The design of social minds involves two key tradeoffs, which interact in an important way. The first tradeoff is that social minds must both make good decisions, and present good images to others. Our thoughts influence both our actions and what others think of us. It would be expensive to maintain two separate minds for these two purposes, and even then we would have to maintain enough consistency to convince outsiders a good-image mind was in control. It is cheaper and simpler to just have one integrated mind whose thoughts are a compromise between these two ends.
I don’t think dieting fits with the rest of the examples, because with dieting people are usually acutely aware that their short term and long term desires are inconsistent, and frequently talk about it. It also makes less sense that declaring the intention to diet is about presenting a good image, as it mostly benefits the individual.
"I expect we have a continuum of systems for varying levels of detail."
Is this supportable? We expect memory and expertise to be encoded in a connectionist fashion. How unfortunate it would be to have to continuously transfer these memories to new systems (read: new areas of the brain) as more detail is made available or our interest in the subject increases. A property of all learning is starting with little detail and going to more detail. Our minds will be optimized for that pattern. In your description this learning process cuts across systems every time.
My opinion here is that near/far (really just concrete/abstract) are a matter really of amount of resources devoted, which in turn is a function of the details available to consume and whether the problem merits the effort. However, it's all applied on the same basic machinery. If abstract thinking is as far as you get on a subject it is either because:a) You are unwilling to process the details available.b) There are no details available.c) You are unwilling to make-up details.d) The subject is too complex for you (your concrete conclusions repeatedly fail verification)
The bias towards image making rather than accurate projection of current behavior in far-off descriptions of self is real but seems to require a different explanation. We admit that to synch these up is a hard problem and that may be precisely why it exists. Eating pizza today does not directly falsify the goal to be skinny.
A lovely post!My own conclusion is that" getting things right" is much more important than "being seen as nice". Decisions are more important than image. And then I think both systems (near and far) can be used. A map of the underground is a wonderful example of "sparse thinking" and it is at the same time wonderfully useful in the near decison of which platform to stand on.Thus I would agree with Hal, that the issue of tortue (in this context) is image and should be rejected as a near solution to the question Retired posed on how to use the model in practice. When Hal ends his post with "IMO" this should be regarded as a social (far) signal of near/decisional modesty. When Retired ends his post with "IMO" this should be regarded as a social signal of social imodesty.In other words my answer to your question, Retired, is to throw away the harvesting of social benefits.
What of the intermediary decision zone between near and far? e.g. We often make quite detailed plans for vacations and decisions about them having only fuzzy, limited details and vague notions of what we think we would like. Compare to Chess or Go strategy with their opening, middle game, and end game, or even to Delany's concept of thinking borrowed from earth science: simplex, complex, multiplex.
Chad, thank you for posting the link to A Breakdown of Will. Thank you for posting it twice, because I only bothered to follow the URL after you reiterated. Interesting stuff.
Retired U: "So, again, is near-far only descriptive, or is it a mechanism that can be consciously controlled to improve decisions?"
Most people have near-far conflicts they have trouble resolving consciously. I have known for a long time that I should exercise more. But, near-far is a description of our thinking that might lead to useful progress, where we learn better tricks for resolving these conflicts, and also identify more cases where either near or far thinking tends to be in error.
I didn't say I was concerned about the torture issue; I said it was an example of an issue where near and far thinking might give conflicting results. I asked how one should deal with issues that have such conflict. Your ad hominem comment does not advance the so-called "intentions" of this blog. IMO.
Retired U, I'd suggest that it is a waste of time to worry about what your policy should be towards torture. You probably aren't wondering if you should start (or stop!) torturing people. You probably aren't even in a position to materially influence whether anyone else is torturing people. The belief that you should spend time on this issue is exactly the sort of self-serving bias that this blog is intended to eliminate. IMO.
This is fascinating! I like how this theory fits with other, more specific examples of bias, such as _Stumbling On Happiness_, and voting as signalling tribal identification rather than being accurate. And the research on deliberation where people's views become more extreme after talking to like-minded people. If deliberation were an exchange of details using the "near" system, it should make people's views more informed, but if it is an exercise in proving one's values by using the "far" system, it makes sense that it drives views to the extreme.
And in any case, if you haven't read "A Breakdown of Will" before, highly recommend it (http://www.picoeconomics.co... -- most thought provoking writing I've come across recently (OB aside, of course).
Robin: Chad, will power makes no sense without a conflict between different internal systems.
My point was not that there wasn't a conflict between two internal systems in will power tradeoffs. I was trying to point out that calling all such conflicts as being between "better decision vs better image" seemed like an overgeneralization.
"It can make sense to have specialized mental systems for these different approaches, i.e., systems best at reasoning from detailed representations, versus systems best at reasoning from sparse abstractions. "
I would question this hypothesis. I find it perfectly reasonable to expect that the same basic mental architecture grows from abstract to concrete performance with improved information quite easily on a continuous spectrum. Perhaps an unconvincing analogy, but OOP has the same basic architecture regardless of where in the super/sub class hierarchy you are operating.
My general impression is that concrete thinking is akin to a system with more extensive 'training' on more 'data'. Thus, any comment of pulling out abstract thought and plugging in concrete thought seems nonsensical aside from the process of training or learning to get from one to the other. It is also nonsensical to approach a new subject from the concrete-thinking mode. One starts at abstract, and moves to concrete.
Likewise, I think one could easily concrete-think on distant future topics but you are not guaranteed to have 'trained' your mind to operate on 'data' (which can be your own prior conclusions) that is provably connected to reality. In order to concrete-think about the future all you have to do is practice. However, it is quite easy to build a castle on sand and be unaware of it. To have a detailed, but erroneus, account of any subject. I hope there is no irony there. :p
I don’t think dieting fits with the rest of the examples, because with dieting people are usually acutely aware that their short term and long term desires are inconsistent, and frequently talk about it. It also makes less sense that declaring the intention to diet is about presenting a good image, as it mostly benefits the individual.
I don't know what happened there - the URL was http://www.economist.com/sc... .
This sounds like an example:Why people procrastinate | Motivating minds | The Economist' rel="nofollow noopener" title='http://www.economist.com/science/displaystory.cfm?story_id=12971028" rel="nofollow">Why people procrastinate | Motivating minds | The Economist'>
"I expect we have a continuum of systems for varying levels of detail."
Is this supportable? We expect memory and expertise to be encoded in a connectionist fashion. How unfortunate it would be to have to continuously transfer these memories to new systems (read: new areas of the brain) as more detail is made available or our interest in the subject increases. A property of all learning is starting with little detail and going to more detail. Our minds will be optimized for that pattern. In your description this learning process cuts across systems every time.
My opinion here is that near/far (really just concrete/abstract) are a matter really of amount of resources devoted, which in turn is a function of the details available to consume and whether the problem merits the effort. However, it's all applied on the same basic machinery. If abstract thinking is as far as you get on a subject it is either because:a) You are unwilling to process the details available.b) There are no details available.c) You are unwilling to make-up details.d) The subject is too complex for you (your concrete conclusions repeatedly fail verification)
The bias towards image making rather than accurate projection of current behavior in far-off descriptions of self is real but seems to require a different explanation. We admit that to synch these up is a hard problem and that may be precisely why it exists. Eating pizza today does not directly falsify the goal to be skinny.
Thanos, I expect we have a continuum of systems for varying levels of detail. I've tried to write everything to be consistent with that.
A lovely post!My own conclusion is that" getting things right" is much more important than "being seen as nice". Decisions are more important than image. And then I think both systems (near and far) can be used. A map of the underground is a wonderful example of "sparse thinking" and it is at the same time wonderfully useful in the near decison of which platform to stand on.Thus I would agree with Hal, that the issue of tortue (in this context) is image and should be rejected as a near solution to the question Retired posed on how to use the model in practice. When Hal ends his post with "IMO" this should be regarded as a social (far) signal of near/decisional modesty. When Retired ends his post with "IMO" this should be regarded as a social signal of social imodesty.In other words my answer to your question, Retired, is to throw away the harvesting of social benefits.
What of the intermediary decision zone between near and far? e.g. We often make quite detailed plans for vacations and decisions about them having only fuzzy, limited details and vague notions of what we think we would like. Compare to Chess or Go strategy with their opening, middle game, and end game, or even to Delany's concept of thinking borrowed from earth science: simplex, complex, multiplex.
To be consistent, estimates made by sparse approaches should equal the average of estimates made when both sparse and detail approaches contribute.
Isn't this goal recursive? Why not just say "estimates made by sparse approaches should attempt to approximate estimates made by detailed approaches"?
Chad, thank you for posting the link to A Breakdown of Will. Thank you for posting it twice, because I only bothered to follow the URL after you reiterated. Interesting stuff.
Retired U: "So, again, is near-far only descriptive, or is it a mechanism that can be consciously controlled to improve decisions?"
Most people have near-far conflicts they have trouble resolving consciously. I have known for a long time that I should exercise more. But, near-far is a description of our thinking that might lead to useful progress, where we learn better tricks for resolving these conflicts, and also identify more cases where either near or far thinking tends to be in error.
@Hal Finney:
I didn't say I was concerned about the torture issue; I said it was an example of an issue where near and far thinking might give conflicting results. I asked how one should deal with issues that have such conflict. Your ad hominem comment does not advance the so-called "intentions" of this blog. IMO.
"the sort of self-serving bias"
I think you must mean self-deluding; worrying about something you can't evenaffect doesn't fit any use of "self-serving" I've ever come across.
Retired U, I'd suggest that it is a waste of time to worry about what your policy should be towards torture. You probably aren't wondering if you should start (or stop!) torturing people. You probably aren't even in a position to materially influence whether anyone else is torturing people. The belief that you should spend time on this issue is exactly the sort of self-serving bias that this blog is intended to eliminate. IMO.
This is fascinating! I like how this theory fits with other, more specific examples of bias, such as _Stumbling On Happiness_, and voting as signalling tribal identification rather than being accurate. And the research on deliberation where people's views become more extreme after talking to like-minded people. If deliberation were an exchange of details using the "near" system, it should make people's views more informed, but if it is an exercise in proving one's values by using the "far" system, it makes sense that it drives views to the extreme.
And in any case, if you haven't read "A Breakdown of Will" before, highly recommend it (http://www.picoeconomics.co... -- most thought provoking writing I've come across recently (OB aside, of course).
Robin: Chad, will power makes no sense without a conflict between different internal systems.
My point was not that there wasn't a conflict between two internal systems in will power tradeoffs. I was trying to point out that calling all such conflicts as being between "better decision vs better image" seemed like an overgeneralization.
"It can make sense to have specialized mental systems for these different approaches, i.e., systems best at reasoning from detailed representations, versus systems best at reasoning from sparse abstractions. "
I would question this hypothesis. I find it perfectly reasonable to expect that the same basic mental architecture grows from abstract to concrete performance with improved information quite easily on a continuous spectrum. Perhaps an unconvincing analogy, but OOP has the same basic architecture regardless of where in the super/sub class hierarchy you are operating.
My general impression is that concrete thinking is akin to a system with more extensive 'training' on more 'data'. Thus, any comment of pulling out abstract thought and plugging in concrete thought seems nonsensical aside from the process of training or learning to get from one to the other. It is also nonsensical to approach a new subject from the concrete-thinking mode. One starts at abstract, and moves to concrete.
Likewise, I think one could easily concrete-think on distant future topics but you are not guaranteed to have 'trained' your mind to operate on 'data' (which can be your own prior conclusions) that is provably connected to reality. In order to concrete-think about the future all you have to do is practice. However, it is quite easy to build a castle on sand and be unaware of it. To have a detailed, but erroneus, account of any subject. I hope there is no irony there. :p