SOAS Research Online

Abstract

Chapter 1 summarizes the debate over the age and authorship of the Arthasastra and proposes to test the common assumotion that it is the work of a single author. Chapter 2 analyzes the five versions of the story of Candragupta and Canakya or Kautilya and finds that the Jain version best preserves the original legend, being closely paralleled by the Pali; that the Kashmirian version is late, and the Mudaraksasa largely fictive; that the Classical version, while betraying its Indian origin, gives uncertain testimony as to the content of the original legend; and that Canakya is an historical figure. Chapter 3 finds, in the structure of the Arthasastra, a priori grounds for supposing a composite authorship; summarizes some previous studies of authorship using statistical methods; and reports the results of a pilot study of the Arthasastra which throws doubt on the assumption of a unique author. Chapter 4 examines the distribution of certain words in Sanskrit works of known authorship, and having found that eva, evam, ca, tatra and axe safe discriminators of authorship, examines their distribution in the Arthasastra. Books 2, 3 and 7 of the Arthasastra, by this test, are homogeneous within themselves but are the work of three different authors. The affiliations of the shorter books are discussed. Chapter 5 inquires whether sentence-length and compound-length may he used to discriminate between different authors, and finds the former unacceptable but the latter promising. Chapter 6 exajnines Arthasastra passages used by Bharuci and Medhatithi in their commentaries on Manu and finds in the latter's reference to an Adhyaksapracara a possible predecessor of the Artharsastra. Chapter 7 reviews the conclusions as to the composition of the Arthasastra in the light of a statistical study of Vatsyayana's Kamasutra and briefly comments on the date and authority of the Arthasastra.