This paper aims at exploring the polysemy of the Modern Greek prepositional prefix iper- within Lieber's (2004, 2007) semantic framework, by using corpus-based findings on its meaning and productivity. On the basis of our data, drawn from the Corpus of Greek Texts, two large meaning super-categories are identified, namely SPATIAL LOCALIZATION and NON-SPATIAL LOCALIZATION; the latter is further distinguished into EVALUATIVE and NON-EVALUATIVE, of which EVALUATIVE is further divided into several meanings. The measurement of productivity across meanings has shown that NON-SPATIAL and EVALUATIVE meanings are more productive, while SPATIAL and NON-EVALUATIVE meanings tend to be much more restricted. Following Lieber's model, we suggest that iper- displays a unitary skeleton with the feature [+Loc], which accounts for the non-evaluative meanings of the prefix, while the evaluative meanings can be analyzed as an extension of non-evaluative meanings. The meanings identified in the corpus are regrouped in the light of this analysis and are further distinguished within each body by the different features they display in their bases. Lieber's model is found to have cross-linguistic validity and to be useful in organizing a large variety of meanings into a few main structures. The fact that the semantic analysis is informed by corpus-based findings on the meaning and productivity of the prefix seems to provide a firmer ground than intuitive or lexicographic examples.
