Definite articles in Ancient Greek

The creation of a definite article is a relatively recent phenomenon in many Indo-European languages: most of the attested ancient varieties lack definite articles, which developed at later stages. Greek is one of the few Indo-European groups where a definite article is attested at very early stages. This paper investigates whether its distribution and syntactic properties in Ancient Greek comply with crosslinguistic standards. It will be shown that a definite article was in fact fully developed in Classical Greek and in the Hellenistic koiné of the Gospels, where its distribution conforms to syntactic restrictions visible in many other languages; cases of free alternation between articulated and articleless nominals, in these varieties, are entirely predictable from the interaction of specific properties that affect the representation of formal features in their nominal structures. The apparent contradictory evidence found in Homer is usually regarded as the expression of a diachronic stage at which the item ὁ, ἡ, τό was embrionically a definite article but not yet generalized as such. Here, instead, the Homeric evi- dence will be connected, following Bozzone and Guardiano 2015, with the composite diachronic stratification of the language of epic, and in particular to the combination of (at least) two opposing grammars, one more conservative, where a definite article did not exist, and one more innovative, where ὁ, ἡ, τό was fully developed as such. The paper is structured in five sections: after the introduction (§1), §2 defines the theoretical and empirical background, §3 analyzes the uses of ὁ, ἡ, τό in Classical and Hellenistic Greek, §4 focuses on Homer, and §5 sums up the main conclusions.