Punishment is unreliable, especially in a noisy environment, precisely because it’s a powerful motivator.
You’re right: if I use a powerful aversive, then each individual correction has a large impact on future behavior. This is sometimes known as “positive punishment” in behavior modification.
That’s great, if I never make mistakes, and I only want to eliminate behaviors: the subject quickly stops doing the things I don’t want it to do. Wonderful.
There are three major difficulties with this approach.
“I never make mistakes” doesn’t describe the real world, and mistakenly punishing subjects can cause them to stop doing things I want them to do. If I think your comment was stupid and I give you a hard time and make you feel like a fool, your brain will form a connection between your comment and the negative utility, as you say. This happens even if I was wrong and your comment was actually not stupid at all. So if I make mistakes, I can inhibit useful contributions as readily as useless ones.
Even if I’m right, I can’t control how your brain generalizes that connection. That is, I can’t control what class your brain sorts the punished comment into, so I can’t control (or even predict!) what classes of behavior I am inhibiting. For example, suppose someone challenges a widely held belief with a poor argument, and I give them a hard time about it and make them feel like a fool. They might generalize to the class of poor arguments, in which case great! Alternatively they might generalize to the class of belief-challenges, or the class of commenting-on-LW, or all three.
Even if I’m lucky, and only inhibit the behaviors I want to inhibit, all I’ve done is get rid of things I don’t want. I haven’t done anything to encourage getting more of what I do want. That’s fine if the baseline is something I want to preserve, but it’s counterproductive if it’s something I want to improve on.
For my own part, if I say something stupid I expect people to either ignore me or correct me politely. The second of those is a favor they do me; I’m not entitled to it. But I don’t expect them to make me feel like a fool. If I started getting that response, I’d leave.
I do not mind occasional false negatives in criticism because I do not instantly accept every counter argument I hear. Mindless trolling or insults have almost no effect on my certainty that an argument is valid, and I do not radically alter my original belief until new information allows me to reproduce an error, or someone points out a logical inconsistency.
The odds that I will say something correct, and that someone will be able to persuade me that I am probably wrong using a faulty argument are insignificant, since I would be unable to find the same error upon reflection no matter which way I thought about it, and would be naturally skeptical of new information unless the source was at least as reliable as the source of my original information. However, If I was missing or ignored information that affects my point, arrived at my conclusion through a logical fallacy, or used information to arrive at my conclusion that I did not tell my critic, then I attempt to fix these problems in future posts or edits.
As for worrying about someone being wrong, and actually reaching a less rational state of mind in response to your criticism, this is why being careful about how you criticize someone is important. If you specifically tell someone what parts of their argument are flawed, then you can minimize collateral damage to correct ideas and beliefs. The connotation of “politeness” is extremely negative to me, but if you are going to deliberately modify someones thoughts than you should do so as non-invasively as possible.
In short, if someone has not learned how to react to and judge the validity of criticism, then they will have to learn at some point. My ignoring it in the short term is only going to leave them vulnerable to the next person who persuades them to change their beliefs based on a flawed argument.
I don’t object to pointing out flaws in arguments, and I endorse being careful about how you do it in order to “collateral damage”.
That isn’t what you seemed to be championing originally (e.g., “If I say something stupid, I expect other people to give me a hard time about it. My brain forms a connection between the stupid comment and the negative utility, and I develop techniques to avoid making a fool of myself the same way in the future. ”). But perhaps I misunderstood you.
Insults don’t have much effect on my confidence in arguments, but they have a lot of effect on my willingness to have the conversation in the first place.
I’m willing to avoid the word “polite” if it has negative connotations for you. Do you prefer “non-invasive,” then?
In retrospect my first comment was misleading, I edited too much.
I have heard politeness used as a blanket statement which covers telling lies to make people feel good, never objecting to something someone considers part of their identity, along with being careful about how you communicate criticism and avoiding swearing. There are a lot of additions and variations between cultures, but these are the mental tags I have built up. So yes, I think politeness has too broad of a meaning to describe how people should interact if they actually want to accomplish something. I have rarely used non-invasive or collatoral damage in this context before, so I am open to more widely used alternatives.
If I’ve understood you, you endorse “being careful about how you criticize someone” as important, but reject “being careful about how you communicate criticism”. Can you say more about what distinction you’re trying to convey there?
Re: politeness...
For my own part, I think of politeness as a natural consequence of signal processing considerations. Speech and body language are noisy, and when a communication channel is noisy it’s helpful to establish standard protocols for redundancy and error-correction, and standard signals for initial handshaking and for key operations.
For most mammals, this includes being careful about when and how to challenge others, which is what most of your examples are about, because status challenges are expensive enough that unintentional ones waste crucial resources.
And, yes, I agree that “politeness” is a blanket term for that and similar things.
If we had better channels, we could engage in continual precise negotiations about status, and we wouldn’t need mechanisms as clumsy as explicit status challenges… and we wouldn’t need social rules for keeping those challenges under control. But most of us don’t have good enough channels for that, and all of us have emotional responses that evolved in bodies and environments that didn’t.
And, yes, if I lack sufficient finesse to avoid issuing unintentional status challenges, I will get slapped down for it until I either learn or leave the tribe (possibly horizontally).
I do not think everything which is considered polite is bad (such as careful criticism and minimal insults), I was just trying to show that it covers a wide range of behaviors, some of which suppress rational debate.
I agree with the rest of your points, but even if society currently works that way, I still wish people’s happiness was not affected by winning (or avoiding losing) arguments. Perhaps my resistance to social norms is hopeless, but intelligent criticism improves my perception of reality, and I appreciate others doing it in the fastest way possible.
Punishment is unreliable, especially in a noisy environment, precisely because it’s a powerful motivator.
You’re right: if I use a powerful aversive, then each individual correction has a large impact on future behavior. This is sometimes known as “positive punishment” in behavior modification.
That’s great, if I never make mistakes, and I only want to eliminate behaviors: the subject quickly stops doing the things I don’t want it to do. Wonderful.
There are three major difficulties with this approach.
“I never make mistakes” doesn’t describe the real world, and mistakenly punishing subjects can cause them to stop doing things I want them to do. If I think your comment was stupid and I give you a hard time and make you feel like a fool, your brain will form a connection between your comment and the negative utility, as you say. This happens even if I was wrong and your comment was actually not stupid at all. So if I make mistakes, I can inhibit useful contributions as readily as useless ones.
Even if I’m right, I can’t control how your brain generalizes that connection. That is, I can’t control what class your brain sorts the punished comment into, so I can’t control (or even predict!) what classes of behavior I am inhibiting. For example, suppose someone challenges a widely held belief with a poor argument, and I give them a hard time about it and make them feel like a fool. They might generalize to the class of poor arguments, in which case great! Alternatively they might generalize to the class of belief-challenges, or the class of commenting-on-LW, or all three.
Even if I’m lucky, and only inhibit the behaviors I want to inhibit, all I’ve done is get rid of things I don’t want. I haven’t done anything to encourage getting more of what I do want. That’s fine if the baseline is something I want to preserve, but it’s counterproductive if it’s something I want to improve on.
For my own part, if I say something stupid I expect people to either ignore me or correct me politely. The second of those is a favor they do me; I’m not entitled to it. But I don’t expect them to make me feel like a fool. If I started getting that response, I’d leave.
I do not mind occasional false negatives in criticism because I do not instantly accept every counter argument I hear. Mindless trolling or insults have almost no effect on my certainty that an argument is valid, and I do not radically alter my original belief until new information allows me to reproduce an error, or someone points out a logical inconsistency.
The odds that I will say something correct, and that someone will be able to persuade me that I am probably wrong using a faulty argument are insignificant, since I would be unable to find the same error upon reflection no matter which way I thought about it, and would be naturally skeptical of new information unless the source was at least as reliable as the source of my original information. However, If I was missing or ignored information that affects my point, arrived at my conclusion through a logical fallacy, or used information to arrive at my conclusion that I did not tell my critic, then I attempt to fix these problems in future posts or edits.
As for worrying about someone being wrong, and actually reaching a less rational state of mind in response to your criticism, this is why being careful about how you criticize someone is important. If you specifically tell someone what parts of their argument are flawed, then you can minimize collateral damage to correct ideas and beliefs. The connotation of “politeness” is extremely negative to me, but if you are going to deliberately modify someones thoughts than you should do so as non-invasively as possible.
In short, if someone has not learned how to react to and judge the validity of criticism, then they will have to learn at some point. My ignoring it in the short term is only going to leave them vulnerable to the next person who persuades them to change their beliefs based on a flawed argument.
I don’t object to pointing out flaws in arguments, and I endorse being careful about how you do it in order to “collateral damage”.
That isn’t what you seemed to be championing originally (e.g., “If I say something stupid, I expect other people to give me a hard time about it. My brain forms a connection between the stupid comment and the negative utility, and I develop techniques to avoid making a fool of myself the same way in the future. ”). But perhaps I misunderstood you.
Insults don’t have much effect on my confidence in arguments, but they have a lot of effect on my willingness to have the conversation in the first place.
I’m willing to avoid the word “polite” if it has negative connotations for you. Do you prefer “non-invasive,” then?
In retrospect my first comment was misleading, I edited too much.
I have heard politeness used as a blanket statement which covers telling lies to make people feel good, never objecting to something someone considers part of their identity, along with being careful about how you communicate criticism and avoiding swearing. There are a lot of additions and variations between cultures, but these are the mental tags I have built up. So yes, I think politeness has too broad of a meaning to describe how people should interact if they actually want to accomplish something. I have rarely used non-invasive or collatoral damage in this context before, so I am open to more widely used alternatives.
Re: criticism...
If I’ve understood you, you endorse “being careful about how you criticize someone” as important, but reject “being careful about how you communicate criticism”. Can you say more about what distinction you’re trying to convey there?
Re: politeness...
For my own part, I think of politeness as a natural consequence of signal processing considerations. Speech and body language are noisy, and when a communication channel is noisy it’s helpful to establish standard protocols for redundancy and error-correction, and standard signals for initial handshaking and for key operations.
For most mammals, this includes being careful about when and how to challenge others, which is what most of your examples are about, because status challenges are expensive enough that unintentional ones waste crucial resources.
And, yes, I agree that “politeness” is a blanket term for that and similar things.
If we had better channels, we could engage in continual precise negotiations about status, and we wouldn’t need mechanisms as clumsy as explicit status challenges… and we wouldn’t need social rules for keeping those challenges under control. But most of us don’t have good enough channels for that, and all of us have emotional responses that evolved in bodies and environments that didn’t.
And, yes, if I lack sufficient finesse to avoid issuing unintentional status challenges, I will get slapped down for it until I either learn or leave the tribe (possibly horizontally).
I do not think everything which is considered polite is bad (such as careful criticism and minimal insults), I was just trying to show that it covers a wide range of behaviors, some of which suppress rational debate. I agree with the rest of your points, but even if society currently works that way, I still wish people’s happiness was not affected by winning (or avoiding losing) arguments. Perhaps my resistance to social norms is hopeless, but intelligent criticism improves my perception of reality, and I appreciate others doing it in the fastest way possible.