Skip to content
Concepts

Stylometry

The statistical analysis of writing style to identify or distinguish authors, using measurable features such as sentence length, punctuation habits and how often particular function words appear.

Why it matters

Stylometry is the established discipline that AI detection borrows from, and knowing the difference explains why detection is weaker than it sounds: stylometry works on substantial bodies of text with a known set of candidate authors, and AI detection has neither.

A concrete example

Scholars have used stylometry for decades to argue about disputed authorship — comparing a contested text against large samples from a handful of candidates, looking at habits writers cannot consciously control. Applying the same idea to a 500-word essay with an open set of possible authors is the hardest version of the problem, not the easiest.

How to use it

Stylometry is the older discipline behind AI detection: identifying authorship from measurable style features such as sentence length, function-word frequency and punctuation habits. It has a long history in literary and forensic work, where it is applied to substantial bodies of text by a known set of candidate authors. That context matters, because it is exactly what AI detection lacks — a short passage and an open set of possible authors is the hardest case for these methods, not the easiest.

The common mistake

Assuming techniques that work on a novelist's collected works transfer to a 500-word essay. Stylometric confidence depends heavily on sample length and on having candidate authors to compare against; strip both away and the method is being asked to do something it was never reliable at.

Related terms

Put Stylometry into practice

Access 750+ AI models and 60+ tools through Vincony — start free with 100 credits.