The first thing that is odd about this book is its title and, especially, its subtitle. Most readers would probably expect a book with that titIe to show how recent developments in the MT community as a whole have led to new insights concerning the relationship between syntax and semantics. The more semantics-minded among them might even be tempted to believe that this book contains the deathblow to all syntax-oriented research in MT. All of them will feel misled to find that the book offers nothing of the sort What the book does offer is a collection of papers all of which describe research carried out by members of both the German Language Group of the EUROTRA project (at the Institut ftlr Angewandte Informationsforschung, Saarbrtlcken) and other, subsidiary, German projects (at the universities of Berlin and Stuttgart). The papers are organized in four sections: (I) "A coherent system -theory and implementation"; (II) "Semantic relations in an MT environment"; (III) "From source language to target language -aspects of transfer"; and (IV) "Explorations."
The introduction (pp. l-9) remarks right away that the book "is not some official report on EUROTRA work" (p. 1). Nevertheless, it goes on to say, "almost all of the ideas presented here have had an impact on the project." At first blush, one might therefore wonder whether a subtitle like "Insights from the EUROTRA MT System" would not have been more appropriate for a book that is larded with the acronyms EUROTRA and EUROTRA-D.
However, a closer reading of the first three chapters already reveals two things. First, the insights reported on are not really insights gained during MT experiments; rather, the insights are from theoretical linguistics (Lexical Functional Grammar, Systemic Functional Machine Tramlation 4:3 (1989)