Abstract
This thesis investigates the diachrony of inflection classes, with a particular focus on which notions of morphological complexity can be relevant as motivating factors for change in the structure of inflectional systems. The inflection of nouns in Greek is taken as a case study offering 2,500 years of relatively well-recorded development The changes directly affecting the inflectional marking of nouns from the (reconstructed) Proto-Indo-European origins of the language through to Modern Standard Greek are examined, together with the shifting relationship between inflection class and gender across this period. To address these issues, the evolution of the Greek noun system is modelled within the framework of Network Morphology, and quantitative metrics of complexity, including both information-entropy- and principal-parts-based approaches, are ap-plied to various stages of the language’s history. This thesis demonstrates that, while complexity does play a role in many instances of "internally-motivated" morphological change, such change cannot be ascribed to a single unified notion of morphological complexity, but there are in fact multiple different types of complexity which can affect inflectional systems in different ways, sometimes in conflict with one another.