Differentielle Linguistik. Entwurf eines Modells zur Messung semantischer und pragmatischer Variation und dessen Anwendung zur Überprüfung der Hypothesen Bernsteins

Abstract

The doctoral dissertation (Regensburg, 1975) which started with empirical tests and observations in a school class (6th grade, Schöllkrippen, Bavaria) in 1969/70 develops a method for the analysis and measurement of code-differences in the sense of Basil Bernstein's hypothesis (distinguished as "restricted and elaborated code"). This hypothesis states that different social classes use more implicit or explicit styles of expression. The key features of such differences concern a measure of semantic information, i.e. how much information in a situation is explicitly uttered by the speaker and how much is left implicit and has to be completed by the hearer. In this work the differences are called differences in "communicative style" (not "code"). In order to measure these differences, a procedure is developed which compares the linguistic reactions to comparable stimuli - in the case of this empirical study the oral and written retellings of two stimulus stories (one simple and one more complex). The lexical and grammatical variations are first classified using a tool derived from Harris' transformational analysis (this avoids the presupposition of a given grammar of German or some socially established norm). The classified variants are then measured applying the model of semantic information proposed by Carnap and Bar Hillel. In order to achieve a detailed analysis, meanings are analyzed using a revised and adapted Montague grammar. As pragmatic constraints and textual structures seemed to be relevant, the measurement is expanded to clause and sentence connectors and to narrative structures. For this purpose the model of narrative analysis proposed by Labov and Waletzky (1967) is further specified and systematically applied to the corpus of retold stories (3 x 26 stories). In the quantitative analysis three levels of communicative style/code are considered: the intra-clausal (subdivided into: nominal/verbal kernel variation and nominal/verbal adjunct variation), the inter-clausal level (conjunctions and embeddings of clauses) and the narrative level (ratio of narrative versus evaluative clauses).The statistical analysis which compares measured (non linguistic) intelligence, social status and semantic information (completed by narrative strategies) could not verify Bernstein global hypothesis (for this social group and for German speakers). It rather showed tendencies of children with different social background to use different narrative strategies and to elaborate their performances differently at the adjunct level (mainly if the stimulus was simpler and in the oral retellings). If the situation asked rather for an explicit style (as in the written retelling of a complex story), the differences vanished. Thus social codes (or communicative styles as we preferred to say) vary with situations and their demands. They cannot be considered as stable "codes" like linguistic or other social systems. Therefore the global hypothesis of Bernstein must be revised (for German classroom communication and the age group considered in this study).The concept of "Differential Linguistics" (parallel to "Differential Psychology") developed in the second chapter could be generalized as a basic tool for Applied Linguistics and a complement to the search for universal features of language, which dominated the Chomskyan paradigm.