Central ideas of Darwin’s theory of natural selection figure prominently in the work of Otto Jespersen (1860–1943). As early as 1886, Jespersen treated linguistic change in Darwinian terms: variation in the pronunciation and meaning of the various units, and factors that raise or lower a variety’s viability. His critique of the neogrammarian principle of exceptionlessness of sound change includes the point that phonologically parallel words often differ in the relative viability of their variants. By 1904, Jespersen was using ‘functional load’ in explaining differences in how much resistance each language offers to various natural phonetic tendencies. He argued that conformity to a sound-symbolic generalization raises a form’s viability and can thus exempt some words from sound changes and accelerate the adoption of novel words and of novel meanings for existing words. Natural selection figures even in Jespersen’s papers on international auxiliary languages, as in his account of why bil, the winner in a contest for a word for ‘automobile’, spread so rapidly in Scandinavia. Jespersen’s speculative scenario for language origins is in terms of Darwinian ‘preadaptation’: conventionalized sound/meaning correspondences can arise in numerous ways prior to the development of anything like a language (Jespersen argues that singing, in all its diverse social functions, developed prior to language), and a language would develop not ex nihilo but by members of a human community segmenting and imposing arbitrary semantic analyses on some of this large body of meaningful sound, and starting to combine the pieces in novel ways, as modern children do anyway (Jespersen argues) in the course of acquiring a language. Jespersen thereby vindicated his unpopular conclusion that early human languages had highly irregular morphology and syntax.