Does speech output control lateralize over time? Evidence from verbal-manual time-sharing tasks