OrangePhilosophy

But, Jeremy, if you believe in libertarian free will then you can believe in fairies and chimeras! And who wouldn't want that? (by the way, most defenses I've seen do make the ad hoc move you speak of).

Is a decision an event? If so, can it be said to cause something, i.e. lead to an alteration of a state of affairs? But then, if a decision is a real event, actually occurs, and has an effect, can it itself be said to be causally determined and yet still be called a decision? I realize that this is just an "analytic" consideration, following from the sense of terms, and not a substantive answer. But it seems to me that one of the problems with the discussions of "free will"- (I don't like the term and prefer to speak of volitional agency)-, in the Western philosophical tradition, is that it is implicitly always defined in terms of causal necessity and thus, due to a bad application of either/or logic, must be considered an absence of causality. (The most famous statement of the problem is Kant's third antinomy.) Another problem is that volitional agency is a finite phenomenon, yet "infinite" considerations are adduced to defeat or "vindicate" it. (Correlatively, there is confusion between being conditioned and being caused, leading to a hyperbolic focus on "unconditional freedom".) It seems to me that the basic idea of volitional agency as a real phenomenon is that it is causally underdetermined and attaches to the emergent organization of an organism/self, which interacts with other such selves on the basis of language and meaning. The idea of choice or decision is that there is some limited range of alternative possiblities, from which a selection is made and effected. Since only one alternative is effected, the others are eliminated, but does that thereby mean that they were never real possibilities in the first place? It would seem that to make purely causal considerations count against the reality of agency, one would have to provide a causal account of meaning.

One of the problems with "compatibilism" is that its sense of "predetermination" fails to consider the sense of time and its openness to the future. (Such a notion of causality is Newtonian, bound up with the notion of "reversibility", that is, with the idea that given adequate data on particles, their relative positions can be back-predicted. But with Einstein, time itself is given physical meaning and such reversibility no longer applies.) Compatibilism amounts to the prediction that my actions will have been determined.

I'm not sure what your problem with compatibilism is supposed to be. A problem with predetermination isn't a problem for compatibilism, since compatibilists don't assert that we're predetermined. They just say that if we are then we're still free. Some of them go further and say that our actions must be caused by the right sort of internal causes (e.g. character), but that doesn't require determinism, just that our free choices must be causes.

As for sense of time and openness to the future, compatibilists believe fully that we have a sense of openness and that we make choices not knowing how the future will turn out. It seems open to us, and our decisions are free. If the future is fully causally determined (which they don't assert), those other claims are still true. This is even one of the main points of the compatibilist.

I'm sorry if I misunderstood some of the terms of your post, since I don't read analytical philosophy and am not familiar with the specific literature cited. (I've never heard of "libertarian free will" before reading this post, but I figured it referred to doctrines emphasizing the "separation" or "autonomy" or "self-determination" of "free will" or volitional agency.) To some extent, I slightly misread the following: "He says that compatabilism has a worse mystery, that we can be predetermined, yet free."

In some sense, the question about "free will" or agency is a question about specifying the meaning of a word or set of words, picking out the phenomenon that is so identified, and addressing whether the phenomenon so identified is in some sense real or efficacious or not. Also one might ask what need or function is adressed by a question or the various positions that purport to answer it; in other words, what is its point? As to the three alternatives on offer, libertarian free will, compatibilism or epiphenomenalist/reductionist dterminism, I would find none of them satisfactory, and I find a philosophical analysis that ends in a claim of mystery odd. (Any mystery should not be in the concepts, but rather in the actual encounters they would underwrite.) I was suggesting a fourth, language-based alternative. Animals, as causal organizations that delimit themselves from their environment, can make behavioral selections that intervene in and alter environmental states of affairs and nexuses of causality, but they do so in response to immediate environmental events or cycles of events. But only with language is it possible to interpret an environmental state of affairs counterfactually in the light non-existent alternative possibilities and select from them to bring about an alteration in the environmental state of affairs, that is not pregiven. The causal component of agency is already given biologically, whereas the volitional component is build over that. That, in simplified form, is the root of human agency.

There is, of course, much else to be said beyond that level, and other sorts of issues that would subtend such an account. But I will draw out just two implications that might give it some point. 1)Language is itself based on interaction between organisms. Though linguistic communication involves an exchange of information, it also functions through bringing about modal shifts in relations between organisms. But such shifts are not describable as an additive relation between the causal organizations of the organisms, and with the advent of syntactically organized, semantically self-stabilized language, they generate a third realm of meaning, which I would maintain is not analyzeable or reducible in terms of causal relations or processes. (I found G.H.Mead's account of "social behavioralism" somewhat crude, but his account of thought, as requiring the capacity to take in and adopt the standpoint of the other, points in the right direction, when it comes to accounting for the apparently "self-determining" capacities of agency.) 2) Volitional agency is not just an individual, factitiously self-intuited property, but must be collectively distributed in a human community.

As for compatibilism, though I'm not familiar with all the p's and q's accounts, the general notion is not new to me, as represented in the classical tradition by, e.g., Hume or Spinoza. I would understand it to mean the identification of volitional agency with a subset of causal processes, which, for all practical purposes, is taken to be sufficient, inspite of the intimation of its illusory character. I would raise two basic objections to such accounts. 1) Are such accounts really operationally adequate? That is, do they really define the phenomenon in question and take into account its specific features? Or rather, under the pressure of a too precipitate logic, do they not resort to vagueness, and "solve" the problem by redefining it? 2) Do they not rely on a notion of causality that is at once too hyperbolic and all-encompassing and too monolithic and undifferentiated? In particular, not only must organismic and environmental causality be distinguished- (this was Kant's great mistake; he effectively introjected environmental causality to cover not just the organism, but virtually the entirety of the empirical person)-, and not only do the generation of emergent properties and capacities from causal processes, which do not "exist" at lower levels of causal description, need to be taken into account, but the rejection of "randomness" and "indeterminacy" and the assertion of the criterion of "control" mischaracterize the importance of stochastic processes in accounts of causality. Specifically, negentropic organization, whereby causal processes themselves alter the basic probability distributions of such processes, need to be taken into account, since, in turn, the interaction between such processes can lead on to an unpredictability, which is not the same as mere randomness. (Here we would be entering into subtending questions involving conceptual/causal mythologies- since it is a matter of doing philosophy and not of doing neurobiology- about body/brain/mind.)

I should emphasize that the account of agency involves a limited, finite phenomenon. It is not a matter of Superman powers or of self-creating Uebermenschen. In any human situation, there is alway a margin of freedom, sometimes more, sometimes less, but it is always just a margin. (To abstract out of human situations and their contexts this small margin and thereby constitute an hypostatized concept of "free will" strkes me as an objectional procedure, which is why I disike the term.) But, in turn, the explanation of such a phenomenon can not be expected to defeat an unrestricted, "infinite" account of causality. (There was a theological influence in the way the problem of agency was traditionally set up: the dependency of finite agency on Infinite Agency. There should be no need to mirror this set-up in a counter-position. On the other hand, my account in terms of "underdetermination" is analogous to some accounts in the Jewish Kabbalah about how God created the world by withdrawing from it.) At any rate, what is at stake in the question of volitional agency, aside from our everyday natural language/common sense attributions, is its power of initiation, what Hannah Arendt once termed "natality".

Compatibilism is the view that we can be both free and determined. It doesn't require a particular account of what freedom is, as long as freedom is consistent with being determined. I agree that reductionistic accounts of free will based on mechanistic categories are inadequate, but compatibilists don't tend to do that. David Lewis was a key compatibilist of the last half-century, and he accepted at least one level of description that turns out to be indeterministic, so there's no insistence on accepting freedom only if the causes are fully sufficient, as Hume thought.

These philosophical analyses don't claim to end in mystery. Peter van Inwagen claims that all three involves mystery, and he finds the libertarian one the most acceptable of those mysteries. You might see it as ending his analysis of free will in a claim of mystery. Compatibilists and hard determinists don't see themselves as doing that, and neither do I. He sees them as doing that.

There is an element of redefining with compatibilism, but compatibilists usually think it's misleading to describe it that way. They prefer to argue that incompatibilist freedom isn't at all what we ordinarily mean by freedom. We normally think of ourselves as free when the causes of our actions come from within ourselves in the right sort of way (some counterexamples of internal causes show that it can't just be any internal cause). To a libertarian or hard determinist, this is a redefinition of what freedom is, since they think freedom requires determinism to be false. But most compatibilists insist that freedom doesn't require that, even in the ordinary conception of freedom.

What does it mean for an event to be in one’s control? ‘Control’ seems to be a load-bearing term in the argument advanced against agent-causation. x controls y iff…? Or is control a sort of place-holder for whatever it is that you need to add to action to get free action--kinda like warrant is just whatever it is that you need to add to true belief to get knowledge? Just wondering, because I don't know what it is.

Let ‘ACE’ designate the event that is the agent’s directly producing/causing some event. In addition to the three you mention, Jeremy, there may be a fourth way for an agent-causalist to respond. Tim O’Connor, e.g., claims that there is nothing that can be such that it is THE direct cause of an ACE. That is, necessarily, nothing can directly produce an ACE. Why? Because it must be the case that to count as being the direct cause Z of a causally complex event X’s causing Y, Z is the direct cause of X. Speaking more loosely, to cause a causing you must cause the leading edge of the causally complex event. Since a substance/agent cannot be the sort of thing that is directly caused, there can be no directly causing an ACE.

The argument I gave assumes libertarianism to show that it leads to unsatisfactory consequences. So 'in your control' in the argument means whatever a libertarian would mean by that. Since the point of the argument is to show that there really isn't any clear notion of what the libertarian might mean by that, thinking of it as a placeholder is probably best.

As for O'Connor's response, that just seems to me to be restating part of the problem but not answering it. The problem is that libertarianism won't work if one of the following is the case:

1. The ACE is caused by something outside the agent's control.
2. The ACE is caused by an infinite series of causes all in the agent's control.
3. The ACE is uncaused.

Then the claim is that there's no further possibility. O'Connor, as you've described him, merely states that options A and B are problematic for the libertarian. That doesn't explain how to get out of the dilemma. He needs to say that 3 is ok or that there's a fourth option. Merely stating that being caused is a problem is not giving a fourth option.

To be clear, the three points you enumerate on the June 16th post are not the same (enumerated) three points on your June 11th post. To say that there is no such event as an ACE (as some interpret Reid, e.g.) is different from saying that ACEs are uncaused. The former vacuously implies the latter, but the latter does not imply the former.

It’s unclear why you say that "O’Connor…*merely* states that options A and B [or 1 and 2] are problematic for the libertarian." He doesn’t merely state this. He argues for it. He says that 3 is ok; indeed, he argues that an ACE cannot in principle be directly produced by anything. I tried to sketch this argument in my June 15th post.

So, is there a problem for a libertarian in holding that an ACE is essentially something not directly produced? If so, why?

O’Connor will want to say that an ACE is intrinsically an event controlled by the agent. Moreover, if one grants that nothing even might directly produce an ACE, there’s no point in objecting that the agent-causal theorist can’t provide an event within the agent’s control that the agent controls in order to control the ACE.

Thad Botham mentions something of what I was trying to get at. There is not really a question of the absence of a causal substrate for agency. It is rather a matter of the insufficiency of causal processes to "determine", (which means ambiguously to cause and to specify), agency. I argued that agency emerges at the conjunction of several different areas and types of causal processes and that some of those processes would have to be-, (in order to be the "right" sort of processes to yield the phenomenon of agency),- what I termed weighted stochastic processes, specifically with respect to the brain processes that would result in the mental and volitional components of agency. "Weighted stochastic processes" would mean that a very large number of sets or pathways of causal events would yield essentially the same aggregate state or outcome, and that, while such processes are steered toward a quite limited number of aggregate states, the outcome is relatively open from the perspective of the mass of underlying causal events. The upshot is that there can be a high degree of relative path-independence between aggregate states and their causal basis, all the more so, if there is interaction between several such processes in yielding higher-order aggregate states. Further agency would involve relating several different sets of causal processes. Hence it becomes questionable whether agency can be identified with some determining factor in its causal substrate, since there can be no isomorphic correspondence between causal events and outcomes. It is precisely the indeterminacy of causal descriptions for such a causally complex phenomenon that leaves room for agency.

The question of causal agency, of the cause that causes agency to cause, amounts to a question about the decision mechanism that would lie at the core of the phenomenon. Agency would emerge and occur as an internesting of feed-back or control loops at the intersection of various causal processes. But no single such loop is agency; rather agency is their conjuncture. Let's take causal processes A, B, and C. With their intersection and interaction, there emerges an organism with a holistic functional closure. (This, in simplified form, would roughly describe cellular metabolism.) But if this system is left to the more or less random interaction and interference of the 3 processes with each other, then their synergistic emergence as an organism will be precarious. So lets add D,- (for "dummy")-, that emerges at the synergistic level of functional closure to balance out A, B, and C. But D also operates under pressure from causal nexus E,- (for "environment")-, and E contains causal organization O,- (for "other")-, which D in oriented toward in its motives and aims, but which can not be incorporated into the causal organization of D, nor the entire complex of A, B, C, D, and E. Now where in all these causes and causal set lies the decision that is at the root of agency? If one answers D, then the problem is that D only effectively exists as a response to the other causal sets, even if it has a distinct causal complexion of its own. It seems that the decision is excentric to the respective causal sets, or, in other words, that the decision is contingent with respect to any causal set. But that contigency is precisely what we mean by the phenomenon of agency,-(in the ordinary everyday sense, not in some extraordinary, hyperbolic, or metaphysical sense.) Of course, the respective causal sets must evolve in relation to each other and interact, which is to say that agency is an acquired capacity. But that just means that it is something that is highly conditioned: to identify agency with causal control and then to assert that such control must incorporate its conditions so as to meet the requirement of being sufficiently "internal" to count as being "free" leads precisely to hyperbolically distorted notions of agency. (Hence I cited the bad example of Kant above.)

So the negative conclusion that agency can not be identified with a distinct set of causes does not imply an absence of causal processes. But I think the compulsion to identify agency with causation, however vaguely, needs to be drawn into question. I think it results from the classical definition of volitional agency, whereby the "arche" of an action must be "in" the agent itself, which combines with the effort to logically account for actions as "in" the world. This makes prima facie sense. However, I argued above that the real locus of agency is in interaction with others and the internalization of relations with others, as established and stabilized through the advent of the realm of language and meaning. Certainly, a distinction between "inner" and "outer",- (a delimiting of boundaries),- is relevant and needs to be established, but it should not thereby be converted into a logical opposition. With respect to the realm of language and meaning, the distinction between inner and outer, a split between subject and object, is not relevant and does not make particularly good sense out of matters. Of course, we are only dealing with agency with respect to causality here, and leaving aside a host of other issues, such as the status of intentions or the identity of agents. But the fundamental issue is why we make attributions of agency and whether our attributions of phenomenal agency hold good, in the sense of identifying an irreducible locus of reality or efficacy. It seems to me questionable whether a direct identification of agency with prior causes would actually be substitutable in such attributions. If we were dealing with a species of highly sophisticated, factitiously plausible robots, would we still feel inclined to make such attributions or would some subtle sense of disturbance get in the way?

Thad, I made those three points in the original post. I just didn't number them and put them in list form. I did number a different set of three points and put them in list form. Still, I made them, and they're the ones relevant to the issue my comment was about, so I numbered them and put them in list form for clarity.

My use of 'merely' wasn't modifying 'states' but modifying what he states. It would have been clearer to say "he states merely that..." The context of his claim was insufficient as an explanation, because it affirms what I already said.

I'm not sure what it amounts to when you say an ACE isn't directly produced. Does it mean that it's uncaused as in random? Then how is it in my control? Does it mean that it's caused by something prior in my control? Then it just takes the problem back a step, and we can ask about the event that caused the ACE. Is it caused by something prior not in my control? Then it's like determinism. Is it caused by me? That also takes the problem back a step, because we can ask about the event of my causing the ACE. I'm not sure how denying that the ACE is directly produced can avoid this dilemma. It hasn't taken an option that I haven't listed, and it hasn't endorsed an option there and defended it. All it's done is agreed that one of the options isn't libertarian free will, but I already agree with that.

Johm, the stochastic thing sounds like what a contemporary philosopher of science calls indeterminism at the ground level consistent with determinism at the macro-level. If so, libertarians would resist it and compatibilist indeterminists like David Lewis would have no trouble. Those who say we have no freedom would be untouched.

I think the main reason most libertarians would want to resist this proposal is that it seems highly reductionistic. The main point of the libertarian view, for many libertarians, is to resist reducing agency to natural processes, even a complex conjunction of natural processes, which is what this view seems to involve. Many of these libertarians are dualists as well, but van Inwagen isn't. Even though I'm a compatibilist, I'm a dualist compatibilist, and I therefore also would want to resist such reductions.

As for whether the view responds to the argument, I'm not sure if it does. You seem to be saying that what causes my actions is a whole set of things, which includes the environment but also internal causes like my beliefs, desires, and character. A libertarian will say that if your actions are guaranteed by the conjunction of those then you're not free. A compatibilist will say that you still can be free despite that guarantee. I'm not sure if you're suggesting this, but I get the feeling you might want to say that this conjunction doesn't guarantee it but just makes it overwhelmingly likely. I think that's enough to make the libertarian worry, and it's close enough to determinism to make the compatibilist happy. So I'm not sure your view is much different, at least on the general issues I'm discussing, than what someone like David Lewis might say.

Jeremy:

I tried to google David Lewis to see if I could find anything succinct to read, but I could only find 2 encyclopedia articles, both of which proved to be forthcoming, so I'm flying a bit in the dark on that one. But since I'm not coming at this from analytic philosophy, I'm not so concerned with reconciling my view with the philosophy of science. At any rate, "indeterminism at the ground level consistent with determinism at the macro-level" sounds like something of the inverse of what I'm saying.

In terms of the 3 alternatives that you allow, of course, I'm closest to some form of "compatibilism" and completely out of sympathy with reductionist/epiphenomenalist determinism. "Libertarian free will" carries the drawback that it risks treating agency as, if not a magical, then an inexplicable phenomenon, hence insisting at once on its reality and unreality, as well as, the danger of proffering hyperbolic accounts of agency. Presumably, the libertarian view rests on a claim that any sort of determination does not accord with the semantics of "freedom" and "agency". However, I would be far more concerned with the natural language sets of words that we actually use to "operate" and attribute the notion of agency, as a source and guide to delineate the real phenomenon. And I am not concerned with whether agency accords with purely naturalistic accounts, so much as whether it can, in some sense, be said to be real, (rather than, e.g., a "necessary" illusion, as in Kant.) As for dualism, I am inclined to be non-dualistic, and, if I admit of a dualism, it is between physical/physiological causality and language/meaning, which I take to be irreducible and non-causal, though I think that the gap can be bridged, though not reduced, through an account of rule-governed behavior.

However, it does seem to me that there is some significant difference between my attempted account and what I can make out of your account. For me, agency is primarily specified at the locus of relations with others, (and hence its selectivity occurs in the medium of meaning, which is built over and interwoven with the pregiven causality of animal motility). Hence, with respect to the causal dimension, I emphasized what I termed its underdetermination, which I correlate with some indeterminacy and "inbetweenness" of causal processes. (Of course, such a social emphasis leads on to an ethical rather than an instrumental consideration of agency.) And as a further point of emphasis, such agency, though real and "free", is a limited and thoroughly finite phenomenon. To put it in the jargon of an alternate philosophical world, agency "ek-sists" and that is the phenomenon I would want to "capture" and conserve in any explanation. Thus my objection to "compatabilism" is in part a semantic-logical one: can that phenomenon of agency really be identified with its causal substrate? (This is a different question from whether there is a causal substrate.) If, in turn, one were to insist that there is nonetheless a rising tide of causal processes that envelops such contingent, limited, finite agency, I would respond that such a notion of causality amounts to an encompassing metaphysical claim that both is not and can not be substantiated and that is incommensurable with the phenomenal level at which the issue of agency arises. That was the point of my initial observation about the oddly subjunctive logic of claims of predetermination. For agency precisely means that the future is not predictable, but will be, if ever so slightly, otherwise,- (back to Lucretius' swerving atoms.) What is at stake in the question of the reality of agency is our confidence in the concept and its applications. But to reduce the question to a matter of causality, in my view, to mislocate and defend against our lack of confidence in the concept. Our lack of confidence in the reality of agency is our lack of confidence in our relations with others.

I'm sorry that I've repeated myself so much here, but I just wanted to pin down the source of any apparent disagreement.

I'm not saying agency can be identified with its causal substrate. I'm just saying certain kinds of causes need to be there and others cannot be there for it to be true agency. That doesn't mean agency is the set of causes. That's part of the dualism I hold. I think there are multiple levels of explanation, and you can't always reduce events or descriptions at one level to those of another. Sometimes there are necessary connections between one level and another, though.

I'm having trouble with talking about agency with respect to our relations with others. Do we fail to have agency when we make decisions on our own that don't involve other people? Am I just missing something in your view?

I’m not sure how O’Connor would respond, but maybe I can be more helpful. Here’s a stab…

Jeremy says, “I'm not sure what it amounts to when you say an ACE isn't directly produced.” O’Connor is a causal realist. So, ‘x’s causing y’ may be abbreviated as ‘x makes it the case that y occurs’ or ‘x produces y’. He defers to Professor Anscombe’s (1971) seminal paper ‘Causality and Determination.’ Causation does not imply deterministic causation. The thesis of determinism is not equivalent to the Principle of Universal Causation (every event has a cause).

Jeremy says, “Does it mean that it's uncaused as in random? Then how is it in my control?” O’Connor will say that it’s uncaused/unproduced, but O’Connor will get off the boat before “as in random.” No one knows what it means to be random, and anyone who uses it as a load-bearing feature in an argument against the libertarian should probably give an account of it. O’Connor will say, I think, that an ACE, if you reflect on the nature of the event carefully enough, is intrinsically controlled by the agent. O’Connor also says that the ACE just is an agent’s free action. So, the agent does not agent-cause an action. Rather, the agent’s causing of an event of the right type just is the action, the agent’s activity par excellence. The agent’s directly producing a state of intention (e.g., a state-of-intention-to-sip-scotch-in-order-to-satisfy-her-desire-to-taste-the-scotch) does seem to be something that is under the agent’s control. After all, it’s the agent’s making something happen and nothing outside the agent’s control makes her perform the action. But, again, I register my complaint (with Peter van Inwagen) that I have no clue what it means to be in one’s control. An ACE meets the libertarian’s condition of not being guaranteed by the hard past. Up until but not including the moment the ACE occurs, the causal future is not closed. That is, if an ACE cannot be produced, then there is no sufficient causal condition for the ACE (where C is a sufficient causal condition of E just if C occurs, C does not imply E, and it is a law of nature that if C occurs, then E occurs but no earlier than C). An ACE also seems to capture an event that implies that the agent originates some event essential to her first directly free action, where an agent performs a directly free action A just if A is free and the freedom of A does not derive from the freedom of any other action the agent performs. So, we’ve got alternative possibilities and origination. Throw some epistemic contraints into the mix and we’ve got a good libertarian stab—so says O’Connor.

Jeremy says, “Does it mean that it's caused by something prior in my control? Then it just takes the problem back a step, and we can ask about the event that caused the ACE.” Right, O’Connor would certainly agree with this, since we’re presumably talking about the intelligibility of an agent’s first directly free action.

Jeremy says, “Is it caused by something prior not in my control? Then it's like determinism.” O’Connor would agree, if by ‘caused’ here we mean full-fledged production a lá Anscombe, van Inwagen, Mele, Kane, Clarke, etc. However, there may be room for what O’Connor calls ‘structuring causes,’ deferring to Dretske’s (1988: chapter 2) distinction between trigger-causes and structuring-causes (see his quotation below). I’m unclear about the distinction, but maybe he just means to say that an ACE may have necessary causal conditions (where C is a necessary causal condition of E just if E occurs, E does not imply C, and it is a law of nature that if E occurs, then C occurs but not later than E). In this way, an ACE may have causal contributors that, from a causal realist’s perspective, just aren’t enough to make the ACE occur.

Jeremy says, “Is it caused by me? That also takes the problem back a step, because we can ask about the event of my causing the ACE.” O’Connor would agree. So, we stop the regress straight away.

It seems that the best way to object to O’Connor’s view is to provide a counterexample to his claim that in an event-causal chain, it must be the case that z produces x’s producing y only if z produces x. It might be worth quoting his argument in extenso:

O’Connor (2000: 53) in Person’s and Causes argues: "Consider a familiar sequence of events. My finger presses a doorbell button, the doorbell rings, and your cat jumps in fright. We may sensibly say that my finger’s pressing the button causes the causal sequence, the ringing of the bell’s causing the cat to jump. But what we mean here is simply that it caused the sequence indirectly, by causing the first element of the sequence, the bell’s ringing. We may also sensibly say that the electrician’s wiring of the doorbell system was a cause of the sequence, the depressing of the button’s causing the bell to ring. Following Fred Dretske [1988: chapter 2], we may term this latter episode one of ‘structural’ causation, which consists in establishing a causal pathway—here, the wiring and power supply—between two objects or systems that is subsequently triggered by some appropriate event. Here we mean only that the ‘structuring’ cause provided a context in which some causal factor exerted its characteristic effect. It is not to say that the establishment of an electrical pathway in any way brought about or enabled the button’s depression’s exerting its characteristic influence on its immediate environment, only that it will determine one important wider effect of that influence. Neither of these legitimate ways of speaking of causes of causings within event-causal contexts supports the idea of a cause of an agent’s causing his own intention. The first type of example has no analogue in the context of agent causation because the cause within the causally complex event, agent S’s causing e, is not itself an event but an enduring substance. The second example, involving structuring causes, is clearly applicable to free human agency."

And O’Connor (2002: 135-6) in his ‘Reply to Hendrickson’ reasons: "I argue that an ACE cannot have a triggering-cause—an event that stimulates the cause (in this case, me, a substance) into action. Consider first event causes. Where agent causes are absent, they are linked in constant flow of one event’s giving rise to another. Now, where event B causes C, there is not a direct triggering cause of B’s causing C. Instead, some event A brings about B, which brings about C, and so on. (We might say, if we like, that A indirectly triggers B’s causing C in virtue of triggering A.) In the case of an agent-causal event, however, there is no front-end event to be caused, only an agent qua substance. So there doesn’t seem to be room, as a simple conceptual matter, for an ACE to have a triggering cause."

This seems to be the principle O’Connor employs: Necessarily, if there is some event z being such that (i) (z makes it occur that (x makes y occur)) and (ii) z is finished no later than there begins any [hard] event strictly implied by (x makes y occur) and (iii) x neither occurs before nor simultaneously with z, then (z makes x occur).

Jeremy:

If you are saying 1) that any explanation of agency must conserve the phenomenon, else it risks not being an explanation of the explanandum, and 2) that a causal explanation does not necessarily contradict the phenomenon of agency, if the causal explanation is of precisely the right sort to allow for the phenomenon, then I don't think I have any narrow disagreement here. (I might quibble with the notion that the "right" sort of explanation involves the causes being "internal" to the agent, since I think the right sort of explanation involves interchange between "inner" and "outer".) I don't know exactly where your "dualism" comes in, however, since I've always viewed the notion that reality must be conceived as consisting of any number of emergent levels, for all that they can causally interact, as a way out of dualism.

Perhaps since I set my account in a broader context than the narrow issue of the causes of agent causality, focusing rather on what "constitutes" agency, I've been talking at cross purposes. Since I hold that what renders us the distinctive sort of beings that we are is language, which lifts us out of nature into a socio-cultural form of life, (which, while not prescinding from the naturalistic account of the world, sets limits to it), I derived agency as an "effect" of language, (since I would also hold that our capacities for discursive thinking ingredient in agency are also intertwined with language), and language entails intrinsically relations and interactions with others. So, let's say I'm a joiner and I decide to build a set of cabinets purely for my own use as suits myself, including its design and workmanship. That would be a case where my decisions only concern myself and do not effect any other. But I normally build cabinets to exchange them with others, i.e. to earn a living, and I have acquired my skills as a joiner, including my conceptions of design and workmanship, from interactions with others. So indirectly and circuitously I have the sort of intentions that I have as an agent in this case too through language and social interaction. In addition to the "logical" point that language intrinsically involves others and interaction, I would also maintain that self-relation, selfhood, is intertwined with relations with others, (hence all the issues about the identity of agents, persons, and the relation between agency and the maintenance of personal identity), and that the maintenance of relations with others, (under which I would include privative cases, such as rupture or the pursuit of hostilities), is salient in the orientation and motivation of agency.

I mentioned above that classically agency has been implicitly conceived in relation to causality, as at once its absence and its mastery. The resulting neglect of the relational dimension leads to conceiving agency in terms of control and self-control. This leads to 2 unfortunate tendencies: on the one hand, agency is often imaged in terms of domination, "the triumph of the will", and, on the other, "freedom" has often been conceived as the mirror image of necessity- ("Freedom is the recognition of necessity", a line that I have often seen attributed to Lenin, but that actually comes from a later chapter in Hegel's "Phenomenology", which Lenin was just quoting.) I once checked a book out of the public library on Foucault. The author was an unusually clear and fluent writer, far above the usual academic norm. He began by noting that Foucault is a skeptic, with an irrelevant nod to Pyrrhonism, and then proceeded to argue that the core theme in Foucault's work is "asocial freedom", expatiating on "asocial freedom" for 120 pages or so, without ever noticing that there is no such thing as "asocial freedom", any more than there is such a thing as "dry rain". I suppose the moral of this little story is that such is the extremity one might be led into through considering the apparent paradoxes and reifications of an (implicitly metaphysical) notion of "freedom".

By the way, my answer to the robot question I posed above is that we would feel a sense of disturbance in interacting with factitiously plausible robots. It would gradually grow through our sense of the response patterns we would experience, leading to the realization that these pleasing robots nonetheless lack a capacity for suffering and thus also for the assumption of risk. I think these two features are part of what is criterial for our sense of agency.

Thad, counterexampls abound to that principle. Here's the principle again:

Necessarily, if there is some event z being such that (i) (z makes it occur that (x makes y occur)) and (ii) z is finished no later than there begins any [hard] event strictly implied by (x makes y occur) and (iii) x neither occurs before nor simultaneously with z, then (z makes x occur).

Take the example of an egg's fertilization (z), implantation (x), and beginning development into an embryo (y).

The three conditions of the left hand side of the conditional obtain. Fertilization makes it true that implantation will lead to the beginning of development into an embryo. Implantation is then all that's necessary. Fertilization is certainly finished before the beginning of any event that could be described as "implantation makes the beginning of development into an embryo occur". Then, of course, implantation isn't before or simultaneous with fertilization. Yet the right hand side of the conditional is false. Fertilization doesn't make implantation occur. That kind of failure is extremely common, even without abortifacients.

The line you find in Hegel and Lenin sounds strikingly like Hume and, long before him, the Stoics.

I'm familiar with accounts of identity in terms of relations with other people, but I never thought people who said that sort of thing mean identity the way analytic philosophers do. I thought they meant what might be called your sense of identity, how you see yourself. You're defining agency in terms of the agent and the agent in terms of this sense of identity affecting by all sorts of social relations. Am I reading you correctly?

I sympathize with this way of seeing a sense of identity, but I want to distinguish that from identity proper. I'm not sure you end up talking about the same thing discussed by van Inwagen, O'Connor, Lewis, and other analytic philosophers when they talk about agency. I'd describe the sort of thing you mean as a different level of explanation, with social relations serving as causes and effects. I'd still want to keep agents, agent causes (if there are such things), and anything else about decisions or choices on a different level. They may cause these social relational elements, and social relational elements may cause things to be true of an agent. I just don't want to equate the two. Maybe you'll just deny that agents and agency can exist in the way they mean, but I'd have a hard time being convinced of that.

Jeremy,

Here's the principle again:

Necessarily, if there is some event z being such that (i) (z makes it occur that (x makes y occur)) and (ii) z is finished no later than there begins any [hard] event strictly implied by (x makes y occur) and (iii) x neither occurs before nor simultaneously with z, then (z makes x occur).

Yes, maybe counterexamples abound. But it’s unclear (to me) that you’ve provided a clear counterexample.

First, a caveat: note that O’Connor never formally or explicitly advances the precise principle in question. I just tried to extract it from his text, and I may have gotten it wrong.

I think your example is informative. Instead of being a clear counterexample, it seems clearer that you’ve provided an example of what O’Connor, deferring to Dretske (1988: chapter 2), calls a ‘structuring cause’. To perhaps oversimplify things, your (z) may be only a causal contributor or partial cause or enabling condition of (x’s producing y). E.g., supposing that determinism is true, then (z) may be only a necessary causal condition of (x’s producing y). But, intuitively, (z), if it’s supposed to produce (x’s producing y), should be a minimal sufficient causal condition of (x’s producing y). Again, assuming determinism is true.

So, perhaps I’m slow and need more help in seeing your example more clearly.

I’m also not sure what “that kind of failure” is supposed to refer to. Is it only implantation? Or is it fertilization’s producing the implantation? If the former, then we don’t have a counterexample, since it is necessarily the case that for any events F and G, F produces G only if F occurs. If the latter, are these extremely common cases also one’s where the egg’s fertilization produces the ENTIRE event ‘the implantation’s producing the initial stages of embryonic development’? I think there is room here for O’Connor to say that they are not. Or at the very least, it is not clear that they are, in which case we don’t have a clear counterexample but rather a borderline case. To be a counterexample, the implantation must occur; again, its occurrence is strictly guaranteed by the causally complex event ‘the implantation’s producing the beginning of the embryo’s developing.’ But if the implantation does occur, then something must have produced it. Well, what did? What produces the implantation if not the egg’s fertilization? If the answer is some other event E, then it would seem that the egg’s fertilization didn’t produce the implantation’s producing embryonic development after all. Rather the egg’s fertilization was a causal contributor. Perhaps E in conjunction with the egg’s fertilization produces the causally complex event, but not the egg’s fertilization by itself. But then this would confirm the principle in question.

And if we’re not talking about full-fledged causal production (when we refer to the relation between z and (x’s directly producing y) but rather mere causal contribution, then there is no counterexample to agent-causation. For no one thinks that an agent needs to be in control of (or personally responsible for) every causal contributor of her directly free action. In other words, an ACE may have causal contributors (e.g., necessary causal conditions) without there being something that produces it. O’Connor’s view is not one of a wholly-unmoved-mover, but rather one consistent with being neither a wholly-moved-mover nor a wholly-unmoved-mover.

Jeremy:

That there would be any number of instances in the classical tradition of philosophy identifying freedom and necessity was just the point I was making. By some accounts, Hegel was the last full-fledged metaphysician of the classical tradition, and his tag-line can be taken as a summation or encapsulation of the metaphysical view. But, of course, "freedom" can not be simply identified with "necessity", else there is no criterion for differentiation between the two. There must be some sort of distortion or conceptual gerrymandering at work here, under the metaphysical compulsion to unification or systematicity.

I was not discussing or deriving agency in terms of personal identity. (What the relations are between agency and personal identity is a whole other set of issues.) I am flying a bit blind when you refer to what analytic philosophers mean by "identity" and "identity proper". I take it you mean logical identity, somewhat along the lines of Leibniz' "principle of the identity of indiscernables", given a nominalist caste that it must come down to particulars. But my account located agency at the level of the meaning potentials of language, which allow for a decision between alternative possibilities. (So the question about how agency can act as an "uncaused cause" amounts to a question about how language maps onto the world and how the world emerges into language, that is, a question about the adequacy of natural language. On more or less pragmatic grounds, I would make a reasonable assumption that, because of its recurrent, well-worn usage, natural language does possess a high degree of adequacy in discriminating the world of our phenomenal experience and that "things" themselves really do possess suitabilities that support our language usage in how we speak of them.) The claim then is that meaning is a set of non-causal relations that is nonetheless intricated in a causally determined real world, and this would account for the apparently uncaused status of causal agency. Hence the round about explanation of agency leads to interaction and relations with others as the "source" of language. "I" only have a language, if "you" have a language and so does "he". This is a level of reality that is not describable in terms of a subject/object split, but which generates and is generated by collectively distributed potentials and which is also the level at which discriminations as to reality are made. The upshot is that, if one hasn't taken this level of reality into account in explaining agency, a causal account of agency is really just an account of organic causality, not of agency proper. So perhaps your surmise that I would be inclined to deny that agents can exist in the way the analytic philosophers want to talk of them is correct: such explanations of agency might strike me as short-circuited. And I don't think that matters can be sifted out to get down to an explanation only in terms of particulars. Rather I think a meaningful explanation involves a complex of items that belong together in a conjunction, so a certain degree of holism is required to arrive at an explanation.

I might add that I think most of the really interesting philosophical issues concerning agency are on the other side of the question of its causal basis. I did drop a hint at the end there that to consider agency with respect to causation is implicitly to emphasize only the active aspect of agency, whereas passive features are just as ingredient in the phenomenon. Agency is at bottom a mode of orientation in and to the world, and, in this, responsiveness to others is a key background feature, which is perhaps missed, if one addresses the matter in too "logical" a fashion. But the fact of the matter is that we do routinely make attributions with respect to agency. (Whether correctly or not is a matter of cases and not a singular generalization.) This would be the case, even if agency would have no causal basis, (though I don't think our ordinary conception of it entails any such claim), and, presumably, it would still be the case, if we had some sort of perfect causal account. Which is to say, that certainty as to the causal basis of agency is not relevantly at issue in our considerations of it, but rather uncertainty, a certain indeterminacy, is a key feature of our notions of agency. Perfect clarity on the matter may amount to a kind of blindness that loses the phenomenon. Going back to Aristotle, the intrinsic uncertainty attaching to action was the prime criterion for differentiating practical from theoretical reason. (Whether he meant the uncertainty of outcomes or the uncertainty with respect to the variable situational determinants of action is not clear: most probably the latter, but possibly both.) Perhaps part of the problem here lies in the attempt to give a theoretical answer, (with requirements such as finitely prestateable conditions), to a question of practical rationality. Something of the stakes involved can be gotten at with the notion that the capacity to exercize agency is dependent on mutual recognitions between agents, that is, to some extent, agency is real, only if it is recognized. Negatively put, the failure to recognize agency can do actual injury to the capacity for agency.

John, I think you're confusing what I'd want to say with what analytic philosophers in general would say. Many analytic philosophers look at language to figure out the best candidate for what our terms refer to. Some conclude that there's nothing like the traditional entity thought to be referred to by terms like 'I' and 'me'. They think there are processes or entities extended over time or sets of indivisible simples or regions of spacetime or mereological sums of indivisile simples, and these are the best candidates for our terms. Some think the terms are indeterminate in some fashion between the different candidates. What you're saying is of a piece with all that, because it takes social and relational factors into account in terms of what really exists and what goes into making whatever realities are there that our terms refer to. I think analytic philosophers could develope a view along these lines fairly consistently with the regular analytic metaphysical tradition.

I'm saying I don't like that sort of thing, because I think the traditional notion of an agent is correct. It has nothing to do with analytic philosophy. I think the realities you're describing exist, and I don't doubt that they affect the very nature of human agency, but I guess I want to resist identifying agency with those realities.

Jeremy:

Well, I'll try once more, before leaving you alone, if only to clarify the difference between our apparent views. I was surprised in the first place that "agent causation" could be at issue, which I probably why I responded to your initial post. Agents, after all, do something, that is, produce an effect, cause something to happen; if there is something definitional for agency, that is it. (I just looked up "do" in the dictionary; all but a very few of the many definitions are basically predicated of agents.) If I can construe the issue you are addressing, it is the question of the very intelligibility of agency, of whether and how agency is at all intelligible. And since this question is conceived to be the fundamental and prior one, it is taken to be the properly philosophical question. And this question is set up in terms of the cause of causal agency; that is, if agency produces effects and thus is a cause, what, if anything, is the cause of agency? But, (leaving aside any question of whether this implies a too linear conception of causality), if this question amounts to something like, e.g., "How does my "will" cause my arm to move to pick up and drink from the cup of coffee on the table in front of me?", then that is not really different from a question about how a dog sniffs out and digs up a buried bone. That is, as I stated above, it is not really a question about (human) agency, but rather one about animal motility and the functionning of the "voluntary" nervous system.

To review briefly, I stated above that if the question about the causal determination of agency is legitimate, then it must involve what I called underdetermination; that is, the causal substrate of agency must be sufficiently open or indeterminate in its outcome to allow for a loop whereby a decision by an agent can feed back into the causal substrate and cause it to cause. And I specified the decision of the agent as occurring on the level of meaning interpretations in language or language-thinking, which is overlaid upon causal processes. (This decision as meaning interpretation, leaving aside any quibble about whether interpreting something is doing something, would, if anything fits, be Thad's ACE above.) And the feature of language that allows for its interpretation of a given state of affairs in terms of alternative possibilities is its recombinant symbolism. Thus far, everything specified is internal to the immanent "moment" of agency. But the next step I make departs from philosophical orthodoxy. Since the rules of that recombinant symbolism must ultimately be generated by and derived from the interaction by which language is exchanged and transacted across the world, and since language intrinsically involves a relation to the other, I postulate another loop intertwined with that between the meaning-decision and the causal substrate, which effects, however indirectly, the decision loop. It is through passing through the relation to the other, in what the French/Jewish phenomenologist Levinas would refer to as the "exteriority" of its existence, that the agent is "bent" or "folded" into the kind of self-relation that allows for the decision of the agent, which has a causal effect, to amount to a "self-determination". This move abstracts from the psycho-physical immediacy and interiority of the "moment" of agency itself, and that may be part of what you resist. But it is proffered as a condition by which to account for what seems to bother or puzzle you about causal agency ,namely, its apparently self-produced quality. I am arguing that only through the relation to the other and through real or potential interactions with others is that apparently self-produced quality itself "produced". (It could be objected that I am already attaching agency to selfhood, but, if so, only minimally and not to any higher order conception of self that would entail coherence in the present or consistency over time, such that the focus on the moment of agency itself is not here violated.) Perhaps a further resistance to this condition would be that the relation to the other involves some sort of influence or constraint by the other that "contaminates" the agent and violates or puts into question the immanence and interiority of the choice of the agent and its possible motives within the moment of agency. But I think that, while such "contamination" does occur, it by no means impugns the separateness of the agents choices, decisions, and acts. (And besides, I am not required, nor would I want, to produce an entirely non-conflictual account of agency.)

Now it might be complained that I am merely displacing the question of agency onto the question of language and meaning and the socio-cultural world that language generates, structures and mediates, with all the tangle of problems and issues that involves. And that would be correct. (I'm pretty much transferring the issue squarely into Wittgenstein country.) And whereas I think the basic set-up of the world of language and meaning can be described, I don't think that it can be derived from or reduced to causes. (Causal accounts of meaning have been tried, mainly in the empiricist tradition, but I don't think that they work or can work. I've no doubt that attempts or projects for a causal account still occur, whether under the auspices of philosophy or cognitive psychology. I don't know if that is your cup of tea, but you could offer such an account, if you are so inclined.) So the question about the cause of agency has been evaded. But my answer is that the socio-cultural world of language and meaning is precisely the one in which the phenomenon of agency occurs and one can not, as it were, step out of that world and look at agency from without, so as to to examine agency "in itself". That would be metaphysical, in the perjorative sense. And if one were to examine the causes of agency with a microscope, as it were, one may find such causes or one may not, but what one would not find is agency. I don't know if you are operating from the assumption that it is only the natural world that is truly real and that the socio-cultural world only has a lesser, derivative degree of reality. Obviously, I do not hold to such a view. But if we are dealing with the phenomenon of agency and trying to conceive of and account for it, I don't think we want a conception that is "more" than we actually need, that is, other than what the concept is good for and how we use it. Now a common sense conception of agency, carried by natural language, is ingredient is a vast variety of customary practices in the ordinary world, in which we variously assign intentions and ascriptions. But it is the problematic cases of conflict or contention that come to the fore. However, even then, we still use our assumptions about agency in identifying and handling them and our conception of agency remains ingredient in how we construe even the most difficult cases. That is, though there may be a wide variety of cases in which one or another of the criterial features of agency is found wanting, there is no case where they all are found wanting. (This is just an anti-skeptical point. An analogy with neurology might help. One way that neurologists investigate brain function is by examining instances of specific brain damage which correlates with dysfunction or absence of function. But complete brain damage that correlates with a complete absence of brain function would not be instructive.) The point here is that the question of the reality and efficacy of agency is not a theoretical matter that can be generally decided once and for all. Our intentions and ascriptions may be real or may be illusory. But what tests this is the ongoing course of events and interactions following on in particular cases. This resembles a basic pragmatic point, except that the criterion provided by the follow through is not the success of actions, but rather the possible counterpositions of interactions. This is why I suggested at the end of my last comment that the question of agency is really a matter of practical reason rather than of theoretical reason, that is, of the orientation of reason toward resolving conflicts, (whether within or between agents), rather than toward solving problems.

Finally, I don't understand why you take my account to be impugning the reality of agents or persons, of "I/me", "you" or "he/him". I don't take my account to tend toward the dissolution, depersonalization or derealization of human beings, whether on post-modernist or objectivist grounds. Such instances can happen, but not as a general implication or tendency of my account. What my account tendentially is leading to is to criticizing and drawing into question a deeply embedded traditional philosophical assumption about agency, namely its identification with autonomy. (But that would also be its self-producing quality). I think that the notion of "autonomy" confuses both the existential separateness of human beings and their collective distribution in a community and leads on to the dominative and necessitarian conception of "free will" that I mentioned above. But it would not be out of any desire for the dissolution of human beings that I would draw out this criticism. It is rather out of concern for their actual capacities and freedom and out of respect for human finitude.

At any rate, such is the rough-hewn account of agency that makes sense to me. I don't think it leaves any mystery to the concept or why it would occur. Again, any mystery would pertain to what we may encounter through and on the other side of the concept.

Jeremy,

You seem to be one of the few bloggers I read regularly who isn't on vacation so I thought I'd bother you here too.

As for the 'silly' objection to agent causation. I'm not sure how silly the objection is. It looks like when we say 'The tree caused the river to flood', someone is going to say 'Well surely the tree didn't cause the river to flood, it was the tree's falling across the river that caused the flood or it was the tree's being so massive that caused the river to flood'. Some will be unimpressed with the original statement for failing to highlight the salient features of trees that make them good flood starters and in order to supply a causally relevant feature will insist that strictly speaking the true causal statement should cite a state (i.e., the tree's being green and the tree's being lovely won't explain this), others will be unimpressed with the original statement's inability to explain an event that was a change (i.e., the tree's remaining placidly in place or existing doesn't explain why the world was suddenly flooding). So at best, it looks like a causal statement involving a term for a substance is acceptable only if something in the neighborhood can figure in a 'real' causal statement. I think these sorts of problems provide a more serious way of making the silly point. I don't happen to think that these points are fatal, but it explains why someone would think to say a substance caused an event or a state is a category error.

The other point is just that it seems that your animus is directed to libertarianism and I'm not sure that every defender of agent causation is a libertarian. John Hyman and Maria Alvarez try to shape up the notion of agent causation but there is nothing in their writing that suggests they are the least bit sympathetic to libertarianism. In fact, they argue that the standard formulations of agent causation are committed to infinite regresses and think their account gets around this worry.

I think there's an easy response to the silly objection, even when you've put it the way you did. Some people really want to think of causation as coming from substances, because that's how our language works. The tree does cause the river to flood. It does so in virtue of being so big and having fallen. This doesn't get the kind of agent causation libertarians want, but there is an ordinary language reason for preferring to see substances and not events as causes. That's why I don't think this objection is that strong.

Is this the sort of thing Hyman and Alvarez are up to? Or does it have to do with something like dualism or non-reductive materialism (which for these purposes don't turn out to be much different)?

Jeremy,

Part of their reason is 'ordinary' language, but the means to their ordinary language end is rather complicated. I think what makes their paper interesting is that their defense of agent causation doesn't require us to put our faith in dualism or any substantive view about the nature between the mental and the rest as far as I can tell. (While I'm giving the paper a plug, I might as well give the reference: Hyman and Alvarez, "Agents and their Actions", Philosophy (1998): 219-45).

Perhaps the most significant facing challenge facing the agent causalist is the Davidson inspired thought that actions can be thought of as events with special causal ancestry. They are interested in showing that actions cannot be understood in terms of event causation, in part, by joining Kent Bach in denying that actions are a species of event.

The nice thing about this strategy is that if successful it frees the agent causalist from what Strawson would call the 'panicky metaphysics' of the libertarian. Worries about getting one's swerve on, as the kids say, don't even enter the picture. It might be an interesting question whether we can make sense of unconditioned choices and the libertarian's conception of agency (I say not), but it seems that on their view, settling this issue has little if anything to do with the question as to whether there is real agent causation.

Regardless of the nature of time, from everything we can percieve (which is what our knowledge is limited to), the universe does operate only in cause-effect chains.

None of the explanations thus far presented for agent causation are satisfying because the simple fact still remains that our "decisions" are caused by chains of events that always end outside of ourselves (in fact, they can be traced back to the very beggining of "time"). When I make a decision, I take into account the relevant factors (which lie outside myself), and wiegh them based on the bias of my "personality", which was created entirelly by the collective experiences and environments of my whole life (again, things I had no control over). We don't have control of our own personality (bias) because, though we do make some decisions about it, those are merely the effects of causes that can be traced back to events outside our control.

Here's the question you have to answer to disprove absolute causality: "What isn't caused by something before it?"

Actually, your last question assumes something the libertarian won't grant, i.e. that X's causing Y always involves X's necessitating Y. Even most compatibilists have no trouble saying that a cause doesn't necessitate its effect. As long as a cause only makes it extremely likely or even somewhat likely, the libertarian can say that everything is caused, and that doesn't mean we don't have libertarian freedom.

Anyone who wants to comment further on this post or the ensuing conversation can do so at the cross-posting at my personal blog. I'm closing the comments here due to spammers trying to take advantage of posts with open comments at a currently non-operational blog.

Due to a bug in the new software for this blog, disabling comments stops all the old comments from showing. I've reenabled comments on this post for that reason.