There’s a perfectly good definition of free will implicit in the name: any act performed due to an agent’s volition without impediment—any act performed according to some will, freely, in other words. Choosing such that the choice (and nothing beyond) determines the outcome. Etc. One can quibble about the details, but this discussion hasn’t yet reached that level.
Otherwise, all that you raise in objection against free will holds just as much against causation.
I should note, perhaps, that I’m not claiming to make an argument in favor of free will; as I said, I don’t really have a settled stance on the matter. All I’m doing is pointing out that the arguments raised against it simply are ineffective, because the notions on which they are based are subject to exactly the same worries.
So yes, it’s perfectly possible that there might be no free will, I’ve never claimed otherwise. But to establish that, more work has to be done than just noting ‘it can’t be reduced to causality or randomness’, as that carries implicit metaphysical assumptions—that everything ought to be thus reducible—that themselves are questionable.
Just to be clear, my research has nothing to do with free will. My research is to model neural activity in order to provide a diagnosis of neurological disease and injury. It was just working on the model inspired me to think about what it may or may not imply about free will, if anything. And I know this board likes to chew on these sort of subjects so I thought it may make a good thread.
I can see why you might say that but no, it doesn’t seem odd to me at all. Because I guess I do think of myself as the ghost in the machine. Perhaps because I’ve always thought of free will as being non-deterministic. Forget for a moment whether it is or not, and put yourself in my shoes of say free will is non-deterministic, you can perhaps see why I would also believe it must be outside of the brain machinery, if the brain machinery is deterministic. My previous view of free will is that the “will” part is the most important element of it. That the will of the free will can alter the deterministic fate of the machine.
I wrote an entire D&D campaign about this idea called the “Against the Forge of Destiny” and the players attempts to overcome their destiny to die. If destiny is a real thing that is created (in the Forge of Destiny), can you overcome destiny if you never destined to overcome destiny in the first place?
This thread has made me consider the possibility of predictable free will. I need to think about it some more, because it still feels contrary to my own personal belief system, but at the same thing it is compelling.
You might be interested in so-called ‘sourcehood-accounts’ of free will, then, the idea being that it’s not whether one could have done otherwise in a given situation that’s important, but rather, where a given choice, or action, originates. I’ve alluded to this above somewhere, mentioning ‘Frankfurt cases’, which are examples intended to show how taking away alternatives not necessarily removes meaningfully free choice.
So suppose you have an agent who’s free to choose between two alternatives, say peanut butter or jelly for breakfast. Now suppose they’ve recently undergone brain surgery, where, unbeknownst to them, the surgeon, due to a deep-seated hatred of peanut butter, implanted a device that, should the agent intend to choose peanut butter, comes online and influences their brain activity in such a way that they end up choosing jelly instead. Now suppose that they, in fact, choose jelly—that is, the device never comes online. Now, they couldn’t really have done otherwise: had they chosen peanut butter, the device would have come online to influence their choices. But it’s really odd to say that a device that never comes online—that doesn’t do anything—could be meaningfully efficacious in such a way as to rob the agent of free will.
However, we may agree that if the device comes online, and changes the agent’s choice, then they certainly aren’t acting freely—but then, it’s not the presence of alternatives, but rather, the source of the decision that’s relevant: the agent is the source in the case the device never comes online, but if it does, then the device itself is the source of the action, and hence, the agent no longer free in that case.
This is true and in your scenario the device did not remove their free will. But if the reason that they prefer jelly is that they were given a jelly sandwich every morning as a kid and developed a liking for it, are they really choosing freely, or did their parents inadvertently program their brain to like jelly sandwiches from a young age?
Addressing your #4, the brain pattern is responding to sensory input. We can observe the results of that response externally. Your ability to see the neuronal pattern has finer granularity than my ability to observe the response “and the tlithy toves”. But, both methods observe a deterministic pattern. In fact an associated pattern is invoked in most of the brains that have read the post. That’s a reasonably deterministic response to an external stimulus.
However, some folks with a different cultural background would not get any response. So, the response is not hardwired.
As you point out above, the brain can generate an internal stimulus and determine the resulting response. That is at least volition. Whether or not it’s free will depends on the definition of the day for ‘free will’.
I need clarification to ease my ignorance. You have sequential images of neural patterns. What is the time between two sequential patterns? The initial pattern is the result of some dendritic input. Is that known to be constant between patterns N and N+1?
1.96 seconds is the average temporal resolution (variance is pretty low) of the fMRI volumes we have. The EEG data is at 5000 Hz.
Again, for clarity, the research has nothing to do with free will. It is intended to diagnose neurological injury and disease. There is plenty of research that does something similar, we have a novel approach by treating the neural activity as a causal temporal process. Previous works have largely treated the data as point-in-time (or short windows for EEG). The participants are in an fMRI machine in a “rest state”, i.e., they are asked to not think about anything in particular. They are not given any tasks to perform. Etc.
I discussed the modelling aspect purely from the hypothetical point of view that if there were a model with sufficient resolution and accuracy, what would it mean to the members of this board with respect to free will. This research will not produce such a model, even if it meant anything.
You’re going to like this - My analysis of free will doesn’t rely on it being doing by the physical brain! Which is to say, even if your mind is housed in a supernatural soul/spirit/ghost, my analysis still applies!
So let’s talk about what free will being non-deterministic actually means. I think we can agree that it means that the outcome is not predictable based on existing state - and state inside a ghost still counts as state. If you have a soul, and that soul has thoughts and opinions and personality and memories, then those parts of the soul are determining factors. So how are things nondeterministic? The short answer is randomity. Some amount of randomity perturbs the system while the decision-making process is occurring, causing potential variance in the outcome.
It’s worth examining what kind of variance you’d expect to occur. I personally thing the process would work something like this - your (deterministic) preferences are used to weigh the options via internal heuristics. Noise in the system can perturb the weightings a little, up or down. Then the weightings are compared, and the best-rated one wins. The effect of this is that randomity only effects things where you’re almost completely ambivalent about the results - If you have any preference at all for one option or the other at the time of the decision, the randomity wouldn’t be able to overcome it. It couldn’t make you choose to drink sewage instead of tea - and it probably wouldn’t be able to make you choose coffee over tea either. Only when you really don’t care at all would the coin flip happen, essentially.
I’ll note that given that model of how cognition works, I consider the randomity/nondeterminism to add nothing meaningful to the decision-making process. Which is to say, I think you’re not actually making choices - to whatever degree nondeterminism influences the outcome, you lose that much volition. In other words the more random you are, the less of a mind you have, because that mind isn’t doing anything but mindlessly flipping coins. So I really, really don’t value nondeterminism in cognition - to whatever degree it is present, it contributes nothing valuable at all.
It’s worth noting that in the absence of randomity resolving the tie, under pure determinism you’d instead get an arbitrary resolution of the tie - whichever option was considered first or whichever. This also wouldn’t be a meaningful decision, but in this case we can be certain it would only impact decisions where the volitional parts of your minds are completely ambivalent anyway.
And, again, I’m talking about any mind here, be it physical, spiritual, magical, or anything inbetween. There could be randomity in physical reality, derived from quarks or whatever. There could be randomity in your soul, thanks to soul gremlins slowly eating it. There could be noise along the spiritual cord connecting your soul to your appendix. I don’t care - I’m considering how cognition works, not where it’s located.
In my opinion that the device must come online, or constantly be online, in order to detect whether it ‘needs to come online’. Which means it did do something, even if that ‘something’ had no consequential effect (since the device already had the outcome it liked).
So, in my opinion, your free will is being abrogated either way.
I think you and I had this conversation before… if I remember correctly I was unable to persuade you that there is a distinction between nondeterminism and stochastic nondeterminism. I wonder if BeepKillBoop believes the idea of non-random nondeterminism is coherent?
I’m not sure. I guess it depends on what that means.
In computer science, an algorithm can be non-deterministic with random elements using parallel processing. I somehow doubt you’re referring to that though.
I suppose my last post was a bit too flip. Could you please describe how the nondeterministic aspects of a process could be generated, if not randomly?
I was using nondeterminism as “not determinism”, “not able to be determined”; indeterminism. But in the analogy of an algorithm, supposing your algorithm contained steps which waited on a human being to enter a value at his or her discretion, such an algorithm is nondeterministic (within the realm of computer science; a person analyzing the algorithm in isolation could not predict the output).
Like so with libertarian free will, where physics stands in for the algorithm. It is hypothesized, by libertarian philosophers, that the physical universe is not deterministic. It is also assumed that free agents (souls) external to the physical universe step in and interact with the physical universe (mind-body dualism), usually restricted to abide by (indeterminate) physical laws. By what exact mechanism the free agents decide, it is impossible to tell based on physical evidence alone - they exist in an external (nonphysical) “substance” just as the human being is external to the computer algorithm being studied in the analogy. The preferences of the agent are analogous to the preferences of a human being who dresses up a virtual avatar, you can still make pretty good predictions based on past behavior but there is not necessarily a direct causal chain. And so libertarian free will is an emergent phenomenon between the physical and nonphysical substances.
It’s also based on a falsifiable premise - that the physical world is indeterministic.
Sure, and that kind of free will exists of course. The data suggest we make decisions based on our past experiences and neurology, which may be entirely deterministic or may have a random component.
This is demonstrable since, while we don’t understand everything about memory for example, we know enough to recognize it as a brain function, know the organelles involved and know obviously it plays a critical role in our decisions.
Anyone that wishes to propose any other inputs is welcome to do so.
No; according to you, one argument against spooky free will can be applied equally to causality.
I don’t believe that’s correct for the reasons given (eg falsification), but even if true, I wouldn’t agree that that’s the only argument against it.
(Also, technically speaking of course, saying “that argument can be applied elsewhere too” does nothing to counter the argument, but let’s put that aside for now)
I know I said I’d retreat from this discussion, but that’s factually not correct, and since I’m not in the habit of acting all scandalized upon being misrepresented without at least trying to clear things up, I think it’s in the interest of level debate to correct this.
The main arguments, as proposed in this thread so far, were:
The consequence argument: the future is as it is because the future is determined by the laws of physics and the past state, and neither the laws nor the past is subject to my will; hence, the future is not subject to my will. The counter to this is to point out that on, say, a Humean account, the laws do not dictate what happens in the future, but what happens (the quilt of local matters of fact) determines what the laws are (as a highly compressed account of this quilt).
The regress argument: if the will determines what is chosen, what determines the will? If it’s chance, it’s not free; if it’s determined by something else, it’s also not free. So we enter into an infinite regress. The counter to this is that any causal view leads to the same problem, or necessitates proposing a ‘first mover’, or must allow some things to be self-caused. But then, there’s nothing that says the will can’t be self-determined, either.
The reducibility argument: free will can’t be given an explanation in terms of more fundamental notions (in this thread, mostly causation and/or randomness), as any such explanation will simultaneously yield an account of how the will is determined, and hence, not free. This is sound, but the premise that everything should be reducible is not generally accepted—notably, causation and randomness are typically held to be irreducible. But if you admit these as irreducible, you can’t validly hold that free will must be reducible.
The predictivity argument: we have reason to believe that causation exists, since causation makes predictions (while free will doesn’t). The counter is simple: it doesn’t, the same predictions are made without appeal to causation, as shown by the examples given. On my challenge to provide any sort of prediction solely due to causation, none has been forthcoming so far.
The falsifiability argument: we have reason to take causation more scientifically serious, because it is falsifiable (unlike free will)—there are certain observations we could make that would convince us that the world isn’t causal. Again, the rebuttal is that this just isn’t true: I’ve given an explicit procedure above how any sort of data generated within an indeterministic system could be obtained from a deterministic one, instead.
The exclusivity argument: everything that happens, does so either causally or randomly. Again, this is shown not to be true by the above examples. Any appeal to the fact that, say, in a block universe, causality must be at work in setting it up is just circular: we’re trying to see whether causality is necessary, and hence, one can’t appeal to the fact that it surely must be at work in some way in these settings in order to show that it is. The block universe’s existence is equivalent to the brute notions of causality or randomness, in this regard.
The exclusivity argument, redux: everything either happens for a reason, or it does not happen for a reason: this is, to my best reconstruction, the argument that the preceding one tries to be. However, in this formulation, it’s clear that it’s not apt: the will is the reason things happen in models including free will, and is not further reducible to causation or randomness (just as causation and randomness are not further reducible to other notions).
So in particular, my critique is not that any given argument can be applied ‘elsewhere’, it is that it can be applied to the alternative that the argument was meant to support—if you have two options, X and Y, and you say you choose X because Y is susceptible to Z, then showing that X is susceptible to Z means that this can’t validly be a reason for making this choice.
So, it would be reasonable to expect that you could create a model that imitates some pattern. But, the existence of a pattern would not be deterministic. It’s just part of the mechanism. Much like measuring the osc frequency of a computer doesn’t reveal anything about the program it is running.
You are certainly involved in a fascinating project.
That doesn’t strike me as being particularly controversial if I’m understanding what you’re saying. Godel’s Incompleteness Theorems (no idea how to do the o with the double dot accent) tells us that not everything can be proven (given some set of axioms that comprise an algorithm). And we see that in computer science with things such as the halting problem. In physics, there is the uncertainty principle, which would seem to be the mathematical equivalent of incompleteness.
I’ve seen people bat this rhetoric around a couple times in this thread, and I think it’s worth putting a name to a face: Compatiblism
Personally I find this position most convincing too. Things are determined and all free. If someone holds a gun to my head and says “give me your money”, then I am not making a free action according to my will by giving him money, but if I buy a jar of jelly because I like jelly and I like jelly because I grew up eating it and because I am a carbon based lifeform that comes from a line of lifeforms that are evolved to extract energy from hydrocarbons which conveniently are naturally produced by the fruiting ovaries of other carbonbased lifeforms: then yes it is free AND yet determined.