Hacker News Reader

Comment by naasking | Hacker News Reader

Yes, hard determinism is very problematic, and this thread covers some of the reasons. I'm not sure every conception of moral realism is incompatible with hard determinism, but I haven't given it much thought as I don't see much redeeming value in hard determinism.

As a hard determinist, you'll have to re-invent concepts of pseudo-moral blame in order to justify rehabilitation for those who break the law rather than simply changing the law itself to accomodate rule breakers, at which point you've just reinvented free will and called it something else.

Moral error theory has problems of its own, so I don't think these all hang together very nicely.

View on HN

First, to answer your definitions ask in the other thread.

"Practical" free will, as described by both of us:

> When people say that free will exists, they're saying that there is coherent notion of control over one's actions

> When humans say "free will", they are referring IMO in actuality to a real thing - a combination of many things, but basically agency + lack of prediction powers

You added a very specific phrase outside of the definition though, with no justification: "that grounds moral responsibility". That both implicitly implies moral responsibility, and also claims that it is grounded by this particular definition of free will. I'd be interested in how you got to those conclusions.

The "stricter/formal/whatever label you add here" version, that some try to argue for, including some compatibilists, goes more or less like this:

> Humans have the power to change the state/future of the universe and the universe is not "set" in any way

This feels a bit logically impossible to me, and it may also be something you don't agree with either, but it is still a definition that people use to mean free will, I have seen it way too often. However, the lack of this type of thing still has meaning is what I am trying to say. It is what leads to error theory for me, and then effectively what leads away from it. I wonder if that "away from" part might also solve some of your qualms with error theory.

-----------------------------------------------------------------------------------------------

In terms of hard determinism, it seems to me you only listed off reasons that hard determinism doesn't sound easy, but made no argument against it itself. Just because an idea invalidates many ones that were previously popular does not make it any less true/false, nor does it prevent it from establishing new and helpful ideas on top of it, even if they do not really exist currently.

> you'll have to re-invent concepts of pseudo-moral blame

This is exactly where I think hard determinism has its value. There is no moral blame - hard determinism requires a morality without it, and that's okay. It simply is new because people have focused so hard on blame, but I think that "formal" definition of free will is useful to show precisely that blame is not in play.

> at which point you've just reinvented free will and called it something else

I do think much of this will happen, but the distinction between the definitions has consequences. I already argued about for blame, but I think there is much more to explore there.

> Moral error theory has problems of its own, so I don't think these all hang together very nicely.

Again, I'd love to hear the argument rather than a simple refutation, truly, in a non attacking or confrontational way. At this point I think the HN format is hurting more than it's helping, so if you're interested in talking more feel free to shoot me an email, @gmail.com. You seem well versed in this area and I've love to hear more details on your perspective and reasoning/principles.

> You added a very specific phrase outside of the definition though, with no justification: "that grounds moral responsibility". That both implicitly implies moral responsibility, and also claims that it is grounded by this particular definition of free will. I'd be interested in how you got to those conclusions.

I always qualify my use of free will as being specific to the philosophical use in ethics because there's so much conflation around the term [1]. I hold that the term "free will" as used in various contexts are not necessarily the same, and that conflating them or trying to define a more general notion of free will that encompasses them all obscures more than it helps.

Per [2], I then conceive of free will in this context as "moral responsibility as accountability". I am not referring to free will as used in any other context, and I hope I've been strict about preserving that distinction.

> > Humans have the power to change the state/future of the universe and the universe is not "set" in any way

This definition compounds exactly the conflation that I think hurts progress in this debate. People often want more general abstractions to think about entire classes of problems rather than specific instances, but in this context, I think it obscures more than it clarifies.

> This feels a bit logically impossible to me, and it may also be something you don't agree with either, but it is still a definition that people use to mean free will, I have seen it way too often.

Compatibilism is necessarily compatible with determinism in which the universe is "set" in a specific way, that's the whole point. So I disagree with the definition. That said, just because it's "set" doesn't make it predictable, in which case it's outcome is not "set" in a way that can be known in advance.

So I would revise the clause "the universe is not 'set' in any way" to read, "the universe is not 'set' in a way that's relevant to responsibility", or something along those lines. But again, maybe you're thinking along more general lines that conflate the ethical properties with properties needed for other purposes (like an experimenters' freedom in science).

> Just because an idea invalidates many ones that were previously popular does not make it any less true/false, nor does it prevent it from establishing new and helpful ideas on top of it, even if they do not really exist currently.

I agree. I don't see anything in hard determinism which doesn't already have a satisfactory answer in Compatibilism though. The only quality which people seem to find appealing is a lack of blame, but as you later say, pseudo-blame will also likely be a feature of hard determinism in order to make it workable.

In the hundreds of debates I've had on this topic, pretty much the only reason people have used to justify hard determinism is that they don't think punishment is justified, and they see eliminating blame as a way to avoid punishment. This conflates justice with the simpler question of moral responsibility, and as a result, they erroneously believe that free will and Compatibilism necessarily entail punishment.

> Again, I'd love to hear the argument [refuting error theory] rather than a simple refutation, truly, in a non attacking or confrontational way.

Here's a rough summary of my thoughts when I first read it:

1. The Argument from Relativity: we're apparently not troubled at all by disagreement over natural facts (see: flat earthers), but disagreement over moral facts is somehow more problematic? We distinguish "fact" from "belief" for exactly this reason. I don't see any motivation for error theory here.

2. The Argument from Metaphysical Queerness, aka, moral properties are "qualities or relations of a very strange sort, utterly different from anything else in the universe", aka are moral prescriptions categorical or hypothetical imperatives: this appears to be the most subtle and misunderstood part of error theory, so maybe I'll delve in deeper at some point.

3. The Argument from Epistemic Queerness, aka, we would need "some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else": the epistemic "strangeness" of moral properties partly follows from the strangeness of qualia. For instance, we have moral prescriptions around causing pain. I'm an eliminativist about qualia, so I think the apparent strangeness are cognitive tricks that we will unpack in time. The only other faculty needed is reason.

Bambrough addressed most of these [3] almost a decade before Mackie's paper on error theory. Bambrough's working example of a child undergoing surgery is precisely an example of a logical argument depending on qualia (epistemic queerness), to establish a moral fact. I'm sure Mackie would find that result very queer indeed!

Which isn't to say that error theory wasn't a valuable contribution, because it's absolutely true that most alleged moral prescriptions have been false. I simply don't think there's any conclusive reason to think they're all false, and per Bambrough, I think there are reasons to think we already know some moral facts.

[1] https://news.ycombinator.com/item?id=23480626

[2] https://plato.stanford.edu/entries/freewill/#FreeWillMoraRes...

[3] A Proof of the Objectivity of Morals - Bambrough (1969), https://www.reddit.com/r/philosophy/comments/3etl9b/a_proof_...

A think we're getting to a stuck point here, but to wrap up some of my main qualms:

1. I feel like we keep talking in circles on the definitional aspect. I'm with you on the many definitions causing confusion and I would love to see expanded vocabulary around this. But you haven't said anything to convince me against the idea that all three have moral relevancy in various ways.

2. To be clear, I do not believe hard determinism needs any sort of pseudo-blame to be workable, at least how I understand you are using that term. The only thing needed is causality in order in order to find the ideal "organization" of the world to optimize for whatever factors are settled on by other parts of the argument.

3. I actually agree with you on the lack of convincing of 1 and 3 for error theory (slightly different reasons but details not needed), but the argument from queerness seems to fit quite well. I'm not sure what part you're claiming is misunderstood.

4. The mentioned "justifications" on hard determinism are going the wrong way logically - you don't decide if something is true by if you need it to make morals that look nice, you follow what is true and see the results, and only then can you "check your work" with a sanity test. You've seem to cut off your thoughts on determinism before ever getting to a point where it actually produced a full set of moral ideas to check in favor of compatabilism simply because it produced something that passed said check, but if the underlying principle is false that's simply a coincidental state. Obviously you've also personally eliminated error theory, which is the core of our disagreement.

As to justice, I fully agree on that being way oversimplified. For context, looking ahead, my theories on morality out of error theory basically lead to a consequentialist version of Rawls, so that's about where I stand re justice, with some nuance.

> I think there are reasons to think we already know some moral facts.

Can you give examples? Single or multiple, whatever you think is more useful in this context. I don't see how we get to moral facts without them being tied back subjectively to humans, which either invalidates the objectivism tract (now we're back at arguing metaethics fully) or you'd have to get there from the error theory route I've been talking about.

To Bambrough's child surgery example, I think that only shows that that moral fact is true only subjectively to humans - it's not an "objective" moral truth. It is derived from the traits of a human, and the example generally works because it takes a clear cut case we can all agree on.

All of this reasoning is still useful to an error theorist, but it is laid on the groundwork that morality is subjective to the state of humans/beings. If you mean to say that this is objective, then there is basically infinite moral facts in existence, we just debate them because humans and society are complex. That is to me the precise meaning of that "optimal organization" I reference, which is essentially that collection of all true moral facts, which I would argue are always subjective to the deterministic state of the world. Change anything about the world, and you likely change ethics itself. The argument from queerness highlights that without our perspective and values, objective morality wouldn't exist at all.

I suspect definition will be the issue here again - are "moral facts" derived from the combination of human traits and our perspective considered objective or subjective? To me, that is subjective and then the combinations of subjectiveness along with the lack of inherent meaning or value produces a single objective level practical morality. I can understand the argument to classify that set of things itself as objective morality, but I think that weakens the base of the argument needlessly.

> But you haven't said anything to convince me against the idea that all three have moral relevancy in various ways.

Because I think only one has moral relevancy, not all three. I've been saying it's a mistake to conflate the others with the one specific to moral responsibility.

> The only thing needed [in hard determinism] is causality in order to find the ideal "organization" of the world to optimize for whatever factors are settled on by other parts of the argument.

The "ideal organization" necessarily designate faulty parts which don't conform to the ideal, and so the parts that require rehabilitation. That's blame however you want to spin it.

> To Bambrough's child surgery example, I think that only shows that that moral fact is true only subjectively to humans - it's not an "objective" moral truth. It is derived from the traits of a human, and the example generally works because it takes a clear cut case we can all agree on.

I think you're missing the forest for the trees by mistaking an existential claim for a universal claim. Suppose we swap out the human child with an alien that doesn't feel pain as we do, but has an adverse reaction to some other stimulus. The same argument applies by parity of reasoning. You're focused on the specifics of the human case, but the moral fact in that case is just an instance of a more general principle from which this fact derives.

It's like I'm pointing out 1+2=3, that's a fact, an instance of the general rule of addition. You then agree that 1+2=3, but gotcha, 1+3!=3, therefore 1+2=3 is not a general fact applicable to all numbers. No argument there, but it was never suggested as such.

> I suspect definition will be the issue here again - are "moral facts" derived from the combination of human traits and our perspective considered objective or subjective?

"Objective" means "mind independent", which is to say, that any form of intelligence will reach the same conclusions given the same understanding.

Any intelligence will reach the same conclusion that a human child undergoing a painful surgery ought to be given anesthetic, all else being equal, assuming they fully understand the meaning of every word in that sentence.

> Because I think only one has moral relevancy, not all three. I've been saying it's a mistake to conflate the others with the one specific to moral responsibility.

And that's exactly where we're stuck, which is all I'm saying. I think all three affect moral responsibility.

> 1+2=3, etc

> Any intelligence will reach the same conclusion that a human child undergoing a painful surgery ought to be given anesthetic, all else being equal, assuming they fully understand the meaning of every word in that sentence.

There's another assumed aspect though that you have not separated, which is that the judging intelligence has empathy if it agrees. You're claiming only using pure logic, but you're using aspects of that subjective human perspective to imply that the reduction of pain/whatever stimuli. Without the human perspective, the example fails to be a "moral" fact. It's just a configuration that raises x, lowers some y maybe, and the beings prefer more x and less y. But without value being projected on x and y, it is not inherently "good". We require the human perspective to create human morality. What if the intelligence viewing this entire scenario loves to see pain/y? Would it not rationally conclude the anesthesia is bad? Only if it assumes or judges based on the values of a human does it get the same conclusion. Yes, within that, it's objective.

I find that distinction is key in how we approach the derivation later. A lack of such distinction can, for one example, enable using the idea of objectivity here to assign value to the entities (more ideal = more value), which then can shift practical effects possibly. And maybe you disagree with this specific example on deriving value from objective nature, but that is one of the many doors the lack of distinction opens. I'm sure if we went far enough down the road, there is likely an argument that you do agree with that comes from the lack of distinction.

You say I'm missing the forest, but I think you are minimizing the importance of the trees through all of this and disagreeing that the trees can be important, even if you zoom out and the forest is the same. You're saying that the rule in play here is "addition", but I'm saying I think you're missing some rules/definition/workings of the "operator" in this metaphor that lead to different results later down the line, even if it works in this case.

> The "ideal organization" necessarily designate faulty parts which don't conform to the ideal, and so the parts that require rehabilitation. That's blame however you want to spin it.

I typed up a longer distinction here, but again this seems to come down to distinguishing between two definitions and if that distinction means something. See importance of trees. In this case, blame being the tree.

Reply on news.ycombinator.com