This paper is concerned with an area of second language learning which has long been regarded as idiomatic and therefore not amenable to systematization.' Specifically, we have in mind the lexical combinability of words, i.e., the tendency of one word to occur with another word or words to convey a particular meaning. This area, which involves questions of style and word-usage, is an extremely difficult aspect of language learning; it requires not only knowledge of the dictionary definitions of vocabulary items, but also information about their combinatory properties. One of the most effective methods of dealing with this problem derives from a principle first developed within the context of machine translation theory.2 This principle has come to be known in linguistic literature as the theory of lexical functions. The present paper will demonstrate the applicability of lexical functions to the problem of word-collocations in the acquisition of German as a second language. The term lexical function refers to the regular semantic relations which exist between certain key words and other words (or wordcombinations) in a language. According to this theory, words are not viewed in isolation, but rather as lexical correlates of other words. The formalism used to express a lexical function consists of three elements: the name of the function, the key word (in parentheses), and the word or word-combination value of the function. An example of a lexical function is the semantic relation between a (key) word denoting a certain phenomenon and the designation "to a very high degree" (of the phenomenon). This function has been called "Magn" (from Latin magnus) and is illustrated as follows: