Ghenadie Mardari on Free Will

Ghenadie Mardari, a grad student at Rutgers, is currently working on a dissertation on free will and has asked us to share a bit about the project to see what Gardeners think about it. So, without further ado, I post below the fold Mardari's own description of the project. I encourage you all to provide any comments you think would be helpful.

"I am working on a
libertarian model of free will, with the ambition to stay within the
confines of the intuitive (folk) understanding of the concept. I have
summarized my work on this topic in the third chapter of my dissertation
(see the link below). In a previous chapter, I have concluded that
agent-causal types of freedom are "worth wanting", because they are
indispensable for the operation of rational self-consciousness. Hence, the
goal of this one is to prove that such a model can be formulated in a
meaningful way. The strategy is to attempt a reverse engineering of free
will, in order to identify its verifiable preconditions. The end result is
a contingent description, whose validity rests on the outcome of relevant
empirical observations.

Here is a raw presentation of my approach. At first I assume that - in
most cases - my decisions are determined by well-defined anteceding
operations. Emotions, desires, reasons, inborn and acquired predilections
(and almost everything else that could matter) somehow add up to a net
preference, whenever several options are presented to me. As a result, my
decisions are free only to the extent to which compatibilist models are
true. Yet, there are also special moments when past histories are
indeterminate. For example, I have to choose between two alternatives, but
the net sum of all efficient considerations is nil. Suddenly, I am like
Buridan’s ass, my mind frozen by the resulting inability to break the tie.
This is where compatibilism fails, and something else has to intervene.

In order to break the deadlock, I suggest that a transcendental source of
inputs must be introduced into the picture. My mind is assumed to contain
an isolated cell, in which it stores a catalogue of so-called tags.
Whenever I am stuck, a transcendental link to the catalogue is open. If
there is a relevant tag, the outcome is a collapse of the indeterminate
state in favor of the matching alternative for action. If there is no such
tag, I pick one alternative at random, in order to see what follows next.
I experience the consequences of the ensuing action and another tag is
added to the catalogue. It should be clear that this is a developmental
model, in which I start with an empty catalogue and keep extending and
revising its contents throughout my lifetime. All tags result from
perfectly random choices, colored by the accidents of the context of
action. On the one hand, the contents of the catalogue are well-defined
and highly relevant for me, because they come out of my experiences. On
the other hand, the same contents cannot be predetermined by earlier
states of the universe, because of the indeterminate state that precedes
every tag-inducing operation. In short, the final choice is up to me, and
it is meaningful without being externally predetermined. Note that the
tags are not reasons or explicit arguments. They are simply that part of
me that reveals itself at the moment of indecision, and is formed by my
past history of similar moments. Though, modifications of the model are
possible in which this "part of me" becomes instrumental for other mental
functions. For example, my brain could be perpetually undecided whether to
think about specific issues or not. In this case, my thoughts would truly
originate within my mind. Once I start thinking, the logical conclusion
may be predetermined, but it was my decision to begin the process that
ended with my possession of its outcome.

This model is not compatible with absolute determinism, but it does
require the operation of unfailing material laws in order to yield what I
call the "best kind" of free will. In other words, the laws of physics
should be deterministic but insufficient to govern all physical outcomes.
The crucial insight is that fundamental material processes must be
configured in a very special way, in order to accommodate this type of
freedom. Transcendental influences are unobservable, but their
facilitating material features are a different story. (If archeologists
were to find a city with houses that have water spouts in the Sahara
desert, they would have to assume that it rained there at some point in
the past). That is why it is not necessary to debate the plausibility of
the proposed model, or the likelihood of its actual manifestation, as long
as it is self-consistent. Once the relevant physical markers for it are
identified, it is up to the scientists to determine if those properties
are present in our Universe. Accordingly, the second part of my chapter on
free will is dedicated to the derivation of the required observable
qualities.

The text can be found here.

Best regards,
Ghenadie Mardari"

Comments

[John Fischer]

Ghenadie Mardari,

Thanks for sharing your project with us. It seems to me very interesting, but (of course) I have various questions.

Let's suppose we are in a Buridan's Ass type situation--so that, as you say, the "net preference is nil". Why exactly do we now need to posit something "transcendental"? Why can't I simply decide to flip a fair coin, or otherwise make my decision arbitrarily when I am literally in such a situation?

Also, why exactly does it HELP to move from coin-flipping to the "transcendental" approach you sketch? How is it really any different? And how does the agent remain in control, under such circumstances?

[Clark]

This sounds very similar to Kane's view where we have a mix of deterministic physics and then indeterminism to produce Libertarianism. Of course lots of Libertarians argue Kane doesn't escape the problem of luck. So I'll be curious to see if you can handle those criticisms.

[Ghenadie Mardari]

Prof. Fischer:

Thank you very much for breaking the ice with your question. My answer is that we do not need to posit transcendental operations in the ideal (single) case of indetermination. Indeed, this situation must be resolved arbitrarily. Nevertheless, I claim that it helps a lot to introduce this new "mechanism", for two reasons. Firstly, it seems to offer an interesting solution to the intelligibility problem, when we move from isolated cases to developmental scenarios. Secondly, it leads to a verifiable model of free will. I will address just the first point in this message.

When I am faced with a set of perfectly equivalent alternatives for the first time, I cannot make a choice that represents me in any way. I simply decide to go with one of them, at random. Still, I do not flip a coin, because that would make the outcome reducible to an earlier state of the Universe. The indeterminate situation must produce an unstable equilibrium in a dedicated part of my brain (somewhat like in Kane’s model, described in the "Significance of Free Will"). By destroying this equilibrium with a transcendental "tip", rather then a quantum fluctuation, I limit the cause to the domain of my mind. In the first iteration, it is random because I have no sense of identity relative to this particular situation. After being in similar situations more than once, I acquire a sense of identity relative to it (as marked by the appropriate "tag"). So, next time around I will choose an outcome that represents me. I will be the kind of person that does in . This part of my identity is well-defined, by it is acquired through a succession of indeterminate steps. Its effect is exercised through a mechanism that is – again – independent from external causal processes.

[Ghenadie Mardari]

It seems that one sentence was inadvertently cropped at the end of my previous post. It should read:

"So, next time around I will choose an outcome that represents me. I will be the kind of person that does [this] in [these kinds of situations]. My identity is well-defined, by it is acquired through a succession of indeterminate steps. Its effect is exercised through a mechanism that is – again – independent from external causal processes."

[Ghenadie Mardari]

Re: Clark’s post from 11/25/08

I read Kane’s book on The Significance of Free Will after I developed my model, at the recommendation of my advisors. After that, it was easier for me to describe my work, because I could start by saying: "It is almost like Kane’s model, except it is agent-causal".

As far as I understand, Kane felt constrained to accept the implications of the so-called problem of luck, because of the intelligibility issue. If the history of my own predilections and reasons for action is indeterminate, how can I also have a preference for a single alternative at the same time? (And, if it was not indeterminate, then it could not be up to me at all). I address this problem by showing that it is possible to have well-defined motives for action, without external determinism – be it material or mental – and also without internal randomness. (This is why I spent so much time on the "sorting procedure" in the text). Then I assume the existence of a dedicated mechanism in the mind, whose sole function is to solve the problem of indecision. At the point where Kane would have reduced the outcome to the random effect of a quantum fluctuation, I would say that the result is not arbitrary. It is either a random choice that is meaningful, because it is subordinated to a process of trial and error, or it is a well-defined choice on the basis of an already existing tag. It is almost like saying that a quantum state collapses in a pre-determined fashion, whenever I have a relevant tag. Though, in the second part of the text I show that quantum fluctuations do not work for this model. A well-defined pre-quantum process has to be at work.

[Clark]

It is either a random choice that is meaningful, because it is subordinated to a process of trial and error, or it is a well-defined choice on the basis of an already existing tag.

I guess I don't see how that ends up practically different. Even if the random choice is made deterministically then a Libertarian will say you aren't responsible for making the choice for randomness and further the randomness also doesn't give you responsibility. i.e. they'd say for responsibility or meaningful choice you need at some point something that is neither luck nor determined. Saying that a luck based decisions is external and chosen due to determinism doesn't make it any less luck. Why would that make a difference? I assume there's a key appeal to some intuition at this point. It seems to me that there may be a crucial disagreement about what makes choice meaningful. (Perhaps this is just a semantic question - I can't really tell)

I rather like Kane's position but I can understand why it doesn't really address the concerns of most traditional agent-Libertarians. But I don't think the problem many had with Kane was whether the chance was internal or external. Just that it was pure randomness.

Now I don't really agree with the concerns by most agent-Libertarians but as I said earlier, I suspect this model won't assuage them more than Kane's did. Personally I can't see an account of agent-Libertarianism that doesn't end up having some dualistic "choosing" stuff that is either substance dualist or ontologically emergent in some fashion.

[Ghenadie Mardari]

I have a short and a long answer to the preceding post by Clark (dated 11/26/08).

The short answer is that you are right: randomness is randomness, and it being internal or external makes no difference. The main point of departure from Kane, in my model, is the developmental effect of those random choices. Granted, the first iteration is qualitatively similar to Kane’s scenario, but the second time around there is no randomness. The tie is broken by a well-defined internal cause (the tag).

The long answer involves a thought experiment. Suppose, for simplicity, that all event-causal motivations for action are of two-types: reasons and desires. What is a person? Is it simply the sum total of all reasons and desires, or is there more to it than that? What is left of us, when all of these motivations cancel out? If the answer is "nothing", then an indeterminate situation can only be resolved by pure randomness. However, I claim that there is something about agents, over and above the propensity to reason and desire. This undefined property is manifest in certain situations that cannot be reduced to known factors. Sometimes a person does something for no reason but the fact that "this is the kind of person that he or she is". The slightest motivation can hide this propensity, but it is there, in the background, nonetheless (or so I suggest).

In my developmental model, self-awareness begins "tabula rasa". I do not know what kind of person I am from the start. So, the first time I experience a certain type of indeterminate choice, I have no sense of identity relative to it. Still, I decide to experiment, in order to get a stronger sense of self. I choose one option at random and, based on the effect of its consequences on me, I have an immediate reaction of identification (such as, "this is not me", or "this is me"). It may also happen that I have a clear tag for action, but then the consequences induce an opposite reaction in me, so I have to reconsider my tags. I am not committed to any view about identity (whether it is fixed or flexible). I am just trying to suggest that a sense of identity can be independent from all reasons and desires, and also that it can emerge into self-awareness from external experiences without being random or reducible to an earlier state of the Universe. Moreover, this sense of identity can act as a demonstrable source of autonomous mental processes.

[John Fischer]

Ghenadie,

Thanks for your further clarifications and reflections. I think it will be helpful, as you develop the project, to compare and contrast it with similar approaches, including (especially) Kane's.

One thing I have always been puzzled by in some of Peter Van Inwagen's work and Kane's on this topic is whether they really think it is a necessary condition for freedom (in the sense linked to moral responsibility) and even personhood (in the model you are sketching) that one face difficult choices. Isn't it conceputally possible or imaginable that an individual always face relatively clear choices and still have free will (in the relevant sense) and develop into a person? Why exactly does one give theoretical hegemony to "hard cases"? Is there a more fundamental, or underlying picture, that drives this theoretical priority of hard cases?

[Ghenadie Mardari]

This is a response to the preceding comment by John Fischer (dated 11/28/08).

I am reluctant to make assumptions on behalf of other libertarians, but I can explain the need for hard cases in my work. What I have described so far is not the actual model that I have in mind, but only the conceptual platform for it. By stripping away as many factors as possible, I hope to convey a better understanding of my solution to the intelligibility problem. If I were to start presenting the actual model, while everybody else thinks that a solution like that is impossible, I couldn’t possibly get very far. People would just say to themselves: "Oh, this guy thinks that he has discovered the wheel, because he is not aware of the difficulties involved". The current position in the relevant literature seems to be that only two types of causes for action are possible: those resulting from well-defined considerations that are pre-determined, or those resulting from arbitrary considerations. I think that a developmental model like mine can produce a transient state in which a third alternative is possible: well defined considerations that are not pre-determined. If my proposal survives the objections that you might raise, then it will require just a few modifications to allow for implications about responsibility. (As it is, you may have sensed, no meaningful responsibility can be derived from it).

Libertarian free will cannot be only about hard cases, from my point of view. Yet, there are certain realities to be taken into account. For example, the idea of "mind over matter" does not seem to be immediately compatible with modern scientific knowledge. As long as this is an open problem, agent-causal contributions are likely to command less respect than other approaches. In other words, libertarian models need a stronger conceptual foundation, and "the land of hard cases" is the most likely place to find it.

[Mark Smeltzer]

Ghenadie,

Could you say a bit more about why your model is not compatible with "absolute" determinism?

You say that,

"If there is no such tag, I pick one alternative at random, in order to see what follows next."

Why does "random" have to mean metaphysically indeterminate? Why can't "random" simply indicate that which bridges the volitional gap between the identical propensities?

On the face it seems that your model, despite your bias toward Libertarianism, could be adopted by a Compatibilist. The same charge can be levied against Kane's model. Kane himself is resilient to the charge since Kane accepts the consequence argument. However, there is nothing in Kane's model that prevents a Compatibilism from subsuming it wholesale.

[Ghenadie Mardari]

This is a response to the preceding comment by Mark Smeltzer (dated 12/01/08).

When I started thinking seriously about free will, I asked myself: What would it take to satisfy the intuition that my decisions are irreducibly "up to me"? The unavoidable answer was that I – the mental entity – should have the power to influence the behavior of physical entities, when I chose to initiate action. By default, this raised two problems. Number one, how does mind act upon matter? Number two, would this process enable freedom in a meaningful way, even if it was possible? It is the answer to the first problem that made my approach incompatible with determinism. It involves the possibility of voluntaristic elements in the degeneration of unstable equilibriums with indeterminate outcomes (which I invoked as a transcendental effect in the introductory statements at the top of this thread). If you wish, I can supply additional comments to clarify this in a future post.

My use of randomness, as discussed above, was intended to work as a solution to the intelligibility problem. Whenever I have a clear reason for action, it must come from somewhere. Either I have inherited it, or I have acquired it, but it must have a cause in any case. Accordingly, it would seem that I have no way to ground my choice on considerations that are not reducible to pre-existing external factors. However, my choices can be "mine" only insofar as I am aware of them. I cannot be responsible for any behavior that originates outside of my sphere of conscious control. Therefore, the most important question is: how do those grounds for action enter the realm of my consciousness? According to my developmental model, deliberation requires rationality, and rational self-consciousness emerges well after birth. Hence, I probably have access to a very small number of inherited considerations for action. Moreover, some of the acquired factors (by my conscious mind, not by my body) are irreducible to external well-defined states, if they emerge from indeterminate processes. For example, I may overcome an instance of indecision with a random choice that has well-defined effects on my consciousness. Because the physical past is not sufficient to determine the exact content of such an experience, the resulting component of my awareness is also not reducible to an earlier state of the Universe. As a corollary, it is possible for me to have well-defined considerations for action that are not fully reducible to external causes.

Please note that this mechanism does not automatically solve the stated problem. It only works if applied to the so-called sense of self, insofar as it is an acceptable cause of action. In the case of other elements of awareness, the argument seems to fail. In the final analysis, it does not matter how the grounds for action are acquired. As soon as they are part of active memory, they are "there" before the moment of action. If they determine the outcome, then there is no choice.

[Mark Smeltzer]

Ghenadie,

Thanks for the comments. So, I see now that you see the transcendental states as being the reason that "absolute" determinism and free will aren't compatible.

You require that the physical laws function deterministically but that the agent has the power (through these transcendental states) to overcome the laws of physics. In other words, the laws of physics are deterministic, but not complete.

Am I on track so far?

Do you think that this set of beliefs requires a denial of physicalism? Tim O'Connor has a very similar theory to yours (it is also has a developmental account). According to his theory, the locus of free action is a mechanism in the mind that constitutes a material substrate that has an emergent power to add to the laws of physics. So, O'Connor's theory is compatible with physicalism.

Putting the question aside, let's suppose that physicalism is false and that there is at least something else besides physical matter to consider in our ontology -- be they substances or properties. It is possible that the interactions of the elements in our ontology still function deterministically.

By extension, it is possible that the interaction between physical matter in the brain and the transcendental states your theory requires could still function deterministically. To see how this would work, we need to apply Fischer's "rollback" thought experiment:

1. In the actual sequence agent S freely performed action P at time T1 after facing a dilemma between P and Q at time T0.
2. If time were rolled back to T0 (e.g. *ALL* variables at T1 are reverted to their state at T0) when S was still facing the dilemma between P and Q, S would either choose P or Q.
3. For the incompatibilist, it must be true that S would perform Q at least once after a series of successful rollbacks.
4. If S could, but never would, actually perform Q after a series of infinite rollbacks, then compatibilism is true.
5. If S does perform Q instead of P after a rollback, there needs to be a reason to explain why S freely chose Q instead of P when *ALL* of the variables were held fixed.

As far as I understand things, the dilemma presented in rollback thought experiment is where the bulk of the intelligibility problem lies for the incompatibilist... and if I understand your position correctly, your solution to this dilemma is to say that S acts freely at T0 if S performs (P or Q) at T1, and thus both outcomes are equally attributable to the agent (and I get your point about this not being sufficient for responsibility).

You take that stance because those are the defining moments where the agent has no evident reason to perform either action, so will choose one at random in order to find out whether the agent identifies with the outcome of the chosen course of action. (On the basis of that identification -- or rejection, the agent would form a new tag.)

I think that's enough background to get to my questions...

1) Do you think that the agent-self is a substance? If so, can we still make sense of a version of the "rollback" (viz., determinism) that applies to every variable in our ontology including physical matter and agent-stuff? (Even if agent is a substance, absent physical indeterminism, it is confusing to me why the agent would sometimes choose P and sometimes choose Q if we rollback *ALL* of the variables.)

2) If the agent-self is not a substance, would you adopt an emergentist model like O'Connors? If not, what is the agent-self and how does it get the ability to transcend the laws of physics? (Even on the emergentist model, the rollback argument is still a problem... even though the substrate adds to the laws of physics in ways that the laws of physics in the distant past cannot predict, it still seems that, abscent of physical indeterminism, rolling back *ALL* of the variables to T0 should produce the same outcome at T1 since the substrate already exists at T0.)

3) Regardless of the composition of an agent, can't we take a minimalist stance towards what qualifies as sufficiently "random" to allow the agent to pick between P and Q in tough calls?

[Ghenadie Mardari]

This is a response to the preceding comment by Mark Smeltzer (dated 12/02/08).

The agent-self is not a substance, in my model, and I would say that its emergence has many things in common with O’Connor’s description. However, I am not yet ready to admit that my approach is compatible with physicalism. Perhaps there are conceptual nuances that escape me, and I trust that you will correct me, if the following comments contradict my claims.

Let us consider an isolated physical system, consisting of three bar magnets that are arranged longitudinally on a surface, well separated from each other. The marginal magnets are fixed to the surface, and the central one is frictionless. Moreover, the central magnet is placed such as to experience an equal pull in both directions, being precisely balanced by these two opposite forces. The slightest action on the central magnet would cause it to slide in one direction or the other, until it connects with one of the other magnets, but there is no such force in our system. If this was the whole story, the arrangement should be perpetual. To make things more interesting, let us assume that the elementary particles of this system have a special property: they have well-defined positions, but at every moment they can have one of two adjacent coordinates. The physical laws do not determine which one is actually going to be materialized, but this is usually inconsequential. The distance between the two positions is too small to have any macroscopic effect, except in unstable arrangements where the slightest fluctuation can initiate a butterfly effect. Accordingly, a fleeting asymmetry in the distribution of the particles from the central magnet may cause it to start moving spontaneously in one of the two possible directions.

Let us assume that an arrangement like this is present in my brain, acting as a switch. If the state degenerates in one direction, I end up doing a "good" thing (however defined). Alternatively, I do the opposite. The question is: what causes the switch to activate one channel, rather than the other? According to Kane’s model, it has to be a fundamentally random phenomenon, out of my control. If my understanding is correct, this explanation should be compatible with physicalism. In contrast, I assume that I live in a Universe where metaphysical effects are possible. In other words, I postulate that my brain also has a special device, physically independent from the switch, which can induce telepathic effects. Hence, the particles in my switch do not choose their physically undetermined positions at random. When I have a tag that is consistent with a preference for the "bad" thing, for example, they will assume an arrangement that must initiate the corresponding action with certainty. This is how my mind, otherwise perfectly reducible to physical considerations, induces autonomous effects on matter. Is this a physicalist model? The propensity of elementary particles to conform to mental tags is a fundamental universal property, but the non-physical essence of this causal mechanism would probably separate it from any monistic interpretation.

Here is my response to the "rollback" problem, in light of the above. Whenever I have an unshakeable sense of self relative to a problem, there is no free will. All rollbacks should look identical. Nevertheless, I would like to suggest that such instances are very rare. In most cases, I am undecided. It is absolutely up to me which way to go, and the switch will follow my will, but I must choose first. Will I be the saint or the sinner this time? Because I actually choose, and because I can choose either way, the rollbacks will not always give the same outcome. One way to think of this is to distinguish between emotions that have "drives" with variable intensity, and reasoned conclusions that are detached and physically neutral. It is only my will that can add a "drive" to a rational prescription, but it is perfectly up to me whether to "turn it on" or not. My choice is not determined by something from the past, to the extent to which it is a choice. I can inform it by considerations about the future (e.g., what kind of person I want to become), or I can "just choose", but I can certainly act differently next time around. This is where the educational effects of retribution come in. (Or else - if I have a strong antisocial identity - I cannot change or act differently, but the actions are still mine and punishment without "education" is my share when I get caught.) In short, "free will" is a transient developmental stage between infantile ignorance and "good will"/"bad will", provided I live long enough to reach one of them. Responsibility follows from it (the will) being mine, rather than from it being free.

[Clark Goble]

Coming late. Sorry for the delay.

I agree with Mark that much of what Kane argues can be accepted by a compatibilist. I think though that the element of randomness present in both Kane's as well as Ghenadie's models makes it difficult for compatibilists who see the determinism they are concerned with as being causal determinism. An other choice though is that one rejects causal determinism but allows for a block universe of some sort (due to foreknowledge for some religionists or due to some reconciliation of GR & QM in which GR's substantial space-time is maintained). This would allow there to be facts about the future while maintaining randomness. Those sorts of compatibilists would probably see a lot in the theories of Kane and similar ones.

(I recognize this is just an expansion of the point Mark was making about randomness - I just wanted to make the connection to physical theory a bit more explicit)

One other quick thought. Mark wrote:

According to his theory, the locus of free action is a mechanism in the mind that constitutes a material substrate that has an emergent power to add to the laws of physics. So, O'Connor's theory is compatible with physicalism.

Is O'Conner's particular kind of emergence compatible with physicalism? Admittedly physicalism as a term is a a bit of a moving target in terms of meaning. If we take it to entail that there are physical laws though that are knowable by science then the very notion of "adding to the laws of physics" seems problematic. I think at a minimum most physicists would cringe at the idea.

Isn't this kind of emergence usually excluded by physicalism?

[Mark Smeltzer]

Clark,

Regarding the emphasis on indeterminism in Kane's view: from the compatibilist perspective, the indeterminism isn't required to be metaphysical so much as it is required to fulfill the operative function; namely, resolving the dilemma for the agent. I don't really find that view appealing, but its there for any compatibilist that wants to take it up.

I tend to use physicalism in the "traditional" sense: that everything that is supervenes on the physical. It can certainly get narrower from there, but the one thing that definition does is that it stipulates that there isn't anything that
does not supervene on the physical. So, physicalism entails that there are no uninstantiated universals.

Views similar to O'Connor's could be construed to suit various concepts of physicalism, regardless of O'Connor's own take. I'd have to review his exact formulation but I don't recall making an association between his account and an incumbent necessary connection to the existence of uninstantiated universals.

To say that an O'Connorian substrate could "add to the laws" is just to say that once it came into existence the set of laws becomes larger -- not that the old laws no longer apply. In fact, I would think that the old adage for determinism could still apply: the present = the past + the laws of physics. If the past + the laws contains the future potential for the substrate and its additive capacity then the past would still be predictive of the future past the time when the substrate comes into being. I don't think O'Connor would like that interpretation very much...

On a side note, physicists ought to be used to cringing by now anyway. The string theorists of yesteryear are becoming today's brane (or M-theory) theorists. And the M-theorists say that the physics of each universe in the metaverse is governed by the distinct kinds of interactions between the colliding membranes that formed each particular universe -- so each universe would have its own separate physics. So here's a case where the same idea that makes the M-theorists giddy also makes the quantum theorists cringe; all of them physistics. So, whether or not some group of physicists cringe can hardly serve as a litmus for truth about physics. Maybe if we got all of the physistics together and they all cringed at the same time then it would count for something?

[Clark Goble]

Well I think there's a big backlash against string theory. (I consider M-theory basically string theory.)

I'm not sure I agree with how you characterize the branes. I think there's some master physical law governing existence. You can then talk about other "laws" which may be (in M-theory or even the multiverse form of loop quantum gravity) different. But then that's true even in more traditional physics as difference symmetries emerge in the early universe leading to different "laws" of physics. For that matter you can derive most thermodynamics on the basis of symmetries in a system.

So I think physicists will often use "law" rather loosely. But in the sense of an ultimate law I think their fair consensus even though I don't think anyone honestly has a clue what that ultimate law is.

That said, I think physicists cringing is a better guide than philosophers cringing. (grin) One day I'd love the experimental philosophy folks to go out and measure physicists' intuitions to see how they match up to philosophers.

But beyond all that thanks for the expansion. That clarifies things for me. I'd always taken physicalism as a theory about ultimate physics rather than tied to the older question about scholastic realism.

[Clark Goble]

One last point I missed.

To say that an O'Connorian substrate could "add to the laws" is just to say that once it came into existence the set of laws becomes larger -- not that the old laws no longer apply.

Is this right? I admit it's been a few years since I last read O'Connor and I don't have his writings handy. But isn't his position much stronger than just a few determinate laws emerging ontologically? I thought he had emerge a substrate that was neither random nor determinate causally.

[Mark Smeltzer]

Clark,

I don't put a lot of stake on philosophers cringing either. So, at least I'm fair.

Regarding O'Connor's view, I am definitely paraphrasing rather liberally. No doubt O'Connor would have lots of disagree with, but the gist is that a compatibilist could offer a theory of emergentism that is compatible with determinism even if O'Connor's specific account isn't (and O'Connor would certainly argue that his isn't).

Lastly, don't take anything I say about physics as implying that I consider myself an authority on the matter. My knowledge of contemporary physics extends to the kind of dumbed down formulations that usually wind up in an article's abstract. I don't even pretend to understand the mathematical underpinnings. Maybe one day...

[Ghenadie Mardari]

Dear Gardeners:

I feel that it is time for me (G.M.) to finally confront the big question:
Who Am I – a "libertarian" or a "necessitarian"?

I did announce myself as a libertarian, at the top of this thread. However, the issue now is: libertarian according to whom? Definitely, not according to Hume! The classical distinction between libertarians and necessitarians is based on the idea that freedom cannot be incompatible with determinism. Arbitrary spontaneity, or pure indeterminism, violates our intuitive understanding of what it is to be free (and therefore responsible for our actions). It has to be the case that freedom involves at least a partial role for determinism. After all, we expect the agent to be the cause of well-intended actions. As I recall it, Hume believed that people only insist on separating free will from determinism because of their confusion with respect to conceptual matters. If one assumes that "determinism" is synonymous with "predetermination" then any discussion about free will becomes problematic. Still, there are no irrefutable empirical or rational grounds for the belief that such a restrictive definition of the concept is required for a proper description of our Universe.

Hume appeared to believe that his notion of conjunction, in contrast to usual assumptions about causality, may lead to a better understanding of freedom. According to some interpretations, this project created a set of problems of its own, paving the way for what we now know as modern approaches. I wonder, though, if this was not due to another case of conceptual confusion. Hume was an empiricist who cautioned against all sorts of ontological assumptions about the world. Accordingly, one should avoid even the assumption that "conjunctions" are more real than "classical causes". It is because of our weak foundations for knowledge that we should refrain from restrictive definitions of causality of any sort. For all we know, determinism could very well work in a manner that does not predetermine the content of our autonomous decisions. This is what I take to be the spirit of Hume’s contribution on this matter, and that is why I now consider myself a classical compatibilist.

As shown in the preceding posts, I assume the experience of willing something to be identical with the act of determining the outcome of a physically indeterminate configuration (unstable equilibrium). This makes every act of will potentially independent from physical processes, and also from predetermined mental processes. The sense of self is formed relative to rational considerations, but it is also autonomous from them. It develops on the basis of conscious formative experiences, involving indeterminate steps. Thus, it is never completely arbitrary. It is either well-defined at the moment of decision, or informed by developmental considerations. Consequently, I favor greatly the concept of freedom as independence from external constraints, but only in this mechanical sense (as opposed to the socio-political descriptions). In other words, freeedom as independence from external physical and metaphysical causes.

Unfortunately for me, this is not the modern position on compatibilism. Contemporary "necessitarians" appear to have reverted to pre-Humean notions of determinism in their implicit adherence to physicalism. I think that the Consequence Argument would never gain any relevance in a Humean context. One has to adopt a Laplacian view with regard to both mind and matter before taking seriously the notion that decisions cannot be up to the decision-maker. I think that Hume would not call the physicalist position "pro-scientific". He would insist that our models about freedom should be compatible with the appearance (!) of physicalism, given the successes of science, rather than with any sort of literal ontological interpretation. In the modern context, intuitively satisfactory definitions of free will can be formulated only in opposition to compatibilism. Then again, incompatibilists are bound to face the intelligibility issue as soon as they also subscribe to physicalism. I think that one can stay true to the empirical facts of modern physics with minor deviations from physicalism. Yet, this compromise is sufficient to undermine the link between agency and the "could have done otherwise" problem.

To sum up, my model is incompatible with Laplacian determinism (which I also described as "absolute determinism" in earlier posts). This is what makes me a libertarian, according to the terminology that apparently governs the modern debates on free will. Though, I resent having to be confined to this position. I would rather call myself a Humean compatibilist, but this – most likely – would lead to misunderstandings and confusion among fellow philosophers of action. Am I out of line with these comments? Where have I gone wrong? Please help!

[Mark Smeltzer]

There's plenty of room in the Agent-Causal Compatibilist camp.

I maintain a seemingly awkward position:

(1) an agent must be the final cause of action to be free or responsible
(2) an agent being the final cause of action is possible
(3) an agent being the final cause of action is actual
(4) a non-physicalist metaphysic may be required given (1) - (2)
(5) a free agent can exist in a world with N-layer determinism (where N crudely represents the total number of physical/metaphysical dimensions operative in the world)

So, if free agency turns out to be compatible with physicalism in response to (4), then my position would be compatible with physical determinism. In the mean time, I take an agnostic stance until more work on (4) is accomplished (mainly in the area of the metaphysics of personhood).

I am not sure that (1) - (5) are necessary to be an agent-causal compatibilist, but they do seem sufficient!

[Clark Goble]

I did announce myself as a libertarian, at the top of this thread. However, the issue now is: libertarian according to whom?

Interesting question. I tend to think names are over-emphasized a little. I think Kane comes closest to my own view but I have a hard time calling myself a Libertarian.

[bramblecrest]

I'll be curious to see if you can handle those criticisms.