Complexity in learning and typology

(3) An example of simplicity bias in learning Saffran and Thiessen’s (2003) phonotactic learning/word segmentation study: a. Training Phase 1. 9-month-olds exposed to isolated words of shape C1V C2.C1V C2, where C1 and C2 are each limited to a set of three consonants. b. Training Phase 2. Exposed to 4 new words in a continuous stream, with only two fitting the pattern from Phase 1. c. Test Phase. Listening time measured for each of the words from Training Phase 2, presented in isolation. d. Results 1. When the pattern was featurally simple, with voiceless [p, t, k] in one position, and voiced [b, d, g] in the other, the infants displayed a novelty preference for non-conforming items. e. Results 2. When the pattern was featurally complex, with [p, d, k] in one position, and [b, t, g] in the other, there was no significant difference in listening time between conforming and non-conforming items.