Blog Coordinator

« Neuro-interventions and the Law Conference | Main | New Blog for Undergraduates »

08/21/2014

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

A reminder that square circles are possible -- you just have to change the metric from Euclidean to taxi-cab or chess-board.

It seems to me that the laws of the Pac-Man universe are not fixed. For example, if I learn a new strategy (or switch strategies in the midst of a game) then the probability wave-functions for the Pac-Man change. It may be that in a situation where the up/down choice had always been a 50/50 split, the probabilites have changed to 100% up (perhaps because I've learned that up is the only choice that works in those situations). If the laws are not fixed -- and the changes are not predictable from the earlier states of the world -- then in what sense is the Pac-Man universe deterministic?

Another way of stating the same thing -- there is not enuf information in the state of the Pac-Man world at time T0 to fix the state of the Pac-Man world at time T1 -- not even probabilistically -- because the evolution of the laws is not determined by the state of the Pac-Man world at time T0.

Hi Mark: Thanks for your comment!

I think yours is the right worry to have, but I still want to resist it, and think that I have the resources to do so.

Here's how I want to understand things.

First, on the P2P Hypothesis/Libertarian Compatibilism, each of our libertarian choices in the higher reference-frame at any point in time violate the normal Schrodinger equation in our brains in some small way.

Second, each of these choices (combined with the choices of all other libertarian wills and the normal Schrodinger equation for objects outside of our brains) causes a global "quantum collapse" from many previously available "paths" through the multiverse (see "A New Theory of Free Will" on this) to a single, intersubjectively experienced reality.

Third, each libertarian choice leads, in the next time stage T of the multiverse, to observations of small neural-network alterations that reflect your playing habits up to T (i.e. one's brain comes to *reflect* one's playing habits).

Fourth, in the way that Kant imagined sensible inclinations/desires impinging upon us from the outside (and which our wills can always resist), these neural changes impose inclinations on our wills (much as gobbling up a "power-dot" in PacMan alters how quickly one can move...though one can still make any choices one likes).

Here, then, is what we have so far. Libertarian choices cause brain changes which in turn (insofar as those changes result in physical inclinations our wills experience as outside "nudges") tempt us to make similar libertarian "moves" after To.

Hence, Libertarian Compatibilism has a mechanism to explain how we (A) develop inclinations which habitually tempt us in ways that reflect earlier choices, while (B) still always retaining radical libertarian freedom to choose otherwise.

This is, of course, roughly Kant's picture: (1) Our libertarian choices affect our sensible inclinations as imperfect agents in the phenomenal world, (2) our sensible inclinations in turn *tempt* us to make certain choices as noumenal agents (beyond the phenomenal world), but (3) despite having those inclinations, we always fundamentally retain radical noumenal/libertarian freedom.

Now, you say, if we always retain radical libertarian freedom, what happens if one starts making different choices all of a sudden at time Tn (viz. one begins to adopt a different strategy)? Well, in that case, one's "personal wave function" -- as observed by observers in this physical world reference-frame -- will begin to change at Tn, showing a new collapse profile at Tn. But then, at Tn+1, one's brain will alter in a way that makes *that* new collapse profile a bit more probable than before. And so on!

How will this appear to observers within the physical reality? Will it appear to be a violation of causal determinism? No! Why? Because, prior to Tn, the choice one made at Tn was always possible (given one's personal wave-function), just *improbable*. So, the new collapse at Tn won't violate their physical laws up to Tn. But now, since at Tn the person's brain will update its neural pathways to make that new collapse value (the new choice made at Tn) *more* probable than before, if a similar choice is made at Tn+1, that won't violate the deterministic (though probabilistic) laws they had formulated up to Tn or Tn+1, since now the person's neural-pathways make (from their perspective) that new collapse value even more probable than at Tn!

And so, even though one's "playing habits" have radically changed at Tn, at every instant after Tn one's new habits result in updates to the brain that make those very playing habits appear more probable...and so consistent with any physical laws observers in the world had formulated up to that time (whatever those formulas may be).

In other words, physical changes corresponding to libertarian playing strategies will mask the libertarian nature of their choices in their reference-frame, leading them to conclude that their reality is deterministic after all. In which case my answer to the Consequence Argument still holds: (1) from our reference frame in our world, we observe deterministic processes(thanks to physical changes masking the libertarian elements of choice in the higher reference-frame), but (2) from the standpoint of the higher reference-frame, our world is non-deterministic/libertarian.

Hope this makes sense (I had to write it relatively quickly, as I have to head off to dinner with the wife!). Anyway, thanks again for your comment!

Hi Mark: there's one addendum I'd like to make to my previous comment to head off a worry about it that I expect you might have, but I will be out for the night, so look for it tomorrow!

Joe and I have had much back-and-forth over the last few years on BeeBee and Mele's ingenious "Humean Compatibilism" (Mind 2002). Using a background of Humeanism about laws across possible worlds, they argue that one particular deterministic world may include agents like us whose actual choices include "dual-able" counterfactual powers to have chosen and acted otherwise than they did. Thus though a particular world turned out deterministically, it need not have (due to a lack of a necessitarian account of laws under Humeanism). They also go on to argue that their view is subject to charges of luck, due to a metaphysical resemblance of sets of Humean worlds to indeterministic ones. Thus Humean compatibilism also resembles libertarianism (qua indeterminism) in offering a form of counterfactual dual-ability. (They also claim that their view side-steps the CA by making the laws "up to us", somewhat as you claim for your view.) So B&M claim that (i) determinism can be true of a world and (ii) agents in such a world still have (Humean) dual-ability, and not unlike libertarian dual-ability.

Your view is different because it relies on different frames of reference of explanation to account for the reconciliation of determinism and libertarianism. But I wished to offer B&M as another approach that offers many of the same claimed features on different grounds. (Not that I endorse Humean compatibilism, as Joe can certainly tell you!)

Hi again, Mark: Here's the addendum to my initial reply to your comment.

In my reply, I described a physical mechanism (neural-network updating) that might mask libertarian free will, leading observers in our reference-frame to observe the world to be causally deterministic (relative to our reference-frame).

The process went like this. Take any time, T, you like. At time, T, a person's neural network will instantiate certain quantum-feedback mechanisms that entail violations of the normal quantum wave-function (i.e. Schrodinger equation). [See Hameroff and Penrose's work, as well as recent observations that seem to support their view of how these violations emerge at the microtubule level. http://www.sciencedaily.com/releases/2014/01/140116085105.htm ).

Now, as you might know, the Schrodinger equation states that a fundamental particle can, in principle, appear *anywhere* where measured, but that certain probabilities are higher than others.

So, whatever quantum-violation mechanism brains instantiate (as a result of libertarian choices in the higher-reference frame), they will lead observers in our reference-frame to observe (A) subtle shifts in the probability function of quantum collapse in said brains, where (B) any possible collapse-value is possible, but some more probable than others, where finally (C) these quantum-violations *appear* to them to be the result of neural-network quantum feedback.

Whatever particular collapse occurs in a brain at T+1, then, will be *consistent* with everything they have observed about the causal structure of their physical world up to time T. Even if (as a result of "new strategy" of libertarian choice in the higher reference-frame), improbable post-collapse values keep recurring, at each successive time-stage, this new post-collapse value will appear to be made more probable by neural-network updating...and so everything will still appear, from their reference-frame, to have a fully deterministic physical explanation.

Here's the schema:

(1) Observers in higher reference-frame H will observe libertarian processes causally determining Schrodinger violations V, which in turn update neural-networks to make V appear more probable in the future, etc.

(2) Observers in lower reference-frame L will observe neural-network quantum-feedback mechanisms appearing to causally determine Schrodinger violations V in manners otherwise consistent with all known physical laws in L, etc. (physical determinism appears true-relative-to-L).

So far, of course, this is the story I already told in my previous comment. But there is one "hanging thread", so to speak, and it has to do with your point about the evolution of the laws in this model.

You write: "There is not enough [sic] information in the state of the Pac-Man world at time T0 to fix the state of the Pac-Man world at time T1 -- not even probabilistically -- because the evolution of the laws is not determined by the state of the Pac-Man world at time T0."

Now, I don't think the probability claim here is right. If quantum-violations caused by libertarian free will in the higher reference-frame give rise to neural-network updating in the lower reference-frame making those very post-collapse values more likely in the future, then the state of Pac-Man at T0 will appear to *probabilistically* fix the state of Pac-Man at T1 (though, again, these probabilities will only be observed in L, whereas observers in H would know they are caused by libertarian choice).

Your point seems far more worrisome, however, if we consider more carefully the notion of "the evolution of the laws."

In our world, of course -- as far as we presently understand it -- physical laws do not evolve at all. Rather, they are entirely fixed (and were fixed at the Big Bang).

On my model, physical laws will *not* appear to observers in L to be fixed. In human brains at least, they will appear to dynamically evolve (or change!) over time, in virtue of whichever Schrodinger violations occur in those brains (leading to whatever neural-network updating making further such violations to appear more likely, etc.).

Moreover, due to the probabilistic nature of whatever Schrodinger violations there are (whatever violations occur will appear to have some probability distribution), observers in L will not be able -- in principle -- to specify in advance *which* ways the laws will evolve.

For instance, suppose, to take your example of "strategy change" again, my libertarian choices in H up to time T have led my brain to develop in ways that make a certain type of Schrodinger violation, call it the "Marcus wave-function", more probable. Then, however, suppose (in H) I make a drastic strategy change at time T+1, such that the post collapse values in L suddenly appear to observers there to take on *improbable* results vis-a-vis the Marcus wave-function. Notice, first, that although those post-collapse values are improbable from their reference-frame, they don't strictly violate the Marcus wave-function.

What will happen on my scheme, however, is that my "new strategy" (in H) will cause my brain to evolve so as to make a NEW type of quantum-violation more probable, call it the "Marcus' new strategy wave-function."

Now, although this new wave-function (in my brain) will appear to have a physical-causal basis to observers in L (viz. quantum feedback mechanisms in my updated neural networks), nothing in the world up to T -- not the ordinary quantum laws, nor the "Marcus wave-function" -- *necessitated* the dynamical shift to "Marcus' updated strategy wave-function" at T+1. That shift will appear to observers in L to be an *improbable* dynamical shift in the laws. Moreover, prior to time T, observers in L would have to say, "There are many ways the laws in our world could evolve from here. If Marcus' brain keeps to its present wave-function, certain actions will be more probable -- however, his brain *could* shift to a new wave-function, in which case things will go differently."

Now, I expect you'll say, that's just a denial of causal determinism! If, for all anyone can ever know in L, the physical laws in their world (in my brain, your brain, etc.) can suddenly shift, and go in one of many different ways, then this is just to say that causal determinism doesn't hold. Later events are not *determined* by earlier events!

Although this might appear to be the case, let me explain why I think it's not right.

The correct definition of "determinism" is a tricky issue. I think it is often used in two slightly different ways:

(A) Determinism = necessitation of particular events from earlier ones (i.e. one can deduce what will happen in the future for certain if one knew everything about the universe, qua "Laplace's demon").

(B) Determinism = necessitation of future events by *physical law* (including "indeterministic" laws).

Determinism, as understood in (2), might also be termed "physical causal closure."

Why is this distinction important?

Here's one reason why it's important: determinism in (1) isn't the real issue in the free will debate. We can see this simply by considering ordinary quantum mechanics. The laws of quantum mechanics are indeterministic -- and so determinism qua (A) is false in our world (I set aside determinisitic interpretations of QM). Why, though, doesn't the problem of free will simply go away with quantum mechanics? The answer is simple: because, even though QM is indeterministic in sense (A), it is *deterministic* in sense (B). It says that everything that happens in our world is physically determined by the Schrodinger equation -- that our world *follows* that physical law inexorably.

This is why the real issue is determinism in sense (B). The question is how we can be free in a world *governed* by physical laws (even indeterministic ones).

This brings me (finally, and apologies for being so longwinded!) to my big point.

On my model, just as with ordinary quantum mechanics, determinism in sense (A) is false. Future events in our world cannot -- either on standard QM or my model -- be predicted with certainty, even in principle.

However, on my model, just as with ordinary quantum mechanics, determinism is sense (B) -- i.e. physical causal closure -- will appear to observers in L to be true...even though the laws of their world evolve dynamically in ways they cannot predict.

How so?

Well, contrast the following two cases:

CASE 1: Up to time T, I have made choices (in the higher reference-frame) according to a certain Normal Marcus Strategy. After T, I continue to make choices according to that very same strategy.

CASE 2: Up to time T, I have made choices according to the Normal Marcus Strategy, but after T (at T+1, T+2, etc.) I adopt a New Marcus Strategy.

Observers in L cannot *predict* in principle which case will obtain before it obtains. Moreover, whichever the two cases does obtain will lead to neural network changes that make that strategy appear to observers in L as different *laws*. So, there's a clear sense in which observers in L cannot know for sure in advance how the laws of their reality will evolve.

All that being said, this does not contravene determinism in sense (B). It doesn't contravene it any more that the inability to predict the future does in ordinary QM! How so?

Consider again ordinary QM. The Schrodinger equation asserts that an electron has a certain probability of being observed at point A, point B, point C, point D, etc. If the electron is observed at point A, the universe will evolve in some particular way. If it is found at point B, it will evolve in a *different* way. Etc. Despite all this, we say the universe is "causally closed under QM" because all post-collapse values are consistent with the Schrodinger equation.

The very same thing is true of my model at the level of the dynamical evolution of the laws!

If Case 1 obtains -- the case where I keep choosing the Normal Marcus Strategy -- that case will appear to observers in L to be possible and consistent with everything they know about physics up to T. Observers in L will say, "The laws of our universe and neural-network updating entail at T that the Normal Marcus Wave-function is the *most likely* law of nature to continue to hold in Marcus' brain. However, there are other less likely ways that the Marcus Wave-function could evolve consistent with our laws of physics. It could, in fact, evolve according to a New Marcus Strategy...though that is unlikely."

Similarly, if Case 2 obtains -- the case where I in fact choose the New Marcus Strategy -- observers in L will tell the same kind of story. They will say, "The laws of our universe entailed at T that the most likely dynamical evolution of the laws in Marcus' brain will continue to be the Normal Marcus Wave-function. However, there were always *improbable* post-collapse values in the Normal Marcus Wave-function that could lead the laws in his brain to dynamically evolve to a New Marcus Strategy Wave-function -- and this is exactly what occurred!"

Both stories respect determinism in sense (B) (i.e. "physical causal closure"). Although people in a world governed by my model could not predict with any certainty how the laws of their world *will* evolve in human brains, however they *do* evolve will appear (in any path through the multiverse) to have a fully physical explanation in terms of earlier laws and probability functions.

Does this make sense? Sorry it was so long-winded. I'm writing it first thing in the morning, and my brain has been a little fuzzy. :)

Hi V. Alan: Thanks for your comment! I wasn't aware of that paper by Beebee and Mele. Thanks for bringing it to my attention. I will definitely check it out. It sort of brings to mind this recent paper by Christian List (http://philpapers.org/rec/LISFWD ).

The way you describe it, there are certainly some broad similarities between their approach and mine. However, I have some real problems with their approach. Let me briefly explain a few.

First, I don't think the Humean picture of reality ("the Humean landscape") is correct. On the Humean picture, a reality is nothing more than patterns of otherwise disconnected objects and properties in spacetime bearing certain regularities to one another but no object/property *necessitating* any other. Reality, in other words, is like a "painting" where no pixel of the painting necessitates the color of any other pixel.

I don't think this picture is right for a number of reasons, not the least of which are those Gregg Rosenberg gives in Part II of his fantastic 2005 book (http://www.amazon.com/Place-Consciousness-Probing-Structure-Philosophy/dp/0195168143 ).

Second, I don't think counterfactual theories of causation are right (I have a paper arguing that which counterfactuals we identify as "causes" trace back to either judgments of mere Humean regularities or to Kantian necessitation relations).

Third, even if counterfactual theories of causation were right, the standard analysis of counterfactuals (in terms of possible worlds) wouldn't suffice to show -- on Beebee and Mele's analysis -- that people in our world have a genuine *capacity* to do otherwise. It would merely show that our counterparts in nearby possible worlds could do those things (raising the famous problem of counterparts).

Finally, I don't think the Humean theory of motivation -- the only theory of motivation that makes sense on Beebee and Mele's picture -- can come to terms with the "Mind Argument" (another argument against free will). For, on the Humean theory of motivation, all of our actions are determined by our beliefs and desires...which in any Humean world are fixed by physical facts. Accordingly, even if a particular agent had the "dual-ability" Beebee and Mele talk about, that dual-ability would still be the result of desires and beliefs out of the agent's control.

I'd be happy to discuss these worries further if you'd like, but they are the ones I have offhand. Anyway, thanks again for your comment and for drawing my attention to that paper!

Mark: Sorry for yet another addendum, but an analogy occurred to me that might make my distinction between the two types of determinism simpler and clearer.

In his work on "cosmological natural selection", Lee Smolin suggests that new universes (with new laws of nature) are borne out of the black holes of old universes. In other words, what appears as a black hold in one universe is a "white hole"/Big Bang of another universe.

On this picture, everything is physical. Each universe is *caused* by some previous universe. (No violations of causal determination).

However...the laws of each universe *cannot* be predicted, even in principle, by the laws of the previous one. This is because black holes, by their very nature, "tear apart" all information.

So, there's a *sense* in which what happens in a previous universe does not determine what happens in the later universe (the sense in which determination=necessitation).

However, there is also another sense in which the previous universe *does* determine what happens in the later one...as the later one is the causally downstream result of (nothing more than physical) processes of the previous one, the black hole, etc.

And, I want to say, it is the latter type of "determinism" ("causal closure of the physical") that is really the crux of things in the free will debate.

Does this help?

Hi Marcus, I'm afraid I'm not keeping up with everything in this thread, so sorry if my comment is missing something crucial, but I noticed this sentence: "even if counterfactual theories of causation were right [it] wouldn't suffice to show... that people in our world have a genuine *capacity* to do otherwise."

Well, I think such theories (as well as interventionist theories I favor, plus the sort of account C. List develops in the paper you cite) do show that people in our world have a *genuine* capacity (or ability) to do otherwise. They analyze that capacity in terms of possible worlds (and nearness to them or accessibility relations). You and other incompatibilists likely find such analyses wrong or unsatisfying. That's fine, but you don't get to say what they are trying to do. As far as I can tell, these analyses do not say that it is merely "our counterparts in nearby possible worlds could do those things [i.e., something other than we do]." They say *we* could do otherwise (have the relevant capacity for choice, etc. necessary for free will and moral responsibility).

What we *can* do (are able to do, have the capacity and/or relevant opportunity to do) is analyzed in terms of counterfactuals (or nearby possible worlds, perhaps with only micro-physical differences, or appropriate interventions, etc.). One reason I like such accounts is that they seem highly plausible to understand capacities (abilities, etc.) *in general*. If one is not inclined to think human capacities, abilities or powers (including free will) are metaphysically different in kind from others--of course, they are different in lots of interesting ways and degrees!--then one will want an account of capacities (etc.) in general, rather than one account for everything except for this very particular (ad hoc?) capacity that humans may have.

Hi Eddy: Thanks for your comment!

You write: "You and other incompatibilists likely find such analyses wrong or unsatisfying. That's fine, but you don't get to say what they are trying to do."

I don't see myself having said anything in my comment what people are "trying to do" (besides simply referring to the content of their views). I merely pointed out that I believe those analyses to be incorrect/unsatisfying (which you think is fine!).

In any case, I think you are absolutely right that these are legitimate matters of philosophical debate. I realize that I am very much in the minority in opposing the Humean worldview, counterfactual analyses of causation, etc. Lots of people find these views attractive/well-defended. I don't. These are the things that healthy debate are made of. You think counterfactuals are a highly plausible way to understand capacities in general. I don't, as I think counterfactual accounts end up presupposing either primitive necessitarian relations or Humean regularities in order to "make the counterfactuals come out right." Further, since I think the Humean regularity theory can't be right, I believe that all causation is essentially primitive necessitation -- which, finally, is why I don't believe there's anything especially ad hoc about the particular primitive capacity of libertarian free will that humans have (it's just another type of something--primitive necessitation--that already exists throughout nature).

You, of course, might have problems with each step I want to make here...in which case we sure have a lot to debate and talk about! ;)

Just to be clear, yes, we obviously disagree about legitimate matters of debate, and that's fine. But you were not simply referring to the content of these views in what I quoted, because *if* these theories are right, it would "suffice to show that people in our world have a genuine *capacity* to do otherwise." Perhaps I'm being persnickety and you just mean to say that they may show people have some sort of capacity to do otherwise but not a *genuine* one (but then "genuine" seems to mean "only one that has some sort of libertarian content"?). More substantively, are you suggesting that causation in general and causal capacities in general are primitives and then libertarian ones are just one more, among many sorts of, primitive causal capacities?

Hi Eddy: Thanks for your reply.

You write: "Just to be clear, yes, we obviously disagree about legitimate matters of debate, and that's fine. But you were not simply referring to the content of these views in what I quoted, because *if* these theories are right, it would "suffice to show that people in our world have a genuine *capacity* to do otherwise"."

I think I was referring just to the content of those views, and that we have a deeper substantive disagreement over the notion of a capacity. I actually don't think that if the theories you mentioned are right (at a substantial level), it follows that people would genuinely have the capacity to do otherwise.

I'll say more about this momentarily, but let me first speak to your next remark. You write: "Perhaps I'm being persnickety and you just mean to say that they may show people have some sort of capacity to do otherwise but not a *genuine* one (but then "genuine" seems to mean "only one that has some sort of libertarian content"?)."

This is one way to put it, but not the way I would prefer. I don't want to distinguish between "some sort of capacity" and "genuine capacities." I want to say there are just capacities simpliciter...and I want to give a very different gloss as to what capacities are simpliciter than you (and many other philosophers) appear to want to.

Let me explain.

The most standard way of understanding capacities, as you note, is in terms of counterfactuals. This is even true in the literature on dispositional properties (i.e. things like fragility, and capacity of a vase to break when struck hard by another solid object).

Although this way of understanding capacities is common (as are counterfactual theories of causation), I believe they are fundamentally mistaken, and that whether counterfactuals express (or represent) capacities depends on how we understand their truthmakers (i.e. the metaphysics of the world beyond them). Allow me to tell a very abbreviated story as to why I think this, and if it makes sense to you, then perhaps we can discuss/debate it in more detail.

To keep things simple, let's work with two toy metaphysical models:

(1) The Humean Model: the world is composed of objects and properties with no necessary relations between them (and so which, modally speaking, could be rearranged in any logically possible configuration).

This model has been very popular philosophically at least since Wittgenstein's Tractatus and David Lewis' work, and is sometimes called "the Humean landscape."

That's one metaphysical way the world could be. Here's another (very roughly):

(2) The Necessitarian Model: The world is not merely a bunch of independent objects and properties, but rather a bunch of objects and properties that bear certain primitive *necessitation* relations to one another.

Notice that these two models entail very different ultimate explanations of fundamental physical "laws."

The Humean Model will say that the ultimate reason that electrons orbit the nucleus of atoms is that this is a fundamental regularity in nature.

In contrast, the Necessitarian Model implies that this is false as an ultimate explanation. It says that the way electrons orbit nuclei is not just a regularity in nature, but the result of primitive relations in nature necessitating that behavior.

Now let us think about counterfactuals.

On the Humean Model of reality, my brain will instantiate some actual sequence of events (following from more basic Humean regularities in nature), thereby "causing" me to behave in various ways.

I believe that, on this model, there is strictly speaking a deep metaphysical sense in which *this* universe could not have been different. This universe is the series of regularities it is...and by Leibniz's law (different propertiesnon-identity), a universe with a different series of regularities would be a different universe.

Now, on this model, there are still counterfactuals true of this universe. The statement, "If X had not occurred, Y would not have occurred" (for some X, Y) could well be true, semantically speaking. Metaphysically speaking, however, (given possible-worlds semantics for counterfactuals) these counterfactuals are made true by other universes, not this one.

In short, I want to say that on a Humean model of reality, true counterfactuals are true only in a "vulgar" sense. They are not ultimately about our universe (which is, by identity, the only way it can possibly be), but rather statements the semantics are made true of by counterpossible universes.

In other words, I want to say that in a Humean world, counterfactual semantics is a kind of philosophical trickery, and counterfactuals do not express genuine capacities embedded in our universe (again, rather, they point to capacities in other universes). The only genuine capacities that exist in a Humean world are simply the regularities the universe actually instantiates.

Now, I expect you will have problems with this analysis -- and I am happy to discuss those -- but before I move to the second part of my analysis (the Necessitarian Model), let me extend my analysis to the case of free will.

On my analysis, if we live in a Humean world, then metaphysically speaking, there is one (and only one!) sequence of actions that my brain has the capacity to lead to: the actions I actually take.

Yes, there will be true counterfactuals about my brain in this world -- but, on my analysis, those counterfactuals are made true not by any (genuine) capacity I have, but rather a counterfactual capacity (i.e. another sequence) that a counterpart of me has in some other possible world. There is nothing *I* can do (strictly speaking) to bring about any other counter-possibility. I am ultimately metaphysically stuck, in a Humean world, instantiating whatever actual sequence I instantiate.

This, in a nutshell, is why I do not think that counterfactuals express capacities in a Humean world. The only capacities that exist in Humean worlds are the actual sequences that occur within them. All other "counterfactual capacities" are a play on language referring to counterfactual worlds no one and no thing can ever bring about.

Now turn to a Necessitarian world. Here, things are very different.

Counterfactuals in a necessitarian world are not ultimately made true by counter-possible worlds, but rather by necessitation relations in this world. This is why I think all and only necessitarian relations are genuine capacities. They *bring-things-about* in an "oomphy" manner that does not exist in any Humean world.

Long story short, I side with a necessitarian worldview. I think we have to posit primitive necessitarian relations in order to explain why electrons orbit atomic nuclei the way they do, and that libertarian free will is a special type of primitive necessitarian capacity -- a primitive capacity to choose between different options, in a manner that "oomphily" pushes our consciousnesses into a future path of the multiverse (as diagrammed in Fig. 1 of "A New Theory of Free Will").

So, to conclude:

(1) I think real capacities exist in Humean worlds, but that such capacities are exhausted by (or identical to) mere regularities.
(2) I think counterfactuals are true of Humean worlds, but do not express (or embody) capacities in those worlds (rather, they refer to actual sequences in other worlds).
(3) I think real capacities exist in Necessitarian worlds, but that these are primitive relations in those worlds.
(4) I think counterfactuals are true of Necessitarian worlds, but in virtue of necessitation relations, not possible worlds.
(5) I think libertarian free will is a special type of necessitarian relation in an otherwise Necessitarian world.
(6) I think the Humean model of reality is (probably) false and the Necessitarian model (probably) true.

Does this clarify things?

Hi, Marcus,

I'm not sure I'm getting the theory right, have a couple of doubts, questions, and worries:

On the issue of whether this will be libertarian freedom, my main worry is that I still do not see how this would be freedom, or how determinism of either sort would preclude it, or it would be worse, or less free, etc.
As for the "up to us" argument, I don't see any good reason to accept it: just because my actions were caused, does not mean that my choices are not up to me, as far as I can tell.
I'm not sure in which sense they can be up to me on indeterminism if they can't on determinism.
I read your reply here, but I would have to say I'm puzzled in the other direction: regardless of whether there are earlier determining causes, we (assuming there is a continuing "we", and the indeterminism doesn't get too far) would be causes of what we bring about in either case, and also, we wouldn't be all of the causes of whatever we bring about, but some of them.
We aren't be the causes of earlier events, but I don't see how that's a problem, and we're not compelled just because we're caused, in my assessment. Is this one of the issues where you have different intuitions?

On the issue of causal closure, you mention that the causal closure of the physical is the crux of things. But why would that be so?
If there are laws of the non-physical, it seems to me the arguments could be run just as well - and I would similarly reject a non-physical parallel to Van Inwagen's argument, but physical vs. non-physical does not appear to be the issue to me.
Would you propose that there would be no laws of the non-physical, or that somehow they would be relevantly different?

On the issue of determinism, if past states of me plus other past states plus any laws of the world (physical or not) do not bring about (at least, usually) my present state, what brings it about? And if they do, how is that libertarian freedom?

On a different note, do you think your view is compatible with theistic views under which God knows everything that will happen?

Hey Marcus,

Thanks for another helpful post and discussion! I think your approach to the consequence argument is interesting but a bit unsatisfying since ultimately you accept that the consequence argument is sound relative to our reference-frame, L. I don't want to concede that much to the libertarian! Not that it isn't tempting!

I'm especially interested in your responses to Al and Eddy. Both Humean compatibilism and Eddy's capacities view are the kinds of compatibilist views that I find more satisfying since they allow us to give a more resounding criticism of the consequence argument.

One quick point about your critique of Eddy's view is that his view does not appear to depend on a Humean view of laws. There are alternative views of laws of nature beside what you call the Humean and Necessitarian views. So I'm not sure how your critique of Humeanism would sweep Eddy's view along with it.

Marcus, thank you for your in-depth consideration of my comment. I think I understand the difference between the A and B versions of determinism. But I don't agree that your proposed model gives us B version determinism. It may well *appear* to us that we have B version determinism, but that would be an illusion. And I think you need (or at least want) it not to be an illusion.

To see that it is an illusion, consider the standard gendanken experiment of resetting the universe to its previous state and recording the outcomes. Because of the quantum nature of the universe, we won't just be looking for a change of outcome (that's for A version determinism). Instead we'll be looking for a change in the probability of outcomes. The in-universe observer will look at the physical state of the universe and predict that the ratio of "up" to "down" results will be p to (1-p). But if the Marcus mind has libertarian freedom, the ratio could be *anything* -- Marcus-mind could always choose up; Marcus-mind could always choose down; Marcus-mind could flip a mental coin. The Marcus-mind could choose "up" for the first ten million iterations and "down" for the following ten million -- and there is no value of p that works as an explanation for *that* result.

Thus the *actual* probability of "up" is not a function of the physical world at all; it is a function of the Marcus-mind. Thus the future is not necessitated by the *physical laws* -- which is what B says it is. It's only the in-world observer's inability to carry out the experiment that prevents discovery of non-physicalist causation.

Now I haven't been following the whole debate (I've been away on vacation), so I don't know whether apparent determinism satisfies your goals, but it doesn't fit with what you wrote in the article above:

//Such finding would mean that determinism is false relative to the higher reference-frame. But, determinism would still be true relative to our reference-frame//

Determinism is not true (not even in the B version) in our reference frame; it just *appears* to be true.

Hi Angra: Thanks for your comment!

You write: "On the issue of whether this will be libertarian freedom, my main worry is that I still do not see how this would be freedom, or how determinism of either sort would preclude it, or it would be worse, or less free, etc."

Can you say a bit more about why you don't think it would be libertarian freedom or freedom? I've said in my responses in previous threads why I think ordinary deterministic compatibilism would be worse/less-free. People have long worried that compatibilism is "freedom lite", for the simple reason that it says we are free even though our actions are (1) entirely determined, by (2) laws of nature out of our control. If X is out of my control, but X *determines* my behavior, how is my behavior under my control...except in the (metaphysically weak) sense that it is executed by my brain and body. Now, as I said in my post on free will contextualism, I'm happy to admit compatibilist freedom as a "vulgar" kind of freedom. But, I think, along with entire hordes of incompatibilists, that it is a pretty weak sense of freedom.

Anyway, I want to emphasize that my theory is an empirical one. One of the things I don't much like about existing free will debates is that they are about which conceptions of free will are "satisfying" and "unsatisfying." I don't like this because science is often "unsatisfying" even when it is true. If the predictions my theory makes are verified, then it is *true*...even if it doesn't satisfy people!

Next, you write: "I'm not sure in which sense they can be up to me on indeterminism if they can't on determinism."

I think I've addressed this point in the original post and my earlier reply to Mark Young. My account denies that reality is indeterministic or deterministic simpliciter -- just as the theory of relativity denies that there is such a thing as absolute motion. On my account, the world appears causally deterministic from our reference-frame, with *indeterministic* laws appearing to causally determine our actions. But this is only true relative-to-our-reference-frame. From the standpoint of the higher reference-frame, our world is partly deterministic, but libertarian freedom intervenes to *comprise* some of its "laws."

You write: "We aren't be the causes of earlier events, but I don't see how that's a problem, and we're not compelled just because we're caused, in my assessment. Is this one of the issues where you have different intuitions?"

My theory is not based on "different intuitions." It's a model of reality entailed by a number of independent hypotheses from physics and metaphysics, and it makes empirical predictions -- which, if verified, would verify the model.

You write: "If there are laws of the non-physical, it seems to me the arguments could be run just as well - and I would similarly reject a non-physical parallel to Van Inwagen's argument, but physical vs. non-physical does not appear to be the issue to me. Would you propose that there would be no laws of the non-physical, or that somehow they would be relevantly different?"

Great question! You are right about this much: *if* there are non-physical laws, then the problem of free will arises again on my model. I address this worry in "A New Theory of Free Will", claiming that because the non-physical exists in a reference-frame inaccessible to us, we can never know either way. The most we can ever know is the disjunctions: either (1) our wills are entirely self-caused in the higher reference-frame, or (2) our wills are determined by non-physical laws. This disjunction may not be satisfying, but again, I don't think it's the job of philosophy or science to be satisfying.

Lately, though, I've begin to toy with the idea that non-physical laws are impossible. Roughly, the argument goes like this: laws have to be in some sense "programming." Remember, on my account, reality is functionally identical to a P2P videogame. Programming, however, is fundamentally quantitative. Our world's laws conform to equations that can be expressed in terms of numbers. Their following those equations is what makes them *laws*. Consciousness, however, is fundamentally quantitative, and so, by definition, cannot be reduced to any kind of "programming" or equations. Thus, I want to say, there cannot be non-physical laws. Non-physical laws are like square-circles. They're impossible. Saying that something qualitative *has* to obey something quantitative (i.e. laws) is a category-mistake.

Your second-to-final comment is: "On the issue of determinism, if past states of me plus other past states plus any laws of the world (physical or not) do not bring about (at least, usually) my present state, what brings it about? And if they do, how is that libertarian freedom?"

I'm having trouble wrapping my mind around these two questions. Could you perhaps clarify them a bit?

Finally, you ask: "On a different note, do you think your view is compatible with theistic views under which God knows everything that will happen?"

Unfortunately, I've never thought of the issue of divine foreknowledge very carefully, and I don't feel prepared to give a good answer at this moment. Here, though, is one way to think about it. Once we've finished playing a P2P online videogame, we can rewind the game and play its recording. Every move we made will play out as though it is unavoidable. If we think of God as the person watching the recording, he can know every libertarian choice we made beforehand.

Thanks again for your comment and questions!

Hi Joe: Thanks for your comment, and for the kind words. I'm glad you're finding things interesting!

You write: "I think your approach to the consequence argument is interesting but a bit unsatisfying since ultimately you accept that the consequence argument is sound relative to our reference-frame, L. I don't want to concede that much to the libertarian! Not that it isn't tempting!"

As I note in my comment to Angra, I think it's really important to emphasize that my account is not intended to be "satisfying." It is intended to be true. And sometimes true things are unsatisfying. For instance, general relativity is unsatisfying. Many people are wont to protest that space and time *must* be absolute. But they're not. General relativity (as far as we now know) is true, whether we like it or not. And I want to say something similar here. Insofar as my model makes predictions, the question of whether it is "satisfying" is irrelevant. If its predictions are verified, it is (probably) true. And if they are falsified, then is false!

You write: "One quick point about your critique of Eddy's view is that his view does not appear to depend on a Humean view of laws. There are alternative views of laws of nature beside what you call the Humean and Necessitarian views. So I'm not sure how your critique of Humeanism would sweep Eddy's view along with it."

Thanks for pointing this out! As I mentioned in my reply to Eddy, I adopted the Humean/Necessitarian dichotomy as a simplification to bring out certain distinctions I think are philosophically important. I'm not sure whether Eddy's views depend on a Humean model of reality (though, could you tell me which model of reality he's working with?). However, insofar as Eddy appears to want to make sense of capacities in terms of counterfactuals, I still think my main point is relevant -- and that was that counterfactuals do not, in my view, necessarily express genuine capacities, and they only do if necessitation relations obtain (in which case it's not ultimately the counterfactuals doing the metaphysical work, it's the necessitation relations).

Hi Mark: Thanks for your reply, and for pressing me on this!

First, you write: "It may well *appear* to us that we have B version determinism, but that would be an illusion. And I think you need (or at least want) it not to be an illusion."

Truth be told, I keep going back-and-forth on this (i.e. I don't have my mind entirely made up, I'm still working through it). In "A New Theory of Free Will", I make the move you state here -- saying that determinism (in sense B) will *appear* to be true to observers in the lower reference-frame even though it's not (i.e. determinism as illusion). I think -- the present post notwithstanding -- this is still the route I want to go, and that I don't really need (or should want) the stronger view I express in the current post. After all, the move I make in this post is that there is no privileged reference-frame (determinism is true-relative-to-L, libertarianism true-relative-to-H). This sort of seems wrong, doesn't it, as unlike in relativity (where there is no privileged reference frame), in my model it would seem as though H is the "correct" reference frame and L merely an impoverished reference-frame in which observers cannot see everything that's actually going on (i.e. the stuff in H!).

If this is right, then I should fall back to the illusion view (the view which, again, I've stated in my published work). And I think that's probably good enough for all of my purposes! I guess the reason why, in this post, I adopt the stronger view -- all truths are relative-to-reference-frames (i.e. L and H, respectively) -- is that I'm sort of drawn to Einstein's Machian postulate, that we should always define truths-relative-to-observations, since it is the denial of this postulate that has so often gotten philosophy and science in trouble.

Anyway, this what I'm inclined to say for now:

(1) I think you're right, and that I should probably stick with the "illusion of determinism" view that I defend in "A New Theory of Free Will", but

(2) I'm not entirely convinced of this, and that *perhaps* the view I adopt in this post (no illusion, just different frames-of-reference) might be correct.

How does that sound?

Next, you write: "To see that it is an illusion, consider the standard gendanken experiment of resetting the universe to its previous state and recording the outcomes. Because of the quantum nature of the universe, we won't just be looking for a change of outcome (that's for A version determinism). Instead we'll be looking for a change in the probability of outcomes. The in-universe observer will look at the physical state of the universe and predict that the ratio of "up" to "down" results will be p to (1-p). But if the Marcus mind has libertarian freedom, the ratio could be *anything* -- Marcus-mind could always choose up; Marcus-mind could always choose down; Marcus-mind could flip a mental coin. The Marcus-mind could choose "up" for the first ten million iterations and "down" for the following ten million -- and there is no value of p that works as an explanation for *that* result."

I still think you're dealing with A-determinism here. Changes in probabilities are just changes in which *laws* the observers will posit. In your gedanken experiment, observers will say, "It is a law of nature in our world for quantum wave-functions to evolve and change unexpectedly in response to neural changes resulting from the results of earlier collapses." As such, they will still think their universe is causally closed under the laws of physics (B-determinism). They will just think their laws of physics are very strange, and can dynamically evolve in a wide variety of possible ways in human brains.

Here's a way to bring this out (a way that I allude to in "A New Theory of Free Will"). Consider an ordinary videogame like "Halo." In Halo, you control your character from the outside -- so there's a sense in which observers in the game cannot predict your behavior (i.e. A-determinism will appear to them to false). Still, once you're done playing a game of Halo, you can rewind the game you just played, press play, and watch a recording of the game. And of course *every* event will play out inexorably, leading to the end result of the game you just played. Let's think about this, then, from the standpoint of observers in the game. Although A-Determinism appears to be false to them (i.e. indeterminism is true vis-a-vis laws), B-Determinism must appear to be true to them. Whatever indeterministic laws they observe -- including, shifting Marcus-wave-functions, Mark-Young-wave-functions, etc. -- will *have* to appear closed under the laws of physics. Why? Because their world really is executing an unbroken series of information (a series which, again, if you were to rewind it to beginning, would lead inexorably to one, and only one, end result). Now again, you're right, observers would not be able to predict the evolution of the laws (in human brains). But this is just A-determinism. Whatever dynamic evolution of the laws in fact obtains (including any shift from a Normal-Marcus-wave-function to a New-Marcus-Strategy-wave-function) will appear to them to have a purely *physical-causal* explanation. Again, the explanation will go something like this: "At time T, Marcus' brain instantiated the Marcus-Wave-Function. The Marcus-Wave-Function makes collapse into state A the most probable--in which case his brain will continue to obey the Marcus-Wave Function. However, collapse into state B is very improbable but possible on the Marcus-Wave-Function, and if state B occurs, Marcus' brain will shift from the Marcus-Wave-Function to a New-Marcus-Strategy-Wave-Function in virtue of neural-updating caused by [improbable] state B." Suppose then, to keep things simple, A and B are the only possible states. Notice, finally, that:

(1) If A obtains, people in the world will say everything was *causally* determined by nothing more than A and the Marcus-Wave-Function.

(2) If B obtains, people will say that everything was *causally* determined by the Marcus-Wave-Function (at T), improbable collapse-event B occurring (at T+1), which in turn physically caused neural updating, which in turn physically caused Marcus' brain to instantiate the New-Marcus-Strategy-Wave-Function, etc.

In other words, although their world will be entirely unpredictable in sense A (A-Determinism is false), no matter how things turn out B-determinism will appear to be true to them. No matter how things turn out, they will have *some* purely physical-causal story to tell about why things turned out that way!

Does this make sense? Thanks again for your comment!

Hi Marcus,

Thanks for your reply, and sorry for the long-windedness of this one - but I'm trying to clarify my questions, points, etc.
On the issue of why it's freedom, you say " People have long worried that compatibilism is "freedom lite", for the simple reason that it says we are free even though our actions are (1) entirely determined, by (2) laws of nature out of our control. If X is out of my control, but X *determines* my behavior, how is my behavior under my control...except in the (metaphysically weak) sense that it is executed by my brain and body."
I would say that if there are laws of nature - or laws of supernature, or laws of all reality, etc.; whether it's brains or souls isn't the point in my view - but we're not compelled (either by external threats or mental illness, etc.), we (whether it's brain or soul) make free choices. Also, we have control over our behavior, since we can freely choose what to do (e.g., I freely choose to write this post. I control what I write).
Now, I get you may say that's "metaphysically weak", but I do not understand why it would be weak, and what alternative would be stronger. But I don't see in which sense any alternative would be stronger. In fact, if there are no such laws and "choices" just may happen to me, then it seems to me that that's not control, or freedom. It's an unfortunate sort of randomness.
Yes, granted, hordes of libertarians may say that choices don't just happen to us, but we make them in a self-caused way. But I don't grok what it would mean to make a choice in a self-caused way, or something like that: if the choices are not brought about by my previous brain/mind/soul states (including evaluative attitudes, deliberation, probabilistic assessments, etc.), in which sense am I making a choice?

Regarding the predictions of the theory, if they were true, that wouldn't make me conclude that there is some stronger freedom than we would otherwise have, but rather, than the laws are different from what QM postulate. I don't see that as more or less free. While I think indeterminism may preclude freedom depending on how it works (since a "choice" might happen to us regardless of our previous states, etc.), that depends on how the indeterministic stuff happens (i.e., there are forms of non-determinism that do not affect freedom), and it's not more of a problem under the new laws than under indeterministic QM laws - in fact, without any good evidence that indeterminism of the wrong kind is getting in the way of freedom, I would say that we do have freedom.
But I'm not sure why to call this libertarian free will, but rather, our brains/souls are working differently from the way we thought they were.

Regarding my not being sure in which sense they can be up to me on indeterminism if they can't on determinism, I'm afraid I still don't grok it, though this might not be fixable (i.e., so far, I've not been able to grok it despite reading what many libertarians say). I get that there would be two different levels, etc. (though I would probably characterize the whole thing as indeterministic, with the lower-level appearance being an illusion, but let's say otherwise, since it's not the issue I'm getting at here), but I still do not see why indeterminism on the higher reference frame would give us any more freedom than determinism on the higher reference frame as well.

With respect to my point about different intuitions, what I'm trying to get at is that if the predictions are correct, intuitively I don't see why that would mean greater or lesser freedom than there would be if they're not - again, except if the indeterminism is getting in the way of freedom, but that depends on the sort of indeterminism, and I'm not sure why the proposed rules would be worse or better, freedom-wise.
It may be simply that I'm not able to grasp what sort of thing you (and yes, hordes of incompatibilists ;)) are describing.

On the issue of non-physical laws, you consider the two options of our wills being "entirely self-caused", or "determined by non-physical laws". But the difficulty is that I don't understand what "entirely self-caused" would mean, or in any case how the non-physical laws scenario would make us unfree, free-lite, etc.; on the contrary, I see the alternative between laws (physical, non-physical, or laws of any sort) and lawlessness as an alternative in which the latter is a problem for freedom, if it doesn't entirely preclude it. The difficulty for me seems to be one of understanding of what it means to say an agent is free in a self-caused, libertarian sense.

On the issue of the impossibility of non-physical laws, I take it you mean consciousness is fundamentally qualitative, not quantitative?
If so, I'm not sure in what sense that would be so, but laws seem to me to be possible as well. Say, a combination of experiences e1, ...., en (and/or experiences plus some interaction with particles, or whatever) bring about experience e(e1, ..., en), etc.; the laws may be laws of the non-physical and also the physical; e.g., I would expect that if we were souls, the laws would not be separate sets of laws for the physical and non-physical, but some laws of reality encompassing both, and regardless of whether they're programmable in a finitistic manner.

As for what brings about my present state, what I'm getting at is that, on a deterministic account, my previous states (conscious states, deliberations, etc.), together with other states of the world (physical or not, etc.), bring about my present state (including my conscious experience). But on the situation you describe as "self-caused", what would bring them about, and why would it be called "self-caused" rather than random?

On the God issue, maybe I'm missing something (if so, could you clarify, please?), but it seems to me that when we consider all of reality (not just the physical universe) we would then get determinism of type (A), because there is necessitation of later events from previous events - which include God's mental state, regardless of whether his knowledge takes the form of beliefs, some intuitive apprehension of things, etc.
In other words, in any world with the same past (and so, in which God was in the past in the same mental state), everything will happen in the same way (God can't be mistaken), so previous states necessitate all later ones, including our free choices.
So, it appears to me that if your theory is compatible with that, it's compatible with type (A) determinism, at least if God exists and we're talking about determinism of all of reality, rather than just the physical universe. But I don't think God would make a difference in terms of compatibility with type (A) determinism.
For example, we may consider your analogy with P2P games. Once we've finished playing a P2P online videogame, we can rewind the game and play its recording. Every move we made will play out as though it is unavoidable, as you say.
But if someone (God, demon, human, or any sort of agent) could watch a "recording" (or whatever we call it, since the game hasn't been played yet) before we play the P2P game, and such that the "recording" in question cannot be wrong, then every move would be necessitated by previous states (even though we would still have a choice under a compatibilist understanding); more precisely, the "recording" would necessitate our choices while playing the game.

Thanks for the explanation, Marcus. I'm satisfied that you can get by with the mere illusion of determinism, so I won't press further on that.

One more question, tho', relating what we've been discussing to the Humean/necessitarian discussion. Wouldn't the laws discovered by the in-world observers be partly (or even largely) of a Humean nature rather than necessitarian? That is, the laws (some of the laws?) reflect only regularities in the non-physical world, with nothing in the physical world necessitating those regularities. And, yes, the non-physical world may necessitate some changes in the physical world, but from the in-world point of view wouldn't the hypothetical freedom of Marcus-mind to have done otherwise fall prey to the same objections you raised to Humean laws -- namely that if it had made a different decision it would have been a different Marcus-mind -- not *you*, but your counterpart in another possible (but non-physical) world?

Thanks again for an entertaining discussion.

Hi Mark: Thanks for your reply, and I'm glad to hear you're satisfied by that part of my response!

You write: "One more question, tho', relating what we've been discussing to the Humean/necessitarian discussion. Wouldn't the laws discovered by the in-world observers be partly (or even largely) of a Humean nature rather than necessitarian? That is, the laws (some of the laws?) reflect only regularities in the non-physical world, with nothing in the physical world necessitating those regularities."

No, they wouldn't be mere regularities -- and indeed, I think the P2P Model shows what is wrong with the Humean theory of reality. Consider what goes into an online simulation. On the one hand, you have the game DVD. The game DVD is just digital data that encodes possible features of the game environment (i.e. possible physical states-of-affairs, e.g. rocks, people, etc.). All by itself, the data is like a Humean landscape. Nothing necessitates anything else.

But now notice: the DVD alone is *not* enough to comprise a world. It is just data. The DVD has to be *read* by an observation device (the game console/computer) processing the information in order for a simulated reality to emerge. And the processing device instantiates clear analogues of necessitation. It does not produce mere regularities. If the game DVD encodes X-->Y, then anytime the console reads X, its processor *necessitates* that Y comes next.

This entire picture illustrates -- very broadly -- why I (following part II of Gregg Rosenberg's 2005 book) don't think a Humean landscape alone can comprise a reality. A Humean landscape is, strictly speaking, just "static", motionless information. In order for time to pass, there must be something else -- an external game processor or consciousness -- to *process* that information.

Your final question is: "Wouldn't the hypothetical freedom of Marcus-mind to have done otherwise fall prey to the same objections you raised to Humean laws -- namely that if it had made a different decision it would have been a different Marcus-mind -- not *you*, but your counterpart in another possible (but non-physical) world?"

I don't think so, because if we understand libertarian free will (the brute capacity to choose among different possible futures) as part of the *essence* of nonphysical minds, then (by Leibniz's law) different possible futures are consistent with the same me making different choices (i.e. if I had chosen otherwise, the mind in that alternative timeline would still be me).

Thanks again, and I'm glad you've found it an entertaining discussion. I've found it entertaining -- and helpful! -- as well.

Hi Angra: Thanks for your reply, and for clarifying your questions.

You write: "I would say that if there are laws of nature - or laws of supernature, or laws of all reality, etc.; whether it's brains or souls isn't the point in my view - but we're not compelled (either by external threats or mental illness, etc.), we (whether it's brain or soul) make free choices."

I think that you and I just have different intuitions here (like most compatibilists and incompatibilists do). You think that we make free choices if we're not compelled. I agree. But I think if we lack libertarian free will, then we *are* compelled to act by external forces...namely, laws of nature. In contrast, you -- as a compatibilist -- want to say this isn't the right kind of compelling to undermine freedom. Like other incompatibilists, I disagree. I think it is absolutely the right kind of compelling to undermine freedom. It undermines it (unless my libertarian compatibilist theory is true) because: (A) We can't choose the laws, and (B) the laws compel our behavior.

Anyway, I think you and I will probably come to a dialectical stalemate here. You have the compatibilist intuitions, I don't. But fortunately, my theory doesn't turn on your or my intuitions. It turns on the *facts*. If the theory is verified, it's true. If it's not, it's false!

Next, you write: "Now, I get you may say that's "metaphysically weak", but I do not understand why it would be weak, and what alternative would be stronger. But I don't see in which sense any alternative would be stronger. In fact, if there are no such laws and "choices" just may happen to me, then it seems to me that that's not control, or freedom. It's an unfortunate sort of randomness."

The stronger alternative, I believe, is our choices being determined by us as libertarian, nonphysical minds (i.e. genuine uncaused causes). Now, your worry is, this is just randomness. To which I reply: no, it's not. Libertarian free will is a *brute* capacity to make choices on the basis of conscious experience -- experiences which have a great deal of content (we have conscious beliefs, experiences of sensible inclinations, etc.), but which do not *necessitate* our behavior by themselves (i.e. our conscious experiences give us reasons to act, and we choose which reasons to act upon).

Which brings us to the crux of your worries. You write: "Yes, granted, hordes of libertarians may say that choices don't just happen to us, but we make them in a self-caused way. But I don't grok what it would mean to make a choice in a self-caused way, or something like that: if the choices are not brought about by my previous brain/mind/soul states (including evaluative attitudes, deliberation, probabilistic assessments, etc.), in which sense am I making a choice?"

I think this is the best worry about libertarianism to have, but that I can meet it. First, although on my account our choices are not "brought about" (or causally determined) by antecedent mental states (evaluative attitudes, probabilistic assessments), these states nevertheless play real roles in my picture. We have beliefs, desires, evaluative attitudes, etc., and *choose* which ones to act upon. So, the claim isn't that we choose entirely from nothing. Rather, the claim is that although there are certain inputs to our deliberative processes (sensible inclinations, beliefs, etc.), we have a brute ability to decide which inclinations, beliefs, etc., to act upon.

Now, you might say, "I can't grok what you mean by that brute ability." To which I reply: you're not supposed to be able to "grok" it! It's a brute capacity that is unexplainable in terms of anything more basic.

Now, this might not satisfy you, but again, I say it is not the job of theory to satisfy. It is the job of theory to be *true*. And, I think there are reasons to think it is true that similar brute, inexplicable capacities suffuse nature at a fundamental level.

Indeed, I think this is something that physicalists/naturalists do not reflect upon nearly enough. What *makes* electrons orbit the nucleus of atoms the way they do? I do not think a Humean regularity theory can be correct, as there must be some explanation for why the regularity obtains (there is literally an infinite number of ways a given regularity can fail to hold, compared to only *one* way in which it can hold!). The only way to explain Humean regularities, I believe, is in terms of brute necessitation. So, the kind of brute necessitation that you claim to have trouble "grokking" in the libertarian case is -- as much as it might not satisfy you -- something that I think we have to attribute to *every* fundamental dynamical law of nature...and for which libertarian free will is merely a very special case.

As you write: "The difficulty for me seems to be one of understanding of what it means to say an agent is free in a self-caused, libertarian sense."

I agree that it is difficult to say, precisely because this is where we run up against the limits of explanation! As Wittgenstein famously said, all explanations must end somewhere. Physicalists (such as yourself?) end up positing unexplainable "magic" whether they like it or not. Why is there a universe rather than none? The physicalist must answer: it's magic/unexplainable. Why are the laws what they are? (The answer, again, seems to be: it's magic!). Why do regularities of natural laws hold without fail (given that there are an infinite number of possible ways any given regularity could fail to hold)? Again, "magic."

There is no way to do physics/metaphysics without positing unexplainable "magic" somewhere (and indeed, everywhere!). The way I see it, the real question is what kind of magic best explains the world we perceive. And, I will claim, if the predictions of my theory are verified, we should believe in libertarian magic.

Which addresses your question: "As for what brings about my present state, what I'm getting at is that, on a deterministic account, my previous states (conscious states, deliberations, etc.), together with other states of the world (physical or not, etc.), bring about my present state (including my conscious experience). But on the situation you describe as "self-caused", what would bring them about, and why would it be called "self-caused" rather than random?"

My answer is: *you* bring them about. Your present states (your beliefs, desires, inclinations, etc.) are the result of your past, brute libertarian choices made in response to your experience of earlier beliefs, desires, inclinations, etc.

This brings us to your worry of libertarianism as lawlessness. You write: "...and lawlessness as an alternative in which the latter is a problem for freedom, if it doesn't entirely preclude it."

Well, the picture I'm giving isn't lawless. There are lots of *physical* laws, including laws about how one's brain will develop as a result of physical perception, earlier libertarian choices, etc. These lawful elements are then, on my account, experienced by you as a conscious subject, and you have a brute ability to choose *which* beliefs, desires, inclinations, etc., to act upon. This is not randomness/lawlessness. It is *lawfulness* experienced by a subject with brute abilities to decide which of the "lawful" things (beliefs, desires, etc.) provide the best reasons to act.

Your final set of questions concern divine foreknowledge. You write: "But if someone (God, demon, human, or any sort of agent) could watch a "recording" (or whatever we call it, since the game hasn't been played yet) before we play the P2P game, and such that the "recording" in question cannot be wrong, then every move would be necessitated by previous states (even though we would still have a choice under a compatibilist understanding); more precisely, the "recording" would necessitate our choices while playing the game."

My reply is that no one -- not even God -- can watch a recording before it is played. So, what I want to say is something like this:

(1) If this is the first time we're playing the game, then no, God cannot know how the game will play out (since we have libertarian free will, a brute, unexplainable ability to choose among options).

However, (2) If we have already played the game, then God *can* know how the game will play out (just as we can when we rewind a videogame).

Thanks again for your comment/questions. They're very helpful!

Hi Marcus, and thanks again for your reply.

A couple of issues for the purposes of clarifying some of my views (sorry if I wasn't clear enough) and addressing a question.

a. On the issue of compulsion, just to clarify, as a compatibilist, I wouldn't say that laws of nature, or supernature, or generally whatever laws of reality there are, do not compel us. In other words, I'm not only inclined to say that that's not the right kind of compulsion, but that that is compulsion at all - i.e., that's not what "compulsion" means, as I grasp the term. Rather, that's we making choices, as I see it.

b. I'm not a naturalist/physicalist, or at least I don't describe myself as such. In fact, I have issues with the definitions of the stances in question, so I take no stance. I do believe substance dualism is very probably false, but I'm not assuming that in this context.

c. With regard to the issue of whether the theory is verified, it's true, else it's false, and it's not supposed to satisfy, that's all fine.
The difficulty for me is that if the experiments yield the results you predict, I still wouldn't see that as confirmation of libertarianism. I would see it as confirmation that things work differently in brains (perhaps, just in human brains, though I find this extremely improbable), and even evidence that consciousness has something to do with the difference, but on the other hand, I wouldn't think that that confirms substance dualism, and - more importantly in this context - even conceding that the results would establish substance dualism, souls, etc., I still would be a compatibilist, and I still wouldn't be able to make sense of libertarian choices, in the wrong way of not making sense (i.e., not just mysterious processes, or primitive concepts, but a conflict).

d. Whether our behavior is necessitated is not the issue, in my view.
For example, let's say I freely choose to pick a number from 1 to 1000. I pick...527. How did I pick it? I consciously made a choice to pick a number, and then the number 527 "appeared" in my head. It seems to me that my choice (among other facts) - a choice made by (part of) my brain, but that is not the point here; we may assume I'm a soul - caused part of my brain to do some unconscious processing and pick that number.
Now, perhaps, that part of my brain used an entirely deterministic mechanism to pick 527, or at least one that is deterministic when working normally (in particular, assuming no absurdly improbably quantum freak events). That seems probable to me, but that is not the issue. If it turns out that indeterminism is true, and my brain has an indeterministic random number generator, and my picking 527 was not determined by previous states, then my behavior is not necessitated by previous events or states, and the choices based on that choice are not determined by anything earlier than that choice, etc., but this sort of indeterminism wouldn't be a problem for freedom, since my choice was to pick a number - with no requirements as to whether my brain picks it indeterministically or deterministically.

As to your point about previous states and a brute choice, you say "We have beliefs, desires, evaluative attitudes, etc., and *choose* which ones to act upon. So, the claim isn't that we choose entirely from nothing. Rather, the claim is that although there are certain inputs to our deliberative processes (sensible inclinations, beliefs, etc.), we have a brute ability to decide which inclinations, beliefs, etc., to act upon."

I do agree that we make choices based on previous beliefs, desires, evaluative attitudes, etc., of course, but given a sufficiently long "etc." including the mechanisms for choosing between them, thus generating new beliefs, etc., I don't see what it would mean for there to be a "we" beyond that that could make a choice in the first place.
Even if we're souls, we (i.e., the souls) have previous states of the soul, etc., and a mechanism to make choices, which works in some regular way, deterministic or not (by "mechanism" I don't mean it has to be finitistic, just lawful).

Also, you say I'm not supposed to "grok" this capacity, since it's "a brute capacity that is unexplainable in terms of anything more basic."
I see that my saying I don't grok it wasn't clear; I will address this point in greater detail below:

On the matter of our present-day states, you say "My answer is: *you* bring them about. Your present states (your beliefs, desires, inclinations, etc.) are the result of your past, brute libertarian choices made in response to your experience of earlier beliefs, desires, inclinations, etc."

I would understand that my present state is the result of my past beliefs, inclinations, and a mechanism for picking new ones based on those - more precisely, the states of that mechanism up to the present -; let's even grant it's some mysterious thing, say element X.

So, my present state is brought about by my previous desires, attitudes, beliefs, etc., and element X - or the previous states of element X, it seems -; so far, I see no problem.
But then, element X would also have to work according to some laws - deterministic or not - as far as I can tell, else there would seem to be no connection between the present state and the previous ones - element X's outputs need to be connected to inputs -, and then intuitively I would say under that scenario that we make not choices but "choices" happen to us, assuming there is even a "we".

So, the difficulty here is this: according to the libertarian position, we only have freedom if element X is indeterministic, and also there are no laws that regulate the behavior of element X. But if I try to analyze the matter using my intuitive grasp of the relevant concepts, I assess that conceptually, we could have (and apparently do have) freedom regardless of whether element X is indeterministic or not, and also that there must be some laws (necessities or not) to connect the previous states with the present ones via element X so that our choices may properly be said to be our choices - and this, of course, is regardless of whether we call element X "a soul making a libertarian choice", or whatever we call it.

So, either libertarianism is false (because of the analysis above), or I failed to grok libertarianism, but this is the wrong way of failing to grok it, not the sense in which there is a primitive concept, or some entity we don't understand how it works, etc.

Rather, I failed to grok it because I didn't even understand what the libertarian is saying, because if I did understand it and I also grasp the relevant colloquial concepts about freedom, choice, etc., then libertarianism is false regardless of how the empirical evidence turns out to be. It's wrong on the basis of conceptual analysis alone. (okay, alternatively, I failed to intuitively grasp the colloquial concepts of freedom, choice, etc., but that seems very improbable for any adult human; alternatively, different people have concepts that are so different that we're talking past each other, and the matter can't be resolved, but that also seems very improbable).

This sort of problem is not at all what happens with a theory about electrons, regardless of how weird its predictions are; sorry if I didn't explain my worry clearly in my previous post.

Granted, this is not a difficulty I have with your particular libertarian theory, but with libertarianism in general: I've tried for years to understand what they're saying, and I don't - or else, I do and libertarianism is false.

But then, this is why I don't accept libertarianism, and wouldn't regardless of the empirical results (which, nevertheless, would be a major breakthrough if you're correct. I mean, libertarian or not, that would be a huge achievement in my view :-)). I conclude that either libertarianism is false on conceptual grounds, or I don't understand what it is. But if I don't understand what it is, it may or may not be true, but I have no good evidence that would make me conclude that it is true, so I don't accept it (and okay, I accept compatibilism because I tend to rely on my intuitive grasp of the concepts, but I think we all tend to do that when we do conceptual analysis; libertarians do so as well;)).

On the God issue, I'm not sure I'm getting your point correctly, but I'm getting that before we play, it seems God can't know what will happen - only after we do -, so libertarianism is incompatible with divine foreknowledge (for that matter, divine foreknowledge implies determinism of type A in my assessment, and for the reasons I briefly sketched in the previous post). Still, this was just my curiosity, so thanks for addressing it.

Also, again, thank you for your reply, and for addressing my points, questions, etc.

Hi Angra: Thanks for your reply, and sorry for the delay in responding. I was traveling all day yesterday and returned home to find my house's internet knocked out from a storm (it's still out). I will try to get a detailed response out to your latest comments tomorrow when I'm on campus and have web access again!

Just one quick point, Marcus. I don't think that classical compatibilist theories of the sort that Eddy and Al are talking about -- views that hold that (1) free will requires the ability to do otherwise and (2) that one has the ability to do otherwise is consistent with determinism -- require counterfactual analyses of ability let alone counterfactual analyses of causation. I might think that the truth of some counterfactuals are necessary for free will (or free action) and work a theory off of this insight.

Perhaps better: I might think that certain counterfactuals are sufficient for certain ability claims. Suppose it is true that if I tried to long jump at least 2', I would do so. That seems sufficient for my being able to long jump at least 2'. And if the former counterfactual is consistent with determinism, then classical compatibilism is true. (See Keith Lehrer's "Preferences, Conditionals and Freedom" for a similar example.)

Angra: Sorry again for taking so long to reply. It's been a rather rough week: first week of class, home internet knocked out, and now the flu! (Yes, the first week of class)

Anyway, thanks again for such a detailed reply. Here are my thoughts in response.

You write: "a. On the issue of compulsion, just to clarify, as a compatibilist, I wouldn't say that laws of nature, or supernature, or generally whatever laws of reality there are, do not compel us. In other words, I'm not only inclined to say that that's not the right kind of compulsion, but that that is compulsion at all - i.e., that's not what "compulsion" means, as I grasp the term. Rather, that's we making choices, as I see it."

Of course - I realize that's what you'll say as a compatiblist! But, as I argued in my post on contextualism about free will, I think there's clearly a deep metaphysical sense in which laws do compel. "Compel", in ordinary usage, means "force/oblige". But now, if all of my actions are determined by laws of nature, then there's a clear literal sense in which I'm forced or obliged to act as I do. The laws of nature include forces (of gravity, electromagnetism, etc.). The forces function, in a deterministic setting, to literally *force* our actions.

Now again, I'll agree with you that in a more vulgar sense (ordinary everyday contexts of praise and blame), they're not the "right kind of force" to amount to compulsion. But to say -- as you suggest -- that there's no sense in which laws compel/force behavior seems to me to do incredible damage to language. You want to say (1) laws don't force me to act in any sense, when quite literally (2) laws of nature include *forces* that determine my actions. That sounds like a plain contradiction to me.

But, it's also neither here nor there, at least in terms of Libertarian Compatibilism -- as my theory does not depend on intuitions about this.

Next, you write: "b. I'm not a naturalist/physicalist, or at least I don't describe myself as such. In fact, I have issues with the definitions of the stances in question, so I take no stance. I do believe substance dualism is very probably false, but I'm not assuming that in this context."

Fair enough on the first two points (proper definitions here are complicated, and we probably can't settle them here). Obviously, we'll have to disagree on the latter point (the truth of dualism).

Your next points are: "c. With regard to the issue of whether the theory is verified, it's true, else it's false, and it's not supposed to satisfy, that's all fine. The difficulty for me is that if the experiments yield the results you predict, I still wouldn't see that as confirmation of libertarianism. I would see it as confirmation that things work differently in brains (perhaps, just in human brains, though I find this extremely improbable), and even evidence that consciousness has something to do with the difference, but on the other hand, I wouldn't think that that confirms substance dualism, and - more importantly in this context - even conceding that the results would establish substance dualism, souls, etc., I still would be a compatibilist, and I still wouldn't be able to make sense of libertarian choices, in the wrong way of not making sense (i.e., not just mysterious processes, or primitive concepts, but a conflict)."

Here's why I don't think this is right. You would take verifications of my theory's predictions to simply show (without any explanation) that "brains just work different." The problem here, however, is the lack of explanation. Natural scientists and philosophers rightly balk at failures of explanation, precisely because when something in the world obtains, the natural question to ask is, "Why?" My theory not only gives an explanation (in terms of nonphysical processes in a higher reference-frame). That explanation predicts and r provides a unified explanation of a wide variety of other puzzling phenomena outside of human brains (quantum phenomena more generally).

An analogy: Ptolemaic astronomers could well maintain that some planets' epicycles "just work different" than others -- but this is precisely why we reject Ptolemaic astronomy in favor of Copernican astronomy. The latter provides a unified explanation where the former does not.

Your next point is: "d. Whether our behavior is necessitated is not the issue, in my view...I don't see what it would mean for there to be a "we" beyond that that could make a choice in the first place. Even if we're souls, we (i.e., the souls) have previous states of the soul, etc., and a mechanism to make choices, which works in some regular way, deterministic or not (by "mechanism" I don't mean it has to be finitistic, just lawful)."

Again, I don't blame you for having trouble understanding what it would mean for there to be a "we" beyond deterministic or indeterministic mechanisms for making a choice. That's because, on my account, the we is *primitive*. On my account, you are a non-physical mechanism of sorts -- a mechanism that is neither deterministic nor indeterministic, but rather fundamentally *self-causing*.

I do realize how this can seem unsatisfying, yet I've tried to explain why I think the world -- throughout nature -- must contain this kind of primitive necessitating "magic" at a fundamental metaphysical level (I don't even think we can explain why electrons do what they do, or why time seems to pass without primitive, necessitating magic). If you're not buying it, that's cool. I'll keep thinking about it, and try to come up with better arguments. ;)

Let me reply, however, to your worry towards the end of your comment. You write: "So, my present state is brought about by my previous desires, attitudes, beliefs, etc., and element X - or the previous states of element X, it seems -; so far, I see no problem. But then, element X would also have to work according to some laws - deterministic or not - as far as I can tell, else there would seem to be no connection between the present state and the previous ones - element X's outputs need to be connected to inputs -, and then intuitively I would say under that scenario that we make not choices but "choices" happen to us, assuming there is even a "we"."

I just don't see why element X "would have to operate according to some laws." On my picture, there are a lot of mental phenomena that do operate according to laws (i.e. our neural networks update to present our consciousness with desires, beliefs, etc.). However, the unknown X -- the nonphysical subject -- is a brute capacity to *decide* which of those desires, beliefs, etc. (if any) to act upon. Why must this X operate according to laws? What is incoherent about the proposition that this X -- me -- has the brute capacity to *choose* the laws it acts upon (i.e. in a Kantian way, willing oneself to act according to laws of pure practical reason)?

Your answer seems to be that you think libertarianism is false or incoherent because it doesn't jibe with the "relevant colloquial concepts about freedom, choice, etc." My reply, though, is that it's not *supposed* to jibe with colloquial concepts -- any more than, say, relativity jibes with colloquial concepts of space (it is really, really hard to wrap one's mind around relativistic space). In my view, it is not the job of philosophy or science to jibe with colloquial concepts -- as both have refuted colloquial concepts many times, and provided us with grounds to adopt new, unfamiliar concepts.

The way I see it, we should believe in whatever fundamental constituents of reality -- relativistic space, qualia, etc. -- our best verified, most explanatory theories affirm. I then believe that our best theory of reality may require us to believe that several aspects of reality -- necessitation of all physical laws, qualia (redness, yellowness), nowness, persons, and choice -- are all brute features of reality. Features that admit of no further explanation, but which explain lots of other things that need explaining.

Those brute things may seem bizarre -- and impossible to "grok" -- and yet, I want to say, our best theories may drive us to them. Maybe. We'll have to wait and see. ;)

The comments to this entry are closed.

Books about Agency


3QD Prize 2014: Marcus Arvan