100% Austen

sty·lo·met·rics |stīˈlämitriks|
1. A method of studying literary style, frequency of word usage, etc by means of the statistical analysis of a text, usually with the aid of computer programs, which allows for the charting of historical changes in style, the investigation of questions of disputed literary authorship, the authentication or otherwise of written evidence, police statements, etc.

That expert you heard on npr discussing an unattributed text fragment found in a crypt in Bath, England was a Stylometrist. She transcribed the text of the text fragment and used a computer program to analyze the style of the text comparing it to contemporary authors George Crabbe, Mary Wollstonecraft, and Charles Lamb. Given the sample, the text fragment is determined to have a 98.3% chance of having been written by Ms. Austen.

A similar method was used to attribute a poem to Shakespeare and to determine the authorship of the book “Primary Colors” (originally attributed to “anonymous”). Stylometrics measures the frequency of certain words, contiguous word patterns, word length, etc., and produces a signature of the author’s style by subjecting all known writing to this analysis. Then new, unattributed texts may be compared to these signatures and the likelihood percentage is calculated.

What would happen if you asked the Stylometric software to produce text that has a 100% chance of having been written by Jane Austen? Such a text has been found. “It looks like Austen, it smells like Austen, so it must be Austen.”

100% Austen (detail)