Corpus-Based Mechanistic Analysis of Cyberbilling Metaphor Framing in American Social Media Discourse

Where digital commerce intersects with linguistic expression in American social media discourse, a distinct phenomenon has taken shape—one marked by the widespread use of metaphor to conceptualize financial interactions; this thesis explores the hidden mechanisms behind these cyberbilling metaphor frames, using a corpus-based approach to unpack how language shapes user perceptions of abstract digital financial realities. Cyberbilling encompasses the full spectrum of online financial exchanges, stretching from familiar traditional electronic banking and routine e-commerce transactions to emerging decentralized economic activities within the expansive Web3 ecosystem. These metaphorical frames serve as core cognitive tools, not just superficial stylistic choices to dress up dry online financial discourse. By leaning on a corpus-based research method, we move past isolated anecdotes to deliver both quantitative and qualitative breakdowns of the specific linguistic patterns that dominate these online financial conversations.

This analysis rests on a core rule from Conceptual Metaphor Theory, which posits that human understanding of abstract domains is rooted entirely in connections to more familiar, concrete experiences; on American social media, the invisible, abstract mechanisms of digital finance are regularly framed using concepts tied to physical force, warfare, long journeys, or the flow of liquids and gases. We examine the “how” of this framing process, pinpointing the exact lexical choices and grammatical structures that trigger these specific metaphorical frames in everyday user posts across major social media platforms. Verbs tied to physical manipulation or nouns drawn from conflict zones signal a clear view of financial transactions as battles or struggles. This framing process follows consistent, set rules, rather than unfolding at random, and it mirrors a shared cultural understanding of value and risk prevalent in American digital society; we focus closely on the source domains that supply metaphorical imagery and the cyberbilling target domains, examining the specific mapping relationships that give structure to this widespread online discourse.

To carry out this study, we use a strict, multi-step process crafted specifically to ensure the reliability and validity of our linguistic data; first, we build a specialized corpus pulled from major American social media platforms, selected precisely because they host huge volumes of ongoing financial discourse and user interaction. We follow tight, pre-defined inclusion rules when gathering data, filtering out posts unrelated to cyberbilling activities while preserving the metadata that gives critical context to each user’s online interaction. After collecting all relevant data, we clean and tag the corpus to get it ready for deep linguistic analysis of metaphorical frames. Our main analytical method combines concordance analysis and semantic tagging to isolate metaphorical expressions related to cyberbilling; we identify key metaphorical keywords and examine their immediate collocational environments—the words that sit right next to them—to determine their specific semantic prosody and framing effects on readers. This step matters a great deal because it lets us clearly distinguish between straightforward, literal discussions of billing and metaphorical uses that carry extra social or emotional layers of meaning for everyday users.

The value of this research goes far beyond theoretical linguistics, touching on user experience design, fintech development, and digital literacy programs for the general public; knowing how users frame cyberbilling through metaphor gives fintech companies key insights to explain complex products to everyday people. If most users frame cryptocurrency interactions using war-related metaphors, this tells us they see these assets as high-risk and highly unstable, a perception marketing strategies need to address head-on to build trust. Educators and policymakers can also use these findings to spot key public misunderstandings about complex digital financial systems. By clarifying the specific linguistic structures that underpin these user perceptions, stakeholders can design more effective educational tools and regulatory frameworks that align closely with the public’s existing cognitive models of digital finance. This thesis makes clear that the language used to talk about cyberbilling isn’t just a passive tool to describe economic reality—it actively constructs how users engage with and experience the entire digital economy.

Building a specialized corpus to analyze the mechanics of cyberbullying metaphors forms a necessary base that links raw social media data to focused linguistic inquiry, and we use a systematic approach to build a robust dataset that begins with selecting American social media platforms—Twitter (now X), Instagram, TikTok—since Twitter hosts high-volume, text-heavy interactions ideal for syntactic study, Instagram, which blends visual context with written commentary, often carries language that mirrors social exclusion patterns, and TikTok’s short videos feature captions and comments that capture fast-changing youth language. We collect data over three years to track longitudinal shifts in metaphorical usage, a timeframe that keeps the corpus relevant while capturing enough linguistic variation to support meaningful analysis. This three-year timeline ensures the corpus stays relevant to current discourse while meeting the need for sufficient linguistic variation in collected data.

We use stratified random sampling instead of basic random collection to ensure statistical representativeness, a choice that cuts down on overcounting viral outliers or bot-generated content and keeps the natural spread of real user interactions intact. We start data retrieval with advanced search queries built from a seed list of cyberbullying-related keywords and hashtags, terms that range from direct insults to culture-specific slang tied to harassment. This targeted sampling method keeps the dataset aligned with natural, real-world patterns of user interaction on social media platforms. After gathering raw data, we run a strict filtration process to isolate relevant discourse entries, using both automated algorithmic filters and manual checks to sort content accurately, keeping only posts that carry clear hostility, aggression, or dominance toward a specific person or group while scrubbing entries that are merely satirical, playful banter, or news coverage of bullying that doesn’t include actual bullying discourse. This dual layer of checking ensures the data stays pure enough for later analysis of metaphor framing in cyberbullying.

After filtering, we compile basic statistics to define the corpus’s core parameters, finding the final dataset holds a substantial total word count and a set number of distinct posts, with content topics spread across appearance-based shaming, political harassment, and other forms of targeted abuse, and user data sorted into individual aggressors, bystanders, and defensive responses. We also follow strict ethical research standards by anonymizing all collected data, scrubbing or replacing usernames, handles, and exact timestamps to protect the privacy of both victims and those who carried out the bullying. This data anonymization step aligns with strict current institutional review board requirements for ethical social media research projects.

To get ready for metaphor analysis, we run the corpus through a series of preprocessing and annotation steps: we normalize text by converting all characters to lowercase, stripping out URLs, and expanding common abbreviations to standard forms, then split the text into individual units of meaning through tokenization. Our annotation system is tailored to spot metaphorical expressions tied to cyberbullying, with annotators using a coding framework to tag source domains like war, disease, or physical assault and link them to bullying-related target concepts. This coding framework lets us quantify exactly how different types of metaphors structure cyberbullying discourse in the corpus. By carefully building and annotating this specialized corpus, we create a reliable empirical base for future work, making sure our upcoming analysis of metaphor mechanics uses authentic, representative, ethically sourced linguistic data that boosts the validity and real-world applicability of our findings on cyberbullying dynamics.

We base our approach to identifying and categorizing metaphorical frames in cyberbullying discourse on the combined insights of Conceptual Metaphor Theory and Framing Theory, where the first argues that people grasp abstract, intangible ideas by linking them to tangible physical settings, turning complex social harm like cyberbullying into something interpretable through everyday bodily experiences that shape their core understanding of the world. The second theory adds to this by showing how these specific mental links shape how people organize and make sense of their experiences, drawing focus to certain parts of online harm while sidelining other critical details. For cyberbullying specifically, we define a metaphorical frame as a coherent structure that links online harm to a specific tangible source domain. This definition lets us move beyond single, isolated metaphorical phrases to uncover the hidden cognitive patterns that guide how social media users see and judge aggressive, harmful online acts; it ensures our work targets shared cognitive tendencies rather than just random linguistic choices.

To spot these metaphorical frames, we use a strict mix of manual and computer-assisted methods, starting with building a specialized collection of American social media posts that serves as our core empirical data, then using tools like AntConc and Wmatrix to run quantitative keyword checks and semantic tags that help pull out potential metaphorical phrases and highlight words that appear more often than others. But computer tools alone can’t do the job, since words often have multiple meanings that shift entirely based on the specific context of the social media post they appear in. We need to manually check each candidate phrase to confirm it uses a cross-domain metaphor rather than literal meaning. To make sure our identification process is reliable, we have two separate analysts code a random subset of social media posts from our corpus, and we bring in a third independent expert to work out any disagreements through detailed discussion, which helps us build a dataset that’s consistent and valid for all subsequent classification work.

After we’ve identified all potential metaphorical frames from our curated American social media corpus, we group them systematically based on the tangible source domains they draw from, which reveals the distinct mental lenses that social media users rely on to make sense of and talk about online harm and victimization. We then count how often each frame appears in the corpus, calculating its share of all metaphorical uses to figure out which conceptualizations are most common in public talk. Most of our statistical data shows frames tied to physical violence or disease pop up more often than others. For each widely used frame, we sum up its core conceptual meaning and the typical words people use to express it in our curated social media corpus; a physical violence frame, for example, casts cyberbullying as a deliberate, harmful attack that needs self-defense, using words tied to hitting, fighting, or striking others online. A natural disaster frame, by contrast, casts cyberbullying as an unstoppable, overwhelming force that people can’t control, using words like floods or storms to describe its rapid spread online. These patterns show how each frame shapes public views and responses to cyberbullying, highlighting the quiet power of cognitive framing in online discourse about harm.

To study the inner workings of metaphorical frames in cyberbullying, we begin with a core definition that sees metaphor not as just a decorative language tool, but as a cognitive device that structures people’s sense of reality, and we apply this to American social media by closely examining how linguistic metaphors taken from a focused corpus actively shape the common view of online aggressive behavior, drawing on a theory that links abstract concepts like cyberbullying to tangible, everyday domains to explain user reasoning and reactions. We put this theoretical framework into practice through rigorous checks of corpus word pairings and contextual word use, moving past surface-level observations to uncover the hidden processes that drive online interaction and conflict growth. This deep dive lets us grasp the real, unobserved mechanisms behind how these frames operate.

One key way these metaphorical frames act is by shaping how people see social roles and who gets blamed for harm, as linking cyberbullying to ideas like physical fights, hunting, or disease creates clear, often one-sided positions for those who carry out harm and those who suffer it—such as when someone calls a bully a “predator” to highlight their intentional harm, while labeling the victim “prey” to emphasize their lack of control. This way of using language directly affects how onlookers judge who’s at fault, as framing conflicts as fights can lead people to unfairly question a victim’s ability to defend themselves, shifting blame away from the bully. These frames also change how social media users feel about the events. Frames that use illness or infection imagery often make social media users feel disgust toward victims, pushing them to exclude those who have been harmed, while frames tied to aggression or war can normalize harmful acts, lead some onlookers to openly back the person carrying out the bullying, and even stop others from stepping in to offer support to the victim. This ability to shift how people feel directly affects whether onlookers step in to help, stay completely passive, or even join in the harassment directed at the victim.

Beyond shaping individual roles and emotions, these metaphorical frames drive how cyberbullying talk spreads online and how public opinion forms, as metaphors that spark strong reactions act like fast-spreading signals that push hostile content to reach more people quickly, especially when they use ideas of immediate threat or spreading evil to make users feel urgent about sharing and commenting. This spread isn’t a straight line, as repeated use of the same frames in comments and shares makes a group’s view of the event stronger and more fixed over time. As incidents spread further, metaphors shift from describing events to judging them, losing nuance as they grow more intense. This shift makes it harder for people to find common ground or work toward resolving the conflict, and data from targeted corpus studies shows this is exactly how small, localized online disputes can blow up into large-scale social media scandals that reach thousands of users across different platforms and spark widespread debate. Each repeated use of a single frame strengthens the group’s shared view of the event, pushing the conflict to become more extreme and harder to calm down or resolve through dialogue.

These frames also act differently at each stage of a cyberbullying incident, as during the initial push to start harm, they set rules for what’s seen as acceptable and justify the first attack, while as the incident spreads, they keep the momentum going and get more people involved by dehumanizing the target or framing harm as a rightful act. When the incident starts to wind down, certain frames can keep stigma alive and block healing, while others can help people see the event as a lesson to avoid future harm. All these processes tie closely to American social and cultural values. Ideas like individualism, competition, and a focus on free speech shape which frames resonate most with American users, as those that fit with cultural stories of standing up for oneself or seeking fairness are more likely to catch on and change how people think about harm or even influence views on related policies. Understanding these links helps build better tools to spot harmful frames online and create educational lessons to teach people to question manipulative language.

When we analyze cyberbullying metaphor framing in American social media discourse through a corpus-based, mechanistic lens, we find metaphor acts not as a simple linguistic flourish but as a central cognitive and communicative tool that shapes how people perceive online harm, react emotionally, and form behavioral intentions toward it. Framing through metaphor, at its root, involves mapping cyberbullying’s abstract, intangible experiences onto concrete, familiar concepts—like warfare, disease, or criminal victimization—to make the phenomenon understandable, judgeable, and responsive, a process rooted in conceptual metaphor theory which holds that human thought is naturally metaphorical, with abstract ideas structured by embodied, sensory experiences; on social media, this means users draw on existing mental frameworks to make sense of a new, technology-mediated harm lacking traditional bullying’s physical signs. This core principle ties every metaphor use directly to how people process and act on online harm experiences.

Our corpus analysis clearly outlines the framing mechanism’s operational path as three closely linked stages that don’t follow strict, separate, ordered steps. Users choose source domains aligned with their emotional take on cyberbullying, for example calling it a “war” to emphasize aggression and conflict, or a “disease” to highlight spread and vulnerability; these domains then activate connected mental scripts, leading others to apply the concrete domain’s reasoning to online harm, like framing it as a “crime” which brings up ideas of perpetrator blame, legal fixes, and victim innocence that shift how audiences assess responsibility and back intervention. Repeated use of these dominant metaphors across the corpus normalizes specific, narrow views of what cyberbullying is and how it unfolds. Over time, these normalized views harden into shared discourse rules that shape institutional actions, such as social media platform policies or school educational campaigns, and individual choices, including whether bystanders step in or victims reach out for help.

This analysis holds practical value because it uncovers the hidden mental structures that shape how people understand cyberbullying, giving evidence-based insights to groups across different professional fields. For social media platforms, spotting common metaphor frames can help design more intuitive content moderation systems and user support tools that match how people actually think about and report harm, instead of relying on vague, technical definitions of bullying; for educators and mental health workers, recognizing how specific metaphors resonate emotionally and cognitively can help build targeted interventions that validate victims’ experiences, like using disease-related language to address the quiet, growing spread of harmful content, or criminal justice terms to empower victims to seek formal help. Policymakers can use these findings to craft regulatory language that reflects how people actually experience and talk about cyberbullying. This mechanistic, corpus-based look at metaphors isn’t just an academic exercise—it’s a practical tool to bridge gaps between linguistic discourse, cognitive understanding, and evidence-based action to reduce online harm.