Compatibilism: What’s Wrong, and How to Fix It

This is a critical review of the Stanford Encyclopedia of Philosophy (SEP) article on “Compatibilism” [1], written by Michael McKenna and D. Justin Coates in 2004 with revisions in 2015.

Before we begin, it might be helpful to briefly explain how free will and determinism are compatible concepts.

We observe a woman going into a restaurant. She picks up the menu and considers her options. She has several goals in mind. She wants to satisfy her hunger and her tastes. But she also wants to eat healthy and avoid excess calories. She reads the menu to see her options. She finds that the Chef Salad best suits her goals and her reasons. So, she tells the waiter, “I will have the Chef Salad.”

We observe that she is not a child, subject to her parent’s will. And we don’t see anyone holding a gun to her head, telling her she must order a cheeseburger instead. Her choice seemed reasonable given her goals, so we can rule out the influence of mental illness. And, as far as we can tell, no one has placed her under hypnosis.

So, we empirically observe that she made this choice herself, of her own free will.

Free will is when a person decides for themselves what they will do, free of coercion or undue influence. It is literally a freely chosen “will”.

Determinism asserts that her choice was causally necessary, from any prior point in eternity. If we assume a world of perfectly reliable cause and affect, then it is a logical fact that every event that ever happens, or ever will happen, is “causally necessary”, and inevitably must happen.

Okay. So, assuming a deterministic universe, we can assert that it was causally necessary, from any prior point in eternity, that this woman would order the Chef Salad. And some people might stop there and insist that she had no choice. But they would be wrong, because they’ve left out a few things.

The following were also causally necessary, from any prior point in eternity:
1. That she would feel hungry around lunchtime, and decide to try out that new restaurant.
2. That she would be faced with a menu of options, requiring that she make a choice.
3. That she would weigh her options in terms of her own goals.
4. That she would conclude that the Chef Salad was the best choice.
5. That there would be no one forcing her to choose something else.
6. That she would therefore make this choice for herself, of her own free will.

Deterministic inevitability doesn’t change anything. It is still her, going into the restaurant and choosing what she’ll have for lunch. No other object or force in the physical universe is making this choice for her.

This event is an authentic example of free will. It is also an authentic example of determinism. The two facts, (1) that the event was reliably caused, and (2) that it was reliably caused by her, are obviously compatible.

Okay, so let’s see what the SEP authors have to say about Compatibilism. I’m going to use their numbering scheme, to make it easier for anyone following along.

1. Definitions

1.1 Free Will

(1) The SEP authors suggest what they call a “theory neutral” definition: “free will can be defined as the unique ability of persons to exercise control over their conduct in a manner necessary for moral responsibility“.

To this I would add an operational definition of free will: “free will is when a person decides for themselves what they will do, free of coercion or other undue influence.” And while we’re here, we can also define its components:

(2) A person’s will is their specific intent for the immediate or distant future. The woman’s immediate intent is to have the Chef’s Salad for lunch. At home she may also have a “last will and testament”, specifying her intentions for distributing her assets after her death.

(3) One’s will is often chosen from a number of possibilities. Choosing is a logical operation which inputs multiple options, applies some criteria of evaluation, and outputs a single choice. Her choosing sets her intent upon having the Chef’s Salad, and she acts upon that intent by placing the order with the waiter.

(4) Free refers to the absence of some meaningful and relevant constraint. To be meaningful, a constraint must prevent us from doing what we want. To be relevant, a constraint must be something that can either be present or absent. Here are some examples:
(a) A bird may be set free (of its cage).
(b) We enjoy freedom of speech (free of political censorship).
(c) The bank offered a free toaster (free of charge) to anyone opening a new account.
(d) The woman chose to have the Chef Salad of her own free will (free of coercion or other undue influence).

So, what is the impact of causal necessity upon our freedom? Surprisingly, it turns out to be neither a meaningful nor a relevant constraint.

First, causal necessity is not a meaningful constraint because, while some specific causes may be experienced as a constraint (the cage, the censorship, the cost, the coercion), other specific causes are the very sources of our freedoms (the wings, the voice, the money, the mind). To be “free of reliable cause and effect” would not only remove all constraints, but it would also remove all the mechanisms of our freedom. Thus “freedom from causation” is an irrational concept.

Second, causal necessity is not a relevant constraint, because it always applies to every event that happens. It is never absent. It is like a background constant that always appears on both sides of every equation, and can be subtracted from each side without affecting the result. What you will inevitably do is exactly identical to you just being you, doing what you do, and choosing what you choose.

Causal necessity is a “logical” fact, implied by the presumption of reliable cause and effect. But it is not a meaningful or relevant fact. All it can tell us is what Doris Day told us when she sang, “Que sera, sera. Whatever will be will be.” Nothing new or useful there. The reasonable mind simply acknowledges and then ignores it.

All of the utility of reliable causation comes from knowing the specific causes of specific effects. We know for instance that a virus causes polio. We also know the body’s immune system can be trained to fight it by a simple vaccination. With that knowledge we gain control over this and many other viral diseases–diseases that otherwise might control our lives. But the logical fact that all of this was causally inevitable from any prior point in eternity tells us nothing new, and nothing we can put to any practical use.

1.2 Moral Responsibility

(1) The SEP authors operationally define a “morally responsible agent” as someone that we hold “accountable” for their behavior. If the behavior is especially good, then we encourage it, using “praise” or other “reward”. If the behavior is bad, we discourage it, using “blame” or some other “punishment”.

(2) A key concept that the SEP authors overlooked is correction. Correction includes those rehabilitative steps that encourage and support the offender’s ability to behave better in the future. One cannot speak of what a person deserves without including some reasonable opportunity for redemption.

(3) Nor do they mention that moral responsibility operates within a system of justice. A system of justice protects and restores rights. A just penalty would naturally include (a) repairing the harm to the victim when possible, (b) correcting the behavior of the offender, (c) protecting the public until the behavior is corrected, and (d) doing no more harm to the offender or his rights than is reasonably necessary to accomplish (a), (b), and (c).

(4) Returning to the SEP, the authors suggest that there is an “intimate connection between free will and moral responsibility”. But that is a bit off the mark. Whether the cause of the harm is deliberate, accidental, coerced, or due to a mental illness, we are justified in taking some corrective action. It is the harm, not the free will, that justifies our action.

(5) But free will does play a key role in choosing the means of correction. For example, if the offender was coerced to participate in the crime, by someone holding a gun to his head, then the offender is easily corrected by simply removing the threat. If the offender’s ability to reason was impaired by mental illness, then the correction would involve psychiatric treatment in a medical institution. If the behavior was deliberately chosen, for reasons of personal benefit, then correction will require addressing how the offender thinks about these choices in the future.

1.3 Determinism

The SEP authors make three assertions in their attempt to define determinism.

(1) First, they correctly note that determinism asserts that “every event (except the first, if there is one) is causally necessitated by antecedent events.”

(2) Second, they suggest that determinism means that “the facts of the past, in conjunction with the laws of nature, entail every truth about the future”. But that will depend upon what is implied by the “laws of nature”. Are these limited to the laws found in a textbook on Physics? Or do they also include the reliable patterns of behavior described by the Life Sciences (e.g., Biology, Physiology, Genetics) and the Social Sciences (e.g., Psychology, Sociology)? And would they also include the laws that societies create for themselves?

All four types of law are required to explain why a car stops at a red light. We have the physics of the red light and the physics of applying the brakes. But between these two physical events we have the biological purpose to survive, thrive, and reproduce, and the rational calculation that the best way to do this is to stop at the light. There’s no way to get there through Physics alone, without first building a causal agent that has both motivation and reasoning.

So, their second assertion is only provisionally true. It is only true when we include all three levels of causation: physical, biological, and rational.

(3) Third, well, at this point their train derails. They say that “if determinism is true, there are (causal) conditions for that person’s actions located in the remote past, prior to her birth, that are sufficient for each of her actions.” Such language gives the impression that something in the past can magically bypass her, and bring about her actions without her participation or consent.

What they fail to recognize is that all prior causes must first become a part of who she is, before they can have any impact upon what she does. And, once they are her, then it is she, herself, that is doing the choosing and the acting and the causing. Prior causes can account for how she happens to be who and what she is, but they are never “sufficient” to do anything without her.

(4) Determinism then, is simply the assertion that each state or event has reliable causes that bring it about, and that each of these causes is also a prior state or event with its own causes. Within this overall scheme of perfectly reliable causation, we find ourselves in the role of causal agents, motivated by goals that exist only within ourselves and our species, and with the capacity to imagine, evaluate, and choose how we will go about achieving them. Once determinism is correctly defined, it no longer poses any threat to free will.

(5) Any version of determinism that fails to take into account all three levels of causation, or fails to recognize the causal agency of living organisms, would be incomplete, and thus false.

1.4 Compatibilism’s Competitors

(1) In this section, the authors correctly point out that incompatibilists come in two flavors: “hard” determinists who hold that determinism is true but free will is false; and “libertarians” who hold that free will is true but determinism is false.

Both incompatibilists appear to view causal necessity as some kind of force that compels us to do things. Some “hard” determinists foster this illusion by suggesting that reliable causation is like a puppet master pulling our strings. Or that determinism is the driver of a bus on which we can only passively ride. The point of such imagery is to shift control from us to our prior causes. This “boogeyman” version of determinism scares the theist into seeking escape through the supernatural, and sends the secularist seeking sanctuary within quantum indeterminism.

But determinism is neither an object nor a force. It cannot do anything. It merely asserts that the actual objects and forces that make up the physical universe interact naturally in predictable ways. Scientists observe reliable patterns of behavior and document them in the Physical Sciences (for inanimate objects), the Life Sciences (for living organisms), and the Social Sciences (for intelligent species). Science embraces determinism because it supports their hope that there are always reliable answers to the question, “Why did this happen?”

We happen to be one of the real objects that make up the physical universe. We are specifically structured as living organisms of an intelligent species. Unlike inanimate matter, a living organism is animated to survive, thrive, and reproduce. With life came purposeful behavior. And, after sufficient neurological evolution, we acquired the abilities to imagine, evaluate, and choose. These capabilities aid our survival as a species. Instead of just instinctual behavior, we now have rational calculation and judgment, the stuff of free will.

Knowing the specific causes of specific effects empowers us. We see in the toddler a miniature scientist, uncovering the secrets of gravity, through trial and error, with each tiny step. Soon, this knowledge will be ingrained in his neuromuscular memory, and he’ll be hard to keep up with. Knowledge is power. But if the force of gravity were unreliable, perhaps reversing at random intervals, then none of us would walk. All of our freedom and all of our control is dependent upon specific causes reliably bringing about specific effects. Our ability to control ourselves and our environment requires that our actions have predictable results.

But the “hard” determinist turns the glass upside down. He takes the one logical fact of causal inevitability, and turns determinism into a master that enslaves us, robbing us of all our control and all our freedoms. It is a perverse viewpoint.

(2) The SEP authors make a serious blunder when they say, “Neither compatibilism nor incompatibilism as such is committed to the further claim that any human persons ever do, in fact, have free will.” This suggests that they are biased against compatibilism, and not the best choices to explain the compatibilist viewpoint.

All compatibilists, by definition, assert that people do in fact have free will, and that free will operates within a determinmistic universe. Compatibilists reject the word games played by the incompatibilists, their figurative speech, and their blatant substitution of “freedom from reliable causation” for “freedom from coercion and undue influence”.

“Freedom from reliable causation” is an irrational concept. Without reliable cause and effect we could never reliably cause any effect, and would have no freedom to do anything at all. Free will requires a deterministic universe. Without it, the will would be impotent to effect any intent.

“1.5 The Free Will Problem”

To the compatibilist, of course, there is no “free will problem”. We observe reliable cause and effect in everyday events. Among these events are our own choices. We observe ourselves making these choices according to our own goals and our own reasoning. Thus our choices are both reliably caused (deterministic) and reliably caused by us (free will).

So let’s see how the SEP authors lay out the incompatibilist arguments.

(1) The first argument, which they call “the Classical Formulation”, is that free will implies that “Some agent, at some time, could have acted otherwise than she did.”

Our woman in the restaurant had a literal menu of options to choose from. She could choose the cheeseburger, or anything else listed, instead of the Chef’s Salad. She had that ability.

However, it was inevitable that, given her goals and her reasoning, she would choose the Chef’s Salad.

The incompatibilist, conflating the “could” with the “would”, claims she “could not” have chosen otherwise, based upon the fact that she “would not” have chosen otherwise. However, this subtle slip of semantics creates a logical dilemma, one that breaks the choosing operation, as illustrated here:

Waiter (a “hard” determinist): “Welcome, Sir. What will you have for dinner tonight?”
Customer: “I’m not sure yet. What are my possibilities?”
Waiter: “There is only one possibility, Sir.”
Customer (disappointed): “Oh. Okay then, what is that possibility?”
Waiter: “How should I know? I’m not a mind reader!”

The logical operation of choosing always begins with multiple, real possibilities, any one of which can be selected. However, only one possibility will be selected. The fact that only one will be selected does not conflict with the other fact that any one of our possibilities could be selected.

Our “can’s” and “could have’s”, our “possibilities” and “options”, exist within the context of our imagination. This is the place where we think, envision, evaluate, and choose what we will do next. This feature of intelligent species allows them to adapt successfully to a greater variety of conditions, improving their odds of survival.

Whenever we speak of what we “could have done” we revert to that context. Whenever we speak of what was “possible” we revert to that context. A possibility is considered real if we could implement it, should we choose to do so. An option that we could not implement, even if we chose it, is not a “real” possibility, and we discard these from consideration.

Our “hard” determinists (a.k.a. “free will skeptics”) throw a monkey wrench into this operation by insisting that we move the end of the process to the beginning. They tell us things like, “you really only have one possibility” or “you don’t really have any choices” or “your experience of choosing is an illusion”. All three statements are literally false. All three are forms of figurative speech, where they have left out the implied “as if”. We know this because it is literally true that you have multiple possibilities at the beginning, and that you are literally making a choice, and it is literally you, and only you, that is performing the choosing operation, specifically through a mental process running upon the physical structure of your own brain.

Our “could have’s” are also useful when our choice doesn’t turn out as we expected. We revisit that scenario in our memory, and consider how things might have turned out if we had made a different choice. This is how we learn from our mistakes. This is how we make better choices in the future. Yet our “hard” determinists erroneously insist that “you could not have done otherwise”, as if our behavior were instinctual, involving no rational calculation at all.

When someone asks us whether it was “possible” to have chosen a different option, we answer “yes”. We logically revert to the context of the choosing process, and correctly recall that we had several options at the beginning of this process, any one of which “could have” been chosen.

The authors make the same error when they say, “what it means to claim that an event is causally determined—that, if it were, then given the antecedent causal conditions for the event, it was not possible for it not to have occurred.” That certainly “sounds like” commonsense. But, even clouded in the double negative, we find the authors’ implicit suggestion that determinism eliminates possibilities.

(2) And that would be false. The correct statement of deterministic inevitability is this: “If an event is causally determined, then it will happen.” What will happen poses no limits on what “can” or “cannot” happen. What will happen has no impact upon what is “possible” and what is “impossible”.

How can we say this? Because within the domain of human influence through deliberate action, we always begin the process with multiple possibilities, followed by an evaluation, followed by a chosen will to bring about a single specific event. That is the inescapable nature of this series of events in the causal chain.

(3) The living organism is a process running upon a physical infrastructure. A multitude of reliable sub-processes enable it to live and to think and to act. Each process depends upon reliable causation to work. Reliable causation then, is not a constraint upon our freedom, but rather a description of how our freedom operates.

2. Two Models of Control

In this second section, the authors correctly point out that the central issue is who or what is controlling the choices, and thus the actions of the person. The compatibilist points out the empirical fact that the choosing is being performed by a specific, intelligent living organism. The woman in the restaurant, choosing from a menu of options, according to her own goals and her own reasons, ordered the Chef’s Salad. She has controlled this choice to suit her own interests, and is held responsible for her action later, when the waiter returns with the bill.

So, let’s see where the SEP authors go with the concept of control.

2.1 “The Garden of the Forking Paths”

What the fork?! Pardon me while I set this garden metaphor upon the compost heap. Now, let’s continue.

This section basically repeats the issue of “alternate possibilities” and “could have done otherwise”. We’ve dealt with that in detail in section “1.5 The Free Will Problem” above. Just to summarize then:

(1) The woman in the restaurant literally had a menu of alternate possibilities, any one of which she could have chosen. However, given her goals and her reasoning, it was inevitable that she would choose the Chef’s Salad. The fact that she inevitability would select the salad does not contradict the other fact that she could have chosen the cheeseburger.

(2) The SEP author’s suggestion that, “if determinism is true, no one has access to alternatives”, is literally false (as is the case with all figurative statements). There’s the woman, sitting in the restaurant, literally accessing a list of alternatives on the menu!

“2.2 Source Models and Source Worries”

(1) They define the “source” problem this way: “If an agent is not the ultimate source of her actions, then her actions do not originate in her, and if her actions are the outcomes of conditions guaranteeing them, how can she be said to control them?” You’ll recall this as the false suggestion that prior causes are sufficient to bring about the event without her.

But what about these prior causes? Don’t her prior causes also have their own prior causes? If their argument were valid, then the control would continue to shift back through the chain of prior causes all the way to — well, that depends upon your cosmology, but general practice is to stop when we get to the Big Bang.

By their logic, the Big Bang is the “ultimate source” of all events, and thus it should be “held accountable” for everything that happens. But how is that done? How do we perform the operation of holding the Big Bang responsible? We don’t, because we can’t! So the “ultimate source” argument is a useless dead end.

The operation of “holding responsible” only works with a causal agent over which we can exercise some control. A criminal offender can be arrested, imprisoned, and offered rehabilitation programs to correct his behavior. We may (and should) also address the social conditions that breed criminal behavior, but the offender himself will still require correction.

(2) The requirement that the woman must have controlled or caused herself, before she can be said to control or cause anything else, is irrational. If we apply that test to the woman’s control, then we must also apply that test to her prior causes, which the “hard” determinist claims are the real sources of control. Each prior cause must have controlled its own causation. The problem is that no prior cause can pass that test. If we insist upon such a test, then the causal chain simply disintegrates. The requirement, then, that she must have first caused herself before she can be said to cause or control anything else, is irrational and can be discarded.

(3) The core issue of “holding responsible” is itself a matter of control. To hold the offender responsible for his actions implies that we are able to control our own actions. The concept of responsibility contains its own operational imperative: that we can in fact control what we do.

In summary then, the “ultimate source” arguments have no valid foundation. They are at best impractical, and at worst irrational. They can be dismissed.

2.3 “Compatibilists’ Ameliorating Efforts”

In this section, the SEP authors muse over how their imaginary compatibilist might respond to the incompatibilist arguments. Since I’ve laid out how an actual compatibilist deals with these matters above, nothing more needs to be said here.

3. Classical Compatibilism

Before we go on, let’s remind ourselves of the correct operational definition of “free will”.

Free will is when we decide for ourselves what we will do, free of coercion or other undue influence. This is the common understanding of free will used by most people. It requires nothing supernatural. It makes no assertions against reliable cause and effect. And yet it is sufficient for both moral and legal responsibility.

Admittedly, not all undue influences are as clear and explicit as a gun to the head. In cases of mental illness we must consider how the specific condition affects a person’s ability to control their own behavior. The impairment may be to the person’s perception of reality, causing delusions or hallucinations. Or the impairment may be an abnormally compelling desire. Or it may disrupt the function of reasoning itself. In complex cases the courts and juries, informed by expert psychiatric testimony, can only do their best. The problem for a system of justice is how to effectively protect the public from harm while providing the best opportunity for rehabilitation.

Okay, so now let’s explore the SEP authors’ “compatibilist definitions” to see how they stack up.

3.1 Freedom According to Classical Compatibilism

The SEP authors cite examples of freedom being defined as the unimpeded ability to do what you want, and free will as freedom from being “forced by some foreign or external source to act contrary to one’s will” (coercion).

They also cite some who suggest that “freedom of will” might be replaced with “freedom of action”. But free will is not a general freedom of action. Free will is concerned with the specific action of “choosing what I will do”.

They point out that external coercion alone is not a sufficient criteria, because it does not cover the internal coercion of a psychosis accompanied by “full-fledged hallucinations”. The person may be doing exactly what she wants, “unencumbered”, but is still not considered by most people to be acting of her own free will.

That’s why we include other forms of “undue influence” in the operational definition of free will. A severe psychosis is an extraordinary (undue) influence upon the person’s ability to choose for themselves what they will do. Other examples of undue influence may be less dramatic. In law, certain relationships of unequal power result in one person having an undue influence upon someone else, which comes up in contract law [2]. We’re using a more general sense here to cover any influence that could reasonably be said to remove the agent’s own control of their choice.

Reliable cause and effect, in itself, cannot be viewed as an undue influence. We exist as a package of reliable processes running on a physical structure. Without reliable cause and effect, we could not exist.

3.2 “The Lasting Influence of Classical Compatibilist Free Will”

Here the SEP authors continue to list other undue influences, such as a severe addiction, or an irresistible neurotic compulsion, or an incapacity to make rational judgments. All of these can be gathered together under the category of “undue influence”.

They also briefly revisit the “ultimate source” notion, this time noting that the agent, if not the ultimate source, is still the “mediated” source. What they mean by that is unclear. What should be clear, at this point, is that the intelligent living organism performing the choosing operation is the most relevant and meaningful source of its deliberate actions. The final responsible prior cause, of any deliberate act, is that act of deliberation, uncorrupted by mental illness, that chose to perform the act.

3.3 The Classical Compatibilist Conditional Analysis

(1) Here the SEP authors revisit the issue of “the ability to do otherwise”, citing various authors’ opinions as to how this might be “possible” or “impossible” in a deterministic universe. The result of this cross talk is a bit of a Gordian Knot. Hang on a second while I find my sword.

The “ability to do otherwise” is embedded in the logical operation of choosing. Choosing is a deterministic process performed by an intelligent living organism. Choosing is a real event that actually takes place in physical reality. It is a real process running upon the “hardware” of the organism’s neurology.

The process is modeled in our minds as
1. the consideration of multiple real possibilities which could be actualized if chosen,
2. a comparative evaluation of these options in terms of the person’s goals and interests,
3. a single choice setting our intent (will) to achieve a specific outcome.

The “ability to do otherwise” is logically required at the beginning of the choosing operation. The single inevitable outcome does not arrive until the end. All three steps are essential to the operation. And the operation is essential to the survival of the species. So, this is the most “pertinent sense” of the concept of alternate possibilities.

(2) The SEP authors introduce the term “counterfactual”. A counterfactual is “counter to the facts”. They’ve already given us many examples of these. The claim that a woman reading a menu has no choices is a counterfactual. The suggestion that she could not have chosen something other than what she did choose is a counterfactual. The assertion that determinism is incompatible with free will is a counterfactual. I suppose one could say that the whole set of incompatibilist arguments are built upon a flimsy structure of counterfactuals derived from figurative speech.

But what they are talking about in this section is a “counterfactual condition”, which is simply a conditional statement containing a counterfactual (something known at the outset to be false) within it.

For example, “If she had chosen the cheeseburger, then she could have had a cheeseburger for lunch.” We know that the woman did not choose the cheeseburger. So, the SEP authors are suggesting that the conditional “if she had chosen the cheeseburger, she could have had the cheeseburger” is somehow also false. Well, no. That’s not how conditional statements are judged to be true or false.

“If she had chosen the cheeseburger, then she could have had the cheeseburger for lunch” is a true statement. It is just as true as the statement, “she did not choose the cheeseburger for lunch”.

A false conditional would be, “If she had chosen a tyrannosaurus steak for lunch, then she could have had dinosaur meat instead of the salad”. It is false because she could not have that for lunch even if she chose to have it. It was not a real possibility.

The authors give some hints that they understand this distinction, yet they return to saying that she “was not able to do otherwise in the pertinent sense”, providing us with another example of the “counterfactual”.

(3) The SEP authors suggest that the compatibilists are arguing that “Since determinism is a thesis about what must happen in the future given the actual past, determinism is consistent with the future being different given a different past.” Well, no. There is only one past, and that’s where the future must eventually go, so there will be one and only one future.

However, in the context of our imagination, we create many possible futures. There can be as many futures as we can imagine. This real process of imagining different possible futures is the mechanism by which the single inevitable future is causally determined.

This real process is a logical manipulation of options, conditions, and scenarios, that would produce different outcomes for different choices. The outcome that we find most desirable at that moment becomes our choice. Our choice becomes our specific intent. Our specific intent guides our actions. Our actions (hopefully) produce the future we envisioned or something similar to it. If it doesn’t, then we go back to the drawing board to see what else we might have done instead.

The authors ask, “How is this counterfactual ability more than a hollow freedom?” The answer is that the ability to reason by conditional logic is all that we require to set our intentions upon accomplishing any one of a wide variety of goals that suit our interests. Do they consider it a “hollow freedom” when the Wright brothers imagined a flying machine? Or didn’t that ultimately give us all the freedom to fly?

When the SEP authors claim that “Despite the classical compatibilists’ ingenuity, their analysis of could have done otherwise failed decisively”, they clearly have no idea what they are talking about. They repeatedly tie the shoestrings of their “could’s” to their “would’s” and stumble over themselves.

(4) They bring us this thought experiment: There is a girl who has an odd psychological condition, one that causes her to avoid dogs with blond hair. We offer her a choice between two Labrador puppies, one chocolate and one yellow. But, she is unable to pick up the yellow Lab. When she imagines trying to pick up the yellow puppy, some unexpected feeling would occur, and she would immediately reject the idea.

Perhaps a similar puppy had attacked or bitten her when she was younger. But when we ask her why, she doesn’t know. With psychoanalysis we might discover the source of this undue influence upon her choices. And with further treatment her aversion may be diminished, just like any other irrational fear. But, for now, all we know for certain is that she cannot pick up the yellow puppy. It is not an option that is available to her.

Oddly, the SEP authors suggest that this example proves that “could have done otherwise” fails. They seem to think that there is no distinction between a neurotic compulsion and a reasoned choice. They even suggest that their imaginary compatibilist would be silly enough claim that she could have picked up the yellow Lab puppy despite her neurosis.

3.4 The Lasting Influence of the Conditional Analysis

We’ve demonstrated that “could have done otherwise” is compatible with perfectly reliable cause and effect (determinism). Within the domain of human influence, our ability to imagine multiple possibilities and the likely outcomes of different choices, is the very mechanism by which the single inevitable future is causally determined. It’s us, and we’re doing it.

The process relies upon conditional logic, not “counterfactuals”, but real possibilities that might in fact be actualized if we choose to implement them.

4. Three Major Influences on Contemporary Compatibilism

4.1 The Consequence Argument

The consequence argument is more wordplay attacking the ability to have done otherwise. But it uses different language that opens the door to a key clarification. Let’s quote their logic here:

1. No one has power over the facts of the past and the laws of nature.
2. No one has power over the fact that the facts of the past and the laws of nature entail every fact of the future (i.e., determinism is true).
3. Therefore, no one has power over the facts of the future.

The issue is this: How does a person gain the power to determine the future?

According to the Consequence argument, all of the power required to determine the future resides in two places: (1) “the facts of the past” and (2) “the laws of nature”.

If that were the case, then what would we be? Well, we would be (1) a current fact and (2) a natural object. It is not necessary to transcend nature, when we are an embodiment, at least one piece, of nature. Whatever “power” is contained in nature is also contained in us. We are able to determine the future simply by being a natural object, a living organism, and an intelligent species. And we simply go about doing what such an entity naturally does.

One of the things we naturally do is make choices. And what we choose to do next determines what happens next, at least within our human sphere of influence.

The SEP authors say, “According to the Consequence Argument, if determinism is true, it appears that no person has any power to alter how her own future will unfold.” But that’s clearly wrong. Every person is a package of those laws of nature in action, and whatever that specific package of reliable causation controls, the person controls.

There’s a tendency among hard determinists to place “me” in one corner of the room, while placing all the parts of me that control my choices, such as my genetic make-up and my life experiences, my beliefs and values, my thoughts and feelings, and everything else that makes me uniquely “me”, in another corner of the same room. Then they try to convince me that the “me” in one corner is controlling the “me” in the other corner, and somehow “I” have lost control.

The problem is that one of those two corners is empty. And the only way to have a “me” that controls “me” is to embrace some form of dualism. A bit ironic, don’t you think?

4.2 A Challenge to the Principle of Alternative Possibilities

The SEP authors bring up the so-called “principle of alternate possibilities”. It makes the common sense assertion that if a person had no choice, then it would be wrong to hold them responsible. Someone forced to participate in a crime because someone else held a figurative or literal “gun to his head”, is not usually held responsible for his actions.

For example, after bombing the Boston Marathon in 2013, the Tsarnaev brothers hijacked a car, and forced the driver at gun point to assist them in their escape to New York, where they planned to set off the rest of their homemade bombs. The driver was not charged with “aiding and abetting”, because he was forced to assist them, against his will. He was not “held responsible” for his actions, because he had no choice but to obey or die.

The hard determinists (a.k.a., “free will skeptics”) attempt to convince us that reliable cause and effect is some kind of “gun to the head” coercion. But cause and effect in itself is neither coercive nor undue. It is just how everything works, including us. This too is common sense, which is why most people are compatibilists, recognizing both reliable cause and effect as well as our ability to make choices for ourselves of our own free will.

The “Frankfort Cases” introduce examples where we would, by common sense, still hold the person responsible even if it were not possible for him to have done otherwise.

In their example, Jones has decided to shoot Smith. A third party, Black, wants to make sure that Jones carries out his plan. Black secretly installs a device in Jones which, if Jones should change his mind, will compel Jones to carry out the murder. Thus, whether Jones changes his mind or not, Jones will still commit the murder.

Jones doesn’t change his mind, and goes ahead and shoots Smith. The device which Black implanted is never activated. But it did eliminate the possibility that Jones would not shoot Smith.

Because that possibility was eliminated by Black’s device, does this mean that Jones was not responsible? Of course not. Jones decided of his own free will to shoot Smith, without coercion or undue influence.

Now, if Jones had changed his mind, and decided not to shoot Smith, and the device was activated so that it caused Jones to shoot Smith against his will, then that would be an undue coercive influence. The device, and Black who installed it in Jones, would be held responsible, but not Jones.

Note that our operational definition of free will holds, even when the principle of alternate possibilities is called into question.

4.3 Focus upon the Reactive Attitudes

Here the SEP authors introduce Strawson’s arguments deriving moral responsibility from an evolved emotional reaction that occurs when we witness one person deliberately and wrongfully harming another.

One could say, though, that a rational system of justice attempts to replace this emotional response with a morally pragmatic one, one that does no more additional harm than is necessary to repair the harm done, correct the offender, and protect society from further harmful acts.

5. Contemporary Compatibilism

Their final section is a survey of recent authors and how they’ve expressed their arguments for and against compatibilism, each one employing slightly different variations of the arguments listed above. I’ve no intention of tracing down these variations, nor do I feel it necessary to respond to each rephrasing of the same old arguments.

But you may find that the various authors do provide enlightenment on other related issues and are both helpful and interesting books to read.

I’m mainly concerned with the issue at the center of the puzzle. So, let me explain how I came to see through the paradox.

I was a teenager in the public library, reading about the determinism “versus” free will paradox. The idea that everything I did was inevitable bothered me, until I ran across this thought experiment. Suppose I have a choice between A and B. I feel myself leaning heavily toward A. So, just to spite inevitability, I’ll choose B instead! Seems too easy. But then I realize that my desire to spite inevitability just made B the inevitable choice. So now I have to choose A to avoid the inevitable. But wait, now A is inevitable again … it’s an endless loop! No matter what I choose, inevitability always switches to match my choice! Hmm. So, who or what is controlling the choice, me or inevitability?

It had to be me, of course, because I was the only one with “skin in the game”.


[1] McKenna, Michael and Coates, D. Justin, “Compatibilism”, The Stanford Encyclopedia of Philosophy (Fall 2018 Edition), Edward N. Zalta (ed.), forthcoming URL =

[2] See section titled “Duress and undue influence” under “Contract” in Wikipedia