Blog Coordinator

« What It's Like to Be a Mele | Main | Call for Papers: Oxford Studies in Experimental Philosophy »

07/05/2015

Comments

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

Helen,

I agree with you that the entailment definition of determinism is inadequate – it’s missing something. I’m thinking that the metaphysical influence you’ve mentioned is caused by new life. Said life exerts new forces as an emergent property, and those forces add together with other preexisting forces thereby helping to determine the path forward.

When compatibilists and incompatibilists adopt a definition of “determinism” which includes the concept that new forces are exerted by life and those forces affect the path forward, I believe both camps will converge on the truth.

The laws that science has deductively come up with over a long period of time don’t exist independently on their own merit. Instead, the laws are simply ideas in human minds that model how reality works. Yes, there’s a lot of consistency in actions at many different system levels, thereby resulting in predictability to a certain extent (i.e., so the laws have value), but there’s also new life that emerges at many different system levels, and said life is fundamentally indeterministic in nature, thereby making the overall path forward indeterministic (free) in nature.

Thanks for the interesting post, Helen!

I have sympathy with your view but I'm still on the Vihvelin side of the debate, which is partly motivated by van Inwagen's (1983) observation that to define “determinism” in terms of causality is to define the obscure in terms of the truly incomprehensible (or words to that effect). In any event, here are some of the worries I have.

1/ I'm not at all sure what (MD) means. I know what the kind of necessity involving entailment means, so I understand (ED), but I'm not sure what a non-logical necessity is apart from reducing it to some kind of logical necessity.

2a/ Related to the first worry is a set of concerns about how to frame the problem of free will and determinism since if you can't adequately motivate that problem than that is a strike against (MD). But why think that (MD) is even provisionally incompatible with free will? Historically folks have used something like the consequence argument, which usually requires grounding principles (like the fixity of the past and the laws) together with transfer principles. I can understand this as propositions or relations between propositions but it is not clear how to understand these parts of the consequence arguments in terms of events or whatever corresponds to the items of (MD).

2b/ Of course, Fischer believes that there are non-transfer versions of the consequence argument that utilize an extension principle: “an agent can do X only if his doing X can be an extension of the actual past, holding the laws fixed” (1994, 88). This creates another problem though since this definition looks an awful lot like what I imagine a fleshed out version of (MD) would look like. Thus, on this approach the consequence argument appears to be question begging, which is worse than the result of (2a). According to (2a) we can’t formulate the incompatibilist’s worry; according to (2b), we can’t avoid it.

3/ This last worry is an appeal to authority. “Determinism” is a concept in physics. But everyone that I know who works in philosophy of physics accepts (ED) over (MD). See Hoffer’s “(Causal) Determinism” (Stanford EP) plus his references.

I’m sure most folks don’t have these worries but it might help get the conversation going!

Hi Helen,

Thanks for this interesting and thoughtful post, and question. A few brief remarks. I might have used the term "definition" (I'm sure I'm guilty of this in at least some contexts), but when I'm being more careful, my view is that it is quite difficult to "define" causal determinism (see Earman's book and Carl Hoefer's work in the Stanford Encyclopedia of Philosophy and elsewhere), and ED is not supposed to be a full or complete definition. My view is that whatever the proper definition of determinism is, it implies ED. And ED is is all that we need to run the Consequence Argument. Actually, I think I borrowed ED from Peter van Inwagen, if I'm not mistaken.

But perhaps someone will worry that even though ED is all that's required for the CA, the "metaphysical" part might be more relevant to Sourcehood worries. I'm still not convinced by Source Incompatibilism, but it is a view worth taking very seriously.

Thanks again.

It sure does! Thanks for the post, Helen. I am looking forward to your posts this month.

A quick question: have you come across arguments against (MD)? I know you mentioned Vihvelin finds this way of thinking "confusing and misleading" but has she, or anyone else for that matter, given an argument for why (MD) is not warranted, or, why (MD) is not to be preferred over (ED)?

Justin,

I'm not sure if this helps, but it's from the SEP link Joe referenced:

"For a variety of reasons this approach is fraught with problems, and the reasons explain why philosophers of science mostly prefer to drop the word “causal” from their discussions of determinism. Generally, as John Earman quipped (1986), to go this route is to “… seek to explain a vague concept—determinism—in terms of a truly obscure one—causation.” More specifically, neither philosophers' nor laymen's conceptions of events have any correlate in any modern physical theory.[1] The same goes for the notions of cause and sufficient cause. A further problem is posed by the fact that, as is now widely recognized, a set of events {A, B, C …} can only be genuinely sufficient to produce an effect-event if the set includes an open-ended ceteris paribus clause excluding the presence of potential disruptors that could intervene to prevent E. For example, the start of a football game on TV on a normal Saturday afternoon may be sufficient ceteris paribus to launch Ted toward the fridge to grab a beer; but not if a million-ton asteroid is approaching his house at .75c from a few thousand miles away, nor if the phone is about to ring with news of a tragic nature, …, and so on. Bertrand Russell famously argued against the notion of cause along these lines (and others) in 1912, and the situation has not changed. By trying to define causal determination in terms of a set of prior sufficient conditions, we inevitably fall into the mess of an open-ended list of negative conditions required to achieve the desired sufficiency."

http://plato.stanford.edu/entries/determinism-causal/

Hi Helen--great post. I love this kind of conceptual challenge.

Everyone here has said something of interest; I hope I can contribute something too.

"Determinism" is only a belief moniker and consistent with all sorts of uses both general and specific. The most familiar one I'd think is generalized as with ED--involving the universe as a whole and as time-sliced to produce the requisite deductions or at least a large system of events so sliced. I suppose those of us in FW/MR tend to think this way about determinism. But there isn't anything inconsistent with calling a specific relation between two events "deterministic" either. However, that last particular usage is inconsistent with ED as stated--the particular relation is the issue, whether as deduced or whatever, and there are no implications for subsequent events (in terms of true propositions as well) in this usage. So some uses of the term don't comport with the stated definition ED, and these particular uses seem connected with something like "caused". This is just a terminological point, but worth noting.

I think in fact (a useful form of) determinism can be defined as a generalization of instances of causation, and in ways that handle some of the concerns Izzy cites from Joe's reference. I won't bore people now with it, but I will say it might be able to handle all those negative conditions that actually are considered by physicists as ceteris paribus standard conditions held (hopefully) epistemically (and maybe metaphysically) fixed. In experimental set-ups you control all the standard conditions you can, and cross your fingers for the uncontrollable ones. If the conditions hold, then experimental results--even particular causes--are statable as observed. Statistical analyses stray from causal certainty as those conditions meander from epistemic norms--studies of billard-ball collisions in a lab versus anti-bacterial results in petri dishes to smoking and cancer to inflation and wages to. . .

I'll weigh in later after I see more responses to Helen's post.

Thanks Izzy for correcting my paraphrase against causal views of determinism, which was from Earman and not van Inwagen (I likely got it from the above passage in fact). Here is another point relevant to this issue.

If we accept (ED) as our definition of "determinism," then we have a very nice program. We have a way to identify metaphysical issues in terms of logical issues, since we can transfer talk of free will and determinism into talk of propositions. We have a way of breaking down the consequence argument into identifiable parts with which philosophers might agree or disagree (depending on which version we endorse): fixity of the past, fixity of the laws, transfer principle, extension principle, etc. We can then clarify the debate more fully and see what's what. So there is something we loose if we switch from (MD) to (ED), as I see it. I want to know whether we can gain these back in a new way once we endorse (MD) before I'm willing to make the switch.

John is right that van Inwagen (1983) endorses (ED). And it would be interesting to pursue John's thought about whether (MD) might (better) support source incompatibilist worries.

Helen,

Thanks for the post. I am really looking forward to your blogging this month.

I am inclined to think there is no one correct way to define determinism for those interested in the free will debate since those interested in the free will debate are interested in it for different reasons. Some may be interested in the debate because they worry that certain physical theories are incompatible with free will. In that case, the right way to define determinism will be in the way those theories define it. Others might be interested in whether certain conceptions of determinism that might be implied or suggested by certain physical theories are compatible with determinism. In this case, these folks should just tell us as clearly as they can what they mean by determinism. Others might be interested in a kind of determinism that is continuous between physical and theological species. This might lead to yet a different definition of determinism. I suspect that there are more options than this.

So two things seem important to me. First, when presenting an argument that determinism is (in)compatible with free will one should be clear about the kind of determinism one is interested in. Second, when responding to others argument one better be sure that one is operating with the same conception of determinism as those one is responding to, otherwise we are likely to end up talking past each other.

Thanks, everyone, for very interesting and astute responses. There's a lot here. Let me respond to one particular theme that seem to be emerging, and provide some contextual background. I'll try to pick up some of the other themes in the days ahead.

I think it's probably true (as Joe suggests) that many have wanted to eschew a definition of determinism along the lines of (MD) because of general worries about the unclarity or obscurity of the notion of causal necessitation. It is an interesting question when precisely the doctrine of determinism understood by way of the entailment definition began to be substituted for the more metaphysical-sounding thesis that used to be called the ‘doctrine of necessity’, and why. Although Hume’s views on the absence of any proper idea of causal necessity must be a very important part of the story that needs to be told, the full adoption of the entailment definition of determinism does not seem to have occurred until much later, after various developments in logic and mathematics encouraged philosophers to begin to appreciate the usefulness of the concept of a function. I suspect Russell was very likely a very powerful voice in the shift made by philosophers to definitions of determinism resembling (ED). According to his highly influential paper, ‘On the Notion of Cause’, Russell claims that “A system is said to be ‘deterministic’ when, given certain data, e1, e2, …, en, at times t1, t2, …, tn respectively, concerning this system, if Et is the state of the system at any time t, there is a functional relation of the form
Et = f(e1, t1, e2, t2, …, en, tn, t).”
On this conception of determinism, determinism holds wherever the way things are at a given time is a function of the way they are at another time – a function that will be encoded by some relevant law of nature. But there need be no implication at all that the law constrains anything; for all Russell’s definition of determinism implies, the law which encodes the functional relation may just be a Ramsey-Lewis law. The functional relation is entirely descriptive, and merely expresses the fact that where such a relation exists, one can in principle infer the state of the world at one moment in time from the state of the world at another. Moreover, the relationship between past and future states is completely symmetrical – one can infer past states from future ones, as well as the other way around. Determinism thus ceases to be merely a doctrine about the development of the world through time – and becomes instead a doctrine about the relationship between descriptions of the world at different times, a relationship which has no particular implications pertaining to constraints on the development of reality.

What motivated Russell to formulate the concept of determinism in this way? Earlier in the same paper, Russell expresses his suspicion of the idea that causes compel their effects – compulsion, he insists, is “a complex notion involving thwarted desire” – a concept of psychology and not of metaphysics. One can see in this critique the influence of the Humean idea that we have no clear idea at all of necessitation, and that in so far as the idea has a source in our experience, that source is in the mind and not in external objects. Empiricism had no room for the unverifiable existence of necessary connections between states of affairs, and the concept of determinism developed by adherents of empiricism unsurprisingly followed suit. The entailment definition is the product of a long and important period of philosophical history in which realistic ideas about powers and causes were tested against an empiricism that refused to countenance them, and a mathematicization of parts of metaphysics promised new ways of thinking about the concept of law. The doctrine of necessity was thought to be dead owing to a bad case of unintelligibility: long live the doctrine of determinism, understood according to the entailment definition!

However, unless I am much mistaken, I think it is fair to say that philosophy has now entered a new phase of development, in which that empiricism itself is on the back foot. Those who believe in the reality of such things as powers and natures and objective laws which constrain reality have been once again admitted into the fold of philosophical respectability. Of course, it cannot yet be said that they have won the day. There are still those who do not think that the idea that laws truly constrain the development of reality can ultimately be made sense of; regularity theories of causation and Ramsey-Lewis theories of law remain quite widely held. But there are now many adherents of more realistic conceptions of physical and causal necessitation. We cannot, then, simply assume as perhaps Russell once did, that determinism as defined by MD, is the mere product of silly confusions of psychological concepts with metaphysical ones. There are plenty of philosophers these days who are absolutely happy to sign up to what Galen Strawson calls ‘Producing Causation Realism’, the doctrine that causation is something that exists in the world, either because there exist constant objective forces, such as the fundamental forces postulated by physics, or because there are intrinsic natures which dictate how objects and substances will behave with respect to one another. But the discussion of determinism in philosophy has simply not caught up with these changes in the Zeitgeist. If it is respectable once again to believe in real, producing causation and objective, metaphysical laws, it should be respectable once again to formulate determinism in ways closer to the ways in which the doctrine of ‘necessity’ used to be formulated – as a firmly metaphysical thesis about the necessitation of the future by the past. (MD), then, should be in contention once again.

Great to see you at Flickers, Helen! I find the ED view a bit baffling myself. Van Inwagen says he uses it in part to avoid stepping into the 'morass of causation', but I am not sure how the concepts used in ED don't drag us into stickier swamps. If we're going to use complete statements of the entire state of the universe at a given time (an idea I'm not sure relativistic physics can make sense of) and of all the laws of nature (or is it just physics?), then what's wrong with using an unanalyzed notion of causation and saying something about what a causally deterministic system (or universe) is:

(CD) for each event E, given the laws of nature, some set of events prior to E are such that these events cause E to occur with probability 1.
[causal (law-governed) indeterminism can then be defined precisely and contrastively in terms of probabilities less than 1.]

CD is similar to your MD but (correct me if I'm wrong) MD adopts a necessitarian view of laws and a 'Producing Causal Realism.' CD is consistent with less 'pushy' views of causation (which I like, since I think interventionism is the best theory of causation and it remains neutral on that issue).

I also think that, to the extent that ordinary people are concerned with some concept of determinism, and when scientists talk about determinism threatening free will, they clearly seem concerned about a causal notion like MD (or CD) and not ED. But the scientists (and some folk) often seem to have in mind something like universal causation governed by the laws of physics (whether deterministic or indeterministic). Sometimes the scientists talk about all of our actions being determined by the laws of physics.

Thanks, Helen, for your historical account of ED, which strikes me as correct. Eddy's definition seems promising. Very interesting discussion so far!

One thing I'd add Helen to your historical analysis is so-called "Laplace's Demon" and its predecessors, which pushes the lineage back from Russell some 100 or more years. One recent resource for this:

http://www.sciencedirect.com/science/article/pii/S0039368113001325

Abstract:

"In this paper I examine the foundations of Laplace’s famous statement of determinism in 1814, and argue that rather than derived from his mechanics, this statement is based on general philosophical principles, namely the principle of sufficient reason and the law of continuity. It is usually supposed that Laplace’s statement is based on the fact that each system in classical mechanics has an equation of motion which has a unique solution. But Laplace never proved this result, and in fact he could not have proven it, since it depends on a theorem about uniqueness of solutions to differential equations that was only developed later on. I show that the idea that is at the basis of Laplace’s determinism was in fact widespread in enlightenment France, and is ultimately based on a re-interpretation of Leibnizian metaphysics, specifically the principle of sufficient reason and the law of continuity. Since the law of continuity also lies at the basis of the application of differential calculus in physics, one can say that Laplace’s determinism and the idea that systems in physics can be described by differential equations with unique solutions have a common foundation."

I'm a bit skeptical that "the law of continuity" here is well-framed as literally a law, but clearly the foundation for ED has historical roots probably hundreds of years old as an heir of Newton, Leibniz, and others. I take ED as employed by the CA as a "logicized" version in this tradition that is useful in examining broad principles like incompatiblism, transfer principles, and the like.

However, all of this rich history certainly doesn't entail that ED is either a necessary or sufficient component of an apt ascription of "determinism" to some system of events. And I do not think it is either of those. I think that "determinism" is a term that generalizes prototypical specific types of rigid causation, and those types are empirically specifiable with explicit ceteris paribus conditions (even if they constitute a complex open-ended set). My sense of causation is thus not far from what Eddy describes above with CD.

I wanted to let people know that I have a workshop today and tomorrow and then the Joint Session Fri-Sun, so my opportunities for posting on the blog may be a bit thin on the ground. Apologies if I don't put anything much up over the next few days - I'll make up for it next week.

I like Eddy's (CD). And I also agree with Eddy that even if (MD)can be accused of reliance on unclear notions, the very same is certainly true of (ED). AS Eddy suggests, the idea that a ‘complete statement’ of the nonrelational facts about a time could ever be given, even in principle, seems deeply problematic. And which, exactly, are the ‘nonrelational’ facts, and do we know how to distinguish them properly from the relational ones? I am very sceptical that these questions can be given acceptable answers. It may be, in fact, that determinism (however formulated) is basically an incoherent doctrine - I have often been tempted by that view. But of course its incoherence mightn't stop it from bothering people!

I also agree with Eddy that the 'folk' worry about determinism is a worldly one - a worry about causation, or the settling of the future by the past, not a worry about entailment. And this actually turns out to be very important, I think. Chris has to be right that one might be worried about determinism for different reasons - and that one might be worried about it even in connection with free will for different reasons - and that one's definition should be sensitive to the root source of the worry. But that is in fact precisely where I am coming from. My worry about (ED) is that it just doesn't get at what bothers most people about determinism - as I think is pretty much shown by the consistency of (ED) with Ramsey-Lewis style deterministic laws. When I ask myself whether I am worried about determinism, given a Ramsey-Lewis conception of law, my answer is simply a resounding 'no'. No one should be worried about determinism if *that's* all it amounts to! If the laws are just Ramsey-Lewis laws and something like a Humean regularity theory of causation is true, there's just no need to worry! Call me a compatibilist, then, if you like - I *am* one, I think, if (ED) is the definition! But it seems to me that this result should rather make us take another look at (ED). Surely what we ought to be concluding is that it just doesn't capture precisely enough what worries people about determinism? As John suggests, (MD) probably entails some thesis like (ED) - but the crucial question, in my view, is whether (ED) entails (MD) - and I don't think it does. For (ED) is compatible with views of the Universe (what one might call 'non-necessitarian' ones) that (MD) excludes; but it's precisely those necessitarian views that worry people!
Now of course it could be (as I think maybe Joe is suggesting) that necessitarianism is nonsense. Perhaps it depends on notions of causation that cannot be upheld in the face of modern physics, or perhaps cannot be upheld in the face of philosophical arguments (such as Hume's, for instance). But that wouldn't stop it from being the source of the worry about determinism. And so shouldn't our response be at least to *try* to formulate properly what it is that people feel worried about using notions like causal necessity - and then to show how the central notions are to be undermined?

Thanks for the great post Helen! I am really enjoying this discussion. And though I know you’ll be away for a bit, I did want to throw in my two cents nonetheless.

I just want to reiterate something that Helen says that I think is getting a bit lost. Helen says “When I ask myself whether I am worried about determinism, given a Ramsey-Lewis conception of law, my answer is simply a resounding 'no'. No one should be worried about determinism if *that's* all it amounts to!”

I agree with Helen here. To make her point in a different way, it is useful to notice that ED by itself can’t generate certain classic problems for free will that are the bread and butter of incompatibilist arguments. Earlier John says “ED is all that we need to run the Consequence Argument.” I don’t think this is correct. The CA (all versions of it as far as I can tell) relies on the crucial premise that it is not up to us what the laws are (i.e., the Fixity of the Laws premise). But as Lewis correctly points out in “Are we free to break the laws?”, on a Humean conception of laws, that something is a law is after all -- in a certain *very specific sense* -- indeed up to us. So if Humeanism is true, then even given ED, we don’t get the incompatibilist conclusion of the CA.

The same point applies to Eddy’s CD. Eddy says it is supposed to be genuinely neutral between Humean causation (i.e., causation that is grounded in a Humean conception of the laws in the way that Lewis envisions) versus more pushy causation. If that is in fact the case, then Eddy’s CD too can’t generate classic incompatibilist worries about free will. For example, you can’t use the fact that CD implies ED and then get the incompatibilist conclusion of the CA for just the reason I said in the previous paragraph (if Humeanism is true, then even given ED, we don’t get the incompatibilist conclusion of the CA).

So that is all another way of saying Helen is right that if we are Humeans about laws, we should not worry so much about determinism.

Helen,

Thanks for this great thread. You mention that "the relationship between past and future states is completely symmetrical – one can infer past states from future ones, as well as the other way around." Well, you just let my hobby horse out of the barn.

This is indeed a feature of ED, but I don't think it comes from ED itself. It comes from the laws that science has actually discovered. In fact the laws meet the even stronger condition of CPT-invariance - meaning that if we invert charge, parity, and time, we get a perfect mirror-image of the laws-as-we-use them, governing backward-in-time processes.

Even if one thinks, as I do, that laws don't "just happen to be true" but reflect the modal properties of the things they describe, I see no reason to think the time-reversibility goes away. Science is our best way of getting at the modal properties, and CPT-invariance is a big fat clue that we can't ignore.

Furthermore, thermodynamics can explain (here, explain *away*) our intuition that the present does not affect the past. It does so by showing that the macroscopic processes that inform our intuitions are "irreversible", meaning overwhelmingly more likely to go one way, than to go back. We overgeneralize - which for all *practical* purposes is innocent, but for some *philosophical* purposes is a fatal mistake. The macroscopic features of the past are immune to present action - this is the truth behind our intuitions - but the microscopic details are variable, as implied by the fundamental, CPT-reversible laws. The past is not fixed. It's broken - and so is the Consequence Argument. Because the CA needs more than just that the macroscopic past is fixed. The macroscopic past is not a sufficient cause of your present action.

Suppose we built into our definition of determinism a governing conception of the laws. Then we'd get a more deeply metaphysical notion of determinism, while also avoiding the sorts of difficulties thought to afflict the likes of MD. This conveniently bypasses Humean responses to the consequence argument, since governing laws aren't up to us. What's not to like? No, seriously, this seems too easy. What's wrong with this maneuver?

Chandra,

Thanks for the clarification. I didn't mean to say that there are no other elements of the CA! That is, I know that the CA needs fixity of the past and fixity of the laws premises, and perhaps a transfer principle (although that is disputed). All I meant was that ED is the definition, or implication, of determinism that PvI believes is all we need to generate the CA, when combined with the fixity premises. That is, we don't need anything further built into the definition. Of course, I suppose the fixity assumptions could be put more explicitly into the definition, but PvI didn't go that way, and I don't why we would need to.

Thanks again.

Justin: I think you are correct. On the other hand, when you put it that way, it might seem like you are changing the definition of “determinism” just to avoid a criticism of compatibilism: If we define determinism as ‘X,’ then compatibilist response Y is no longer an issue. Or do you see it more as getting to the real heart of the incompatibilism issue? If the latter, how so?

Chandra: I agree that given the new (and old) way of framing the debate the Humean is looking good. But what about this argument for incompatibilism?

1/ Determinism is the thesis that for each event E, given the laws of nature, some set of events prior to E are such that these events cause E to occur with probability 1 (Eddy’s MD).

2/ If an event E is caused to occur with probability 1, then E is not a free action (cf. John’s extension principle).

3/ Thus, given determinism, there are no free actions.

Oops, (3) should read:

(3) Thus, there are no free actions, given determinism.

Joe: on a non-governing view of the laws, there's a sense in which the laws depend on what I do. But what seems troublesome about the (inchoate) idea of determinism is that which actions I'll perform is (in some sense) settled or fixed in advance by things over which I have no control, things that don't depend (in the relevant sense) on which actions I perform. Thus, determinism, given a non-governing view, turns out not to be the sort of thing that is troublesome for free action.

I agree with Helen that if all determinism amounts to is ED combined with a non-governing view of the laws, then call me a compatibilist. Indeed, I can't think of anyone who accepts a non-governing view of the laws who is an incompatiblist. Why would you be?

Joe,

Could this be a reasonable way of looking at things?

Given *predeterminism* is true, there are no free actions. (i.e., If the fundamental forces of physics are solely responsible for controlling everything in a bottom-up manner, then there are no free actions.)

Given *determinism* is true, there are free actions. (i.e., If life exerts new emergent forces that affect the path forward, then there are free actions.)

The Hempelian covering-law concept is in keeping with ED in terms of the logical symmetry of prediction and retrodiction. Salmon famously explanatorily reframed causation in more asymmetric empirical terms that tracked earlier to later events as what counts. I'd think that what Helen is after is more like baked Salmon than cold-cut Hempel in terms of explanation.

But PvI's CA wants the best of both worlds: preserve the abstract account of laws as Hempel invoked, but presume entailment as based on a Salmonic asymmetry of cause and effect to produce incompatibilism as a future-forward thesis of what we cannot do given what is done. Is this the sole role of the fixity of the past, or do transfer principles assume a unidirectional temporal asymmetry of explanation, or does the logical property of entailment do that job given the fixity of the past?

@John, thanks, that is a helpful clarification.

@Joe, the problem in your argument seems to be premise 2. My understanding (and I am not a metaphysician, so what I say here could be wrong) is that Humean-types like Lewis have both a non-governing conception of laws as well as a non-governing view of causation. Laws just describe and summarize events -- they don’t restrict or constrain. So too with causation. Causation is analyzed in terms of counterfactuals, and counterfactuals are analyzed in terms of certain sorts of descriptive similarity relations between possible worlds. It is hard to see how this descriptive non-oomphy understanding of causation could justify premise 2.

Thanks for the replies! This is a helpful discussion.

For one thing, it helps me to better understand why Humeanists about laws who are also compatibilists -- like Lewis and Hume -- don't always resort to their Humeanism when responding to arguments for incompatibilism. It also helps me to see why folks regard Kant as an incompatibilist, for his view of determinism rests on a "non-oomphy understanding of causation" and laws, as well.

But what is it about deterministic "oomphy" causation that makes it a threat to free will? Is it the mere deterministic "oomphiness" of the cause? That is doubtful (think of tracing issues, drunk-driver cases, or my Analysis criticisms of CA).

James suggests and Justin explicitly states that the threat from determinism is that things are "settled or fixed in advance by things over which I have no control." So it isn't just the deterministic "oomphy" causes; it is that together with the fact that the initial causes of our actions were in play prior to our existence, prior to any potential causal control on our part (Sartorio and Pereboom, among others, argue for this view).

What do you all think?

Alan,

Consider Alexander Pruss's Beta-2 from http://philosophycommons.typepad.com/flickers_of_freedom/2014/06/the-consequence-argument-exciting-news.html :

Beta-2: For any propositions p and q: Np, p entails q ⊢ Nq.

Where Nα means α & There is nothing that anyone can do, such that had she/he done it, it would falsify α. And p is going to stand for some conjunction about past and laws.

But let r stand for some proposition (e.g. about the future) and let Sr mean r & there is something someone can do, such that if she did it, it would falsify r. Then Beta-2 implies

Gamma: Sr, r entails p ⊢ Sp.

More precisely, let r stand for a conjunction of a statement about the future, plus the laws. Given the actual, bidirectionally deterministic laws that science is considering, r really does entail p! (Recall that p is a description of a past time slice, plus the laws.) And now we can play burden tennis. Which way should we infer: modus ponens Beta-2 and modus tollens Gamma, or ponens Gamma and tollens Beta-2? For a hint, see J. Ismael's http://mediasite.capd.fsu.edu/Mediasite/Play/4936b3a64b754898aad036545e9dc0ca1d (00:55:00+)

So, to answer your question: the asymmetry comes from the mention of the forward entailment while neglecting the backward in time entailment. Telling half-truths is a good way to create apparent asymmetries.

The comments to this entry are closed.

Books about Agency


3QD Prize 2014: Marcus Arvan