One of the first problems confronting infant language learners is word segm
entation: discovering the boundaries between words. Prior research suggests
that 8-month-old infants can detect the statistical patterns that serve as
a cue to word boundaries. However, the representational structure of the o
utput of this learning process is unknown. This research assessed the exten
t to which statistical learning generates novel word-like units, rather tha
n probabilistically-related strings of sounds. Eight-month-old infants were
familiarized with a continuous stream of nonsense words with no acoustic c
ues to word boundaries. A post-familiarization test compared the infants' r
esponses to words versus part-words (sequences spanning a word boundary) em
bedded either in simple English contexts familiar to the infants (e.g. "I l
ike my tibudo"), or in matched nonsense frames (e.g. "zy fike ny tibudo").
Listening preferences were affected by the context (English versus nonsense
) in which the items from the familiarization phase were embedded during te
sting. A second experiment confirmed that infants can discriminate the simp
le English contexts and the matched nonsense frames used in Experiment 1. T
he third experiment replicated the results of Experiment 1 by contrasting t
he English test frames with non-linguistic frames generated from tone seque
nces. The results support the hypothesis that statistical learning mechanis
ms generate word-like units with some status relative to the native languag
e. (C) 2001 Elsevier Science B.V. All rights reserved.