Adult learners’ use of lexical cues in the acquisition of L2 allophones

Adult second language (L2) learners gain knowledge of L2 allophones with experience. However, it is not well understood how this knowledge is acquired. We investigated whether naïve subjects use lexical cues in the form of visual referents to acquire L2 allophones. We exposed native English speakers to one of two artificial languages where novel words containing two acoustically similar sounds, [b] and [β], occurred in an overlapping distribution. The words were paired with an image that either did or did not reinforce the contrast (i.e., [bati]—"apple" and [βati]—"penguin" (DiffImage group) or [bati]—"penguin" and [βati]—"penguin" (SameImage group)). Participants completed three tests to determine whether the exposure phase impacted their ability to perceive and lexically encode the [b]-[β] contrast in trained and untrained words. If subjects use lexical cues in the form of visual referents to infer the phonological status of [b] and [β], then participants in the DiffImage group, but not the SameImage group, should discriminate and lexically encode the distinction. Data from 40 participants suggest that lexical cues may impact participants’ lexical encoding, but not perceptual sensitivity, to the [b]-[β] contrast. We discuss our findings in relation to proposed learning mechanisms.