Frankfurt's New Principle and its Conflict with the Compatibility of Moral Responsibility with Determinism

In Harry Frankfurt’s Essay Alternate Possibilities and Moral Responsibility, Frankfurt denies “the principle of alternate possibilities” (PAP), which states that “a person is morally responsible for what he has done only if he could have done otherwise” (486). Frankfurt goes further to suggest a replacement of the principle of alternate possibilities with his own principle. In the following essay, I will concur with Frankfurt’s denial and suggested reformulation of the principle of alternate possibilities. I will then respond to Frankfurt’s claim that his principle “does not appear to conflict with the view that moral responsibility is compatible with determinism” by setting out to demonstrate that the principle does in fact conflict with the compatibility between moral responsibility and determinism (492). Seeing as Frankfurt’s aforementioned statement is where he and I have conflicting views, the main focus of my paper will be on the repudiation of this statement. Let us now take a closer look at Frankfurt’s argument.

Frankfurt begins his paper by introducing the principle of alternate possibilities and declaring it to be false. He then argues his case by introducing several scenarios of action and questionable morality. The first few, serving merely as a warm-up exercise, involve a man Jones who is about to perform an action of his own accord but is threatened to perform that same action. As Frankfurt notes, the threat can have one of three effects: either (1) it does not affect him whatsoever and so he acts just as he would have without the presence of the threat, (2) it affects him tremendously and so he acts in response to the threat, forgetting his previous decision to act in this same way, or (3) it affects him somewhat and factors into his decision-making, but ultimately Jones acts because he already resolved to do so (488). It is evident that Jones is morally responsible in cases 1 and 3, in which he acted ultimately of his own accord, but is not morally responsible in case 2, since he acted in response to the threat and not because of his previous decision to do so (remember that he forgot his decision). In each case Jones performs the same action, unable to have done otherwise, but only in some cases is Jones morally responsible. This is Frankfurt’s first step toward showing that a person can be morally responsible even if that person could not have done otherwise (and thus refuting the principle of alternate possibilities).

Frankfurt then continues, with a set of more developed scenarios that address any possible criticisms of the warm-up exercise. In these scenarios, Jones is again set to perform an action of his own accord, but this time another man Black is prepared to force Jones to perform the action in the event that he deviates from his decision/resolve. In scenario 1, Jones does not deviate from his intentions and acts of his own will: he “decides to perform and does perform the very action Black wants him to perform,” so Black does not intervene (490). In scenario 2, Jones begins to deviate from his original intentions and so “Black takes effective steps to ensure that Jones decides to do, and that he does do, what he wants him to do” (490). As with the warm-up exercises, it is clear that Jones is not morally responsible in scenario 2 because he does not act due to his own intentions, but instead acts because he is forced in this direction. In scenario 1, on the other hand, Frankfurt insists that Jones is morally responsible since he acts because he previously decided to do so, as a direct result of his intentions: “his moral responsibility for doing it is not affected by the fact that Black was lurking in the background with sinister intent, since this intent never comes into play” (490-91). Again we see that Jones performs the same action in all cases and could not have done otherwise, although he is morally responsible in one but not in the other. The principle of alternate possibilities, however, would dictate that neither be morally responsible. Therefore, as Frankfurt would say, a person can still be morally responsible even if he could not have done otherwise, and so the principle of alternate possibilities must be false.

Frankfurt moves on from establishing that the principle of alternate possibilities is false and proposes that it “be replaced” by the principle that “a person is not morally responsible for what he has done if he did it only because he could not have done otherwise” (492). We will refer to this principle as PAPF. I am in agreement with Frankfurt’s conclusion that the principle of alternate possibilities is false and assent as its proposed reformulation. While I would have suggested a different modified principle, Frankfurt’s principle still seems to be a good improvement over the principle of alternate possibilities. It prescribes the expected moral outcomes of Frankfurt’s scenarios (it appropriately dictates moral responsibility in scenario 1 but not scenario 2) and so seems to be more fitting for determining the existence of moral responsibility.

Let us now turn to look at Frankfurt’s claim regarding his new principle: “This principle does not appear to conflict with the view that moral responsibility is compatible with determinism” (492). In the paragraphs to follow, I will show that Frankfurt’s principle does in fact conflict with the compatibility between moral responsibility and determinism. We have dealt with several scenarios of action and questionable morality thus far, all involving Jones and Black. The two main ones considered are the scenarios in which (1) Jones acts of his own accord, without intervention from Black, and (2) Jones acts because Black steps in to ensure that he performs the appropriate action. We will consider an additional scenario, (3) in which a man performs an action in a deterministic universe, but first we will return to the first two to clarify a few concepts.

In the Jones and Black scenario 1, Jones acts, without the intervention of Black, for two reasons, which we will call reasons “a” and “b.” He acts (a) because he could not have done otherwise and (b) because he forms the intentions/resolves to do so, independently of the fact that he could not have done otherwise. No force influenced him prior to the intention formation, and therefore the formation of his intentions is solely due to his own will. Black waits readily to force Jones’ decision, just in case the event arises that Jones deviates from his intentions, but Black still has no actual effect on these intentions. As Frankfurt insists, “The circumstances that made it impossible for him to do otherwise could have been subtracted from the situation without affecting what happened or why it happened that way” (490).

In the second Jones and Black scenario 2, Jones acts for only one reason. He acts only because he could not have done otherwise. He does not act of his own accord, as he deviates from deciding to perform the action that Black wants him to perform and Black is required to force Jones’ decisions and actions. In this case, Jones is not morally responsible, since his decisions were forced and he acted only because he could not have done otherwise. Furthermore, he did not act because he independently formed the intentions to perform the action.

Now we will move on to considering the deterministic universe scenario (scenario 3). To clarify, what I mean (and is typically meant) by a deterministic universe is one that is governed entirely by causal laws such that the initial conditions of the universe can only lead to one particular state at any subsequent time t. Therefore every state of the universe, every particle position, and every event that ever occurs is predetermined from some previous state or set of initial conditions in the distant past. This means that, up until this point in time, everything had to have happened exactly the way it did and the future can only have one possible outcome. This also means that our actions and thoughts are not free and any choice or freedom that we may think we have is nothing more than an illusion.

Now that we have clarified enough about what is meant by a deterministic universe, we will return to our consideration of scenario 3. It is obvious here that the man acts at least for the reason that he was unable to do otherwise. It may seem unclear at first however, whether he also acts for another reason in the same sense that Jones acts for more than one reason in scenario 1. First consider the possibility that the man acts only for the reason that he was unable to do otherwise, which is a perfectly rational conclusion for an unspecified action in a deterministic universe. If this conclusion is true, then in accordance with Frankfurt’s principle, the man is not morally responsible. This in turn would mean that no action in a deterministic universe could entail moral responsibility. Expressed another way, if this is true, moral responsibility is not compatible with determinism. Since my view is that the man acts only because could not have done otherwise, I support this conclusion. Furthermore, anyone who holds the same view would either have to come to an identical conclusion or refute Frankfurt’s principle (or else find issue with my progression of logic). Frankfurt believes the converse of this conclusion, that moral responsibility is compatible with determinism. It seems that in order to maintain this position, he would have to support the possibility that Jones acts for a reason in addition to the fact that he is unable to do otherwise. I will now argue against Frankfurt’s statement and demonstrate its fallacy by setting out to show that, in the sense that Frankfurt describes it, Jones acts only because he is unable to do otherwise. In the Jones and Black scenario 2, it is clear to us that Jones acts only due to reason (a); he does not act of his own resolve. This is because there is something that influences his decision-making. He is not able to form intentions completely of his own will, without any forcing hand, that would result in the relevant action. Although it may be unclear at first, the man acts only due to reason (a) in the same fashion as in the deterministic universe scenario 3. One could say that he acts with his own intentions and resolves, but not in the sense that we require for application to Frankfurt’s principle. In a deterministic sense, the man is not able to form his intentions or make his decisions completely of his own will; they are not free or under his control in any way. His intentions and brain states simply follow from the conditions of the previous states of the universe. This is the deterministic “forcing hand.”

In fact, this is an even greater example of a forcing hand than the Jones and Black scenario 2. In scenario 2, Jones is at least uninfluenced in forming his own thoughts and deliberations and is even given a split-second to waver in his intention before he is snapped back into line. The forcing hand acts after the process of deliberation and only exists just in case deviation of intention occurs. In scenario 3, however, every event and action has been previously determined by the past, so every cognitive function, every intention and every thought is under the greatest level of influence possible. This is comparable to a scenario (4) in which Black, instead of altering Jones’ mental state at the time of the decision (as in scenario 2), finds a way to access Jones’ mind at some time before decision formation and program in every one of his subsequent thoughts, decisions, and actions. Black makes sure that environmental factors do not influence Jones’ decision, but just in case develops a way to influence all the environmental factors that affect Jones. If we consider Black accesses Jones’ mind a split second before decision formation, or seconds, days or years before decision formation, this will not make Jones any more morally responsible, since Black still has complete control over all of Jones’ thoughts and intentions. This scenario demonstrates an all-encompassing forcing hand. Therefore, if we are to say that Jones acts only due to reason (a) in scenario 2 (and thus is not morally responsible in this scenario), then we must also be driven to say that the man acts only due to reason (a) in scenario 3 (and likewise is not morally responsible).

At this point, Frankfurt’s only option is to maintain that in scenario 3 (and likewise scenario 4), the man acts due to both reasons, the justification for reason (b) being that no force stepped in and directly influenced his decision to act. So, in some sense he did act because he wanted to, or because he formed the appropriate intentions. This pathway leads to some quite extreme conclusions, however, and would not be advisable for Frankfurt to take. It would mean, for example, that in scenario 2, Jones is not morally responsible because Black influences the decision that leads directly to his action, whereas in scenario 4, Jones is morally responsible because Black influences Jones at a previous time. To clarify, in scenario 4, even though Black is able to determine the exact chain of thoughts and decisions that lead to the action, it is not a direct intervention. This would hold Jones morally responsible even though neither his actions nor his intentions nor his thoughts on the matter would be under his control. But this seems absurd. Black could have intervened 20 years, a week, 5 minutes, or even 3 seconds before Jones performed the action and Jones would have still been held accountable, morally responsible for his state at the time, even though his state was not at any time under his control.

To take this further, suppose Black intervened at a time before Jones’ birth, such that he was able to configure all the factors, internal and environmental, that would go into developing Jones’ mind in a certain way and ensuring that Jones performed a certain action later in life. Then, Jones would have been held responsible for the person he was born as and the person he developed into, both pre-determined by Black and completely out of his control. This seems absurd as well. How can any of us be said to be responsible for how we were born? This would result in a terrible and unfortunate view towards children with birth defects. Thus, we have no choice but to reject this idea that Jones is morally responsible in scenario 4. But this extreme case of scenario 4, with a far back pre-determination, is essentially equivalent to scenario 3. This means that we must reject the idea that Jones is morally responsible in scenario 3 as well. As a result, we once again come to the conclusion that moral responsibility is incompatible with determinism.

In this paper, I introduced Frankfurt’s refutation of the principle of alternate possibilities. I concurred with his criticisms of the principle and also accepted his reformulation of the principle. I found fault, however, with his statement that the principle does not conflict with the compatibility of moral responsibility and determinism. I reasoned that in order for moral responsibility to co-exist with determinism, a person acting in a deterministic universe would have to act for a reason other than the fact that he could not have done otherwise, in the sense that Jones acts in scenario 1 for more than one reason but does not in scenario 2. I contested Frankfurt’s insistence of compatibility by rationalizing that if we consider the possibility that Jones acts only for the reason that he could not have done otherwise, then we must either say that a man in a deterministic universe acts for only this reason or we must come to some unlikely conclusions. From this we see that the most likely possibility that we can accept is that Frankfurt’s statement is false. As I set out to show, Frankfurt’s principal, just like the principle of alternate possibilities, does in fact conflict with the view that moral responsibility and determinism are compatible.


Frankfurt, Harry. “Alternate Possibilities and Moral Responsibility.” Reason and Responsibility: Readings in Some Basic Problems of Philosophy. 13th ed. Eds. Joel Feinberg and Russ Shafer-Landau. Belmont, CA: Wadsworth Publishing, 2007. 486-492.