Productivity in Historical Linguistics: Computational Perspectives on Word-Formation in Ancient Greek and Sanskrit

Abstract

``Productivity'' is a simultaneously familiar and fraught topic for all linguists. Every linguist believes that he can recognize it, yet grasping what properties characterize a ``productive'' process in opposition to a non-productive one is difficult. The historical linguist ought to be particularly desperate for a definition that can be operationalized -- distinguishing archaism from innovation depends upon it. This dissertation is therefore first concerned with transforming ''productivity'' into an object that can be interrogated, and then seeking tools that can provide a useful characterization that object. On the explicitly diachronic side, I am concerned with how to measure, diagnose, and motivate changes in productivity. Corpora from two the oldest-attested Indo-European languages, Ancient Greek (here, mainly the Iliad and Odyssey, as well as the New Testament) and Vedic Sanskrit (here, mainly the RÌ¥gveda) will guide and define these explorations. Within the realm of morphology and word-formation especially, I will argue that concerns about productivity rightly take pride of place in diachronic discussion, and that those discussions become more meaningful when made precise and psychologically motivated. This increased precision offers hope of accounting for diverse linguistic phenomena for which the presence or absence of morphological structure is a crucial determinant. Overall, this work calls for the increased usage of corpus-based quantitative methods and computational modeling in the study of language change.