Testing the bootstrapping hypothesis of infant-directed vocabulary: A longitudinal individual-difference analysis

Infant-directed speech contains a substantial number of lexical items characterized by sound-symbolism (e.g., moo, choo-choo), full/partial reduplication (e.g., night-night, daddy), and diminutives (e.g., doggy, blankie). It has been proposed that such register-specific words are easily-acquired and facilitative of further vocabulary acquisition because of their nonarbitrary sound-meaning mappings, phonological repetition, and edge invariance. If this is true, we expect initial vocabulary growth to be boosted in infants whose lexical input has a higher incidence of such characteristics. To test this prediction, we examined speech samples addressed to 47 English-exposed infants at 9 months, and calculated the proportions of lexical input featuring sound-symbolism, reduplication, or diminutives. Mixed-effects models showed that sound-symbolism was not related to later vocabulary growth of the infants, but the proportions of reduplication and diminutives contributed significantly to variance in infants’ vocabulary growth from 9 to 21 months as measured by CDI data.