In four experiments a large temporal effect is demonstrated for alternating between different response modes of musical output: sing—whistle, sing–play. This effect, demonstrated for college students with some musical training, has theoretical implications for the setting of response mode parameters. Over the series of four studies, it is shown that the time to alternate between musical output modes is not due to peripheral competition for the components of the same motor system, or to central competition produced by concurrently trying to use the same central resources for several things at once. Syntactic disruption explains some of the alternation effect, but the most complete explanation is attributable to an output parameter setting mechanism in which different response modes may be set to ON/OFF. Both evidence and argument suggest that these parameters are set globally, and it takes time to change their values (about 200 msec per switch).