Here’s a bunch of analogies, some good, some not so good: 1. People in the West often think of ramen as a defining Japanese food, but that’s not how the Japanese see it. Ramen only arrived in Japan in 1910 when a shop in Tokyo
Interesting read. I think analogies reflect important parts of human cognition esp. related to learning. If it wasn’t for our ability to extrapolate, to infer from one instance to others, to explain the unknown in/with terms of the known, applying analogies that is, we wouldn’t have come so far (or short) of understanding the universe as we did. The really interesting part to me is: how do we know to what we can apply what analogy, how do we perceive/know off the shared characteristics/structure that make analogies applicable and useful (especially when what I am applying the analogy to is as of yet unknown to me, e.g. the mechanical analogies of the body in the 19th century, or the computer/brain analogies in the 21st). What is that faculty that tells us that some similarities exist that could be exploited by analogies?
A support for the idea that analogies are fundamental is that when there really is no good analogy (like quantum mechanics?) it seems to drive us a bit crazy.
I figure the inadequacy of analogies is an essential ingredient: after all, you want something explained in terms of what it is not. I’m not even sure a perfect analogy would do any work at all, it would presumably be plain & obvious and add nothing to the understanding of its analogue.
Yes. Abortion is another topic for which there are no good analogies. Philosophers have tried for decades to find one, and the old Judith Jarvis Thomson article about being connected to the famous violinist for 9 months is the best anyone's come up with. Back in the 80s, when I was in grad school, you were scolded if you couldn't find a way to express your thoughts completely literally.
I agree that using analogies in debates can be extremely helpful. I do think that picking apart the analogy can be valuable as well. The shared topology is a useful framework, but being able to examine the differences can be incredibly valuable for understanding why someone disagrees with you. In your India-US example, I would describe this as saying something like "but what if the US invented dishwashers during the party". Big caveat is that its not clear my debate partners feel this way.
Regarding if your debate partners feel that way... My instinct is that analogies typically are much more useful for "explaining" stuff than "debating" stuff. It's hard to draw a clear line between the two, but my instinct is that once you reach a point that you're debating the analogy rather than the original subject, you should pretty aggressively drop the analogy.
I think the better AGI:Mars analogy would be “Worrying about AGI is like worrying about climate change on Mars.”
Planning for adversarial AGI is like planning a pandemic response in advance: humans are terrible at it.
Interesting read. I think analogies reflect important parts of human cognition esp. related to learning. If it wasn’t for our ability to extrapolate, to infer from one instance to others, to explain the unknown in/with terms of the known, applying analogies that is, we wouldn’t have come so far (or short) of understanding the universe as we did. The really interesting part to me is: how do we know to what we can apply what analogy, how do we perceive/know off the shared characteristics/structure that make analogies applicable and useful (especially when what I am applying the analogy to is as of yet unknown to me, e.g. the mechanical analogies of the body in the 19th century, or the computer/brain analogies in the 21st). What is that faculty that tells us that some similarities exist that could be exploited by analogies?
A support for the idea that analogies are fundamental is that when there really is no good analogy (like quantum mechanics?) it seems to drive us a bit crazy.
I figure the inadequacy of analogies is an essential ingredient: after all, you want something explained in terms of what it is not. I’m not even sure a perfect analogy would do any work at all, it would presumably be plain & obvious and add nothing to the understanding of its analogue.
Yes. Abortion is another topic for which there are no good analogies. Philosophers have tried for decades to find one, and the old Judith Jarvis Thomson article about being connected to the famous violinist for 9 months is the best anyone's come up with. Back in the 80s, when I was in grad school, you were scolded if you couldn't find a way to express your thoughts completely literally.
I love it, the ideas are presented beautifully, thank you.
I agree that using analogies in debates can be extremely helpful. I do think that picking apart the analogy can be valuable as well. The shared topology is a useful framework, but being able to examine the differences can be incredibly valuable for understanding why someone disagrees with you. In your India-US example, I would describe this as saying something like "but what if the US invented dishwashers during the party". Big caveat is that its not clear my debate partners feel this way.
Regarding if your debate partners feel that way... My instinct is that analogies typically are much more useful for "explaining" stuff than "debating" stuff. It's hard to draw a clear line between the two, but my instinct is that once you reach a point that you're debating the analogy rather than the original subject, you should pretty aggressively drop the analogy.