Perceptual learning in speech is phonetic, not phonological: Evidence from final consonant devoicing

Listeners flexibly recalibrate the perceptual categorization of sounds in response to speakers’ unusual pronunciation variants. Recent studies have shown that generalization of this recalibration can inform us about the nature of prelexical units used for speech perception. The present study tested whether this generalization is sensitive to phonetic or phonological properties of speech. Using the example of German word-finally devoiced stops, we found that generalization does not extend to sounds that match in phonological specification (here: voiced) when they are dissimilar in their acoustics (phonetically voiceless in word-final position to phonetically voiced word-medially). However, phonologically voiceless stops in word-final and word-medial position did show the effect. This supports suggestions that listeners extract segments of sufficient acoustic similarity from the input and use them for generalization of learning in speech perception. The units of perception thereby appear context-sensitive rather than abstract phonemes or phonological/articulatory features.