|Title||Pontic Greek as a conservative innovator in the development of gender agreement in Asia Minor Greek|
Alongside the syntactic gender agreement system that it inherited from earlier stages in its history as a Modern Greek (MGr) dialect, Pontic Greek exhibits a semantic agreement system (Koutita-Kaimaki 1988/1989; Oeconomides 1890; Papadopoulos 1955) that represents a clear diachronic innovation when compared to the situation found in the overwhelming majority of MGr dialects, in which gender agreement can only be syntactic. This paper provides (a) a theoretically-informed description of the synchrony of Pontic gender agreement, and (b) a historical account of the origin and subsequent diachronic development of the innovative semantic agreement system, both adopting Corbett’s (1991, 2006) typological framework.
The synchronic analysis shows that the distribution of the two agreement systems in Pontic—syntactic and semantic—is conditioned by the morphological and semantic properties of agreement controllers, viz. their morphologically-assigned gender value and the position their referents occupy on Dahl’s (2000) Animacy Hierarchy, as well as by the position agreement targets occupy on Corbett’s Agreement Hierarchy. Human nouns, whose referents are found at the high end of the Animacy Hierarchy, trigger syntactic agreement on all kinds of agreement targets: η μικρέσσα η νύφä έτον κι άλλο πονηρέσσα (Chaldía Pontic; Drettas 1997: 684). On the contrary, the overwhelming majority of targets controlled by low-end, inanimate nouns—which can be morphologically-assigned to either the masculine or the feminine gender—appear in their neuter form to agree with the semantic properties of their controllers. The singular forms of the definite article that agree with their controllers syntactically when immediately preceding them are the only exception to this pattern that is otherwise found in all agreement domains: τ’(ο) ασημένιον ο μαστραπάς πάλι κρέμεται (Oenóe Pontic; Lianidis 2007 : 228). The combined effect of animacy and gender is particularly well-illustrated by nouns denoting non-human animate entities. As shown in the following example, targets controlled by masculine nouns of this type agree with them syntactically and appear in their masculine forms; targets controlled by feminine nouns agree with them semantically in appearing in the neuter form: τα κάτας εξενίτεψαν κι οι πεντικοί χορεύνε (Papadopoulos 1961: 215). Therefore, the distinction conditioning agreement with masculine nouns is animate vs. inanimate; semantic agreement with feminine nouns is based on the human vs. non-human distinction.
This difference is taken here to suggest that the innovative semantic distinction that first became operative in Pontic agreement was between animate and inanimate, with the neuter gender expressing the part occupying the lower end of the Animacy Hierarchy. This original distinction was later redefined as human versus non-human with the neuter being again associated with the expression of the lower-end part of the distinction, as shown by feminine nouns. The preservation of syntactic agreement on the definite article—the target that is found closest to the controller—for all semantic types of nouns in the singular suggests that semantic agreement initially applied in a domain outside the NP. Drawing further on the evidence from Óphis Pontic in examples such as πούλησο με αού το σ̑κύλλο· κ̔ι πορώ να πουλώ ατό (Lianidis 2007 : 242) that show that the only target with which semantic agreement is possible when controlled by a masculine noun denoting a non-human animate entity is the personal pronoun, I argue that that exactly was the first target to express the distinction between animate and inanimate when this became initially operative in Pontic agreement—a proposal that complies with the typological findings of Corbett (1991, 2006) and Greenberg (1978). From there, I further postulate that semantic agreement was extended to more targets along the path defined by Corbett’s Agreement Hierarchy. As for the origin of semantic agreement in Pontic, I propose that it is to be found in the conflict between the semantic and morphological properties of inanimate masculine and feminine nouns, which were thought by speakers to belong to the ‘right’ gender for their morphology but to the ‘wrong’ gender for their semantics (see Topcharas’s comment in 1998 : 12) thus triggering the resemanticisation (in the sense of Wurzel 1986) of the Pontic gender system evidenced in (2-3). What is more, I show that this system can be posited to have preceded other, more advanced gender developments attested in the MGr dialects of Asia Minor, most notably the loss of gender agreement in Cappadocian and Pharasiot. Pontic therefore figures as a conservative innovator in the development of gender agreement in Asia Minor Greek.
|Conference||5th International Conference of Modern Greek Dialects and Linguistic Theory|
|Accepted author manuscript|