ERROR ANALYSIS - ENGLISH AS L2 AND GERMAN AS L3

Most analyses of linguistic errors set up categories of grammatical, lexical and semantic deviations. Such types of unidimensional classification are far too undifferentiated, particularly for didactic purposes. A two-dimensional classificatory system is described here, which formed the basis of an investigation into phenomena of foreign-language interaction. This system proves much more informative than the unidimensional model. Thirteen classes of linguistic deviation emerged in the course of the investigation. These classes are set up on the basis of the criteria of location and extent of the errors and of their specific type. The criterion of location refers to the particular hierarchical level of language on which the error occurs, these levels being those of the grapheme, moneme, syntagma, sentence and text. The resulting fivefold grid appears on the vertical axis of a matrix. The error type is established by assigning the error to a particular semiotic relation. The threefold grid is represented on the horizontal axis of the matrix. The twofold classification permits the assignment of any error to a particular point in the matrix; this makes for a more differentiated and effective treatment of deviations than is possible with the usual systems of error classification.