Access to talker-specific representations is dependent on word frequency … See original article here: Access to talker-specific representations is dependent on word frequency