An exploration of the accentuation effect: errors in memory for voice fundamental frequency (F0) and speech rate

The accentuation effect demonstrates how memory often reflects category typical representations rather than the specific features of learned items. The present study investigated the impact of manipulating fundamental frequency (F0) and speech rate (syllables per second) on immediate target matching performance (selecting a voice from a pair to match a previously heard target voice) for a range of synthesised voices. It was predicted that when participants were presented with high or low frequency target voices, voices even higher or lower in frequency would be selected. The same pattern was also predicted for speech rate. Inconsistent with the accentuation account, the results showed a general bias to select voices higher in frequency for high, moderate, and low frequency target voices. For speech rate, listeners selected voices faster in rate for slow rate target voices. Overall it seems doubtful that listeners rely solely on categorical information about voices during recognition.