A certain mother habitually rewards her small son with ice cream after he eats his spinach. What additional information would you need to be able to predict whether the child will: a. Come to love or hate spinach, b. Love or hate ice cream, or c. Love or hate Mother?
When you interact with someone, you may think, I will do this, so that they will do that, or think such-and-such, or feel thus-and-so; but what is actually going on for them may bear no resemblance to the model of them that you have in your head. If your model is wrong at the meta-level—you are wrong about how people work—then you will either notice that you have difficulty dealing with people at all, or not notice that the problem is with you and get resentful at everyone else for not behaving as you expect them to.
Here, Mrs. B.F. Skinner imagines that she is reinforcing the behaviour that she desires, of eating spinach, by providing the reinforcer, ice-cream. Or is she really punishing the consumption of ice-cream by associating it with spinach? Or associating herself with an unpleasant situation? Or any number of other possibilities.
Sure thing. For me, it was the sudden realization that I had made assumptions from the very start of reading it, and that I had ranked certain outcomes far lower than the problem—taken in isolation—would justify.
When I read it, I immediately thought, “Okay, rewarding a kid for eating spinach, same ol’ same ol’ …”; then when I got to the end, I—very quickly—absorbed the insight that, in order for the process not to result in the child hating the mother, certain conditions have to hold, which are probably worthy of probing in depth.
I know all of this may sound obvious, but I really had an aha!/gotcha! moment on that one.
For you, understanding what was asked. The question is not, “what will happen?” The question is, “What information do you need in order to know which outcome will happen?”
Can someone explain why the parent is upvoted? Is everyone just assuming that the Bateson quote is just a sarcastic, roundabout way of asking what will happen?
ETA: In case you weren’t aware, cousin_it is not joking with his comment.
I downvoted you because you’re either completely missing the point of the quote, or you’re unsuccessfully trying to be funny.
In case it’s not the latter: Yes, since you already know the answer, it’s easy to “infer” the result from the givens. But the question is, what additional information are you using that constrains your answer to that? That’s what you need to say to solve it, not just repeat back from the answer key.
Furthermore, it’s not at all clear that children get the result you claim.
I don’t quite understand your objection. “Love mother” was an unconditional answer, yes. Most people love their mothers, even though the mothers did try to “shape” them in childhood with rewards and punishments. But “hate spinach” and “love ice cream” were inferred from the information in the question. The kid dislikes spinach, or the mother wouldn’t need to reward him; but he does like ice cream, or the mother wouldn’t use it as a reward. And I haven’t heard of any cases where the mother succeeded in “shaping” the kid’s food preferences like this.
If I’m not allowed to use real-life common sense, it’s not clear how I would even understand the question, let alone solve it. Okay, what additional information do you think one should need? Why?
If I’m not allowed to use real-life common sense, it’s not clear how I would even understand the question, let alone solve it. Okay, what additional information do you think one should need? Why?
Are you serious? The problem is to specify which “common sense” reasoning leads you to which conclusion! Yes, now that you’ve explained one reason why one outcome holds (even though it doesn’t account for children who grow up recenting their mothers and so isn’t even right on its own terms), you’ve given the kind of information the question is asking for.
Stating which outcome your common sense tells you would result—which is what you did—is non-responsive. And even now, you haven’t told what conditions determine the 27 possible outcomes—just one reason why one outcome would result.
Black-box “common sense” reasoning is exactly how you stray from rationality. You should open the box, and see what’s inside.
8 possible outcomes, not 27. But I think I see your point. Let me ask some more questions in this vein:
A man jumps off a 100 foot tall bridge. What additional information do you need to determine if he’ll die?
I have just washed my cup. What additional information do you need to determine if my cup is clean now?
What additional information do you need to determine whether the Sun will rise tomorrow?
If all such questions are effective in making you open your eyes, question your assumptions and upvote away, well, then I guess we’ll have to agree to disagree about the nature of “rationality”.
A man jumps off a 100 foot tall bridge. What additional information do you need to determine if he’ll die?
Within what degree of confidence? He could have a parachute of some form, or a bungee cord or there could be some form of trampoline to break the fall.
Moreover, you miss the point of the original quote. The question relies on standard assumptions about how humans learn and absorb values. Since humans are very complicated entities, understanding explicitly what assumptions we make about them can be helpful.
27 if you allow for “no effect”, which you should.
If all such questions are effective in making you open your eyes, question your assumptions and upvote away, well, then I guess we’ll have to agree to disagree about the nature of “rationality”.
It’s true that you can construct similar questions in other domains.
But the questions you posed are different from that in the quote because it refers to a:
-more common situation with a -more common inference that is -more often poorly grounded and hinges on complex aspects of human sociality, which are -more relevant to our everyday lives because of the -more frequent occurrence of similar situations.
The rationality issue involved in the quote is one of how you come to a conclusion, and I think it’s fair to say you might have missed some of the factors that come into play regarding manipulation of children, which Richard explains. There’s a difference between
a) “What does your gut tell you would happen?”, and b) “What information should you use to justifiably reach a conclusion about what would happen?”
You were answering a), while the question was asking b).
Understood; I’ve edited the GP comment to be more diplomatic and improve the formatting. Let me know what you think.
However, regarding the other comment, my question “Are you serious?” is an honest question. I don’t see how cousin_it could misinterpret the question as “What is X?” when it’s clearly asking “How do you know what X is?” So I don’t see why his answer of “X is …” got modded up.
I think I’d eventually come to hate ice cream in that kid’s situation. A treat is no longer a treat when it’s systematically used to manipulate you into eating something you hate.
I think it depends on how much you hate the spinach compared to how much you love the ice cream. People’s memories of an experience are strongly affected by the last bit, so the love of the ice cream may do quite a bit to overwrite the memory of hating spinach. Almost certainly not enough to affect feelings about spinach, but probably enough to not interfere with love of mother.
-- Gregory Bateson, “Steps to an Ecology of Mind”
Great quote! (Talk about be smacked with your own hidden assumptions...)
Got a link to the surrounding text?
Care to explain to me what you got out of it? I think I might be missing the point of this quote.
When you interact with someone, you may think, I will do this, so that they will do that, or think such-and-such, or feel thus-and-so; but what is actually going on for them may bear no resemblance to the model of them that you have in your head. If your model is wrong at the meta-level—you are wrong about how people work—then you will either notice that you have difficulty dealing with people at all, or not notice that the problem is with you and get resentful at everyone else for not behaving as you expect them to.
Here, Mrs. B.F. Skinner imagines that she is reinforcing the behaviour that she desires, of eating spinach, by providing the reinforcer, ice-cream. Or is she really punishing the consumption of ice-cream by associating it with spinach? Or associating herself with an unpleasant situation? Or any number of other possibilities.
Sure thing. For me, it was the sudden realization that I had made assumptions from the very start of reading it, and that I had ranked certain outcomes far lower than the problem—taken in isolation—would justify.
When I read it, I immediately thought, “Okay, rewarding a kid for eating spinach, same ol’ same ol’ …”; then when I got to the end, I—very quickly—absorbed the insight that, in order for the process not to result in the child hating the mother, certain conditions have to hold, which are probably worthy of probing in depth.
I know all of this may sound obvious, but I really had an aha!/gotcha! moment on that one.
On Google Books (limited online availability, but search for “habitually”).
If you Google the title, you’ll find the full text on a Brazilian website; whether legally or not I don’t know.
See also http://lesswrong.com/r/discussion/lw/dxm/clarification_behaviourism_reinforcement/
Hate spinach, love ice cream, love mother. What’s so difficult?
For you, understanding what was asked. The question is not, “what will happen?” The question is, “What information do you need in order to know which outcome will happen?”
Can someone explain why the parent is upvoted? Is everyone just assuming that the Bateson quote is just a sarcastic, roundabout way of asking what will happen?
ETA: In case you weren’t aware, cousin_it is not joking with his comment.
How dare you dispel Deep Wisdom of Master Bateson?
I try not to downvote people when they are right.
Nor, apparently, when they’re not even wrong. cousin_it’s reply was non-responsive.
Thanks. I was surprised to be downvoted, but decided not to ask why.
I downvoted you because you’re either completely missing the point of the quote, or you’re unsuccessfully trying to be funny.
In case it’s not the latter: Yes, since you already know the answer, it’s easy to “infer” the result from the givens. But the question is, what additional information are you using that constrains your answer to that? That’s what you need to say to solve it, not just repeat back from the answer key.
Furthermore, it’s not at all clear that children get the result you claim.
I don’t quite understand your objection. “Love mother” was an unconditional answer, yes. Most people love their mothers, even though the mothers did try to “shape” them in childhood with rewards and punishments. But “hate spinach” and “love ice cream” were inferred from the information in the question. The kid dislikes spinach, or the mother wouldn’t need to reward him; but he does like ice cream, or the mother wouldn’t use it as a reward. And I haven’t heard of any cases where the mother succeeded in “shaping” the kid’s food preferences like this.
If I’m not allowed to use real-life common sense, it’s not clear how I would even understand the question, let alone solve it. Okay, what additional information do you think one should need? Why?
Are you serious? The problem is to specify which “common sense” reasoning leads you to which conclusion! Yes, now that you’ve explained one reason why one outcome holds (even though it doesn’t account for children who grow up recenting their mothers and so isn’t even right on its own terms), you’ve given the kind of information the question is asking for.
Stating which outcome your common sense tells you would result—which is what you did—is non-responsive. And even now, you haven’t told what conditions determine the 27 possible outcomes—just one reason why one outcome would result.
Black-box “common sense” reasoning is exactly how you stray from rationality. You should open the box, and see what’s inside.
8 possible outcomes, not 27. But I think I see your point. Let me ask some more questions in this vein:
A man jumps off a 100 foot tall bridge. What additional information do you need to determine if he’ll die?
I have just washed my cup. What additional information do you need to determine if my cup is clean now?
What additional information do you need to determine whether the Sun will rise tomorrow?
If all such questions are effective in making you open your eyes, question your assumptions and upvote away, well, then I guess we’ll have to agree to disagree about the nature of “rationality”.
Within what degree of confidence? He could have a parachute of some form, or a bungee cord or there could be some form of trampoline to break the fall.
Moreover, you miss the point of the original quote. The question relies on standard assumptions about how humans learn and absorb values. Since humans are very complicated entities, understanding explicitly what assumptions we make about them can be helpful.
Thanks—point taken.
27 if you allow for “no effect”, which you should.
It’s true that you can construct similar questions in other domains.
But the questions you posed are different from that in the quote because it refers to a:
-more common situation with a
-more common inference that is
-more often poorly grounded and hinges on complex aspects of human sociality, which are
-more relevant to our everyday lives because of the
-more frequent occurrence of similar situations.
See also Richard’s further remarks.
The rationality issue involved in the quote is one of how you come to a conclusion, and I think it’s fair to say you might have missed some of the factors that come into play regarding manipulation of children, which Richard explains. There’s a difference between
a) “What does your gut tell you would happen?”, and
b) “What information should you use to justifiably reach a conclusion about what would happen?”
You were answering a), while the question was asking b).
Silas, you’re making strong arguments but mixing in emotion that makes it harder for your interlocutor to change their mind.
Understood; I’ve edited the GP comment to be more diplomatic and improve the formatting. Let me know what you think.
However, regarding the other comment, my question “Are you serious?” is an honest question. I don’t see how cousin_it could misinterpret the question as “What is X?” when it’s clearly asking “How do you know what X is?” So I don’t see why his answer of “X is …” got modded up.
I think I’d eventually come to hate ice cream in that kid’s situation. A treat is no longer a treat when it’s systematically used to manipulate you into eating something you hate.
I think it depends on how much you hate the spinach compared to how much you love the ice cream. People’s memories of an experience are strongly affected by the last bit, so the love of the ice cream may do quite a bit to overwrite the memory of hating spinach. Almost certainly not enough to affect feelings about spinach, but probably enough to not interfere with love of mother.