Well, it’s not invalid. But it may be beside the point. Or it may not! It depends!
There would be two claims on the table:
Brown bunnies exist.
Brown bunnies are what we’re after when we talk about freedom of the will.
Claim 1 is trivially true. Claim 2 is trivially false. The brown bunny puzzle is solved!
What I’m saying is that that when it comes to the analogous claims:
1’. Uncausedness in Dennett’s sense allows for assignment of personal responsibility
2’. Assignment of personal responsibility is what we’re after when we talk about freedom of the will
there’s nothing trivial about the question whether these claims are true or not. 1 and 2 are trivial. 1’ and 2’ are not. This puzzle is definitely not solved. The brown bunny puzzle, which has been trivially solved, seems to have nothing to do with this puzzle at all, so I’ve forgotten why it was brought up.
It looks like you’re sympathetic with 1’, and disagree with 2’. Is that right?
Frylock, I don’t think it’s worth continuing to argue about this; you have a valid point, but, IMO, the two definitions of free will have sufficiently little in common that using them alongside one another breaks the discussion. It’s like having two conceptions of god, one of which is a transcendent, all-powerful being and creator of the universe, the other is, I don’t know, a Spinozan embodiment-of-nature kind of being, then arguing when we talk about the former (or think we do), we really talk about the latter, then providing an argument for the latter, and then going on to claim, since all talk about the former was really about the latter, the former has been vindicated as well. This, not the question of whether Dennett’s conception of free will is legal, was the point of my analogy – I have no problem with his or any conception, it describes an interesting concept that is well worth thinking about, but so does the ‘could have done otherwise’ concept, which seems to be crowded out in compatibilist arguments (since everybody is ‘really’ talking about another kind of thing).
Pbear, going back to your OP, Do I have you right when I summarize your post thusly?
Libertarianism doesn’t work because it’s clear that our choices are determined
Dennett’s compatibilism doesn’t work because, since according to compatibilism everything is determined, and since assignment of responsibility requires that an action not be determined, compatibilism doesn’t allow for the assignment of responsibility–but the ability to assign that responsibility was supposed to be one of the selling points of Dennett’s view.
Soft compatibilism doesn’t suffer from either of these problems.
Half Man Half Wit, I don’t know if you’ll respond to this, but I really don’t see how the definitions of “cause” are so different. We agreed in post 23 and 24 that “cause” means necessary condition. Since then, I’ve only proposed one other thing:
There are some proposed causes that, while technically are necessary conditions, are so vacuous and/or tautological that they would be rejected in any serious attempt to actually determine causation. Examples
[ol]
[li]What caused the 69 deaths on Utøya?[/li]46 deaths per hour for 1.5 hours.
(Actually, this isn’t a necessary condition)
[li]What caused the recession?[/li]A sustained decrease in gross domestic product.
[li]What caused our high unemployment?[/li]Lots of people being unable to find employment.
[/ol]
An event is “uncaused” if all possible necessary causes are either vacuous or tautological.
I think that defining cause as necessary condition is a non-starter. I can cause a match to light by striking it–but it’s not necessary for me to strike the match in order to cause it to light.
In fact I’m not sure causes are ever necessary conditions.
The problem with the notion of vacuousness is that it lifts up the decision of whether or not an event is caused to the level of human reasoning, which I think is a massive level-confusion. Whether or not something is caused or not should not depend on whether or not we feel its putative cause is enlightening.
As for tautologies, ultimately, all explanations are tautological – I don’t see why that would render an event uncaused, either.
Perhaps it helps to regard causes informally as reasons why something happens. The fact that 46 people died per hour for 1.5 hours is very much a reason for the fact that 69 people died in total; that the GDP decreased is a reason for the recession; that people find no jobs is a reason for unemployment. The reason why we don’t think of these things as ‘causes’ is that we don’t think of either them or their ‘effects’ as events, but in more abstract terms; causes in a more specific sense could then mean reasons for events.
Frylock, the problem with leaving out the ‘necessary’ is that if we do, we can’t point to anything as a thing that must have occurred in order to have another thing occur – but if there’s nothing that must have occurred prior to some thing’s occurrence, then in a sense there’s ‘no cause’ to some event. (That’s a rather clumsy sentence, but I’m too tired to go back and rephrase it…)
But I still see no problem with just forming the disjunction of sufficient causes to find a necessary cause, because, in order for the match to be on fire, it must have been the case that either you struck it, or you ignited it with another match, or you subjected it to heat beyond its ignition point, or…
And in general, the more in detail you describe your system, the less possible causes will there be, up eventually to microscopical time-reversibility, were knowledge of the state and the dynamics of the system allows you to calculate every prior state, yielding a necessary cause.
I don’t see a way to avoid intertwining causation with reasoning. Nature by itself gives us very little information on causation. If all you know is that determinism is true, and that the universe went through state A and then state B, you can only conclude that B will follow A again, if A should ever recur. If determinism works in the other direction, you can only conclude that A will always precede B. Even with more information, I don’t see how it changes much. The laws of physics are mute when it comes to necessary conditions. By themselves, they tell us only sufficient conditions (if the system starts like this, and evolves for some amount of time, the we get that, or vice versa). To get necessary conditions, we have to look for them—through experimentation or by using our mathematical intuition to discern which terms are important and which are not. This whole undertaking is governed by our own sense of what is “reasonable.” For example, by examining the mechanical equations of a coin flip, we might conclude that the mass of the Andromeda galaxy is not a causal factor. Why? Because the outcome wouldn’t change if we altered the mass of Andromeda, or even if we remove it entirely from consideration. Unless we made it much, much larger. Is this counterfactual sufficiently different from the real world that we can ignore it in our causal calculations, or do we need to admit that Andromeda’s mass is a causal factor in the coin flip? Unless we have some idea of what is a reasonable counterfactual, it seems like our necessary condition will quickly encompass the entirety of the universe.
I can’t imagine why all explanations would be tautological. There are statements in logic that aren’t tautological, so why do explanations comes exclusively from the ones that are? How is the statement “The diversity of species is explained by evolution via natural selection” tautological? There’s not even a purely logical reason why it should be true!
I’m not sure what to say about my three examples, particularly the GDP one. A definition of a recession is falling GDP (for two quarters). Saying that a recession is caused by its definition being true strikes me as incorrect use of the word “cause.” It’s like saying that the Super Bowl is caused by the national championship game of the NFL. If 46 deaths/hr and 1.5 hours caused 96 deaths, could we also say that 96/n deaths per grain of sand and n grains of sand on Utøya is the reason? What if I multiplied and divided 96 deaths by the temperature that day? Or the number of active volcanoes in the solar system?
It’s just that, if A happened because of B, C, and D, and if B, C and D are not the kinds of things we would like to call ‘causes’ for whatever reason, to me, it still just doesn’t seem right to me to call A therefore uncaused – it didn’t happen for no reason, or ‘just happen’, after all. That would only be the case if there were no B, C and D. The notion or definition of ‘causation’ you employ when you exclude certain causes is not the same notion used in calling something uncaused, at least to me – that’s where I think things get somewhat word-gamey.
As for tautological explanations, I gave an example earlier – wrt natural selection, since it is a mechanism leading to diversity, one could say that ‘species evolving according to a mechanism leading to diversity leads to diversity’; to me at least, this does not seem substantially different from ‘flipping the coin so that it lands heads causes the coin to land heads’. But you wouldn’t now call biological diversity unexplained.
As for the examples, it’s a question of what constitutes an explanation – what is relevant to the to-be-explained. The number of grains of sand isn’t, for example: a difference in that number would not lead to a difference in the number of deaths, while a difference in the time the massacre lastet likely would have.
As mentioned upthread, the idea of using emergence to solve the free will puzzle occurred to me as I was reading the passage on causation in Dennett’s Freedom Evolves. I thought, hey, isn’t that an emergent property? What’s funny is that, at the time, I couldn’t have strung together twelve words describing what the phrase means. It’s something I’d bumped into from time to time but never had need or occasion to understand. So, I turned to my internet machine to see whether my hunch was right. What I discovered, based on the Wiki and SEOP articles linked in the OP, was that he was making a more-or-less emergence argument. (This was based on my in effect reading, which as I’ve said may well have been wrong.) I thought, how clever. That’s a compatibilist stance which might go through and went back to reading the book. Only to discover Dennett doesn’t develop the argument along those lines after all. Frankly, I didn’t find his thesis persuasive but I wondered whether something similar could be made to work by exploiting my initial hunch. So back I went to the internet machine and read everything I could find, of which the three best papers were those linked in the OP. But best of all was the book Emergence (2008), edited by Mark Bedau and Paul Humphreys, collecting previously published essays by various philosophers and scientists.
At first, when I hadn’t yet discovered the book, I developed the argument in the alternative. That is, I wrote an argument showing how easy it is for compatibilism to go through in a naturalistic way if one assumes strong emergence is true (as several philosophers have argued, especially Paul Humphreys). Importantly, no magic is involved. It’s a real interaction of real things. Whether the interaction can have causal effects not accounted for by the parts is the question. In the alternative, I composed an argument based on weak or descriptive emergence, the form which eventually became the OP. I abandoned the strong argument for several reasons. First, upon finding and reading Emergence, I discovered no scientist will have anything to do with strong emergence or irreducibility in principle. Second, no one seemed to be able to point to a clear example of it happening, or even a reasonably contested or disputed example. All I could find was speculation without evidence that consciousness might be such a thing. Third, I learned a good argument can be made for downward causation in a weakly emergent system. That seemed to me enough to put through a compatibilist thesis. Fourth, as mentioned in the OP, although scientists assume reducibility in principle, irreducibility in practice (in a complete sense) and inability to work reductionism in reverse are the norm. In such cases, a macro theory based on observation is as good an explanation as we’re going to get.
The third point is based mainly on Bedau’s work, Chapter 8 of Emergence, originally published in 2003 (see here) (marked as a draft but I’ve noticed no differences from the version published in Emergence and the link is to his college website) (the original published version is available here, but it’s a much larger download). This article was an extension of a 1997 paper (see here) in which he apparently introduced the phrase “weak emergence,” which others sometimes call descriptive, predictive or epistemological. His further argument in this article is that weakly emergent systems can have downward causation. (He also argues weak emergence can have explanatory autonomy in some cases, e.g., traffic jams, but I don’t think emergence of the mind is one of those cases.) The full argument is too long to give here, so I’ll quote what seems to me the crucial passage (p.177):
The fourth point, irreducibility in practice and inability to work reductionism in reverse in most cases, is made many times in the science essays in Emergence. I particularly like the statement of the problem by Philip Anderson, winner of the 1978 Nobel prize for his work in solid state physics. The original essay (see here) was published in 1972 and appears as Chapter 10 of the book (p.221):
To similar effect, see Robert Laughlin & David Pines (also physicists), Chapter 14 in Emergence (originally published in 2000) (see here) at p. 261, “The emergent physical phenomena regulated by higher organizing principles have a property, namely their insensitivity to microscopics, that is directly relevant to the broad question of what is knowable in the deepest sense of the term.” They then illustrate by various examples how, as a practical matter, almost all advances in physics and other sciences are possible only by experiment and approximation, rather than reduction to fundamental equations. In sum, they say at p. 265, “Rather than a Theory of Everything we appear to face a hierarchy of Theories of Things, each emerging from its parent and evolving into its children as the energy scale is lowered.”
With that as background, what I think emergence brings to the table is a useful way of looking at complexity. On one hand, if we accept (as I do) that the brain is a natural system and reducible in principle, it’s ultimately determined in a Laplacian sense. On the other, if we accept that this system exhibits volition, we can look to emergence as an explanation, or at least the best description we can manage, of how this ultimately determined system acts as it does. I’m happy to leave that as a compatibilist compromise. But, if we are to apply Occam’s Razor as HMHW suggested earlier, what we should cut is the useless abstract deterministic reduction. Rather, it’s the observable volition (subject to observable constraints) which matters. Whether the mind observably demonstrates volition (and the constraints, for that matter) is, of course, a legitimate subject for debate.
Sort of, but not quite. My problem with libertarianism, as I explained to Dr. Love in Post #18, is that I think behavior is influenced and that when we give those influences their appropriate weight, we end up with a form of compatibilism. As for Dennett, I was using him in the OP as an example of traditional compatibilism. In that sense, you’ve fairly characterized my objection. Or rather the objection of the SEOP article. Stated a little differently, I don’t think Dennett delivers on the promise of giving his verson of compatibilist free will (or what I think can only be called responsible agency) the same content as what what most of us call free will, or even what I define as volition. Having failed in this regard, he doesn’t answer the ethically hard question (which he acknowledges) of whether it’s fair to punish (legally or socially) those who do wrong.
Whether my model of soft compatibilism solves this problem is, of course, what we discussing.
But the question I have is still: How does volition emerge, and what does it do? What, to the best of my ability to tell, you seem to be doing is to assert that volition is an emergent notion, to then proceed to elaborate on emergence in general. That’s all very fine, but it doesn’t give me a better an understanding of your model – indeed, lacking such justification, there’s hardly a model at all.
Take snowflakes. Their unique shape is an emergent property, but just to say so isn’t very illuminating; rather, the interesting thing is to describe why and how such shapes emerge, and while we in general can’t as a practical matter predict the shape of a snowflake, we understand very well how it comes about.
As for volition, it clearly isn’t present in the fundamental description, so it may well be emergent. But it also may well be illusory – we may simply be mistaken about possessing volition. Absent any justification for the emergence of volition, both possibilities remain open – and absent any need for volition, Occam would have us conclude that it’s just not there. Moreover, weak emergence can’t introduce any wiggle room in order to make genuine choices – since the micro level always determines the macro level, determinism on the micro level means determinism on the macro level, even if it’s not obvious.
The need for volition is the other thing I don’t see in your posts. You’ve stated that it’s ‘just a label’, which would suggest we can do away with it without consequences, obtaining a description of the world that is just as good. Also, and I’m sorry if I just missed this, but I don’t see where you have given an argument to the effect that there are problems determinism can’t solve, while volition does, and discussed how it solves them. And moreover, the stances that volition is just a label and that it has any explanatory utility determinism alone lacks seem to be contradictory to me.
Speaking very opportunistically, why should I believe in volition? What do I gain? Why should I believe it exists?
As for my stance on Dennett, I think the problem I have is really twofold. One, it seems to me that the argument proceeds somewhat as follows: 1) lack of strict causality allows freedom, 2) let’s define strict causality in such-and-such a way, 3) plainly, the just defined thing doesn’t exist, 4) hence, we have freedom!
But the causality defined in the second step is not the causality in the first one, or at least need not be. And if Dennett uses the notion of causality as it was presented in this thread (or rather, as I understand what was presented in this thread), I don’t think it is. Freedom is granted by the existence of uncaused events only if uncaused means that there is no reason for the event to happen, if it ‘just happens’, or something like that. But in Dennettian causality, things always do happen for a reason, we just don’t call all of these reasons causes. Equivocating one notion of causality with the other is, I think, where things go wrong.
The second problem is that then, it appears, one goes on and defines an appropriate notion of freedom allowed by the Dennettian notion of causality. So, Dennettian-uncaused events grant Dennettian freedom – which I grant. But what have we learned? The conclusion is trivial, following purely definitionally: Dennettian freedom is defined such that it is granted by Dennettian causality. So all that we get out of it in the end is that there is a definition of the words ‘causality’ and ‘freedom’ such that we have this particular kind of freedom in a deterministic world. But that’s hardly a revelation – we could have just defined freedom as whatever it is we have, and come to the same conclusion. The interesting questions are completely missed by this exercise in definition-making (well, one set of interesting questions, anyway; as I said, I think that he makes great strides on the front of moral responsibility, which is of course a pretty interesting question as well).
So you’re worried that people could be wrong in their assessment of causation under this theory? That will happen with any reasonable theory of causation. Besides, I’ve given conditions about when B, C, or D won’t count.
I assume you mean the gravity example from upthread. You said “And of course, ultimately, all explanations can be cast into such a tautological form.” A statement like this potentially confuses two distinct propositions:
[ol]
[li]For any event, there is a tautological explanation.[/li][li]All explanations are tautological.[/li][/ol]
Proposition 1 is true, and obvious. One can always say “Event E was caused by the event E occurring.”
Proposition 2 is false. In order to get some explanations into tautological form, you have to remove information. You simplified natural selection to “a mechanism leading to diversity,” but the theory of natural selection is considerably more complicated than that! In the other example, the statement that “gravity is an 1/r attractive potential” is not a tautology. Not all vector fields are conservative (e.g., magnetic fields are not). Of those that are, not all of them have a potential that decreases with 1/r. A priori, there’s no reason that either of these things should be true of gravity. Thus, the statement is not a tautology.
Compare this to your suggestion to “forming the disjunction of sufficient causes to find a necessary cause.” Logically, this statement will be true for every possible event! This is a tautology. Do you see the difference?
In science, the concept of falsifiability is a pretty popular one. Tautologies are one kind of unfalsifiable statement. Why is falsifiability important in science, but not when investigating causation? Or are you attempting to argue that there are no falsifiable statements?
Isn’t this rather question begging? You seem to be assuming that deaths/hour is some fundamental property of the massacre, but deaths/grain isn’t, and basing your conclusion on that.
Or, consider another thought: suppose that I take all possible values of (deaths/grain) and (number of grains), and form a disjunction of them. By your own reasoning, haven’t I just created a necessary cause? What makes the disjunction of (deaths/time) and (time) to be “causier” than (deaths/grain) and (grains)?
No, I’m not worried that people may consider the wrong causes, I’m worried that different ‘cause-assignments’ yield different judgement on whether or not something os caused or uncaused – that causation becomes a subjective matter, from which allegedly objective conclusions (such as whether or not we are free) are drawn. Whether you call something a cause or not does not have any bearing on whether it happened for a reason or not.
Regarding tautologies, you claim that being able to formulate the reason for heads coming up as ‘throwing the coin such that heads comes up’ means the description is void, and hence, no cause can be pointed to; I merely pointed out that you can do an analogous thing with every explanation, but we don’t consider these explanations void for that.
And not every conceivable event can be given a necessary cause by disjoining sufficient ones – those that don’t have any, that I would consider uncaused, can’t.
And regarding what constitutes an explanation: do you think varying the number of grains of sand on the beach would have made a difference in the number of deaths? Do you think varying the amount of time wouldn’t have?
You’ve talked about varying the comditions of some event counterfactually to find its causes, so I don’t see where the problem lies.
Anyway, I’lö be away on holiday for the next twp weeks or so, so don’t be offended if replying might take some time…
Half Man Half Wit: I realized earlier today that I’ve not thanked you for taking the time to think about and comment on my model. As I said to Dr. Love, I was hoping for push-back in this thread and you’re certainly doing that. So, thank you. Also, I will mention that I have a humble skepticism about my model for the simple reason that so many great minds have wrestled with the problem over the years without achieving a resolution. I have to assume there’s a pretty good chance I haven’t either. On the other hand, bear in mind that hard determinism hasn’t carried the day with the majority of careful thinkers. It’s very much a minority view. All that said, let’s turn to substance.
First, you’re reading too much into “just a label.” In context, this was in response to your question of what I thought the concept added. I acknowledged, “of itself nothing. It’s just a label.” But, I continued in the very next sentence to explain, “Whether it’s a useful label depends on how one considers the observational evidence discussed in the OP.” As a matter of principle, I think it’s inappropriate to load one’s theory into the definitions, which is pretty much the objection you’re making to Dennett (with which I mostly agree). I was pointing out that I was doing no such thing.
Second, what it seems I’m not communicating effectively is that one of my core claims is that we should view the free will problem as an empirical question rather than an analytical one. That’s the point of my relying on emergence. Since descriptive emergence stipulates reduction in principle, it has no magical causal powers and I’m not asserting any. On the other hand, reduction in principle doesn’t generally give us a useful way to get at problems, whether they be in solid state physics, chemistry, biology or psychology. We figure those things out by observation. I’m saying we should use the same approach with the free will problem. Perhaps surprisingly, the Internet Encyclopedia of Philosophy article on Free Will says something similar. Section 3.b summarizes the current state of the debate as follows: “Most philosophers agree that whether or not determinism is true is a contingent matter; that is, determinism is neither necessarily true nor necessarily false. If this is so, then whether or not determinism is true becomes an empirical matter, to be discovered by investigating the way the world is, not through philosophical argumentation.”
Third, as for why you should care about volition, you captured it pretty well yourself yesterday. Volition (or free will) is a key concept upon which turns the issue of moral responsibility. Yes, one can justify punishment and social sanctions on utilitarian grounds without volition, but it’s hard to defend that position ethically. It’s all very well to say punishing misconduct will deter others, but what of the perp in the dock? He didn’t heed the sanction and, if determinism is true, couldn’t. Would you say it’s okay to punish the perp’s mother if that would deter misconduct by other perps? I assume not, even though arguably this would be very effective. Substituted punishment is frequently used in war for that very reason. So, we’re relying on more than utilitarian justification. We’re relying on the perp’s owning the misconduct. And have to, I think, to justify deterrence ethically. Straight determinism won’t get us there, which is why traditional compatibilists spend so much time trying to give deterministic free will morally relevant content. See the SEOP article on Compatibilism linked in the OP.
Fourth, and this is the problem which particularly interests me, there are cases such as senility, homelessness and mental illness where judging people by their actions is palpably unfair. IMHO, punishing such people in the interest of deterrence is just as wrong as punishing the perp’s mother. In real life, this is the side of the issue I find myself most often defending and wonder whether I’m not more of a determinist (in this sense) than are you. Abstract determinism is all very fine, but where do you come out on these difficult “edge cases”?
On preview, I see you’re about to leave on holiday. Have fun. If you have a chance to reply before you leave, so much the better for us.
That is absolutely not my claim. My claim is: if ALL (correct) necessary conditions are vacuous or tautological, then the event is uncaused. You think my claim is: if ANY (correct) explanation is vacuous or tautological, then the event is uncaused.
Let me amend this clam. I should say “Logically, this statement will be true for every possible event in a deterministic universe.” Now we’re back to tautology.
I’ll try to explain how I see the situation. If we assume that the ratio deaths/time is constant*, then changing the time will change the number of deaths. Similarly, if we assume that the ratio deaths/grains is constant*, then changing the number of grains will change the number of deaths. Or we could look at it from the other direction. If we assume that time is constant*, then changing the deaths/time ratio will change the number of deaths. If we assume that grains is constant*, then changing the deaths/grain ratio will change the number of deaths.
You probably think that discussing the number of grains of sand in this context is ridiculous. Of course it’s ridiculous! But if we appeal to it’s being ridiculous then we’re bringing the question of causation “to the level of human reasoning.” I’m trying to figure out how to avoid such a “massive level-confusion.” I’m trying to figure out why multiplying and dividing the number of deaths by “time” counts as an explanation/necessary cause, but multiplying and dividing by “grains” apparently doesn’t.
Given this, I really would like an answer to my last question… Suppose that I take all possible values of (deaths/grain) and (number of grains), and form a disjunction of them. By your own reasoning, haven’t I just created a necessary cause? What makes the disjunction of (deaths/time) and (time) to be more valid than (deaths/grain) and (grains)?
As far as I can see, the mathematics and logic used to generate a necessary cause with time is identical to that with grains.
