Perceiving non-native consonants that utilize articulators not employed contrastively in the native language: Revisiting PAM and the Articulatory Organ Hypothesis

We perceive non-native speech in terms of similarities to our native
phonology, which makes many non-native contrasts diffi cult to discriminate
(e.g., Speech Learning Model [SLM]). However, discrimination is poor
mainly when contrasting non-native consonants are both mediocre exemplars
of the same native consonant. Discrimination is much better if they are
similar to different native consonants, and good if they are nativelike versus
deviant exemplars of the same native consonant (Perceptual Assimilation
Model [PAM]). The Articulatory Organ Hypothesis (AOH) offers
orthogonal predictions that con sonants produced by different articulators
should be discriminated better than consonants using the same articulator.
To compare these models, we tested Italian listeners on non-native English
and Nuu-Chah-Nulth fricative contrasts differing in perceptual assimilation,
articulatory organs, and articulator use in Italian. Results support PAM and
pose challenges for AOH and SLM.