So, for quite a while the dominant paradigm in criminal justice has been one of "crime and punishment". Why not use positive reinforcement?
I remain skeptical of the presented statistics:
According to Clapham, youth recidivism was reduced from 60% to 8%. Overall crime was reduced by 40%. Youth crime was cut in half. And it cost one-tenth of the traditional judicial system."
but the overall idea still seems like it would actually work.
It seems to me that humans respond much better to Skinner boxes than they do to punishment; I think Scott would agree, given his Skinner box obsession. At the very least, the fact we are using such a limited set of the techniques psychology gives us for social control should give us pause.
There are counter studies showing that the moment you reward someone for "good" behavior, they begin acting toward the reward, not the behavior itself. It makes the motiviation extrinsic, and can actually lead to reduced "good" behavior in the long term despite short term gains.
There was a TED talk on just this topic some time ago.
More importantly, initiatives like this give the police an excuse to harass anyone they want, since now beginning an encounter can involve both probable cause or, as cause, a complete lack of probable cause. Do you believe, in the US for example, that a cop talking to a kid who "did the right thing" (I don't know, didn't jaywalk, for example, or hadn't robbed the bank he walked past), and then smelled weed on his breath/clothes, wouldn't end up citing/arresting him? In a previous world, he wouldn't have had the excuse to even talk to the kid in the first place.
There are counter studies showing that the moment you reward someone for "good" behavior, they begin acting toward the reward, not the behavior itself. It makes the motiviation extrinsic, and can actually lead to reduced "good" behavior in the long term despite short term gains.
That's interesting. I think it could depend on how it's done, though. If you use techniques more along the lines of encouragement and social interaction rather than wordless giving of a reward, you can perhaps internalize some component of the reward loop - i.e. the person will think to themselves, "hey, I just did the right thing" and feel good about it.
Given that the current system of social control is pretty shit and there's other approaches to be tried, I think we should be looking into them a lot more than we currently are.
More importantly, initiatives like this give the police an excuse to harass anyone they want, since now beginning an encounter can involve both probable cause or, as cause, a complete lack of probable cause. Do you believe, in the US for example, that a cop talking to a kid who "did the right thing" (I don't know, didn't jaywalk, for example, or hadn't robbed the bank he walked past), and then smelled weed on his breath/clothes, wouldn't end up citing/arresting him? In a previous world, he wouldn't have had the excuse to even talk to the kid in the first place.
Good point. However, I think you can resolve this by making the "positive reinforcement" branch entirely separate from normal policing - for example, you could set it up such that it's done by normal citizens rather than police.
The extrinsic problem is sort-of always a problem, in rewards or punishment. Some incentives probably wouldn't be a terrible idea... I wouldn't have any problems with something like "If you havn't had a ticket in x years, your license and plate tag renewals are at a reduced rate."
More importantly, initiatives like this give the police an excuse to harass anyone they want, since now beginning an encounter can involve both probable cause or, as cause, a complete lack of probable cause. Do you believe, in the US for example, that a cop talking to a kid who "did the right thing" (I don't know, didn't jaywalk, for example, or hadn't robbed the bank he walked past), and then smelled weed on his breath/clothes, wouldn't end up citing/arresting him? In a previous world, he wouldn't have had the excuse to even talk to the kid in the first place.
Good point. However, I think you can resolve this by making the "positive reinforcement" branch entirely separate from normal policing - for example, you could set it up such that it's done by normal citizens rather than police.
The thought of governments expanding to pay people to hand out gold star stickers, when most can't even manage their own budgets as is, makes me sick.
So, for quite a while the dominant paradigm in criminal justice has been one of "crime and punishment".
Crime and punishment is a novel by Fyodor Dostoyevsky. The social aim of the judicial system is one of prevention, correction, and rehabilitation, not punishment. At least not in a western civilised nation. Oh wait...
Yeah, I phrased that badly and I should really read that novel sometime. However, my main point was not about the stated goals of the system (though I would say that those stated goals are often contradicted by what actually gets done), but rather the techniques used to achieve them.
Also, it definitely seems to me that a retributive theory of justice is dominant in countries like the U.S. Some effort is put into rehabilitation, but a lot of the time it seems like mere lip-service.
Yeah, but the techniques, as the article says, are well known. Psychology is not a mystery science anymore and many ways for rehabilitation are known to work. However, those methods have as prerequisites a society that is willing to see murderers go (eventually) free.
I find it a much more interesting topic of discussion and debate whether that is ever going to be the case for the US.
Yeah, but the techniques, as the article says, are well known. Psychology is not a mystery science anymore and many ways for rehabilitation are known to work. However, those methods have as prerequisites a society that is willing to see murderers go (eventually) free.
Not always. Perhaps it's better to use "social control" rather than justice as the overarching discussion topic, but for prevention of crime there are lots of methods that can be used without triggering concern over murderers going free.
I find it a much more interesting topic of discussion and debate whether that is ever going to be the case for the US.
It will definitely take a while, especially given that there seem to be basic neurological impulses at play behind the very idea of punishment. However, I think it's rather unlikely that this is something that will never change in the U.S.
Obviously much study in criminal justice and psychology has been done, and I am not knowledgeable in those areas at all.
But, from my layman's perspective, the key difference between positive and negative reinforcement lies with the consistency of application.
See, positive reinforcement, when applied, is usually applied with a guarantee. That's the situation where food comes out every time you pull the lever. In those cases, the rat only pulls the lever when it is hungry. But if the food only comes out some of the time, the rat pulls the lever like crazy. That is how you get a slot machine.
With negative reinforcement, you get the opposite effect. If the punishment is guaranteed, almost nobody will perform the negative behavior. Imagine if you got a speeding ticket automatically if your car went even one fraction over the speed limit. You would see all the cars, except the very rich people who don't care, going 5 below the limit. But when the punishment is dealt out inconsistently, people will perform the negative behavior constantly, because they think they can get away with it.
It seems like it's the opposite, but it's also the same. When the reinforcement, whether negative or positive, is applied inconsistently, the associated behavior is expressed more frequently. When the reinforcement is consistently applied, the behavior is expressed less frequently.
I think it's rather unlikely that this is something that will never change in the U.S.
Our prison and criminal justice systems are too closely tied to local economic and political interests. Considering the increasing congressional deadlock and lack of any actual discourse, I honestly don't foresee significant correctional reform in my lifetime.
Obviously much study in criminal justice and psychology has been done, and I am not knowledgeable in those areas at all.
But, from my layman's perspective, the key difference between positive and negative reinforcement lies with the consistency of application.
See, positive reinforcement, when applied, is usually applied with a guarantee. That's the situation where food comes out every time you pull the lever. In those cases, the rat only pulls the lever when it is hungry. But if the food only comes out some of the time, the rat pulls the lever like crazy. That is how you get a slot machine.
With negative reinforcement, you get the opposite effect. If the punishment is guaranteed, almost nobody will perform the negative behavior. Imagine if you got a speeding ticket automatically if your car went even one fraction over the speed limit. You would see all the cars, except the very rich people who don't care, going 5 below the limit. But when the punishment is dealt out inconsistently, people will perform the negative behavior constantly, because they think they can get away with it.
It seems like it's the opposite, but it's also the same. When the reinforcement, whether negative or positive, is applied inconsistently, the associated behavior is expressed more frequently. When the reinforcement is consistently applied, the behavior is expressed less frequently.
I wonder if this has factored into automatic traffic tickets at all.
I remember there being a big freak-out when automatic toll collection was first starting up. People were afraid that the government's ability to time your travel from point A to point B would allow them to prove you traveled over the speeding limit, and mail you a speeding ticket. It never happened, even though the means are right there. Probably b/c they want automatic toll collection to be a success and for people to use it, but now we're at a point where it is mandatory in a lot of areas. You still don't get a ticket in the mail, perhaps because they realize if you will get caught every time, nobody will ever speed again? All that revenue lost.
Would anyone here speed if you were 100% guaranteed to be caught?
And if such a system existed, how would it even work? Let's say the speed limit is 75. I'm cruising at 70 and hit 76 for a little bit when the road starts to go downhill. I dip back under 74 and then back up to 76. Do I get two speeding tickets? What if there is one of those rare legitimate reasons for me to go fast, such as an emergency? Do I get ticketed?
You'll probably have to change it to charge a fine based on the number of exact mph you are over the limit, and for the duration. Graph your speed over time. Draw a horizontal line at the speed limit. Integrate to get the area above the speed limit line, but below the curve of your graph. Multiply the area by some number of pennies to get your fine. They could just integrate the fine right into your car registration or taxes.
Or... raise the speed limit to a reasonable level on freeways and only auto-ticket edge cases based on an RMS of the speed over time.
If the speed limits were reasonable, there's little liklihood that anyone speeding long enough to breach a well-thought-out RMS threshold was doing so for any good reason.
You are correct that the current fine structure wouldn't work. It'd only be able to tell your average speed over that leg of the trip. That's why I still pondered the question. You could still speed for just a little bit and not raise any flags. So would people still do it? Would they have that feeling of being able to beat the system, or would they all just accept the law?
And I agree with Rym that if the laws actually made sense, that we wouldn't even be having this conversation.
Texas and Ohio both completely fucked over a few towns that took it too far and disincorporated them.
Tangentially, the only places I've ever seen people get tickets are semi-poor states like Rhode Island. I've seen cars regularly break 80 or 85 MPH in Massachusetts, and nobody gives a shit. I think tax revenue as a percentage of state budget is probably one of the biggest indicators of probability of getting pulled over for speeding.
Comments
There was a TED talk on just this topic some time ago.
More importantly, initiatives like this give the police an excuse to harass anyone they want, since now beginning an encounter can involve both probable cause or, as cause, a complete lack of probable cause. Do you believe, in the US for example, that a cop talking to a kid who "did the right thing" (I don't know, didn't jaywalk, for example, or hadn't robbed the bank he walked past), and then smelled weed on his breath/clothes, wouldn't end up citing/arresting him? In a previous world, he wouldn't have had the excuse to even talk to the kid in the first place.
Given that the current system of social control is pretty shit and there's other approaches to be tried, I think we should be looking into them a lot more than we currently are. Good point. However, I think you can resolve this by making the "positive reinforcement" branch entirely separate from normal policing - for example, you could set it up such that it's done by normal citizens rather than police.
Also, it definitely seems to me that a retributive theory of justice is dominant in countries like the U.S. Some effort is put into rehabilitation, but a lot of the time it seems like mere lip-service.
I find it a much more interesting topic of discussion and debate whether that is ever going to be the case for the US.
But, from my layman's perspective, the key difference between positive and negative reinforcement lies with the consistency of application.
See, positive reinforcement, when applied, is usually applied with a guarantee. That's the situation where food comes out every time you pull the lever. In those cases, the rat only pulls the lever when it is hungry. But if the food only comes out some of the time, the rat pulls the lever like crazy. That is how you get a slot machine.
With negative reinforcement, you get the opposite effect. If the punishment is guaranteed, almost nobody will perform the negative behavior. Imagine if you got a speeding ticket automatically if your car went even one fraction over the speed limit. You would see all the cars, except the very rich people who don't care, going 5 below the limit. But when the punishment is dealt out inconsistently, people will perform the negative behavior constantly, because they think they can get away with it.
It seems like it's the opposite, but it's also the same. When the reinforcement, whether negative or positive, is applied inconsistently, the associated behavior is expressed more frequently. When the reinforcement is consistently applied, the behavior is expressed less frequently.
I remember there being a big freak-out when automatic toll collection was first starting up. People were afraid that the government's ability to time your travel from point A to point B would allow them to prove you traveled over the speeding limit, and mail you a speeding ticket. It never happened, even though the means are right there. Probably b/c they want automatic toll collection to be a success and for people to use it, but now we're at a point where it is mandatory in a lot of areas. You still don't get a ticket in the mail, perhaps because they realize if you will get caught every time, nobody will ever speed again? All that revenue lost.
And if such a system existed, how would it even work? Let's say the speed limit is 75. I'm cruising at 70 and hit 76 for a little bit when the road starts to go downhill. I dip back under 74 and then back up to 76. Do I get two speeding tickets? What if there is one of those rare legitimate reasons for me to go fast, such as an emergency? Do I get ticketed?
You'll probably have to change it to charge a fine based on the number of exact mph you are over the limit, and for the duration. Graph your speed over time. Draw a horizontal line at the speed limit. Integrate to get the area above the speed limit line, but below the curve of your graph. Multiply the area by some number of pennies to get your fine. They could just integrate the fine right into your car registration or taxes.
If the speed limits were reasonable, there's little liklihood that anyone speeding long enough to breach a well-thought-out RMS threshold was doing so for any good reason.
And I agree with Rym that if the laws actually made sense, that we wouldn't even be having this conversation.
Texas and Ohio both completely fucked over a few towns that took it too far and disincorporated them.