files. The teacher in this case often learns as much as the student.

I guess that is one way to describe it! I bootstrap people to help them become self-directed learners. I get them past the hardest part of learning something new: not knowing where or how to start.

I would never want to hamper the flow of information, or bootstrapping people into a new subject. But bootstrapping is somewhat different than having the deep knowledge that is required for really teaching a subject.

If the skill or area of knowledge lends itself to constructivism, then this style of pedagogy is a perfect match.

Boots on!

Perhaps someone can sort this into the fallacy in which it belongs.

Experts make bad teachers -- Most people make bad teachers -- easily explained by teaching being a skill separate from other skills.

"Technical people make bad managers" is another one I see put into use all the time.

The article is saying that having something explained to you by someone who's abstractions are closer to your level will be, in the default case, better than by someone who has more "advanced" abstractions. Additionally, great teachers are ones that can match abstraction levels with their students and slowly advance those abstractions as the student learns.

So, no it's not a fallacy, and yes the process of becoming an expert reduces the likelihood that someone will be a good teacher (of non experts).

> and yes the process of becoming an expert reduces the likelihood that someone will be a good teacher (of non experts).

I agree, but would add that it loops around. Somebody with a very good grasp of the abstractions at some point can go back and review all of the intermediate ones. They can hold them simultaneously and help guide students through them. That makes them excellent teachers.

I've been thinking about this through the day and it really may depend on the domain being taught.

A non expert may be able to teach a well defined skill step by step better than someone else who performs those steps without thinking about them anymore.

I've done too many years of math tutoring to know that the non experts in math teaching math are not doing a consistently good job. A non expert in a complex field has no way of teaching off script -- which is where a lot of students will end up entirely by accident.

I see it all the time tutoring math. I'm helping students who wandered from the one way that it was being taught. I help them get back some understanding when they feel lost trying to execute algorithms to solve problems.

The article is very long on assertions and short on everything else. The final argument against this in my experience is that experts blow away non experts on fundamentals. Sports, math, science, writing, cooking, construction, painting etc...

You may be looking for Sturgeon's law: "ninety percent of everything is crap."

> The phrase was derived from Sturgeon's observation that while science fiction was often derided for its low quality by critics, it could be noted that the majority of examples of works in other fields could equally be seen to be of low quality and that science fiction was thus no different in that regard from other art forms.

https://en.wikipedia.org/wiki/Sturgeon%27s_law

Teaching is a skill that is orthogonal to the subject being taught. Like patience (closely related) and emotional intelligence (tangentially related), you either have it or you don't.

Purely anecedata, but I don't know of anyone I'd consider a good teacher that was taught to be such. It seems innate to them.

When that one teacher that everyone loves, but always seems exhausted, hears your request for help, takes a long tired sigh, then turns on a smile and spends the next 30 minutes walking you through the material....That's someone who had to learn to be a great teacher. It obviously costs them a lot to teach, yet teach they do.

There is one reason that can make it harder for "experts" to teach a subject that others may teach better.

Best I can explain this is from a personnal story. I used to do rollerblade... I mean a lot of it. It was my main mean of transportation ans I could do easily 20-50km per day. Obviously after a year or two of this I got pretty good and many of my friends would ask me to teach them how to do it. And I did, relatively well, they got the hang of some subtelties of breaking and turning.

However, I noticed that as time went on, I found it more difficult to teach it. It got so natural for me that decomposing the movements in their atomic parts was difficult.

I cound identify two factors that contributed to me going from a decent teacher to a lousy one while at the same time I went from a decent rollerblader to a pretty good one.

For one, the abstraction went from a concious one to a subcouncious one. I no longer had to think about doing it right, I just did. Second I had not taught anyone for a while, so I did not keep in touch with how I built these abstractions in the first place. Both together contributed in me forgetting how to build these abstractions.

To teach one does not have to be a foremost expert in a field but just ahead of the target audience to be in control of the material they need to absorb. The greater the difference in knowledge between the target audience and the teacher the harder for the teacher to "bring down" his thaught processes to their level. That is unless you 1 - were teaching the whole time between getting from pretty good to expert. This way you've kept contact with the different steps and breakdown you'va had to go through yourself while learning. Here is probably where most have had their bad experience when an expert tried to teach them but miserably failed at it. 2 - are a natural pedagogue, in other words you are a genious at making things around you look simple. I think Richard Feynmann would be an stellar example of this.

That being said, as a corrolary to point 2 above, there are people that are just bad at teaching reagardless of any other factors.

I've had this notion for a while that the most qualified person to teach you something is often a thoughtful person 1 or 2 levels above your current level of understanding, not someone all the way at the top who can no longer relate to your level of understanding.

All this said, being a truly good professional educator is a learned skill.

I second that. I remember having tutorials with professors that were highly decorated. FRS and a whole list of acronyms after their names.

I always found it a struggle to ask them questions that helped my understanding, because they were talking the expert language and I was talking layman. Before you know how a transistor works, how do you know to use words like "gate" in the way that explains how it works?

By contrast, the guy who was best at teaching was a phd student. He seemed to know exactly where I would struggle, and he had a very simple manner of explaining the next step.

I disagree with the premise:

> "We’d all agree that to teach a subject, you must know the subject. So you’d think that experts would be the best teachers, but they’re not. The question is why?"

That is like saying that dogs are animals so you'd think all animals are dogs.

Moving past of it, I took a class on education which is probably the best class I've ever taken in my life. First they made us create a 1 sentence stating our goal for the teaching material we were going to create. Then we created the content outline and then they made us ask ourselves how the student would profit about each of them for the course objectives. We realized how many of the things we added were abstractions we learned through the years and totally unnecessary for a beginner in the subject.

To make the group assignment, we created a series of concrete tutorials where, through concrete examples, we tried to add a specific (or more) new material on each one while strengthening the previous learned lesson (each would build on each other).

I am interested mainly in group learning (though I do some one-to-one), so another important point is the speed of the learner due to previous abstraction models. While on one-on-one you can tune up/down to the person, with group learning the most you can do is to put them in knowledge groups or try to automate it (which is quite hard). For instance, unless we are talking about kids, most people know that you read top to bottom in English/most languages and left to right, so explaining that about programming is not only unnecessary but it makes your course boring. Then depending on the level you can skip variables, or variable types.

Lessons don't need to be 1:1 to be tuned up/down to the learner. For example, check out oppia.org, which has a range of lessons which students can take themselves without supervision.

They learn by answering questions, and the flow changes depending on their answers. Like an robot teacher using the socratic method.

The OP makes the false assumption that the goal of the teaching is to generate other experts. That may be true of some courses at some schools, but I would say that most teaching does not involve trying to turn a student into an expert.

For instance, I am a lawyer. I am an expert in certain areas of law. I teach/sell a lecture on intellectual property law, mostly to startups full of young people who don't know IP from IP addresses. I have no intention of turning them into IP experts let alone lawyers. That's not why they hire me. My job is to give them a taste of some basic rules, to give them enough knowledge so they can spot some red flags. And I'm there to answer specific questions, to address those red flags. That;s most teaching. You want the expert to relate a tiny piece of their knowledge, the focused bit you actually need, so that you don't have to spend years at law school. Half of the task is relating knowledge, the other half the selection of which knowledge to relate. The expert then goes away to continue learning in their field, returning as needed to keep you abreast of changes. The relationship remains asymmetrical.

"If you can't explain it to a six year old, you don't understand it yourself."

This quote is attributed to both Feynman and Einstein. Both were very good in breaking down complexity into chunks digestible by mere humans. Ironically the author choose an Einstein image.

It's a common sentiment. [Ernest Rutherford](https://en.wikiquote.org/wiki/Ernest_Rutherford) is known for the quote "If you can't explain your physics to a barmaid it is probably not very good physics. "; the point being that if you can't find and explainable version of your theory, you don't really understand your theory yourself. There are also (potentially apocryphal) stories of Rutherford talking over his work with the cleaning staff as means to help himself make his own theories clearer -- the challenge of explaining it was useful in understanding it.

Unexpectedly, it turns out to be pretty hard to explain anything from the adult world to a six year old.

It's not so hard to come up with some massively simplified analogies that can give an adult the illusion they understand a topic they remain fundamentally clueless about - pop science does this professionally.

It's something else entirely to teach specific professional domain competence to the point where someone can work as an expert among other experts.

Oversimplification and popularisation do STEM no favours. It helps no one to believe they understand string theory or climate change when they know barely enough math to add two numbers together.

A more realistic understanding of how hard physics and engineering are would mean more respect from the public, from law makers, and from other professionals.

>Oversimplification and popularisation do STEM no favours. It helps no one to believe they understand string theory or climate change when they know barely enough math to add two numbers together.

While I understand your general point, I think that oversimplification is vital in order to convey the importance of scientific work to the (mostly) non-scientists that often control the purse strings.

Though, if you can ONLY explain it to a six year old, then you still probably don't understand it yourself.

Very odd article to me and not particularly fond of what it's reaching for as a concept. I mean, I think the premise being promoted is debatable from an Educational Research perspective, which I got familiar with in graduate school.

Honestly I kind of bristle at the implication being done at the bottom of basically letting a child/student/trainee wander through the forest of music in the name of learning abstraction the hard way, because, you know, experts aren't good teachers. With something like guitar, I'd argue anything but and expert makes a mediocre or even poor teacher because of the nature of the objective.

It's not against the rules to reply a student's question with "I don't know the exact answer to that, but I will get back to you" but I'd like to think those are infrequent asides in an otherwise useful exchange where the teacher/prof actually knows the ins and outs of things and can explain them because they had to learn them too.

Then we start getting into issues of diversity in learning styles and the conversation gets a lot more complex than what this piece asserts. Good food for thought and a neat topic, but I'd pass on endorsing its conclusion(s) as stated herein.

I like this post, but perhaps it'd be more accurate to say that expertise is a necessary but not sufficient condition to teach a subject? I do find that the very best teachers have mastery of the material.

On a related note, when I posted this notebook about the sigmoid function to HN

http://karlrosaen.com/ml/notebooks/logistic-regression-why-s...

it was upvoted to the front page (presumably by non-experts like me) but derided in the comments by some experts, I think in part because it seemed so obvious to not bare explaining. So I think a challenge of having expertise is to remember what it's like to not understand something and take the beginner on the journey somehow, as the article gets at. And from this perspective a non-expert or aspiring expert can be in a sweet spot of just having learned the material him/her self. But it can be dangerous if you end up with the blind leading the blind.

Maybe this came up in the earlier comments, but... why are you calling w_0 + w_1·x_1 + w_2·x_2 + ... + w_n·x_n a "linear combination" of the x values? It isn't. (Nor is the constant term really a "weight".)

How is is w1x1 + w2x2 +... wn*xn not a linear combination? Or is your objection that, upon adding w0 it is no longer a linear combination of x1,...xn? In that case would it be more accurate to say, "a linear combination of xs plus a constant"?

Correct, the constant term makes the function nonlinear. A linear function must necessarily satisfy f(0) = 0 (or, I guess, f(0) = infinity).

The author makes some very good points, but seems to have missed the real punchline. Instead of making this about experts-vs-non-experts, the real lesson is the value of pretesting and similar methods.

http://www.nytimes.com/2014/09/07/magazine/why-flunking-exam...

The basic idea is simple. Instead of "telling people the answers", let them bang their heads against the wall first, trying to solve a problem that they have never studied before. Once they've realized that they don't know how to solve the problem, then teach them how to do so. This way, because the students have experienced the frustration of solving the problem themselves, they'll truly appreciate whatever knowledge you provide that can help them out.

There are experts who are bad at teaching but be assured all good teachers will be good enough "experts".

I don't agree that experts are necessarily bad teachers. Some of the best experts in physics have been those who were widely regarded as great teachers (Feynman and Heisenberg being great examples).

However, I think there is something to the article as applied to the software industry. Often times we have come to understand solely through abstractions and abstractions are dangerous in this regard particularly when they don't match reality.

A really good example of this was the effort to apply the OSI model of abstractions to TCP/IP back around the turn of the century. Without an understanding of what made the ITU OSI effort different from the DARPA TCP/IP effort, at best one got an overly confusing and misleading understanding of what was actually happening from a networking perspective.

I do a lot of teaching these days as I have changed from being a self-employed consultant to being an employee at a Swedish consulting and training firm, and I do a lot of thinking as to "what are the right abstractions to push in this course?" In IT, that is a remarkably difficult question and one I answer largely through my love of history. For example, when I talk about object-orientation, I start with procedural programming, discuss why structural programming largely replaced it, and then discuss the problems of having your code contracts based on your data structures in C. From there I go to my main abstraction: Object orientation is an effort to hide state changes behind interfaces, to make the interface responsible for the state rather than the outside routine responsible for it. (This also makes it easy to start discussing functional programming.)

The danger of the expert is that the expert may become complacent about this question but the best experts are those who have not. Feynman wrote about insisting that he could explain his theories to his non-physicist relatives. Heisenberg wrote a book on the history of philosophy and how this applies to physics arguing the whole time that data does not imply theory.

If our experts aren't good teachers, that says more about what we want from experts and about our industry than it does about expertise generally.

This article would be stronger if it were just about the importance of learning from concrete examples.

After all, education is all about taking shortcuts compared to figuring everything out from first principles.

Richard Feynmann was one of the best teachers ever!

>> Binkles have long plackerts and whipitat their snoblats when they get excited

What I was most interested by was how much of this example you could understand. A binkle is some kind of animal, a plackert is some kind of body part, i'd imagine a whipitat is some kind of action, and I'd imagine a snoblat is another kind of body part. Already I know more about binkles than I did.

You're also, however, a long way from a dog simply by describing a 'barking thing'. There are things that bark that are not dogs, and there are somewhere, dogs that do not bark. Wittgenstein demonstrated that there are an entire set of items that you could call a table than share no common features whatsoever.

I'm a fan of the dialectic method for this reason: it encourages people to go beyond the limits of their own knowledge by applying first principles to things they don't understand. This tends to take you a lot further than "here, learn this."