Pii: S1053-8100(02)00022-3

We have only limited awareness of the system by which we control our actions and this limited awareness does not seem to be concerned with the control of action. Awareness of choosing one action rather than another comes after the choice has been made, while awareness of initiating an action occurs before the movement has begun. These temporal differences bind together in consciousness the intention to act and the consequences of the action. This creates our sense of agency. Activity in the anterior cingulate cortex and medial prefrontal cortex is associated with awareness of our own actions and also occurs when we think about the actions of others. I propose that the mechanism underlying awareness of how our own intentions lead to actions can also be used to represent the intentions that underlie the actions of others. This common system enables us to communicate mental states and thereby share our experiences. 2002 Elsevier Science (USA). All rights reserved. 1. Awareness of action Will, the sense of being in control of our own actions is a major component of consciousness (along with emotion and cognition). But are we aware of all aspects of our own actions? Milner and Goodale (1995) have intensively studied a patient, known as DF, who demonstrates a striking lack of awareness of certain aspects of her own action. As a result of damage to her inferior temporal lobe DF suffers from Consciousness and Cognition 11 (2002) 481–487 www.academicpress.com Consciousness and Cognition * Fax: +44-207-813-1420. E-mail address: [email protected]. 1053-8100/02/$ see front matter 2002 Elsevier Science (USA). All rights reserved. PII: S1053-8100(02)00022-3 form agnosia. In other words, she is unable to perceive the shapes of things. She cannot distinguish a square from an oblong and cannot report the orientation of a slot. On the other hand, when she picks up a square or puts her hand through a slot she orients her hand and forms her grasp appropriately. DF is able to use information about the shapes of objects to guide reaching and grasping movements, but she is not aware of this information. This unconscious guidance system is not unique to patients. It is simply revealed more starkly in the case of DF by the damage to the system that normally brings certain kinds of information into consciousness. A number of studies have shown that we all can make rapid and accurate grasping movements without being aware of the information that is being used to control these movements (Pisella et al., 2000). In some cases, we are not even aware of having made the movement. This unconscious and rapid system for controlling movements is particularly involved in making corrections to actions that are already underway. These are automatic responses driven by our environment rather than by our will. By definition, we must be aware of our willed actions and yet there is evidence that this awareness is too late for any involvement in the selection of the action. In Libet s much discussed experiment (Libet, Gleason, Wright, & Pearl, 1983) subjects made the willed action of lifting a finger, whenever they felt the urge to so so. The time at which they felt the urge to make the action occurred some 100s of milliseconds after the first detectable changes in brain activity (the beginning of the EEG readiness potential). This result has been replicated by Haggard, Newman, and Magno (1999) who have also extended the finding to show that the time of awareness of the urge to act is correlated with the time of onset of the lateralised component of the readiness potential (Haggard & Eimer, 1999). This is the time at which the brain activity ceases to be centrally located and shifts towards the side of the brain that will control the movement, contra-lateral to the hand that will move. This observation suggests that we are not aware of the movement we have selected until that movement has been precisely specified. If awareness comes after movement selection then presumably this awareness has no role in making that selection. This principle of awareness coming after the event does not just to apply to the very simple responses studied in these experiments, but also to the selection of complex actions (Bargh & Chartrand, 1999; Velmans, 1991). Indeed it has been argued that we are never conscious of cognitive processes, but only of the results of those processes (Nisbett &Wilson, 1977). If consciousness has no role to play in the short-term control of our actions, then what role does it play? Why does our subjective experience so strongly suggest that we are in control of our actions? Libet also measured the time at which his subjects were aware of initiating their actions. This is a distinct and later event from the time of the first awareness of the urge to act. In contrast to the awareness of the urge, which is late, the awareness of the initiation to act is early. The time at which we think we initiate the act occurs some 80ms before the limb actually starts to move. This result has also been replicated by Haggard et al. (1999). Note that any sensory feedback that results from the movement of a limb will occur a further 100ms or so after the movement has actually started. Thus, the awareness of initiating the movement is much too early to be based on any of the sensations associated with the movement. What we are aware of 482 C. Frith / Consciousness and Cognition 11 (2002) 481–487