About Authorship Analysis

 

Authorship analysis is one of those areas in forensic linguistics that sounds deceptively simple: “just figure out who wrote the thing.” Easy, right? After I explain it to you, you'll say that you could do it. But language isn’t handwriting, and it’s not DNA. It’s slippery. We change how we speak and write depending on who we’re talking to, what we’re talking about, or whether we’re hungry and cranky when we hit “send.”

So how does it work?

At its core, authorship analysis is about identifying patterns of linguistic behavior. Sometimes call idiolect, the idea that everyone has a unique language fingerprint. Think of it like Gollum and Sméagol: same body, same vocal cords, but different word choices, different rhythms, even different quirks in grammar. Those subtle differences give them away.

When a suspicious ransom note, threatening email, or questionable confession shows up, analysts don’t look for single “tells” (“this person always uses semicolons”). Instead, they build a profile out of many small, often unconscious features:

• Vocabulary (do you say soda or pop?)

• Spelling & punctuation quirks (Oxford comma forever—or never?)

• Syntactic habits (short, clipped sentences vs. sprawling ones)

• Functional words (tiny glue-words like and, the, but that are nearly impossible to disguise)

A corpus lets us compare these features systematically. If someone claims, “I didn’t write that creepy message,” we can measure how their writing lines up with the suspect text. It’s never about 100% certainty (that's not how science works), but it can help courts, investigators, or historians weigh the evidence.

And here’s where it gets a little… creepy. Language leaves traces. Even if you delete a tweet or scrub your email drafts, the way you write carries fingerprints you didn’t mean to leave. You can change your name, your IP address, even your entire online persona, but your syntax has longevity, and tendency of following you around like a shadow.

Authorship analysis lives at the crossroads of linguistics and human behavior. It’s not just about what was written, but how—and sometimes, the how tells us more than the author ever intended.