ABSTRACT
The design of vocabulary tests has traditionally been governed by frequency of word occurrence. In multilingual contexts, however, this one-directional approach appears far too simplistic. This chapter proposes a view in which frequency as well as cross-linguistic similarities provide a more detailed understanding of vocabulary acquisition and testing. Specifically, the chapter reviews cognates from a linguistic, cognitive, and testing perspective, attempting to account for learners’ background language influences beyond the first language. Such influences are especially the case for typologically proximate languages. The importance of this consideration is additionally supported by two cognate-controlled studies that are designed to explore the cognate facilitation effect. The studies show that a systematic understanding of the proportion and distribution of cognates in so-called real language should be adopted as a test-design criterion. By way of a response, the current chapter presents a cognate list of Romanian, Spanish, and Hungarian in the 10,000 most frequent words of English. Given that such a list allows for quantifying cognates, false cognates, and non-cognates in these languages, popular vocabulary tests are reassessed in terms of their “cognacy”. Aside from providing implications for vocabulary testing, the cognate list informs the teaching of English to speakers of Romance languages.
