Taal en Tongval - Volume 61, Issue 1, 2009

Southern Brabant dialects of Dutch provide constructions such as: (a) ik heb iemand niet gezien <I have anybody not seen> (I did not see anybody) (b) Ik heb niemand niet gezien <I have nobody not seen> (I did not see nobody) The traditional view is that (a), with a positive quantifier expression developed from (b) with a negative one (cf. Van der Auwera & Neuckermans 2003). In this article the opposite development is proposed. Arguments are derived from earlier language stages in which the negation particle en/ne played an important role, and from the present distribution of both construction types, which are currently found in the central (Brabant) language varieties . In western (Flemish) and eastern (Limburg) dialects (b) is found, but is apparently borrowed from central dialects; (a) is not found at all. The article ends with the suggestion that the same development path might account for Brabant constructions with universal quantifiers, as well. In fact (c) is common in Brabant, and much less so (if at all) in the other southern Dutch dialects: (c) Ik heb al die mensen niet gezien <I have all those people not seen> (I did not see all those people)

Frisian nouns ending in schwa (henceforth: schwa nouns) like brêge (‘bridge’) show unpredictable behaviour when used as a first member of a nominal compound. In some cases the schwa is retained (e.g. brêgeman ‘bridge man’), in other cases the schwa can or must be dropped (e.g. bûsjild ‘pocket money’, from bûse ‘pocket’). Although there are no fixed rules two factors turn out to be crucial: the (non)optionality of the schwa in the simplex word (e.g. brêge ‘bridge’ vs mis(se) ‘church mass’) and the similarity to the Dutch synonymous counterpart (e.g. F. skoalle, D. school ‘school’). In a particular subset of the data the interaction of these two factors has a non-trivial and surprising effect on Frisian word formation.

This study discusses Versloot 2008, an important contribution to the study of language variation in older Frisian, based on charters and codices. Special attention is paid to (a) how theses charters and codices are used by Versloot in order to obtain linguistic results, and (b) to the statistical treatment of the data. It is argued that with respect to 2 x 2 tables the Fisher Exact test one-tail should be preferred to X2.

boekbesprekingen