1100. This week, we talk with popular online linguist Adam Aleksic, also known as Etymology Nerd. We discuss "algospeak" and how social media algorithms are changing language. We discuss euphemisms like "unalive," the spread of trending words, and how content creators adapt their speech to go viral.
1100. This week, we talk with popular online linguist Adam Aleksic, also known as Etymology Nerd. We discuss "algospeak" and how social media algorithms are changing language. We discuss euphemisms like "unalive," the spread of trending words, and how content creators adapt their speech to go viral.
Adam Aleksic - The Etymology Nerd
Adam's book - "Algospeak"
🔗 Share your familect recording in a WhatsApp chat.
🔗 Watch my LinkedIn Learning writing courses.
🔗 Subscribe to the newsletter.
🔗 Take our advertising survey.
🔗 Get the edited transcript.
🔗 Get Grammar Girl books.
🔗 Join Grammarpalooza. Get ad-free and bonus episodes at Apple Podcasts or Subtext. Learn more about the difference.
| HOST: Mignon Fogarty
| VOICEMAIL: 833-214-GIRL (833-214-4475).
| Grammar Girl is part of the Quick and Dirty Tips podcast network.
| Theme music by Catherine Rannus.
| Grammar Girl Social Media: YouTube. TikTok. Facebook.Threads. Instagram. LinkedIn. Mastodon. Bluesky.
Mignon Fogarty: Hey, just a quick heads up that we talk about a euphemism for suicide at the beginning of this episode.
Mignon: Grammar Girl here. I'm Mignon Fogarty, and this is our Thursday interview show. Adam Aleksic is a linguist and content creator, best known for posting videos as Etymology Nerd to an audience of nearly 3 million. He's the former president of the Harvard Undergraduate Linguistic Society. He's written for The Washington Post and is a frequent contributor to NPR. And today we're gonna talk about his new book, “Algospeak,” which is about how social media is changing our language. Adam, welcome to the Grammar Girl Podcast.
Adam Aleksic: Thank you. I'm super excited to be here.
Mignon: Super excited to have you. So let's start with the basics. I mean, what is algospeak?
Adam: Yeah. Well, traditionally, the term has referred to words used to evade algorithmic censorship. The classic example is the word "unalive." So we have actual kids in middle school now saying the word "unalive" because you can't say "kill" on TikTok and other social media platforms. So people find creative ways to circumvent these guidelines. Yeah, there's a lot of examples like that. There's creative respellings people do. You'll see like comic book-style asterisks and at signs and dollar signs. But you'll see words like "seggs" instead of "sex." You'll see words like emojis, like the eggplant emoji. These are all evasive ways to talk on the internet, and the idea here is that algorithms are shaping how we speak; therefore, algospeak.
My book kind of starts there, but then it dives further because I think algorithms are now shaping every aspect of how we communicate and relate to each other online.
Mignon: Right. I mean, it kind of reminds me, you know, kind of old-school leetspeak, you know, where you'd substitute letters for numbers and things like that. So it sort of has an in-group aspect to it as well, doesn't it?
Adam: Absolutely. So for algospeak, the in-group is people on social media. It's this shared vernacular we all kind of have together. And I do draw leetspeak as an analogy in my book because leetspeak also started as a way to avoid content filters in early internet chat rooms. So there are certain words that were censored, and people would come up with these “leet” ways of respelling words that the chat rooms couldn't filter through. And this is sort of the next step in that.
Mignon: Yeah. So you said it's changing like every—
Adam: It also is genuinely was a subculture. This "leet" subculture, and algospeak too, is the subculture of speaking on the internet now.
Mignon: Right, because I'm sure a lot of my listeners who aren't as active on TikTok or Instagram Reels as we are might not be familiar with "unalive" and these substitute words. So besides these sort of substitution words to get around the censorship and the algorithm, like what are some of the other ways that the internet is changing the way we communicate?
Adam: So as a content creator myself, I can tell you that every word is also a piece of metadata for the algorithm. It's not just hashtags anymore. These algorithms are looking at every single word that's used. They have NLP transcriptions of the audio you use. They have the computer vision analyzing every word on the screen, and all this is taken together and embedded and turned into like this computer representation of what the video's about.
So the words you're using yourself are now part of this process used to target viewers. And so people will very deliberately use words to try to reach a specific audience that they think will appreciate that word. And when a word's also trending, because algorithms perpetuate trends through trending songs, trending hashtags, but also words themselves. If a word is trending, creators will try to lean into it. And so we have kind of this perpetuation of a lot of trending words and some famous examples here like "rizz" and "skibidi." And once these words started trending, creators intentionally used those words because they were hopping on the trend, but in doing so, they helped the word spread.
Mignon: Right, and like you are creating content specifically for the algorithm. With the algorithm in mind. I mean, it's for people. You want people to enjoy it, but it was really like I think about the algorithm, but I don't think about the algorithm to the extent that you think about the algorithm, and it was really fascinating.
Adam: It starts with a different style of communication altogether. What we're using right now is broadcast communication. That's literally the etymology of podcast, right? iPod broadcast. And what that means, it's drawing on the analogy of broadcast farming, where they put seeds everywhere and disperse it everywhere.
So the idea here is that when we talk, everybody who tunes into this podcast will hear it. There's no filter. With algorithms, we use a different style of viral communication where my message will only reach a few hundred people initially, and then only if it's compelling and attention-grabbing to those few hundred people will it spread beyond that. And then the algorithm continuously reevaluates whether this message is good at capturing our attention. And that means that what goes viral is also what captures our attention. And when we have new words enter the zeitgeist, these words are somehow good at being compelling to us as humans, and the algorithms pick up on that and perpetuate that.
Mignon: Right. And you talked about the Matthew Effect in your book, which is just, you know, if a piece of content does just even a little bit better in the algorithm, it has more, a lot more potential to go viral. Did you start like, thinking about virality and the algorithm, or did you start as someone who loves etymology and then just sort of discover how to manipulate the algorithm?
Adam: A little bit of both. I got very into etymology in 10th grade. It started with Mark Forsyth's book, The Etymologicon, which I highly recommend to beginners.
Mignon: I love that.
Adam: Then I just started reading a lot more about, you know, words, word origins. I started a little blog for myself back in the day called etymologynerd.com. So I was, that was just for me, you know, I was just doing that. And then eventually down the line when I graduated with a linguistics degree, I was like, I should make content. So the etymology was always there. But, also the knowledge of virality kind of evolved separately. So around the same time, I was separately interested in going viral on Reddit, and I was like a Reddit influencer back in the day. But the tactics are kind of the same. It's how you get people's attention and how you can kind of capitalize on that. And then when I graduated college, I was like, I'm gonna combine these, and I started making these TikTok videos about etymology. But it's hard to separate both linguist brain and content creator brain from when I make stuff and kind of seeing those evolve together and how they very much kind of are the same thing led to the creation of this book.
Mignon: Yeah. It's funny, to prepare for this interview, I watched about an hour of your videos on Instagram Reels.
Adam: I’m so sorry. They’re so intense.
Mignon: No, they're so good. But they are, that's what I was gonna say. They're so intense and you talk really fast, and I've actually, even in this interview, I find myself talking faster than I normally do in an interview because you're talking fast. So maybe first, let's talk about the sociological aspect of like why I am talking so fast right now.
Adam: Well, I am a fast talker, and you know, there's this linguistic phenomenon of communication accommodation where we change how we speak depending on our perception of what the audience is. So right now, maybe we're talking, you know, you're talking slightly faster to me simply because I am naturally a fast talker. But you also notice that if you compare how I'm talking now to how I'm talking in videos, it's a different style of communication. I'll stress more words to keep you engaged. And that's because my perception of what audience I'm accommodating for is the algorithm's part of that, and my knowledge of how people need to like be drawn into the video, how I need to capture their attention plays a part in how I'm gonna speak to you.
So right now, I know I have through broadcast communication this access to a longer form style of attention. But with viral communication, if I'm not maximally compelling to you in the moment, you're gonna scroll and you're gonna get a more compelling video. So I have to get your attention, and that changes how I communicate. So my style of, I break down a few different kinds of online accents, and we do code-switch into these accents. So I have like what I call an educational influencer accent. You have also like the lifestyle, beauty influencer accent, like the "Hey guys, welcome to my podcast, get ready with me," kind of, but they're also accommodating for their perception of what audience they're talking to. And it also works as an attentional tactic, because it kind of draws you into this lullaby-esque trance where you watch the video and forget that you can scroll.
Mignon: Yeah. And another thing, you talked about in the book is, so at the very beginning of the interview, so I'll, you, you reminded me, I'll never forget my editor when I was starting to do videos, and he said, “Mignon, you can't breathe at the beginning of the interview. You can't, you can't take that breath.” That's an old person thing. And you know, it's so stressful. And I'm like, good God. Like, I can't even breathe now. But like, what is the deal with like that first like half second of the video and all the tricks to keep people watching?
Adam: You can't have dead air because the first second, 50% of your viewers will scroll away. So if you have dead air for that first second, that's already lost half your audience. So a common tactic for especially younger creators or more internet-savvy creators is to just launch right into it and edit it down such that you get bombarded right from the start. And that does affect both the verbal and visual hooks that we're going to use. A visual example is there's like the so-called Gen Z shake, where somebody will have a visual disruption of putting their phone down as they start speaking, and that interrupts scrolling patterns. And now you're watching the video. But that's only a visual indicator, right? We can also see the same thing happen linguistically. There are a few set hooks that all creators will rotate through, and once you start paying attention to it, you can't stop looking at it. How we all have the same 30 intros that we all use in our videos because that's what works best at capturing your attention initially. You have to rely on that. So maybe it's a little reductive. We've always had attention grabbing techniques, whether you're writing a journalistic article or starting a lecture or something; you need to capture people's attention. But it's just, I feel like the algorithm compounds natural human behavior.
Mignon: Yeah. And so also thinking about the algorithm, there was an interesting section in your book about the suffix “core,” like “Barbiecore” and “cottagecore,” and how that plays into the algorithm and how it works. So it'd be great if you could talk more about that.
Adam: Yeah, so when I say the algorithm compounds natural human behaviors, another thing here is that as humans, we naturally separate ourselves into in-groups and out-groups. We find who we're alike and we try to be more like them, and we find who we're not alike, and we differentiate our identity so that we're different from them. Algorithms play into this natural human process by making you feel like you're part of an in-group by pushing you into certain echo chambers. They also homogenize some trends. So it makes you feel like you're in the general in-group of social media, but there are very specific aesthetics that it tries to push you into. One thing you can't separate is that these platforms are commodifying our speech. So when you have like niche echo chambers like cottagecore and goblincore and angelcore or whatevercore, here being a kind of lifestyle kind of aesthetic. You have these sort of echo chambers where you feel like a cottagecore person because that's your natural instinct.
But at the same time, the algorithm's profiting off of that because, you know, you're one click away from cottagecore clothing on the TikTok Shop, or you wouldn't have been a cottagecore person unless you knew that label was there. So there's something about the label and how it's applied to you, and then how you identify with it. And the algorithms like to push labels also because they're more metadata. If they can identify you as a cottagecore person, then they can target more videos to you, and then you feel like, “Oh, you know, the algorithm really knows me.”
Mignon: Yeah, no, I went down some rabbit holes. Like, I started searching for different kinds of core after I read your book. And it's like these whole subgroups that I didn't even know existed.
Adam: Yeah, there’s over like a hundred.
Mignon: It's amazing. And so then when you say that the platforms use NLP (natural language processing) to like pull out words from the videos even—so if you say cottagecore in your video, but you never put it in any of the text-based metadata, it still knows that that's what, what category?
Adam: Yeah. It’s all still metadata. I think there's this outdated notion of just hashtags and titles being metadata, and it kind of comes out of like—it's the same legacy as when people would use SEO (search engine optimization). They'd put keywords in their website metadata to make it rank higher on Google or whatever. We still kind of have that, but we have algorithmic optimization now as well, where you say things, and it's not just the keywords and the hashtags, it's everything. Because the algorithm really is looking at everything. When you upload a video, they simultaneously analyze every visual object, analyze the audio, analyze all the words that are on screen, and they take all that into account when they make their kind of cluster of numbers of what this video represents, and then they decide to send it out to an audience.
Mignon: Is it looking at the titles of the books in my background?
Adam: Maybe those are too far back, but it definitely could pick out that those are books and maybe that this is like a, I don't know, more wordy sort of discussion happening. And it might take that into account.
Mignon: Wild. So if I had a big poster that had certain words on it, then that might even be like more like picked up by the algorithm if it could actually read it.
Adam: It can pick all of that up. Yeah.
Mignon: That's amazing. I never would've thought of that. It is so much more advanced than it used to be. It's just astonishing.
//
MIgnon: You know, you were talking earlier about in-group and out-group, and one thing that I've known but also you covered really well in your book is that so much of what we think of as Gen Z slang and things like that come from African American English and other cultures, and you traced it all the way back to Vine, which is one of the few social media networks I actually haven't been on. Can you sort of talk about the slang pipeline and how words, you know, get in, you know, end up on TikTok and then in regular language for, you know, people like me who, 20 years after it’s cool, say, “Yeet,” like, “I’m gonna yeet myself into the sea?”
Adam: Well, one thing you can't separate here is that the algorithm makes you feel like you're in an in-group, and then it makes you feel like you can use the language that you want to use with your in-group, but it might send that video to another audience. You really have no say once you upload the video who it gets sent to. And there could be a disconnect in the intended audience and the actual audience, and then you, as, I don't know, like a white person, see African American English being used, but you think the video's targeted to you because it's on your For You Page, right? And you internalize that and you start using the word, and I do think the algorithms are like accelerating how language transcends linguistic boundaries, because there's no policing effect, right? If you're in person and you hear a certain in-group's language being used, there might be a reaction to it, and then you're not going to use that language as readily.
But algorithms make it easier for people to think that they're speaking to their community, but then opening up those communities just enough to allow their slang to spread. What you were touching on earlier is there is an element of how slang moves across platforms. My general rule of thumb is, if it's not African American English, it's probably from 4chan. And there's this whole pipeline of like 4chan really was this unique community in creating new words. A lot of those words then shifted to Reddit, and then they shifted to TikTok, and then reached the mainstream.
Vine is actually a really interesting example that you brought up because they were not algorithmic; whatever the most popular post was ranked on top. And so you would have the broad perpetuation of trends like “yeet” and “bae,” and “on fleek.” These were words that were compelling to the general populace. And everybody learned these words at the same time, and they became huge trends. And, you know, as somebody who was in middle school during Vine, it was really everywhere; you could hear it everywhere. But algorithms are different because they personalize the experience, and that means that several different trends and words and memes can incubate at the same time.
So not everybody's feed is dominated by a single “yeet” style video or a single like “on fleek” video. But instead, these separate echo chambers are fostering their own linguistic creativity. And I think there’s more linguistic innovation happening perhaps more than ever before in these separate in-groups. It's really hard to quantify something like that, but there's definitely some compounding effect of algorithmic echo chambers in language creation.
Mignon: You mentioned middle school, which reminded me, I think for “Algospeak,” you did a large survey of middle school parents and teachers, and I was wondering sort of what the most interesting thing you learned from that survey?
Adam: Yeah, I feel like if you wanna actually understand how words are being used offline, you have to start with young people who are usually the ones who are fastest to adopt new language. It's, you know, it's one thing to have our separate online space, this online domain of use, but when the words start to jump boundaries, we really need to be on the ground and seeing how middle schoolers are talking.
And so I interviewed, I think it was 1,700 middle school parents and teachers about these words, and some of them are really being used. So like "unalive" is actually being used in middle schools. There was a student who submitted an essay on Hamlet "unaliving" himself, or thinking about "unliving" himself. There was a classroom discussion on the "unaliving" that happens in Dr. Jekyll and Mr. Hyde. And it's, I think I like "unalive" as an example because I think it's like the prototypical example of how algo speak is really bleeding into our daily lives. And I think this is really going to be how people talk in the future. And you know, you might not like it, but you're gonna die out, and these kids are gonna keep saying "unalive." It might be like slightly ironic, but there is a genuine euphemistic function. Like these kids are using it as a serious synonym for "kill" or "commit suicide" because they feel comfortable talking about it that way.
Mignon: Do they actually even know that adults think of it as slang?
Adam: There are degrees of it. It's hard to tell how much of it is ironic and how much of it is real, but there seems to be some degree of real, and definitely from the parents and teachers I surveyed, a majority of them said that at least, you know, some of their kids did not know that these were online slang words. You know, you just hear your friend who's more online using it, or you think there is an aspect of how young people differentiate themselves from adults, so they want to be using new words. But I think a lot, a lot of "unalive" especially, people, I actually think it's a euphemism in some cases.
Mignon: Other youth. We have so many other euphemisms for death, and kids don't talk about death that much, probably.
Adam: I mean, it's a classic, like we're all, we're all terrified of it. Yeah, "passed away," "kicked the bucket," "deceased." It's all euphemisms all the way down. If you keep tracing back the etymologies of those, like they're also euphemisms of previous words. Any word we have for death at some point was probably a euphemism because we're all terrified of it, and middle schoolers are too, and they're, you know, they're drawing on the same process as we are when we say stuff like "passed away." But it's really happening. There was an incident last summer when the Seattle Museum of Pop Culture did a commemoration for the 30th anniversary of Kurt Cobain's suicide, and they didn't say that Kurt Cobain killed himself. They didn't say that he committed suicide. They said that Kurt Cobain "unalived" himself at 27. So it's really, really being used.
Mignon: Wild. Wild, and it was kind of controversial, wasn't it?
Adam: Yeah, and there's a huge backlash, and they actually took down that placard and replaced it with "committed suicide." But I think it was also the first legitimizing use of "unalive" by a serious authority rather than just these middle school kids. And we might see more and more of it as there's less grounds for a backlash. But I think the backlash was specifically because people are concerned about online words bleeding into the offline, more than just a euphemism for death because that's a normal thing. It's the perception of what the word is.
Mignon: Right. Was there anything else about this survey that you found especially surprising?
Adam: Just the broad, like the speed at which these trends occur. I did a few surveys at different timestamps asking, "Are your kids saying “gyat” and '”rizz” and “skibidi”? And when I first was testing these words out, some of the kids knew what some of them didn't. And six months later, every kid in every middle school was saying “gyat” and '”rizz” and “skibidi.” So I mean, it's kind of...
Mignon: I have to tell you, I tried to search to see how to pronounce "gyat." In your, it's G-Y-A-T and oh my God, don't do it. The things I saw.
Adam: There's a lot of acronyms of gyat too. So they'll say like, "girl, you ate that," or "girl, your ass thick." But it's from an exaggerated African American English pronunciation of "god damn." And so you'll have words like that move between in-groups and then become broader memes. And that's another thing that words are spreading as memes.But that's how these changes are occurring through the algorithm now.
Mignon: Yeah, no, that's another one that came up through African American English. You know, I forgot to ask you. When we were talking about talking so quickly, there was another thing about the way people speak online and there's the supposed TikTok accent, and it has purposes, and if you could sort of describe what the TikTok accent is and then the purpose that it actually serves. I thought that part of the book was really interesting too.
Adam: Yeah, so I started to get into this a little bit, but there is an intentional tactic first and foremost. So when I'm stressing certain words to keep you engaged, that's to make sure you don't scroll away. And if you do scroll away, I lose views, I lose revenue. So I want to speak in a way that's gonna be compelling to you, and everybody does this.
You look at an interview of Mr. Beast, the most popular YouTuber, speaking in real life, and he doesn't speak anything like he speaks in his videos because he’s very meticulous. He figured out exactly what goes viral in his type of videos, and he's like practically screaming at you in those. And that's because it works to keep your attention. It's all about retention, which is how long viewers keep watching your video.
So that's partially how the accents evolve. Now, there are probably a few different ways that you could get people's attention, but there's also an element of prestige, which is how much we value accents. And over time, if an accent is the accepted way to speak online, other people just replicate it. So regardless of whether it is deliberately an intentional tactic, at this point, it's just people see other people talking that way. The videos that appear on your For You page inherently already are successful, and so you're a little bit biased in your perception of what's going on, and we have the replication of these attention-grabbing tactics.
Mignon: Yeah. And then there's the, that sort of ties into the linguistic context of floor holding. Like trying to make sure it's like always your turn. “It's still your turn. I'm still talking. Here we go.” And then the uptalk aspect of that also, that a lot of the online TikTok accent utilizes uptalk at the end of everything. So it sounds like you're not finished, and so that keeps people listening because they think there's more coming. And I thought that was so interesting.
Adam: Yeah, that's literally floor holding. In this case, when I have a short video on your screen, I'm holding the floor from you trying to swipe it up and replace me with another video. But the analogy comes out of you're on a stage and you're trying to keep people's attention, and you gotta hold the floor. Like a comedian, it's not just about the jokes; it's about the timing and the pacing and how you're interacting with the audience. So this has always been. I'm not, I haven't been saying anything of what I've been saying to fearmonger. We're using time-honored linguistic tactics like euphemizing death or trying to keep people's attention. This is just a normal thing we do. I just do think it's accelerated by the algorithm.
Mignon: Absolutely. Well, at the end of our Tuesday show, we always play a familect, a family dialect from one of our listeners. And so to finish up, I thought it was really interesting and cool in your book that you talked about fanilects, which I have heard before, but then there's like some sort of ludolect that is from gamers. So if you could talk about those different aspects and what purpose they serve for online people, and I'd love to know where ludolect comes from, even, that word.
Adam: Well, "ludo" from Latin "ludu" meaning play. But yeah, like what I was saying, anytime humans communicate, we form in-groups and out-groups, and we like to behave more similarly to the people that we find similar to us. It's called “homophily,” loving the same. And so when you're in a gaming community, you'll also have, like for video games, accents perpetuate for the specific function of the game. So there's like a Fortnite accent now where people will kind of have these sharp, kind of cutoff ways of talking because you need to communicate quickly, and there's certain slang words and certain pitches that they use that are ... again, there's an aspect of prestige, accents replicating themselves simply because other people are using them. But it's also an aspect of community. It's the fact that this is how your friends talk, so you're gonna talk that way. That's what we have with the influencer accent. That's what we have with familects; you talk more like your family. When a group of people is together, they're gonna behave more similarly to each other. And you also mentioned fanilects, which is how fans talk similarly. If you're deep into like Swifty culture or K-pop culture, the deeper you get, the more exclusive the language becomes. And they start speaking in a language no one else can understand, really. That's what a fanielect is. And there's a growing kind of study of fanthropology. Another fun word.
Mignon: Awesome. Well, Adam Aleksic, thank you so much for being here. The book is “Algospeak.” I thought it was fascinating. If you're interested in how things work online, if you watch a lot of online video and want to understand how you're seeing what you're seeing, or if you're interested in language change, I think you'll love the book. Adam, where can people find you?
Adam: Well, thank you, Mignon. I'm EtymologyNerd on all platforms, and I guess the book is a new way you can find me.
Mignon: Wonderful. Thanks so much. This is the end of the show for regular listeners, but Grammarpaloozians, you're gonna get a bonus episode. We're gonna talk about some of my favorite videos that Adam has done about different words and their meanings, etymology, we're gonna dig into his Etymology Nerd stuff. For everyone else, that's all. Thanks for listening.