Tekstdekking, Tekstbegrip En Woordselectie Voor Het Vreemde-Talenonderwijs

Abstract

This article deals with an alternative approach to the problem of the selection of words in foreign language teaching at higher levels. Since frequency as a selection criterion is inadequate beyond the first 2.500 words, it is argued that the question as to how many and which words can only be answered on the basis of quantitative data on the relationship between lexical coverage and text comprehen-sion. Only after establishing what coverage goes with what level of comprehension will it be possible to come forward with valid suggestions concerning the number of words required for reading authentic L2 texts. As to which words, it is argued that these should be selected from a corpus of 10.000 words (beyond the first 2.500) to be constructed on the basis of a combination of selection criteria such as frequency, valency etc. Knowledge of any 5.000 words from this corpus (combined with the first 2.500) will yield such a dense coverage (at least 95%) of general L2 texts that the meaning of any remaining unknown words can be deduced via contextual clues utilisation. To substantiate this claim the results of some experimental investigations of the relationship between coverage and text comprehension are reported.