Let’s Separate Free Will from Moral Responsibility

This post will argue in favour of a separation between Free Will (FW) and Moral Responsibility (MR). It will start by examining the traditional relationships between these two and upon what premises they rest in terms of definitions and resulting truth values. It will then make a decision as to what attributes of MR are worth wanting and attempt to show that FW is not a condition for such an endeavour to succeed if we adopt a consequentialist view of MR. I shall then attempt to bring forward an example from Scot’s law to illustrate and then discuss a challenge by Strawson against this system. Afterwards it will briefly speculate that tensions could be relieved by rejecting a unified theory of MR in general. This post shall conclude that although such a distinction is prima facie reasonable, further research is required to see whether it is a first- best account of MR.

The problem of FW and MR arises out of a tension between three definitions: That of FW, MR as well as of determinism. Depending on what attributes we ascribe to each of these, their truth values are going to turn out differently. What this means is that for all three definitions to be logically consistent, only certain attributes can be incorporated. By the same token – and this is mostly the case in the debate – if you want specific attributes in your definitions, for instance libertarianism in your FW and the LaPlacian demon in determinism, your overall truth values between the three definitions are going to give an inconsistent picture. This is already the most classical example in the ancient debate. Let us assume 1) a rigid definition of determinism wherein the same past always brings about the same future; 2) a libertarian FW in which an agent is able to bring about causae sui and requires alternate possibilities; as well as 3) a picture of MR that holds people responsible iff they acted freely under a libertarian notion of FW. In this picture 2) and 3) are logically consistent, however, the three together are not. The task of the philosopher is to modify one or more of these three definitions in order to make them consistent.

This is where the debate between the compatibalists and incompatibalists sets in, during which mostly determinism and FW are attempted to be moulded into a coherent picture. This essay will not examine that particular debate in great depth, but suffice it to say that for instance the classical compatibalists tried to give a sophisticated account or new definition of FW in terms of constraint that they thought would be consistent with determinism (Kane, 2005: ch.2). However, if determinism is interpreted as fatalism, this consistency falls apart. Some others have created increasingly complicated and absurd accounts for alternate possibilities and coulds to somehow make the picture consistent (see exemplary Moore, 1912: 130). Philosophers around Frankfurt on the other hand showed forcefully that alternate possibilities or indeterminism is not required for a consistent picture with FW (see Frankfurt, 2003). Again the question arises, what precisely FW then becomes to mean. To cut a long story short for the purpose of this essay: The entire debate in this subject can be seen as philosophers trying out new definitions of the three terms and testing their truth values when paired up. As a result, the debate has brought forward some convincing definitions and possible interactions between them.

This leads us to the big question any philosopher must answer if she enters this debate: What kind of definitions does she want to defend? If it is libertarian FW, then something within determinism has to be changed to accommodate that. If the answer is the LaPlacian determinism or quantum indeterminism, then not all versions of FW can be defended at the same time. This essay however will start from the point of MR: I wish to debate some account of MR that I desire first and then “work myself backward” towards FW and determinism as it were. Fisher and Vargas correctly point out that as a result freedom becomes “whatever condition” is required to be morally responsible (Fischer, 2005: v.I xxiii; 2007: 128). I wish to do that for a simply reason: Unlike FW or determinism, MR must exist, just in virtue of us ascribing it to agents on a daily basis. Whether or not FW or determinism under any definition turns out to be false, we still require a notion of MR to keep society running. This comes close to Strawson’s claim that in any case people will want to hold on to their reactive attitudes (Strawson, 1962), a concept which I shall examine momentarily. As a normative study which is in its methodology at least in some ways different from empirically informed metaphysical considerations, it is a requirement for the entire field of ethics and separated from FW not at all unintelligible. I am not alone in thinking this (see Waller, 1990; or Zagzebski, 2002) My aim therefore is to find an account of MR that gives truth values that stay positive regardless of whether FW and determinism turn out to be true or false; a notion of MR that is decoupled from the metaphysical debate so that it can operate independedly and, to be somewhat colloquial, keep society running. To clarify, I am alas attempting to go even a step further than Dennett who also starts with some notion of MR and asks which the varieties of FW are worth wanting (Dennett, 1984), by finding a notion of MR that does not require any FW.

What could such a notion of MR be? What attributes would it require to give judges the power to ascribe MR to people and at the same time not appeal to FW? Let us start with some simple definition of MR: Sometimes when a agent acts or refrains from acting, we believe it to be appropriate to react in a certain way that indicates moral judgement. Praise and blame are the most commonly associated terms with this behaviour. I wish to define MR as being “worthy” of such a response (Eshleman, 2009: 1). To give an example: If I break a promise in cold blood, it may be appropriate to call me a traitor and hence blame me for my actions; or perhaps one should praise me for picking up and giving back £50 that a man walking before me dropped last month. Contrary to that I may say that the storm last week was responsible for uprooting trees in the Meadows, but it would seem absurd to blame Scottish weather for acting in such an outrageous manner. The intuition here seems to be that there is something special about agents that enables us to praise and blame them instead of natural phenomena. For many philosophers this turned out to be FW (Eshleman, 2009: 2). (For simplicity, this essay shall ignore the nuances between agents and phenomena, for instance in relation to animals as well as concerns whether praise and blame are the only warranted reactions in these cases.)

I disagree deeply here. I wish to propose instead of FW that we blame agents and not phenomena for consequentialist reasons that are independent of FW. Take the example of the promise breaker. In the case of Robertson v Anderson [2002] Scot SC 312, Mrs Anderson had repeatedly assures Mrs Robertson that if they ever won Bingo night in Glasgow, they would share the prize, which Mrs Anderson did not when finally she won. A court ruled that the behaviour of both parties indicated an intention to create legal obligations and that Mrs Anderson must share the winnings. (Note that this is a Scottish contract formation, not promise case, however the legal nuances shall be ignored for the purpose of this essay.) Under our definition we can say that Mrs Anderson is worthy of blame. The philosophically relevant question is now: What is it about Mrs Anderson that makes her worthy?

A consequentialist may answer that unlike the weather, Mrs Anderson can be made to change by such a ruling. In the future she will have been influenced not to repeat this behaviour iff, had the ruling not occurred then she may have done it again. Furthermore it becomes part of the general case law and is future reference to agents on how to behave so not to be blamed. Rulings under this reading ascribe MR as measure to give normative advice to society. We can even go a step further and ask what is it about the wind that makes it immune to blame? The consequentialist has an answer again: Blaming the wind cannot have any effect on future happenings since no ruling can be enforced against the weather. If you take this ad absurdum then the thing that distinguishes agents from other phenomena is their influenceability. Under this reading of MR, FW is no longer a requirement and it can operate independently. What this points at is the so called consequentialist view of MR under which praise and blame are appropriate iff they are “likely to lead to a desired change” in an agent. The contrast to this view is the merit based approach according to which praise and blame are appropriate iff an agent “deserves” or “merits” such a response, typically in reference to some idea of freedom or freedom to do otherwise (Watson, 1987: 258; Eshleman, 2009: 12).

Let me summarise where we are in the essay. We started out by putting MR in a context that interconnected it with FW and determinism. In an effort to decouple these, we took a closer look at MR and saw that the proposition that there is something special about agents let many philosophers to believe that FW must be part of the equation. Yet, I brought forward a positive case for why that is not necessarily so in consequentialist terms, where for a ruling in a court to assign blame, there need not be anything special in agents. This is opposed to the merit-based view of MR wherein an agent ought to “deserve” a ruling. However, this account of MR is far from unproblematic and I shall spend the remainder of this essay bringing forth counterexamples and negative cases in an effort to relativise this prima facie reasonable notion.

Let me begin this venture by looking at another notion of MR by Strawson: He argues that both, the merit as well as the consequentialist approach that I advocate are mistaken in his paper “Freedom and Resentment” (1962). This is the case since both warp the concept of MR by over-intellectualising it and assuming that there is some intelligible, formal basis on which to judge people. Under the consequentialist view for example one may blame a person only if a positive outcome is expected from doing so. Yet, people according to Strawson do not assign moral judgements in this manner. Rather, people have reactive attitudes towards other agents, this means that an if an agent’s act is morally significant, it will cause some reply in others, for instance, blame or praise. (According to Strawson there are many others, but for the purpose of this essay they shall be ignored.) Unlike the consequentialist view, these attitudes have the advantage of being modifiable and allow for justifications and excuses of an agent (Strawson, 1962: 7). This exposes a critical error in the consequentialist view, namely that there are counterexamples of reactive attitudes that intuitively are moral, however are not consequential in that these emotions are not controllable by a consequence conditions, i.e. just because a positive result cannot be achieved, I cannot stop feeling a certain way (Bennett, 1980: 22).

One could argue against this position by saying that Bennett misinterprets the consequentialist view in that it is not concerned with the attitudes of people or the “intentions of holding people responsible” but rather the “overall practise” and “function of holding people responsible” (Kupperman, 1991: 60-64, see also Eshleman, 2009: a11). However, I am not going to enter this particular debate here for the simple reason that I do not want to discard Strawson that quickly. After all, even though he goes about it in different terms, he is still on a very similar quest to my own in that he also has a strong intuition that FW and MR need separation to at least some extent. However, there is a great difference that must be discussed: Strawson goes about arguing that they are separate concepts by trying to show that no rational basis on which to judge people exists (as is the case in my consequentialist view). What he abstracts from that is a distinction between the mistaken idea of being responsible to holding responsible. For the latter he then argues there is no point of asking whether FW is true or false since in the process of holding an agent responsible FW does not appear as premise (Eshleman, 2009).

Here I am very sceptical of Strawson’s approach. Shifting the issue from being to holding responsible is admittedly not a formal equivocation, but one can charge him with missing the issue. After all, the holding responsible of people is a very practical idea and to build all of MR on it means that being responsible has to go. This is because Strawson as afore mentioned rejects any idea of a logical framework that allows people to assign responsibility to agents. If such a system cannot exist, neither can being morally responsible. This does not seem to be very philosophically pleasing. As Chisholm points out, such a notion of MR avoids the deeper issue of finding an account of the conditions an agent has to have in order to be responsible (Chisholm, 2002: 55). This leads us back to a point I have tried to make earlier in the essay. Since I defined MR as worthiness of praise and blame, the question of what attributes an agent must fulfil is raised. I am against of that being FW, but if we follow Strawson, then we just state how we morally feel about the behaviour of agents, not what makes them worthy of our reactive attitudes. For this reason I reject Strawson’s critique of the consequentialist view, since the alternative he offers is even more philosophically problematic and his argument that the consequentialist over-intellectualises seems void on light of the proposition that intellectual bases for assigning responsibility seem necessary so not to avoid the issue of what it means to be responsible.

This closes a circle with the earlier mentioned statement of Dennett who inquired what varieties of FW are worth wanting for the MR we seek. My answer is none since a consequentialist account can be separated from FW. Where does this leave us? It seems evident through Strawson’s arguments that there is something about MR that the consequentialist view is not able to capture, namely the intuition that we wish to assign MR also if no positive outcome can be achieved through it. However, we have seen that Strawson’s own view of how this can be rectified distorts what we want MR to do, i.e. point out the people the people that are responsible.

Philosophers have tried to find common ground between these two positions by pointing out that they rest upon the premise that there is one unified notion of MR, and that perhaps the problem can be solved by rejecting that premise (see exemplary Watson, 1996). Watson draws a distinction within responsible in respect of attributability and accountability. This distinction is related (but not parred!) to that between the more intellectually based account of MR (for instance the merit-based and consequentialist accounts) and a more social view of responsibility (with Strawson’s intuition of attitudes). An agent can be attributed responsibility iff the praise- or blameworthy act “belongs” to that agent (Eshleman, 2009: 25). What this means is that we hold an agent against some sort of “standard” to show whether her behaviour is praise- or blameworthy (Watson, 1996: 235). On the other hand the accountability of an agent is a purely social notion wherein an agent is justifiably met with some reactive attitudes by others (Watson, 1996: 235) (for simplicity this essay will ignore the nuances between different readings of these two terms, for a general overview see Eshleman, 2009: 24-27).

This this gives the following view of the debate: In an effort to relief tension between FW and MR, Strawson developed reactive attitudes, which in turn may appear as shifting the debate to holding people responsible and thereby avoiding the question what it means for an agent to be responsible. Watson entered the debate by proposing that MR can be understood as both attributability and accountability. These two terms tie back into the discussion of FW in the following way: Both seem to be at least partially connected in that attributability seems to require some idea of agent causation to be able to par an action to agent thereby determining their “belonging” together, while accountability has been argued to even require libertarian FW to succeed (Eshleman, 2009: 28).

This means that we are back to square one. We have tried to decouple FW and MR through consequentialist terms yet saw that Strawson has grounds on critiquing that view. We charged Strawson with missing the issue by shifting the debate from being to holding responsible. We thought that maybe this could be reconciled with the consequentialist view by letting go of a unified theory of MR, which led us to attributability, accountability; and eventually back to FW! We have hence made a journey trying to separate FW and MR that led us right to bringing both back together. What this means in conclusion is that although the separability of FW and MR seems intelligible, finding an account of how exactly MR would have to look is highly problematic. The consequentialist, Strawson and Wallace have had very different accounts of MR, yet none seem at this point apt to give a purely ethical notion of MR that is independent of the underlying metaphysical discussion of determinism and FW. To tie this back in with my view that this entire debate can be seen as the interconnection and truth values between different definitions of determinism, FW and MR: Without further research a decoupled account of MR gives a prima facie inconsistent picture.


Bennett, J. 1980. Accountability. in van Straaten, Z. Philosophical Subjects. Clarendon Press.

Chrisholm, R. 2002. Human Freedom and the Self. in Kane, R. Free Will. Blackwell. 47-58.

Dennett, D. 1984. Elbow Room: The Varieties of Free Will Worth Wanting. MIT Press.

Eshleman, A. 2009. Moral Responsibility. in Stanford Encyclopedia of Philosophy. URL=[http:// plato.stanford.edu/entries/moral-responsibility/]. access=[16.11.2011].

Fischer, J M. 2005. Free Will: Critical Concepts in Philosophy. Routledge.

Fischer, J M. et al. 2007. Four Views on Free Will. Wiley- Blackwell.

Frankfurt, H. 2003. Alternate Possibilities and Moral Responsibilities. in Watson, G. Free Will. Oxford University Press. 2nd ed. 167-76

Kane, R. 2005. A Contemporary Introduction to Free Will. Oxford University Press.

Kupperman, J. 1991. Character. Oxford University Press. Moore, G E. 1912. Ethics. London.

Strawson, P F. 1962. Freedom and Resentment. Proceedings of the British Academy 48: 1-25.

Waller, B N. 1990. Freedom Without Responsibility. Temple University Press.

Watson, G. 1996. Two Faces of Responsibility. Philosophical Topics 24: 227–248.

Zagzebski, L T. 2002. Recent Work on Devine Foreknowledge and Free Will. in Kane, R. The Oxford Handbook of Free Will. Oxford University Press.