Funktionsverbgefüge und automatische Sprachverarbeitung

Funktionsverbgefüge und automatische Sprachverarbeitung. By Stefan Langer. (Linguistic resources for natural language processing 3.) Munich: LINCOM Europa, 2009. Pp. 207. ISBN 9783929075649. $87.64.

Reviewed by Louisa Buckingham, Sabanci University Writing Center, Turkey

This monograph investigates how Funktionsverbgefüge, commonly known as support verb constructions (SVC) in English (verb-noun combinations with a delexicalized verb), can be identified in electronic corpora by automatic language processing. The book is the product of Stefan Langer’s Habilitationsarbeit at the Ludwig-Maximilians University, Munich, Germany. Consisting of six chapters, the main body of L’s investigation is found in Chs. 3–4. In the introductory chapter, L outlines the importance of morphological, syntactic, and semantic characteristics of these constructions in the context of automatic language processing, and goes on to discuss the position of SVCs in machine translation as well as automatic text creation and retrieval.

Ch. 2 is dedicated to an overview of the variation in terminology used to refer to SVCs. For scholars with a general interest in this construction, L’s summary of the reasoning behind each of the terms used to refer to SVC is both useful and illuminating. He offers a coherent explanation for the differences in the conceptualization of closely related phenomena, specifically Funktionsverbgefüge, Stüzverbkonstruktion, complex predicates, and light verbs. His explanation of Funktionsverbgefüge and Stüzverbkonstruktion is of great importance for his analysis of SVC in the remaining sections of the book. He essentially distinguishes between two types of SVC; those with a nominal complement in accusative case and those with a prepositional phrase complement. L demonstrates these nominales Funktionsverbgefüge (e.g. Kritik üben ‘criticize’) and adverbiales Funktionsverbgefüge (e.g. in Beziehung stehen ‘either’) respectively.

In Ch. 3, L discusses the degree to which the extraction of SVC through statistical methods from electronic corpora is effective. He concludes that although such methods may be used to identify word combinations, they are unable to distinguish between different types of combinations, such as idioms, semi-formulaic expressions, and free combinations. Statistical methods are more equipped at eliminating semantically meaningless combinations than as an aid for the classification of word combinations. In light of this, the author examines the effectiveness of linguistic criteria to identify SVC in Ch. 4. The twenty-two test criteria he applies (only some of which are able to be used in automatic language processing) consist of the most common characteristics used to define SVC, focusing in turn on the nominal phrase, the support verb, and the degree of compositionality of the SVC.  L demonstrates their application on two SVCs: nominales Funktionsverbgefüge (Vorlesung halten) and adverbiales Funktionsverbgefüge (in Schwerigkeiten stecken). The author concludes that purely automated methods are not sufficient to identify SVCs and a manual analysis is still required. In the final chapter, the author briefly considers the analysis of SVCs from a frame semantic perspective.

The book is unusual for its breadth of treatment of SVCs. The author addresses general concerns, such as the validity of the concept of SVCs (the question has received some attention in recent publications in German), the methods used to investigate SVCs in corpora, and the relative difficulties involved in working with electronic corpora comprised of texts from the internet.  Additionally, L examines specific questions regarding SVC structure and how these combinations have been conceptualized in different languages. This work will appeal to students and scholars of syntax and phraseology, specifically those conducting research on SVCs in German and other languages.