ABSTRACT
Jakobson (1941/1968) proposed that markedness predicts the types of phonemes and word structures that appear in children’s early phonologies; that is, they contain the least marked phonemes/word structures crosslinguistically. Other phonologists following in the generative grammar tradition have further developed this learnability hypothesis: that a child will produce structures with unmarked values until positive linguistic evidence triggers the setting of the marked value for a grammatical parameter (White, 1989). Research in the past five decades, although somewhat encouraging initially (e.g., Cairns & Williams, 1972; Leopold, 1947; Menyuk, 1968), has failed to confirm this prediction (e.g., Ferguson & Farwell, 1975; Macken, 1980; Pye, Ingram & List, 1987; Stoel-Gammon, 1985). Both individual variation and lack of predictive clarity weaken the markedness hypothesis substantially as an explanation for patterns seen in early phonology.
