This post provides my account of how choices are causally produced given that libertarian free will is the correct view of how we make choices.
For our purposes, Libertarian free will says that the prior circumstances of a choice do not necessitate that choice. This formulation follows from the guiding libertarian thought (in my view), that even if all prior circumstances of a decision I freely made were the same, I could have chosen differently. Let me explain the meaning of libertarian free will.
First, the prior circumstances of a choice are all the elements that causally produce the resulting choice and that are not identical with the choice itself. I also mean to exclude any causal elements that produce those causal elements that produce the choice—that is, I’m not concerned with a choice’s entire causal history from the dawn of time, but just those elements that are the immediate cause. For example, if feeling hungry influences or is my primary reason to get up and eat something, then feeling hungry is to be included as a part of the prior circumstances of the choice to get up. Any number of things can be included in a prior circumstance (examples include a ringing doorbell, feeling tired in addition to hungry, just having the physical ability to get up, and much more).
Next, necessity will be understood as a relation between the prior circumstances to the resulting choice that is visible or manifest for further causal interactions—for example, a choice to get up. In the context of libertarian freedom, necessity means that, given the prior circumstances, the choice that is made has to come about.
Now, I know what some of you philosophically savvy folk will say to this. You’ll say something like: “Oh yeah, well what if I block the choice with my sophisticated brain laser!—not so necessary now, is it!?” The short answer to this is that I mean to fit prior circumstances into a causal context that would exclude blockers of the choice made, such as that of your laser (This is accomplished by specifying the causal context such that the laser or its beam would contradict that causal context. You can read the longer version of this response here). However, the objection perhaps warrants making explicit that such blockers, called causal defeaters, are excluded:
Necessity means that, given the prior circumstances that are embedded in a causal context that excludes causal defeaters, the choice that is made has to come about.
With this in mind, the libertarian says that, in freely choosing, given the prior circumstances, the choice made does not have to come about. That is, the choice can be different, given those same prior circumstances. And so the libertarian rejects necessity, so defined.
If this sounds causally impossible, then this post is for you. There is a ready model for such a causal relation in indeterminism, into which libertarian free will can entirely fit. I will try and clearly show this, but before I begin, perhaps some people, even libertarians themselves, take the notion of libertarian free will to end at the view that there is no causal relation between the prior circumstances and the choice made. However, this is not the correct outlook to have, and can be a reason why libertarian free will sits so uneasy with many people (including philosophers). Let me piece together the proper understanding of the causal structure of libertarian free will.
To begin, let me show how indeterministic causation is not the same as no causation. G. E. M. Anscombe showed this by her geiger-counter-hooked-up-to-a-bomb illustration (a thought-experiment she credits to physicist Richard Feynmann).
Consider a geiger counter that is connected to a bomb, which is attached to a bomb that will explode only after a certain measurement of radioactive material is made. Let’s say that at a specific time, the bomb explodes.
There’s no doubt that the explosion was caused. The radioactive material caused radioactivity (or was radioactive, if you prefer), which caused the geiger counter to measure it according to how radioactive it was, which caused the bomb to explode for being the right measurement to set it off. Nevertheless, that the bomb would go off at just this instance, instead of a few moments before or after (depending on the half-life of the radioactive substance—let’s just say it’s long), was not determined to go off at that time prior to its going off. That is, repeat the same prior circumstances (by setting up circumstances that are type-identically the same), and the bomb could go off at a different time. (This, as far as I know anyway from my layman’s perspective, is how sub-atomic phenomena are taken to be indeterministic in the first place: the prior circumstances (experiments) are type-identical, and yet they yield different results. And the causal context of the experiment is considered type-identical when its results are probabilistically constrained in the way it is known to be regardless of its context. I am excluding deterministic interpretations of quantum events by default, not just because there is plenty of sound scientific support for the indeterministic view, but that, regardless of this, indeterminism in quantum events doesn’t have to be true in order for understanding to be gained for how libertarian free will could work. My purpose here is mainly to provide the model, and only incidentally, if at all, to argue its truth.)
Avoiding the details of indeterminism’s causal structure for now, notice how under indeterminism, the bomb’s going off at the specific time it does did not have to come about. That is, given the prior circumstances, the bomb’s going off at a particular subsequent time does not have to come about. That is, it is not necessary, given the prior circumstances, that the bomb goes off at a particular time. Notice how this mirrors the free will claim, that in freely choosing, given the prior circumstances, the choice made does not have to come about, or that, which is the same, the prior circumstances of a choice do not necessitate that choice. So this is the first substantive result, that libertarian free will, as stated, is already a form of indeterminism, and subsequently should not be confused with the view that a resulting choice lacks a cause. This is key to understand, as I’m guessing there are libertarians out there who claim this very thing. If such views are out there, let me make it clear that I diverge from them on this point. Let’s now dig deeper into the causal structure of libertarian choosing.
Recall, in brief, how I understand indeterministic causal relations to work (the lengthy version of what I cover next is found here). Take the radioactive material hooked up to a geiger counter. Given a moment of time, let’s say that the radioactive material will cause a certain range of radiation to to be emitted, to varying probabilities. There are some readings around the edges of this range, low or high, that have a fairly low probability, let’s say, while a reading in the middle of the range has a good chance of occurring.
The way to understand the causal structure here, I believe, is to view the radioactive material as necessitating a probabilistic effect that is this very range of readings with its probabilities for occurring, all of which add up to 1 in guaranteeing that one and only one of the possible readings of the range will come about. Now, this is what is necessitated, this very range and its accompanying probabilities. And, just to repeat, such an effect further necessitates that one and only one of the probable or potential effects occurs. So when one reading subsequently comes about, it is caused by the probabilistic effect that was necessitated by the radioactive material. But the ’caused by’ here is not for any other external element landing on that particular amount of radiation, but because it is just in the nature of the necessitated effect to land on one of its probabilistically constrained results (and with the probability with which it is constrained). There is no further reason to be had for why a particular reading results.
That objects have what we may term internal causation should not be controversial. This just means that, regardless of what is causally interacting with an object, it is nonetheless constituted such that it will behave in certain default ways. Sometimes it is put that an object has a certain nature, and an internal nature would just be whatever its nature is when it is not causally interacting with anything outside of it. There is, furthermore, the notion that some objects are atomic—that is, that they are not further composed of simpler elements, or, that for an object, there exists a simplest element that composes it. Such atomic objects would also have their own internal natures. And what explains why it has one nature as opposed to another (e.g., perhaps the basic nature of one is to move around in a small circle) is little beyond that it is in its nature to be one way as opposed to another (with the possible exception of causal origins of such a nature to begin with—e.g. found in God or the Big Bang). Such is the level of explanation for indeterministic effects. It is just in the nature of such effects to settle on one occurrence from a range of possible occurrences that are probabilistically constrained.
The Indeterministic Production of Free Choices
Now to fit libertarian free will into indeterministic causal relations. Let’s say that I consider changing my course of action. For example, let’s say I’m outside walking, and I consider turning around and heading back home. And since we’re examining libertarian free will, let’s say that I’m free in the libertarian sense to either choose to continue or to choose to head home. So what would this look like? How can we understand the libertarian sense of this choice?
First, to make the picture more clear, let’s say I have good reasons for doing one or the other, and that my desire to do one over is best described as ambivalent. My reasons to continue include that I haven’t exercised at all in a few days, that I can benefit from stretching my legs some more, that I am already in a good rhythm, that it’s as good as doing anything. My reasons to head back include that it is getting hot out, that I am tired, that I can always walk more tomorrow, and that it’s as good as doing anything (this later one is the perfect reason for every list of reasons). Let’s also say that neither my moral principles nor my political ideology provide any good reasons to choose one side over the other. So each side has good reasons, as sometimes occurs when we make a choices.
In making a choice, we are confronted with two or more possible courses of action to take. We are nudged, influenced, or cajoled by various reasons or desires to act in one direction, or they sort of come out even as in the case I’m outlining for the moment. More specifically to my illustration, I have two possibilities to choose from, only one of which is to be chosen. The next question to answer is, in general, do the possibilities that are entertained by the chooser have a probabilistic constraint? I cannot conceive of a cause without some constraint that bears on how things turn out for its effect, since to suppose no constraint here implies no causation at all—I won’t get into this too much here, but if you want, read more on necessary connections in causal relations. Perhaps a causal relation with no constraint on how things turn out from the cause might be thought to mean that any possible event whatsoever may be produced by that cause. But if this were so, then we’d have to answer how likely it is that a particular event occurs as opposed to some other. I suspect the most chaotic formulation would give equal weight across every possible event. Yet this result would constrain what occurs by the cause to favor every possible event equally, as opposed to some other probabilistic distribution. We should therefore conclude that, yes, the two possible courses of action that are before the chooser are to be understood as probabilistically constraining in terms of which one it is that occurs. How are they probabilistically constrained?
The probabilistic constraint, to my estimate, fits nicely with the thought that our feelings and reasons nudge us along during the course of decision-making. Sometimes they even do so in conflicting directions. Furthermore, such feelings and reasons are not taken, absent some prior theoretical commitments, to determine a singular course of action, but, when taken together, suggest multiple possibilities. How greatly feelings and reasons favor one side or another lends itself well to a probabilistic interpretation: when one side has stronger reasons or desires, then that side will be more likely than the other to occur. It is merely more likely, as opposed to inevitable, because such feelings and reasons, even when they show their greatest favoritism, never seem to play their hand in a way that makes the favored outcome a singular inevitability.
It may be difficult to put an exact probability for a choice given a set of feelings and reasons, as well as how they cancel or amplify one another. But this shouldn’t indicate that the material basis for the feelings and reasons are not without some probabilistic contribution (for example, as quantum or other indeterministic events of the brain). This will depend on how we are to identify feelings and reasons, particularly as they causally manifest in the world. The claim I make is just that the probabilities involved in feelings and reasons that we sense from the first-person in everyday decision-making are causally manifest as probabilities of indeterministic causal relations toward a set of choices. This isn’t to say that we can infallibly assign probabilistic numbers from a first-person perspective, but perhaps we’d be able to do so more or less reliably anyway. After all, it’s plain to see how when we desire and have reason for one choice over another it tends to happen, despite some potential hick-ups here when acting irrationally or apart from one’s best interests. That even when we strongly favor one side it does not seem inevitable is, I think, indicative of other reasons and feelings that favor different options, although perhaps negligibly so. They are never to hard to find with quick reflection. Probabilities can be reliably assigned from the first-person, but the final verdict is always in the causal manifestation itself of reasons and desires as they probabilistically constrain one choice or another (perhaps there can be a sense that the two are one and the same). The way that the idea, estimate, or feeling about what the likely choice is can come apart from what the likely choice actually is is just when the idea, estimate, or feeling does not render fully explicit to the experiencer the probabilistic distribution that it causally imposes.
So reasons and feelings provide probabilistic causal constraints on the possible outcomes that present themselves before us. Are these the source of all probabilistic constraints of a choice? Perhaps not, but I don’t know of any alternative source. And does every possible outcome that we entertain need some reasons or feelings for it to be? I’m not sure, but a choice that is devoid of any reason or feeling for it would seem to be, roughly, an imagined scenario for which there is not even the consideration of choosing. So if someone tells me to fly to New York, slap a street mime, and fly back, I would think it was just a strange consideration, not something I’d even be remotely considering doing. And the reason for the lack of any possible free action on my part for such a choice just seems to be that I have no reason or desire to do such a thing. Maybe I want to visit New York. But I wouldn’t intend to carry out the side mission.
Returning to my walking illustration, we now see that, even if I had stipulated that my desires and reasons were an exact 50 / 50 split in probability between nudging me home or suggesting that I continuing walking, we don’t require that they be split evenly in order for the choice to sometimes be different given these same prior circumstances (as with the geiger counter example, which can be set up so that, even if the bomb blows up at some particular time, it was quite unlikely that it would occur at that time given its prior circumstances).
Let’s say I decide to go home. When I do decide after all to head back home, I have all sorts of reasons to do so. I think about how hot it is, how I’m tired, and how I can just run tomorrow. (What a loser, he should quit whining and get in shape.) And I also think that, everything else being equal, perhaps I could have gone on with my run. Oh well, what’s past is past.
The indeterministic causal structure of decision-making shows how it can work that these same prior conditions can produce a different choice. But how is the choice produced? An astute critic might readily think that I’ve just made choices arbitrary and accidental. Let me address this concern.
The pool of choices is not arbitrary and accidental, since they are (always as far as I know) tied to one’s reasons and desires. And reasons and desires should not be understood as accidental (lest this same criticism be applied to the compatibilist position).
The astute critic would likely respond that they did not mean that the desires, reasons, or even the resulting pool of potential choices was arbitrary, but rather that what is arbitrary is that, given such desires and reasons, the choice that is produced is the one choice as opposed to the other.
But what, exactly, is arbitrary here? An agent chooses some choice C in the end (and so causes it) from a non-arbitrary set of reasons and desires that constrain what pool of possible choices are available, that probabilistically constrain them, and then that remain with the choice once it is made.
If what is arbitrary is considered to be just that a singular choice is not necessitated given the prior conditions, then the term is merely acting as subterfuge for indeterminism. But what is arbitrary in the context of a choice one has reasons or desire for making? If the thought is that people sometimes freely choose things that lack reasons and motivations, well what’s so unheard of about that!? Moreover, everyday experience would seem to favor the indeterministic outlook. The universe does not show up in experience as landing inevitably upon one choice at every fork in the road we consider, and the choice certainly does not appear as being set from the beginning of the universe (or some faraway back time). Indeterminism is the model to look to to validate such notions of a universe open to different possibilities. Determinism must end with the view that the experience of a universe open to different possibilities is illusory (of course, this does not in any way prove that determinism is false, just that some of our experiences of making choices are illusory—explicitly, sometimes a choice that seems possible to make is sometimes not actually a possible to make for its prohibition by prior circumstances. This is about as straightforward an illusion as you can get, but it’s not the death-knell for determinism. “So what if some elements of experience are illusory?” the determinist can say, “Big deal!”).
That one choice is made rather than another should not be viewed as arbitrary under the indeterministic model. A person’s taking up a choice sometimes seems to be the primary reason that one possibility is chosen rather than another. The taking up of a choice in the probabilistic possible choices that are considered should be considered the active force of choice-production. A choice is, after all, taken up. And why is the choice taken up? As we have seen, there are reasons and desires, but, in addition, there is just the reason that it’s the choice I took up. It’s just what I chose. This ability to take up of one and only one choice from amongst the possibilities is what constitutes the will. This is about as clear a conception as one could hope for, but perhaps I can say more.
One difficulty with this view of the will is that I already said that the reasons and desires set probabilistic constraints on the possible choices. This would imply that the will cannot take up a choice that would contradict this probabilistic constraint. So if one choice is only lightly favored, then the will cannot take it up over and over again, lest the probabilities come out reversed, or contrary to the ones that reasons and desires set. One heavy-handed response to this implication is to say that new probabilistic distributions show that the reasons and desires must have changed. However, there is no reason to suppose this is so other than just to uphold the model as so far presented in an ad hoc manner. Alternatively, we may just be content not to expect the will to be capable of violating the probabilistic constraints set by reasons and desires. Either way, that reasons and desires are the final determiners of the probabilistic constraints would spell trouble for the notion of the will as the primary reason for a particular decision (the I chose it because I chose it mentioned previously). The primary reasons would have to always fall on the totality of reasons and desires to act.
I do not have all of these notions set in stone. The good news is that I the structure of indeterminism leaves ample room to modify the way everything fits together. Perhaps the answer is that the will rather sets a probabilistic constraint of its own in its making a choice. That is, the will itself bears on the probability that one choice comes about or not. It may be that the only necessity that falls upon the will before it chooses is that one and only one of the possibilities (which are set by reason and desire) is to occur.
So perhaps we should hold fast to the notion that the will takes up the particular choice it does simply because it is in the nature of the will to cause one of the possibilities. The explanation would bottom out in this basic nature of the will to have such a causal ability. If this be cogent, which I don’t see a reason to think it isn’t, then the picture goes that the reasons and desires set up the possible choices, while the will just picks one. It can both be true that every choice has its reasons or desires for action, and that every choice is primarily brought about for being chosen. One difficulty with such a picture is that I have lost some clarity provided by reasons and desires alone as contributing to the probabilistic distribution of the possible choices. In that version, it was clearer how, causally speaking, reasons and desires nudge the will in different directions. Yet if now the will overrides or sets the probabilistic distribution in choosing, it may be less clear how reasons and desires causally contribute to the choice that gets selected, beyond setting up the initial options.
Or is it? Perhaps reasons and desires influence the will toward certain possibilities in a way that results in a certain likelihood, but that this resultant likelihood, which would be revealed by a frequency of trials, actually gets passed off to the will first in a serial causal chain that then is moved in a certain probabilistic way, to be sure, but which also has overriding powers. That the will can override the probabilistic constraints set by reasons and desires for what occurs could be a feature of the will. So the probabilistic constraints would be there, as our desires and reasons seem to be when engaged in decision-making from the first-person perspective, but they can be overridden by the will in the final result of what choice is made.
Perhaps an additional way to think about what I am proposing here is to think of desires and reasons as contributing to the probabilistic constraint in ways that are fixed. Will, on the other hand, is not to be considered as fixed, but dynamic. The will can change the probabilistic constraint itself, and thus its contribution to that probabilistic constraint, just for taking up the choice regardless of what reasons and desires favor. When does it do this, and what is the picture more clearly? Well, that’s just it: giving will this causal ability would indicate that it cannot be given when (i.e. under what conditions) the will modifies the probabilistic constraint. But the causal capacity of the will just includes that it can take up a choice just for taking it up. Such a causal capacity bottoms out in the constitution of the will itself. It is just in the nature of the will to be capable of this feat. In the end, reasons and desires constrain the options of choices to make, they move the will in certain directions to choose, but the will has an overriding say in which of the options plays out.
There is a theoretical problem concerning the possibility of measurement given our model. If it’s true that free actions cannot be had without all three components: reasons, desires, the will, then we could never measure the individual causal contributions to the resultant choices (or their probabilistic constraints). This is just because we cannot remove an element without also removing the possible choice, and removing a choice could not indicate how probable it was to occur for the other two non-removed elements contributions. That is, we could not measure a difference between the probabilistic contributions of all three and the probabilistic contributions of just two of them in order to get the total probabilistic contribution of the one element removed. However, if reasons, desires, and wills are possibly transparent in their contributions to first-person experience, then a measurement would be possible through first-person experience.
The indeterministic causal model of the will fits well with the libertarian notion that we sometimes choose differently, given the same prior circumstances. I have tried to show how it can work that such a will can produce a choice. I have even tried to provide a causal model for the notion that I make a choice primarily just because I made it, and not primarily for a particular reasons or desires (even if they are always present).
Such notions of the will do not have to lack a basis in material reality. Indeterministic notions have an established use in science to explain quantum events. Hopefully I have not only shown how indeterminism is quite capable of causally explaining the production of choice in libertarian free will, but also that this possibility should open up the possibility of empirical verification. That is, however, another topic. And I don’t think I have exhausted the possibilities for giving a basic model of the indeterministic causal production of choice. I have maybe only just scratched the surface.
Several follow-up questions interest me, including: How does consciousness relate to the will I’ve described? Is conscious experience possible for those without a will (e.g. for those that have merely singular wants and desires)? Can consciousness follow (causally or logically) from having a will? Is it possible that every indeterministic system in our universe has a will?—And, if not, then under what conditions do indeterministic systems have a will and under what conditions do they not have a will?