What Are Phonetic C

This paper reviews some of what is known about the acquisition and mental representation of phonetic categories. They in part arise from the language learner’s experience, but not entirely. They are, moreover, governed by functional principles, which must nonetheless be instantiated in the mind and grammar. 1 AN EXAMPLE TO START In many languages, tones spread from earlier to later in a word, but they rarely, if ever, spread from later to earlier in a word [1]. Why should tone spreading be perseverativebut not anticipatory? One possible answer is suggested by another crosslinguistically robust observation: speakers reach the F0 targets that realize tones late, often not until after the tone bearing unit that sponsors the tone [2,3 & many others]. When the F0 target is reached after the sponsoring tone bearing unit, listeners may mistakenly interpret the next tone bearing unit as the tone’s sponsor. Then when they take their turns as speakers, they will produce it one tone bearing unit later than the original speakers did. As they will also reach tone’s F0 targets late, other listeners may mistakenly interpret yet the next tone bearing unit as the tone’s sponsor, and so on. These successive misinterpretations progressively displace the tone to later and later tone bearing units. Tones are displaced in this way to the next tone bearing unit in Yoruba [4], where an input high.low (H.L) sequence is pronounced high-falling (H.HML) and an input L.H sequence is pronounced L.LMH, and in Kikuyu [5], where each input tone is pronounced on the next tone bearing unit. Tones in Isixhosa [6] are displaced as far as possible, up to the antepenultimate tone bearing unit or to next tone-sponsoring unit. Yoruba and Isixhosa illustrate the minimal and maximal phonological consequences of mistaking which unit sponsors a tone. Displacement is not, however, spreading. If listeners have morphological evidence from alternations that the tone is sponsored by the first tone bearing unit in the sequence, then the lateness of F0 targets will lead them to interpret the tone’s domain as extending from the sponsoring unit to the following unit(s), and the tone will then spread to rather than being displaced to later units. Spreading, like displacement, propogates because when listeners take their turns as speakers they also reach F0 targets late. Because tones’ F0 targets aren’t reached earlier than the sponsoring tone bearing unit, a listener has no occasion to mistake the preceding tone bearing unit for the tone’s spons spread