Notions of free will were systematized very early on, for example, by Aristotle, and then incorporated into Christian theology and law by Thomas Aquinas (Keys, 2008). Aristotle held that virtue was the product of action and habit, and action was a result of free will (Sorabji, 2006). Pretty much all philosophers until the late twentieth century have agreed with this. We all experience the sense of free will and so the idea that free will exists appears to be confirmed by our everyday experience. This sense seems to result from the fact that before we act we have the sense that we are about to act, and we usually even anticipate the action we are about to perform. Because of this, we think that these prior thoughts have caused our action. As we will discuss, this is an illusion, but one that is very hard to overcome. Commons and Armstrong-Roche (1985) carried out an experiment on this topic. Participants were randomly presented with two kinds of stimuli. In the first condition, it was very easy for participants to tell which stimulusduration was long and which was short. The participants were asked, "How much of a sense of free will did you have?" And they said, "Oh, a great deal." In fact, it was the stimulus conditions being easy that made it easy for them to' make the right choice. The stimuli controlled their behavior. In the second condition, they were given two stimuli very close in duration, which was a hard choice. Following this condition, they reported having no experience of free will. In this latter case, their choice could, if such a thing existed, be under the control of a free will, since the stimuli were so difficult to tell apart that their choice was essentially random, and so relatively free from the stimulus conditions. Instead, they experienced themselves as helpless and as not being able to determine their actions. A related example may be seen when one asks people about when they are asleep and dreaming and there is nothing impinging on them that would determine their behavior, "How much of a sense do you have of free will controlling your dreams?" They say, "None" or "almost none." Whereas, if they are asked the same question when they are awake, they will answer that they experience a high degree of free will. In both of these examples, the sense of free will seems to occur exactly when participants do identify the environmental events leading up to their own behavior. The report of the sensation of free will does not occur when they do not identify specific events related to their behavior. This is paradoxical, since in the situation in which there is less discriminable control, one should feel more free. Neither the sense of free will nor the doctrine that flows from it can be disproved. Everyone, even the most empirical scientist, experiences the internal thoughts that precede their action and so could potentially have the sense that those thoughts cause their actions. The sense of free will is like an independent self—a "god" or homunculus within us. There is no way to collect evidence for or against such a homunculus. The only evidence that can be collected is interms of the actions themselves. Individuals' statements about why they engaged in those actions cannot provide empirical evidence that they are the causes in and of themselves. Their self-report cannot be detected by two independent paths but only by one path, the self-report itself (Commons, 2001). Despite a pervasive illusion of free will, and the fact that this illusion has become an integral part of much Western philosophical, religious and legal thinking, it is not a view that is scientifically supported. Since the end of World War II, many different areas of behavioral science have shown the many ways in which behavior is controlled by factors such as past history, current environmental factors, and broader contextual factors. We will provide a number of examples. Researchers in the field of learning, including Skinner (1971) have argued most strongly against the idea that free will exists. Skinner asserts that the fundamental mistake made by all those who choose weak methods of control is to assume that the balance of control is left to the individual, when in fact it is left to other unexamined factors. For an extensive review, see Wegner (2002). In addition, there is considerable research on predicting psychopathology that shows that, if one knows enough about someone's background, one can often predict which people will end up in trouble. Garbarino (1999) for example, argues that we could predict with at least some degree of certainty for about 90% of boys who become violent. They would have a number of risk factors, such as a history of abuse and/or neglect, exposure to community violence, and a history of difficulties in social interactions. For the other 10% or so, it would be much more difficult to predict, although according to Garbarino, many of these individuals do have serious disorders at an early age.