stilometry

Stilometry is a quantitative method used to identify the author of a text. It involves analyzing various statistical features of writing style, such as sentence length, word frequency, punctuation usage, and the frequency of specific grammatical structures or function words. The underlying principle is that authors have unconscious, consistent habits in their writing that can be measured and compared.

The process typically begins with a corpus of known texts by different authors. These texts are analyzed to extract a set of stylistic features. Then, a questioned document, whose author is unknown or in dispute, is analyzed using the same set of features. The results from the questioned document are then compared to the stylistic profiles of the known authors. Sophisticated statistical techniques, such as discriminant analysis or machine learning algorithms, are often employed to determine the most likely author.

Stilometry has been applied in various fields, including literary studies to attribute anonymous or disputed works, and in forensic linguistics to identify authors of anonymous threats or legal documents. While powerful, stilometry is not infallible. Its effectiveness can depend on the size and representativeness of the texts analyzed, the number and relevance of the features chosen, and the complexity of the writing styles being compared. Additionally, co-authorship, editing, and deliberate attempts to mimic another's style can pose challenges to accurate attribution.