[atq2119]

As usual, these kinds of discussions are really about what we mean by words.

What does "free will" mean? I think technically inclined people like us are easily led down the path you describe, but over the years I've become convinced that the kind of definition that leads you to "free will = random choice" is pretty pointless because it's so vacuous.

I like to take a step back and ask: why are we interested in free will in the first place?

In practice, what tends to be really relevant are questions like: are criminals responsible for their actions?

Free will is better defined as a shorthand for these kinds of questions. And in that context, it seems clear to me that even if our actions are purely deterministic, we can still have free will. What matters isn't whether our will is free of physics or determinism; what matters is whether or to what extent our will is free from the influence of other people and society.

Once you look at it this way it's also clear that there are no absolutes. We all influence each other, which puts limits on our free will, but we aren't just puppets, either.

[stareatgoats]

Agree it's about word meanings, and superimposing different meanings of the same word on each other (often in an attempt to close any cracks in our worldview that might allow religious concepts to take hold). That's probably why "free will" discussions are such a mess.

> In practice, what tends to be really relevant are questions like: are criminals responsible for their actions?

Exactly what I've landed in too, i.e. does free will have any real every-day meaning. The answer must be an unequivocal "yes", otherwise questions like these would be meaningless (which they are not):

- "did they force you to take that apple or did you do it of your own free will?"

- "was the sex voluntary or were you forced?".

- etc

The problematic word/concept here is really "force" and the different meanings of that word. The force of causality is fundamentally different from the force of coercion, and they can not be used interchangeably.

And so the question of "free will" should really be disentangled from physics, once and for all.

[naasking]

> And so the question of "free will" should really be disentangled from physics, once and for all.

Welcome to Compatibilism!

[stareatgoats]

Yeah, maybe. Although (from what I can see) compatibilism is more like carving a niche for free will in an otherwise deterministic universe. I find that the "deterministic universe" is a concept without sufficient empirical support. It is a nebulous abstraction of reality, a faith good as any, but even as such it has a structure (a meaning) that disqualifies it from interfering with any practical notion of "free will" as outlined above.

A more pertinent question is really why the question (of whether "free will" exists) is posed at all. I find that it is often for anti-religious reasons, sometimes for moral responsibility evasion, and sometimes because people want to avoid the all too common cases where a single individual (or group of) is chastised when other factors (like circumstances, other people's actions) are just as important. The former two I don't care for much, that latter is more likable, even if none of them accomplish what they seek using this line of reasoning IMO.

[naasking]

> A more pertinent question is really why the question (of whether "free will" exists) is posed at all.

Good question. See my reply to this post for an answer: https://news.ycombinator.com/item?id=18608440

[gpderetta]

I think this practical definition is quite interesting, useful, and worth discussing about, but I suspect that most people have a more nebulous, metaphisical idea of free will.

[ajuc]

Yup, free will is meaningless, there's no measurable difference between an universe which has a free will and one that doesn't.

> are criminals responsible for their actions?

Doesn't matter, what matters is - is a commitment to punish certain behavior decreasing number of people behaving that way? If so - punishing crimes makes sense, no matter if there's a free will or not.

[coldtea]

>Doesn't matter, what matters is - is a commitment to punish certain behavior decreasing number of people behaving that way?

If there's no free will obviously not -- since any crime decision is predetermined upon the earlier states of the universe (even before we appeared on Earth), and any drop in crime in correlation with a "commitment to punish certain behavior" is just correlation itself, and not causation (both things being caused by those earlier states).

In fact, if there's no free will, then there's also no free decision of a "commitment to punish certain behavior" or not: all those decisions to make laws, punish certain behaviors, etc are already determined.

Without free will nobody can influence whether we create this or that law.

In essence, you started that "free will is meaningless" and then your whole argument was like as if free like totally exists and we can take actual non predetermined decisions.

[ajuc]

> If there's no free will obviously not -- since any crime decision is predetermined upon the earlier states of the universe (even before we appeared on Earth), and any drop in crime in correlation with a "commitment to punish certain behavior" is just correlation itself, and not causation (both things being caused by those earlier states).

An asteroid hitting Earth is predetermined by early stages of universe too. Is the collision just correlated with dinosaurs extinction, or was it the cause?

When a ball hits another ball in snooker - is the movement of the second ball caused by the collision, or just correlated with it? In deterministic universe it was known since the start of time it will happen after all.

Why is it different when people are involved?

If everything is determined by early stages of universe - then you can still discover how exactly that determinism works between 2 particular events (no matter that they share a common causation chain higher up).

In the case of criminals and punishement assuming determinism - our brains deterministically evolved to deterministically respond to punishement by deterministically changing the behaviour when punished. This deterministically caused the brains to come to conclussion, that punishement makes sense. Then they deterministically punish crime, and criminals deterministically respond by avoiding being caught which reduces the crime.

No need for the free will at all.

BTW even if universe is not deterministic it doesn't mean free will exists. After all you wouldn't say dices have a free will.

[naasking]

> No need for the free will at all.

You're mistaken. Free will is needed to identify who the criminal is in any given situation, ie. who are the morally responsible parties. You can't escape that with the arguments you presented, and you just skipped it to talk about justice, which is a whole separate matter.

[ajuc]

> Free will is needed to identify who the criminal is in any given situation, ie. who are the morally responsible parties.

Belief in free will is required (in most common ethical systems at least - there were whole religions believing in deterministic salvation, where God made people who will sin and can't do anything about it - and people carried on, punishing crimes as usual :) - instead of striving to be good to deserve the salvation people were supposed to strive to be good to prove they are these that will be saved).

Existence of free will isn't.

In universe without free will there can be people who believe in free will and "decide" who is criminal based on that. It's their brains causing them to do so, but in reality neither they, nor the criminals make any decisions - it's all predetermined. But everything still works the same way, and their consciousnesses feel in control.

BTW even the common moral systems doesn't always care about responsibility - for example if you steal because you're a drug addict - nobody searches for the first guy that gave you the drugs and got you addicted. It's assumed it's impossible to find, and to keep society functioning they put you in jail (even if you had no choice).

[naasking]

I suggest reading my other reply here since it covers the same ground: https://news.ycombinator.com/item?id=18608505

[cthalupa]

>Free will is needed to identify who the criminal is in any given situation,

No. Causality is all you need to identify who the criminal is.

Physical stimuli are basically all we need as intelligent beings to come to a conclusion about things we don't like happening to us, and thus ultimately make into moral laws or codified ones. These are the inputs, these are the causes. Then, once enough human beings have had those inputs result in a large enough collective coming to the conclusion that stuff is Bad, we end up with laws.

Then, whenever someone acts in a way that we have determined is Bad, they are the criminal. They do not need to possess free will for the rational thing to do to be to provide punishment for them, because punishing them ultimately changes the inputs used by others to reach decisions. Nothing about this requires free will - it's all just as applicable if you just believe the laws of physics and causality determine every choice we make.

[naasking]

> No. Causality is all you need to identify who the criminal is.

Unfortunately not. If your gets car stolen, without free will, you can't assign responsibility to the "thief", because you yourself were a causal factor in your car getting stolen: had you parked somewhere else, your car wouldn't have been stolen. Had society or his parents better supported the "thief", he wouldn't have stolen that car. Had your city placed that street somewhere else, your car wouldn't have been stolen.

To designate the "thief" as the "singular" cause that's relevant, you need free will.

[coldtea]

>No. Causality is all you need to identify who the criminal is.

Causality just determines who did something criminal. The word criminal though also invokes a moral judgement (of character) and intent, which presupposes an agent, which presupposes free will.

[coldtea]

>An asteroid hitting Earth is predetermined by early stages of universe too. Is the collision just correlated with dinosaurs extinction, or was it the cause?

It was the cause but it was predetermined itself.

Note that the free will of the dinosaurs or the asteroid doesn't enter the picture at all here. Similarly, and this is the argument, in the drop-in-crime case, our "commitment to punish certain behavior" is not any more of a commitment than the rock was "committed" to extinguish the dinosaurs.

>Why is it different when people are involved?

It's not different physically. The actions can't be said to be decisions is all or to have caused a particular chain of events is all (that chain was the only possible one).

>In the case of criminals and punishement assuming determinism - our brains deterministically evolved to deterministically respond to punishement by deterministically changing the behaviour when punished.

The point is that they couldn't have evolved any other way, and the crime couldn't have been any less or more than what it is. So those measures can't be said to have dropped crime (since it could never be anything else).

[ajuc]

> So those measures can't be said to have dropped crime (since it could never be anything else).

That's redefining words to the extreme.

A rock has dropped because of gravity and lack of support. It doesn't matter that in this universe with the starting conditions as they are - the rock had to drop. It still dropped.

Homo sapiens (or whatever ancestor it was when it happened) - evolved to punish crimes. It lowered the crime, or had other beneficial effects on the population, that's why it was preserved by the evolution. It doesn't matter that it couldn't evolve in any different way given the starting conditions of the universe. It still happened.

[coldtea]

>That's redefining words to the extreme.

The fact that crime can't drop or rise depending on prior decisions, and will always be what it's predetermined to be at any point in the future is pretty non-controversial if you accept determinism.

And requires not "redefining" at all.

When we say "we've dropped crime" we don't just describe a casual relationship, but that we achieved something that otherwise would not be. We understand it as if free will is given, if you like.

>It lowered the crime, or had other beneficial effects on the population, that's why it was preserved by the evolution. It doesn't matter that it couldn't evolve in any different way given the starting conditions of the universe. It still happened.

Only, "that's why it was preserved" has no meaning anymore. It wasn't preserved because of any beneficial effects (which weren't a concern at all). It was preserved because initial conditions were so and so.

[szemet]

If I write a program that scans for counter example of Fermat-theorem it will run until physics is kind to my machine.

The non-halting property of this setup is caused by the physical arrangements/movements of particles and the laws of nature. Fermat-theorem is an abstraction having no real causal power.

I would say, that for us the proof of Fermat-theorem is the right abstraction level to explain the general behavior of the machine: The machine does not halt because Fermat theorem is true

Similarly "punishment causes less crime" may be the right level of abstraction if the correlation is there, even if it can't be reduced to physical terms. And "Responsibility for action due to free will" may be worse abstraction if agent causation[1] is impossible (I am not saying it is)

So even if increasing/decreasing punishments is determined by low level laws of nature, we still can use the higher level explanation if it suits us better.

[1] https://en.wikipedia.org/wiki/Agent_causation

[coldtea]

>Similarly "punishment causes less crime" may be the right level of abstraction if the correlation is there, even if it can't be reduced to physical terms.

If it can't be reduced to physical terms (and there's no free will) it doesn't matter if its the "right level of abstraction" or not.

We have no other alternative but to think of it at whether level of abstraction we're determined to think of it. We don't have a say in the matter for our thoughts on it to matter at all.

(In fact both our thoughts of it and this comment and your responses will also be pre-determined).

>So the future increasing/decreasing in punishment is depend on elementary particles, also the effect of those, but we still can use the higher level explanation if it suit us better.

This presupposes some "us" that can or cannot use the this or that explanation at will. Which is exactly what we assumed doesn't exist.

If free will doesn't exist, then we don't have any say on whether we "use the higher level explanation if it suit us better". We use it or we don't use it as was predetermined by the causal chain of the universe -- not because it "suits" us.

[szemet]

Ok. I see your point. So I just hope (not actively as a free agent - of course this was determined too) that we are determined to use logically more consistent explanations for our behavior than classical libertarian free will. ;)

[naasking]

> Yup, free will is meaningless, there's no measurable difference between an universe which has a free will and one that doesn't.

Actually there is: in a universe with free will, you must have intelligent agents that can understand their own actions and learn from their choices. A universe lacking free will would lack these qualities, and be observably different as a result.

[ajuc]

> A universe lacking free will would lack these qualities

Why?

[naasking]

Because free will merely requires the ability to understand the choices we're making and learn from them, like I said.

[szemet]

AlphaZero made choices in chess, and learned from them. So did I get you right: it had a free-will because it is a necessity for this behavior according to you? Because mere "machines" can't make decisions, and can't learn from consequences.

[naasking]

> AlphaZero made choices in chess, and learned from them. So did I get you right: it had a free-will because it is a necessity for this behavior according to you?

That's domain-specific learning, not general learning. If you could then take AlphaZero out for coffee to talk about a match and why it made certain choices, I'm fairly certain you'd be less confident that it wouldn't qualify.

> Because mere "machines" can't make decisions, and can't learn from consequences.

I disagree. Humans are a type of machine.

[intralizee]

No that is absurd and dehumanizing towards the victims.

[cthalupa]

Plenty of things about reality are unfair. I'm not quite sure how free will being nonexistent is dehumanizing towards victims, but even if you feel it will be, that isn't an argument that free will exists.

[intralizee]

No your last part it was directed towards.