Seldom does one appreciate that one is often unaware of what word one will say next. It seems that, while speaking, one is aware of what one has just uttered, and that one has a general inkling regarding what one intends to say, but that one is often unaware of the exact words that will be selected by our brains when trying to convey some information.
This is most obvious with respect to speech errors, as when one says "please close the window" when one intended to say "please close the door." (Such an error, in which the unintended word is semantically-related to the intended word [a "semantic error"] is the most common kind of speech error.) In many cases, one is not only unaware of the words that happen to be selected for speech production, but one is also unaware of how the vocal apparatus-- including the voice box, lips, and tongue--produce the sounds that they produce. When one speaks, one is unconscious of the motor codes telling the lips, jaw, and mouth to move the way that they do. These things are so unconscious that it is often only from reading textbooks about linguistics that one realizes that, regarding what is going on in the mouth, /b/ and /p/ are articulated in the same manner (both are bilabial stops), and so are /d/ and /t/, and /g/ and /k/.
This limited awareness of the ongoings of action programs is not restricted to just speech. The brain's motor-programs, those programs which tell the muscles what to do when, are all largely unconscious. These processes (telling which muscle fibers to contract for a certain amount of time) are far from dumb or inflexible. The computations involved in unconscious motor control are often more complicated than the conscious algorithms we use to solve puzzles. The great motor researcher David Rosenbaum stated that, though one could train a computer to play chess and beat humans at the game, as in the case of IBM's Deep Blue, one still needs a human to move the chess pieces for the computer during a human-cyber match-up. This is because motor control is far more complicated than the algorithms used to win a game of chess, even though it is largely unconscious.
It has been proposed by the William James, the father of American psychology, that action guidance and action knowledge are limited to perceptual-like 'representations' of action outcomes (e.g., the ‘image' of one's finger flexing), with the motor programs/events actually responsible for enacting the actions being unconscious. (The drawing above of the pain withdrawal reflex is by James.)
Recent evidence for such a dissociation between 'conscious action knowledge' and 'unconscious action processing' comes from studies involving brain stimulation. As mentioned in my previous blog, Wilder Penfield (1891-1976) pioneered a technique for the treatment of severe epilepsy which required damaging the brain areas responsible for the onset of the seizures. One obvious concern in carrying out this procedure is that the surgeon may be damaging a brain area that is critical for the welfare of the patient. Penfield devised a technique to assess whether the area to be damaged was critical for brain function. While the patient is awake (there are no pain receptors in the brain, so neurosurgery can be carried out painlessly while patients are conscious), Penfield would mildly stimulate the targeted brain area with an electrode and would note the effects of the stimulation. With this technique, which continues to be used today, the surgeon can evaluate whether the electrical stimulation leads to something notable (e.g., a visual hallucination, auditory hallucination, or movement of a finger) or whether the stimulation disrupts function (e.g., the patient can no longer utter a word). Either kind of effect suggests that the region being stimulated should not be damaged.
Regarding action, Penfield noticed that activation of some brain regions led to overt actions (e.g., a finger moving) of which the patient was unaware. This is consistent with recent pioneering experiments by Sirigu, Desmurget, and colleagues. Striking findings from their laboratory reveal that direct electrical stimulation of motor areas of the brain (e.g., in premotor areas) can lead to an actual action, but the patient believes that he or she did not perform any action. Conversely, activation of parietal areas of the brain (which are historically associated with high-level perceptual processing) makes the patient believe that he or she executed an action even though no action was performed. Interestingly, these findings are consistent with James's ideomotor theory: Awareness of our actions tends to occur only in a perceptual-like, ‘Sensorium' (a term used by the great nineteenth-century physiologist Johannes Müller) with motor processes, belonging to the world of 'The Motorium,' being largely unconscious.