I shall argue that there can be no such thing as free will and moral responsibility. My argument is a metaphilosophical one that holds that neither concept can have discrete reference, Instead, these terms are merely honorific and subjective; they cannot be legitimized by appeal to the nature of extralinguistic reality, Free will and moral responsibility, as they are viewed in philosophical discourse and everyday life, are not to be counted as candidates among the class of real entities.

I have come to the non-realist position grudgingly.(p.5)

The ideas of free will and moral responsibility play a deep and central role in the way that we view ourselves. Because of this, almost all of the great figures in Western philosophy have addressed free will and moral responsibility, most of them attempting to vindicate these beliefs. In recent years, the importance of freedom and responsibility have been well documented. The belief that we often act freely and are at those times morally responsible for our behavior is an integral constituent of the Manifest Image of Man, the view of ourselves and the world that we hold before we entertain the postulation of theoretical entities that constitutes the Scientific Image (Sellars, 1963). The belief that persons are capable of moral responsibility and deserve moral consideration comprises (along with the logically prior intentional stance) the personal stance that we take toward each other (Dennett, 1978, Chapter 12). This personal attitude is characterized by "reactive attitudes" such as gratitude and resentment, which stand in distinction to the "objective" attitude that we take toward beings that we believe are not endowed with free will (Strawson, 1962; Nagel, 1986). The capacity for free will has been alleged to be a criterion for distinguishing persons from non-persons (Frankfurt, 1971). And in law, besides the obvious issue of retributive justice, the ideas of free will and moral responsibility provide the moral rationale for legal responsibility. (p.3)

Free choice (decision), freedom, free, free will, free act, free agent — Because the aim of this book is to see whether any theory of free will is acceptable (what "free will" really means), I cannot produce a complete definition here prior to that investigation. It is clear, though, that free will has to do with making choices that have the desirable property of being free, which enables agents who make such choices to be more worthy of dignity than agents who cannot. Free will seems, at first blush, to be something without which our moral responsibility for our actions will be jeopardized. At the same time, free will has connections with other highly desirable things like independence, autonomy, activeness rather than passivity, and rationality.

There is reason to believe that these general desiderata might be generated by the satisfaction of three basic conditions that have produced much of the historical debate between the compatibilists and incompatibilists. The first condition is the requirement that free agents have the ability to choose and to act differently than they actually do. The point here is that free agents do not have to make the choices they do; they have the ability, in a sense to be established, to choose otherwise. The second condition, often subsumed under the first, is that free agents control what their choices shall be. The intuitive idea, accepted by compatibilists and incompatibilists, is that if I choose freely the choice that I make is 'up to me', again in a sense to be established. The third component of free will is rationality. Free choices are reasonable and sensible in the light of a belief-desire psychology where we choose in order to maximize the likelihood of achieving our goals. (p.12)

I hope to show that a libertarian account of dual rationality, control, and the ability to choose otherwise also fails.

The most important [condition] for the libertarian is the requirement that free agents have the ability to choose otherwise under the exact conditions that obtain at the moment of choice. The libertarian's interpretation of this requirement signifies the clearest demarcation between the incompatibilists and the compatibilists and is probably incompatibilism's greatest selling point. The compatibilist's argument that the concept of "can choose otherwise" is analytically definable in terms of the hypothetical notion that agents would have chosen otherwise had some condition been different seems dubious to me. Chisholm rejects the hypothetical analysis by arguing that the categorical statement (a) "He could have done otherwise" is not logically equivalent to the hypothetical statement (b) "If he had chosen to do otherwise, then he would have done otherwise". I think that Chisholm is correct in his rejection of the compatibilist's claim when viewed as linguistic analysis. (p.191)

the libertarian's commitment to indeterminism guarantees a dual interpretation of the control and rationality conditions. For the libertarian, free agents not only have the ability to choose in two directions, but they must be in control of either choice. This consequence could be deduced from the demands of moral responsibility. If an agent is to be truly responsible for either choice, A or not A, then both outcomes are under the agent's control. The case is likewise with rationality, although a predictable amount of strain is likely to arise. Libertarians, at the very least, need to show how indeterministic choices can be rational whichever way they turn out. Ideally, libertarians will show us how such a dual ability might be, contrary to appearances, a desirable or even necessary element of rational choices. (p .192)