Processing syntax and morphology: A neurocognitive perspective (review)

Ina Bornkessel-Schlesewsky and Matthias Schlesewsky have, in recent years, contributed prolifically to the literature on the cognitive neuroscience of language. In this book, they step back and review the work that they and others have done. They emphasize research that uses the techniques of cognitive neuroscience, but discuss research adopting behavioral techniques when it is appropriate. The authors bring sophistication in both linguistics and neuroscience to the task, and the result is a book that linguists can both resonate to and trust.

The book begins with a concise discussion of the methods most widely used in cognitive neuroscience— electroencephalogram (EEG), event-related potential (ERP), magnetoencephalography (MEG), functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and transcranial magnetic stimulation (TMS)—and useful descriptions of some behavioral techniques, such as eyetracking. The presentation is remarkably clear and informative. Careful readers will realize that it is more than a list of procedures. For instance, it contains a table listing the ERP components (N400, P600, etc.) and their interpretations that have become familiar to psycholinguists who have taken even a quick look at the cog-neuro literature, but by the end of the book, the authors make it very clear that these ‘classical’ interpretations are in dire need of updating. The mapping between the typically reported components and domains of processing is far less simple than generally believed.

The substantive core of the book is divided into four sections. Part 1, ‘Syntax and morphology at the word level’, is the least satisfying. It is rather sharply focused on the question of whether inflected and derived forms are ‘rule-based’ or ‘network-based’. The authors provide clear summaries of a fairly large number of examinations of what parts of the brain are involved in processing inflected and derived forms and when different patterns of electrophysiological activity occur, but they are unable to come to any satisfying conclusions. This is not the authors’ fault. They make a convincing case that most studies of the topic have used linguistic tasks that are not themselves well understood, and worse, that differ in multiple ways from study to study so that one cannot make much sense out of apparently conflicting results.

Part 2, ‘Syntax and morphology in sentence processing’, is far more compelling. The authors provide selective but useful reviews of the theoretical and behavioral literature on sentence processing, reflecting the field’s concentration on the processing of complex sentences (sentences with relative clauses, long-distance dependencies, etc.). The first novel contribution of this part of the book, however, is its emphasis on how cognitive neuroscience techniques allow researchers to study simple sentences that do not disrupt processing in ways that show up in behavioral measures. The second novel contribution is the weight the authors place on crosslinguistic research (emphasizing work on German, but extending to work on Turkish, Icelandic, Hebrew, Japanese, etc.) and how the book uses results from this research to qualify and correct generalizations that have been made on the basis of English. The third novel contribution is how the book’s discussion of the cognitive-neuroscience findings reflects the authors’ theoretical stance. This is a strength, in that data are presented from a coherent theoretical perspective. But many readers will find it to be a serious weakness, because the authors do not make clear what their perspective is. They assume, without much justification or explanation, a functionalist approach to grammar (Van Valin 2005); they place great reliance on a hierarchy of argument types and its congruence with syntactic relations and semantic categories, but do not discuss the hierarchy until much later in the book; and they implicitly assume the distinction between processing stages assumed in their own model of sentence processing, the extended argument dependency model (eADM), which they have not yet discussed at this point in the book. These latter two weaknesses are corrected in a brief but lucid discussion of the eADM that...