Investigating the linguistic representativeness of Early Modern Greek Corpora

Following a poorly documented period in the history of vernacular Greek (6th-12th c.), the late 15th century sets the beginning of a linguistic era characterized by a quantitatively and qualitatively incomparable production of prose texts written in “common” language. It is at this point that classicizing Greek stops dominating in writing, and a new linguistic variety – albeit a very diverse and fluid one – Early Modern Greek (EMG) starts growing rapidly as a literacy language. The development of this new variety is manifested in its widespread use as literary language (in texts with aesthetic function), as well as in its use as a simple scripta, namely a written vernacular for legal, administrative, commercial, and other functions. Despite its significance in the history of Greek, this period remains to a large extent unexplored and underrepresented in Greek language corpora. On this view, our understanding of EMG depends crucially on the representativeness of the few available corpora. The aim of this paper is to investigate the linguistic representativeness of EMG corpora, and to explore possible associations between observed linguistic patterns and corpora design. Focusing on the distribution of contrastive and reformulation markers, our study reveals that the linguistic data illustrated in the available EMG corpora are divergent and largely dependent on the representation of variables, such as text form (poetry/prose), period, geographical region, and genr