← The Hardcore Self-Help Podcast with Duff the Psych
The Hardcore Self-Help Podcast with Duff the Psych
420: Productive Failure with Dr. Manu Kapur
00:47:27
Loading summary
Transcript105 lines
- [00:00]
A
So productive failure is the idea that if failure is good for learning, then we shouldn't wait for it to happen. We should deliberately design for it, understand how, when, and why it works, and bootstrap that for deep learning. If you're about to learn X, and I've have had you think about all the knowledge that's relevant to X in different ways, and especially in ways that it doesn't work right, it activates the cognitive system, and then you're ready to just process new information. And here we found something very interesting that, you know, it's okay to fail a little. It's okay to feel a little bit of these negative emotions, because the idea that only positive emotions correlate with positive learning outcomes is not true. Failure prepares the ground for you to then start making these contrasts. What does not work with what works in what ways, and that leads to deep learning.
- [00:53]
B
Negative emotion is a potent memory tool, right? That's why trauma happens. We are very good at remembering traumas or embarrassing things that we've done, et cetera. But it's in that the context is different.
- [01:05]
A
We looked at more than 160 odd experimental effects across 50 odd studies in different contexts, and we found that the average effect of productive failure over direct instruction is quite robust. It's equivalent to twice the amount that you might learn with a good teacher in a year. If that is true, if failure can be powerful, then how do we actually design for it and use it in a way, in a safe way for deep learning?
- [01:41]
B
All right, everybody, I'm excited to have today's guest. I have Dr. Manu Kapoor. Dr. Kapoor, how would you like to introduce yourself to the audience?
- [01:49]
A
Well, I'm a reluctant or an accidental academic, but, yes, I'm a professor of learning sciences and higher education in. At ETH Zurich in Switzerland.
- [01:59]
B
What is that? So what is that eth what is that school?
- [02:03]
A
Well, it's one of the top. It's the top continental university in continental Europe. It's the home of 25, 30 Nobel laureates and field medalists, including Albert Einstein, this chap you might have heard of. And yeah, no, it's not very well known outside, but it's one of the top universities consistently ranked amongst the top 10 in the world. So. Wow.
- [02:26]
B
So you have some big shoes to fill.
- [02:28]
A
Yes, it is humbling. Yes.
- [02:32]
B
Awesome. And you've been doing some other podcast interviews today. How are you feeling right now? Just kind of coming into this one?
- [02:39]
A
No, feeling really good. I mean, I think each has its own energy, and so far, you come out with paradoxically more energy talking about. I mean, when you talk about the stuff that you like and you're passionate about, I guess it just energizes you. Anyway, so now I'm feeling good. Looking forward to this.
- [02:54]
B
Yeah, absolutely. We're going to talk all about sort of your ideas about productive failure, things within that framework. I intentionally didn't learn too much about it because I didn't want to play it straight like I didn't already know what you're talking about. So what context do we need about you as a person to understand how you came into your line of work?
- [03:17]
A
Yeah, so like I mentioned the accidental part. I mean, in my teenage years, I was training to be a. A soccer player. Then an injury, surprise, surprise, derailed all that. And then I had to figure out something else to do. And I finished up my engineering bachelor's. Yeah, not wanting to become an engineer per se, but that was my plan. Plan B. Then I ended up being in the startup world, went into industry for a while, and then finally got around to teaching math to kids, to high school kids. And that I really liked. And the idea that you can understand something abstract, or how does one understand something abstract, such as math, sort of caught a hold on me. And so then I moved from teaching to research on learning abstract concepts in mathematics. And once I ventured into that, that was my Ph.D. at Columbia University. That's when I started to look at, okay, what are the best ways to design? First of all, how does one learn? What is the mechanism by which someone understands something new?
- [04:23]
B
Right.
- [04:24]
A
Especially if that new thing is abstract. And if we understand those mechanisms, how can we design interventions or applications or learning methods where we help people learn abstract, new abstract concepts better? There. I built up this entire theory and architecture around using failure as a mechanism for deep learning.
- [04:46]
B
Okay, awesome. Yeah, we'll get into the mechanism for sure. What was like, where did you grow up? What did your parents do? Was there expectation for you to go into soccer or academics or what was their expectation?
- [05:00]
A
Well, look, I grew up. I was born in a very humble family in India, in the northern plains of India. And then we moved up north to Chandigarh, which is where I picked up soccer. I mean, which is a strange choice because if your audience knows India, then India loves cricket. And so here I was playing soccer and my parents were pretty cool. They just let me do what I really want, what I was passionate about, and I became good at that pretty quickly. Yeah, my parents supported me throughout, regardless of which path I took. I just ended up Taking many parts throughout my teenage years and throughout my 20s and eventually landing as a professor.
- [05:44]
B
Yeah. So obviously failure is. It's a loaded term, but it's kind of central to your approach. How has the role of failure kind of taken place in your life or maybe changed throughout your life?
- [05:59]
A
Yeah, I mean, you know, the dots always make sense looking back. So when the soccer ambitions or dreams crashed, that was not nice. Right. I mean, you trained for years and years and that doesn't happen. And then you try multiple things each. You give it a good shot in their own right. But either your passion is not there or, you know, you're not quite up to the same kind of metal for it. Either way, you know, things don't work out. And that was. That's been the story throughout my 20s, is to basically try all the things that I could try given the opportunities that I had. And. Yeah. And each time when things didn't work out, I moved on to the next thing. It. Yeah, it really sucked because you have to start all over again and try a new thing all over again. Yeah. And somehow after trying four or five different parts, so to speak, each revealing something to me about myself, which was. Which was, on hindsight, very nice. Not in the moment, but on hindsight, nice. And eventually research and academia, scientific research, actually just stuck and. Yeah, I mean, that's been my career for the last 20 years.
- [07:19]
B
Yeah. You know, that's good to hear. A lot of people underestimate how often people actually do pivot in their life, whether it's due to choice or due to circumstances kind of out of their control. And they're really hard on themselves, I think. And so it's always good to see somebody who's in a position that they really care about and they're very good at and the some esteem there who has gone through a lot of different, you know, failures, if you want to call them that.
- [07:44]
A
Yeah. Then it's easy to get. If you don't. I mean, I. Things just happened to me and it was not by design, but if I think about it looking back, I think I'm glad that they happened to me because at least I did not settle into something, a suboptimal career or suboptimal path that, okay, I could do this, I could be happy, I could live a decent life. But, you know, so on hindsight, it's really good that I tried many things and was. Remained true to what I really wanted to find and do in the end after my first love of football or soccer.
- [08:18]
B
Can you imagine Yourself as an engineer.
- [08:21]
A
No, I don't. I. That's why I didn't. That that path didn't work out. But in the work that I do now, the training as an engineer comes in handy because you're designing systems, right? Learning systems you're designing. So that kind of training kind of helps you right now, even though at that time, and I think that's probably true for how, how the old varied experiences where, which in and of themselves didn't go anywhere but the things you learned, there's a latency, there's a, you know, usefulness to even the things that went nowhere. And you never know when in your life they're all going to come back and say, let's pull them all together now. And that's the stage I've been. And so in that sense, it's been a really healthy dose of. Big, strong dose of productive failure itself.
- [09:11]
B
Yeah, yeah, yeah, that's. I mean, I, I completely agree. I relate to a degree in terms of, you know, I have two failed podcasts before this one, other YouTube channels that I tried to start and abandoned and stuff. And skills you gain along the way are instrumental to, to what becomes later on.
- [09:28]
A
Yeah.
- [09:29]
B
So hit me with the definition. What is productive failure?
- [09:34]
A
So productive failure is the idea that if failure is a learning, it's good for learning, then we shouldn't wait for it to happen. We should deliberately design for it, understand how, when and why it works, and bootstrap that for deep learning. So if you're about to learn something new one way, one intuitive way that we've all learned for this, I will just tell you exactly what this thing is, try to explain it to you in as engaging and logically structured and clear manner as possible so that with the hope that you will learn what productive failure says, that's not the most optimal way of learning. It's not a bad way of learning, but it's not the most optimal way of learning. And so in productive failure, instead of teaching you or explaining to you, first up, what exactly this thing is, we have a framework, a set of design principles for designing challenging problems or tasks for you, which are designed in a way that are intuitive. They give you access to the problem, but they're designed for you not to be able to solve them correctly, because you can design 2, 3, 4, 5 different ways of, you know, approaching the problem, trying different things, different representations, different solutions, and so on and so forth. But over time, it's calibrated in a way that you won't be able to solve that completely. And it is this failed exploration or generation that the theory suggests that this is. And there are mechanisms why that happens. This is why this, this failed generation or exploration then sets the stage for you to then learn from an expert or an instruction or a video. So that explicit instruction, you know, which we, we normally go straight into, is important. Just that in productive failure, we first designed a safe curated way of you to fail via exploration and then it's followed up by explicit sort of explanation instruction, productive failure. Yeah.
- [11:35]
B
Again, this is such a weird analogy for it, I guess, but it reminds me of the idea of like exfoliating your skin before you use some sort of treatment on it. Like you like rough it up first and then you do something to kind of like let it sink in.
- [11:51]
A
Yeah, in a way, that's how. I mean, you're a neuroscientist yourself. That's how activation works, right? It's preparing the neural networks as well as the cognitive networks to activate relevant knowledge so that new knowledge can be processed better and integrates into your new schemas better. So, yeah, in a way, that's a very clever analogy.
- [12:14]
B
That's just what came to mind. Can you talk about that a little bit more though? Because I actually had a guest years ago, Dr. Jared Horvath, and he talked about how, you know, you're most primed to learn when you've made a mistake, which sounds very similar to what you're describing. So what are the actual mechanisms there, as far as you understand?
- [12:36]
A
Yeah, so priming is one of the sub mechanisms. But when I talk about the four umbrella mechanisms, one is activation, both physiological arousal and activation, but also cognitive activation. If you're about to learn X, and I have had you think about all the knowledge that's relevant to X, different ways, and especially in ways that it doesn't work. Right. It activates the cognitive system and then you're ready to just process new information because the more you can activate, the easier the processing. Right. That's how a working memory works. But also integrate this new information into what you already know. So activation is key. That's one big mechanism. Second, what this activation does is it makes you aware that there is a gap. No, you've tried to solve the problem in several ways. They've not worked fully or suboptimally. Then you know, well, this is what I can do. I have. These are the limits of my knowledge. But there's a gap. And what people generally want to do after having invested in some kind of an effort to solve A problem is they know there's a gap, there's an awareness of a gap, then they want to bridge the gap. And this is when what? Affect. When affect comes in. And by affect I mean, you know, your motivation to bridge the gap, your interest in finding what's the correct way or the more optimal ways of solving the problem.
- [13:56]
B
Frustration, maybe.
- [13:57]
A
Yeah, I'm coming to that. Yeah. Or your, your orientation towards the same explanation that the, the expert is going to give you. You know, if the explanation was given up front, your orientation may not be exactly the same as having tried and failed. And then you listen to the same experiment. Your orientation, your orientation is really about, I want to understand why, you know, and we've studied that. And then there's the emotional aspects which you mentioned. Frustration, anxiety, shame, guilt. And here we found something very interesting that, you know, it's okay to fail a little. It's okay to feel a little bit of these negative emotions because the idea that only positive emotions correlate with positive learning outcomes is not true. You know, some amount of frustration, struggle, anxiety, shame. If it's done in a safe space when you're trying to learn something new initially, then that can be a powerful emotive charge for learning the outcome. Right. So, yeah, and so emotionally as well, the emotional roller coaster is pretty interesting as it correlates with learning out. So if your knowledge system is activated, it. You're aware of a gap and you're in an affective state which is primed, as you said, it's primed for learning. If at that point an expert or a mentor or a teacher or an explanation comes, that assembles it all together. And really the idea is assembly because you want to take bits and components from, you know, the, the exploration of the learner and tie it to the, and integrated with the components that go into making the correct solution or strategy. And it's that assembly, you know, that's critical. So this is what I call like the four A's of activation, affect. Activation, awareness, affect, and then assembly.
- [15:50]
B
Assembly. Walk us through. Maybe an example like, how do you actually construct this into. I know you've done a lot of this in teaching, for instance, so how does that actually integrate in practice?
- [16:02]
A
Yes, in teaching, for example, if kids are learning statistics or some basic concept and you know, say standard deviation or, you know, we give them a data set that's very carefully designed and we give them, we ask them question like, for example, if it's standard deviation, we ask them questions like, can you, here's three or four Football players. Can you decide? You're looking at the data, who's the most consistent of them all over the years. Right. And it has an intuitive appeal. People are engaged and they start looking at the averages, the totals, who's gone the most, and the least and very intuitive ways also to describe the data using their bodies and gestures. And the data is designed in a way that if they were to look at averages, we keep the averages the same so that they try it. And so it doesn't work. I can't decide using averages. If they look at the extreme points, they keep them the same too. So that, oh, I tried the extreme points and still can't. So that's the iterative calibration process by which we make these tasks invite your prior knowledge, but make them not solvable fully by that. And therefore, each time you try something, you reach an impasse, you get stuck, or you fail, and then you have to try something else. And so this process goes on for 20, 30, 40 minutes, depending on how much time is available. And once that is done, that's when a teacher could come in the classroom context at least, and say, oh, let's look at what you've done. This strategy. Does it. Why does it work? What part of it works? What part of it does not work? Let's compare your strategy with your strategy. And oh, okay. Then you start to compare and contrast. But the biggest components, the biggest sort of benefit comes when you compare the student generated or learner generated solutions or representations with the canonical, the expert strategies. Because that's when students see the contrast between here's what I did and here's what the expert. That's when they really see, ah, this is why this is done in this way. Yeah. And that leads to deep understanding because they start to notice what is critical if you just give them the correct one. You know, it's like looking at a chair for the first time in your life and figuring out what is what makes a chair a chair. Right?
- [18:24]
B
Yeah.
- [18:24]
A
Which is very, very hard. Or looking at a color like red for the first time in your life and wondering what that. What makes this color red? Unless you have contrasts of red and other colors to contrast with, you can't really understand what that is. And it's the same with concepts and ideas. Failure prepares the ground for you to then start making these contrasts. What does not work with what works in what ways, and that leads to deep learning and knowledge.
- [18:54]
B
That sounds almost opposite from the way most classes are. Right. You usually start with the introduction of the material, the Broad concepts. And then you might break out from there and try to apply it in some sort of way. Only after you've done the learning.
- [19:11]
A
Exactly. So it's always learn first and then apply. Which is, like I said, I don't want to say this is totally bad.
- [19:18]
B
No, that's how we all learned. It obviously works to some degree.
- [19:21]
A
Yeah. And the question is, can we make it better? Are there better ways now that we understand the science a little bit better, are there better ways to optimize that? I mean, obviously that instructional component is important even in productive failure. But you derive a bigger bang for the buck if you precede this learning with failure driven sort of activation. So you do that, then you learn and the application is still important. Right. You still have to apply and achieve mastery and so on and so forth. But if you proceed all of this with a failure driven activation, then you learn deeper so that when you apply and in education contest is critical, we are using failure early in the learning process where it is safe so that you learn deeply. So that when failure is to be avoided, like in exams or tests, where the stakes are high because you've learned it deeply, the likelihood of failure when it actually matters is lowered. Right. So this is not so productive. Failure is not saying all failure is okay. Yeah, yeah. I'm not saying. Yes, when failure happens, it's fine. Productive failure saying, let's be intentional about designing for it in the initial safe space of learning and then learn so that you learn whatever, you learn deeply so that the stakes, when the stakes are high, that failure is reduced.
- [20:51]
B
Absolutely. I imagine there's a sweet spot as well when you are designing these challenges. Because if the frustration level, if the emotion is too high, then you might shut down a bit. But you're right, you know, as a psychologist, I'm thinking about it and you're very right. That negative emotion is a potent memory tool. Right. That's why trauma happens. We are very good at remembering traumas or embarrassing things that we've done, etc. But it's in that the context is different.
- [21:21]
A
Exactly, exactly. So if you can design this failure or these, which may invoke this negative emotion in a way that it's safe, then you know, two things are happening. You're emotionally charging the situation, the learning itself, you know. Right. So from a neurocognitive standpoint, your amygdala and then you have frontal parietal networks and so on. So attention network and then the processing, so all that kind of gets triggered and geared up nicely. But Also you're teaching this learners that it's okay to be frustrated. And how do you learn how to deal with that? Right. So a, to, you know, yes, I'm feeling frustrated right now, I'm feeling anxious right now. But I'm also in the learning zone and that means it's okay. I must learn to manage it and maybe derive the most out of it.
- [22:06]
B
Right.
- [22:07]
A
That to me is a bigger aspect of learning as well. Like if you let first and then you can raise the bar of frustration and you know, negative emotions that you can sort of tolerate and deal with up to a point, obviously. So that you can always keep using it to learn. To use it to learn.
- [22:30]
B
Yeah, yeah, absolutely. And it makes me, you know, I have, I have a couple small kids, you know, six and eight years old. And so, you know, there are frustrations with the way that elementary school works here. A lot of stuff from the time that they have to go to the way things are taught and tested and things like that, it, it seems so counter to the way that we work most effectively. And this, this highlights that further for me. You know, it's like, yeah, it's, it's kind of the opposite of what is done where there's stakes and frustration feels scary and stuff like that.
- [23:07]
A
Yeah. And also in the children, like one of some of our steepest learning curves are the first seven, eight, ten years of our lives. Right. And there before, prior to schooling at least. I mean, if you look at how children naturally learned when we were children, there's a lot of productive failure. Children are not afraid of failure. They haven't been enculturated into a system of failure. Right. Or that sort of doesn't look at failure as something that could be positive. Yeah. The way they learn how to make sense of the world, how they learn to talk, walk in every aspect. You know, they, they try, it doesn't work out, they try again. You know, if it's a safe space at home and in kindergarten or wherever, then this is a mechanism that's very well supported. Just that when we reach school and throughout the schooling years and after, I think we get enculturated into thinking that failure is bad. Of course failure and high stakes can really sucks and it's bad. But now we can use that early on in a safe way so that, you know, high stakes failure can be reduced.
- [24:09]
B
I think a lot of us do it without even realizing it. Because I was just thinking about how, let's say that I am trying to. Recently I was trying to fix A plumbing issue in my bathroom. Right. So I go, I try it, I try to see if I can figure it out. I do this, I do that, I fail a few times, and then I go to YouTube and look at an expert teaching me exactly what I need to do in that case. And then I go do it. I think there's a lot of things in our lives that we naturally do. That way we come to some sort of learning or expert after we have already gotten in there and tried ourselves.
- [24:42]
A
Yeah, yeah, exactly. As long as when you tried, whatever you did to fix the problem doesn't flood your entire.
- [24:47]
B
Right, exactly. The stakes have to be right. Absolutely. Yeah. So you've actually done some research on this, right?
- [24:56]
A
Yeah.
- [24:57]
B
What. What are the highlights of. Of that research? You know, empirically? What is that showing?
- [25:02]
A
Yeah, so remember we talked about what's what. The method that's commonly used in the classrooms when you. When we were young. And even now, it's like, it's what we call direct instruction. The idea is, if you don't know something, let me teach it to you very clearly in a very engaged manner, in a structured manner, Give you lots of examples to help you understand, and then you apply. So basically, the model is learn and then apply. And we've compared this model with productive failure, where you first solve this problem and then you learn. And then we've tested this on three different kind of kinds of knowledge variables. You know, after students go through one of these two conditions, we test them. We test them on basic procedural knowledge, the kind of stuff that comes out in standardized tests and exams. But we also design research instruments on conceptual understanding. So do you understand the formulation, why it is the way it is, and so on, and then your ability to transfer it to learn more advanced concepts. Because to me, that's the ultimate goal is if you learn something and you can only apply in the context within which you learned it, then that's very shallow learning. But if you deeply understand it, you should be able to apply it in more novel contexts. And so the transfer. And transfer is something you can't really test in a standardized exam because I test what you know, and then it's kind of bounded by that. So we test on three different variables. Procedural knowledge or basic knowledge, conceptual understanding and transfer. What I call, you know, the. Yeah, basically, basically, let's give the basic procedural knowledge, conceptual understanding, and transfer.
- [26:49]
B
Yeah.
- [26:49]
A
And what we found is if you compare the two methods over a number of studies, the performance on basic knowledge is equal and very high. So both methods are very Good at achieving high levels of efficiency and effectiveness on learning basic knowledge. Right. But the differences are in the conceptual understanding and transfer, where productive failure far exceeds direct instruction on that. So you have two methods that you know, achieve the same level of procedural knowledge. So you've got two groups of people who look very similar on basic knowledge, but their understanding of it and their ability to use this knowledge to novel is very different. Right. And so the idea is that learning path matters because sometimes people argue that, oh, it doesn't matter how you learn the basic, let's just get give you the basics and then you can later apply it and be more creative and inventive and transfer it. It turns out how you learn even the basics actually influences how well you can understand it, how well you can transfer it. So that's been the, the major finding. And it's not just a finding that's coming out of my lab or my studies. You know, a couple of years ago we did a meta analysis which you're familiar with, and you know, we looked at what's the state of the art in productive failure research when productive failure has been compared with direct instruction. And we looked at more than 160 odd experimental effects across 50 odd studies in different contexts. Only about 15% of them were mined. So that means the rest of them, much of the work is independent replications and reproductions of the work. And we found that the average effect of productive failure over direct instruction is quite robust. It's equivalent to a year of twice the amount that you might learn with a good teacher in a year for direct instruction. Yes, exactly right. But if you do productive failure, well, this effect could be three times. Sure. So it's, it's a, it's a very strong and robust effect. And yeah, those are the main kind of big ticket findings. Actually there's another bit that's really, we were not looking for it, but we also found that students who learn through direct instruction, they were given the same sort of productive failure tasks after they'd learned. So in the application mode, you know, you learn it and then you solve the problem. Which productive failure students were given up front. Right. And we found that, yes, direct instruction students were able to solve the problem completely and correctly because they'd learned at first. But when we asked them to, you know, just solve as many ways as possible, think of other solutions, they just could not get out of the box. Right. So again, the idea that knowledge facilitates application in the domain, but somehow sometimes if you learn it too early, then it kind of constrains the creativity, kind.
- [29:46]
B
Of solidifies your understanding in a very concrete way.
- [29:51]
A
Exactly. I mean, there's notions of the functional fixedness and. And other aspects that come into play as well.
- [29:58]
B
Wow. So that makes a lot of sense about teaching the classroom, things like that. Have you thought much about how this is applicable in daily life? So, say an adult listening to this, how might somebody keep in mind this approach?
- [30:16]
A
Yeah, I think it's a bit harder, to be honest. If it's designed for you, it's a lot easier because you don't have to.
- [30:22]
B
Think about it, just go through it.
- [30:23]
A
Right. But yes, personally, I think it requires a lot of discipline. But maybe some people do it naturally, like when you were trying to fix your bathroom. Right. And so just if people can just have this idea that. And I devote a chapter in the book to talk about this as well. If people can just have this mindset to just resist when it's possible and when it's safe to do so, resist reaching out for the correct or canonical ways of doing certain things right. Like you did. Like YouTube is always there, or some manual is always there, some expert I can always talk to, you know, then I think, and then just to try, engage or try something. Try to do something with whatever activity, whether you're learning a language or trying to fix the bathroom or make a meal, as long as it's safe, just try to do yourself in different ways before you consult expert knowledge. And like you said, you can do this. If you have this discipline, you can do this in a month. You can make it a habit, even.
- [31:26]
B
Yeah.
- [31:27]
A
Including. Including in trivial situations like when you go networking and you want to learn how to. When you meet so many people and you want to remember their names, you know the best. I think we do party introductions completely wrongly. So if I were to.
- [31:41]
B
How should we do it?
- [31:42]
A
Exactly? So how. You know, I. I think party introductions should be. You know, we normally, what we do is the direct instruction. Hello, I'm Manu. You are. Then you say your name. In fact, you should. The first job, the first thing we should do is don't tell me your name. Let me guess your name. So I just do a random guess. Hey, are you Tom? And they said, no, I'm not Tom. I am so and so and so. And then I say, oh, okay, you're not Tom. And so then you ask me, hey, are you Mark? And. And say, no, I'm not Mark. I'm Manuel. And so the idea that you just tried and guessed, and even if the guess is wrong, you know, you kind of invoke a mini productive failure event, and then chances are you remember those names a lot better because you made that extra initial failed effort. Right.
- [32:31]
B
Yeah. It also provides more association for you, too.
- [32:34]
A
Right.
- [32:34]
B
Because when you say, hey, are you Mark? And it's a no, but I did have, you know, a teacher named Mark or blah, blah, blah, and then that becomes context that you store the actual information with when you learn it.
- [32:46]
A
Yeah, I mean, in psychology, that's called a retrieval part. So you're creating multiple retrieval parts to reach that target knowledge when you're trying to retrieve it. Yeah, exactly. Yeah.
- [32:57]
B
That's great. And I guess in. In applying to daily life for adults, I guess the thing also to keep in mind from what you've described so far is the stakes idea, the pressure that's there. Like, I was thinking about cooking a recipe. You could totally freestyle it and try to do it on your own and learn through productive failure. That's harder to do when there are five kids waiting on you. And this is actually higher stakes. Much easier to employ that when you're just chilling on a weekend or something like that.
- [33:27]
A
Yeah, yeah. And here I make. And this applies not just in your own life, but even in workplace when you're learning. I make a distinction between learning, identifying, like what is a learning task and what is a performance task. If your kids are waiting and they're hungry and you want to cook something that is a performance task, use all your capabilities to cook the best thing as, you know, as delicious and timely in a timely manner as possible. There's very little room for failure. Right. Whereas if you're on the weekend trying to plan about, you know, you just feel in an experimental mode where, oh, let me try a new recipe or something that I want to concoct by myself before trying to see how things are made, then you have a learning space within which you can. There isn't. There's an. You have the space to explore. The stakes are not high. You can fail, something can get burnt, and that's fine. You try it again, you know. So I think just as a habit, all of us, if you can try to identify or signal, is this a learning space I'm in or is it a performance space? Right. And then depending on that decision or some combination of that, that will then determine whether you go ahead and if it's safe to go ahead and do things. So, yeah, so that's how it kind of advise people to go about it.
- [34:48]
B
That's a great awareness to build. Anyway, I Think, for example, people with social anxiety have a very hard time discriminating between the two. What is a learning context? What is a performance context? Everything feels like a performance context. So maybe leaning more into intentionally building those learning periods of productive failure that might help build that tolerance you were talking about.
- [35:10]
A
Exactly. And if you're not clear yourself, you can ask people or your environment can help you answer that question. Question. So if you get a task from your boss to do something, or, you know, you're doing something on your own, together with your family or friends, you can just ask. Like on this, on the spectrum of learning to performance, where am I? How should I approach this? What. What degrees of freedom do I have to go more the productive failure route or the more, you know, direct and get it, get it done route? You know, so.
- [35:40]
B
Right. How important is it that I nail this?
- [35:42]
A
Exactly right. So. And I think oftentimes we can get those. We. The thing is, we just go into situations making our own assumptions, and oftentimes we make assumptions to take the safe route. When I think in most situations, we have greater degrees of freedom, either strictly by our own agency or feedback from the environment.
- [36:04]
B
Great point. Now, you grew up in India, you live in Germany, correct?
- [36:10]
A
I grew up in India, then lived in Singapore for the biggest portion of my life and went to New York for my PhDs, returned to Singapore a year in Hong Kong, and for the last eight years, I've been in Switzerland as a professor at eth and then January this year, I took the directorship of my university's research center in Singapore, which is called the Singapore ETH Center.
- [36:38]
B
Okay, so obviously.
- [36:40]
A
Well, sorry.
- [36:41]
B
Yeah, obviously you've had some experience with different cultures. Thoughts about how culture applies to this.
- [36:49]
A
I think this is a case where failure naturally happens because we get. You make certain assumptions of ways of thinking and being. And, you know, when you are. When you are in the new environment and you very quickly run. Learn that, oh, you're making the wrong assumption. This is not how. This is not what is acceptable. So your own limits, your own understandings of your own boundaries of what is acceptable, what is not, ways to think, ways to behave, ways to believe and just be. Perhaps, you know, they get calibrated because you have these sort of, let's say, things going at loggerheads with each other or, you know, people. Yeah. So I think going. Every time I've gone into a new culture, a new work, a new place, the biggest learning has been for me to learn about myself. You would think that you would actually learn about the new culture. Right. And you do. I am not saying the biggest thing, the biggest growth that happens is, is the mirror that gets shown to you and you come to realize, oh, I always taught him this. I was limited in one perspective, that actually there are different ways of doing, of approaching life and work and play and friendships and relationships. And so that's the biggest growth is because it kind of sets up productive failure for you, just naturally. Yeah.
- [38:10]
B
There'S, there's a. I mean, obviously I'm just connecting a lot of dots here, so forgive me always intruding with, you know, ideas and stuff like that. But you're talking about the. Getting feedback from the environment and understanding if something is more of a performance moment or a learning moment. That seems especially relevant when you're experiencing new cultures. Right. Say you're at a dinner, you don't know how to navigate a certain utensil or a certain, you know, whatever happens, you could ask, I guess, hey, is this important for me to get right or can I fumble around a little bit?
- [38:43]
A
Yeah, yeah, yeah, exactly, exactly. And there's so many of these little moments, Right. Some of the, most of them are funny, thankfully, because most of the spaces within which people are generally nice. Right. And they, they recognize that you're new to the situation. So yeah, you have the space to learn.
- [39:00]
B
So for people like me that have kids that are in school and their schools don't necessarily embrace this philosophy very much, do you have any suggestions for how we might approach that and try to trickle it in, try to help round out the way that they're being educated?
- [39:19]
A
Yeah, I mean, I think the school obviously controls the formal curriculum. Right. But the expectations and the norms, the culture that you create as part of your family and what's valued in terms of effort and growth and trying, I think parents and families can play a big role in that. Right. So students can still approach the same performance driven school culture from a more growth and effort perspective so that they don't have the same stressors and anxieties that would otherwise. I did that with my, I mean, my eldest son grew up in Singapore and Singapore has a very good education system, but it's also a high stakes system. Right. And so part of our role as parents was to really tone that down, to just make sure that the process, they focus on the process, they focus on the efforts that it's okay if, you know, you're struggling a little bit, like normalizing these and saying that that part of learning is to feel these emotions, to feel these things kind of helps your own children. So I think just in those framing effects and setting the right expectations and knowledge, I think parents can make a big impact on how your kids approach school, even if that is not, as you said, an ideal situation. Ideal, like, yeah, they're following practices that may be not consistent with deep learning, then also after school. Right. So if I don't know how much you help with your, you know, parents normally end up helping their kids, but school, school and homework and so on and so forth. And to the extent that people do that, I think deploying sort of strategies that are more consistent with productive failure. So if you know you're trying to explain something to your kid how to solve a problem, it's best not to show them the solution. But first have them try it out and see how far they get. And they get stuck. And when they get stuck, then you come in with your explanation as opposed to just explaining everything to them. So again, so these are some strategies that you could deploy. Or if they think that they've solved a problem or they've got the idea, then you can facilitate deeper understanding by saying, okay, can you think of a situation where your own thing would not work? You know, again, pushing them to hack or fail their own idea so that, that broadens and deepens their exploration and understanding and not to be satisfied with, with the, with the box that they get put in.
- [41:49]
B
Does that, does that fall into that generative problem solving term that you use?
- [41:54]
A
Exactly. So part of the facilitation of the generation is, you know, sometimes people think they've got a solution, but if you, you know, I call it counterfactual reasoning. I mean, so it's to, to hack your own idea and say, can I fail this? Can I, Can I make it? Can I, when does it fail? Basically? And that's again a mental habit of the mind. So the scientists obviously have. Right. So you know, we never satisfied with, okay, this is it. But maybe there's a boundary condition here. Let me look for when this fails. Even science, for example, if you have a new study with a new effect, people don't just accept it, people try to test it, replicate it. And the logic of replication again is, let me prove you wrong on, right. And, and then if the null hypothesis.
- [42:40]
B
Right, we try to reject that.
- [42:42]
A
Yeah. So, and if, if repeated attempts to fail a hypothesis fail, then you start to say over multiple studies, ah, maybe there's something there. Right. Yeah. So again, that, that kind of mindset, building that mindset with kids after, after work, after their school.
- [42:59]
B
Yeah.
- [43:00]
A
Can Be quite helpful.
- [43:02]
B
You just released your book. Congratulations on that, by the way.
- [43:06]
A
Thank you.
- [43:07]
B
Can you tell us what the book is and what people might expect, who should be picking it up?
- [43:12]
A
Yeah. So the book is on productive failure. It's about unlocking the science of failure for deep learning and growth. In the book I set it up, I describe what productive failure is, obviously, and what evidence we have for it and. And how it works. Then I dedicate this middle part of the book to explaining the science behind it because I really want people to understand. Why does it work? Why is it different from other ways of learning? And what are the mechanisms? So the four A's, I mean, each A is a dedicated chapter in the book. And the third part of the book is about. Okay, so now that you know what it is and why it works, can you use it? So if you are designing, if learning for somebody else, be it for your students, your employees, or you're training somebody or your kids, how do you design for productive failure given a learning goal? Right. And then the last chapter is really about, okay, how do you use it for yourself for your own learning and growth? So that's the. That's the book, really. Yeah.
- [44:19]
B
I can imagine it would be helpful for people like me also that give talks, you know, to, you know, various other professionals and stuff like that, or in I do clinical supervision. I can see that being very applicable there too.
- [44:35]
A
Yeah. I mean, the initial feedback that I've gotten, like, people are trying, like, several areas, like, you know, it's not just educators and parents and professors and teachers or trainers, but also athletes and coaches and business leaders. A big thing in business, actually. You know, HR and entrepreneurship and. Yeah. So wherever there is a chance to learn and grow, I guess these principles can be. Can come in handy. And I'm very glad that it's resonating on multiple fronts rather than just in the traditional sense of education.
- [45:12]
B
Yeah. Well, in the interest of starting to wrap up here, is there anything else that you want to make sure that people know about? Are there any final kind of thoughts that you want to give to everybody listening?
- [45:25]
A
Yeah, I mean, only that, you know, productive, like the main idea. I just want to reiterate that it's not just another conversation about failure. Like we've always said, failure can be a powerful teacher. It's good to embrace failure. We can learn from failure when it happens. So I don't want people to think productive failure is just more of that, when in fact what it is, is, as I said at the outset, is if that is true. If failure can be powerful, then how do we actually design for it and use it in a way, in a safe way for deep learning? It's like sun has energy. If you only capture that energy when the sun comes out. Well, okay, that's great. Make hay while the sun shines. Right? But what happens if the sun is not there? So can we, can we find a way to actively design tools and technologies to trap the energy and use it as and when we like for our use, for our needs? I think that's the next step. That's where productive failure comes in. So it's taking the conversation forward from that perspective.
- [46:31]
B
Good. So not just learning from your mistakes and your failures. How do you design for failure in a safe way to bring about the deepest learning?
- [46:39]
A
Exactly. And what is the science behind it? What is the architect? What is the design framework for doing things like that and not just waiting for it? Good.
- [46:48]
B
Is there anything else you want to point people toward? Where else should people be looking for, you know, aside from your book, other information from you, should they be following you anywhere?
- [46:57]
A
Yes, you can follow me on my LinkedIn, you know, on. On Instagram and of course on X as well my website, manukapur.com or the books website, which is productive failure.com and if you're interested in a couple of TED talks on failure and learning, then just type productive failure and you'll have the TED talks as well. That may be a good entry into the, into the work and thinking.
- [47:22]
B
Awesome. Thank you for your time.
- [47:24]
A
Thank you so much for having me. I enjoyed this conversation. Thank you.