In lieu of an abstract, here is a brief excerpt of the content:

18 SHOFAR Summer 1993 Vol. 11, No.4 JOSEPHUS, SlYLOMETRY, AND JEWISH STUDIES by David S. Williams David Salter Williams earned a Ph.D. from Hebrew Union College-Jewish In'stitute ofReligion in 1988 and has served as an Assistant Professor of Judaica' at the University of Georgia since 1989. Stylometry is the numerical study of textual style.l It is a valuable tool in many endeavors ofliterary criticism, such as analysis of the development of a language over time, or close study of an author's style to discern shades of meaning. Stylometry is perhaps most often used, however, for the evaluation ofauthorship. This is the application of stylometry on which I wish to focus in this paper. In recent decades, scholars from various fields have made use of stylometric techniques to make judgments about authorship problems. Well accepted studies have determined that an eighteenth-century essayist known as Junius was actually Sir Philip Francis,2 that twelve unattributed essays of the Federalist Papers should be assigned to James Madison,3 and lThis paper has been inspired by J. Allen, "Methods of Author Identification Through Stylistic Analysis," French Review 47 (1974), pp. 904-916. For a general overview of stylometry, see A. Q. Morton, Literary Detection: How to Prove Authorship and Fraud in Literature and Documents (New York, 1978). Two helpful studies for understanding stylometric procedures are by A. ]. P. Kenny: The Aristotelian Ethics: A Study of the Relationship Between the Eudemian and Nicolnacheaen Ethics ofAristotle (Oxford, 1978) and A Stylornetric Study of the New Testament (Oxford, 1986). See also my Siylometric Authorship Studies in Flavius]osephus and Related Literature (Lewiston, NY, 1992). 2A. Ellegard, Who Was ]urlius? (Stockholm, 1962); idem, A Statistical Method for Determining Authorship: The]unius Letters 1769-1772 (Gothenburg, 1962). 3F. Mosteller and D. L. Wallace, Inference and Disputed Authorship: The Federalist Papers (Reading, MA, 1964). See also idem, Applied Bayesian and Classical Inference: The Case of the Federalist Papers (New York, 1984). Josephus, Stylometry, andJewish Studies 19 that Mark Twain did not write, as sometimes thought, ten letters published in 1861 under the pseudonym of Quintus Curtius Snodgrass.4 Similar issues in Jewish Studies have yet to receive stylometric attention. It is desirable, then, for those who pursue Jewish Studies to discover the value of stylometry and to put its techniques to use. One should not expect statisticians to identifY problems which are of interest to scholars ofJewish Studies. Optimally, authorship questions should be handled by persons with primary training in literary fields, who are willing to learn about basic statistical procedures. While a background in statistics may be necessary for developing new methods of measurement, if one applies techniques already proven effective by others, then extensive statistical knowledge becomes less important.5 The goal of this paper is to demonstrate a simple, yet effective, method for stylometric authorship evaluation. While I will concentrate on ancient Greek texts, the approach I will discuss is transferable to texts· of other languages. My hope is that if I show how powerful and straightforward stylometric analysis can be, then perhaps others will become interested in utilizing it as a supplemental research tool. I. Before turning to my demonstration, I should discuss first some background matters. To begin With, "the use of stylometry in authorship attribution studies depends upon the hypothesis that there are quantifiable features of style which are characteristic of [individual] authors.,,6 In particular, word selection and the frequency of usage of common words are often habitual choices made by writers. A given habit of usage for one author-which feels entirely "right" to that author-may well feel uncomfortable to another writer. An example of a stylistic choice is the option in English of using "on" or "upon." This alternative offers a range of possibilities: a writer may utilize one of the two terms exclusively, use both more or less equally, or favor one over the other. Stylometry provides ways to isolate and to measure such selections. 'c. S. Brinegar, "Mark Twain and the Quintus Curtius Snodgrass Letters: A Statistical Test of Authorship," Journal of the American Statistical Association 58 (1963), pp. 85-96. 5Cf. Allen, "Methods of Author Identification," pp. 905-906. 6Kenny...

pdf