Reviewed by Katrin Hiietam, Manchester, UK

In this extremely thorough and detailed account of homonymy in two-argument agreement paradigms, Trond Trosterud explores languages from three Uralic language families: Mordvin (Erzja, Mokša), Ugric (Hungarian, Mansi), and Samoyedic (Nenets, Selkup, and Kamas). In these languages, the verbal morphology incorporates not only the person and number of the subject but also the person and number of the object. In contrast to more simplex systems, in which there is distinct morphology for each person-number combination, two-argument paradigms often contain verb forms that are ambiguous between several readings. This homonymy, T argues, is typical of object paradigms because it would be difficult to fill each paradigm cell with a distinct form.

In general, T aims ‘to develop a theory that can restrict the set of possible two-argument paradigms’ (11). Because these paradigms inflect for multiple occurrences of the same type of morphosyntactic properties (e.g. person, number, tense, mood, case), T concludes that such complicated subsystems tend to collapse more readily than other parts of the inflectional system.

Within the framework of generative grammar, T explains the modular architecture of grammar and his lexeme-based approach to morphological components. He concludes that homonymy exists in almost all paradigms and that homonymy increases in cases in which more morphosyntactic categories are incorporated and in which properties have more than two values. Moreover, throughout the paradigm, the shape of paradigms is restricted and the homonymy is not randomly distributed: homonymous forms occur in the same paradigmatic positions across multiple languages. Based on these patterns, T proposes a geometrical homonymy condition, which states that homonymy can only occur in neighbouring cells within a multidimensional paradigm.

T illustrates that homonymy is dependent on semantic markedness and usage frequency: more marked forms are systematically used less often than less marked forms (307). Furthermore, T does not exclude the possibility of phonology playing a role in homonymy. He hypothesizes that phonological and morphological changes may work together to facilitate speech production and speech comprehension.

This volume will be a useful tool for Uralists with a special interest in morphology and homonymy. The book includes plenty of graphs, schemas, and tables that contribute to the discussion.