What is understanding?
January 27, 2006 8:40 AM Subscribe
According to Leibniz, a theory must be simpler than the data it explains. In other words (for the nerds amongst us), understanding equals compression.
I challenge you to help me understand this via a thought-experiment.
Let's take two experiences we are all familiar with: Pleasure and Pain.
Explain the difference between these two experiences.
There is (of course) no "correct" answer; however, I ask that you attempt to use as few words as possible to explain your answer.
No need to spend too much time/effort pondering your response. What pops into your head first should prove itself the easiest to explain.
Let's take two experiences we are all familiar with: Pleasure and Pain.
Explain the difference between these two experiences.
There is (of course) no "correct" answer; however, I ask that you attempt to use as few words as possible to explain your answer.
No need to spend too much time/effort pondering your response. What pops into your head first should prove itself the easiest to explain.
This post was deleted for the following reason: askme is not for thought experiements or meta questions, what is the problem to be solved here?
Pleasure and pain are imprecise words describing subjective experiences. If you limit the domain to physical sensations, you can posit a hypothesis that pleasurable and painful experiences each excite specific regions of the brain in particular ways.
You could test this hypothesis by sticking people in fMRI machines, exposing them to various sensations, tracking the resulting brain activity, and asking them "would you say that's pleasurable or painful?"
Eventually, patterns of brain activity might emerge; if so, you'd be able to say "pleasurable experiences excite the XYZ region in this manner, painful experiences excite the ABC region in this manner." That's your theory. The data is the gigabytes of information you collected in testing it.
A theory doesn't just rehash data; it elucidates a pattern in it.
posted by adamrice at 8:51 AM on January 27, 2006
You could test this hypothesis by sticking people in fMRI machines, exposing them to various sensations, tracking the resulting brain activity, and asking them "would you say that's pleasurable or painful?"
Eventually, patterns of brain activity might emerge; if so, you'd be able to say "pleasurable experiences excite the XYZ region in this manner, painful experiences excite the ABC region in this manner." That's your theory. The data is the gigabytes of information you collected in testing it.
A theory doesn't just rehash data; it elucidates a pattern in it.
posted by adamrice at 8:51 AM on January 27, 2006
Pain hurts. Pleasure don't. Unless you're a freak.
Now THAT is compression! :)
posted by antifuse at 9:06 AM on January 27, 2006
Now THAT is compression! :)
posted by antifuse at 9:06 AM on January 27, 2006
"Pleasure rewards, pain punishes."
Pleasure and pain are the neurocognitive functions that have evolved in the brain to reward or punish behaviors that are likely to increase or decrease an animal's procreative success. Pleasure is the reward, and pain the punishment, and that's the only objective difference.
posted by nicwolff at 9:08 AM on January 27, 2006
Pleasure and pain are the neurocognitive functions that have evolved in the brain to reward or punish behaviors that are likely to increase or decrease an animal's procreative success. Pleasure is the reward, and pain the punishment, and that's the only objective difference.
posted by nicwolff at 9:08 AM on January 27, 2006
Response by poster: >Simpler doesn't necessarily mean smaller. [yesster]
I concede that this is true; but I have found that an economy of words will often lead to quality explanations.
> A theory doesn't just rehash data; it elucidates a pattern in it. [adamrice]
Wonderful feedback. This sentence also reinforces what I was looking for: few words, but I learnt a great deal.
posted by stungeye at 9:13 AM on January 27, 2006
I concede that this is true; but I have found that an economy of words will often lead to quality explanations.
> A theory doesn't just rehash data; it elucidates a pattern in it. [adamrice]
Wonderful feedback. This sentence also reinforces what I was looking for: few words, but I learnt a great deal.
posted by stungeye at 9:13 AM on January 27, 2006
You might be interested in the "Define blah in 10 words or less" [google.com] series that 37 Signals run(s) at their blog.
posted by misterbrandt at 9:13 AM on January 27, 2006
posted by misterbrandt at 9:13 AM on January 27, 2006
I'd say asking for the difference between pleasure and pain is much like asking for the difference between red and blue. They're primitives. No definitions are possible; hence, the differences cannot be articulated.
posted by bricoleur at 9:17 AM on January 27, 2006
posted by bricoleur at 9:17 AM on January 27, 2006
Response by poster: bricoleur: What would be a better question for this thought experiment? (I'm not being defensive here, just curious.)
posted by stungeye at 9:22 AM on January 27, 2006
posted by stungeye at 9:22 AM on January 27, 2006
One-byte difference.
You could say the pleasure/pain axis is a tristate input for a biological computer. -1 = pain, 0 = neither, 1 = pleasure.
"I sense injury. The data could be called pain."
posted by kindall at 9:23 AM on January 27, 2006
You could say the pleasure/pain axis is a tristate input for a biological computer. -1 = pain, 0 = neither, 1 = pleasure.
"I sense injury. The data could be called pain."
posted by kindall at 9:23 AM on January 27, 2006
pleasure and pain both are shorthand for (compressions of) complex chemical, biological environmental feedback loops. pleasure describing the subjective experience of taking part in a positive feedback loop and pain describing the subjective experience of taking part in a negative feedback loop.
The terms already represent compressions of more complex realities.
posted by jouke at 9:36 AM on January 27, 2006
The terms already represent compressions of more complex realities.
posted by jouke at 9:36 AM on January 27, 2006
stungeye, I think you just need to pick more complex phenomena to have explained. Pain and pleasure are already conceptually simple, so it isn't possible to simplify them further.
How about an explanation of the retrograde motion of the planets? Complex phenomenon in terms of data points, (relatively) simple geometric explanation.
posted by bricoleur at 9:49 AM on January 27, 2006
How about an explanation of the retrograde motion of the planets? Complex phenomenon in terms of data points, (relatively) simple geometric explanation.
posted by bricoleur at 9:49 AM on January 27, 2006
You could say the pleasure/pain axis is a tristate input for a biological computer. -1 = pain, 0 = neither, 1 = pleasure.
What about both?
posted by nomad at 9:50 AM on January 27, 2006
What about both?
posted by nomad at 9:50 AM on January 27, 2006
Pain and pleasure are the same thing.
It's your interpretation of them that differs.
posted by five fresh fish at 9:52 AM on January 27, 2006
It's your interpretation of them that differs.
posted by five fresh fish at 9:52 AM on January 27, 2006
Best answer: A theory is a tool, it's not something that's meant to explain things or provide a basis for philosophy. A tool is something that increases your abilities. If the theory is more complicated than the data, you are better off just using the data.
In your thought experiment, you defined a question, but not the data, so we have no way of determining if any proposed theory is more complicated or not.
posted by 445supermag at 9:58 AM on January 27, 2006
In your thought experiment, you defined a question, but not the data, so we have no way of determining if any proposed theory is more complicated or not.
posted by 445supermag at 9:58 AM on January 27, 2006
Why do you assume that words, specifically the English language, is the appropriate medium for stating this particular theory? Why not use numbers or paintings or, better yet, why not come here and let me slap you and then offer you some chocolate. (Actually, I won't share my chocolate with you. It's mine!)
Your "understanding equals compression" is a bad simplification of Leibniz's thought. Leibniz's desire to break everything down into the simplest possible terms was a metaphysical attempt (to prove God's existence) not a true scientific theory of understanding. One of these things is most definitely not like the other.
posted by nixerman at 10:12 AM on January 27, 2006
Your "understanding equals compression" is a bad simplification of Leibniz's thought. Leibniz's desire to break everything down into the simplest possible terms was a metaphysical attempt (to prove God's existence) not a true scientific theory of understanding. One of these things is most definitely not like the other.
posted by nixerman at 10:12 AM on January 27, 2006
I wouldnt go too far with the compression analogy. As 445supermag says, a theory is just a mental model in which we can make simplifications in order to make progress. In most cases, this means losing data. For example, in examining planetary motions, a basic Newtonian model of two interacting bodies (planet and sun) works rather well but the truth is that planets feel each other's gravity too.
Luckily the model is tolerant to these perturbations and still allows stable regimes where the theory works. So a lot of this is finding places where the Universe will tolerate stable regimes and finding good-enough rules that help us navigate complexity. Moral rules such as "Do unto others.." are still around because they provide decent approximations to the otherwise unfathomable complexity and dynamics of relationships. But they are not a "compression" of morality.
Rules and theories are good but dont put too much faith in them - just as much as you need to. Einstein, of course, said it best: "everything should be made as simple as possible, but not simpler."
posted by vacapinta at 10:13 AM on January 27, 2006
Luckily the model is tolerant to these perturbations and still allows stable regimes where the theory works. So a lot of this is finding places where the Universe will tolerate stable regimes and finding good-enough rules that help us navigate complexity. Moral rules such as "Do unto others.." are still around because they provide decent approximations to the otherwise unfathomable complexity and dynamics of relationships. But they are not a "compression" of morality.
Rules and theories are good but dont put too much faith in them - just as much as you need to. Einstein, of course, said it best: "everything should be made as simple as possible, but not simpler."
posted by vacapinta at 10:13 AM on January 27, 2006
I disagree with your premise. The main value of theories is not the abstraction of data into principles that they may attempt, but their power in so doing to help us link and organize other data and theories into a corpus, which can then be conveniently "walked" both from the specfic to the general (induction), and from the general to the specific (deduction). No point in developing an infinity of disconnected theories over an infinity of data sets, if the theories don't build into something still greater and more abstract, that greater thing being human knowledge.
posted by paulsc at 10:17 AM on January 27, 2006
posted by paulsc at 10:17 AM on January 27, 2006
This thread is closed to new comments.
posted by Leon at 8:51 AM on January 27, 2006