Codeswitching by African language speakers in South Africa (whether speaking English or the first language) has been extensively commented on and researched. Many studies analyse the historical, political and sociolinguistic factors behind this growing phenomenon, but there appears to be a little urgency about establishing a database of new lexicons to inform first-language interventions in education and public information programmes. In this article I use research conducted in three townships in Cape Town to focus on one aspect of codeswitching, namely lexical borrowing, in one of South Africa's languages, Xhosa. The central argument tested in this article is that lexical borrowing in Xhosa is driven by a need to simplify expression by using syllabically shorter English words. I test this hypothesis by (a) establishing the average syllabic difference between Xhosa and English; (b) seeking lexical responses to a set of pictures; (c) analysing lexical borrowing in the main Xhosa radio station. In addition, the article uses the data collected to elucidate the link between the Xhosa speakers' lexicon and the concepts it needs to describe, and then suggests further possible reasons for the choice of English over Xhosa. (Contains 23 notes and 7 tables.)