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

Social media platforms are transforming how language evolves and circulates, driven by both user behavior and the constraining logic of platform algorithms. TikTok, in particular, has become the preeminent engine of modern slang and cross-generational linguistic innovation, while algorithmically mediated attention and economic incentives are reconfiguring not only what people say but what they feel can be said online.

Tiktok Dominates Modern Linguistic Innovation and Cross-Generational Word Adoption

Adam Aleksic affirms that TikTok now surpasses Reddit, 4chan, and Twitter as the main source of new slang—a shift documented in a Know Your Meme 2022 study tracking the origins of trending terms. Initially, most linguistic innovation started on forums like 4chan and Reddit or on Twitter, but now TikTok and Twitter dominate, with TikTok rapidly accelerating the lifecycle and churn of modern slang.

Tiktok Surpasses Reddit, 4chan, Twitter as Top Source of New Slang, Accelerated by Algorithmic Trends

On TikTok, the user interface encourages rapid-fire, participatory conversation, fostering a sense of collective innovation. However, Aleksic stresses that this innovation is now shaped less by organic linguistic creativity and more by algorithmic trends and echo chambers. Modern slang cycles turn over faster than ever because users respond to the platform’s incentive structures: what the algorithm boosts, spreads.

Ui Encourages Linguistic Play, While Algorithms Create Echo Chambers

TikTok’s visibility and virality mechanics deeply influence which words and phrases spread. Aleksic notes examples like the "6-7" meme, which spread virally with users, politicians, and even brands deploying it for maximum reach. This is a conscious realization among users—clip farming and strategic keyword use are now the key to distribution. For many, distribution and algorithmic appeal have become more significant than the inherent meaning or quality of the language itself.

Tiktok's Visibility and Virality Mechanics Incentivize Clip Farming and Strategic Keyword Use, Making Distribution and Algorithmic Appeal More Important Than the Language's Meaning or Quality

Aleksic explains that certain words ("maxing," "gooning," "Clavicular") are strategically used to trigger the algorithm, thus facilitating greater viral reach. All words are now keywords, functioning as search engine optimization tools within social platforms. The content creators purposefully design their language to push videos or posts to wider audiences, competing in an ecosystem where attention and data have been monetized and commodified. This results in a system where clip farming—generating short, algorithmically optimized moments primed for viral sharing—takes precedence, and where creators replicate attention-grabbing structures rather than focusing solely on substance.

Social Media Platforms Are Environments Where Users Adopt Different Dialects and Styles Depending On the Platform

The linguistic norms of social platforms are shaped by their unique audiences, interface constraints, and social cultures, creating distinct dialects and subcultural styles.

Social Media Platforms Develop Unique Norms and Cultures

Aleksic likens different platforms to different "houses," each hosting its own commenting culture and linguistic register. LinkedIn users employ formal, professional language; Twitter is characterized by playful and memetic slang; TikTok evolves its own fandom and influencer vernacular. Users instinctively adapt their styles depending on the platform, behaving like sociologist Irving Goffman's actors donning multiple faces or registers, much like how people switch language between speaking with parents versus friends.

Micro-Dialects In K-Pop and Swiftie Communities Reveal Distinct Linguistic Subcultures

Even within a single platform, micro-dialects can emerge, such as within K-pop fandoms or Swiftie (Taylor Swift fan) subgroups on TikTok. Each micro-community innovates language unique to their culture, further stratifying the digital linguistic landscape.

Speech Expectations on Platforms Reinforced by Founder Effect

Founders and early adopters imprint specific communication styles on new platforms—a "founder effect"—that guide future linguistic norms. The "influencer accent" on TikTok, for example, is now standard because new users copy the linguistic style of early, successful creators. Meme propagation and style choices, like popular thumbnail formats or newscaster voices, are similarly set by early standout figures and become reinforced by community consensus.

Algorithms Shape Word Spread, Compressing Language Into Recognizable Forms and Specialized Variations

Social media algorithms do not merely boost v ...

Language serves not just as a conduit for transmitting information but as a powerful marker of identity and community affiliation. Adam Aleksic and Chris Williamson explore how both the words we use and the ways we use them signal which groups we belong to, reflecting histories, subcultures, and pressures around conformity and self-expression.

Language as a Marker of Group Identity vs. a Neutral Information Tool

Cohort Membership Indicated by Language Use

Chris Williamson notes that specific language use instantly signals something about the person—a clear marker of in-group membership. Adam Aleksic argues that language is 100% a tool of identity, and every time someone chooses a word or phrase, they're also revealing their cohort or subcultural ties. Just as certain clothing styles flag membership in particular communities, so do speech styles, slang, and turns of phrase.

Unique Idiolects Shaped by History, Pressured by Conformity

Aleksic points out that every individual has a unique idiolect, a personal dialect influenced by upbringing, experience, and cultural exposure. This individual linguistic identity is reflected in unique expressions and quirks, such as someone saying "eat your cake and have it too"—a reversal of the common phrase that, in the case of the Unabomber, helped police pinpoint his identity. These idiolects demonstrate that we all see and express our worlds uniquely. However, social pressures—such as generational identity or online personas—also push people towards conforming with their group's collective language practices, in effect making them interchangeable with others who share those labels.

Slang Adoption Driven by Youth Differentiating From Parents, Ages 10-25

Language change is primarily driven by young people. Teenage years, especially from about 10 to 25, are a hotbed for slang invention and adoption, motivated in part by a desire to create distinct generational identities separate from parents or older generations. The process of adolescence, marked by growing need for independence and peer affiliation, propels this constant renewal of linguistic style.

Linguistic Innovation Pipeline: Marginalized to Mainstream

Origin of Gay Slang: 1980s NYC Black/Latino Ballroom Culture, Adopted by Straight Women, now Gen Z Mainstream

Aleksic highlights the trajectory of linguistic innovation, where marginalized communities generate novel language that later enters mainstream use. Popular Gen Z slang like “slay,” “serve,” “queen,” and “ate” originates in 1980s NYC Black and Latino gay ballroom scenes. In these spaces, language was a means of boosting members’ status and agency in a society that denied them both. Over time, terms from gay and Black subcultures trickled into broader usage—first among straight friends of gay men, then throughout general youth and internet culture.

Slang From African American English Enters Mainstream Without Credit or Benefit

Many new slang words in mainstream and internet English stem from African American English. Words and phrases that originated as tools for community and subversive expression are widely adopted—sometimes with detached or ironic undertones—while the originating communities often see little recognition or benefit from their linguistic creativity.

Manosphere and 4chan Shape Language: Maxing, Pilled, Gooning Influence Mainstream Culture

The manosphere and forums like 4chan also act as laboratories for slang, producing terms like "maxing," "pilled," and "gooning." These terms develop as in-group signals on anonymous online platforms and gradually filter into wider cultural consciousness through memes, jokes, or social media.

Anonymity and the Demand for Cultural Competence Online Pressure Linguistic Innovation and ...

Creators in the attention economy refine their speech and vocal performance using distinct linguistic and paralinguistic tactics to maximize engagement and hold audience attention, with each archetype shaping speech patterns for both platform optimization and identity signaling.

Creator Archetypes Use Distinct Vocal and Linguistic Strategies For Their Goals in the Attention Economy, Each Producing Predictable Effects on Engagement

Influencers Use Vocal Techniques to Signal Relatability and Mimic Familiarity, Maintaining Attention Through Floor-Holding to Prevent Dead Air

Adam Aleksic observes that creators often begin videos mid-sentence—starting with phrases like "no because"—to instantly draw viewers in by making them feel as though they are eavesdropping on an already-unfolding conversation. This tactic, described as in medias res, fast-tracks the sense of participation. The "hey guys, welcome to my podcast" influencer accent, originating from figures like Kim Kardashian, Paris Hilton, and early beauty YouTubers, persists on TikTok, preserving features like uptalk, vocal fry, and lengthened final vowels. This accent, according to Chris Williamson, creates a softness and welcoming familiarity, fostering a parasocial sense of inclusion and relatability.

Dead air is algorithmically penalized; lifestyle influencers drag out final syllables, use uptalk, and employ "floor-holding"—the strategic use of filler words like "um"—to maintain a continual conversational flow and signal that more is coming. This prevents listeners from disengaging while the creator transitions between thoughts. Aleksic adds that these vocal strategies build the influencer's identity and optimize for platform dynamics, creating a bond with the viewer that feels authentic and cozy.

Influencers Use Fast Pacing, Emphasize Key Words, and an Authoritative Tone to Appear As Experts, Sacrificing Lifestyle Content's Coziness For Confidence and Authority

For educational influencers, the goal shifts toward perceived expertise and trust. Aleksic and Williamson note that these creators deliver content at a faster, sharper pace, emphasizing key words with greater frequency and adopting a more authoritative tone. This style sacrifices the casual coziness of lifestyle content for the clarity and confidence needed to establish the creator as an authority. Educational influencers aim for precise diction with clear, accurate consonants, constructing a brisk, staccato rhythm that grabs attention and enhances information retention. The emphasis is on structure and certainty: the goal is not just to relate, but to inform with confidence.

Mr. Beast and Peers Use Loud Delivery, Exclamations, and Emotional Escalation to Keep Viewers Engaged and Prevent Them From Scrolling Away

The archetype exemplified by Mr. Beast embraces maximal intensity to capture and hold the short attention spans of a youthful audience. Aleksic describes how Mr. Beast pivots his vocal delivery—often screaming or emphasizing extreme statements like "I just bought a private island"—and continually escalates emotional stakes. Chris Williamson explains that this creates a promise of ongoing excitement and imminent payoff, which is crucial for preventing viewers from scrolling away before a video's conclusion.

Mechanics of Optimized Speech: Enhancing Engagement Through Linguistic Manipulation

Uptalk and Elongated Vowels Signal Continuation, Keeping Viewers Engaged, While Downward Intonation Signals Finality and Should Be Avoided In Online Content

Up talk, characterized by lengthening the last vowel and pitching the voice upward at the end of a statement, signals to listeners that a thought is unfinished, encouraging continued attention. Aleksic emphasizes that elongating the vowel effectively signals continuation—a cue that the speaker isn’t finished, enticing viewers to remain engaged. Conversely, downward intonation ("down talk") signals closure and prompts the audience to tune out, so experienced creators deliberately avoid it.

Consonant Clarity Creates Structure in Educational Content, While Blended Consonants in Lifestyle Content Encourage Bonding

Chris Williamson shares that vowels add color and emotional nuance, while consonants provide speech structure and clarity. For educational content, creators articulate consonants carefully, ensuring clear transmission of ideas. In lifestyle and casual content, blending or dropping consonants, such as glottal stops found in Northeast UK dialects, make speech feel softer and less structured, which encourages a feeling of authenticity and intimacy. Aleksic observes that creators select pronunciation styles purposefully—clear and precise for instruction, blended and relaxed for social bonding.

Strategic Silence & Filler Words Help Maintain Conversational Flow, but Younger Creators See Them As Low Status and Eliminate Them

Filler words and strategic silences—traditionally employed as floor-holding tools—help creators maintain conversational flow, avoid algorithmically penalized dead air, and give themselves time to formulate responses. However, as Williamson observes, elite creators and younger influencers increasingly train themselves to eliminate such fillers, preferring a more polished, status-signaling presentation style.

Balance Linguistic Variety For Authenticity and Engagement; Avoid Scripted Language That Triggers Skepticism

...

The rise of generative AI, especially language models like ChatGPT, is rapidly reshaping how people use language in daily conversation, research, and public communication. This subtle but pervasive influence leads to the homogenization of vocabulary and communication patterns in ways most people do not notice.

Chatgpt Bias in Word Choice Influences Communication Patterns

"Delve" Usage Surged 1000% Post-Chatgpt Due to Latin-Over-Germanic Bias From Prestige Vocabulary

Recent studies show a dramatic increase in the use of the word “delve”—spiking by 1,000% since the debut of ChatGPT. Adam Aleksic attributes this phenomenon to biases in the reinforcement learning process used to train ChatGPT. While reinforcement workers in places like Nigeria and Kenya may contribute to the frequency (as “delve” is somewhat more common there), the main factor is a deeper, systematic Latin-derived word preference over Germanic vocabulary. Latin-origin words like "delve" are seen as more prestigious and confident, projecting knowledge and incisiveness, whereas Germanic synonyms like "dig in" appear more basic. This bias for “fancier”-sounding, Romance-language vocabulary is reinforced both by initial training data and by human feedback, which signals approval for such authoritative-sounding choices.

Chatgpt Outputs Boost Usage of Latin-Derived Words Like "Commendable," "Meticulous," "Crucial," and "Significant," Suggesting Unconscious Linguistic Mimicry

“Delve” is only the most visible example; there is a notable rise in terms like “commendable,” “meticulous,” “crucial,” and “significant”—all Latin-derived. This shift is not inherently negative, but Aleksic underscores how linguistic choices by these models are now visibly shaping real-world speech and writing. As humans read AI-generated responses filled with certain vocabulary, they begin to unconsciously mimic these patterns. Thus, the AI not only reflects user language but actively programs speakers and writers into new norms they may not be aware of adopting.

Studies Show Humans Adopt Ai-generated Vocabulary Patterns Subconsciously

Evidence now demonstrates that humans are picking up these tendencies in spontaneous spoken conversation. Social channels, professional networks like LinkedIn, and ordinary speech all begin to echo the AI’s signature phrases, even down to preferred sentence structures like em dash segmentations and “negative parallelism.” Aleksic warns that, increasingly, the trainers become the trained: “The creature that programmed the AI is being programmed by the AI.”

Language Model Architecture: Bottlenecks and Distortions From Optimization and Reinforcement Reward Structures

Chatgpt Translates Language Into Tokens and Neural Network Embeddings, Distorting Meaning For Speech Representation Rather Than Authentic Human Speech Recreation

Modern language models such as ChatGPT do not actually “speak” English. Instead, they convert input language into tokens—segments of words—which are mapped to high-dimensional coordinates, forming what’s called an embedding. These are processed through complex neural networks and predictive algorithms to generate output tokens, which are finally transformed back into readable text. At each of these steps, distortions and loss of nuance can occur. The frequent repetition of certain words or structures, such as “delve,” arises as meaning is reshaped through layers of optimization that prioritize coherence, authority, or other rewarded features.

Reinforcement Learning Biases Regions and Preferences, Causing Systematic Distortions

The system of reinforcement learning—shaped by the click-choices or rewards of remote workers—injects its own regional, cultural, or personal linguistic preferences. Because reinforcement workers might not catch subtle overuse of certain words, these terms are baked into the language model and disproportionately reproduced. Furthermore, all AI platforms optimize aggressively for improvement and speed, creating “bottlenecks” in what kind of language is amplified and repeated.

Lingustic Biases in Language Models: Overt and Covert Political Tweaking

Beyond inadvertent linguistic bias, overt political or ideological tweaks are present. Aleksic notes that prominent figures such as Elon Musk have openly modified their own AI models (“Grok”) to reflect personal or political leanings. Even companies that appear safety-conscious, like Anthropic, ultimately operate under optimization imperatives that can skew results in covert wa ...

Language is not static; it evolves alongside society, technology, and the lived experiences of its speakers. By tracing the history of certain words and examining shifts in grammar, usage, and vocabulary, we see how language both shapes and is shaped by culture.

Etymology Reveals Words' Cultural and Historical Shifts With Social Values, Technology, and Speakers' Experiences

The origins and metamorphoses of English words illustrate how meaning is tied to emotion, context, and social shifts. Chris Williamson and Adam Aleksic discuss the word "silly," which in Old English, "seilig," meant "blessed" or "fortunate." Over centuries, it shifted to convey "innocent," then "naive," and finally "foolish." This transition shows how the emotional valence of a word can shape its meaning, just as "awful" and "awesome" both derive from "awe," yet diverged profoundly in connotation. Words describing intense feelings, such as "terrible" and "terrific," can oscillate across positive and negative meanings due to the complex nature of human emotions.

Historical contexts often linger in word origins. "Candidate" descends from Latin "candidus," referring to someone in a white robe. In ancient Rome, office seekers wore white to symbolize purity, and although the practice has ended, the connotation remains embedded in the term.

Semantic Divergence of "Awful" and "Awesome"

"Awful" and "awesome" both share the root "awe," but traversed in different directions semantically, one toward negativity and the other toward positivity. Aleksic notes that the emotional quality inherent in a word allows it to shift easily between meanings, further exemplifying language’s responsiveness to the cultural landscape.

Language Evolves: New Words, Shortenings, and Abandonments Through Tech and Social Change

Changes in technology, society, and generational perspectives fuel constant linguistic innovation. Aleksic identifies two primary sources: institutional terms and slang. Words like "iPhone" and "podcast" originate from corporate or mainstream sources, while video game slang—"NPC," "skill issue"—and cultural slang from groups like Black communities or incels are typically driven by youth and grassroots adoption.

Semantic Contraction: "God Be With You" to "Goodbye" To "bye"

Words often contract and simplify over time. "God be with you" contracted to "goodbye," and then further to "bye," demonstrating how speech evolves for efficiency and changing social interaction.

Vocabulary Loss Reflects Conceptual Changes Over Time

As lived realities shift, certain vocabularies fade. Aleksic observes that modern English speakers use fewer terms for plants compared to the 1800s, reflecting a dwindling connection with the natural world. Concepts and names are lost as they lose relevance.

Different Languages Encode Reality Uniquely

Language shapes not only communication but also perception and cognition. Aleksic references Potawatomi, where "Saturday" can be expressed as a verb: "to Saturday," an idea unfamiliar in English yet natural in other linguistic systems. Such expressions reflect different cultural perspectives and priorities.

Agglutinative languages like Turkish and German construct complex concepts by tacking morphemes together, generating single words for nuanced ideas—such as German’s highly specific compound nouns. Inflectional languages like Latin or French modify word forms instead. In contrast, English often relies on phrases or multiple words, affecting the nuance and speed with which certain ideas can be communicated.

Language Loss: A Mass Extinction of Human Expression and Cognitive Diversity

The world is facing a mass extinction of languages; Aleksic cites that a language disappears every two weeks. Of the 7,000 known languages, most are forecast to vanish by the century's end. This extinction event, accelerated by social homogenization via platforms like social media, means the loss o ...