Choosing a compatibilist free will perspective

brainby Dwayne Holmes

[This essay is part of a special “free will week” at Scientia Salon. The Editor promises not to touch the topic again for a long while after this particular orgy, of course assuming he has any choice in the matter…]

Despite the question having been around forever, the topic of Free Will (FW) has been pretty hot lately, including several entries at Scientia Salon [1-4]. Personally, I find the philosophical aspect of the topic of Free Will a bit pointless (either we have it or we don’t), but references to questionable “findings” in neuroscience and claims of how FW beliefs impact ethics and social policy do interest me (quite a bit) and force me to beg everyone’s patience for yet another essay on FW.  If you will, think of this as an entry on how not to interpret findings from neuroscience (or biology).

Let’s begin by cutting to the chase on the matter of definitions. First there is Libertarian FW (LFW), which holds a dualistic view of mind and brain and defines FW by the causal relationship between the two. Minds are situated somewhere other than the physical brain (perhaps being immaterial) and so can make choices independent of physical states related to the brain.  You could say, “mind writes to body” and so it is as free an agent as you can get.

Then there is Anti-Free Will (AFW), often known as hard determinism. AFW also defines FW by the causal relationship between mind and brain (with a distinct nature for both), which is why I consider AFW a form of “scientific dualism.” In this case, the brain processes inputs/outputs based on physical conditions that are deterministic and so (they argue) exclude choice and FW. In contrast to LFW, AFW proponents would say “body writes to mind,” trapping mind as an epiphenomenal rider, a transitory set of hallucinations (a sense of control and freedom of choice) that are mere by-products of normal brain function.

Finally, along comes Compatibilist FW (CFW), which maintains that the mind and brain are essentially one and so defines FW based on the relationship between an intentional agent’s natural potential to obtain personal desires and their practical ability to do so. An agent’s natural potential can be restricted in practice due to physical impairment of normal processing in the brain (i.e., tumors), or the activities of other agents. Since multiple agents with cross-agendas are an extremely common phenomenon, FW is commonly interpreted as one agent’s freedom from coercion/manipulation by others.

Note that I am leaving out the kind of FW-skepticism (FWS) described by Gregg Caruso in the last two FW essays at Scientia Salon [3,4]. This is because FWS seems to fit somewhere between CFW and AFW (or CFW between FWS and AFW).  Since I am still not confident about discussing FWS positions, it is up to such skeptics to figure out where they stand on my descriptions/arguments.

Redefining decisions as having no choice

Before getting to science, I need to address a definitional issue that tends to sidetrack discussions of FW and moral agency. While AFW and CFW agree on physical determinism, they may disagree whether “choice” exists in a deterministic Universe. While there may only be one choice possible at a certain moment given all inputs (as AFW proponents argue), isn’t that exactly what one would want from any decision-making system (ask CFW advocates): one choice (hopefully the best) from all alternatives?

According to CFW, in a deterministic system FW involves the ability to process a sufficiently large amount of information from potential inputs such that one can define the best choice from sufficiently large numbers of potential actions to obtain one’s desires (aka actualize one’s will).  Lack of FW (or loss of agency) involves the absence or reduction of potential by restricting possible actions, information, or processing capacity such that agents can’t manifest their will as they would under normal conditions. From this perspective, the locus of control in decision-making (choice) is placed firmly within a brain’s processing capacity and its potential to actualize the will of the agent, except where other agents, direct physical impairments, or dramatic historical events significantly reduce that capacity.

AFW proponents (and some FWS’s [3,4]) maintain that this processing of information to actualize desire does not result in a “true” choice but merely “selection,” since in some “ultimate” sense all prior events leading to the selection of an action had to happen, making an agent’s selection pre-determined. Notice that this notion of an “ultimate” sense of choice subtly shifts the locus of control from the potential provided by a brain’s processing capacity, to the factors it had to process. Your brain does not process inputs; inputs process your brain. Thus an intentional choice (even if vastly more complex in its mechanics) is no better than a falling rock changing direction because it happened to hit another rock. Both simply had to go in the direction events demanded they must. Such proponents often go further, pointing out that even if selection of an action could be considered a choice, the desire (and processing capacity) underlying the choice was not itself chosen. One is still a prisoner of one’s nature (a “will” and decision system constructed by prior events), and so all choices are chained to that (not to you).

Fair enough. Placing the locus of control is largely a matter of perspective, and so it is not objectively wrong to take either view. However, if this “ultimate” view of choice is accepted, no one can ever be said to “truly” choose or exhibit FW, even in principle, as the level of freedom demanded by this perspective is not even granted to Gods!

Let me unpack that claim and its consequences. Gods are presumably eternal and live beyond time and space, always existing. That means they always had a nature (Will) and it was not something they could have chosen for themselves. So, whatever nature a God has, its relation to that nature has to be the same as humans to the nature they were born with: their initial desires were not of their choosing, and can only change based on subsequent events shaped by their initial desires.

Then, at some point beyond time and space (whatever that means), Gods do something. Once a God pushes that first domino (a product of its unchosen Will), the next act must either be wholly random or in response to that first cause-effect event. All that follows (until the end of that God’s existence, if any) must inherently be a response to the inputs that God receives from the unwinding chain of causes and effects. Otherwise, a God is nothing more than a random number generator hooked up to an action lever.

Thus the concept of “true” choice advanced by AFW advocates (free of all prior events) is a myth. It is an impossible expectation, whether one has a physical brain or a mind composed of magical energy beyond time and space. It might as well be called “fictional” choice, because a choice fitting those criteria will never be found.

All real choices require a confluence of events impinging on an actor such that a decision is mandated. That would seem to be a defining character of any choice, and LFW proponents do not argue otherwise. What they argue is that minds as nonphysical entities are impervious to certain physical events, while the nonphysical mind can cause physical events, and (let us not forget) nonphysical events can affect the mind and the physical world. They do not demand choices be context free.

It is then questionable to use the very defining characteristics of a choice as criteria for excluding something from being a “true” choice. It also smacks of a true Scotsman fallacy. And in any case, it seems unnecessary to use a fictional entity to bring down a (arguably mistaken) theory, particularly when that theory does not use or demand such an entity.

The following sections explain why, when we move beyond fictional capacities for making “ultimate” choices (that even Gods do not possess) to factual capacities involved with making real choices, CFW accounts bring something to the table for discussions of FW and moral agency.

Mistaking Experimental Signals for Decisions and Awareness for Mind

Proponents of AFW like to reference data emerging from neuroscience. Studies, from Libet onward, purportedly show that the brain is making decisions before our mind is aware of them, thus nixing LFW outright and even suggesting that brain ≠ mind (thus nixing CFW) [5].

I am not going to address these studies directly as that would take a small book. Suffice it to say that the correlations found do not necessarily license the conclusion that a point of “decision” was identified. And even if correlations were sufficiently high to suggest a causal relation exists between a specific brain region/activity and a specific decision/action that would still not undercut CFW… or even LFW!

That last point might be surprising, but on reflection it should be obvious. Just because the (disembodied) mind feels the decision later than a physical manifestation in the brain does not mean that the brain caused the action.

Taking a devil’s advocate position on behalf of LFW, the brain region identified could be the spot where the mind connects to the physical brain (like a string to the puppet). That the mind only “feels” the decision later is perhaps because its awareness is so linked to the body by habit that it has an errantly delayed sensation of the self. In short it confuses the physical sensation of a decision imposed on the physical brain (when the brain feels it) as being the “real” point of having made the decision (even though it was made earlier, elsewhere).

Or, it could be argued the other way around, that the region found is the “string” of sensation leading from the body to the disembodied mind which really does make the decision (later than the recording in the brain region). It is just that we can reliably match powerful “sensations” from that region, to acts the mind is likely to make. This is analogous to finding sufficient nerve activity in a toe (by a fire), before the foot is retracted. Did the nerve in the toe make the decision? No, but it is a reliable predictor of what the body will do next [6]. And the retraction may even come before awareness of the pain.

Both LFW explanations for these findings align with the existential problem most LFW theists seem to have with the mind, that it is so drawn to the body by habit that it has lost awareness of itself. Hence the need to separate one’s true self (the mind) from bodily sensations.

Of course I’m not advocating LFW since there is plenty of evidence against disembodied minds from less demanding experiments than timed neural activity during decisions [7]. But I do like CFW.  So what do I make of those studies, assuming for the sake of argument that their correlations were meaningful regarding causation? My reaction is that it is interesting information but so what? What would a neuroscientist expect to find other than signals in a brain correlated with decisions?

On the issue of mind vs brain, just because the brain might process a decision for action before processing it for conscious experience, does not mean that the mind was not involved in the decision process. That is to conflate mind with conscious awareness, which CFW can reject as it links mind (or at least “you” as an agent) with the physical brain. Some, arguably most, decisions regarding actions are likely to be processed subconsciously, with those that are more important or require greater consideration/attention making it to higher levels of awareness.

Lack of awareness for some decisions would not negate the importance of such decisions (like they are done “without thinking” or more importantly “beyond your will”), nor does it suggest that conscious awareness has no influence in all decision-making. One criticism of such studies (and Libet’s in particular) is that they often do not demand much in the way of having to deliberate on a set of options that are important for you or your future [5]. Libet-style experimental settings don’t elicit much concern, and so do not trigger the need for higher order consideration on what to do.

Still, we shouldn’t assume that scientists have actually observed decisions being made prior to conscious awareness, rather than detecting elements that must be present prior to making the decision or registering awareness. For example, the brain region responsible for the message “press button” might be active a while before a command decision “ok” (made in another region) allows it to go ahead (perhaps by relieving repression of the signal at a third brain region).

For experiments keyed to changes in signal strength what may be detected is the precursor of a specific action as it is offered for decision (which would change over time), rather than the area that is consistently active in making decisions (and so presumably holding a near constant signal). Or (for any experiment) perhaps it is the area with the greatest coherent signal (or change in signal) that makes it likely the decision encoded will win out over other options being offered for deliberation across the brain.

In both cases it may take time to “tally” the inputs, via activation/repression over different networks, before the system “recognizes” the final decision so as to report it (which it should be noted is a completely different action than simply acting on the decision itself). The brain is not a single wire affair, instead displaying supplementary systems and feedback loops running across many regions. A decision, then, may be best understood as a set of activities, which might or might not include awareness, rather than any single signal correlated at some earliest time point with some subsequent action.

This can be analogized to how sounds are processed by the brain. A sound close to one ear will enter that ear first and begin to be processed by the brain before the same sound enters the opposite ear and begins to be processed on that side. Yet you hear the sound only once. How?  Mechanisms within the brain can modify signals (strength and speed), and perceptions are not necessarily reported in absolute real time (first come first serve). There can be a delay as a “representation” of the experience is constructed by hashing out signals. In this case you “hear” a construct of one sound with an added recognition that the source is located closer to one side than the other.

Presumably experiments would detect activity in the neurons that process information for the ear closest to the sound first. That would not mean that the “decision” the sound came from that side was “made” by those neurons. Rather, organized processing further down the chain in another region, using all relevant signals from both ears, creates the “decision” regarding location. (For those curious about how the brain accomplishes the feat of sound location, it involves two separate regions (the medial and lateral superior olive), which use two separate comparative criteria (timing of reception and loudness) [6].) It stands to reason that decisions of what to do could have equally complex decision matrices over diverse locations.

To be clear, I am not dismissing scientific investigations of decision-making, involving timing or locations of signals. They are important in trying to understand how the brain works. What I am trying to point out is that it is premature at best, and completely missing the target at worst, to use this data to advocate AFW positions on FW or choice.

Losing the mind for the neural activity

Hard determinists like to reduce everything to baseline causes and effects. While that is wonderful for accurate discussions of simple reactions, it is not as useful when discussing reactions taking places at higher levels of organization. For example, it works for how ion channels operate on the surface of a neuron, but not as well for how someone reading a book will come to understand (or feel about) a different culture, or decide the best way to install a new sink.

The human brain is an evolved system, with brain cells (both grey and white matter) networked in a way that their associations build representative models of the world around them. Learning involves the construction of extensive, integrated models of physical entities, internal states, and abstract concepts, by physical embodiment in cellular networks.

This system can identify objects and project potential responses based on their natural relationships in the world, and not merely “because neurons fire.” That means unless one is a neuroscientist studying specified brain activity, one is arguably better off discussing thoughts related to choice at the level of what the brain depicts rather than the biology lying below it, or one misses the “true” or functional cause-effect relationship in play.

Admittedly neuroscience has revealed that our awareness (external and internal) is not as great as we might believe, and that much of our decision-making is influenced by or takes place within subconscious systems. In other words, the model we hold of ourselves as agents in the world is not complete or wholly accurate. But humans were aware of that long before neuroscience emerged [8]. And more importantly, that does not remove the fact that, subconscious or not, the brain is acting on collective representational information and not simply on the fact that “a bunch of neurons are firing.”  Even the “tricks” mentioned by some determinists, like the smell of baking bread or certain colors influencing decisions, describe associations feeding into relations considered at the level of the represented subject.

While neuroscientists discover how the brain accomplishes these feats through the activity of neural networks (and that is important), that should not be confused with or used to dismiss what feats are being accomplished. Indeed, by limiting one’s interpretations to brute neural activity one fails to recognize the model of the self (mind) being created and utilized in meaningful decision-making by that neural activity. This would seem to be a crucial oversight!

So what is being accomplished? I would argue that the human brain’s evolved capacity to build extensive, complex, and highly interactive models of the world provides enough degrees of freedom regarding potential actions to warrant the term “choice” and so a FW that is worth talking about. This is especially true as the brain runs simulations that allows the prediction of future events and so the generation of novel behavior to obtain goals. This is a point AFW proponents seem to downplay in discussing choice. Human brains are to some degree capable of moving outside immediate input-output causal relations, by generating their own inputs (of estimated future events) for consideration in a recursive feedback mode.

This is arguably why people take time and seek relaxing environments (even sleep) before making important decisions or while working out complex problems. Sleep (some neuroscientists claim) helps consolidate experiences for better model building and resting states allow for more free association of thoughts with minimal distractions (external inputs). Time spent in consideration (simulation mode) potentially reduces the effects of the irrelevant, background noise of immediate experience (fresh bread, colors) in our decisions.

And the extent of human agency doesn’t end there. Our new fangled brains have evolved such sophisticated simulation capabilities that they can generate novel information by mixing and matching experiences to produce entities and associations that we can’t directly experience, have never existed, and possibly can’t exist (yet are captured in our cellular networks). In short, we exhibit creative thought.

Wholly fabricated entities and associations can develop and be considered in the mind, and manifestations of those things created in the world where they certainly never existed before.  Whether real or not, it becomes a practical reality as if they did exist. People can sacrifice large portions of their lives to these things, and force others to sacrifice their lives to servicing them as well. From religious beliefs to popular works of known fiction (such as a scifi TV show), people can be so inspired by events within their minds that they shape the world around them (and so the future of the world) in their image.

And let’s not forget that these “fabricated entities” are not always wholly fictional! In addition to theoretical concepts about the world like general relativity and quantum mechanics (which are not directly observable but result in actual technology) we have engineered artificial elements, molecules, and even living organisms based on our imaginations to suit our interests.

With creativity, humans extend degrees of freedom of thought and (to a limited extent) action beyond their immediate environment, even beyond reality itself. This, if nothing else, should grant credibility to a concept of functional CFW. Human agents are more than simply the sum of their experiences, because they process experiences into novel beliefs and potentials not possible without their personal agency. One might say this presents a new option for causal relationships: “mind writes to the world and across other minds.”

After considering the how and what of brain activity, one might ask why brains developed (or require) this level of function?

While speculative, it seems likely that our vast processing capacity (with real world modeling and creative thought in support of decision making) developed as part of an “arms race” of agency. Dan Dennett suggests this sort of pressure for increased agency in a recent interview [9].

In a world of inanimate objects, there is no need for advanced processing, much less a sense of agency. Once systems emerge that are capable of reacting to the environment, better processing to enhance response naturally evolves. A crucial problem is then posed by the need to distinguish these reactive, animate objects from inanimate objects in the environment. Animate objects can be modeled with how they interact with the world (behaviors), which naturally lends itself to viewing others as intentional beings, acting toward goals. In short, we come to ascribe agency, and that sort of modeling easily extends to the self.

Dennett intriguingly posits defense against manipulation by others as an important reason for evolving advanced concepts of (self) intentionality. Agents need to take into account (model) what their own desires and required goals are, and what others might do in reaction to such knowledge, in order not to be manipulated away from obtaining them.

For AFW scolds of CFW, this is where the concept of CFW exhibits explanatory power and utility that AFW lacks. Evolutionary pressures drove increased degrees of freedom (potential ways to act) in organisms, leading to modeling of competitors as agents capable of intentional action (to be opposed), and ourselves as agents with our own intentions requiring freedom from coercion/manipulation by opposing agents.

Our sense of choice (locus of control placed within ourselves) and CFW are part of that model, and arguably an important part of the recursive simulations we run in order to make decisions in support of our own goals over those of others. To view oneself and others as “meat puppets” (as some AFW advocates state we are) places the locus of control entirely outside the actors working in the world and so is unlikely to generate accurate simulations. At the very least it would seem to make such simulations unnecessarily complex and communication about our simulations unwieldy.

Given all this, it is not clear what reason there is to artificially limit discussion of FW to the relative relation between one’s mind and physical brain (arguably an errant — inspirational if not very useful — product of our creative thinking). This is especially true when it is clear that competitive organisms actually exist with varying degrees of freedom in obtaining desires and that our brains have developed efficient modeling techniques regarding that agency.

Take home message

CFW, of course, provides relevant context and meaning for discussions of moral agency and social policy. In fact, it is the only theory whose definition of FW involves interactions between minds, a seemingly crucial part of any consideration of ethical and social policy.

From the point of view of CFW, LFW appears bankrupt given the decidedly strong physical connection between the brain and manifestations of the self. While AFW/FWS concepts of “true” choice are inherently flawed, CFW proponents can take onboard the understanding that prior events shape the course of events that follow, including decisions. Indeed, in considering manipulation by other agents CFW inherently possesses that perspective. As with FWS and AFW proponents, things like luck can be factored in while considering policies, only without the need to disregard humans as “intentional” agents.

The locus of control (at least for major decisions) is best considered internal to the agent due to 1) the recursive simulations and 2) the capacity for creative thought (novelty) that our brains deploy to attenuate external inputs and to oppose manipulation by others in pursuit of our goals.  This is not invalidated just because some decisions may be made subconsciously or because the representative models are not wholly accurate. Using conceptions of ourselves as intentional agents our brains process information into action, including attempts to make personal dreams a reality.

Finally, human communication regarding future action (to ourselves and others) normally references what the brain produces (meaningful representations) and not how the brain produces it. It seems a mistake, or at least impractical, to demand we consider how to the exclusion of what. The strength of compatibilist FW is that it (as the names suggests) it allows one to use both sides of brain activity in their appropriate frame of reference.

Epilogue (fanfare for a cell-based soul or requiem for a meat puppet)

We are not meat puppets, but living, cellular decision engines. These cellular souls are poised at the culminating edge of all events that have come before. The history of the universe has gone as far as you, and while you live your little corner of it won’t go further without a decision made by you, even if that choice is to do nothing at all. From this edge onward, you peer into a wholly unshaped universe. So it is a certain mistake to view yourself trapped like a fly in the amber of a life story already petrified from beginning to end, as dead flotsam pushed about on the tide of history, or an object wholly manipulated by forces beyond your control. You hold a privileged role in this universe’s unfolding, having been allowed to see the possible directions it might go and selecting whatever path seems best to you. What you can choose may be limited compared to your greatest dreams and desires, but wherever your corner of the Universe goes next is unquestionably up to you.

And if, after all, the doomsayers are right and you are fated in whatever choice you must make, fate has to work through you all the same. So you might as well decide that fate favors your best deliberated opinion. Fate will never know the difference, but you will.


Dwayne Holmes is a PhD student studying Neuroscience at the Free University of Amsterdam in the Netherlands, with prior degrees in philosophy and molecular biology. He is particularly interested in how science and philosophy impact our understanding of ethics (from molecules to social norms). He writes about brain and mind at Emerging Mind.

[1] Free will and psychological determinism, by Steve Snyder, Scientia Salon, 21 October 2014.

[2] Back to Square One: toward a post-intentional future, by Scott Bakker, Scientia Salon, 5 November 2014.

[3] Free Will Skepticism and Its Implications: An Argument for Optimism — Part 1, by Gregg Caruso, Scientia Salon, 22 December 2014.

[4] Free Will Skepticism and Its Implications: An Argument for Optimism — Part 2, by Gregg Caruso, Scientia Salon, 23 December 2014.

[5] See this video for a nice quick explanation of Libet.

[6] Principles of Neural Science, 4th ed., McGraw Hill (esp. Chapter 23 Touch, Chapter 24 Pain, Chapter 30 Hearing).

[7] As has been pointed out in prior FW threads at Scientia Salon, the fact that physical alterations of brain regions effect not only perception but also action and personality in predictable ways suggests a strong, meaningful connection between brain and mind.

[8] The description of subconscious activity from the brain was recognized as early as Hippocrates and extensively explored by Freud well before formal neuroscience began as a field of study.

[9] The truth about free will: Does it actually exist?, Salon, 28 December 2014.


Categories: essay

Tags: , ,

116 replies

  1. Hi DM,

    With regard to the AFW (and CFW? ) concept of free will you wrote:

    it doesn’t invalidate moral responsibility altogether, but it does undercut any justification for retribution or hatred.

    I heard this claim quite often from the “AFW camp”, but I am not sure whether I understand it.

    I mean it might certainly be that some (all?) forms of retributive punishment are unjustified, but your claim seems to be that this was shown by a rejection of LFW. It is this latter part of the claim which doesn’t follow for me. It rather seems to me that if retribution is unjustified under CFW then it is even more so under LFW be it in its “miraculous” (religious) form or its natural “randomness” form. After all these views of free will seem to give the perpetrator new (ultimate?) outs. “I tried to respect the laws, but then a miracle happened and my soul decided otherwise.” or “I tried to respect the laws but the quantum dice rolled against me and I couldn’t”. How can you hate on caprice? Against what do you retaliate in this picture?
    On the contrary, it is the proper understanding of CFW that gives license to punishment in the first place, as far as I can see. We typically can understand the rules and factor that in to our behavior. We have that kind of competence. That is why we can take blame for our misdeeds. We are wired right for blame and appraisal, punishment and reward.

    Maybe this still doesn’t license retribution or (cosmic) “just desert”. But how would LFW change that? Could you try to unpack that for me?


  2. Hi Dwayne,

    > the notion that things like consciousness, free choice, and even the idea of ‘me’ are but convincing illusions fashioned by natural selection

    Even Dennett could be characterised as describing conscousness as an illusion. This is simply orthogonal to free will. And when Coyne says free choice is an illusion, this is because he interprets “free” in the libertarian sense of having no external causes. Again, I think your example supports rather than refutes my claim that there is little substantive difference.

    > it has a neuroscientist explaining conscious components to decision making as not happening

    If there exist AFW proponents who believe that conscious processes have nothing to do with decision making (as opposed to being identical to or an epiphenomenal reflection of underlying neural processes) then I agree with you that they are confused and completely wrong. I don’t think Coyne and Harris (or my former self) are such people. Nor is Patrick Haggard (from that BBC programme). What he actually says (rightly or wrongly) is that Libet disproves classical free will, where conscious thoughts are the causes of our actions. By this, I think he means ultimate causes, i.e. independent of preceding neural activity.

    Later on, he even articulates very clearly that what he disputes is classical free will, which he describes as the view that our minds are exceptions to the rule that everything is determined by the laws of physics. The kind of free will defended by Al Mele is CFW, i.e. not that rejected by Haggard. The two men are simply talking about two different ideas by the same name.

    Nowhere does he describe conscious components of decision making as not happening.

    > Why wasn’t the one removed attempting to provide an alternative (now cancelled) for the one remaining?

    Because of history and everyday experience, LFW is the “default” position and what many people understand by the term “free will”. The familiar illusion is that we are completely free to choose — all choices are totally and ultimately possible, none are determined. The problem is that this idea (LFW) is untenable. AFW reacts to this by calling it an illusion, that free will doesn’t exist. CFW reacts to it by trying to explain what is actually happening and calling *that* free will.

    > That is to use the fact that there are similar sounding words to mask the definitional differences.

    And that is the heart of the disagreement. Honestly, I think few AFW proponents would have a serious issue with CFW if they scrupulously always prefaced the free will they believe in with “compatibilist”. Likewise few CFW would have a problem with AFW if the latter always clarified that they disbelieved in “libertarian” free will. The problem is that both camps have different ideas about what actual free will (needing no qualifiers) is.

    > You are incorrect on this assessment. You can look at Harris’s own site to see their dispute range beyond this

    I’ve already read the whole exchange. Actually, I was summarising Harris’s view of the disagreement. Search for “However, it seems to me that we do diverge at two points:” on this page.

    > If this isn’t pure Cartesianism, I don’t know what it is. ”

    It isn’t. Harris does not require that there be a homunculus in his mind to be the witness, pace Dennett’s assertions to the contrary. A more charitable interpretation would take him to mean his holistic self.


  3. Hi Wm.Burgess and Marko, in your latest replies you both touched upon a similar observation, regarding the relationship between CFW and LFW.

    “Right now I am thinking that there is no substantive difference between CFW and LFW… I’ve never heard anyone articulate an LFW position that didn’t end up being a form of CFW, nor do I think anyone can.”

    I largely agree with this statement by Marko, but I want to explore how/why this is the case. There is some added nuance to this claim when one looks at the definitions being used between the two.

    Basically anything considered having FW using the LFW definition, would by necessity (based on its level of agency) count as having FW using the CFW definition. The reverse is not the case, since CFW would not expect (if the tape were wound back) that given all the same physical conditions you could make a separate choice.

    LFW proponents who are dualists certainly can articulate a difference between LFW and CFW even if their actions meet the CFW definition, on the basis of the nature of the conscious mind itself.

    However, and this is where your point comes home, if one begins discussing non-dualist LFW it becomes vanishingly small to differentiate the two, beyond the ‘play back’ situation (which arguably we can never test). This is partly why I was so puzzled by the strong reaction of some that I missed some ‘other’ version of FW that would have made my case for CFW harder.

    And of course your point was mirrored in commentary by Marko…

    “This is arguably a matter of terminology distinction between LFW and CFW. But if we roughly agree on the terminology, my argument is that presence of nondeterminism implies that LFW and CFW are operationally impossible to distinguish… The physics of the brain is the same, while the metaphysical interpretation of uncomputability is different. Since this type of metaphysics is essentially a matter of taste, one can argue that LFW and CFW are basically two points of view for the same phenomenon.”

    Agreed. I also agree that CFW cannot eliminate LFW though I would argue metaphysical versions can have questions raised about it by facts like specific lesions in the brain generating discrete/predictable behavior/cognitive effects. Those sorts of tests would not effect physical versions of LFW.

    “I am not familiar with the details of the brain physiology, so I cannot answer in detail. But the mechanism for amplification is quite generic, arguably ubiquitous in nature,”

    I would be interested in seeing this ‘fleshed-out’ as it were, before becoming more confident about its possibilities. I’m not saying it is wrong or impossible, and would encourage people who understand the phenomenon better to pursue that investigation.

    I wonder if this would have more ‘purchase’ if the quantum effect was on the electromagnetic field oscillations produced by the brain itself, than functions of any particular cell.

    “The word “random” in QM is very loaded”

    I wasn’t sure what other term I could use in this case. I was not trying to argue the ‘disorganization’ concept of ‘random’. Rather from the point of view of the brain, and all of its compartments, I am not clear how these effects would consistently be isolated to the appropriate compartments to simply generate novel thoughts/options as compared to other novel (perhaps counterproductive) effects.

    “Second, on ontological grounds, full-blown determinism would imply that all our actions are just a playback of a predetermined movie”

    While I get that QM would generate novelty moving forward, I do have a question about whether it actually has power in the ‘could choose to do otherwise if we wound the tape back’ plank which is commonly held in LFW. That plank is what prevents an agent with CFW from being considered having LFW.

    If we ‘wound the tape back’ to a moment of decision, wouldn’t the same QM effect be in the process of playing out. If so, then the same decision would always be made, and physical LFW seemingly collapses into CFW.

    Or is it the case that one can never ‘wind the tape back’ to a moment of decision as winding it back would entail different QM effects appearing in different locations/moments?

    Especially given the fact that QM has to work through the speed of biological organisms, and the idea is of having created novel choices (now set in the brain), it would seem that any decision that is ‘rerun’ from the moment it was originally made, would have to result in the same choice.

    Hope these points make sense.


  4. Hi Marko,

    However, a nondeterministic source is already there (whether FW needs it or not) and can arguably have large impact on the functioning of the brain.

    I agree in principle. But, if we think about what primitive brains would be doing, they’d be doing things like sensing the scent of blood in the water, evaluating its direction, and then sending signals to the fins to cause swimming in that direction. To do that you want a deterministic device; you don’t want to swim in a random direction owing to quantum non-determinacy. Thus one would expect brains, at the low level, to have evolved with sufficient redundancy in the signalling pathways to average over any quantum non-determinacy and produce an effectively deterministic device.

    From there I’d put the burden of proof on anyone arguing that quantum non-determinacy propagates as far as macro-scale decisions such as the flapping of fins. Virtually everything done by the brains of virtually all animals would be best done deterministically. Further, fully deterministic systems are plenty good enough to, for example, beat the best humans at chess.

    Admittedly, AI devices haven’t yet written sonnets to outclass Shakespeare, so I suppose one could try arguing that quantum-novelty is required for that, but I’d still place the burden of proof on any such claim.

    The second law of thermodynamics is a *restriction* on the behavior of a physical system (certain initial conditions are forbidden by it).

    Well it is in your conception of the second law! This whole concept still seems weird to me. How does the system “know” about its initial configuration, how does it figure out whether that state is “forbidden” or not and how does it then ensure it is not in that state?

    In my conception of the second law there is no such restriction on starting states. All there is is some probabilistic non-determinacy in the low-level dynamical rules, and that alone is sufficient to result in second-law behaviour. Thus, in this conception, it is a low-level freedom of exactly the sort you are asking for here. And, to quote you :-), “a nondeterministic source is already there”, so why not use it to explain the second law? (Straying rather off topic here, sorry everyone.)


  5. Hi Miramaxime,

    > It rather seems to me that if retribution is unjustified under CFW then it is even more so under LFW be it in its “miraculous” (religious) form or its natural “randomness” form.

    Firstly, I don’t really think randomness is sufficient for LFW. LFW requires something more than either randomness or determinism. There are naturalist libertarians, but my view is that their position is self-contradictory and not well thought out, so I cannot begin to defend it.

    The most sensible form of libertarianism in my view is the supernatural kind, which you have described as holding the soul responsible for decision making. I agree with that characterisation. Where we differ is that you seem to think libertarians ought to hold the soul responsible without holding the person responsible, however to most libertarians the soul *is* the person. It is the soul that is punished or rewarded in the afterlife, after all. Crucially, the soul’s decisions are not held to be the outcome of caprice, nor are they held to be determined. They are held to be the result of something else, although what this is is never really described to my satisfaction although such description might use such terms as character, temptation, willpower and so on. Whatever it is, it is that kind of causality that justifies retribution for some libertarians.

    For the libertarian, the buck stops with the soul, and so souls are either good or evil or somewhere in between. Furthermore, souls freely choose to be good or evil, so they are responsible for their choices. It is therefore justifiable (to some libertarians) to hate evil souls and praise good ones. Hurting evil souls (punishment) is a good in its own right, as evil is the enemy of all right-thinking people. It taps right into stone age psychology, us vs them, the in group vs the out group, but I don’t claim to find it a particularly good justification even on its own terms (as I’ve said before, I think all forms of LFW are incoherent). It is, however, a way of looking at moral responsibility that comes all too naturally.

    But determinism undercuts this questionable justification by eroding the premises it stands on. If we take this ultimate choice out of the equation, then people are ultimately a product of their environment, circumstance and genes. Yes, people can (compatibilist) choose to improve themselves, but if they fail to make that choice then they were determined to do so. Hating people for who they are or the decisions they make is therefore as hard to justify as hating people for being poor or disabled or ugly. Ultimately, they can’t help it. Before they were ever born they were destined to be this way. Punishment is no longer a good in its own right but can only be justified on consequentialist grounds such as deterrence or rehabilitation.


  6. Hi DM, we seem to be hearing what we expect to hear regarding conscious mind I guess, because I do not get some of the statements to mean less than the conscious mind (thought) does not influence behavior. I’ll give some clarifications, but otherwise we can drop this angle.

    “Actually, I was summarising Harris’s view of the disagreement… Harris does not require that there be a homunculus in his mind to be the witness, pace Dennett’s assertions to the contrary. A more charitable interpretation would take him to mean his holistic self.”

    I can only note that does not undercut the point I was making about their dispute. That Harris claims his position does not entail what Dennett claims, does not mean there is no dispute on that issue. Dennett still maintains his evaluation was right. So the dispute still exists, only now the question is, is Dennett right? And I’m not sure why I should take a charitable reading (in Harris’s favor). I’ve seen sloppy back and forth arguments from him before. Though I should point out that Dennet was not suggesting a humunculus-like subunit. Harris’s use of I creates a functional entity separate from the brain which is a passive observer.

    “The kind of free will defended by Al Mele is CFW”

    Right, and actually I liked his ‘grades’ of FW concept. That part was not so much important for our discussion and more for philonous. Regardless of the issue about conscious minds, I understand that the neuroscientist was concentrated on disputing classical LFW and not CFW.

    “Because of history and everyday experience, LFW is the “default” position and what many people understand by the term “free will”.”

    That is disputable. I know for myself, even having been raised in a somewhat religious household, I never heard of “free will” except in the arguably compatibilist sense. This may in part be because some of my childhood was spent around lawyers (relatives). My thought is this is also likely to be a more common concept to our distant ancestors than that they have some soul capable of making different decisions if the tape were wound back. Granted once we started telling stories and letting our minds wander, that greater sense of freedom probably started coming up. My guess is that some people back than, just like today, never bought into that idea.

    “Honestly, I think few AFW proponents would have a serious issue with CFW if they scrupulously always prefaced the free will they believe in with “compatibilist”. Likewise few CFW would have a problem with AFW if the latter always clarified that they disbelieved in “libertarian” free will. ”

    I’ll end with a point of agreement. I think when context isn’t clear (like when one is in a court room) we should distinguish between LFW and CFW by name. There certainly does seem to be a hangup about this, which I don’t find useful. I identify as compatibilist and don’t claim FW only means that.


  7. Hi Everyone,

    there is a lot of criticism in the comments, some of which I can relate to, some not at all, so I would like to expand on why I find this piece valuable (not that Dwayne is not able

    The way I see it, the articles main “flaw” might be that it could have framed the contribution better. Rather than dismissing all versions of LFW and AFW, the arguments target the divide between incompatibilists and compatibilists. This, in my view, explains why it focuses more on the perceived similarities between (some versions of) AFW and (some versions of) LFW instead of their differences, since both camps often rely on similar arguments when criticizing CFW.

    It was said by some in the comments that CFW and AFW are similar since they both (roughly) agree on the natural facts (“on what is going on”) while others said that LFW and CFW are similar since they (largely) agree on the (moral) implications of free will.

    Instead I find some (most?) versions of LFW and AFW are similar/collapse into each other in that they would/should agree about all the relevant consequences given that they just agreed on the facts. If an AFW would not be “converted” to a LFW on showing him “a soul” or the workings of randomness in the brain then she opposes LFW on other (more fundamental) grounds than Libet’s style experiments. It is in this respect that a proper understanding of CFW can show that much of the discussion between AFW and LFW misses the mark.

    I thus understood the aim of the article as to argue that the debate should not focus on determinism vs. nondeterminism nor dualism vs. monism nor neuroscience/naturalism vs. supernaturalism since nothing of importance hinges on any of these topics, when talking about free will.

    Here, I think, Dwayne is absolutely right.


  8. Ok, Dwayne. We can drop it there. I enjoyed the conversation. Thanks for provoking a very interesting comments thread.


  9. Copenhagen interpretation (CI) is all about the complementary, and it is wrong. This wrongness misled not only the entire physics but all other fields (philosophy, etc.).

    This mutable universe is the emergent of the immutability. The arrow of time is the expression of the timelessness. These two are not philosophical concepts but are actual physics-mechanisms, which provide the framework for the calculation of all nature constants. And, they are the mutual-immanence mechanisms which are completely different from the complementary.

    The fact that the skirmish at the frontline (in accordance to the ‘engagement rule book’, the hardwired) a few days before the White House (WH) decision (an expression of a Will) of retaliation does not prove that WH is not a Will-agent. Thus, the Libet-type of result is not even an interesting thing in neuroscience, let alone to have any relevancy to the free-will issue. We are totally off track here.

    However, these two wrongness lead the ‘determinism’ becoming the center point of this free-will issue. In order to rid of these wrongness, we should ask two questions.

    Q1, does determinism of any kind (hard, soft, psychological, nomological or else) have any relevancy to free-will?

    Q2, what the heck is determinism anyway?

    The answer for Q1 is a huge “NO”, and it should be answered on its own. But, by answering the Q2, the answer of Q1 then goes without saying.

    There are zillions different determinisms in the market today. Thus, the best way to examine this issue is by analyzing the ‘Indeterminacy” (the quantum uncertainty, the QU).

    One, the philosophy of QU {delta P x delta S >= ħ}:
    First, without QU, we will never know that electron is unable to stay in the nucleus. QU makes this knowledge known and totally CERTAIN.

    Second, without QU, there is no chance of any kind for us (human mind) to enter into the micro world. QU is the guiding light to reveal the knowledge of the micro-quantum world with total CERTAINTY.

    Two, QU in physics:
    First, the quantum spin (1/2 ħ) defines the structure of the micro world (the fermions) with total CERTAINTY. From this certainty, all nature constants are calculated to their precise values.

    Second, it is the quantum action (ħ) which drives this universe expanding with acceleration. The ruler of this acceleration is the Cosmology Constant (CC), which is totally CERTAIN with 120 ZERO after the decimal point while there could be some uncertainty at 123rd digit (might not be a zero). How can such a CERTAINTY be accused as UNCERTAINTY?

    The quantum UNCERTAINTY paradigm is totally wrong. More detailed elaboration on this is available at .

    Without uncertainty, there will be no determinism. Any free will discussion, based on the determinism or it opposite, is wrong, totally off track. Free will is the reflection of immutability while its ‘expression’ in the human world is decision making (an expression and manifestation, not the essence).


  10. If I counted right, this is my fifth comment in the thread. 🙂


    While I get that QM would generate novelty moving forward, I do have a question about whether it actually has power in the ‘could choose to do otherwise if we wound the tape back’ plank which is commonly held in LFW.

    Yes, nondeterminism is just that — if you “rewind the time” back to some initial moment and let the physics play out again, it will play out differently than it did the first time around.

    Of course, there are both conceptual and practical problems with the idea of “rewinding the time”. But these can be circumvented by discussing a thought-experiment about two atom-for-atom-identical copies of a brain, simultaneously starting from identical states, in an identical controlled environment, etc. After a few Lyapunov units of time, the two brains will display different behavior.

    Moreover, next to these two brains, one can also set up a powerful computer, which simulates of the behavior of every particle in every atom of the brain. After a few simulated Lyapunov units of time, the simulation can tell us only the *range* of possible states that the brain can be in. States of both real brains will of course fall into that range, but the simulation cannot tell us which brain will be in which precise state. This illustrates “uncomputability”.

    Note that I have deliberately avoided to mention any notion of “randomness”, to make the issues at play more clear.


    Virtually everything done by the brains of virtually all animals would be best done deterministically.

    Why would deterministic be best? Because it will be easier to find food? Maybe that’s true if we talk about animal instincts, but instincts are not FW. From the evolution POV, I think FW has an advantage if it is creative rather than deterministic, because it can adapt to changing conditions more efficiently. On the other hand, instincts should indeed be as deterministic as possible, because one wants them to work as reliably as possible.

    Note that increasing complexity introduces stronger nonlinearity, thereby shortening the Lyapunov time, and thus becoming more sensitive to nondeterminism. So it is natural to expect that a “simple” brain of a fish is more deterministic (i.e. it amplifies quantum uncertainties slower) than a “complicated” brain of a human, that is more nondeterministic (i.e. it amplifies quantum uncertainties faster).

    By the way, I agree with you regarding the burden-of-proof issue in this context. It is a scientifically answerable question (one needs to look into the detailed physiology of various brains and estimate relevant Lyapunov times), and it is certainly interesting to pursue. This research seems to have already started, on fruit flies. While the brain of a fruit fly is vastly simpler than the human brain, it is at least a first step. 🙂


    It’s been an interesting discussion, and I enjoyed it! Dwayne, thanks again for a very good essay! 🙂


  11. > While there may only be one choice possible at a certain moment given all inputs (as AFW proponents argue), isn’t that exactly what one would want from any decision-making system (ask CFW advocates): one choice (hopefully the best) from all alternatives?

    A distinction needs to be made between one choice POSSIBLE (you can only walk away with ONE fruit from this box of apples, bananas and oranges) and one choice AVAILABLE (the box only contains an apple).

    If only one choice is available, then there are no alternatives. Having only one choice available means the same thing as having no choice.

    In both a deterministic universe AND a free will universe we can say that only one choice is ever possible at each moment in time, but that has nothing to do with ‘choice’ per se. It’s just an example of our casual use of everyday language muddying the debate. What we mean is there can only ever be one outcome (in retrospect) ie you might be able to choose where you go on holiday, but you cannot arrive at two different holiday destinations at the same time.

    The main problem with the whole debate is that science is no closer to understanding consciousness than it is to understanding some phenomena which has yet to be discovered. But thanks to our ‘multiple choice’ style education system we are trained to think we must always come up with an answer and that “This is my opinion which I am going to defend to my grave” is a better answer than “not enough information, don’t know yet, sorry.”

    In just about every practical sense consciousness has yet to be discovered by science, mainly because modern science is little more than an inventory based on data collected by dead, unconscious instruments and machines. There is perhaps only one instrument available to us which is capable of detecting and handling consciousness and that is the human body itself, and science does not consider this instrument reliable enough to gather data! (Although curiously it is considered reliable enough to interpret data gathered by machines).

    This is why science ended up with a model of the universe which is completely devoid of consciousness in the first place – even though consciousness (and indeed free will) is perhaps the most pervasive, important and OBVIOUS phenomena in the universe. For this reason scientists cannot really deny consciousness altogether, so they try to define it in terms of non-consciousness (electro-chemical reactions in the brain). This is still essentially denying consciousness exists.

    Determinism is an entirely logical conclusion drawn from a dodgy premise. It is like the entirely logical conclusion that we must always encounter giant waterfalls and cliffs if we head in one direction for long enough, based on the dodgy premise that the world is flat.


  12. The discussion of freewill is very different in mysticism and a solution is offered. We find this solution in ‘A Course in Miracles’ where it is stated, ‘Choice is meaningless’. If we grasp this view and what is actually being said then we have reconciled freewill and determinism and all such metaphysical opposites.

    This view is rarely discussed in scientifically-inclined journals, obviously, where the discussions have no end in sight, and it seems to be dismissed prior to analysis, leaving essays on the topic to just trawl through failed solutions over and over again. Hence my jaded reaction to an essay that compares and contrasts various ideas that don’t work. This is not progress. It is not even catching up. .


  13. Hi DM,

    thank you for your response. I can’t say that I understand your claim much better now, but I also realize that it is difficult to defend a position you find incoherent yourself. Let me just say that if one aims to justify retribution on the basis of “good” and “evil”, then this seems to work just as well without free will (the good simply deserves to be rewarded and the evil condemned/punished by definition). And in so far as one aims to ground moral responsibility on the lawful interactions between the “self” and the “environment” (as in willpower, character, etc.) then I do not see how determinism undercuts anything. From where I stand, it simply seems irrelevant whether we are talking about natural or supernatural or deterministic or nondeterministic laws.

    Instead it seems clear to me that some “lawfulness” in “self->environment” interactions is required to assign agency (proximate causation) to a being, which in turn is the basis for assigning moral responsibility. I don’t see how “too much lawfulness” can take it away then.

    Furthermore you seem to somewhat misrepresent the implications of determinism. When you say that (on determinism) “any action is fixed”, you seem to mean to say “any action is fixed by something particular” (an ultimate cause perhaps). But that is not how we usually assign causes, otherwise a stone could never really cause a window to break since – on determinism – the event would have been fixed by some prior cause.

    If you rephrase the implications of determinism as in “given precisely the same internal and external state a being would always take the same choice”, it is much harder to explain why this should undercut moral responsibility.

    Finally, on the topic of desert, I do think that a soccer player “deserves a red card” for using his hands to prevent a goal, as long as he is not the goalkeeper. And I would still be of that opinion if you could prove to me that no bad consequence could ever arise from making an exception in this case. He knew the rules and had the competence to respect them. Given his internal state and the external state at the time he didn’t though. Do you think this might serve as a toy model for desert?


  14. Hi All, I think it is getting close to the end for replies, and it is slowing down with the new essays which came out. So this will likely be my last post for the thread.

    I want to thank everyone for contributing to an enjoyable conversation. I was glad the essay resonated with many people. And I appreciated the criticism as well. I will definitely be taking it aboard when approaching such cross-field topics in the future.



  15. Miramaxime, I’d say I’m out of comments here, but if you want me to reply you can contact me via my blog at


%d bloggers like this: