Lies, Likes & Linguistics

You’ve probably seen it. A social media profile that seems a little too perfect. An online review that feels strangely aggressive or suspiciously glowing. A text message from a “wrong number” that quickly turns into a flirtatious, high-stakes conversation. In a world saturated with digital communication, our instincts often tell us when something is off. But what if we could prove it? What if the words themselves held the key?

Welcome to the digital frontier of forensic linguistics, a field where the humble text message is treated like a crime scene and a Twitter feed is analyzed with the same rigor as a ransom note. Long gone are the days when linguistic analysis was confined to disputed wills and taped confessions. Today, experts are sifting through the sprawling, chaotic data of our online lives to answer one crucial question: Who really wrote this?

Beyond the Ransom Note: The New Digital Crime Scene

Traditional forensic linguistics built its reputation on high-profile, tangible cases. Think of the linguistic analysis that helped identify the Unabomber, Ted Kaczynski, by comparing his 35,000-word manifesto to his previous writings. The principles were sound, but the data was often limited to a single, carefully crafted document.

The internet changed everything. The modern “document” isn’t a single letter; it’s a constellation of data points scattered across platforms:

• Text messages and WhatsApp chats
• Social media posts and comments (Facebook, Instagram, X/Twitter, Reddit)
• Dating app profiles and conversations
• Emails and forum posts
• Online product and service reviews

This digital deluge presents both a challenge and an incredible opportunity. The language is informal, riddled with slang, emojis, and abbreviations. Yet, the sheer volume of text produced by a single individual means more evidence, more patterns, and more chances to uncover the truth.

Your Idiolect: The Linguistic Fingerprint You Can’t Wipe Clean

The core concept behind this work is the idiolect. Your idiolect is your unique, individual linguistic profile—a combination of vocabulary, grammar, spelling, and punctuation that is as distinctive as a fingerprint. You might not be aware of it, but you have one. It’s built from your education, where you grew up, your social circles, your age, and even your profession.

Forensic linguists are trained to spot the subtle components of an idiolect, including:

• Lexical Choices: Do you write “okay,” “ok,” or “kk”? Do you refer to a soft drink as “soda,” “pop,” or a “Coke”? Do you favour certain adjectives like “amazing” or “brilliant”?
• Punctuation Habits: Are you a serial user of the ellipsis (…)? Do you put two spaces after a period (a tell-tale sign of being taught to type on a typewriter)? Do you use the Oxford comma? What are your go-to emojis, and in what combinations (e.g., the 😂 followed by the 😭)?
• Spelling and Typos: Everyone makes typos, but we often make the same typos consistently. Do you always mix up “their” and “they’re”? Do you spell “definitely” as “definately” or “seperate” instead of “separate”?
• Grammatical Patterns: Do you often start sentences with conjunctions like “And” or “But”? Do you use specific phrasings, like saying “I’m needing to” instead of “I need to”?

On their own, these are just quirks. But when collected and analyzed, they form a powerful, identifiable pattern that is incredibly difficult to fake consistently.

Case Study: Unmasking the Catfish

Let’s imagine a classic catfishing scenario. A 60-year-old widower, David, connects on a dating app with “Sophia,” who claims to be a 28-year-old fashion designer from Milan, Italy, temporarily living in his city. Her photos are stunning, her messages are charming, but soon, she needs money for a “family emergency” back home.

David’s suspicious daughter hires a forensic linguist. The expert doesn’t need to see “Sophia.” They just need her words. Here’s what they might find:

1. Inconsistent Language Background: “Sophia” claims Italian is her first language. However, her English mistakes aren’t typical for an Italian speaker. For instance, she mixes up “he” and “she”, a common error for speakers of languages that don’t have gendered pronouns, like many West African languages—but not Italian.
2. Anachronistic Slang: For a supposed 28-year-old, her slang is dated. She uses phrases like “that’s the bomb!” or “talk to the hand”, which were popular in the 1990s, not among today’s twenty-somethings.
3. Punctuation as a Generational Marker: She consistently uses two spaces after every period, a habit strongly correlated with people who learned to type before the year 2000.
4. Revealing Idioms: She uses an idiom like, “We will cross that bridge when we get to it”, but translates it awkwardly from another language, resulting in something like, “We will pass that river when it arrives.” This points to a non-English, non-Italian native tongue.

The verdict? The linguistic evidence overwhelmingly suggests the author is not a 28-year-old Italian woman, but likely an older individual from a completely different linguistic region. The idiolect doesn’t match the persona.

The Troll Patrol and the Science of Stylometry

The same methods used to unmask catfish are also deployed to identify anonymous trolls, harassers, and criminals. In these cases, linguists often turn to a powerful statistical method called stylometry.

Stylometry is the quantitative analysis of writing style. Instead of just looking at qualitative tells, it uses software to measure and compare texts. An investigator will build a “corpus” (a body of text) from the anonymous harasser’s posts. They then compare it to a corpus of writing from a suspect.

The analysis focuses on features that are hard for a person to consciously control, such as:

• The average length of words and sentences.
• The frequency of function words (like the, of, a, to, in, and). We use these thousands of times a day without thinking, and our individual usage rates are surprisingly stable.
• Vocabulary richness (the ratio of unique words to total words).

One of the most famous real-world examples of stylometry was when it was used to confirm that J.K. Rowling was the true author of “The Cuckoo’s Calling”, written under the pseudonym Robert Galbraith. The statistical signature in the novel was a near-perfect match for her other work.

Lies, Likes, and the Limits of Linguistics

Of course, this science isn’t foolproof. People are complex. We engage in style-shifting—we don’t write the same way on LinkedIn as we do on Reddit. A very short sample of text, like a single tweet, might not provide enough data for a confident conclusion.

Furthermore, the rise of AI language models presents a new challenge. Could a scammer use ChatGPT to generate text that mimics a specific demographic or scrubs their own idiolect clean? It’s an ongoing cat-and-mouse game, and linguists are constantly developing new methods to stay ahead.

But for now, our words leave a trace. Every time you post, comment, or send a message, you are contributing to your own digital linguistic fingerprint. It’s a subtle, unconscious trail that tells the story of who you are, where you’re from, and sometimes, what you’re trying to hide. In the world of lies and likes, linguistics is often the last, best hope for the truth.