A historical syntax of late Middle Indo-Aryan (Apabhraṃśa) By Vît Bubeník (review)

BOOK NOTICES 229 heads. One place where L's account differs from Cinque's well-known proposals is in admitting of right-hand specifiers to account for facts of adverbial scope. The main thrust of Ch. 3 is a proposal to recast the familiar three-way distinction among strong pronouns , weak pronouns, and clitics into a two-way grouping between strong-pronouns and 'clitics', with the latter subdivided into syntactic clitics and LF clitics (formerly weak pronouns), whose intermediate nature arises from needing to be 'close enough' to their host overtly that they may undergo covert incorporation . Ch. 4 (on clause structure) deals almost exclusively with German, examining a range oftopics having to do with word order phenomena. Specifically, L deals with scrambling (its status as A or A' movement and its interaction with specificity and focus), adverb positions, the positions of pronouns, and the V2 phenomenon. [Jonathan David Bobaliik, McGiIl University.] Word order in Arabic. By Sven-Olaf Dahlgren. (Orientalia Gothoburgensia 12.) Göteborg, Sweden: Acta Universitatis Gothoburgensis, 1998. Pp. 273. Dahlgren presents a mostly statistical study of word order variation in Arabic, with a focus on the relative order of subject and verb in different kinds of texts. The study is noteworthy in examining texts from different modern dialects of Arabic (leaving out North Africa, but including Bedouin dialects) and one from Classical Arabic as well. Arabic has posed something of a problem for research on word order, as Ch. 1 points out. It has often been stated that Classical Arabic was VSO while most of the modern dialects have become SVO. But the picture is not so simple. Many of the dialects seem to exchange VS for SV order quite freely, and others use VS more often than SV. No one has yet examined the data in an attempt to find a functional distinction between the two orders, or even to do statistical counts. The aim of D's study is to do just that. The first eight chapters present background to the study. Chs. 3-6, for instance, cover information structure, discourse analysis, and the 'functional text perspective' . Ch. 6 is a discussion of 'marked' contexts , where the most basic type of sentence (following T. Givón) is declared to be 'foregrounded, main, declarative, affirmative, active and continuative clauses' (93). Other contexts are marked and less likely to involve the 'basic' word order. While this may be true, it is also striking that D decides to leave subordinate clauses out of his study altogether. We know from Germanic languages, for instance, that basic word order may be revealed in subordinate clauses but not in 'main, declarative, affirmative' clauses. Ch. 9 finally launches into the actual investigation reported in the book, laying out methodology and the texts studied. The dialects under consideration are those spoken in the Eastern Meditenanean (Lebanon, Syria, Palestine, and Jordan), Anatolia, and Egypt; Qeltu and Gilit (both in Iraq), Bedouin (Arabian Peninsula ; Transjordan, Syria, and Palestine; Sinai; and Syrian Desert), and one text from Early Arabic (Ibn Ishaq's Life ofthe prophet). Chs. 10-12 provide the results of the study for the different categories of text (narrative discourse, dialogue and description, and Early Arabic). D decides in Ch. 9, somewhat peremptorily, that the foreground-background distinction is what distinguishes word order types in Arabic. He notes a strong conelation between VS order and the characteristics of foregrounding (narrative sequence, perfective aspect , past tense), and the appearance of SV order and backgrounding (imperfective aspect, present tense, events out of sequence). He then proceeds to examine the texts only with an eye to this distinction. But nowhere does he argue that this is in fact the relevant distinction rather than, say, a grammatical one where VS order is preferred with past tense (suffixal agreement , in Arabic) and SV with nonpast (préfixai agreement). Seeing that one of the few hallmarks of the foreground-background distinction is exactly the grammatical one of tense/aspect, a grammatical underpinning is just as likely. D's study does provide statistical counts on word order variation across Arabic dialects and examines the variation from a functional perspective, something no other study has done. It also gives statistics for definite...