Inside The Viral Words That Make You Click - Etymology Nerd - #1086

“6.7” got named word of the year, but it feels like a word that means nothing on purpose. Is that the point?

Big dictionary knows what it’s doing. “6.7” is a meta-trigger for clips and outrage, a wink at the algorithm that rewards anyone who says it—including the politicians and the athlete who sparked the trend. The joke is the system.

Oxford picked “rage bait,” others pushed “slop.” Feels engineered to rile people up.

Exactly—those choices are built to provoke comments and shares. We project our frustration with social feeds onto words, but the vocabulary isn’t rotten; the incentives are.

Is TikTok now the main engine of new language?

Yes. A shift from forums to TikTok and Twitter sped everything up. Each platform is a “house” with its own voice, and inside them are micro-dialects like K‑pop or Swiftie slang.

That viral “maxing” mash-up—what did it reveal?

It works like “6.7”: sprinkle the right keywords and you hitch a ride on the trend. Some of it’s harmless, some isn’t, but the algorithm doesn’t care—keywords are fuel.

Using those words also flags who you’re aligned with.

Totally. Word choice is identity work.

Break down the influencer accent.

I even consulted on a lawsuit about a copied “vibe.” Today’s lifestyle voice traces back to reality TV and early YouTube; the founder effect keeps its uptalk and vocal fry alive.

What is that voice trying to achieve?

Belonging and retention. Drawn vowels, uptalk, and fillers keep the floor while you think—classic “floor holding,” which existed long before apps.

Uptalk signals you’re not done, so people don’t bounce.

Right. A falling pitch says, “we’re finished,” which online reads as “scroll.” For education, I switch to a quicker, crisper cadence to project authority.

Diction helps too. Vowels add color, consonants give shape; drop them and speech gets mushy.

Algorithms pressure accents toward a standard for reach, so creators often code‑switch to prestige varieties. That leaves many authentic voices sidelined.

And the algorithm acts like a filter on what gets through.

Yes—one big bottleneck, then speciation into niche dialects in subcultures.

How does MrBeast’s delivery fit this?

It’s pure adrenaline—loud, punchy, relentless—to stun you into staying. Different audience, different voice.

Live streams keep promising payoff but not yet.

Same showman trick as street acts: delay the peak, sustain tension. YouTube scripting often withholds the answer until the end for the same reason.

Which online speech communities intrigue you most?

A lot of youth slang traces to AAVE and 4chan. Anonymity and gatekeeping made 4chan a hotbed for coining terms that later leak outward.

Newsreaders and sports commentators also have “voices.”

News mirrors the educational tone; sports mirrors Beast-like hype. Both spread by founder effects and formal training.

Over-sanitized talk feels fake, though.

Media trims real-life fluff for pace. TED’s cadence became cliché; now talks are optimized for clip potential.

So distribution beats message?

On platforms, arousal wins. Anger, awe, fear, and jokes spread; calm content rarely compels a tap, so “wellness” gets staged as glossy performance.

What about the gay male accent and its slang pipeline?

It’s not one voice, but shared cues exist. Ballroom culture fed words like slay and serve; language is both a shield and a beacon, from Polari in Britain to Swardspeak in the Philippines.

And slang usually moves from black and queer spaces outward.

Often, yes. Also, every speaker has a fingerprint—an idiolect—which is how the Unabomber got flagged.

Is there a clear lesbian accent?

Listeners often detect it, but features vary and studies disagree. It resists neat rules.

Are we all performing roles?

Goffman nailed it: we swap faces by context. Even masculinity is a moving costume—from Roman norms to today’s look‑maxed heroes.

Do emojis count as language?

They sub in for words, tag tone, or act as reactions. Courts have even treated a thumbs-up as agreement. Meanings morph fast—today’s cry-laugh reads “boomer” unless used playfully.

Do words naturally get shorter?

Language maps reality more than it follows a single arrow. We contract, blend, and drop words as life shifts; new tech spawns labels while slang remixes below.

Who really drives slang?

Young people. Institutions coin formal labels, but teens set the vibe.

Can authorities force new terms successfully?

Push too hard and people resist. Institutions mostly bless what’s already common.

Filler words have always been with us.

That central vowel shows up across languages when we think. “Like” can quote a mood, not just a sentence.

Creators grab attention by opening mid-thought—“no, because…”

It seizes the floor instantly. Today, every word is metadata for the algorithm, shaping both how it routes content and how viewers read it.

How is AI changing language use?

Models overproduce prestige Latin-leaning terms like delve, partly due to reinforcement choices. We see those tics in posts and even speech as people unconsciously imitate.

Writers now avoid obvious AI tells like double dashes.

Delve is just a mascot; lots of fancy synonyms climbed. The real risk is hidden social biases getting normalized at scale.

Bigger picture: homogenization and language loss?

Languages are dying quickly, taking unique concepts with them—like treating Saturday as a verb. Fewer systems means fewer ways to think and speak.

Why do some languages pack complex ideas into single words?

Different machinery. Agglutinative languages stack pieces; inflected ones reshape forms, so some express long ideas compactly.

What worries you more for language: social media or AI?

Social media. It weaponizes spread, floods zones, ties money to memes, and invites info‑ops to seed narratives.

Does this choke what we can say?

Not a hard 1984 cap; more a Huxley drift. The Overton window shifts with volume, making some ideas feel normal and others unsayable.

Is Gen Z actually a distinct linguistic tribe?

Generational labels are marketing buckets that become roles to play. Each of us has a singular umwelt that resists neat sorting.

We want to belong yet still be unique.

Hold multiple frameworks, seek variety, and touch grass to escape the tunnel.

Hit me with some etymology gems.

Muscle from “little mouse,” salary from salt, assassin from hashish user, candidate from “white-robed,” girl once gender‑broader, silly drifted from blessed to foolish, avocado also meant testicle, penguin likely from Welsh via the great auk, and orange from “a‑norange.”

I saw talk about AI translating bird chatter; have you seen that?

I’m cautious about calling animal signals a human-like language. There’s real progress, from Cornell’s bird projects to whale work, but creatures mix sound with movement and sonar, so meaning is contextual and the very idea of language here is fuzzy.

Wild; it makes me think of QWERTY—do you know the backstory?

Typewriters jammed, so designers spaced frequent letters apart to avoid clashes. Dvorak existed, but QWERTY won because it jammed less.

Other layouts can be 30 to 50 percent faster, which made me wonder if we could design a language that’s brutally efficient yet still expressive.

Treating speech as pure data misses its social job; small talk and goodbyes are rituals that bond us. Languages also balance information flow: Japanese tends to use more syllables while Thai packs more into tones, yet both carry similar content over time.

So what would a maximally efficient tongue sound like?

Conlangs like Ithkuil cram dense meaning into tiny units, which is fascinating but basically unlearnable for humans; we talk to connect, not just compress. Machines could adopt tighter codes, but that’s a different goal.

I once saw a clip of two AIs switching to bleeps because it was faster—maybe fake, but still.

Models like ChatGPT don’t think in English; they slice text into tokens, turn them into high dimensional numbers, predict the next piece, then map it back to words. Meaning can warp during that round trip, which can overemphasize some terms or yield a style that feels misaligned with how people actually talk.

With megaphones getting louder, are we heading into the fastest shift in how people speak and write?

Probably, though it’s hard to measure when ‘language’ is a fuzzy label and everyone speaks a personal version. LLMs train on a smoothed out corpus and output a mathematical dialect that belongs to no one, which could speed a new, homogenized style.

Your work is great; where should people follow along?

I steer folks to longer form: my Substack, Etymology Nerd, and my book Algo Speak about how social media reshapes language; I’m also Etymology Nerd on the major platforms. Thanks for having me—this was a blast.

Quick closer: I’m often asked for reading recs, so I put together a free list of one hundred impactful books with notes and links; you can grab it on my site on the books page.