How can we strategically rethink our basic tools for communication?
Communication is a framework of priority. What’s the most important thing you need to share? How can you learn to share those core, essential, salient points without losing your message in the weeds?
What Ben Whately has found — as a language learning specialist, a memory expert, and founder of the language learning platform Memrise — is that the answer is right there at the tip of your tongue. Here’s how he found his own (and how you might find yours).
Brett: Welcome back to the Founder Vision podcast. Today, I’m speaking with Ben Whately, the Chief Product Officer of Memrise.
How are you doing today, Ben?
Ben: Very good, thanks. Great to be talking.
Brett: Yeah, you too. So tell me a little bit about Memrise. What are you guys doing over there?
Ben: We are helping people learn languages. So far, we’ve got about 50 million users. A couple years back, we won Google’s “App of the Year.” We were the first education app to win it — the first European app, as well — and we’re pretty excited about the mission.
Brett: One of the things I want to dive into today is how tech can change the way we learn and live. First off: People have been using tech to change the way we learn languages for a long time — all the way back from Rosetta Stone and Duolingo. What is it that Memrise is doing differently?
Ben: Actually, if you look at the way we’ve gone about learning languages across the history of humanity: the traditional way to learn languages is to live near somebody who speaks a different language, and to want to trade with them or talk to them or make love to them. From that motivation, you pretty soon work out how to speak their language.
Then there was kind of an unfortunate period in the early modern era, when people wrote textbooks to teach Latin. I call it unfortunate because the original textbooks for teaching Latin were written to teach you to be a really good Latin translator once you already spoke Latin because it was kind of assumed you spoke Latin, because you chatted in it if you were an educated person back then. So they created textbooks in order to help you be a better translator of prose. Then, when Latin died out as a spoken language, we only had those textbooks left, and then people started teaching from those textbooks.
Ben: That is an anathema to the way the human mind actually works. In fact, the only way that you develop an ability in a foreign language is by processing target language input, just above the level you currently understand. You have to be guessing slightly at meaning. Learning grammar structures, what the rules are, learning the translations of individual words, these things in themselves will never get you to the point of actually speaking a language.
In a way, over that period of transition in the early modern era, there was this shift of this construction of an idea of a language as something external to you that you learn, and by learning the language, then you can do and do stuff in Spain or Berlin. But that’s kind of like teaching people the rules of Monopoly — but never letting them actually play a game of Monopoly — and then examining them on whether they know all of the rules of Monopoly. I don’t know if you’ve ever tried to read the rules of board games. I don’t personally because I know I’m not going to understand them. The only way to do is you just start playing and you gradually work it out as you start playing and get a feel for it. That’s how humans learn things. That’s how we approach language learning as well.
That’s the truly traditional way of learning languages — but it’s not the more recently traditional way of teaching languages in order to help you pass exams. To your point, that’s the way the first apps — the first software — went about trying to teach languages: based on the way we teach languages in schools. I don’t know what your experience was like of learning languages in schools, but mine was basically an abject failure.
Brett: Mine, too. My experience learning languages in schools was a total failure, and then I went to Mexico on a trip. I was, like, wow, I want to move to Mexico, so then suddenly I wanted to learn Spanish even though I had taken it for years in middle and high school and I hadn’t learned anything. Then, once I wanted it, I started listening to music.
Ben: Then you could do it, and in fact, it’s even deeper. Right? In one sense, you want to do it and so you do it, but in fact, the pedagogical approach: you pick up grammatical structures and your ability to use grammatical structures in the order in which they can convey meaning for you. As an example of this, when people learn English: in lesson one, they’re taught that, in the present tense, you put an -s on the third person. I eat. He eats. That’s always taught right at the start. English learners don’t do that until they are really sort of upper-intermediate. They are pretty damned good at English before they start doing that, because it doesn’t matter. No one cares. It just doesn’t matter at all. If I say, “The boy, he eat the chocolate,” no one is confused about what is going on.
So: when you actually have meaning to transfer, you care about communicating with someone. That is when you learn, but it is the only time you learn not just because there is sort of a motivational aspect, but because that’s how language learning actually works. Your brain will only do what it is has to do in order to get its meaning understood, and so yeah, when you want to be in Mexico and you want to understand what the songs are saying, when you want to order yourself a beer, that is when you learn how to order yourself a beer.
One of the big issues we have as an app is: if you are in your bar in Tijuana and you want a beer, you have to work out how to order a beer. Then your reward is you get a beer. Whereas if you’re on a mobile app and we’re trying to get you to get good at ordering beer, we can get you to say ‘una cerveza, por favor.’ But all we can give you is a green tick, which is just a whole lot less rewarding than a beer.
Ben: This is part of the crux of the problem for a mobile app as a way of learning languages is that we need to find totally different motivational triggers. We need to understand what the motivations are when you are in that perfect language acquisition environment, which is living in a country where the language is spoken and trying to live your life and trying to get stuff done. That is the perfect environment. What are the motivations there? What can we reproduce on your phone? What can we not?
By understanding those levers, we can actually create. It’s incredibly rich what we can do on a mobile phone in terms of video now — in terms of actually connecting you live with people. We’ve got a huge range of options, but we’ve got to understand what it is that is going to motivate you to engage in those conversations; to try and understand those videos and so on. That’s kind of the heart of what we are.
Brett: That makes a lot of sense. What you’re pointing is the traditional learning of languages, even the way they have been built into apps, is largely intellectual, and language isn’t just an intellectual thing. We actually do it with our body. We develop a muscle memory for the speech patterns we are using, and then we pick up body language responses and backchanneling from the people we are speaking with to constantly gauge how well they are understanding what we are saying or what we think they are understanding of what we are saying.
Ben: An analogy I use for this is this: I’m playing the jazz piano, where there are a few points in there that represent an intellectualized activity and the rest is feeling. That is absolutely right. The way that we tend to think of it is a very kind of left-brain activity if we think of language like a puzzle.
I was taught Latin in school. It literally was like a puzzle. I had never heard it spoken out loud. It was just a puzzle I had to work out the order of the words and then I had to go and look up the declension of masculine nouns and find the right ending for this or that case, and all that. It was simply done like a puzzle, and that’s one part of the way that language works.
But then there’s the other side. When you look at the hemi-spacial specialization in language processing, there is one part that is the direct literal meaning of words and word order that is very much left-hemisphere processed. But then you’ve got the right hemisphere processes, which is a lot more about contextual meaning — a lot more about emotional salience — about the feel of the language. You can’t truly understand the language without having both of these things together. But modern schools only teach the left-brain side of it. This leaves the right-brain side of the equation for those moments when you go out and actually exist in the world and try to tie that intellectual ability back to your real life.
Jazz piano is a great analogy. I was explaining these concepts to a jazz pianist — keeping in mind that I’m terrible at piano and am just relating what he said to me because it was a great analogy. He said the right hand, which is governed by the left brain, plays the melody, which is to say the right notes in the right order. That’s what people think of as the ‘tune.’ But if you just play that, that’s not music. You need to add the left hand to make it music, and that’s where the emotional is. That’s the context for it. It’s all of those parts.
I think there’s a really strong analogy in the way that we produce and process music and the way that we produce and process language. To distill it down, I think the recent traditional school and other courses for teaching languages teach you the right hand. They teach you the melody, but they don’t teach the left hand and to give it the context. We need to bring both of those things together.
Brett: That´s a fascinating metaphor. I had never heard that before. The right hand plays the melody and is connected to the left brain. Language — Broca’s area — is considered to be a language area of your brain, and it’s in the left hemisphere — but a lot of your associative mapping, your semantic web of how the world fits together — for instance, when you’re walking through a bar and your brain is processing people and objects and beers and labels and flavors, it’s doing that by connecting to all of the different parts of your brain and your body that are processing those sensory experiences. They are all a part of what you’re ultimately going to want, and then what you’re going to say.
Ben: Right, absolutely. When you think of just the experience of creating a sentence: As I was saying, the way I learned Latin, and the way a lot of people try to teach languages by translating sentence by sentence — saying, ‘how do you say this in the language’ — that kind of practice is the muscle of having a whole concept or set of ideas that you want to convey and then trying to convey them at once.
The actual lived experience of speaking is that you just don’t know how the sentence is going to end. You just start speaking. It comes from a much more procedural memory of the various chunks that could come next, given what you have said already. You’ve got that feeling of what you want to express, and so you are going to lay those rails in front of you. Sometimes you talk yourself into a dead-end. You are like hold on, that sentence isn’t going to work. I’ve got to slightly back up and go down a different path.
If all you practice is translating a sentence into another language, you aren’t practicing the experience of feeling that you want to express something and trying to lay the track in front of you, and that is actually the necessary and sufficient skill that you need to have. There are just so many areas in language teaching where we traditionally get lost in this loop where we want to know whether the student is able to do something, so we set them a test. We teach them to be able to pass that test. But all of that effort misses the slightly ephemeral, slightly hard-to-grasp concept that what people actually want to do is be able to feel that they want to express something and then pour it out in words — words that point somewhere that the speaker doesn’t quite know — but they go roughly in the right direction, and then they eventually end up in the right place.
Those are just very different skills. Our challenge is to work out the way to train people to do that slightly ephemeral skill, and not to simply do all the things that help them pass a grammar test — not to help them translate individual sentences — and so on. And making that process as efficient as possible.
Brett: And the more emotionally engaged you are in the moment, the more those words have to flow from that place, which is why even in your own language, you can be stupefied if you are in the presence of a romantic interest.
Ben: I am going to riff off that. It is a kind of truism that the best way to learn a language is by having a monoglot lover. I was just riffing off the fact that you were saying that you may be tongue-tied in your own language in the presence of a love interest, but maybe part of the quality of the monoglot lover is that you are kind of forging your abilities in the fire of a difficult situation of talking to somebody you really care about. I don´t know whether there is much mileage in that thought.
Brett: I have had that experience, and it is a fascinating one. It has impacts on both the language learning and on the relationship — — impacts that are very unique as a dynamic.
Ben: What was your experience there? It sounds like there is a story.
Brett: Just a couple of experiences. When I first moved to Mexico, I was like 19 or 20. I met girls there that spoke only Spanish, and I spoke brand-new Spanish. Trying to get to know them was a process. Similarly, in Turkey, I had a couple of experiences there using Google Translate and trying to speak very, very basic Turkish. I studied for a little while. Yeah, I have a little bit of experience with it, but I have never dated a long-term partner that was monoglot in another language. I know some people who have, and it is fascinating.
Ben: I had an experience when I was living in China. One of my daughters spoke Chinese before English, and the other one gradually picked up Chinese pretty quickly. I had the very strange experience that my Chinese was pretty fluent at the time, but I didn’t have any of the language for talking to children and like playing with children.
It was such a weird thing to find this experience where I just couldn’t play children’s games with my child. I thought I was pretty fluent in Chinese, but that area was completely empty. I had nothing. That made it really hard to develop that relationship with my child because I had to keep switching back to English, but then the younger child didn’t speak English. It was just very peculiar: bumping up against the edges of where you think you’re competent in a language.
I didn’t even know a polite childish way to say to do a wee, so when she was asking me about it, I had no idea what she was talking about. It took me a while. She went to the loo eventually, and I was, like, okay — now I get it.
Language learning is so context-specific, as well. When you set that in the context of a specific relationship, that becomes the language that you know. That makes you somewhat unaware, because you don’t keep moving outside that context of how limited your ability in that world is.
Brett: This is a whole other rabbit hole. But I’m curious what it is like to grow up with a father that doesn’t speak your language at the level of speaking to a child as you.
Ben: Both my wife and I are English. Neither of us at the time were speaking the language as well as they were, which was really fascinating. I was really aware of the cultural barrier, and I was also aware that I just didn’t know how I felt about it. I was also aware that I hadn’t empathized with that enough. I knew people of Chinese ethnicity in England, whose parents had moved to England — I just hadn’t really understood that separation for the parents of children being brought up generally in another culture who just don’t have that residue of it. And I felt that happening to myself. It was a very strange experience.
Brett: I can understand now your personal connection to this project.
Ben: Right. Memrise, from my perspective, grew out of the fact that I had a very similarly negative experience to yours of learning languages at school. I just wasn’t very good at it. But then I studied psychology at Oxford, where I met my co-founder. One of the things I noticed — or just troubled me when I was studying how the human brain learns — was that it is a defining feature of humans that we learn languages. If you’re not very good at something that’s a defining feature of your species, it’s sort of hurtful to the ego.
After I left Oxford, I told myself that I was going to go and see if I can learn a language in a context where no one is trying to teach me — if I just go do this in a traditional way, live in China and see if I can make myself understand. Can I make any progress? I moved to China. I moved to a place called Chichi Har, which is in the far northeast corner, on the Siberian border, and then just saw if I could, with minimal teaching, actually find a way to speak the language. What I discovered was yes, by employing memory techniques and just trying to make myself understand, I could actually learn it remarkably quickly to a remarkably good level.
That then led a little time later into thinking: we’ve got to spread this understanding. We have got to reverse this extraordinary situation where people spend years studying languages without ever actually using them. It is like having a school class study the rules of chess, but no one actually plays it.
Brett: Absolutely, so you mentioned using memory techniques, and the company is called Memrise. Is there a connection to that? It’s interesting to me that the name of a language company doesn’t seem to evoke language in the name or the concept of language, but it is more about memory.
Ben: It is kind of about how you use your mind more effectively. Ed, my co-founder and CEO, and I both studied memory.
Ed was a memory grandmaster and got very into competitive memorization. He trained a journalist called Josh Foer to be US Memory Champion in a year. Josh wrote a great book called Moonwalking with Einstein about the experience. We come from that background, but also had a bit of our chip on our shoulder about the way people think about memory.
Through human history, we tend to use analogies for the mind that are linked to the cutting edge of our technology at the time. It used to be clockwork. We used to think of the mind as a sort of clockwork thing. Now we think of it as a computer. We have this model of how human memory works, like files you save on the company and then you can go and open them up, which is just not how human memory works.
Brett: It is more about ecosystem.
Ben: If I ask you to imagine a cow smoking a cigar, you don’t just invent that as a piece of creativity out of nothing. The cow that you are imagining is a cow that is based on a cow that you have seen. It is a memory that is stored somewhere, and the cigar likewise. You are just piecing together memories. In a sense, memory gives us the building blocks that define what we can build with our imagination. But in the popular discourse, we tend to have this dichotomy of memory or creativity: with the implied question, which side do you come down on? We have this viewpoint that creativity and memory are inseparable. In fact, memory itself and memorizing things is an act of creativity. The Memory World Championships are kind of the imagination world championships. It is how you connect meaningless stuff to give it meaning and, therefore, make it memorable.
We were talking a lot about this at the point of founding. Particularly in the light of learning Chinese characters, memory techniques are hugely powerful, but also for the vocab learning aspect of learning a language, which is still useful and a helpful part of it. Memory techniques are hugely important in understanding what’s going to strengthen a memory effectively. These are all kinds of useful things. That was our kind of first step as a product. We probably named the product slightly too much to do with the first part rather than the next horizon, which was already very much in our plans, but we could have made it in our name too.
Brett: I want to get back to something that we were touching on earlier, which is what Memrise does that that incorporates this full, holistic picture of memory learning into an app and into the technology so that people aren’t just getting a dopamine hit of a green checkmark when they get something right, but instead getting something that hits their whole system.
Ben: The motivation has to be actually making progress. That is the only true motivation in learning a language. Green ticks only go so far. As I mentioned earlier, the only way you can acquire an ability in a foreign language is by having target language input just above your current level. On one level, if you just distill down the tasks that we are going after, it is to get learners to just engage in and try and understand content that is just above their current level for a couple of hundred hours and just really focus that. That’s one level of what we are trying to do.
But that begs the question: What things govern what language is just above your current level? The first thing is if you don’t know any words or phrases, there’s going to be very little content you can work with, and the task of trying to interest you in that content is going to be pretty hard. We can help ourselves hugely by teaching you words and phrases very quickly.
That’s where the memory techniques come in, and that’s where we can build up your vocab at a much faster rate. When learning in classes, people learn, long-term, about three to four new words per hour of study time. Compare that to the vocab explosion stage of early childhood, when children learn 10 to 12 words a day. On Memrise, learners over the long term learn about 15 words per hour, so it is way, way faster even than children at their fastest rate of language acquisition. That’s because we are using these memory techniques and applying them effectively.
What we do then is we teach you the words and phrases, sure, but then we have what we call ‘situation tasks.’ You learn the words and phrases, but then you have to go and make sure you can understand them in the real world.
There are comprehension tasks. We call them ‘response tasks,’ and in these you’ve got a video asking you a question and you speak back to it. It keeps pushing you slightly outside your comfort zone. These tasks don’t just say, for example, ‘How do you say one beer, please in Spanish?,’ and require that you answer the question. That’s just a test. Instead, a character says, ‘Que quieres?’ from behind a bar in the video, and you have to think up an answer to it. We might give you a prompt in English saying, ‘You’re thirsty — maybe you feel like a beer,’ and then you have to think: Oh shit, I have got to say that, how do I say it.
It is finding these kinds of experiences that take that knowledge you already have — the left-brain knowledge about the language; how to translate the individual words — and create situations where you need to engage the more right-brained feeling of the context of what’s going on. How do I change that into a desire to express something? That’s the kind of training we are building, and the experience we’re creating.
Built on top of that system is the fact that language, when it’s taught as a thing you need to learn, is taught as a syllabus. You learn certain things, and the all-knowing teacher lays out the order in which you should learn them.
Like your experience in Mexico, that’s not the order in which you will actually pick things up. The order you will pick things up in is the order in which they are interesting to you, and actually, there isn’t very much difference in difficulty between the subjunctive than the present tense. They’re both just chunks of language that you didn’t know and now you know.
The way that we’ve structured our whole content system is that you can learn this in any order, and you can pick up the language that is useful to what you want to do. Whatever it is that you have learned, you will be served with situational tasks based on everything that you have already learned. If you learn a particular phrase and get tested on it and get a situation task for that phrase, the situation task you get will be different from the situation task I get because you will have learned other phrases that make it possible to give you a different situation task.
Brett: That makes a lot of sense now.
Ben: Essentially, we have created a context structure which allows us to lay the rails in front of the learner based on their interest. It is entirely a learner-centric journey of what you’re interested in and what you want to learn, so then we can give you what you need. Whatever it is that you need, we can spin up situation tasks that give you the experience of stretching just beyond your current knowledge of the language and then giving you that experience of applying it.
Brett: Your interest basically navigates the search space of content rather than a set syllabus, which is much better.
Ben: Yes. I think this is a huge problem for EdTech in general, by the way. It’s a problem, and it is a missed opportunity — this kind of paternalistic sense of syllabus creation. It’s only relevant to a very, very few spheres of life.
If I am going to see a doctor, I probably want to know that they have covered the basic syllabus and they know those things. But in most other situations, what you want is the ability to learn rather than already having learned it.
For example: When we’re looking to employ people, we’re not just looking for someone who has done a computer science degree in C because they are really good at programming C. We want to know they were good at learning that, because then they’re going to be able to learn whatever else we need them to learn as they are working it out. I think it applies to languages, and it applies to all realms of education. I think it is something that EdTech has a particular opportunity to change that I think it can change faster than it is.
Brett: I think it really does apply to everything. Yes, I want a doctor to have learned the syllabus of things that they are supposed to learn so that they don´t miss a common thing they could miss if they just learned it through intuition. But I also want them to have a lot of hands-on experience and have a lot of intuition — so that if I’m bleeding out on the table, they know to be checking for the other things that may be going wrong.
The same is even true in language. If I want somebody to write a press release, sure, I want them to have studied to the point that they know the grammatical rules of the business version of the language that they are doing this in — but also they need to have the creativity and everything else required to make it engaging.
Ben: Absolutely. It is an interesting thing that I kind of related to that about, to what you were saying about the context specificness of language as well. You find things in talking to polyglots. One polyglot I know, for example, has an Italian girlfriend. He speaks Italian fluently but cannot read in Italian. He actually hasn’t spoken French since he was at school and finds it very difficult to socialize in French, but can read books in French because he did all of the exams and the grammar in French. Therefore, he just has a much tighter understanding of the structure of the language.
This is the point. It’s not wrong to learn grammar. It’s not unhelpful. But it is not your shortest route to being able to chat with people and to be able to make friends. Just as my children can speak perfectly in English now, but they can’t write a press release very well. They have got a ways to go. They need to learn a little bit more about the structure of language to make them bulletproof; they don’t know grammar rules.
It’s a question of what you need to learn in order to be able to get to the goal that most of our users have: which is socializing in the language. When that is your target, the way you lay out the experience is very different to if your target is passing an exam, or your target is writing a novel in the language. They all require different ways of going about it.
Brett: Right, ultimately we live in a social reality the way that our mammalian brains and nervous systems are constructed.
Ben: Yeah, indeed.
Brett: We are getting close to the end of our time here, so I want to ask just one more question, which is: What is something that you have learned about yourself through this startup journey that has most impacted the business?
Ben: Wow. Definitely, the extent of self-learning is enormous, and it is incredibly frustrating to realize how little attention I paid to things that I kind of knew back when we founded the company. I kind of knew that I was like that, but I didn’t realize the extent to which I was like that, and the extent to which that was going to impact the direction of the company.
My experience of being a founder, which is entirely personal, was that when the company was formed in the beginning stages, I was super intense. I worked all hours and all weekends and expected everyone else to do the same. When things were obviously going badly and we should have given up, it literally never occurred to me to do so. It was just completely relentless, and so we just pushed, pushed, and pushed until the damn thing worked.
As we then started to grow the team, two things happened. One: I knew that I needed to change the way that I was behaving with the team, but I didn’t know quite how to do that. I wasn’t able to give people the space that they needed to do the work at the same time as giving them clear direction on where the work should go.
What I mean by that is I found it hard when I had been working deep in the details, and I had all those details in my mind, and I understand the problem from all sides, and I had total control of it — as I stepped back and allowed somebody else to start dealing with those details, what I found was the need to distill out all of the details which were the important directional, structural pieces — the principles of what we were in the process of building that I really care about, and communicate those, and let people fill in and work out the other parts for themselves. I found that I didn’t consciously distinguish between the really big, important principle pieces and the details, and so I would just splurge everything at people and expect them to understand everything I was saying.
The other thing was that I thought because my team didn’t question me, that they had understood me. When we were a tiny team, we were all totally equal and pitching in. If people thought I was talking shit, they told me I was talking shit — but at some point in our growth, people weren’t working with me as closely and didn’t know me as well. They stopped telling me. Then I would give them a download of detail and they would nod, and then they would go away and do something that wasn´t in line.
Back to your question of what I learned about myself: just learning the extent to which I am not clear in explaining myself naturally. Understanding both the whole gamut of the detail, and the whole project in meta, is the way that I get comfortable with what’s going on. In order to actually explain that to somebody else, I need to distill that down to some basic principles that are the most important parts and then leaves the other bits to them.
Brett: That´s a beautiful reflection. I think it even relates to language learning, too.
A lot of people’s experience in learning a language is that so much is thrown at them, and they are overwhelmed. That rhymes with the experience of an employee in a company where there’s a founder who has the whole vision in their head. The employee is trying to download it all the time.
As you said, as the company grows beyond those people that are close enough to the founder to be a part of that vision, then it grows beyond that and there are now people who don’t have as much direct access to you — but the times during which that they do have access to you, they’re overwhelmed. They could take that to mean that they are incompetent and not able to understand, and then they might not ask questions because they might feel stupid. Then things go in a different direction than intended.
Ben: And if I don’t know which are the important principles that I want to communicate and which are detailed ideas that I don’t feel very strongly about, those will be taken as the same thing. If someone has contact with the founder who just says ‘I don’t think that bit, why is that bit like that,’ that feedback is taken with the same gravity as when I talk about something that is absolutely core strategy, because I haven’t been explicit that they are different. That has been something that I feel like I am still learning.
Brett: I am still learning that, too. I have a lot more to learn in that particular area.
Ben: There is always another order of magnitude of improvement there, I think, in clarity.
Brett: It is asymptotic. Ben, thank you very much. This has been a really, really riveting conversation.
Ben: Very good to chat. I’ve enjoyed it.
Brett: I really appreciate it. Take care.
References
Foer, Joshua. (2012) Moonwalking with Einstein: The Art and Science of Remembering Everything https://www.penguinrandomhouse.ca/books/301277/moonwalking-with-einstein-by-joshua-foer/9780143120537
