Complex behavior requires the control of various bodily effector systems (e.g., manual, oculomotor, vocal), both in isolation (i.e., single actions) or together (e.g., dual actions). Dual-task research has extensively documented performance costs when combining actions, making the implicit assumption that executing a single action is straightforward. We challenge this assumption by investigating scenarios in which multiple effector systems principally play a role but sometimes only one should be activated.

Across multiple experiments participants switched unpredictably between executing a keypress only, a saccade-only, and both actions in response to left/right visual stimuli. The main findings comprised minimal or even absent dual-action RT costs and robust co-execution errors: prepared-but-irrelevant action systems spontaneously intruded when focusing on single responses. These errors showed a striking asymmetry: saccades intruded far more frequently in keypress-only trials than vice versa, revealing effector-specific differences in action prepotency and inhibitory control demands.

We identified several factors modulating this effect. First, effector intrusions increased after any trial involving that effector (regardless of direction), suggesting effector-system-level activation is carried forward across trials. Second, prolonging preparatory intervals between effector cues and directional stimuli decreased errors and speeded required responses, indicating proactive regulation of effector-system readiness. Third, errors diminished substantially with increasing task practice, alongside the emergence of typical coordination costs suggesting the segregation of action representations over time.

Together, these findings reveal that single actions do not always come easy. When multiple action systems are prepared to potentially work together, using just one system alone may become surprisingly difficult.