Modern WisdomTHEY’RE BRAINWASHING YOU! (& other secrets that made you click) - Etymology Nerd
CHAPTERS
“Word of the Year” as virality marketing: the meaning of nonsense
The episode opens by dismantling “word of the year” announcements as marketing tactics rather than linguistic authority. The guests use “six seven” as an example of a deliberately empty phrase whose real function is to provoke curiosity, clipping, and algorithmic spread.
- •“Word of the year” selections are framed as a marketing play by “big dictionary”
- •“Six seven” is a self-aware meme designed for clip farming and distribution
- •Absurdity itself can carry meaning: a meta-commentary on the information ecosystem
- •Words like “rage bait” and “slop” reflect platform incentives more than linguistic decay
TikTok as a global slang engine: accelerated cycles and echo chambers
They argue TikTok is currently the most powerful driver of slang creation and diffusion. The platform interface, comment culture, and algorithmic trend loops speed up adoption and turnover of new terms.
- •Platform share of slang origins has shifted heavily toward TikTok (and Twitter/X)
- •TikTok fosters a sense of ongoing conversation that invites linguistic participation
- •Echo chambers and algorithmic repetition compress slang lifecycles
- •Negative feelings about social media get projected onto the words themselves
Platform dialects and micro-communities: Twitter vs LinkedIn vs fandom speech
The conversation broadens into how each platform shapes expectations of tone and vocabulary, creating recognizable dialects. They emphasize that within-platform subcultures (K-pop, Swifties, etc.) produce even finer-grained micro-dialects.
- •Platforms function like “houses”: you code-switch based on expected norms
- •LinkedIn encourages professional register; Twitter rewards play and edge
- •Within any platform, fandoms and subgroups develop distinct lexicons
- •Slang is also an in-group identifier that signals belonging
Keyword virality and “algorithm wink” language (six seven, incel lexicon, ‘maxing’)
They unpack viral jargon as a set of algorithm-friendly keywords that creators and users deploy to trigger distribution and recognition. Some terms are harmless meta-jokes; others are tied to more toxic subcultures, but the mechanism is similar.
- •Certain phrases are used primarily because they’re clip-and-search friendly
- •“Keywords” act as algorithmic handles and social identity markers
- •Harm level varies (innocuous meme vs harmful subculture language), but distribution logic persists
- •Talking about a viral figure can itself become a virality hack
The influencer voice: founder effects, relatability vs authority, and floor-holding
They analyze “influencer accents” as evolved performance strategies shaped by early successful creators (founder effect). Lifestyle influencers aim for parasocial warmth, while educational creators use speed, stress, and clarity to project authority.
- •Influencer speech styles trace back through a lineage (Kim K/Paris Hilton → YouTube → TikTok)
- •Lifestyle influencer cues: uptalk, vocal fry, softness, familiarity
- •Educational influencer cues: faster pace, stressed keywords, authority framing
- •Uptalk and fillers function as floor-holding tools to prevent audience drop-off
MrBeast, livestream ‘edging,’ and retention-first vocal performance
MrBeast is presented as a deliberate vocal code-switcher: calm in interviews, high-arousal in videos. They connect this to livestream formats that continually delay payoff, using language and pacing to keep viewers from scrolling away.
- •MrBeast voice: loud, urgent, “shock and awe” tuned for attention and younger viewers
- •Bounded videos structure payoff; livestreams sustain endless anticipation
- •Retention optimization influences vocal delivery (no dead air, constant hooks)
- •Auditory clickbait parallels visual clickbait mechanics
Broadcast voices: why newscasters and sports commentators sound engineered
They compare broadcaster speech to influencer archetypes: newscasters resemble “educational authority,” while sports commentators resemble high-excitement performers. These norms persist because audiences expect them and newcomers imitate proven templates.
- •Broadcast registers are conditioned by medium and expectation, not “naturalness”
- •Newscaster voice evolved via training toward standardized accents
- •Sports commentary emphasizes excitement, clarity, and continuous engagement
- •Founder effects explain why one successful style becomes institutional default
Distribution over content: TED Talks, clip farming, and viral misalignment
They argue modern media rewards distribution mechanics more than message quality. Viral spread is biased toward high-arousal emotions (anger, fear, awe), creating a mismatch between what’s good for people and what platforms amplify.
- •“Clip farming is the future”: content gets atomized into shareable fragments
- •TED Talk prestige declines in an oversaturated ecosystem; distribution dominates
- •Algorithms reward arousal; ‘warm and fuzzy’ content struggles to spread
- •Wellness content often becomes performative “aesthetic wellness” rather than real well-being
Can you hear sexuality? Gay speech cues, lesbian accent uncertainty, and coded identity
They explore research and stereotypes around identifying sexuality by voice, noting gay male speech is more recognized than lesbian speech in studies. They frame such features as identity signaling shaped by history, safety, and community norms—not monolithic traits.
- •Gay male speech patterns are recognizable to many listeners, but vary across communities
- •Historical need for covert signaling produced slang and micro-languages (e.g., Polari)
- •Gen Z slang pipeline includes ballroom culture and LGBTQ+ spaces
- •Lesbian accent research exists but results are mixed and less definitive
Emojis as language: substitution, tone-tags, and legal ambiguity
The episode treats emojis as meaningful linguistic units used for censorship evasion and emotional framing. They highlight how emoji meanings shift quickly, creating real-world confusion—including court cases hinging on interpretation.
- •Emoji functions: word substitution (e.g., ice cube for “ICE”), tone-tagging, reactions
- •Meaning is context-dependent and rapidly evolving across age groups
- •Legal disputes show emojis can constitute agreement or imply intent
- •Shifting norms (cry-laugh as ‘boomer’/ironic) demonstrate semantic drift in real time
Etymology as a mirror of reality: shortening, loss, and youth-driven change
They discuss whether language has a direction, concluding it mainly tracks changes in lived experience. Youth are described as the main engine of slang innovation due to identity formation and the desire to diverge from parents, while institutions mostly legitimize after the fact.
- •Language change reflects shifting realities (e.g., fewer plant words as people interact with fewer plants)
- •Processes: contraction/truncation (God be with you → goodbye → bye) and blending/portmanteau
- •Young people (roughly 10–25) drive slang; institutions typically ratify later
- •Top-down forced language often fails due to resistance (‘fetch’ problem)
Filler words and “in medias res” hooks: like, you know, and creator openers
They reframe filler words as functional tools for turn-taking and maintaining attention. Creator patterns like starting with “No, because…” manufacture immediacy and pull the viewer into a story already underway.
- •Fillers are universal; central vowels like “um” appear across languages
- •“Like” has complex functions (e.g., quotative like) and is stigmatized socially
- •Turn-taking cues (“you know,” “well…”) help seize or hold the conversational floor
- •Hook openers create an ‘already in progress’ feeling that boosts attention
AI’s linguistic fingerprints: ‘delve,’ em dashes, and humans learning from models
They argue AI is not just generating language but feeding back into human usage patterns via writing assistants, politicians, academia, and platforms like LinkedIn. “Delve” becomes a case study in how training and reinforcement biases propagate into real speech.
- •Post-ChatGPT spikes in certain words (e.g., “delve”) suggest model-to-human influence
- •Bias sources include reinforcement worker norms and prestige/Latin-leaning vocabulary
- •AI-written texts create second-order influence: you read AI without knowing it
- •Counter-signaling emerges (avoid em dashes/‘delve’) but deeper shifts may be harder to notice
Social media vs AI: bottlenecks, homogenization, language death, and shaping thought
They conclude social media is the bigger driver because it captures and amplifies everything, including AI outputs. Concerns broaden from words to ideas: algorithms shape the Overton window, incentivize manipulation, and may contribute to homogenization amid rapid language extinction.
- •Social media is the primary constraint system; AI effects often route through platforms
- •Algorithms create linguistic bottlenecks while also enabling new niche outgrowths
- •Mass language loss (often cited as one language dying every ~2 weeks) reduces expressive diversity
- •Language influences discourse norms (Overton window) even if it doesn’t fully determine thought
Gen Z as a constructed label: identity buckets, ‘umwelt,’ and resisting commodification
They challenge the reality of generations as natural categories, calling them modern marketing constructs that people are nudged to perform. The discussion ties back to language-as-identity and the tension between individuality and belonging.
- •Generational labels are recent social constructs used for consumer segmentation
- •Labels can be ‘violent’ by forcing identification for/against a category
- •Concept of ‘umwelt’: each person has a unique perceived world (and idiolect)
- •Social media intensifies bucketization (cottagecore/Swiftie/Gen Z) and interchangeability
Rapid-fire etymology and playful linguistics: word origins, conlangs, and QWERTY
The episode shifts into a fast, entertaining segment on surprising word histories, then expands into constructed languages and design constraints. They use QWERTY and Esperanto/Ithkuil to illustrate that “efficiency” is not language’s only goal—human bonding is.
- •Word origin lightning round (muscle/mouse, salary/salt, assassin/hashishin, etc.)
- •Conlanging as exploration: dolphin/bird sound systems and what “language” could be
- •QWERTY as an intentionally anti-jam (historically inefficient) design that became standard
- •Highly efficient conlangs (e.g., Ithkuil) show tradeoffs: density vs learnability/human connection
Does ChatGPT speak English? Tokenization, embeddings, and why meaning can distort
They explain how LLMs transform text into tokens and mathematical embeddings, then back into text—suggesting the model isn’t “speaking” in a human sense. This pipeline clarifies how subtle biases or distortions (word choice, tone, ideology) can slip in and scale.
- •Input text is tokenized, mapped to numbers, embedded in high-dimensional space
- •Model predicts next tokens statistically rather than holding human-like meaning
- •Meaning can be altered through encoding/decoding and reinforcement pressures
- •The output can become a ‘mathematical dialect’ not identical to any human idiolect
Wrap-up: where to follow Etymology Nerd and the core warning about attention systems
They close by pointing viewers to Adam’s Substack and book ‘AlgoSpeak,’ reinforcing the episode’s theme: language is being reshaped by attention incentives and intermediaries. The send-off mirrors the earlier discussion of performative communication and media-aware sign-offs.
- •Adam points to Substack (Etymology Nerd) and book ‘AlgoSpeak’
- •Core thesis recap: platforms monetize attention and shape language/ideas
- •Media literacy and skepticism are framed as necessary defenses
- •Final sign-off nods to “movie phone call” minimalism (click/end)
