Scripture Stylometry

Function words such as articles, prepositions, conjunctions, and particles are used habitually and at rates that depend little on what a text is about, so they reflect an author's ingrained style rather than the topic. They are also used unconsciously, which makes them hard to imitate or disguise deliberately. Content words, by contrast, track subject matter, so two texts on the same theme by different authors can share much vocabulary, while two texts by the same author on different themes can differ greatly. Burrows's Delta and most stylometry therefore build their fingerprint from the most frequent, largely content-neutral words.

Delta is the average, over a chosen set of the most frequent words, of the absolute difference between two texts' standardized frequencies (z-scores) for those words. Each text is first turned into a profile of how far its use of each common word deviates from the corpus average; Delta then sums up those deviations' differences. A small Delta means the two texts depart from the norm in similar ways and are stylistically close, suggesting common authorship; a large Delta means they diverge. The standardization is essential, because it prevents the very commonest words from dominating and lets each feature contribute on a comparable scale.

It can provide strong, independent evidence but rarely a final proof. Stylometry shows whether texts are stylistically consistent with common authorship and can corroborate or undercut traditional claims, especially when validated on known-author samples and tested for significance. But it is limited by short or composite texts, by the confounding effects of genre, editing, and translation, and by the fact that absence of a stylistic match does not always prove different authorship. For scripture it is best treated as one line of evidence alongside historical, source, and redaction criticism, strengthening a case rather than deciding it alone.