On the Origins and Scope of Final Lowering

This experiment examines final lowering, the lower than expected scaling of the last pitch accent in downstepping word lists. In word lists of four or five items two parameters were manipulated. First the number of unaccented syllables between accents was either constantly two (‘‘LIma beans, GREEN beans and SOY beans’’); alternatively, the presence of ‘‘and’’ resulted in an extra syllable between the penultimate and final accent (‘‘LIma beans, NAvy beans and SOY beans’’). Second, the number of unaccented syllables after the final accent was either one (‘‘SOY beans’’) or three (‘‘HAricot beans’’). The results showed evidence of final lowering, in that the final accent in four word lists was lower than the fourth accent in five word lists, but the presence of ‘‘and’’ did not produce extra lowering of the final peak, suggesting that final lowering is an independent phenomenon, not an epiphenomenon of declination (contra-Grabe). On the other hand, the final accent was scaled higher when the number of unaccented syllables following it was larger. This suggests that final lowering is not under phonological control but a physiological phenomenon, since it does not target the last accent but a fixed time interval before the end of the utterance.