Problems of seem/scheinen constructions, and their implications for the theory of predicate sentential complementation By Susan Olsen (review)

988 LANGUAGE, VOLUME 60, NUMBER 4 (1984) improvement on, say, Daniel Jones' Outline of English phonetics (1972), which is certainly the definitive treatment of RP. However, Jassem provides a large number of additional chapters with occasionally useful reference material: a history of the individual phonemes from late Middle English, an extensive discussion ofmodern acoustic phonetics, a long chapter on generative phonology, and a very long chapter on phoneme/grapheme correspondences (283419 ). There is also an extensive list of derivational suffixes, with their effects on stress patterns . J's discussion of generative phonology relies heavily on Chomsky & Halle's Sound pattern ofEnglish (1968), Harms' Introduction to phonological theory (1968), and Schane's French phonology and morphology (1968); he consequently presents GPh (as he calls it) at its most baroque. His criticisms of it stem from the work of B. Derwing, P. Lined, and R. Botha, and deal with such issues as falsifiability: mentalistic theories are not scientific since they are not falsifiable ; the proposed innate structure oflanguage is 'speculative' because no experimental technique exists for testing such a claim. J's arguments are not new; and even if one believes them, they are better stated in the original sources. J's phonemicization ofEnglish uses a TragerSmith flavored analysis, with the vowel nuclei /i e e a a o 3 u 3 9 ij ej aj oj aw/. Aside from the problems ofpsychological reality here, there are also peculiar choices in symbols; e.g., why, use Id to represent the nucleus in bad, sail Neither e nor a? occurs on most typewriters, but surely * is more suggestive ofthe sound involved. Similarly , why /3w/ instead of Trager-Smith's low/ or Jones' /ou/? The book is a full-size photographic reproduction of a typescript, and this creates its own problems. Typographical devices for signaling emphasis include single and double underlining, which is fine. However, citations are signaled by long spacing, which is troublesome to read. In all, though J offers much interesting reference material, this is not likely to be a useful book in a university classroom. [Geoffrey S. Nathan, Southern Illinois University.] Problems of seemlscheinen constructions , and their implications for the theory of predicate sentential complementation . By Susan Olsen. (Linguistische Arbeiten, 96.) Tübingen : Niemeyer, 1981. Pp. viii, 200. This revision of a 1980 dissertation from the University of Cologne is a detailed, methodical, and orthodox treatment of sentence complementation in English and German. The theoretical framework is that of Revised Extended Standard Theory as of 1977. The main issues discussed are the sentential complements of the seem class in English and the scheinen class in German, which appear to have similar exceptional properties. In particular, the verbs of these classes do not allow complement clauses in subject position: That S is likely but not *That S seems. Nevertheless, these verbs (like others) allow sentential complements in postverbal position . Further, the complements of the seemlscheinen type do not function syntactically like subjects, though the complements of verbs ofothertypes ratherclearly have the grammatical roles of subject or object. Olsen's solution uses lexical subcategorizations of a fairly complex nature for the verbs which take sentential complements. The syntactic differences among verbs are represented as different patterns of subcategorization for S' complements, which lack grammatical roles in most cases, and null NP's pro, which do have grammatical roles. The subcategorization information also specifies which S' complements are co-indexed with null NP's having grammatical roles. Seem is represented as having a specification ofa clause complement, but lacking co-indexing for an element with a subject role. In one sense, the problem of seemlscheinen is theory-independent, and no obvious solution results from any particular currently known theory . O shows in great detail just how accounts of all different eras and theoretical persuasions fail to make all the descriptive distinctions required by the syntactic facts, both in English and German. She argues very persuasively that the complements of seemlscheinen are not subjects , though one would expect them to have the same properties as other verbs which permit raising to subject position. The lack of subject properties and the failure to occur as preverbal subjects are exceptional properties of complement clauses, by any account; and...