How the Unabomber Manifesto Exposed Ted Kaczynski

Somewhere between a crime documentary and an English class lies forensic linguistics, and Sam Denby from Half as Interesting shows just how creepy and fascinating it can get. Using the manifesto of Ted Kaczynski as a case study, Denby explains how language can betray a person without them ever realizing it. Every sentence leaves fingerprints.

The manifesto itself is packed with accidental clues about its author. By reading between the lines, experts could estimate Kaczynski’s age, education level, regional upbringing, and even narrow down aspects of his personality and identity. The frightening part is that he was trying to hide.

One of the biggest concepts Denby explores is the idea of an idiolect, essentially a linguistic fingerprint. Every person writes differently. Tiny habits most people never notice become identifiers. Certain phrases get repeated. Grammar quirks appear over and over. Word choices become patterns. Forensic linguists can compare those patterns to letters, essays, emails, or published work to determine if the same person wrote them. It has been used to verify confessions, analyze ransom notes, and identify anonymous authors hiding behind fake names.

What makes this especially difficult to fake is something called function word consistency. People can consciously change their vocabulary or tone, but they usually cannot control the invisible mechanics underneath their writing. Words like “the,” “and,” “of,” or “but” appear with surprisingly stable frequency from one piece of writing to another. Those tiny connective words become statistical evidence.

Denby also dives into the math behind forensic stylometry, which sounds like science fiction but is very real. One major technique is Burrow’s Delta, a system that compares the frequency of the most commonly used words across writing samples to measure stylistic similarity. It famously helped reveal that J. K. Rowling wrote The Cuckoo’s Calling under the pseudonym Robert Galbraith.

Another method is the Jaccard Coefficient, which measures overlap in unique word choices between texts. In other words, it looks at how similar two writers are based on vocabulary patterns. Combined with other linguistic markers, it becomes surprisingly effective at narrowing down authorship.

The unsettling takeaway is that nobody writes anonymously as well as they think they do. Every tweet, email, comment section rant, or manifesto leaves traces behind. Language is personal. Even when people try to disguise themselves, their words keep talking.