Koordinationsreduktion und Verbstellung in einer generativen Grammatik des Deutschen By Manfred Kohrt (review)

682LANGUAGE, VOLUME 54, NUMBER 3 (1978) Koordinationsreduktion und Verbstellung in einer generativen Grammatik des Deutschen. By Manfred Kohrt. (Linguistische Arbeiten, 41.) Tübingen: Niemeyer , 1976. Pp. x, 243. Reviewed by Robert I. Binnick, Scarborough College, University ofToronto When Calvert Watkins wished to criticize the 'typological' approach to IndoEuropean syntax exemplified in Friedrich 1975 and Lehmann 1974, he argued in part as follows (1976: 306-7): 'One wonders, if English or Chinese are clearly and unambiguously SVO, Japanese or Turkish SOV, and Modern Irish or Tongan VSO, why is it so hard to determine what Proto-Indo-European was, or more importantly, Homeric or Classical Greek? Could it be that it means something different to make such a claim about some languages than about others?' Ironically, the underlying word-order type of neither English nor Chinese is unambiguous , even leaving aside McCawley's 1970 proposal that English is VSO.1 Furthermore, German, one of the most thoroughly studied of languages, has never definitively been assigned an underlying word order. If it is true, as Watkins notes (305), that, 'Syntax is now viewed as coterminous with word order, and word order is reduced to the relative ordering of dyads', and that 'Over the past decade one of the most significant advances in the understanding of language has had to do with word order' (cover blurb for Li 1975), one might wonder (in the style of Watkins) why it is that, since Greenberg 1963, it has not been possible to determine what the underlying order in German is—scholars having variously argued for the traditional SVO, for SOV, and even for VSO. The determination of the underlying word order in German is one of the questions (as the title indicates) considered in this revision of Kohrt's 1974 dissertation .2 Following an introductory Chapter 1 (pp. 1-39), which gives K's full generative transformational grammatical apparatus (and which, despite his disclaimer , serves as an excellent if condensed introduction to GTG), Chapter 2 ('Underlying and derived verb position in German', 40-1 17) reviews the numerous arguments which have been brought forward in regard to the word-order question. German exhibits three different word orders on the surface—verb-first in questions (ex. 1), verb-second in main clauses (2), and verb-last in subordinate clauses (3): (1)Klaut Kunibert Kohlen? 'Does Kunibert steal coal?' (2)Kunibert klaut Kohlen 'Kunibert steals coal.' (3)[Jeder weiss,] dass Kunibert Kohlen klaut ' [Everyone knows] that Kunibert steals coal.' Whichever order is considered basic, the other two must be derived from it by a verb-shift rule; and all three orders have had their champions. Some arguments have been silly: since the average sentence has more subordinate than main clauses, to assume underlying SVO order would require more applications of a verb-last shift than underlying SOV would require of 1 Li & Thompson 1975 have argued that Chinese is undergoing a shift from SVO to SOV; but Tai 1 973, 1 976 has argued that it already is SOV. 2 K has thoroughly revised the dissertation, bringing his review of the literature up through 1976. The 300-odd items cited in the bibliography constitute a virtual thesaurus of GTG syntax. Pleasantly, there are no typos in the book, and only one erratum: the penultimate paper listed on p. 226 emanated from the University of Chicago, not UCLA. REVIEWS683 verb-second shift; hence the underlying order must be assumed to be SOV. Others have been sophisticated : an underlying SVO order would demand a treatment of prefix placement which violates Chomsky's A-over-A principle. All these arguments, however, are inconclusive, being based either on insecure formulations of particular rules or on universal principles of dubious relevance and/or universality, or else fraught with counter-examples. K concludes (78) that 'not one of these arguments is fully free from doubt'; but on the whole he favors, on the purely German evidence, the SOV solution. Still, the arguments from German must be balanced against universal considerations. It has been proposed (K quotes from Emmon Bach on p. 224) that 'the order of elements in the deep structures will be the same for every language (by definition)'. Given that, 'as...