Devil or angel in the details?
Perceptual attunement to native speech begins early in life, becoming the foundation for efficient native word recognition, yet simultaneously constraining perception of non-native segmental contrasts. It is less well understood how these two sides of native listening handle natural phonetic variations. To recognize a given uttered token as a particular word, listeners must recognize its specific phonetic details as relevant either linguistically or indexically (e.g., talker identity, mood, accent). Perceivers cannot recognize varying tokens of a word by filtering or normalizing phonetic variation. Rather, they must exploit both types of variability to differentiate the words being said from who is saying them. This requires a grasp of two complementary principles: <i>phonological distinctiveness</i>, i.e., phonetic differences that are critical to lexical distinctions, and <i>phonological constancy</i>, which keeps word identity intact across lexically irrelevant variations. Perceptual attunement supports discovery of those principles, fostering word recognition and the ensuing acquisition of morphology, syntax and literacy.