The following description is taken from an affidavit by Gerald R. McMenamin, one of the leading scholars in the field; the affidavit – from Case 1:10-cv-00569-RJA -LGF Document 50 Filed 06/02/11
As part of his expert witness statement, McMenamin describes the theoretical and practical foundation of the method by which he examined documents and determined that they were not written by Facebook founder Mark Zuckerman. See http://www.nytimes.com/2011/07/24/opinion/sunday/24gray.html, :
These are the principles I apply to my authorship analyuses
“This approach to author identification is based on two principles generally accepted, and well-documented in peer-reviewed contexts: author-specific linguistic patterns are present in unique combination in the style of every writer, and these underlying patterns can be empirically described and often measured by careful linguistic analysis, making author identification possible.
“A language is at one and the same time owned by its whole group of speakers but uniquely used by individuals from that group. Why one writer chooses linguistic form A and another chooses form B has two possible causes: differences in what they individually know of the language, and differences in how each one uses the core of linguistic knowledge they have in common as speakers and writers of English. Individual differences in writing style are also very often due to an individual’s choice of available alternatives within a large, shared common-pool of linguistic forms. At any given moment, a writer picks and chooses just those elements of language that will best communicate what he/she wants to say.
“The writer’s ‘choice’ of available alternate forms is often determined by external conditions and then becomes the unconscious result of habitually using one form instead of another. Individuality in writing style results from a given writer’s own unique set of habitual linguistic choices. Identification and analysis of a writer’s choices, i.e., of his or her style markers, constitute stylistic analysis, which is well established as a generally accepted and peer-reviewed method of author identification in both literary and forensic contexts.”