Hence the study of Greek divinities must not be based on the assumption that the divine personality of a deity was substantially the same throughout the Greek world.

Selections from Christiane Sourvinou-Inwood’s Persephone and Aphrodite at Locri: A Model for Personality Definitions in Greek Religion
The first determining factor is clearly the worshipping group and its specific realities and needs as they develop in the course of time. Deities are shaped by the societies that constitute the worshipping group and develop with them. A second factor to be taken into account is the pantheon to which they belong and the spheres of activity of its members. For a pantheon is an articulated religious system within which divine beings catering for the needs of the worshipping group are associated and differentiated; and this nexus of relationships contributes to the definition of each divine personality

The concepts ‘worshipping group’ and ‘pantheon’ bring us to an important aspect of Greek religion: the fact that the Greek deities existed at two levels-the local, polis level, and the Panhellenic level.

Too often, in the study of Greek divinities, the local personality of a deity is overshadowed by the Panhellenic one and the individuality of the different local deities is ignored. However, it is extremely unlikely that the establishment and crystallization of a Panhellenic persona for a deity so stamped out the local personalities that only insignificant variations remained. For the parameters affecting their definition differed in different cities, and again at the Panhellenic level. The realities and needs of the worshipping groups differed while some were common to all and also operated at the Panhellenic level. The composition and hierarchy of the pantheon also differed in the different cities, and again at the Panhellenic level. Moreover of the spheres in which a divine personality manifests itself3 that of cult would be especially resistent to change under the impact of Panhellenic religion; for cult operates primarily at the polis level, having a function within its structures

Hence the study of Greek divinities must not be based on the assumption that the divine personality of a deity was substantially the same throughout the Greek world. Consequently to avoid the danger of distortions we must study each local divine personality of a deity separately from the Panhellenic one, and not use evidence from the latter to determine the former. Instead, we must recover each local manifestation of the personality, and then relate it to the Panhellenic persona. Moreover, we must not extrapolate from one local cult to another and attempt to interpret an aspect found in one place through another found elsewhere. Nor should we conflate evidence from different parts of the Greek world. The result would be a totally artificial conflation that had no cultic or theological reality. The fact that a given function is, for example, associated with Aphrodite at Sparta only means that this function belongs to her in the context of a particular personality nexus. It is not necessarily found in all, or indeed any, of her other personality nexuses which, I have argued above, had a different profile. Nor is it an inalienable part of an integral complex which included all the aspects of Aphrodite from the whole of the Greek world, and which would be ‘the’ Aphrodite. For divinities only existed at two different levels of cultic reality: local and Panhellenic. No aspect of a deity has any significance when separated from its organic context. Tendencies and aspects common to a deity throughout the Greek world have to be recovered and tested, not assumed or extrapolated.

To sum up, the study of Greek divine personalities should be based on specific local religious units and rely on internal evidence alone. The Panhellenically consistent traits would then be recovered and tested. It should be clear that this method based on the study of local cults is valid whether or not my initial analysis is correct. For it is a neutral approach that does not introduce any preconceived distortions. Moreover, it allows a circumscribed investigation of the circumstances of the worshipping group, necessary for the study of the development of any one deity, and of the other deities of the pantheon to whom it related.