Testing Delta on the «Disputed Federalist Papers»

Abstract

The Federalist Papers stand out as an excellent proving ground in the field of authorship attribution, being nowadays considered a breaking issue in literary detection. The crucial point of the Federalist Papers is the set of the Disputed Papers, twelve articles traditionally attributed either to Alexander Hamilton or James Madison. This authorial obscurity, together with the existence of undisputed samples, surely explains the proliferation of studies trying to spot the hand responsible for the Disputed Papers, particularly throughout the second half of the 20th century, both with traditional and non-traditional approaches. Since the publication of Mosteller and Wallace’s masterpiece, there has been a consensus as to consider them exclusively Madisonian (Mosteller & Wallace 1963: 300; 1964: 16). Notwithstanding this incessant activity on the Federalist Papers as a test probe for authorial purposes, the use of Burrows’ Delta is still deemed a desideratum in the field, a technique proposing that the salient features which characterize an author’s style can be obtained from the hierarchy of the most common function words (Burrows 2002: 267-87; 2003: 5-32). The present paper then proposes the testing of Burrows’ model in a twofold version: a) modified Delta; and b) simplified Delta. The results come to corroborate the lexical differences between Hamilton and Madison, a fact allowing us to validate the hypothesis of the Madisonian composition of the Disputed Papers, exception being made of Paper 55.