MAHIEU, Marc-Antoine and Nicole TERSIS (eds), 2009 Variations on Polysynthesis: The Eskaleut Languages, Amsterdam and Philadelphia, John Benjamins Publishing Company, Typological Studies in Language, 86, 312 pages.

This book contains 18 papers from the proceedings of the 15th International Congress of Inuit Studies. The papers are grouped into three parts. The first part, “Polysynthesis,” comprises 6 papers. In the first paper, Mithun revisits the concept of polysynthesis to show how it has evolved over time and why most scholars working on Eskaleut languages consider them polysynthetic whereas a recent and influential book (Baker 1996) rejects this view. Through a review of the evolution of this concept since Duponceau (1819) and the evolution of the related concepts of noun incorporation and holophrasis, she shows that Eskaleut languages exhibit many features traditionally attributed to polysynthesis. Her conclusion provides a bridge between Duponceau’s mostly semantic definition and Baker’s highly constrained formal definition: “If polysynthesis is defined as having many morphemes per word, Eskaleut languages are clearly polysynthetic, just like Iroquoian languages, if not more so (p. 15).” De Reuse also takes issue with Baker’s restrictive definition. Drawing on examples from Central Siberian Yupik Eskimo, Western Apache, French, and Dutch, he suggests that polysynthesis does not refer to a type of language but rather to the presence of a large quantity of productive noninflectional concatenation (PNC) elements. These elements are distinguished from inflectional and derivational elements by a number of features, the most important one being productivity. According to De Reuse, morphology can be split into three types: inflectional, derivational, and PNC, where derivational morphology is limited to nonproductive derivation as opposed to productive PNC. The distinction between polysynthetic languages and other languages would thus be described quantitatively by the presence of PNC instead of qualitatively as a special type. Fortescue takes a much more functional approach. He points to a parallel between auxiliary constructions in Chukchi and some West Greenlandic suffixes with similar functions. He hypothesises that this parallel came about through the Eskaleut tendency to incorporate nouns and adjuncts. Once absorbed into the verb complex, these earlier auxiliaries “would [lose??] any special discourse function they once have had, but gaining more specialized semantic meaning on the way” (p. 48). Tersis discusses construction of novel lexical entries. Whereas most stems are nominal, adjectival, or verbal and suffixes are either denominal or deverbal, some suffixes seem to escape categorisation or to belong to more than one lexical category. She proposes a continuum ranging from lexical incorporating suffixes to totally grammaticalised suffixes, which would be similar to inflectional suffixes. Vakhtin takes issue with the levelling and ordering of morphemes generally suggested for Eskaleut languages. He suggests that morphemes might have different origins and been “incorporated” into polysynthetic structure at different moments in the evolution of these languages. He presents different phonological phenomena and repetition of some suffixes to support his claim and the possible “auxiliary” origin of some deverbal postbases. Miyaoka looks at comparable constructions in Central Alaskan Yupik (CAY) to reject a simple “slot-and-filler” analysis. He suggests, through a resolutely functionalist approach, that polysynthesis structure is quite different in CAY and Chukchi or Athabaskan for example. He attributes these differences to limited use of prefixation, reduplication, noun incorporation, and a fair number of valency-increasing suffixes in CAY. The second part, “Around the Verb,” contains five papers. In the first paper, Sadock looks at the anaphoric relations between personal markers in verbs and case markings on nominal expressions surrounding them in Aleut. He shows that the anaphoric system in Aleut is quite efficient even though the case system has fewer distinctions in Aleut than in other Eskaleut languages. To achieve this efficiency, Aleut uses “not just the meaning of individual expressions, but the available contrasts with other expressions in the grammatical domain …