Distributionelle Phonemähnlichkeit

Abstract

Abstract The author of this paper develops a measure for distributional similarity between phonemes, which can be applied to dichotomous data. The phonemes of a language are classified on the basis of their distributional similarity. The classification is interpreted by means of a classificatory key in terms of distinctive features. The author develops a measure in order to evaluate the degree of concordance between classification and classificatory key. Finally, the author shows how the feature of distributional similarity between phonemes can be adopted for the comparison of languages. All devices are demonstrated by use of material drawn from the Roumanian language.