THEY’RE BRAINWASHING YOU! (& other secrets that made you click) - Etymology Nerd
In a Nutshell
"67" was dictionary.com's 2025 Word of the Year as a viral marketing ploy exploiting absurdity and algorithms for engagement, mirroring rage bait and slop, while TikTok dominates slang innovation through platform dialects, influencer accents (uptalk, vocal fry), and subcultures like manosphere/4chan terms signaling identity. Language evolves rapidly via youth-driven slang pipelines (AAVE → gay ballroom → mainstream), AI biases (e.g., "delve" spike), and memetics, but homogenizes under social media/AI constraints, prioritizing virality over truth—urging media literacy against manipulation in an attention economy where performance trumps authenticity. Etymology reveals identity truths (e.g., "muscle" from "little mouse"), with conlanging and animal signals highlighting language's ritualistic, human essence amid accelerating change.
These notes were generated by AI and may contain inaccuracies.
67 was voted word of the year in 2025 by dictionary.com. Dictionaries choose words of the year as a marketing ploy to sell more dictionaries. 67 references a viral trend started by basketball player Talon Kenny. Saying "67" aims for clips to go viral for personal gain. Dictionary.com, Connecticut house representative Bill Buckby, Gen Alpha kids like the 67 kid, and others cashed in on it. Clip farming is the future of online distribution.
67 is vacuous, designed to incite "what does that mean?" Absurdity has meaning, critiquing the information ecosystem. Its absurdity is its definition—a meta word, a knowing wink to the panopticon. Oxford's 2025 word was rage bait, also rage baiting for controversy to drive engagement and awareness of Oxford dictionaries. Slop describes sloppy language, prompting complaints about the state of language.
Language is a tool; nothing inherently good or bad in words. Negative associations from social media are cast onto language. Videos may be slop, but words aren't bad for your brain.
TikTok is the most powerful linguistic engine on Earth, shaping language more than anything. 2022 KnowYourMeme study: word origins shifted from 4chan, Reddit, Twitter to mostly TikTok and Twitter (still Twitter). TikTok fosters linguistic innovation from user interface and conversational feel. Echo chambers and algorithmic trends accelerate slang cycles.
Platforms have dialects like houses: different speech expectations. LinkedIn: professional language. Twitter: linguistic play (e.g., jester gooning). TikTok: fandom language. Micro-dialects within platforms (K-pop, Swifties).
"Mid jesteru maxing at the club when the fuds come over. Is it better to be mogging with the bros?" Fallout mirrors 67: knowing wink to algorithm for virality, cashing in on Clvicular's fame (human 67). Keywords like maxing, gooning push through algorithm, resonate personally, signal in-group belonging. Language signals identity and cohort membership.
Influencer speech has defining characteristics. Consulted on Austin court case (sad beige lawsuit): minimalist influencer sued another for stealing vibe, including accent and brown clothing style—claimed as IP. Neither original; homage to older influencer traditions.
Modern "hey guys, welcome to my podcast" traces to Kim K, Paris Hilton, early beauty YouTubers, filtered to TikTok. Features uptalk and vocal fry due to linguistic founder effect—following predecessors.
Social signaling: "I'm an influencer," performing relatability to young women. Optimized for algorithm: drags out words for retention; dead silence bad. Uptalk lengthens final vowel (floor holding) to hook audience, implying more to come. Down talk signals "I'm done," prompting scroll away.
Filler words like "actually" hold floor. Christopher Hitchens tactic: halfway through sentence, pause for sip/thought, continue to keep attention. Uptalk creates unfinished feel.
Educational influencer accent: stresses words, faster pace for scripted videos to hook attention. Lifestyle: soft, welcoming, cozy parasocial familiarity. Educational: sharp, authoritative staccato for teacher-like trust.
Vowels give words color; consonants structure. Northeast UK glottal stop drops T's (e.g., "bu'er" instead of "butter"). Restoring consonants adds clarity, clippiness. Modifying style for online performance creates homogenization. No dialect inferior (North England vs. received pronunciation, transatlantic). Indian creators code-switch to British/American accents due to status bias.
Strong accents (Geordie, Scouse) near speciation. Language evolves like biology: bottlenecks (algorithm compresses to recognizable accents), then speciates in platforms/fandoms.
MrBeast's video accent differs from interviews: screams ostentatiously ("I just bought a private island. I'm giving away a million dollars") for 14-year-old attention spans—excitement, loudness, shock and awe to dumbfound viewers.
Live streams: permanent edging for payoff. Beast videos build to payoff within bounds (e.g., escalating gyms). Live performers edge: dance, collect money, promise backflips. TikTok lives (egg peeling) delay gratification. Videos script resolution at end.
Manosphere language interesting: half Gen Z slang from African-American English or 4chan (incel words like maxing, pled, gooning). 4chan incubator due to anonymity: slang proves non-normie status. Text-only constrains to quick identity signaling via language for in-group belonging. Memes/images aid, but no selfies. Funny language spreads.
Newscasters use authoritative broadcaster voice (mid-Atlantic/Midwestern accent trained historically) like educational influencers. Founder effect: follows predecessors (thumbnails, video styles). Educational accent influenced by Hank Green, Vsauce, Bill Nye.
Sports commentators like MrBeast: excited, clarity to keep viewers engaged in game.
Gymshark ad: hybrid shorts (onyx, gray, navy), geo seamless t-shirt—lightweight, sweat-wicking, fast-dry, 30-day returns, international shipping, 10% off at jim.sh/modernwisdom with code modernwisdom10.
Speaking with clarity and expressiveness involves minimizing filler words, as seen in trained speakers. To maximize relatability, authenticity, and belonging, avoid overly precise language that feels contrived. Strike a fine line between naturalism and conciseness. Phone calls in movies never say goodbye; they end abruptly to save screen time, performing the idea of a phone call without real-life formalities like "All right, I'll see you later. Goodbye."
For a TEDx talk 5-6 years ago, preparation involved draining knowledge of the talk to avoid sounding contrived or performative. TED talks have a recognizable cadence. There's a YouTube video of a guy delivering a TED talk with meaningless content but perfect cadence—recommend looking it up. TED talks have lost prestige; now, clip farming is the future. If booked for a TED talk, focus on clip farming for distribution.
TED talk audiences are aspirational speakers, not ideal for real messages. Transmit online to reach the right audience. 67 and jester maxing emphasize distribution over content. In social media virality, TED talks are a dead format due to oversaturated media ecosystem and fewer views. One video can be broken down like a book sold sentence by sentence, reproduced endlessly.
Example: HH bomb incident in Miami club with Kanye song and insane bar—every second broken down, shared, republished, commented on.
Dangerous misalignment: human preferences vs. what goes viral. Online rewards anger, fear, awe, humor—emotions triggering mental arousal. Contentment feels warm but doesn't activate brain for likes. Likes measure button-clicking willingness, not true liking. Warm, fuzzy ideas don't spread; rage bait and clickbait do.
Meditation teachers reduce brain activation for calm, getting less engagement and virality, disincentivizing content creation.
Wellness influencers perform hyper-aestheticized, sanitized "clean girl" wellness (yoga mat, Pilates), not true wellness. True wellness is relaxing on the couch, watching a movie—not visually compelling.
Gay accent recognizable; adopts strategies similar to women's speech (raised vowels), differing by community. Gay slang (slay, serve, queen, cooked, ate) influences Gen Z; "cooking" from sports lang, "cooked" from gay ballroom speech.
Tops vs. bottoms or masculine vs. feminine presentation may have linguistic identifiers. Straight men notice feminine-coding in gay men. Everything is performing the idea of something: gay men perform gayness, straight men perform straightness.
Historically, closeted gay men used cues to signal to others. Examples: Polari (UK micro-language to evade police), cottaging (foot-tapping), swardspeak (Philippines), South African gay languages—subversions of norms due to marginalization.
Capoeira: Brazilian martial art developed under military rule/slavery bans on fighting. Looks dancy with music to disguise as a party.
Language ties to identity and power. English norms from 1700s white British; others subvert (AAVE slang → Gen Z, gay ballroom slang like king/queen from 1980s NYC black/gay/Latino ballroom elevating community status).
Slang pipeline: black people → gay people → straight white girlfriends → mainstream ("four-person human centipede with normies at the end"). Or 4chan → black people → gay people → cool people → normies. Language spreads when funny or cool.
Hood irony memes parody AAVE (e.g., cap, huz from exaggerated "goddamn").
No monolithic language; gradient influenced by background. Idiolect (Greek idios = one's own): unique linguistic footprint from upbringing, education, peers. Unabomber caught via brother's recognition of "eat your cake and have it too" (logical reversal of common phrase).
Everyone has quirks.
Lesbian accents harder to identify; studies show identification possible, possibly masculine traits (lowered vowels, deeper voice), but mixed results.
Gay fashion more flare/color/accessories (feminine-coded); lesbian fashion more plaid/jeans/loose (male-coded). Language matches presentation, but distinct—not imitating opposite sex to attract partners.
Read Erving Goffman's The Presentation of the Self in Everyday Life (1960s): we adopt front-stage (public roles) vs. back-stage (casual). Adjust register by audience (parents vs. friends, podcast vs. off-mic). Manhood is performance (beards, dress, speech as masculinity symbols).
Masculinity caricatured: visually, action figures like Luke Skywalker evolved—'60s skinny to '90s jacked (search "ChatGPT Luke Skywalker action figure over time comparison"). 37-inch shoulders/32-inch waist → 52-inch/27-inch (inhuman proportions).
Masculinity evolves: ancient Rome, topping men was masculine (Caesar offended at bottom accusation). Long/short hair, togas vs. 2026 ideals—arbitrary, moving target.
Emoji functions:
- Substitutive: e.g., ice cube for "ICE" to evade algorithms.
- Paralanguage: augment tone (laughing/crying/sneezing emoji as tone tags).
- Reactions or standalone.
Linguistic: carry meaning. Court cases interpret:
- Canadian farmer grain contract: thumbs-up emoji ruled as agreement (context of prior yes-signoffs).
- Murder case: emoji intent interpretation.
Fluid meanings: crying emoji → laughing → ironic; laughing emoji now cringe (boomer signal unless ironic). Post-cry-laughing world.
Language reflects current human identity. Etymology from Greek etumos (truth)—reveals truths about us as a tool of identity. Less directional, more mirroring people.
Language lacks a fixed direction and instead reflects current feelings and identity. It categorizes reality and communicates it. If reality changes, language evolves to describe the new reality. Fewer words describe plants today compared to the 1800s due to less interaction with them.
Contraction shortens phrases like "God be with you" to "goodbye" to "bye." Words can also form portmanteaus by combining, such as "podcast" from iPod + broadcast. New words like "jester maxing" emerge, reflecting new realities. Language is a moving, living thing; old words are lost as new ones appear.
Young people, especially ages 10-25 like middle schoolers, drive most slang changes to build shared identity and differentiate from parents. Older people resist due to cemented language ideas. Institutional words like "iPhone" or "podcast" come from top-down sources but slang bleeds into mainstream from youth, including video game terms (NPC, skill issue), Black slang, and incel language.
Institutions struggle with top-down language dictation. Forced words like "fetch" in Mean Girls fail due to resistance. Dictionaries and news legitimize existing words rather than force them. Words must circulate long enough for institutional acceptance.
Filler words are universal; "um" or mid-central vowels like Spanish "e" provide thinking time—a holding pattern or floor-holding. "Like" is stigmatized from Valley Girl association but serves functions: quotative ("I was like...") for indirect quoting or persona adoption—a self-simile implying approximation.
Books and Influences
The Etymologicon by Mark Forsyth sparked interest in etymology in 2016, leading to a blog, linguistics studies, and becoming a linguistics influencer. Elements of Eloquence also praised.
Grandparents' generation used "well you know" or bloating phrases for padding and breathing room. Modern equivalents: "yep," "no worries." These signal turn-taking or landing thoughts. Influencers start with "no because" for in medias res effect, grabbing attention by simulating mid-conversation.
Creators exploit attention economy: platforms monetize attention, commodify data for ads. Language prioritizes attention via SEO—every word signals algorithms for video clustering and viewer priming. Metadata now permeates all content.
"The medium is the message."
Algorithms constrain language, ideas, and reality through biases, similar to AI bottlenecks.
"Delve" usage spiked 1000% since ChatGPT (pre-2022), as AI uses it 10x more. Reasons:
- Reinforcement workers in Nigeria/Kenya favor familiar words.
- Latin bias ("delve" Latin-derived vs. Germanic "dig in") for prestige and sounding knowledgeable.
AI trains to sound confident, sycophantic; discrepancies amplify. Humans now use "delve" more in speech.
Other AI-favored words: commendable, meticulous, crucial, potential, significant—all Latin-derived, increasing post-ChatGPT.
Examples of AI Influence
- British politicians say "I rise to speak" (US colloquialism) in House of Commons, indicating AI-written speeches.
- 13% of research abstracts AI-aided.
- LinkedIn bullet points, negative parallelism ("not X, but Y").
- AI's em dash (M-dash) for segmentation; good writers counter-signal by avoiding to seem human/unpredictable.
Counter-signaling: Ozempic makes natural weight loss less high-status; expect pushback against AI markers like "delve."
AI shapes reality via political/gender/racial biases (e.g., Elon tweaks Grok). All platforms prioritize speed over alignment for profit. Anthropic praised for safety efforts but question all intermediaries—medium constrains speech.
AI/social media homogenize language. A language dies every two weeks; 7,000 languages, most gone by century's end—mass extinction. Lost expressions like Potawatomi "to be a Saturday" (embodying Saturday as verb). Languages vary: agglutinative (Turkish, German) tack morphemes for niche concepts (e.g., German migratory bird restlessness); inflectional (English, French) change forms.
Social media more concerning—it captures AI effects and accelerates via artificial networks mimicking human diffusion (ideas as viruses, per Malcolm Gladwell). Enables rapid misinformation, "flooding the zone." Financial incentives (memecoins, Polymarket) push viral ideas. Remain skeptical; touch grass.
Memetics studies meme language and network diffusion; reverse-engineering occurs.
They used a burner account on Reddit, possibly the UFC subreddit, to post a video supposedly of CCTV footage of Francis Ngannou trying to get into a dry cleaners. The door was locked, he smashes it, and the whole door breaks. The entire thing was constructed: they filmed it with fake glass. Francis Ngannou was wearing a Gymshark logo. The burner account posted it and disappeared.
People close to the ground floor lack resources of a country but can shape language. Governments have the ability to shape language top-down, but people push back if it feels forced. Advice: get offline as much as possible while remaining aware of online happenings. Even offline, algorithm ideas spread through friends. Be highly media literate, knowing information is manipulated by actors with vested interests. This effect amplifies as more people exploit the information ecosystem.
Warning: language is being shaped by malicious actors, companies, and foreign governments with dashboards tracking populations and idea clusters on social media. They seed ideas to spread effectively in active information warfare campaigns.
Living in algorithmic constraint systems changes easily expressible thoughts, feeding back into capacity to think. Reference to 1984: language influences thought (linguistic relativism), but language does not fully determine thought.
Not heading to 1984; more like Brave New World by Aldous Huxley, where people entertain themselves into unawareness of dictatorship via content and drugs. Language cannot be truly constrained: humans are tenacious, creating algospeak like ice emoji or "unalive" to evade censorship.
Some ideas are harder to express; Overton window is the range of acceptable discourse. Example: gay marriage moved from unthinkable to acceptable via gay liberation movement. Now, Overton window moves toward looksmaxing, with rising interest and ideas like using Ozempic (zenic).
Consensus reality is what we think others think is acceptable. Increased discourse on alt-right or looksmaxing ideas shifts the window. Thoughts are not constrained, but consensus reality moves.
Steelman argument for Gen Z being different: every generation thinks kids destroy language, but older people die, kids grow up, cycle repeats.
Generations do not exist; they are weird social constructs. Familial generations exist (older/younger), but broad social cohorts started post-WWI with Lost Generation, then Silent Generation, GI Generation, Baby Boomers, Gen X (meaning unknown), Millennials, Gen Z, Gen Alpha. All made up, commodified for consumer demographics.
People in 1800s did not perform for social cohort; now, self-branding constrains via labels (Gen Z, cottagecore, Swifty). Labels force identification or opposition, combining into identities. Every person has unique idiolect and identity.
Linguist has Umwelt tattooed: the world as perceived by a particular person. Everyone sees the world uniquely; social media buckets make people interchangeable despite desire for individuality and belonging.
Tension: desire for autonomy vs. kinship/support. Child's life cycle: kinship need until 11-12, then rebellion at 13-14-15, coinciding with new slang creation.
Reminiscence bump effect (or memory bump): language, music, movies from ages 12-16 lock in lifelong preferences; highly formative as sexuality comes online. Incest avoidance (Westermarck effect) drives independence from family.
Single frameworks are reductive (e.g., Freudian lens, looksmaxing's focus on attractiveness). Adopt multiple frameworks; too many can be chaotic. Framework maximalist: broaden worldview, consume diverse media (algorithms, regular news, etc.) to counter biases, touch grass for better reality picture (polyconsumption, poly-framework approach).
Online, single explanatory frameworks sound convincing/sexy due to conviction. Human brain prefers simple stories; reality is complex/interrelated.
Muscle: Latin musculus (little mouse), as it looks like a mouse moving under skin.
Salary: Latin sal (salt); wages paid in salt.
Assassin: Arabic hashish (marijuana consumer). Sect in Persia (Hashashin) got high and killed people.
Candidate: Latin candidus (white-robed), associated with purity for noble office.
Girl: Originally gender-neutral for any young child, could mean boy.
Silly: Old English sælig (blessed/fortunate) drifted to innocent, naive, foolish. Semantic drift like awful/awesome, terrible/terrific/terrifying.
Avocado: Nahuatl āhuacatl, also meant testicle.
Nightmare: Germanic demon (mare) sat on sleepers' chests; mare like horse.
Goodbye: "God be with you."
Penguin: Welsh pen gwyn (white head), originally for great auk, repurposed.
Orange: Sanskrit for orange tree, via Arabic naranja; "n orange" became "orange."
Etymology entered for fun facts. Inventing languages (conlanging, constructed language): e.g., Dothraki, Klingon, Esperanto (attempted global, evidence-based, precise/easy language).
Studied conlanging at MIT: built dolphin language (whistles/clicks, e.g., shark = specific sound), went viral on TikTok. Also bird language (whistled); whistled registers exist in human languages.
AI bird song analysis: Cornell tracks migrations. Animal communication ≠ human language; contextual, multimodal (whale body rolls, dolphin sonar). Signals presence/emotional intent, not abstract meaning. Hard to define language; any definition is a framework.
The QWERTY keyboard originated from typewriter jamming issues. Early layouts like ABC caused jams, and alternatives like Dvorak were tried, but QWERTY prevented jamming by placing the most used letters on the edges, away from each other, to avoid simultaneous triggering. It was purposefully built to be slow and inefficient. Even when attuned to it, other keyboard arrangements can make people 30% to 50% faster.
Languages are not just for efficiency but to be beautiful, illustrative, and communicative. A flaw in discussing language is viewing it solely as information bits transferred per second, which can be categorized and commodified. Language also serves ritualistic bonding, like small talk or saying goodbye at the end of phone calls, which establishes social ties and adds humanity—though not ideal for algorithms.
Different languages transfer similar bits per second despite variations: Japanese has more syllables per second than Thai, but Thai is more inflectional with tones, building more on syllables, so speakers adjust pace to convey the same information.
Ithkuil is a hypothesized constructed language (conlang) designed for maximum information density, carrying many meanings efficiently. No native speakers exist; it's impossible for humans to learn because language connects humans to describe reality in understandable ways. Robots could use more efficient languages.
Computers don't process language like humans. AIs like ChatGPT predict tokens statistically, not truly speaking English: input words are broken into tokens (word segments), paired with numbers as high-dimensional embeddings (like coordinates in thousands of dimensions), processed through neural networks, then output tokens converted back to language. Meaning can get lost, leading to misaligned representations like over-representing certain elements or improperly encoding natural speech.
This may be the time in history when language changes most rapidly due to massive influences and broadcasts. Measuring language change is hard since language is undefined—everyone speaks their own version, and languages are spectra of similar speech. ChatGPT is trained on a homogenized English corpus, creating a mathematical representation of speech that's not how anyone actually speaks.
Etymology Nerd promotes longer-form content: Substack at etymology nerd, book Algo Speak on how social media changes language, and social media as etymology nerd.
Keep Chris Williamson in your library
Save the videos and channels worth coming back to, and find them again in one place.
