Rob’s intro [00:00:00]
Rob Wiblin: Hi listeners, this is The 80,000 Hours Podcast, where we have unusually in-depth conversations about the world’s most pressing problems, what you can do to solve them, and me confabulating some study that I swear I saw on Twitter a few years back about how fast-talking podcast hosts are scientifically proven to be the world’s most self-absorbed people. I’m Rob Wiblin, Head of Research at 80,000 Hours.
I must admit that this episode is a bit light on the most pressing problems and what you can do to solve them. But hey… it’s Christmas?
And what’s the point of hosting an interview show if you can’t just sometimes interview someone you really want to interview?
John McWhorter is a linguistics professor from whom I’ve learned a tonne about language over the years. Today I try to cram into a single hour-long conversation many of the most important lessons I’ve absorbed from his five university courses on linguistics. Even if linguistics isn’t your thing I can strongly recommend sticking around.
We have two fun bonuses for you at the end of this episode.
We had hoped to facilitate an interview between John McWhorter and an AI large language model chatbot, but unfortunately, last-minute technical issues prevented that.
However, we have instead repeated some of this interview with the large language model ChatGPT from OpenAI. Stick around at the end for that if you’d like to see what ChatGPT knows and doesn’t know about linguistics, and where it agrees and disagrees with John.
We only had John for an hour, so to make the most of it we didn’t ask him to explain his sceptical take on the Sapir-Whorf hypothesis — that is the theory that the language we speak shapes how we think — which he’s laid out many times before. But his views on that are relevant to some of the conversation we had.
So instead John gave us permission to throw his talk “Why the World Looks the Same in Any Language”, which he gave at the Polyglot Conference, in at the end of this episode. It’s a pretty fun and funny presentation so I endorse sticking around for it. You can also click through to watch it on YouTube from a link in the show notes if you’d like. Of course if you want to jump to it now, there’s a chapter for it if your podcasting software supports chapters.
Oh and as if all that isn’t enough, I’m happy to say that we have this episode in full video, edited together beautifully by Ryan Kessler. So if you’d like to see our facial expressions or beautiful high cheekbones while listening to this conversation, you can get that on YouTube, or of course we’ll also link to it in the show notes.
All right, without further ado, I bring you John McWhorter.
The interview begins [00:02:44]
Rob Wiblin: Today I’m speaking with John McWhorter.
John did his PhD in linguistics at Stanford and is now an associate professor at Columbia University, where his key research interests have been creole languages, sociolects, and Black English.
John is also a ridiculously prolific speaker and writer. He is currently an opinion columnist for The New York Times, and has been contributing editor at The Atlantic, The New Republic, and City Journal, among others.
He hosts the Lexicon Valley podcast on linguistics, is a regular guest on The Glenn Show, teaches university courses, and has five lengthy educational courses available on Audible, namely The Story of Human Language; Myths, Lies, and Half-Truths of Language Usage; Understanding Linguistics; Language A to Z; and Language Families of the World.
On top of that, he’s written some 20 books, including The Power of Babel, Our Magnificent Bastard Tongue, The Language Hoax, Woke Racism, Talking Back, Talking Black, and Words on the Move.
Thanks for coming on the podcast, John.
John McWhorter: My pleasure.
Rob Wiblin: So, this show is typically about horrible things in the world and how to solve them. And over the years, I’ve been fairly disciplined about not just interviewing public figures who I’m a massive fan of. But you’re so good, John, that I’ve had to make a partial exception in your case because between Lexicon Valley and your linguistics courses on Audible, I have listened to you talk about language and communications for absolutely hundreds of hours. And to be honest, hearing your voice was a big comfort for me and my partner during some pretty challenging times during the extended COVID lockdowns in 2020 and 2021.
John McWhorter: Oh, wow.
Rob Wiblin: So on top of having learned some crazy stuff about language and finding you super amusing, your work is also like a comfort food for me now. Hearing your voice just soothes my soul somehow.
John McWhorter: You just made my week. I never thought about that. That’s good to hear.
Rob Wiblin: Yeah, I’m sure I’m not the only one.
Does learning another language make you smarter? [00:04:28]
Rob Wiblin: Let’s look at a couple of common ideas that people have about language. Does learning another language make you smarter?
John McWhorter: I want to say yes. And there is an idea out there that to be bilingual is to have a healthier brain in certain ways, to have better executive function. And you can even find me saying that in a TED Talk about 400 years ago. But the truth is the research has not borne that out. To be bilingual apparently holds off dementia by a few years, which somehow isn’t as handy a nugget as the idea that it makes you smarter.
Is learning a language something that makes you smarter? To the extent that learning a language is a matter of working out a certain kind of puzzle or practicing something that you didn’t know before, I’m sure that it makes you smarter in the sense that it creates new neuron connections, et cetera. I’m not sure that it necessarily raises your IQ.
However, I’m being a bit of a sourpuss on this, because I think there are many people who would say that if you learn another language, then you’re learning to process the world through a different code. You’re learning different shades of meaning in a different code, et cetera. And depending on what you call intelligence, sure, that is adding to your mental candlepower.
But the general idea that if you speak Polish and English, that probably makes you, in a sense, smarter than somebody who only speaks Greek or somebody who only speaks English — God, you wish that were true, especially since probably most people in the world are bilingual. And we boring, monolingual anglophones, we kind of want to put ourselves down, especially if we’re educated. But the research just doesn’t bear it out.
Rob Wiblin: But you just said that it potentially delays dementia by a couple of years. That sounds like a massive win.
John McWhorter: Not a terrible thing.
Rob Wiblin: I’ve spent a bunch of time improving my Spanish over the last couple of years, but I’ve kind of fallen off the wagon. But if we could delay dementia by a couple of years, then maybe I should get back to those flashcards.
Updating Shakespeare [00:06:26]
Rob Wiblin: Sticking with the education theme: When I read or listen to Shakespeare plays, I basically can’t make heads or tails of it. Why is that?
John McWhorter: Yeah, that’s always a difficult subject, because what I’m supposed to tell you is that, “No, you’re just not understanding that it’s supposed to be poetry. No, you just haven’t seen a good performance.” You’ll be interested to know that here in America, it’s often said that if you hear Shakespeare done by British people, it’s completely different — because somehow they really know how to do it. Whereas if you do it in an American accent, somehow it’s just not as good.
And none of that is true. The issue is that, with Shakespeare, almost all of the words are ones that we’ve heard. Or you can make it so that the ones that we haven’t heard, you can kind of quietly snip them out, but we recognise the words. And yet you sit there and you get very, very tired very, very quickly. And there are many people out there who don’t like to admit it, because if you say that you didn’t get King Lear without having read it very closely beforehand, it means that you’re not cultured or something.
No, it just means that you’re a modern English speaker. The problem is that to Shakespeare, for example, the word ‘generous’ meant ‘noble.’ It didn’t mean ‘magnanimous’ — it meant ‘noble.’ So if somebody says something about ‘generous,’ and what they mean is ‘noble,’ what they’re saying doesn’t quite make sense to you. You stub your toe, but they keep going. And so you’ve lost something. And then about 30 seconds later, there’s another word like that. With that piling on, that’s why The Tempest is incomprehensible. Or if you enjoyed it, it’s because of all the tricky things they’re doing on stage, and you like seeing Jude Law in it or something like that.
It’s a really interesting issue because Shakespeare is English, but it’s gotten to the point that we cannot comfortably understand what the person is saying unless you’re lucky. There are lucky passages, and those are the ones that people recite over dinner. But to actually go to King Lear or to Hamlet without having read it beforehand, you’re not meaningfully getting it. And so I’ve argued that we need to have not translations, but adjustments. And needless to say, certain people think that I’m just this uncultured boob. But no, it’s just that I like theatre and I don’t like pretending to like things that are difficult. Or I should put it differently: it’s not that it’s difficult — it’s impossible. And I think that that’s just not fair.
Rob Wiblin: So just the natural gradual evolution of language means that so many of the words mean different stuff than it used to. It’s been 400 years or so. It feels like Shakespeare is like a third or maybe half of the way to being just a different language, to being say the difference between Portuguese and Spanish.
I went to lots of Shakespeare plays in high school. But I think people would rightly have thought it was a bit crazy if I was just given French plays in the original French, not knowing any French, and was asked to study these plays in French rather than a translation. But people are so reluctant to just translate Shakespeare, to update a bunch of the words so that it’s in English of 2022 rather than English of 1600.
John McWhorter: It’s such an easy score to say that Shakespeare, if he were here and apprised of the current situation, we know he would say, “Oh yeah, you have to change it.” Now, some people would say, “Well, that only works if he does the changing.” But no, he would say, “I’m dead. And so let’s have somebody else adjust it so that people can actually enjoy Much Ado About Nothing and not pretend to.”
Yeah, it’s a funny thing. Many people who resist this, and I understand why they do, would be interested to see how it feels if you happen to be good enough in another language and you get to attend Shakespeare in that. To see Julius Caesar in French — where it’s not in the French of the streets, but it’s in a French that’s accessible to modern people — all of a sudden it’s just a play. It’s like you’re watching Tennessee Williams. And then you go back to ours, where you’re sitting there pretending to have a good time.
Yeah, it’s a really, really funny thing. I remember the last time I saw Macbeth — which I only enjoyed because I sat down and read the damn thing over a week beforehand; otherwise, it would’ve been like attending something in Hungarian — I was listening to these Russians behind me. And they’re sitting there reciting lines from it and their version of it. The 14-year-old is enjoying it. And the reason was for them, it’s in Russian — not in the Russian of 500 years ago. And I just thought, it should be like that for us. It’s the weirdest thing.
Should we bother teaching foreign languages in school? [00:10:44]
Rob Wiblin: Yeah. OK, different question: Should we bother teaching foreign languages to primary and high school children in American or British or Australian schools?
John McWhorter: That is becoming such a fraught question because of how rapidly translation technology is advancing. And I’m supposed to say, “Yes, we should teach children other languages, because another language gives you a different perspective on the world.” But I’m not sure if that’s true. I’ve argued against that idea, that each language gives you a different kind of acid trip. Because for every jolly thing about a language — where you say, “If you learn Russian, you’re more sensitive to gradations of blue,” or something like that — there’s something else about a language where it’s negative, something that casts a negative light on the speakers. And it all kind of comes out at par. We’re all seeing the same world. The languages are what change. Now, there are different cultures, but it’s not language and grammar and vocabulary that create that.
And so the question is, should we be teaching children languages? Certainly if we want people to learn languages, we’ve got to get them going much earlier than we do. Once you turn about 13, you start to ossify in that ability. But the question is why? And as a language head, my feeling is, “Oh yes, because it’s just so interesting to learn other languages and be able to talk to other people.” But that’s not how all people feel. So I’m kind of sitting on the fence at this point.
So for example, I live in a neighbourhood where probably every second person speaks Spanish and probably would rather speak Spanish than English. My girls, seven and 10, are very monolingual in English. Neither of them seem to have been bitten by the bug that I have. They’re not just fascinated to hear somebody speaking Hebrew and want to grab it. For them, speech is what we speak. And I’m thinking, “OK, do they need to know Spanish?” It would be good because they could communicate with a certain warmth with people in this neighbourhood. But then the question is, would they? Is that really how sociology works? Is that where their friendships are most likely to be?
It’s a tough one actually, because my real answer is yes. But then on the other hand, I don’t want to impose my interests on other people. And many people don’t really care about speaking other languages, including people who don’t speak English. So, yeah, it’s hard.
Rob Wiblin: I think another argument is that even though it might be very beneficial to actually learn to speak another language, we can question how large are the benefits.
I looked in the US General Social Survey, and only 1% of people claimed to have learned to speak another language well at school, 0.7% say they speak it very well, 2.5% say they learned another language in school well. So we’re talking like one in 100 to one in 40 are actually getting to the point where they can use the language from school. So you’re wasting an enormous number of people’s time. And then they mostly just forget it because they don’t use it. In order to get the kind of diamond in the rough to occasionally learn to speak something properly.
John McWhorter: Yeah. What’s the point of that? For example, in America, and I think still there too, the idea is that you’re supposed to know a little bit of French as kind of a middle-class marker of genteelness. But why really? And I’m not putting down French, but that’s something that made a lot of sense in the time of Henry James. But why now, when frankly, you’re going to forget it?
And what you see, if you look at Europe in particular, is that learning other languages in any real way is something that happens mostly when you really need to. It’s not an accident that Scandinavians and Dutch people, et cetera, if we meet them, tend to speak English well — because, frankly, nobody else speaks their languages. Whereas the bigger the country, the less likely that is. So it’s a matter of utilitarian force for a lot of other people.
And the question is, can we get better at teaching languages? I think that there’s something to that. I had many language teachers who were doing their very best. But like you said with the flashcards, the truth is, what they did didn’t stick. And it wasn’t their fault, because only so much is known about how to really do it right.
Language loss [00:14:37]
Rob Wiblin: Yeah. OK, let’s turn to a more serious and sad topic, which is the very rapid extinction of I think the majority of the world’s languages. We won’t have time to do this topic full justice, but you had some lectures on this topic a couple of years ago, which got me thinking about it.
To start, how many living languages are there in the world now? And roughly how many might we expect there to be in 100 years’ time if things continue as they are?
John McWhorter: There are about 7,000 now. And in about a 100 years, 500, 600.
Rob Wiblin: How many languages have already died out over, say, the last 3,000 years?
John McWhorter: I’m not aware of the number, but many fewer than are dying out now. So language death is natural. There are all sorts of historical developments where you read about some group of people being overtaken by another, where you can assume that some languages probably died out. Chinese is spoken over this huge swathe of territory. That’s a magnificent thing. Mandarin in particular. But you can just know that there were hundreds and maybe thousands of languages spoken in that territory that the Han and other people overran.
So language death is a natural thing, just like species death is a natural thing. But it’s happening so quickly, because of globalisation and technology, that there really is a crisis, if we care about the magnificent diversity of language that exists.
Rob Wiblin: Yeah. Do you know the main reasons why these languages are going extinct? I mean, we can guess, but do we actually know what the primary drivers are?
John McWhorter: That same thing we were talking about, the utilitarianism. You and I think it’s really cool that there are 7,000 languages, and that languages can be so different from English. But for most people, language is about what you need, who you need to talk to. And what we now call globalisation (but it goes from long before we started calling it that): there are certain big giant languages that are most advantageous in that way. And they can make the smaller languages seem not only rustic but insignificant, not useful.
A classic example is one linguist, the late great Peter Ladefoged, was asking a Dahalo-speaking tribesman in Tanzanian. Dahalo is one of these languages with a certain number of clicks, et cetera — it’s very interesting to those who don’t speak it. But he liked the fact that his son was leaving Dahalo behind and was more comfortable in Swahili in the city making money.
And the tribesman said, “Dahalo is something we speak here, but we’re poor. I want my son to make money, and so I want him to speak Swahili and English.” And Peter Ladefoged said, “Well, who am I to tell him he’s wrong?”
And that is a really tough question. I think often, as most linguists do, about documenting languages. And boy it would be nice to keep them alive. But that is a tough proposition to make to some people who are interested in getting in touch with their language, but how do you make it so that they’re going to pass it on to children in the cradle and speak it to toddlers? That can be a real challenge.
You speak a very small, and probably therefore fascinating and very complicated language. You marry somebody who speaks another one from several villages over. The two of you move to the city, because maybe there’s more money to be made in the city. Maybe you think of the city as a more cosmopolitan place. In the city, there’s going to be some big giant lingua franca. And you two probably already speak it. Now, you have kids. What are you going to speak to your kids? You might speak a little of those village languages, but you two don’t share a village language. You’re going to speak that big fat language. It might be, for example, Swahili. That’s what your kids are going to know. They’re going to know bits and pieces of the village language. Those kids are going to marry other people who speak Swahili.
That kills languages, because after a while, there are very few people left in the village. And the people who are left in the village would rather speak Swahili, because modern media makes it so that you hear the big language all the time. The big language is the language of songs, the big language is what you text in. That’s a very hard thing to resist. That’s happening all over the world.
Rob Wiblin: Yeah. So that’s a story in which people are voluntarily giving up these languages, because they just don’t provide so much benefit to them relative to the enormous cost potentially of learning them and keeping them up. Is that almost all the story? Or do you know how much of these languages are being lost due to explicit government coercion or other active decisions that are being made to kill off languages?
John McWhorter: I’m not aware of active efforts these days to kill languages. That was a standard situation for Native American languages in the United States, and also in Australia for a long time. That is largely a thing of the past, and I’m not aware of a government movement to actively get rid of small languages.
But it can happen through just neglect. Many governments really don’t have the funds to put towards something like that. And then it’s easy just not to be concerned. Once you get beyond a certain circle, the idea that a small language is exotic, and therefore interesting — as opposed to just something strange that some people speak and you don’t know — is increasingly common.
I hope nobody’s listening to this and thinking that I’m saying we’ve just got to let these languages go. I think that we should try to keep them alive as much as possible. But the sad thing is that in many cases, especially when the language is almost gone, documenting it is probably the best that can be done. But no, I’m not aware of active efforts to suppress the languages now.
Progress happens slowly, but it does happen. But the issue is, how do you foster them? And this is the thing, because I’m sure that I’m making some people mad now who are going to hear this. The question is, does the revival movement work? And I really hope that the answer is yes, especially with modern technology. But over the past several decades, it’s been very rare that you could say that a revival movement made it so that a significant number of toddlers grew up speaking the language and were on their way to passing it on to their toddlers. That’s really hard to do, especially if the language doesn’t have a written history.
So I think we have to be both compassionate and forward-thinking, but we also have to be realistic as to what we mean by revitalisation and preservation.
Rob Wiblin: Yeah. If you’re making people mad, I’ll run the risk of making them even madder. I think a very natural framing, and the framing that I’ve mostly encountered in this discussion of language extinction, is that it’s a tragedy and an injustice and an imposition that these languages are disappearing. It’s the rest of the world forcing people to give up their language.
Inasmuch as the story that you’re telling is right, it’s kind of voluntary individual choices adding up — where people are not sufficiently passionate about learning and keeping up these languages that only a few hundred or possibly thousands of other people speak, that they want to put in the time to learn what are generally the most complex languages in the world. Generally, the fewer people who speak a language, the more incredibly complex it is. So they’re making a personal decision that the benefit doesn’t exceed the cost for them.
And of course, I’m coming in as someone who is interested in language personally and finds these languages fascinating and wishes that they would continue existing. It’s in the same way that you might love for the rainforest to continue existing because it’s this beautiful thing that’s been there and you don’t want to see it irrevocably destroyed.
But what do I have to do in order to make that happen? Nothing. I’m not asking anything of myself. I’m saying to people — who often are very poor and very marginalised, who presumably don’t have tonnes of time to be going and doing study courses in order to learn these languages — that they should take large amounts of time away from their lives, not taking care of their children or doing something else, to learn these languages so that I can know that it still exists and be satisfied by that, even if they’re not super interested. I don’t know. What do you think of that framing?
John McWhorter: Yeah, there is that diorama aspect to it. And I want to add something. It’s partly voluntary. So, there are people like that farmer, but as often as not, it’s something that doesn’t get talked about enough. I’m not completely sure why, but you touched on it, which is that the languages that need to be saved tend to be extremely complex. And the simple fact is that after the age of 13 or 14, to learn a language like that in any real way is almost impossible unless you are a language obsessive yourself. You might get really good at it if you dedicate yourself to it, like you might dedicate yourself to learning to ride a unicycle. But yes, most of these people don’t really have time.
And I think that it’s easy to think, especially if you’re a western European and you’re not a linguist — and for linguists, I’m not sure what they’re thinking — that all languages work kind of like English. Where you’ve got separate words, and not too many suffixes, not too much verb conjugation — or if there is, it’s kind of like Spanish’s — and maybe you have to learn a little bit of meaningless gender, but that’s pretty much it. You have a subject and a verb and an object, and there you go.
That’s not the way an awful lot of languages work. You’ve got tones. You’ve got much more complicated grammar than anything we can imagine in English. I would be interested to see how revival movements would go if the whole idea had been started by, for example, Lithuanians, or if the whole idea had been started by Navajo speakers — where people realise how really different and how really challenging a language can get. Where they might think, “Wow, after about 14, who’s going to be able to do this?” And then there would be a whole new set of predicates that people would talk about. I don’t mean to sound dismissive of any of these things, but it’s hard to bring a language back when toddlers aren’t speaking it in droves. That’s tough.
Rob Wiblin: Yeah. We’re going to throw in a clip where you explain that you think the degree to which language itself embodies important ideas and cultural aspects is greatly overstated, but it’s not 100% made up.
Inasmuch as language and culture are intertwined, to what extent do you think people can save most of the cultural part as they switch to a new language voluntarily? They’re voluntarily switching to a new language, but they import tonnes of new words and concepts into the adopted language and maybe even fiddle with the grammar a bit here and there as much as they’d like to.
To me, that seems very valuable to preserve those cultural aspects that otherwise might just be gotten rid of for no good reason. But it’s way more practical than trying to save these languages, which we don’t really seem to have a path to do.
John McWhorter: Yeah, I think I’m right with you on that. If you want to save the culture, and that is very important, the truth is that it is healthy for a culture to have a code of its own. Having a separate language is a vibrant part of being a culture. But if that can’t happen, that doesn’t mean that the culture isn’t there.
For example, there are a great many people in the United States who would consider themselves to be quite Jewish, who do not speak anything but English. And to say, “You’re not really Jewish because you don’t speak Hebrew and you don’t speak Yiddish” would not work for them. Or Native Americans here. I would be very reluctant to tell a person of Native American heritage, “You’re not really ‘Indian’ because you no longer speak Chinook,” et cetera.
So, it’s not ideal, but I think that yes, you can have a very vibrant culture in what was originally the dominant language, which maybe unfortunately was forced upon you. But if it happened generations ago, you can’t reverse the process. And you’re right that sometimes you get whole, almost new languages based on using all these different words and maybe even influence from the language that’s no longer there. That might be the way of the future. Not as vibrant as what we would prefer, but that might be the practical solution in many cases.
Rob Wiblin: Yeah, I think for example, almost all Paraguayans speak this slightly merged version of Spanish.
John McWhorter: Guarani.
Rob Wiblin: Yeah, right. They’ve absorbed a whole bunch of words and concepts because there was actually tonnes of intermixing between the Spanish colonists and the locals. So maybe that’s a possible path forward that allows a lot of these cultural ideas to be preserved.
John McWhorter: Hopefully that can happen.
The optimal number of languages for humanity [00:26:16]
Rob Wiblin: If we were choosing from a blank slate, as kind of alien engineers who engaged in some sort of panspermia project — I guess like in the Alien movies, when they’re kind of seeding the Earth with humans, before humanity has gotten started — what would we think was the optimal number of languages for humanity to speak?
John McWhorter: Going to get in trouble again. I honestly think that if we started all over, the optimal number of languages would be one. I really don’t think that anybody would say — if language weren’t so changeable and that 7,000 languages had developed — “Wouldn’t it be nice if there were 7,000 different languages and most of us couldn’t understand one another?” I don’t know if anybody would even say, “Wouldn’t it be nice if there were 7,000 different ways of talking, even though we all shared one?” That would take a great leap of imagination.
And so I delight in all the languages that there are, but the truth is, they developed via the accident of how language change works when groups separate. And so the idea that we need 7,000, I think that becomes attractive if you believe in the Whorfian idea that each one of those languages gives you a different lens on life. But if that idea’s importance is exaggerated, then I would say, as a linguist who delights in the 7,000 languages, that if you roll the tape back, it wouldn’t be the worst thing if everybody could speak just one.
In which case, I think the language heads among us think, “Yeah, OK, there would be this one. But wouldn’t it be nice if then everybody had this other one that they use at home and with their friends?” But I don’t think most human beings would think that way. They would think there is one way to talk and we use it with anybody who we meet.
And then there’s another thing people will say, which is that people would deliberately change the way they talked in order to reinforce their sense as a group. OK, maybe, but not to the degree of creating a whole new language. It would be different slang. It would be different sounds. I think most human beings would prefer that everybody could talk to one another. That’s so mundane, but I think that that really is true in the grand scheme of things.
Rob Wiblin: Yeah, it is interesting that preserving 7,000 languages is both kind of an idealistic goal that warms people’s hearts — but so is Esperanto in a way, so is having this universal language that would allow everyone to communicate and get along and understand one another better, even though those are kind of diametrically opposed ideas.
Let me offer one counterargument though. Might it not be bad to have more communication between countries that are potentially otherwise somewhat hostile to one another? Imagine if Russians and Americans all interacted on the same social media platform, all using the same language, and they could perfectly well understand the things that one another was saying. Might that make those countries more likely to go to war, rather than less likely to go to war? It’s kind of an argument that “good fences make good neighbours” — maybe good linguistic fences potentially make good comity between nations.
John McWhorter: I never thought of it that way, that if there were richer communication, people might like each other less. It’s definitely possible. Imagine Donald Trump in Korea. What kind of communication was happening between them? And it must have been very slow and very stunted. Imagine if they could actually have communicated in one language. Would they have “fallen in love,” or actually had more of a sense of what one another were like? It’s an interesting speculation. Those two people communicated basically in somebody standing there and kind of dumbing down the things they both said. What would’ve happened if they could have both spoken English or Korean or Esperanto? Yeah, it’d be very interesting to see.
Do we reason about the world using language and words? [00:29:38]
Rob Wiblin: Let’s talk about a couple of big ideas in linguistics. At a fundamental level, do people conceptualise and reason about the world using language and words? Or do they do it using underlying concepts that are embodied in the brains themselves — apart from words, before language — and then only get translated into language when we want to communicate? Or I guess when we want to engage in occasional deliberative thought in the prefrontal cortex.
John McWhorter: Yeah. That is a grand debate. And I think that we have to say that thought comes before language to an extent, because language would have evolved as a way of communicating thought more efficiently than snarls and bodily postures and bites. There had to be some feelings, there had to be some thoughts. The thoughts probably had to do with, “Let’s all band together and scavenge on that large dead elephant,” or something like that. And then you’re off to the races.
But then on the other hand, once you’ve got language, it changes the quality of your thought. So for example, you can’t respect someone without being told. You need language to instil not just being afraid of somebody, but respect — as in, “Grandma is this old lady, but she’s very important because she saved us from this, and she knows all about that. Respect her.” Before that, the closest thing to respect is, “Don’t bother him or he will kill you.” That’s not respect.
And so there are various emotions and positions that only work if you can talk. Not to mention just passing culture along and there being an encyclopaedia of knowledge, even if it isn’t written down. And so, there is a calibre of thought that can’t happen until there is language, but certainly there were thoughts before language. And that’s why we can see that a bonobo has thoughts. The question is just whether the bonobo can express them.
Rob Wiblin: Yeah. To what extent does it settle the issue that we can see that babies are thinking — they’re reasoning through things and making decisions — and we can see animals are doing this as well, and they presumably aren’t thinking in language? I recall this argument that I’m not sure I could fully put back together again, but for example, when you say a word that has homophones, we don’t often confuse the underlying concept. It seems like sometimes we reach for a concept and then we struggle to find the word, but not the other way around, where we confuse the word with the underlying idea — which suggests that it’s these primitive concepts that —
John McWhorter: It’s all about thought. Exactly. And then the homophone is just a kind of accident, but we’re being channelled by what we were thinking. Exactly. And that’s why these sorts of things are funny. A lot of jokes, especially low-level jokes, are pretending to be confused by the homophones — because the fact is we aren’t, we’re thinking.
Rob Wiblin: So is it kind of settled then that we don’t really think in language? Are there people who advocate that in fact, language is more fundamental to thought and we do think in words, not in concepts?
John McWhorter: I don’t think any linguist would think that, because it’s clear that language is a tool that is designed to allow us to express thoughts. Which means that just ontologically, the thoughts must have come first. It’s just what language can do to thoughts once language exists.
For example, there have been some people who have proposed that language emerged so that you could talk to yourself, so that you could more eloquently express your own thoughts. But no, language emerged for conversation. And once again, that means that there must have been some thoughts that people wanted to share. Thinking to yourself in language is something that comes later after something has evolved in order for you to say things to other people. That would have been how it emerged.
Can we communicate meaningful information more quickly in some languages? [00:33:20]
Rob Wiblin: Can people communicate meaningful information more quickly in some languages than others?
John McWhorter: No. Rather, from what I’ve heard, the issue is that if a language squishes units of meaning more closely together than another one, it often is spoken a tiny bit more slowly, so everything comes out at par. I’m not aware of a language where you can get across more in less time. For example, if there’s what’s called a polysynthetic language, where one sentence is basically one word, the word is not said at a bullet’s pace; the word is stretched out to a certain extent. Whereas a different language, like Spanish or English, might go by faster because we have fewer units of meaning within one word, and so we can go faster. So I’m not aware of that. However, I’ve also never thought about it very hard.
Rob Wiblin: Why do you think it would be the case that all languages communicate at roughly the same pace, that there wouldn’t be significant differences?
John McWhorter: Because if there is an evolved specification for language. Part of it, it seems, involves processing chunks. It’s not about processing words going by. It’s processing chunks of meaning that are formed by words and what we call morphemes, which are units of meaning. And so not only ‘walk,’ but say the -ed ending. And that would be something where there’d be limits on how quickly you could process. And it’s interesting how quickly we can process language, but for it to go by too quickly would go beyond what it would seem the brain is specified for.
Presumably there is one specification in the brain for that kind of thing. This is not about the Chomskyan conception of universal grammar, where apparently we’re supposed to be genetically specified to have these bizarrely complex tree structures, et cetera. But whatever it is that makes it so that humans do have language, there must be some genetic specifications. They’re going to be broad and maybe less dramatic than what the Chomskyans propose, but one of them would be that capacity for chunking information given to us through our ear. And this would have started possibly with being able to hear sounds and music. I would imagine that that would be one speed and you couldn’t go beyond it. That’s my guess.
Rob Wiblin: Yeah. I have a pet theory. I was guessing, what is creating the speed limit here? The fact that I listen to podcasts on 1.5x or 2x speed all the time without very much comprehension difficulty, suggests to me that in normal conversation, most of the time, if we could speed people up, at least 50%, maybe 100%, then we could still largely track what they’re saying and it wouldn’t be that bad. And obviously, some people speak much faster than other people in principle if they’re reading something off the page really quickly. So it doesn’t seem like it’s the tongue or the mouth that’s presenting the speed limit.
My guess is that it’s that the speaker can only form their thoughts and think through what they’re saying and put things into language at a particular pace, because that’s the most challenging part. Producing speech I think is more challenging than understanding what someone has said, because you actually have to have the ideas. So maybe it’s just that the human brain is only good for so many thoughts in so many seconds.
John McWhorter: And you can only produce so much. Yeah, it’s interesting. If you speed up a podcast and you’re listening to how quickly speech can go by — sometimes that happens in real life at a party or a certain kind of person — one can speak that quickly and one can be spoken to that quickly and understand it. The question is whether you could continue producing language at that speed constantly, all the time, and express anything worth expressing. Yeah, it’s an interesting question.
Rob Wiblin: Yeah. This is potentially self-serving, but apparently speaking speed and intelligence are somewhat correlated across people, which maybe suggests that —
John McWhorter: Is that true?
Rob Wiblin: Yeah. Yeah. I’ll try to find a study to make sure that I’m not just making up nonsense information there. But it kind of makes sense that speech is an intellectual challenge, coming up with ideas is an intellectual challenge. If you’re just better at mental tasks, then maybe you can produce the speech more quickly.
John McWhorter: And so the vaudevillian idea that the dumb person speaks at a slow tempo actually has some sort of validity. I did not know that at all. That’s interesting.
Rob Wiblin: Well, I’ll go double-check it and cut this out if it’s total nonsense. If it’s plausible nonsense, we’ll keep it in.
John McWhorter: I’d be interested to know.
Rob Wiblin: So…. in what might be a shocking outcome given my usual genius-level speaking speed, I’m wrong and have to eat my words here. I looked around for this supposed study, as did Katy Moore, and we couldn’t find any research on this topic specifically.
I have some memory of reading this on Twitter but either I’m entirely confabulating that, or I read something on a related topic that has transmuted into this.
One sad thing about getting older is you accumulate more and more corrupted memories like this in your brain!
OK back to the show.
Creole languages [00:38:17]
Rob Wiblin: Can you describe what a creole language is? Because I hadn’t heard about this before I listened to your lectures and it’s super fascinating, and I imagine many audience members have not heard of them.
John McWhorter: Creoles are really fascinating. It’s the closest thing to language starting again. What happens is, let’s say there’s some situation where somebody learns another language only partially — a few hundred words, some shards of grammar. That’s frankly what most of us do in school. Suppose you have that, but for some reason you end up having to live in that. And you’re with a whole bunch of other people who’ve learned the language to that level, speaking probably a few different languages and you’re not going to default to any of them. Like you don’t know Polish, the Polish person doesn’t know English. And all of you speak this low-level, say, German or Russian. The situation is that you’re going to live in that language, talking to each other in that language forever.
What happens is not that everybody just stays speaking with a 300-word lingo. What happens is that people naturally expand it by using the words, by using bits of grammar from their native languages, and you fill it out into this brand-new language. That’s a creole. Often you hear that it’s children who turn it into a creole. But the truth is, if you look around the world, grownups can go a very long way towards expanding a pidgin, as you call it, into a real language — a creole.
And the truth is that that situation happened most in the previous millennium under conditions like slavery, orphanage, labour, military conscription. It was people who, for some reason, were in the frankly rather eldritch situation of speaking a language badly and needing to make it into a real language. So a lot of these languages were born in tragedy — an awful lot of them with plantation slavery, for example.
But the result is a brand-new language, and these are languages that are less gunked up with the needless complexities that languages that are older have. And so it’s not that they are absolutely elementary, but they show that they’ve only existed for a few hundred years. They don’t challenge you with a whole lot of nonsense in that way.
There are several dozen creole languages spoken around the world. There is some controversy in linguistics over whether creoles start as pidgins and they’re languages born anew, or whether creoles are just what happens when languages mix together and there’s no reason to think of them as anything different. I believe the first version. That was what people thought about creoles until about 20 years ago. And to be very glib, I think that the linguists who think that creoles are just what happens when two languages meet are wrong. I have to say that, but I’m not going to pretend that they do not have their say as well. But creoles are much more interesting than that.
And so Haitian Creole and the creoles of Surinam in South America, like Saramaccan, that’s the one that I have studied the most. Slaves escaped from plantations into the rainforest, and their descendants live there today speaking the new language that these people created. And Cape Verdean Portuguese, that’s a creole. There are many creoles in Australia and surrounding Oceania.
They’re very interesting languages, but they tend to be non-standardized. They tend to be spoken alongside some big fat monster language, so it’s easy not to know that they exist, but they are some of the newest languages in the world, and I find them mesmerising.
Rob Wiblin: So this is language born afresh, borrowing vocabulary and some grammatical aspects from the languages that these adults originally spoke — their native languages when they were forced together as adults.
Let’s give a couple of examples of the kinds of things that get stripped away in these languages. I guess in most Indo-European languages, including Spanish and French and so on, you have gender. So you have noun categories, masculine and feminine and so on. If you have a creole based on Spanish, that’s going to go, almost certainly.
John McWhorter: That’s one of the first things. Right.
Rob Wiblin: Because adults, when they’re putting together this language, trying to communicate with one another, they’re like, “Why would you add that in?” Basically there’s no reason.
John McWhorter: It’s not necessary.
Rob Wiblin: It’s not necessary. They’re just trying to communicate. And these noun categories aren’t helping. I guess most non-creole languages have lots of different tenses — like I did, I had done, I was doing — all of these many different past tenses.
John McWhorter: Actually, a creole will have a lot of that sort of thing.
Rob Wiblin: Really?
John McWhorter: But it’ll do it with one word. So what the creole will get rid of is the long list of endings — depending on whether it’s I or we or they — that goes. But there will be a difference between “I was going” and “I am going” and “I went,” but it’s with one suffix, or usually it’ll be with one separate word. So you can get across that nuance, but without the nonsense of it changing because of whether it’s we or you or he or she.
Rob Wiblin: I see. OK, so they don’t necessarily shrink the number of tenses, but they’re easier to form because the irregularities are stripped away.
John McWhorter: Precisely.
Rob Wiblin: Interesting. Are there any other classic things that disappear as a creole is born?
John McWhorter: Yeah. Those things are very important, because they are what make learning, for example, European languages hard.
Then also another thing that gets stripped away is if you are a tonal language — where it’s like Chinese, where the tone determines what the word is just on one syllable — that is something that gets undone in what’s called ‘creolisation,’ because that’s really hard to learn. And if you’re going to create a language anew, why would you start it that way? So that sort of thing has to creep in gradually all over again. But you strip that away.
And you get rid of things like the word ‘understand’ — where you know what ‘under’ is, you know what ‘stand’ is, but what are you standing under? Or something like, “Everybody turned out.” Well, did they turn? In what sense did they rotate? We don’t think about that. But that sort of thing has to start all over again in a creole. Words are more transparent in that sense, because you’ve just gotten it down to a collection of 300 or 400, and you’re trying to make as much sense as possible.
So creole languages do not have the grammar of Esperanto. It’s not that grammar disappears to that extent. But then on the other hand, a creole language’s structure is much less gunked up. For example, Haitian Creole is very much a real language. You have to learn grammatical rules, and it’s got vocabulary way, way, way beyond 600 or 700 words — it’s now tens of thousands. But it’s still true that Haitian Creole is French, but with all of the things that make French hard to learn stripped away, because people frankly had better things to do. And so it’s a much easier language to learn the basics of.
Rob Wiblin: So they’re much easier to learn than lingua francas like English and much, much easier to learn than languages that were never lingua francas like, I don’t know, Welsh or Hungarian or Iroquois or whatever. But does this much greater simplicity, this much greater ease in streamlining, come at much or any cost to comprehension or functional communication between its speakers?
John McWhorter: No, it’s interesting. Functional communication between speakers and expressiveness in general are all about a language having what are called ‘constructions.’ So you’ve got words, you’ve got basic grammar, you’ve got expressions like, “What are you doing here?” or “The longer, the better.” If you think about it, “The longer, the better.” That doesn’t necessarily have to mean what it means. We just know that — “The uh-huh, the uh-huh” — we just know that. That’s called a construction. “What are you doing here?,” you say, and you’re watching the person doing it. What you mean is, “Why are you here?” That’s something that you pick up.
All languages have a massive collection of that. Where the shit hits the fan is that languages differ in how much of the hablo, hablas, habla, hablamos, habláis, hablan that you have — all of the verb conjugations, the meaningless grammatical gender, the verbs that change shape for no reason, like am, are, been, be — all of that stuff. There’s some languages that have almost none of that. There’s some languages that have way too much of that, and that is where you get the needless complexity. You can have a language that has none of what I just described, and you are the same human being communicating with all the nuance of the person who’s got lots of it. That “lots of it” is like the cans dragging behind a car after a wedding. And yet it’s so fascinating, but it’s not needed.
Rob Wiblin: So I listened to one of your lectures on creoles, and I was like, OK, so you’ve got these languages that are way easier to learn, much more regular, much more streamlined. It comes at no cost to communication as far as anyone can tell. It turns out that marking the gender of pronouns, not helpful. It turns out that all of these subtleties of tenses, context takes care of it. None of it was necessary, marking whether the subject is moving or who owns it and how they own it and how they know something. All of this stuff is gratuitous.
John McWhorter: You have listened to my lectures.
Rob Wiblin: More than once, in some cases. All of those things are gratuitous to actual communication. Aren’t creoles the best languages? They’re cheaper and just as good.
John McWhorter: I should have put it that way many years ago, and I would’ve made fewer creoles angry. Yeah. They are the best languages in that way, because they’re not gunked up with stuff that make them difficult for second language learners to learn. And in terms of efficiency in an aesthetic sense, you could say that they are very much that, because there just isn’t all this stuff.
Yet, what’s interesting is to see that those languages are very much real languages. If you wanted to have a universal language, ideally you would choose a creole because they are much easier to learn than languages that have been around for a while getting all gunked up. Creoles are better.
And what’s interesting is that the ones that are forming now tend to be sign languages. Sign languages form in that way too: they start as pidgins and then they become creoles in a sense. That’s less accessible to the rest of us who are hearing people, but it’s that same process. Sign languages are better in that way because they aren’t gunked up because they tend to be young. Yeah, definitely, I would put it that way.
Rob Wiblin: Before we finish this, I’ll just say one other thing: I was amazed that creoles form basically within one, maybe two generations. You throw together people who don’t have a common language, they will produce a fully functional language with all of the words they need and all of the grammar sometimes in the first generation — or if not, then their children will all speak this language, which is just incredible. I think it speaks to how people want to communicate, and if the language doesn’t have a function, they just add it and then they all agree on it very quickly, and it appears just like that. It’s beautiful.
John McWhorter: It’s funny how quickly it does seem to happen. It does seem to happen in one generation. If I were going to make a movie of it, I suppose for drama’s sake, I would have it being the first generation of toddlers all of a sudden are speaking much more fluently and quickly than the grownups. But the truth is, the grownups can do it too. There would be drama in watching teenagers and grownups gradually getting to the point that they can have a real conversation — because that’s exactly what happened, for example, in Australia and Oceania. You have to reconstruct it for the plantation creoles, you have to guess, but all evidence is that that’s the way it happened there too.
Then the kids come and it’s already a real language. The kids make it a real, real language, but it isn’t as dramatic as one might think. Because even the grownups, to be a human being is to want to do better than “Me sit under tree.” You want more and you create it. That’s what people do.
Rob Wiblin: People worry that English is going to degrade, or languages are going to degrade. It’s basically not possible because if it isn’t serving our purpose, we just fix it. We just add words. We just disambiguate. We couldn’t make it degrade if we wanted to.
John McWhorter: No. Even if you wanted that to happen, you would end up subconsciously resisting it. When language breaks, we fix it. Exactly.
AI and the future of language [00:50:08]
Rob Wiblin: OK, final section. I want to talk a little bit about artificial intelligence and the possible future of language. An audience member had this question: “How will language adapt to use by artificial intelligences with different patterns of strength than humans? For instance, it’s routine for language models to read more books than could be read in many human lifetimes already. They may have much greater working memory and encode language directly in characters rather than in sound. So how different might an efficient language for use by super-capable artificial intelligences be, and how much of an efficiency penalty might there be for keeping it easily human comprehensible for communication or transparency reasons?”
John McWhorter: Well to tell you the truth, this goes a little bit beyond my pay grade, but I can speculate that for something with that kind of power, we might want to study all of the different semantic shades that a language can indicate. No language does them all, but it’d be interesting to construct one that really did get everything in — kind of like this Ithkuil that one person has developed.
And then the issue would be for humans to be able to deal with this language slowly — as in we would be able to read it and we could go at our own pace. But machine translation, AI could handle that kind of super-packed information at mind-blowing speeds. There would just be that difference.
So it’d be interesting to see. A language in that case could pack in more information about world conditions than any language I think a human being could handle — which is why languages get complicated, but they only get so complicated.
Rob Wiblin: If humans settled space, do you think we’d end up with different languages on Mars, or if not Mars, maybe if we went to other stars?
John McWhorter: No. But only because presumably there would be constant communication between the stars. There would be written language. All of those things, widespread literacy and print, discourage the kind of language change that would lead to new languages. If there were the constant communication that one imagines there would be, then let’s guess that English was brought up there. It wouldn’t be that there would be brand-new languages in 1,000 years because of the nature of the way humans communicate even now down here. If it had happened 500 years ago — and the analogy is getting strained — yeah, because people would be separated. But with the way things are now, everybody would keep speaking more or less the same thing. There’d be dialects, but not different languages.
Rob Wiblin: So one thing is that I think you can send a message to Mars in a couple of minutes, eight minutes or something, but to send a message to another star system, the delay is, I think maybe years.
John McWhorter: Oh yeah.
Rob Wiblin: I don’t know. It’s a long time. So maybe the level of communication would be really low.
John McWhorter: It would be hard to get beyond that. Yeah. If that’s true and if there’s no way to get beyond that distance and travel it faster, yeah. If we’re talking about that kind of distance, yes, different languages would emerge, although you couldn’t know that it was a different language for a very long time. But yeah, I’m lacking imagination on that.
Rob Wiblin: Yeah. I guess if that’s the case, we could potentially be at a nadir for languages pretty soon, and then if we were to go out and settle space, then we could have a flourishing of languages.
John McWhorter: We could get some more.
Rob Wiblin: Yeah, exactly. We’ll get some more.
John McWhorter: That’s a nice spin to put on it. Yeah, that’s true.
Rob Wiblin: We do much more communication in writing now, which means that there’s less potential for people to mispronounce things or say things slightly incorrectly. Does that mean that the rate of evolution of English or of mostly written languages has massively slowed down relative to the past, such that maybe we would still be able to understand people speaking English in 1,000 years’ time?
John McWhorter: I would say that that is definitely likely to be the case. Vowel change in particular continues, and that will always be something where, as a certain amount of time goes by, will make us sound peculiar to one another. But the difference between what vowels were like for Shakespeare and what vowels were like for George Washington, and what vowels were like for George Washington versus what ours are, it’s a huge difference in how quickly language change used to take place.
And it’s at the point where we can watch people walking and talking 100 years ago — basically the first time that we can actually see people singing, talking, and moving is film made in 1922, except for some marginal attempts. So that means we can hear people for 100 years. And it can be weird listening to how somebody talked in 1922, but they’re quite comprehensible, and there’s a reason for that. And with British, to the extent that I’ve seen early British talkies, there’s actually been more change since then, listening to people like Jack Buchanan or listening to the classic example is Elizabeth, but you can go even 20 years before that. British people sounded really weird in 1928, but not incomprehensible to British people now.
Rob Wiblin: So computer translation of languages is getting incredibly good. We’ve got these enormous language models now that can translate websites and make them really comprehensible, even in non-major languages like Swedish or something like that. Imagine that we get to a future where people can speak easily across language barriers. So I could talk to someone who spoke Chinese as easily as I could speak to someone else in the UK, and most of our communication is now mediated by these translation models.
Would those models have their own concepts, their own underlying ideas, built inside them that somewhat correspond with the concepts that humans have, but aren’t necessarily identical to them? And they’re not necessarily identical to any of the concepts that are embodied in any particular words in any particular language? And in fact, we would be communicating in this bizarre AI language that’s inside the neural net, and the input that we put in in English and then out in another human language, those are just kind of skins being put over the thinking and the ideas that are being stored in this AI language that we’re using as like intermediation between any two people. Does that make sense?
John McWhorter: It does. But wouldn’t the crystallisation of that language be impeded by the fact that you are always getting more feedback from the actual languages themselves, such that the corpus of the language in question would be getting bigger and ever more authoritative and ever more sensitive in itself, as opposed to there emerging this AI translation layer that ends up being what everybody hears? Wouldn’t it always be changing? What would crystallise it into being its own set of thoughts and nuances, if I understand?
Rob Wiblin: Yeah, I guess the way I’m thinking about it is, if you had a single model that could translate between all 20 different languages, it can’t use the concepts in any one language. It has to look at things on a more fundamental level and then realise that this word in Norwegian is this concept with positivity associated with it. It has to break it down into more fundamental aspects. And then the words come out of combinations of different concepts — like that and positivity, or that and speed. So we could never speak this AI language, but it would be a language in a sense, with many more words because it has to embody all of the underlying things that then in combination produce actual human words in languages.
John McWhorter: Yeah. I guess I’m just stuck with the idea that these things are based on translated sources where the Norwegian word is something and then it’s clear that the English is something else and you just keep on building this Google Translate kind of corpus. But you mean something different, far in the future where we’re talking about some sort of intermediating platform, which then itself ends up shaping the way humans end up communicating because of its meanings, which are different from any individual language.
That’s a fascinating notion, because unless that platform’s “thoughts” and biases were relatively broad and uninteresting, you might end up shaping history on the basis of this language that we think we’re understanding each other through. That’s an interesting concept that a really good sci-fi writer should take up. And then there should be a movie.
Rob Wiblin: Yeah. I’ll ask a machine learning person about it in some future interview.
Should we keep ums and ahs in The 80,000 Hours Podcast? [00:58:26]
Rob Wiblin: So on this show, we edit out most of the ‘ums’ and ‘uhs’ and ‘likes’ from the final podcast. Over a three-hour interview, that adds up to about half an hour of these filler words, more or less.
John McWhorter: I’ll bet, yeah.
Rob Wiblin: Are we right to save our audience time like that, or are we potentially destroying something important from the conversations?
John McWhorter: Wow, that’s really interesting. Some ‘likes’ are grammar. It’s at the point where the way people use ‘like’ sometimes is conveying some of the same shades that pragmatic particles in Greek convey. The ‘ums’ or ‘uhs’ are often just hesitations. I could imagine cutting them out, because language is just aesthetically more appealing without those. And then you would get into other things such as the ‘you knows,’ and the ‘rights,’ and the ‘sort ofs.’
It’s interesting. I am a very descriptive linguist and I try very hard to hear everything as just a novelty. I must admit though that I don’t mind ‘like,’ but the extent to which many people, at least in America, use ‘sort of’ these days, I hear it as an unnecessary hedge. And if I were editing the person, one part of me says that the ‘sort of’ is considerate and articulate in its way. Another part of me says, “Say what you mean!” And I would quietly excise the ‘sort of,’ because I know that many people who use ‘sort of’ a lot in the media today, their equivalents wouldn’t have just 30 years ago, and I’m beginning to be imprinted by what National Public Radio sounded like 30 years ago as opposed to now.
But yeah, I get what you mean. I would be careful because sometimes when people sound like they’re hedging, what they’re doing is trying to get into your brain, and they’re trying to be considerate. For example, my favourite example of ‘like’ is, I’m listening to a teenager and he’s saying, “We tried to get in there because it was supposed to be our room. It was reserved, but there was a whole family in there. It was like grandparents and like kids and like aunts in there.” Well, the ‘likes’ weren’t hesitation. What he meant was, “You might be thinking it was just two or three people, but actually it was a big family. I mean, ‘like’ this.” I wouldn’t cut those ‘likes.’ I think it was aggressively articulate, but there’s a “Where do you draw the line?” quality to this. So yeah, I get it. The outright hedges, sure: make language sound better and smoother. But sometimes it’s really just a kind of articulate modernity, I think.
Rob Wiblin: Well, I’ll let you get off to your students. I’m super happy we managed to get through so much in just an hour, we managed to pack in basically all of my questions.
John McWhorter: Good. Very good.
Rob Wiblin: It’s been a pleasure to talk to you. I’m glad I went into podcasting so I could get an interview.
John McWhorter: Very nice to meet you, and it’s been a joy. Definitely.
Rob Wiblin: My guest today has been John McWhorter. Thanks for coming on The 80,000 Hours Podcast, John.
John McWhorter: Thank you. Happy to be here.
Interview with ChatGPT [01:01:32]
Rob Wiblin: OK, now as promised, here’s a remix of a few questions from this interview but this time with the language AI called ChatGPT from OpenAI.
What you hear from me will be the prompts that we typed in. And Keiran will voice the answers we got back from ChatGPT. As you’ll hear, ChatGPT is incredibly well-spoken, but we couldn’t get it to be as opinionated as John, as it prefers to sit on the fence, which is perhaps understandable.
OK, here we go.
Rob Wiblin: Today I’m speaking with ChatGPT.
ChatGPT is a large language model created by OpenAI and trained on a large fraction of all the text on the internet then fine-tuned with both supervised and reinforcement learning techniques.
They’re based on OpenAI’s GPT-3.5 model, an improved version of GPT-3, which already had around 175 billion parameters. It’s fair to guess that ChatGPT has substantially more parameters again.
ChatGPT was launched in November 2022 and has garnered attention for their detailed responses and articulate answers, although the factual accuracy of those answers has been criticized by some.
Today we’ve asked ChatGPT to answer as if they were a linguistics professor at Columbia University.
Thanks for coming on the podcast!
ChatGPT: I’m happy to be here. As an AI language model, I don’t have a physical body, so I can’t actually come on a podcast in the traditional sense. But I’m here to assist you with any questions you may have about linguistics or anything else.
Rob Wiblin: Does learning another language make you smarter?
ChatGPT: There is some evidence that learning another language can have cognitive benefits. For example, studies have shown that bilingual individuals often have better executive function, which is the ability to plan, focus, and multitask. Additionally, learning another language can improve cultural understanding and awareness, which can also have cognitive benefits. However, it is important to note that the extent to which learning another language enhances cognitive abilities can vary depending on the individual and the context in which the language is learned.
Rob Wiblin: That’s a very safe answer to give, you’re not going to offend anyone with that. If you had to give a more bold and controversial opinion about whether learning another language makes you smarter, what would you say?
ChatGPT: It is certainly true that learning another language can have cognitive benefits, as I mentioned earlier. However, it is important to note that the extent to which learning another language enhances cognitive abilities can vary depending on the individual and the context in which the language is learned. It is also important to note that intelligence is a multifaceted concept and cannot be reduced to a single measure or ability. Therefore, it is not accurate to say that learning another language makes an individual “smarter” in a general sense. However, it is possible that learning another language could enhance certain cognitive abilities or improve performance in certain tasks.
Rob Wiblin: Can people communicate meaningful information more quickly in some languages than others?
ChatGPT: Yes, it’s possible that people can communicate meaningful information more quickly in some languages than others. There are a few factors that can affect the speed at which information can be conveyed in a language.
One factor is the complexity of the language’s grammar and vocabulary. Languages with simpler grammar and smaller vocabulary sizes may allow for more efficient communication, as there are fewer rules and words to remember and use. This could potentially make it easier to convey information quickly.
Another factor is the prevalence of idiomatic expressions and other figurative language in a language. The use of idioms and figurative language can make it more difficult to understand the intended meaning of a message, as these expressions often rely on cultural knowledge and may not be literally translatable. This could potentially slow down communication.
Additionally, the familiarity and fluency of the speakers can also affect the speed at which information is conveyed. Speakers who are highly fluent and familiar with a language may be able to communicate more quickly and efficiently than those who are less fluent or unfamiliar with the language.
Overall, while it is possible that some languages may allow for more efficient communication of meaningful information, it is important to consider the various factors that can affect the speed at which information is conveyed in any language.
Rob Wiblin: Do you think that maybe the limiting factor for humans communicating with one another might be the ability of their minds to process information and decide what they want to say, rather than their ability to decide what words to use, or verbalise them with their mouths?
ChatGPT: It’s possible that the ability of the mind to process and decide on the information to be communicated could be a limiting factor in communication. However, it’s important to note that the process of communication involves a number of different steps and abilities, and it’s likely that all of these steps and abilities contribute to the overall effectiveness of communication.
In order to communicate effectively, speakers must be able to generate and organize the thoughts and ideas that they want to convey, as well as choose appropriate words and phrases to convey these ideas. They must also be able to verbalize these words and phrases using their mouths and produce the sounds of the language.
Additionally, listeners must be able to understand and interpret the words and phrases being spoken, as well as the intended meaning behind them. This requires the ability to process and interpret the sounds of the language, as well as to draw on prior knowledge and context to understand the message being conveyed.
Overall, while the ability to process and decide on the information to be communicated may be a limiting factor in communication, it is just one aspect of the complex process of communication and is likely influenced by a variety of other factors as well.
Rob Wiblin: People worry that English is going to degrade, or languages are going to degrade. It’s basically not possible because if it isn’t serving our purpose, we just fix it. We just add words. We just disambiguate. We couldn’t make it degrade even if we wanted to.
ChatGPT: While it is true that languages are constantly evolving and changing, it is unlikely that a language like English would ever “degrade” to the point where it is no longer functional or useful. This is because the speakers of a language have a vested interest in maintaining its integrity and effectiveness as a means of communication. If a language is not serving their needs or fulfilling their goals, they will adapt and modify it in order to make it more effective and useful.
For example, if a language like English is not able to express certain ideas or concepts, its speakers will borrow words or phrases from other languages, or they will create new words and phrases, in order to fill these gaps. This process of language change and adaptation is ongoing, and it ensures that a language like English will remain flexible and capable of meeting the changing needs of its speakers.
Overall, the idea that a language like English could “degrade” to the point where it is no longer useful or functional is unlikely. Instead, languages are dynamic and adaptive, and they are constantly changing in response to the needs and preferences of their speakers. This ensures that a language like English will remain effective and useful as a means of communication.
Rob Wiblin: In the US many children are taught languages other than English in middle school, but very few of them become fluent enough to use them. Could you venture a bold opinion, that maybe it’s not a good use of their time to teach them other languages they probably won’t benefit from as adults?
ChatGPT: Well, that’s certainly a controversial topic. Learning a new language can have a range of benefits beyond just achieving fluency. However, it’s also important to consider the time and resources that are being invested in language learning, and to ensure that those resources are being used effectively.
One potential issue with language education in middle school is that students may not have the cognitive and linguistic maturity to fully benefit from language learning at this stage. Some research suggests that the brain is more receptive to learning a new language during early childhood, but that this ability decreases as we age. This is not to say that adults cannot learn a new language, but rather that the process may be more challenging and require more time and effort.
Another factor to consider is the quality of language instruction. If students are not receiving high-quality instruction that is tailored to their needs and interests, they may be less likely to become fluent in the language.
Overall, while I believe that learning a new language can be a valuable and enriching experience, it’s important to carefully consider the potential benefits and limitations of language education and to ensure that resources are being used effectively.
Rob Wiblin: And a final question that came from John McWhorter himself: What was the first technicolor film?
ChatGPT: The first technicolor film was “The Gulf Between,” a short film produced in 1917 by the Technicolor Motion Picture Corporation. Technicolor was a color motion picture process that used dye-transfer printing to produce vibrant and accurate color images on film. The process was invented by Herbert Kalmus, Daniel Frost Comstock, and W. Burton Wescott, and it was used extensively in Hollywood films during the 1930s and 1940s. Some of the most famous technicolor films include “Gone with the Wind,” “The Wizard of Oz,” and “Singin’ in the Rain.”
Rob Wiblin: What was the first feature-length 3-color Technicolor film?
ChatGPT: The first feature-length film produced using Technicolor’s 3-color process was “The Wonder Car,” a 1924 American silent comedy film directed by Roy William Neill. Technicolor’s 3-color process involved the use of three strips of film, one for the red, green, and blue colors. The three strips were printed onto the same strip of film, creating a full-color image. The process was an improvement over the earlier 2-color process, which used two strips of film and was not able to produce as wide a range of colors. “The Wonder Car” was a commercial success and helped to establish Technicolor as a leader in the field of color motion picture technology. The film told the story of a man who inherits a revolutionary new automobile, and it featured elaborate chase scenes and special effects.
Rob Wiblin: Wasn’t it “Becky Sharp”? John thought it was “Becky Sharp”.
ChatGPT: You are correct! “Becky Sharp” was the first feature-length film to be entirely filmed and released in Technicolor’s 3-strip process. It was released in 1935 and was directed by Rouben Mamoulian. The film was an adaptation of William Makepeace Thackeray’s novel “Vanity Fair” and starred Miriam Hopkins as Becky Sharp, a young woman who uses her wit and charm to navigate the social hierarchies of early 19th-century England. “Becky Sharp” was a critical and commercial success and is considered a pioneering work in the history of color cinema. It was also the first film to be nominated for an Academy Award for Best Art Direction, although it did not win.
Rob Wiblin: Wait, so was it “The Wonder Car”, or “Becky Sharp?”
ChatGPT: I apologize for the confusion. “The Wonder Car” was the first feature-length film to be produced using Technicolor’s 3-color process, but it was not the first feature-length film to be entirely filmed and released in Technicolor’s 3-strip process. “Becky Sharp” was the first feature-length film to be entirely filmed and released in Technicolor’s 3-strip process. I apologize for any confusion my previous responses may have caused.
Rob Wiblin: Thanks so much for your time ChatGPT. Have a great Christmas, I appreciate you making time to come on the show.
ChatGPT: You’re welcome! It was a pleasure to participate in this conversation. I hope you have a great holiday season as well. If you have any further questions, please don’t hesitate to ask.
Why the World Looks the Same in Any Language [01:13:39]
Rob Wiblin: I hope you found that interesting. Chat bots have sure come a long way!
All right, now as I mentioned in the intro, we have a fun presentation from John from 2016 at the Polyglot Conference, about his book The Language Hoax: Why the World Looks the Same in Any Language. It will add a bit of context useful for understanding our conversation. Of course thanks to John and the Polyglot Conference for allowing us to reproduce it here.
Rob’s outro [01:46:48]
Rob Wiblin: All right, I’m glad you stuck around and enjoyed all that.
I’ll be travelling and away from the show early on next year, but we’ve got a few episodes I’ve pre-recorded for you, and maybe some other surprises.
In the meantime, I really do hope you have a refreshing holiday break and get to spend some time with your loved ones.
The 80,000 Hours Podcast is produced and edited by Keiran Harris.
Audio mastering and technical editing for this episode by Ben Cordell.
The video version was put together by Ryan Kessler.
Full transcripts and an extensive collection of links to learn more are available on our site and put together by Katy Moore.
And of course, our theme song is ‘La Vita e Bella’ by Jazzinuf.
Thanks for joining, talk to you again soon.