The syntax of argument structure by Leonard H. Babby (review)

Leonard Babby is a scholar whose investigations into Russian and Slavic linguistics made the field what it currently is. Over the past few decades, his work has always been a source of inspiration for those who seek a better understanding of the structure of Slavic languages and for those who wish to integrate their material into linguistic theory. The book summarizes much of B's previous work on diverse topics in Russian syntax, which includes, but is not limited to, control, case, structure of adjectives, nominalizations, and verbal adjuncts. Presenting mature results of detailed reflection, however, the book introduces an entirely new perspective on how a number of apparently unrelated phenomena can be approached. B offers a novel theory of argument structure and shows how the most complicated syntactic phenomena can be successfully dealt with by reducing an explanation to argument structure considerations. The reader will benefit from acquaintance with this outstanding piece of research in at least two respects. First, this work seems to have every chance of becoming a bedside book for the next generations of Slavic linguists: so wide is its empirical coverage and deep are its observations and generalizations. Second, B uncompromisingly goes against the current reductionist trend of simplifying presyntactic, lexical computation to the absolute minimum, ideally to zero: he articulates lexical representations of argument structure and defines a number of operations on them. The reader who believes that a model with an impoverished lexicon is to be preferred on conceptual grounds will be interested to learn how far a theory can take us that assumes exactly the opposite.

Most of the book is exclusively based on Russian material. The only exception is Ch. 1, where architectural principles of the theory are established. B assumes a rich lexicon in which predicate argument structure is built up and various structure-changing operations apply. The essential theoretical construct is a DIATHESIS assigned to all lexical stems and most derivational affixes, which stores item-specific properties of argument structure. The diathesis is thought of as a pairing of two tiers, a θ-ROLE SELECTION TIER and a categorial tier, or C-SELECTION TIER. For instance, 1 is a diathesis of a ditransitive verb.

1. 1.

   Full notation				Flat notation
   i	j	k	-	{{i^N}1 {j^N}2 {k^N}3 {-^V}4}
   N	N	N	V
   1	2	3	4

   where i, j, k are θ-roles, N is a categorical noun head, V is a verb-stem head, and numbers represent a relative position of an argument in the diathesis. ^ in the flat notation is a linking relation.

The θ-role tier represents thematic information about arguments of a stem, and the c-selection tier indicates what lexical category the argument is associated with in the syntax. Slots in the two tiers are linked pair-wise; hence an argument (e.g. {i^N}₁) is a pairing of a θ-role (e.g. i) and a categorial label (e.g. N). A stem can have up to three arguments, so the complete diathesis consists of the four pairs of slots, one for the stem and three for its arguments.

Diathetic representations entirely determine grammatical relations within vP, which places the theory in line with projectionist (as opposed to constructionist, e.g. Borer 2005 and Ramchand 2008) approaches to grammar. In the syntax, arguments merge one by one, left to right, so that {k^N}₃ appears as a complement of V; {j^N}₂ is mapped to Spec, VP position; and the external argument, {i^N}₁, merges at Spec, vP.

The two-level organization of argument structure is the most innovative aspect of B's approach to the lexicon-syntax interface (even if superficially B's diathesis bears remote resemblance to Mel'cuk's (1974 and further work) notion of model' upravlenija 'valency pattern'). B's crucial argument for keeping the θ-role and c-selection tiers apart comes from the set of observations suggesting that neither...