|
|
Children around 3;0 have been observed to go through a stage of overgeneralization—producing '*breaked' instead of 'broke'. Does this phenomenon reflect a change in children’s generalizations about the past tense morpheme '–ed'? Given children’s early grammatical abilities in receptive language, it is possible that children generalize about verbs and the morphemes that attach to them before production of overregularizations.
Across three experiments using the Headturn Preference procedure, English-acquiring 16-month-olds preferred previously unheard forms like '*breaked' to either 1) their correct counterparts (e.g. 'broke'), 2) nonce verbs + '–ed' (e.g. 'bruck–ed'), or 3) English nouns + '–ed' (e.g. 'orange–ed'), suggesting relatively early tacit knowledge that '–ed' only applies to English verbs, months before overregularization in production appears.
The data here is also consistent with many observations that production data have a different developmental trajectory than receptive language, and bolster more scant data supporting early segregation of input into (proto-)lexical categories such as noun and verb.
This work is in part funded by NSF DDRI 1729862.
|
|