8 - How Translatable Are Languages?

335 turns

• linkGARY LUPYAN - Yes, that was a very influential paper for me.
• linkHedvig Skirgård - I know!You know, some people read Tolkien every year.I try and read that paper every year.
• linkGARY - Yeah!
• linkHedvig Skirgård - I think it's so good!
• linkGARY - Cool, cool.That's great.
• linkDaniel Midgley - I feel like Ben is fighting his way toward a microphone right now, so that he can call you both NEERRRRDDS.LAUGHTER
• linkDaniel Midgley - THEME MUSIC
• linkDaniel Midgley - Hello, and welcome to this episode of Because Language, a podcast about linguistics, the science of language.I'm Daniel Midgley.And with me today is linguist and morning person Hedvig Skirgård.LAUGHTER
• linkHedvig Skirgård - That is such a bad…This is…listeners, this is an in-joke.I notoriously am really bad with mornings and I'm currently…in my timezone, it's four AM.But actually this is sort of better for me because I know I can go back to bed, so I'm just like, you know.I mean, l'm like in airplane mode.
• linkDaniel Midgley - Exactly.
• linkHedvig Skirgård - It's like I'm out of time.
• linkDaniel Midgley - You're timeless.
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - Ben is taking a well-deserved break from his other life in teaching.He'll be back with us soon.But we're very pleased to have an academic and researcher whose work we really like, and who we've been wanting to talk to for a while.It's Gary Lupyan of the University of Wisconsin at Madison.How's it going, Gary?
• linkGARY - Good, good.Glad to be here.
• linkDaniel Midgley - Thanks for coming on the show.How are things going over there in Wisconsin?
• linkGARY - Well, the university is doing this reopening plan with hybrid instruction.So next week I'm going to be teaching an in-person class.It's a grad seminar, so it's pretty small.The larger classes are all online, and there are multiple betting pools going on about how long in-person instructions are going to last.
• linkHedvig Skirgård - Betting pools?
• linkGARY - Yeah, yeah.
• linkDaniel Midgley - Oh, boy.
• linkHedvig Skirgård - Is that allowed?
• linkDaniel Midgley - This is the darkest timeline.
• linkGARY - Well, not, not…Yeah, yeah.You know, keep people just talking offline.Yeah.About, you know, what's going to happen.So it's, it's a weird time.
• linkDaniel Midgley - So that's…just, just to be clear, that's a that's a betting pool about how long in-person instruction will last.
• linkGARY - Yes.
• linkDaniel Midgley - It's not a betting pool about any particular person getting ill or…
• linkHedvig Skirgård - Catching covid.
• linkGARY - No, no, no.The university is doing what it can.It's a really hard situation.I'm just glad I'm not on the administration side having to figure this stuff out.And I feel so bad for the college students who are, you know, excited to have their college experience.And they're…They've been grouped into quaran-teams, you know, that they can socialise with and it's…yeah, yeah.I mean, I'm personally in a good place.We have small kids.We wouldn't be traveling much now anyway.And so, but, you know, all the people who were looking forward to their study abroad and to their, you know, college experiences…
• linkHedvig Skirgård - Yeah.
• linkGARY - It's sorry to see it.Yeah.
• linkDaniel Midgley - We are going to be talking to Gary about some recent work that he's been a part of about the meanings of words and how that ties into culture.Really interesting work.We're looking forward to chatting about it.
• linkDaniel Midgley - But first, if you are a patron, you have probably already gotten the chance to listen to our bonus episode seven, The Mailbag.In it, we answered such great questions as:how do you communicate expressively if you're wearing a mask?Why is the word CATERPILLAR so long?And what did they call it, if it came up later?And other great questions like that.So if you become a patron at the Listener level, you will get the chance to hear that, and you will also get the chance to support the show, help us keep going, help us keep talking.Thanks to all of our great patrons.You can sign up there at patreon.com/becauselangpod.Thanks to all of our great patrons.
• linkDaniel Midgley - Right.It's time to do some news.First bit is about old mate Steven Pinker.
• linkHedvig Skirgård - PAINED NOISES
• linkDaniel Midgley - I mean, I wasn't going to go here, but since we've devoted an entire episode to this…this thing, whatever this thing is…
• linkHedvig Skirgård - It's fair.
• linkDaniel Midgley - I feel like it's in our line.So last week, people started noticing that they were getting blocked by Harvard cognitive scientist and free speech guy, Steven Pinker.But the strange thing is they hadn't even been interacting with him.And then people started realising that if you tweeted the words "Pinker" and "Epstein", you'd also get blocked almost instantly, and even if you…even if you liked such a tweet, you get blocked.
• linkHedvig Skirgård - Is someone's made a bot?I think that's the answer.
• linkDaniel Midgley - That was a common theory.And that led to other Twitter accounts tweeting things like "Punker Oopstoon", not Pinker Epstein, but just close, just close enough.
• linkHedvig Skirgård - Oh god, yeah.
• linkDaniel Midgley - Okay, here's the surprising thing.Motherboard got in touch with Steven Pinker and he confirmed that it's not a bot.
• linkHedvig Skirgård - Yes, so Motherboard is the hacky podcast from Vice?
• linkDaniel Midgley - Yes.
• linkHedvig Skirgård - It's not a bot!
• linkDaniel Midgley - It wasn't a bot, it was a person.
• linkHedvig Skirgård - He's hired some poor RA or something?
• linkDaniel Midgley - I hope that he's paying them.Yes.Somebody who volunteered.Somebody who said, hey, I'll sit on HootSuite for you.And every time a search comes up or every time a tweet comes up— BOOM —I'll just block them for you.And apparently Steven Pinker was like, yeah, cool.Now let me just say the Epstein connection is relevant because back in 2007, Steven Pinker wrote a document for lawyer Alan Dershowitz that ended up as part of a defense for child sex trafficker and Harvard donor, Jeffrey Epstein.Pinker now says that he regrets the involvement and that he didn't know that it was for Epstein at the time, but you could see how he wouldn't be very happy with people bringing up this association.So what do you think about this?You know, Pinker is a free speech guy— Motherboard says "free speech crusader" —and blocking people?Any takes?
• linkHedvig Skirgård - It's not a smart move.Even if you somehow agree with Pinker and all of this, this is, a…I believe the term is a Barbra Streisand move.
• linkGARY - Yeah.
• linkDaniel Midgley - Tell me about the Streisand effect.
• linkHedvig Skirgård - It's when you try and shut something down, and draw more attention to it.That's basically it, right?
• linkDaniel Midgley - Mhm.I think she filed a lawsuit because someone took pictures of her house and she wanted the pictures down, and it just…it just blew up.So it attracted far more attention than the original photos would have.
• linkGARY - Yeah, it's a bad look for Pinker, especially outsourcing.I think the article said "a colleague" is the word that Pinker used, which is odd.And I I saw something about someone who was one of Pinker's students being blocked.And so kind of tweeting like:That feeling when your former advisor blocks you on Twitter.
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - Geez.
• linkGARY - Yeah, so probably didn't mean to do that, but yeah.
• linkDaniel Midgley - Let me just be clear.I think blocking people is fine, not because I don't want to hear views that I don't agree with.I mean, I like having interesting disagreements, but then, I'm not a big free speech proponent who thinks that removing someone from a list of experts is up there with the destruction of the Library of Alexandria, okay?
• linkHedvig Skirgård - Yeah, it also might not improve his…he's been mentioning, with this LSA letter that there's so many people in the signatures that he doesn't recognise.Like "I don't even recognise the names." Well, you're going to recognise a lot fewer now.You're basically just, like, shutting yourself out.
• linkDaniel Midgley - Would you say that Steven Pinker is creating an echo chamber?
• linkHedvig Skirgård - Is that how we use the term?Yeah, maybe?Yeah.Yeah!
• linkDaniel Midgley - Or perhaps he's just trying to create a safe space for himself online.And if that means canceling a few people, if that means de-platforming a few people, then that's his choice.
• linkHedvig Skirgård - Yeah?There's a really good paper by Kastner et al.that's been circulating called, "Who Speaks for Us?Lessons from the Pinker Letter" where they're not talking about the original content as much, but like, how the discourse has been changed, and what Pinker and associates have done to discredit and shape the discourse.It's very interesting.It's actually well written as well, which is not something all academic articles are.
• linkDaniel Midgley - I thought it was very good.We'll put a link on our blog:becauselanguage.com.Also, I propose that we change the name of "the Streisand effect" to "the Pinker-Epstein effect".
• linkHedvig Skirgård - Yeah, yeah, yeah.
• linkDaniel Midgley - But one thing I will say, I admire the PR media prowess that no doubt landed Pinker on the list of experts at handling the media.
• linkDaniel Midgley - LAUGHTER
• linkDaniel Midgley - Okay, let's go on to the next one.Twitter seems to be pushing automatic translation.Hedvig, you noticed this one.What's going on?
• linkHedvig Skirgård - Yeah.So I don't know if this is something that happens to you as well, Gary, but like on Facebook, for example, when my friends make posts in not English, sometimes they just appear automatically translated without me having to push a button.
• linkGARY - Mhm.
• linkHedvig Skirgård - And Twitter is trying this out as well.And a lot of people are upset and not pleased with it.I'm upset because I don't like things to be monolingual, but also because it's sometimes really bad, and it's bad at detecting what language it is.
• linkGARY - Yeah, oh yeah.
• linkHedvig Skirgård - So I have friends who speak someone and write Samoan and write in Samoan on Facebook, and Facebook will often try to desperately translate the text from Italian, or from Portuguese.
• linkDaniel Midgley - Oh, lovely.
• linkHedvig Skirgård - And like, yeah, this is just gibberish.So we'll see if it turns up…it hasn't turned up on Twitter.I think they're only testing it on some people.Hasn't turned up for me yet.
• linkGARY - I…yeah, I haven't seen it on Twitter, but on Facebook, I guess there's an option to stop it from translating, when the original is in a certain language.But I've only seen it now.And the worrying cases are when I don't realise it's a translation.So it's easy to miss the "See Original".So I speak Russian and so I get…for some reason, they've been auto-translating Russian into English for me and some of the translation, especially when it's just a sentence, it's a perfectly fine sentence.And so you don't realise— there's nothing that gives it away —that it's a translation, that it's an auto-translation.And often they are, at least from Russian, they are quite good.It doesn't get the slang right and the sarcasm.And, you know, sometimes you get the opposite meaning, but yeah, yeah.And I am I'm sure there are lots of other people who don't even realise that they're seeing a translation because it's not very apparent in the interface.
• linkDaniel Midgley - So in other words, this would be problematic even if the translation were perfect, because this is not the person's original words.
• linkGARY - Yeah, yeah, absolutely.
• linkHedvig Skirgård - And also so that you know what to respond, right?So you know what language to respond in.I wonder, I can't help but wonder like who this is for, because people who subscribe to accounts that have posts in not English are presumably doing so, knowing that.
• linkDaniel Midgley - Yeah, who asked for this?
• linkHedvig Skirgård - Yeah, who needs this?Is it for retweets?When retweets occur in someone else's feed?
• linkDaniel Midgley - That's a problem because I want to say my own words.I don't want Twitter's translation to be representing me.
• linkHedvig Skirgård - Exactly.
• linkDaniel Midgley - Ugh.
• linkHedvig Skirgård - Ugh.We'll see.
• linkDaniel Midgley - We'll keep an eye on that one.I was kind of excited to see this one.This is some work from Rishi Rajalingham and a team from MIT.This work was published in Nature Communications.So way back in 2012, during the Talk the Talk days, we did an episode "Words with Baboons".It was episode 69 and there was an experiment with baboons.On a computer screen, they would look at a sequence of four letters, and the baboons would have to decide whether what they were seeing, the sequence of letters that they were seeing was an actual English word or a non-word.Which is a weird experiment, but okay, go with me.So when a real word appeared, they would push a green circle.And if it was a fake word, it was a blue plus sign, just to see if they could do it.And eventually, after enough iterations of this experiment, they started getting it right.They could tell the words from the non-words.But the really cool part was that after a while they could tell words from non-words, even when they hadn't seen those words or non-words before.They could do it for new sequences of letters.
• linkHedvig Skirgård - So they learned that the combination, for example, CZ, never really occurs in English.
• linkDaniel Midgley - Except the word CZAR.
• linkHedvig Skirgård - Is that spelled out that way in English usually?Okay.
• linkDaniel Midgley - It's spelled a good many ways.
• linkHedvig Skirgård - Okay, okay, listen…!
• linkDaniel Midgley - Bad example, bad example.
• linkHedvig Skirgård - But, but they learned that a certain combination of letters, because it was letters, correct?
• linkDaniel Midgley - Yes.
• linkHedvig Skirgård - They learn a certain combination of letters.They probably just learned like a Pavlovian effect to that, right?
• linkDaniel Midgley - That's it, yeah.
• linkGARY - Yeah, so it's more sophisticated than that.It's a form of just statistical learning, and most learning is statistical.And so there are enough regularities in the visual input that without knowing that these are letters or that they make sounds or that the words have meanings, there's enough signal there in just the visual forms to, yeah, to be able to tell…It's a common…it's a really odd task that has a very long history in psycholinguistics— lexical decision task —where, because you don't need to know the meaning of the word to respond, you can test people with nonsense words to look at effects of, for example, word length or certain letter combinations that make it look more word-like or less words-like to see how it influences…in humans, it's usually the time that it takes them to respond to whether it's a word or non-word.Lots of priming studies use lexical decision tasks.
• linkDaniel Midgley - Well, at the time, there was a researcher, Stanislas Dehaene, who found out that when we as humans look at words, and then we look at things that aren't words, the words are picked up by the left visual word form area, the VFWA, and they wanted to know if non-human primates also have a part of their brain that activates when they're looking at words or non-words.That's what they wanted to do.They said, well, we want to do this in the future.Well, now it's the future.And they've done it!
• linkHedvig Skirgård - It's not the future we wanted, but it is the one we have?Question mark?Maybe not deserved, but yes.
• linkDaniel Midgley - Here we are!
• linkHedvig Skirgård - Mhm.
• linkDaniel Midgley - Well, it turns out that there is!In non-human primates, there's a part of the brain called the infero-temporal cortex, which can tell words from non-words.How about that?
• linkHedvig Skirgård - That is fun.I particularly enjoy…I've just been writing my PhD thesis and you get to write a lot of like "in future studies, I'd like to address" because you find out all the things you've done wrong.And it's nice to know that sometimes people do do their future studies!
• linkDaniel Midgley - Instead of saying, I am never having anything to do with this again.Blecch!
• linkHedvig Skirgård - Yeah!
• linkDaniel Midgley - Okay, what does this mean, though, that…Primates don't read, but they have brains that respond to words versus non words.What's going on?
• linkHedvig Skirgård - Isn't reading just hijacking of, like, a pattern recognition center that we usually use for like faces and stuff otherwise…?she says, not knowing things.
• linkGARY - So there's been a long standing controversy about…So people are on board with the idea that there is a kind of recycling.Neuronal recycling is responsible for making us expert readers because obviously, unlike faces which have been in our evolutionary history for a long time, there's good reason to think there are evolved mechanisms for recognising faces.But that's obviously not true for reading.But when you look at the kind of specialisation, neuronal specialisation that happens for reading, it's the same type of specialisation as for faces, in that you seem to have reliably similar areas responding in the same way, specifically to letters and words of one's…of a language that a person can read.And so if we didn't know any better, we would have assumed that, yeah, this is an evolved kind of thing.Right?Which because we know the history of reading, we can rule out.And so there's been, and Stan Dehaene has been one of the key players in this debate, about the extent to which the specialisation is really for reading versus, for example, reading tapping into mechanisms, visual mechanisms that are used to distinguish visually similar objects.So, for example, faces, but also things like car makes, or for people who herd sheep and can tell their sheep apart.People have argued that, well, you know, that uses some of the same mechanisms.And so the debate has been:Is the visual word form area that Dehaene's team identified a long time ago, is that really about reading, or is that showing kind of more general mechanism that just happens to be primarily used in the course of reading in, obviously in literate people?So that's been the kind of a debate going back and forth.
• linkHedvig Skirgård - Yeah, that's really cool.Thanks, Gary.
• linkDaniel Midgley - So it sounds like what's going on then, is that evolution just picks up whatever it can and reuses it.
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - So reading piggybacked on more general processes.Would we kind of agree that that's kind of what language did, too?
• linkGARY - Yeah, I think, you know.
• linkHedvig Skirgård - Oh…LAUGHINGYeah, that's my laughing.That's my nervous laughing!
• linkGARY - So I saw Stan Dehaene at a debate some years ago, "neurobiology of language" with Cathy Pryce.And they've had this ongoing debate about how, whether the word form area is just for reading, which is Dehaene's point, or is more general.And something happened that I wish would happen more at conferences, which is:they have this debate format and they gave fairly short talks, I think 20 minutes each.And then most of the rest of the time was spent audience asking questions.And someone asked a question to both of them:"What evidence would you like to sweep under the rug?"LAUGHTER
• linkGARY - And to their credit, they…you know, they didn't do the politician thing.They really answered the question, you know what evidence is inconvenient for their perspective?And I have so much respect for that.
• linkHedvig Skirgård - Yeah.
• linkGARY - Because you know it and, you know, maybe you, you know, sometimes don't cite it in your papers or you hope the reviewers won't force you to deal with it or, you know.But when asked, you answer.
• linkHedvig Skirgård - Yeah, that's super cool.
• linkDaniel Midgley - Let's move on to our last news story.This one, this is kind of a sad one about Scots Wikipedia.
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - Robyn Speer, who has been on a show with us before, she noticed a post on Reddit where it seems that almost the entirety of Scots Wikipedia has just been created out of thin air by a US teenager— possibly a teenager —who doesn't actually speak Scots.This one was good because it forced me to come to grips with what's the difference between Scots and Scottish English.It seems that Scots is its own language.It came about at the time that Middle English was going on.It became its own language, Modern English became another.Then there was Scottish English, which happened when speakers of Scots started having more contact with speakers of Modern English, after the 1600s.Then Scottish English became a thing.So Scots is a language.Scottish English is a Scottish way of using English.Phew, okay!
• linkHedvig Skirgård - But Scots genetically is Germanic.
• linkDaniel Midgley - It is a Germanic language.It is…oh, that's even closer.It's an Anglic language.Hang on.What's Scots Gaelic?
• linkHedvig Skirgård - I thought Scots Gaelic was a kind of Gaelic and that Scots was a kind of…I'm updating myself now that Scots is a kind of Germanic language and then there Scots English, which is like another thing.
• linkDaniel Midgley - Okay, well, because I get to edit, I'm the one who will say, well, everybody, I'm going to drop in the answer now.Scots Gaelic, which I am saying very intelligently, is a Celtic language, whereas Scots is a Germanic language.LAUGHTER
• linkHedvig Skirgård - Okay.Very good.
• linkDaniel Midgley - Okay, so to back to Wikipedia, we know that there are different language versions of Wikipedia.This was Scots, and this person was extremely prolific.They did about nine articles every day for ten years.By the time they stopped in 2018, they'd written twenty thousand articles and even worse, 200,000 edits.And this person was just all over Scots Wikipedia.And they would just take the English language version of the Wikipedia article that they wanted to do.And then they would, like, write it again with kind of an accent or maybe look up some Scots words.But they weren't speakers of Scots.
• linkHedvig Skirgård - Were they…can I be the potentially person stepping in it and say, were they just a very enthusiastic, nice, nerdy teenager?
• linkDaniel Midgley - Yeah, possibly.
• linkHedvig Skirgård - Because you know, they exist and they might not always get it right, but they have good intentions.
• linkDaniel Midgley - "Oh, wow.I love Scots!I'm going to write Scots' Wikipedia pages!"
• linkHedvig Skirgård - Mhm.
• linkGARY - Yeah, that was my hunch, also, knowing nothing more about it than what I read on, what I saw on Reddit and the few tweets.My assumption was that this is just…as is the case of, you know, with prolific Wikipedia editors, they are kind of obsessive, nerdy, mostly men, you know, who often don't have ill intent.They're just, you know…there's someone who changes "comprised of" to "comprised" in every Wikipedia article.
• linkDaniel Midgley - Yes.That fucking guy!
• linkHedvig Skirgård - Wow.
• linkDaniel Midgley - Well, for computational linguists, this is just a disaster.
• linkGARY - Yeah.Yeah.No, it really is.But on a larger scale, people are using machine translation, of course, to translate often articles, and maybe with some hand edits, and populate different language Wikipedias using machine translations.And so now you're creating this feedback loop, right?where you're using machine translation to create the data and then training the translation algorithms, right?on the translated data.And that's a bad situation.
• linkHedvig Skirgård - It almost sounds like one of the points of having language codes is for machine translation translation purposes.It almost sounds like we should issue a special language code for this time period of Scots Wikipedia.
• linkDaniel Midgley - Oof.Well, there's some debate over what to do next.Some Wikipedians are saying that it might just be easier to shut down the Scots Wikipedia and start over and instead of trying to sort through the trash.Yeah, because, I mean, this person didn't write every article, but they did probably edit most articles, and if anybody did write something real, it might have gotten edited into gibberish.So whatever happens, it's going to require a lot of people to put in a lot of time for free, which is what built Wikipedia in the first place.But will they again?It's just a disaster for this minority language, especially one that has struggled for validity in many people's minds.
• linkHedvig Skirgård - It's very sad.It's very sad.
• linkHedvig Skirgård - TRANSITIONAL MUSIC
• linkDaniel Midgley - We're here with Gary Lupyan of the University of Wisconsin at Madison.Gary, thanks again for hanging out with us.
• linkGARY - Thanks.Thanks.It's fun.
• linkDaniel Midgley - You had a paper recently published in Nature Human Behavior, along with Bill Thompson and Séan Roberts.Séan Roberts is a friend of the show.We've had him on a couple of times.
• linkHedvig Skirgård - Yeah, we need to get this whole trifecta, we need to have Bill on.
• linkDaniel Midgley - It's too much!It's too much knowledge.So, Gary, walk me through this.We know that language and culture interact, and we have looked at work showing how the vocabulary of languages is optimised to reflect things that are culturally important.Like, people who live in deserts may not have words for ice.Or something.What got you into this area?
• linkGARY - Yeah, what got me into the area is thinking about the cognitive functions of language.That's where my research started.So the natural assumption is, of course, that language is for communication and when people study the evolution of language, it's often approached from that perspective.It's studied in the context of the evolution of communication.But there is a long history of the idea in philosophy that language is intimately tied to our cognition, that our thinking is aided by language, that learning a language transforms our thinking.The idea, of course, of linguistic relativity, that learning different languages might shape our thinking in different ways.What might be easier to express in some language might be harder to express in another, and that might have impacts on cognition.And so for the last 20 years or so, most of my work has been doing, trying to test these ideas experimentally.Less on the language relativity side, and more on kind of:how does learning a language, any language, affect our cognition and perception?I've done some work in looking at effects of language on visual perception in particular.So that's kind of the background going into this work.
• linkDaniel Midgley - I think that makes a lot of sense because, like for computer languages, if I wanted to do work on regular expressions and text manipulation, I would…I would use Perl or Python.I wouldn't use Java, because it's just easier in those languages.
• linkGARY - Yeah.So I'm glad you bring that up because in computer languages, because we invented them, and they were designed to be equally expressible.So anything that can be expressed in one Turing-complete language can be expressed in another, and that can be proven.That's mathematically true.But of course, that doesn't mean— as you pointed out —that we use any language for any task, because some languages were designed to be better at certain tasks, because there are certain abstractions built in.So just because Python is written in C, doesn't mean that you would always use C for everything, right?And so in the case of natural languages, human languages, we don't know that they are all equally expressive, that literally anything that can be said in one language can be, with the same level of precision, can be expressed in another language.
• linkGARY - But even if it could be, it doesn't mean that people actually would, right?And so if a language makes some idea more…easier to express because it's a more compact expression, because it's used at a higher frequency and so it's more accessible to a speaker in real time conversation, they are more likely to express it.And it might be translatable.But…so, of course, you know, we can learn other languages.And so in principle, the idea, these concepts, these ideas can be translated.But at what cost?So those are the kinds of questions that got me into this work.
• linkHedvig Skirgård - I think that's a really, really good point.There's this sort of mantra in linguistics which usually comes, usually traces back to Boaz of…
• linkGARY - Right.
• linkHedvig Skirgård - …all…Yeah, you know what I'm going to say.
• linkDaniel Midgley - Yeah, I know
• linkHedvig Skirgård - All languages can express the same meaning, but they don't necessarily…there are differences in what is obligatory expressed and that is grammar.But like you said, even if you can express the same meaning, even if Python and JavaScript can technically solve the same tasks, it might be like pages and pages and lines and lines of code in one of them, and not in the other.
• linkGARY - Yeah.
• linkHedvig Skirgård - And that might make you prone to choose one or the other.In fact, some people would argue that trying to have a programming language doing everything is actually a way of just making a programming language worse, because it tends to clutter it and confuse it and things, so people prefer specialisation.But the idea that even if people can indeed express everything, I think is a very good point, that doesn't mean that they do all the time.
• linkGARY - Yeah, and I think also, when we bring the learning dimension into it…So, as language learners, we learn a certain core vocabulary and basic grammar.And so the question arises, well, okay, sure, you learn some words, but not others, depending on the language you're learning.But are you nevertheless learning different meanings?So do you have kind of a different conceptual repertoire?And that's kind of the crux of the paper, where we talk about:do cultures shape meanings?And I don't blame people for having the snarky responses on Twitter, you know, to the press release and headline because it's like, yeah, you know, of course.And we actually…we struggled with…
• linkHedvig Skirgård - LAUGHTERYeah, we should maybe clarify what the snarky response on Twitter was.
• linkDaniel Midgley - Yeah, I do want to talk about the snarky response, just…not just yet, because I think that's another interesting question.But would you mind if, can I try to summarise what you were doing and you can tell me if I'm wrong?I'm going to be the abstract.
• linkGARY - Mhm.
• linkDaniel Midgley - So we know that you can tell something about the meaning of words by their neighbours.
• linkGARY - Right.
• linkDaniel Midgley - You know a word by its neighbours.So what you wanted to do was:let's take a look at some words and their neighbours in 41 languages, and let's see if a word in maybe English has the same neighbours as that same word in French or Basque.And sometimes they'll have really similar neighbours, and sometimes the neighbours will be really different.
• linkHedvig Skirgård - And the neighbours are…so words that occur in a certain proximity of the word in a set of texts.So you use different text, but also presumably you're filtering a bit for words that occur in general a lot.So the word TUESDAY will often occur with the article THE, but so will every other word in the corpora.So you have to do some filtering for that.
• linkDaniel Midgley - Okay, okay.But then at the same time, we're going to work out a metric of how culturally similar the people are who speak those languages.And now this is you:We think that where the word neighbours are super different, the cultures are going to be really different too.How did I do?
• linkGARY - Yeah, pretty good.Yeah, I mean, that's the basic idea.There were actually more languages.There were around 80 languages in the full sample.We filtered…the analysis we report in the main paper are filtered to 41 to avoid the kinds of things that were mentioned with Wikipedia.So there are some (for example) Wikipedias in different languages that are just too small to be trusted.And there were also certain words that were not present, kind of, with sufficient frequencies in all the languages.And so, yeah, the analysis doesn't change if we use the full sample, but given that we knew that the data for some of the languages was not very good, it didn't seem right to use it.
• linkDaniel Midgley - So let's talk through some of the results.And let me just use an example just to make sure I got it straight in my head.So a word like TUESDAY has direct correlates in other languages.French has a TUESDAY, Spanish has a TUESDAY.Lots of places have Tuesdays because Tuesday is pretty common.And if I look at the words that appeared near TUESDAY in English, it's going to be stuff like NEXT and THIS and it's going to be things like WEEK and DAY and other names of wordsdays — Dof the week.And if I go to other languages, I'm going to find pretty much exactly the same thing, because those concepts are pretty, pretty similar to lots and lots of speakers.
• linkGARY - Yeah, yeah.And what makes the word alignable on this metric is not just that it has a lot of the same numbers, but that the relationship of the word to the neighbours is similar.So, for example, the similarity between TUESDAY and WEDNESDAY is going to be higher than between TUESDAY and FRIDAY, for example, because they're closer in time, in the week.
• linkDaniel Midgley - Oh, wow.
• linkGARY - And that's going to be true across lots of languages.And the same thing for numbers and months of the year, for example.So these kinds of systems are really quite similar.And it's not going to be identical.So, for example, in languages where you have GOOD FRIDAY and, you know, various culturally specific associations with certain words, that's going to make those words be slightly less aligned than they would be otherwise when comparing, you know, language that uses phrases like GOOD FRIDAY to a language that doesn't, because FRIDAY will now have this higher valence, for example.
• linkDaniel Midgley - Well, it sounds like you've had a cultural artifact, is what you found there.
• linkDaniel Midgley - Yeah.So those are the kinds of things that contribute to misalignments, even when you say, well, that's the right translation.That's the best word available.
• linkDaniel Midgley - That's days of the week, but other things didn't align quite so well?
• linkGARY - Yeah.So the words relating to the body.You know, one might assume that a nose is a nose and a neck is a neck.And literally that's true.And if you look up the translation for NOSE, you're going to get a word in lots of languages.But the way that the word enters into other constructions, so the polysemy, the kind of range of meanings of the words, it actually differs quite a bit.And so, for example, in English, you know, you talk about being NOSY, right, and that might not hold for other languages.So NOSE is now having this other meaning.The other thing that varies often is frequency.So one thing that we did— one kind of analysis —is once we computed these alignments for each word and each language pair, we can see, well, what predicts the alignment.And one thing that predicts the alignment is differences in frequency — so this is written frequency.The larger the difference in frequency, the less well the words align.And in some extreme cases, one I use when I talk about translation to undergrads is COMRADE.LAUGHTER
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - So, as a translation from the Russian word товарищtɐˈvarʲɪɕː), COMRADE is the right translation, but it just cannot have the same meaning, if only because that word in Russian, at its heyday in the Soviet times was used…It had roughly the same frequency as words like MAN and WOMAN.Right?Because it was just the form of a gender-neutral form of address, kind of not necessarily formal, but it's how you would address someone.And that's just…it can't have that meaning to an English speaker — like, the word COMRADE.And so, yes, it's the right translation, but it just will not have the same effect.And if you look at its semantic neighbourhood, it's going to be quite different.And a lot of that is driven just by that huge difference in the frequency between the Russian word on the one hand, and the English translation on the other.
• linkHedvig Skirgård - This is where we get into one of the trickiest matters that I imagine you guys have discussed a lot, which is polysemy and metaphor.Because obviously, like you're saying about NOSY, and then there are words that are just straight up polysemous that are in your data as well.So how did you guys think about and handle that?
• linkGARY - So there have been some previous work that we cite in our paper that have used patterns of polysemy to argue for semantic universals, to argue that culture doesn't matter because the patterns of polysemy are the same across languages.
• linkDaniel Midgley - For people who aren't familiar with the concept of polysemy, it's the way that NOSE can have different kinds of meanings, like it could be the nose on your face.It could be the nose on an airplane.
• linkGARY - Yeah.
• linkDaniel Midgley - It can be the nose that wine has.
• linkHedvig Skirgård - Yeah, then we get into the arguments of, like, when those become different words or just senses of one word, but yeah.
• linkGARY - Yeah.Or for example, I was mentioning the body.So it's common across languages that a word like BACK, which names a part of the body, is also used for space.So you talk about MOVE BACK, right?And BACK IN TIME.And you do find— it's not universal —but you do find that lots of languages will extend the body part to these other meanings, these other domains.Yeah.
• linkDaniel Midgley - And it sounds like you found that when there were languages that were more closer to each other, historically related, and if there was a cultural similarity, then the word neighbourhoods lined up a lot more.They were aligned more closely.
• linkGARY - Yeah.We could look at it by semantic domain as well.So, for example, kinship terms in languages spoken, and here we have this languages spoken by cultures.Of course, that's not a one-to-one relationship.But if we take sort of the closest culture for which there is anthropological data that's been input into this D-PLACE database, there's information about kinship systems that are used.And this is not based on the language, this is based on observations.And so, is that reflected in the semantics of the language, in the way that we're quantifying it?And so that might not be surprising to many anthropologists, but even if it's true for a culture that someone studies and someone can say, well, I'm not surprised by that because I've been in the field and that's the case, it's still valuable, I think, to see, you know, does it hold up for different language families?And, you know, to what extent is it correlated, for example, you know, to what extent is it also there for animal terms or terms for natural features?
• linkHedvig Skirgård - That's what I was going to say, as well as the fact that the kinship system, because kinship data is actually in D-PLACE and, you know, they ask questions like, do you make a distinction between gender for siblings, and these kinds of questions?It's not that strange that that would align.But the fact that it also held true…if it only held true for the kinship subset of the meanings you studied, then this wouldn't be as strong a correlation as it is.
• linkDaniel Midgley - Let's talk for a second about the anthropologists, because when this paper broke on Twitter especially, as we mentioned, there was a lot of snark.A lot of snark going on.And there was a lot of hostility.And the tenor of the remarks that I saw were things like:why is this news?We've known this for hundreds of years.One tweet was, "Is computational linguistics Columbusing anthropology?"
• linkHedvig Skirgård - Oh.
• linkDaniel Midgley - I noticed that you were explaining the work, but why was it, why was the reception so hostile among some groups of people?
• linkGARY - I can't know for sure, but I think there were different reasons, right?Different people.The press release one-liner didn't help.It kind of made it seem that we were computer scientists.And, you know, linguists, psychologists, lots of people who study language have developed a kind of allergic reaction to, for example, you know, physicists coming along and saying:Oh, you know, I've got a formula that explains language evolution.
• linkDaniel Midgley - And for good reason.
• linkHedvig Skirgård - Yeah, we've talked a bit on the show about economists.
• linkDaniel MidgleyGARY - Yep.
• linkHedvig Skirgård - That's a known phenomena.
• linkDaniel Midgley - Yup, I mean, that's yeah.There's there are good reasons for that, but.
• linkGARY - Yeah.Yeah.But I think what one has to be careful there because…So as, as a cognitive scientist— is how I identify myself, who studies language, so I don't identify as a linguist for what it's worth, but I obviously collaborate with linguists —So I see value in quantifying things, in running experiments where it's possible.And so if some…so I do see value in generalising, of course, the data — and in this case, the data often come from field work.So in my earlier work where we were looking at relationships between population and grammatical features and grammatical complexity, none of that work would be possible without people writing grammars and then people quantifying these observations in a way that enables these data to be analysed at scale.And so I see value in drawing generalisations and bringing together large…trying to expand the coverage.I don't think that takes away from the value or the need to study individual cases in great depth.But I can see how someone who spends their time doing that would be annoyed.Yeah.
• linkHedvig Skirgård - There's also a great value in estimating the effect size of something.So, for example, my PhD supervisor Andy Palli wrote several papers saying that sort of societal organisation mattered for language diversification in the Pacific.And you can use the Ethnographic Atlas variable for like quote unquote political complexity for this.And I decided to test that and see, sure, it matters, but how much does it matter and how much does it matter when you factor in other effects?And you get a sort of a raw number, you can sort of put that into a context instead of just writing "it matters".You can form, you can form convincing arguments.But it is interesting to attach some numbers so you can sort of rank how relevant it is.That's what I see as the value of this paper as well.
• linkGARY - Yeah.And I think, you know, where linguists, psychologists I think are right to…well, you know, it sounds weird to say our right to be annoyed.You know, people don't need permission to be annoyed, but maybe are right to dismiss certain claims.You know, it's when, it's when the abstraction is so far from the data that it just doesn't tell you anything about, you know, how to understand your data.And I, I would hope that's not the case here, but we really do need the real expertise of translators and multilingual speakers who can give us…because in the end, I mean, this is a subjective measure.It's a psychological construct.
• linkDaniel Midgley - Can you tell me, Gary, one thing that you've learned from the paper itself and one thing that you've learned from the media coverage and the subsequent response?
• linkGARY - Well, I suspected that this might have been posted on Reddit as well.So I searched Reddit last week.And it was, it indeed did make Reddit on the bad philosophy subreddit.And the post said "computer scientists discover language".
• linkHedvig Skirgård - Oh, my god, oh, I'm so sorry.
• linkDaniel Midgley - They didn't read it, did they?They didn't read behind the title.
• linkHedvig Skirgård - No.No, they didn't.
• linkGARY - No.And then there were all sorts of snarky comments.And like, I'm a full professor, you know, I don't really care that much.But Bill, the first author, is a postdoc, and he was really kind of dismayed at kind of being painted as this tech bro, right?Like, who is this?you know.And he cares very deeply about language and has a degree in linguistics.And so I guess I, I didn't expect that…I didn't realise that the coverage would kind of, there would be this computer science kind of headline or data science headline, but I didn't expect that, oh yeah, that people would have this kind of reaction to a well, you know, "what do they know about language?" things, so that you know.
• linkHedvig Skirgård - And yeah, there are two things I noticed about it.The first was some people who are complaining that you weren't citing certain people.And people always complain about that and I think are different styles of writing science articles.There's different journals.And I think people expect…like humanities is notorious for like citing a lot, and citing very old stuff a lot.And I think that kind of critique, I think you guys cited a lot of people and it was fine.And I think people are always happy to jump on that kind of critique.The other thing I noticed, speaking of the thing of, like, being painted as a tech bro…um academia in the last decade and even more so now, has become a very precarious place.Um, a lot of people don't have funding to do the research they want to do.And a lot of people who do qualitative humanities feel threatened, I think, materially for research money, honestly.
• linkGARY - Yeah.
• linkHedvig Skirgård - And I think that's, I think we shouldn't underestimate the kind of anxiety that that brings.
• linkGARY - Yeah.And that was one of the things that I saw right away and people's responses, and in retrospect, I should have expected.Because I know that even within humanities departments, there's tension between, you know, the people doing the digital humanities kind of stuff and the more traditional humanists.And yeah, and I mean, my response is I would love, for example, to collaborate more with people who do translation studies.And I think there is a lot to be learned and a lot of…But again, you know, here I'm thinking as kind of an empirical scientist that I'm excited about, you know, what kind of data they can help me with, right, so that, you know, to advance my research questions and, of course, from their side that that's not necessarily what they are interested in.And that's also been a hurdle.For example, the Cognitive Science Society has anthropology as one of the disciplines.And over the last few decades, you know, there've been fewer and fewer to now, you know, pretty much no anthropologists who go to the cognitive science meeting.And lots of cognitive scientists would love more anthropologists to be involved.But it kind of like, well, on our terms, right?Like, oh, we want to run different experiments, you know, in different cultures.But when the anthropologists say, well, I'm not really interested in running your experiments, they don't make sense to the people I work with, right?that kind of puts a damper on the collaborations.But we really do need to kind of be more interdisciplinary in more than the name, I think, to get traction.
• linkHedvig Skirgård - And realise how things can complement each other.Like you wouldn't be asking these questions and having all these interesting venues…these interesting inroads into this topic if there weren't for the philosophers and anthropologists and psychologists etc, who thought about this.
• linkGARY - Yeah, absolutely.There was one point I wanted to make, and feel free to edit it out, is one critique I saw, is:Well, these are a biased set of languages and so, you know, what can we…what conclusions can we draw from it?It's a biased sample.It doesn't represent all human languages.But I think that misses the point that we're trying to you know, one of the arguments we're trying to advance here is that, even in closely related languages, there are differences that people didn't realise existed.So it's, in fact, a stronger argument to say:look, we're comparing English and German.You know, as far as world languages go, they're basically the same language.And look at kind of…
• linkHedvig Skirgård - Well, I'm trying to learn it now, but yeah.LAUGHTER
• linkGARY - And look at the kinds of differences that exist even in related languages, right?And so, of course, we would love to have more languages to examine in this way.But there is value in kind of…if what you're interested in is finding the extent of a difference, looking at related languages can often be…can make for the strongest case.
• linkHedvig Skirgård - Yeah, no, I thought the same thing, actually, and I was thinking of bringing it up earlier, but we had so many things to talk about!Yeah, precisely.That they do differ still is actually a really good point.
• linkGARY - Yeah.
• linkDaniel Midgley - Well, it's really fascinating work, not only for what it tells us about the world's languages, but for what it tells us about the process of rolling out research, and the kind of pitfalls that can happen.So, Gary Lupyan, thanks so much for telling us about your work today.
• linkGARY - Thank you.Thank you.Really enjoyed it.
• linkGARY - TRANSITIONAL MUSIC
• linkDaniel Midgley - It's Words of the Week, and because nobody's here to complain about it, Hedvig, are you going…would you like to take that on?
• linkHedvig Skirgård - So Gary, our other co-host, Ben, hates Word of the Week.I don't hate Word of the Week.
• linkDaniel Midgley - He doesn't!He loves it!He fucking loves Words of the Week.
• linkHedvig Skirgård - The thing I hate about our show is that we have too smart listeners.They make my job too hard, but that's, you know, we hate different things.I don't have any hate for Word of the Week.I can't muster it up.I'm sorry.
• linkDaniel Midgley - Okay, well, Gary, you've got a couple of words for us.Whatta you got?
• linkGARY - I've got two that we in the US have been hearing a lot lately.So HOAX…
• linkHedvig Skirgård - Uh-huh.
• linkGARY - I looked at the Google Trends for HOAX and it had a huge peak mid-March…
• linkDaniel Midgley - Oh, gee.
• linkGARY - …when the shut down started here, because of course covid-19 is a hoax.
• linkDaniel Midgley - Of course.
• linkGARY - Of course.
• linkDaniel Midgley - So say the living;the dead were unavailable for comment.
• linkGARY - Yes.And I've never looked up the etymology of HOAX.It's a weird word.It doesn't look very English-y.And the consensus, you know, there's controversy as always, but the consensus seems to be that it comes from HOCUS.From HOCUS POCUS.
• linkDaniel Midgley - Oh, wow.
• linkHedvig Skirgård - Oh.
• linkDaniel Midgley - I had never thought of that.I was going to say Greek, but that's really interesting.
• linkGARY - Yeah.Yeah.So mid-1600s, and the verb, which is pretty unusual even now, the verb is attested in 1796, as in "to ridicule".Yeah.
• linkDaniel Midgley - That's pretty new.
• linkGARY - And then the other word was RIGGED.Yeah.Trump is setting us up for the election being rigged.And I had assumed that it was some weird sense of like rigging a ship or a RIG, like a thing.
• linkDaniel MidgleyHedvig Skirgård - Yeah.Me too.
• linkGARY - And apparently it's unrelated to either.
• linkDaniel MidgleyHedvig Skirgård - What?
• linkDaniel Midgley - That's bananas.
• linkHedvig Skirgård - Yeah, that was my thought as well.
• linkGARY - And it comes from a noun, a RIG and in the 1700s that meant something like a trick.And I'm imagining kind of like three-card-monte kind of situations.Right.So like tricking someone to get money out of them.And then the verb comes later.Yeah.So 1600s, first used as a noun, and then early 1700s as a verb.
• linkDaniel Midgley - Wow.So a couple of words from impeached US President…I just always describe him as "impeached US President and Russian asset, Donald Trump".The quote is interesting.He says, "make sure to vote because the only way we're going to lose this election is if the election is rigged".So what he's doing here is he's engaging in a form of confirmation bias, like:how do you know if the election results are valid?Well, if I win, then it's valid.If I lose, then it was rigged.Heads I win, tells you lose.HOAX and RIGGED are interesting, because they are two words that he uses to dismiss and to help others dismiss anything he doesn't like or anything he finds disadvantageous.And there's a term for this in psychology that I love.It's called a thought-terminating cliché.A thought-terminating cliché.
• linkGARY - Mmm, that's wonderful.
• linkDaniel Midgley - And I engaged in this when I was…yeah.Back when I was religious.I engaged in this a bit, like when there were inconsistencies that I felt threatened by, I would say things to myself like "well, we don't know everything" or "God will explain that to me in the hereafter".And so that was a way of stopping thinking, and not going to the conclusion that this guy's lying to you.
• linkGARY - Yeah, yeah.
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - FAKE NEWS is another one.When you say that something is FAKE NEWS, you can just dismiss it.And that's a thought-terminating cliché as well.
• linkHedvig Skirgård - This is going to sound weird, but I think in my I am now in my thirties and I think I'm getting wiser by the minute.It's probably not true, but that I used to have a greater tendency to…I feel like nowadays I'm more happy to leave thoughts be.I'm like, oh, this isn't…I don't need to spend time on this.I'm going to stop thinking about this now.I'm going to do something else.For example, I actually tried to look up this whole negative Twitter kerfuffle and then I read a bit and I was like, okay, I'm stopping.Like, I don't…I've learned enough.I don't need to read any more of this.This isn't useful to me.
• linkDaniel Midgley - And that's fine.
• linkHedvig Skirgård - That's like the positive stop-thinking.
• linkDaniel Midgley - Yeah.You don't need to take it…you don't need to wallow in negativity, for sure.Absolutely.I think it's when you're using it to protect your own opinion and you don't want to change that, it becomes a bad thing.What do you think?
• linkHedvig Skirgård - Yeah, I just think there's…that sounds like a fine line, but yeah.LAUGHTER
• linkDaniel Midgley - Okay, yeah.
• linkDaniel Midgley - Next one.Oh, we may be delving into some toxic masculinity and/or ableism here, sorry about this.The word SIMP.
• linkGARY - SIMP.That's a new one.
• linkDaniel Midgley - I've noticed this word a lot lately.
• linkHedvig Skirgård - Is this someone who is simpering?
• linkDaniel Midgley - No, it doesn't appear to be.Somebody who's simpering, that's unrelated.SIMP originally, and this word is a really old word.it's a simpleton.Okay, so that's kind of not…that's not great.It comes from as early as 1903.The first Oxford reference is:"In circus dialect"…circus dialect…YAP and SIMP indicate a credulous rustic who is easy prey for sharpers".So someone who gets tricked.But this has evolved a new sense and it's a guy who is too much into their crush or their girlfriend, because you're not supposed to like your girlfriend very much, right?You're not supposed to…it's deeming someone's absorption with their inamorata as inappropriate in their view, which is an interesting look at the expectations that some guys have about relationships and the mercantile nature of it all.Anyway, the reason this is in the news is because Bill Shorten, the Labor leader, has applied this term to Prime Minister Scott Morrison in Australia.He said on the show Insiders, "Mr Morrison needs to make sure that he doesn't look like he's just a simp to Donald Trump on this very important issue".And then the news guy said, "Explain simp?" And Shorten said, "Well, soft."
• linkHedvig Skirgård - Okay, hmm.
• linkDaniel Midgley - So I just, you know, just in the same way that guys on the internet started calling each other CUCKS, I find it very much in that gross toxic lexicon.What do you think?
• linkHedvig Skirgård - Yeah, this is the first encounter of it is in this metaphorical use with Donald Trump, I'm like meh, meh, meh, maybe.But yeah.
• linkGARY - Yeah, I have not seen that in American media.But I'll keep an eye on it.
• linkDaniel Midgley - Okay, please do.
• linkDaniel Midgley - Next one, 2020.
• linkHedvig Skirgård - Perfect vision!
• linkDaniel Midgley - Yes.But also in addition, the Economic Times India suggests this one could be not just a reference for perfect vision, but also:a complete shitshow.An unending series of catastrophic events:"Boy, our school bake sale was a real 2020" or "That's so 2020".This year has just been really terrible for a lot of people.
• linkHedvig Skirgård - Do you remember when, that year when all the actors died, like Alan Rickman died and David Bowie died and everyone was like:this is the worst year ever?And it sort of felt like that that was about to happen.I think it was 2016.
• linkDaniel Midgley - It was 2016.Yeah.And we had no idea, did we?
• linkHedvig Skirgård - We had no idea, did we?
• linkGARY - I saw a bumper sticker the other day that said "Giant Meteor for 2020" and then in small letters "Just end it already".
• linkDaniel Midgley - Just end it already!
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - That's a good idea.Let's make 2020 end on August 31st, and then…
• linkHedvig Skirgård - We can all have Christmas.
• linkDaniel Midgley - Everybody can have Christmas and New Year and then we can start with, like, 2020 and a half, or…
• linkHedvig Skirgård - 2020b.We can do like you do references and bibliographies.LAUGHTER
• linkDaniel Midgley - 2020+.I like it!Okay, so this episode is coming to you in September, 2020+.
• linkHedvig Skirgård - Yes.
• linkDaniel Midgley - All right.
• linkHedvig Skirgård - Much better.
• linkDaniel Midgley - Good.Solved that.It's not the dumbest solution to any problem I've heard this year.Okay, last of all…how do you say the bird, Gary the bird that's spelled E-M-U?
• linkGARY - i.mjueemyoo.
• linkHedvig Skirgård - I also sayi.mjueemyoo.
• linkDaniel Midgley - i.mjueeymoo?Okay, U.S.National Public Radio is going withi.mueemoo.
• linkGARY - Hmm.
• linkHedvig Skirgård - Okay.
• linkGARY - I can live with that.
• linkDaniel Midgley - Which sounds strange to me, but yeah, okay
• linkHedvig Skirgård - Yeah, I'm fine with it.
• linkDaniel Midgley - Australians that I talked to on the ABC Radio aren't having any, frankly.LAUGHTER, HEDVIG INADVERTENTLY SNORTSI guess they're not fans.
• linkHedvig Skirgård - Okay.
• linkDaniel Midgley - The question I get is:which way is right?And my response is, well, this comes from Portuguese.It was Tiger Webb that points out:this comes from Portuguese.The word isi.ma,e.ma.And it doesn't have a "ya"jat all.I also just want to do a comparative thing.Gary, if you would help me with these words.Australians say tunetʲun, play us a tunetʲun.
• linkGARY - Oh, yeah.
• linkDaniel Midgley - What do you say?
• linkGARY - tun.
• linkDaniel Midgley - Okay.Do you read the newsnuz, or do you read the nyewsnjuz?
• linkGARY - nyewsnjuz
• linkDaniel Midgley - Oh, interesting.Okay, so what Gary is doing there is including a little yod, a littlejsound.
• linkGARY - Yes.In news.Not in tune.Yeah.Calculatorkæɫ.kə.lei.təris another one, right?
• linkDaniel Midgley - Calculatorkæɫ.kə.lei.tər.Oh, wow.Okay.Oh, cool.What about nudenud?Is someone nudenud, or are they nudenjud?
• linkGARY - Hmm.
• linkDaniel Midgley - Try not to think about it too much!
• linkGARY - nud.
• linkHedvig Skirgård - I would definitely saynud.
• linkGARY - nud.
• linkDaniel Midgley - Okay.In Australia, you'll find people read the newsnjuz, but they are often nudenud.They read the news in the nude.So yod-dropping is not uniform.Some places will keep the yod.Some places will drop it.Some places it's even dropped even more.I pointed out once that in East Anglia people might say, if you look really nice, if you look really cutekjut, they would say that you're cutekutbecause they take the yod-dropping to that extent.I just want to throw this in here.NPR decided oni.mubecause it's been that way in many of their stories.But when I went onto youglish.com, which is a site that allows you to look at a certain word as it appears in YouTube annotations, and it takes you right to that word so you can listen to a lot of pronunciations really quick.
• linkGARY - Oh, cool.
• linkDaniel Midgley - It's really cool!And you can break it down by country, too.So when I looked at only USA videos,i.muandi.mjuwere about 50/50 for the first few that I listened to.Also, the coverage of this, I think, is bury the lede, because when I found out that I was going to go to Australia for the first time, I looked up the word E-M-U because I didn't know how to say it.I hadn't encountered it before.And I found that the dictionary that I looked it up in said that emyooem.juwas also a valid pronunciation.
• linkHedvig Skirgård - Yeah, I've heard that.
• linkDaniel Midgley - Have you heardem.ju?
• linkHedvig Skirgård - Yeah.
• linkDaniel Midgley - Watch out, there's an emuem.jubehind you.
• linkHedvig Skirgård - Yeah.
• linkGARY - Okay, sounds odd to me, but okay.
• linkHedvig Skirgård - Yeah, sounds odd, but I've heard it.
• linkGARY - It would get the point across.There aren't too many similar sounding birds.
• linkHedvig Skirgård - Yeah, exactly.Yeah.
• linkDaniel Midgley - Look out!There's a bunch ofem.juz!You silly sap, it's clearly an em, ahh!!LAUGHTER
• linkHedvig Skirgård - We should say the Daniel's broadcasting from Perth, and that Western Australia was home to one of the few wars that Australia has lost, correct?
• linkDaniel Midgley - Correct.
• linkHedvig Skirgård - Yeah.The war between the white settlers and the emus.
• linkDaniel Midgley - Yeah, terrible bloody conflict.Those birds are relentless.
• linkHedvig Skirgård - Yeah, look it up, Gary.It's technically defined as a war for various reasons.
• linkGARY - I will look it up.
• linkDaniel Midgley - Yeah.And the emus couldn't even hold rifles with their tiny feet…their enormous feet.Oh, God.Okay:HOAX, RIGGED, SIMP, 2020 and EMU:our Words of the Week.Gary Lupyan, thank you so much for hanging out with us today.
• linkGARY - Thank you.My pleasure.
• linkDaniel Midgley - And we will be sure to get in touch again when more research comes down the pike.Quick comments from Nigel.One of our words last week was SYMPATICO, describing the relationship between presidential hopeful Joe Biden and vice-presidential hopeful Kamala Harris.The word was SYMPATICO.Nigel says, "The Spanish use of SYMPATICO and its use as an English loanword always makes me think of sympathetic resonance, particularly in a musical context.The idea of an agreeable person being like strings that cause each other to resonate due to similar underlying harmonic characteristics".That's nice, isn't it?
• linkHedvig Skirgård - That's nice.That sounds like the nice version of dog-whistling.Sorry
• linkDaniel Midgley - Yes!Sympathetic resonance.Yes.Like when you…I've always noticed when you sing into a piano, the same note answers you back.So that's sympathetic resonance, too.Neat.Thanks, Nigel.
• linkDaniel Midgley - END THEME
• linkHedvig Skirgård - If you like the show and you want to give us a shout out and tell us what you think about it, or if you disagree with something, you can always get in touch with us.We are on all the social medias except TikTok, essentially, which means Facebook, Twitter, Instagram, Mastodon, and Patreon.And on all those platforms we have the same name, becauselangpod, which means it's really easy to find us.You can also send us an old-fashioned email at hello@becauselanguage.com.And if you like the show, tell a friend about us, or leave us a review anywhere where there are reviews to be left.
• linkDaniel Midgley - We're really grateful to all our Patreons who are awesome people and who help us, who support the show and keep us talking.So thanks to everyone who's become a patron.Here is a smattering:Termy, Chris B, Lyssa, The Major, Chris L, Matt, Whitney, Damien, Helen, Jack, Elías, Michael, Larry, Kitty, Lord Mortis, Binh, Kristofer, Dustin, Andy, Nigel, Bob, Kate, Jen, Christelle, Nasrin, Ayesha, Emma.Thanks to you all.Our music is written and performed Drew Krapljanov, and you can hear him in Ryan Beno and Didion’s Bible.Thanks for listening.We'll catch you next time.Because Language.