The Forger’s Bane: What is Stylometry?

What if your writing style was as unique as your fingerprint? Every time you craft an email, post on social media, or write a report, you leave behind a trail of invisible clues. It’s not about the content—the ideas you express—but the framework you build around them: the length of your sentences, your choice of punctuation, the little “filler” words you lean on without a second thought. This collection of unconscious habits forms your unique linguistic fingerprint, and the science of analyzing it is called stylometry.

At its core, stylometry is the statistical analysis of literary style. It transforms the art of writing into quantifiable data, allowing linguists, historians, and even forensic investigators to answer one tantalizing question with surprising accuracy: Who wrote this?

The Building Blocks of a Linguistic Fingerprint

Stylometry isn’t magic; it’s meticulous measurement. It operates on the principle that while we can consciously choose our words and topics, the underlying structure of our language is deeply ingrained and remarkably consistent. Computers are perfectly suited for this work, sifting through thousands of words to spot patterns a human reader would never notice. But what are they looking for?

The analysis focuses on a variety of features, often grouped into several categories:

• Lexical Features (Word Choice): This is the most famous aspect of stylometry. Analysts look at vocabulary richness (how many unique words an author uses) and, most importantly, the frequency of function words. These are the small, grammatical words we use automatically, like “the”, “a”, “of”, “in”, “on”, “with”, and “but.” While two authors might both write about dragons, one might subconsciously favor “on” while the other prefers “upon.” These tiny preferences, when measured across a large text, create a powerful authorial signature.
• Syntactic Features (Sentence Structure): How does an author build their sentences? Stylometric analysis measures things like average sentence length, the frequency of different clause types, and where phrases are placed within a sentence. One writer might favor short, punchy statements, while another constructs long, complex sentences woven together with semicolons.
• Character-Level Features: This gets even more granular. The analysis can break a text down into sequences of characters (known as n-grams) to find patterns. For example, does an author frequently use the combination “ing ” or have a habit of starting sentences with “However”,? Even punctuation habits—the use of commas, em-dashes, or exclamation points—contribute to the overall profile.

Think of it like this: a forger can painstakingly replicate the content and even the general “feel” of another person’s writing. But it’s incredibly difficult to fake hundreds of these tiny, unconscious stylistic habits at once. Sooner or later, their own fingerprint begins to show through.

Stylometry in Action: Unmasking Authors and Solving Mysteries

While the theory is fascinating, stylometry’s true power is revealed in its application. To test an anonymous or disputed text, analysts first build a comparison corpus—a collection of texts by known authors. The software then analyzes the mystery document and determines which author’s “fingerprint” in the corpus it matches most closely.

This method has solved some of history’s and literature’s most intriguing puzzles.

The Federalist Papers

One of the earliest and most famous successes of stylometry involved the Federalist Papers, a series of essays written in 1787-88 to promote the ratification of the U.S. Constitution. They were published under the pseudonym “Publius”, and while authorship was known for most, a dozen were disputed between Alexander Hamilton and James Madison. For nearly two centuries, historians debated. In the 1960s, statisticians Frederick Mosteller and David Wallace fed the known works of Hamilton and Madison into a computer. They found that Madison used “whilst” where Hamilton used “while”, and Madison used “by” far more frequently. The analysis of these and other function words overwhelmingly pointed to James Madison as the author of all twelve disputed papers, a conclusion now widely accepted by historians.

The Unmasking of Robert Galbraith

In 2013, a debut crime novel called The Cuckoo’s Calling by an unknown author named Robert Galbraith received critical acclaim. When a journalist received an anonymous tip that it was actually written by J.K. Rowling, a stylometric investigation was launched. Linguist Patrick Juola compared the book to the works of Rowling and several other potential authors. The analysis revealed that the linguistic fingerprint of The Cuckoo’s Calling was a near-perfect match for Rowling’s adult novel, The Casual Vacancy. The statistical evidence was so compelling that, when presented with it, Rowling ‘fessed up. Stylometry had unmasked one of the world’s most famous authors in a matter of days.

The Forger’s Bane: Limitations and the Future

As powerful as it is, stylometry isn’t an infallible oracle. Its accuracy depends on several factors. First, it requires a significant amount of text; it’s nearly impossible to analyze a single tweet or a short email with any confidence. The “fingerprint” only emerges over thousands of words.

Furthermore, an author’s style can change over time or be influenced by genre. Someone’s academic writing will have a different stylistic profile than their personal blog. Sophisticated analysis must account for these variables. And, of course, there’s the question of intentional mimicry. While difficult, a skilled writer could potentially try to imitate another’s style to fool an algorithm, though maintaining that facade consistently is a monumental challenge.

Despite these caveats, stylometry remains a formidable tool. It has been used in court cases to analyze ransom notes (like in the Unabomber case), authenticate historical documents, and expose literary hoaxes. It serves as a potent reminder that in our increasingly digital world, our words leave behind more than just a message. They leave a part of ourselves—a unique, quantifiable, and often revealing signature, waiting to be read.