One of the enduring insights I’ve gotten from elityre is that different world models are often about the weight and importance of different ideas, not about how likely those things are to be true. For instance, The Elephant in the Brain isn’t about whether or not signalling exists, its’ about how central signalling is to the worldview of Simler and Hanson. Similarly with Antifragility and Nassim Taleb.
One way to say this is that disagreement is often about the importance of an idea, not its’ truth.
Another way to say this is that worldview differences are often about the centrality and interconnectedness of a node within a graph, and not its’ existence.
A third way to say this is that disagreements are often about tradeoffs, not truths.
I’ve used all of these when trying to point to this idea, but I’d like a single, catchy word or phrase to use and a blog post I can point to so that this idea can enter the rationalist lexicon. Does this blogpost already exist? If not, any ideas for what to name this?
Yeah. This problem is especially bad in politics. I’ve been calling it “importance disagreements”, e.g. here and here. There’s no definitive blogpost, you’re welcome to write one :-)
Note that I think we’re talking about similar things, but have slightly different framing. For instance, you say :
I’ve had similar thoughts but formulated them a bit differently. It seems to me that most people have the same bedrock values, like “pain is bad”. Some moral disagreements are based on conflicts of interest, but most are importance disagreements instead. Basically people argue like “X! - No, Y!” when X and Y are both true, but they disagree on which is more important, all the while imagining that they’re arguing about facts. You can see it over and over on the internet.
I think “Value Importance” disagreements definitely do happen, and Ruby talks about them in “The Rock and the Hard Place”.
However, I’m also trying to point at “Fact Importance” as a thing that people often assume away when trying to model each other. I’d even go as far to say that often what seems like “value importance” intractable debates are often “hidden assumption fact importance debates”.
For instance, we might both have the belief that signalling effects peoples’ behaviors, and the belief that people are trying to achieve happiness, and we both assign moderately high probability on each of these factors. However, unless I understand, in their world model, how MUCH they think signalling effects behaviors in comparison to seeking happiness, I’ve probably just unknowingly imported my own importance weights onto those items.
Any time you’re using heuristics (which most good thinkers are) its’ important to go up and model the meta-heuristics that allow you to choose how much a given heuristic effects a given situation.
Yeah, I guess I wasn’t separating these things. A belief like “capitalists take X% of the value created by workers” can feel important both for its moral urgency and for its explanatory power—in politics that’s pretty typical.
Just wanted to quickly assert strongly that I wouldn’t characterize my post cited above as being only about value disagreements (value disagreements might even be a minority of applicable cases).
Consider Alice and Bob who are aligned on the value of not dying. They are arguing heatedly over whether to stay where they are vs run into the forest.
Alice: “If we stay here the axe murderer will catch us!”
Bob: “If we go into the forest the wolves will eat us!!”
Alice: “But don’t you see, the axe murderer is nearly here!!!”
Same value, still a rock and hard place situation.
Similarly, we might both agree on the meta-heuristics in a specific situation, but I have models that apply a heuristic to 50x the situations that you do, so even though you agree that the heuristic is true, you disagree on how important it is because you don’t have the models to apply it to all the situations that I can.
If you make it explicit like “X is important” vs “X is not important” I have hard time to use the word “disagree” on it. Like if A and B emphasis and have signaling as similarly central in their worldviews saying “we agree on signaling” sounds wrong. Also saying stuff like “I disagree with racism” sounds like a funky way to get that point across.
I think disagree is not semantically accurate for the thing I’m trying to point at, but it still feels internally often like “We have a fundamental disagreement about how to view this situation”, it make more sense to talk about “our models being in agreement” than us being in agreement.
One of the enduring insights I’ve gotten from elityre is that different world models are often about the weight and importance of different ideas, not about how likely those things are to be true. For instance, The Elephant in the Brain isn’t about whether or not signalling exists, its’ about how central signalling is to the worldview of Simler and Hanson. Similarly with Antifragility and Nassim Taleb.
One way to say this is that disagreement is often about the importance of an idea, not its’ truth.
Another way to say this is that worldview differences are often about the centrality and interconnectedness of a node within a graph, and not its’ existence.
A third way to say this is that disagreements are often about tradeoffs, not truths.
I’ve used all of these when trying to point to this idea, but I’d like a single, catchy word or phrase to use and a blog post I can point to so that this idea can enter the rationalist lexicon. Does this blogpost already exist? If not, any ideas for what to name this?
A Matter of Degree
Yeah. This problem is especially bad in politics. I’ve been calling it “importance disagreements”, e.g. here and here. There’s no definitive blogpost, you’re welcome to write one :-)
Note that I think we’re talking about similar things, but have slightly different framing. For instance, you say :
I think “Value Importance” disagreements definitely do happen, and Ruby talks about them in “The Rock and the Hard Place”.
However, I’m also trying to point at “Fact Importance” as a thing that people often assume away when trying to model each other. I’d even go as far to say that often what seems like “value importance” intractable debates are often “hidden assumption fact importance debates”.
For instance, we might both have the belief that signalling effects peoples’ behaviors, and the belief that people are trying to achieve happiness, and we both assign moderately high probability on each of these factors. However, unless I understand, in their world model, how MUCH they think signalling effects behaviors in comparison to seeking happiness, I’ve probably just unknowingly imported my own importance weights onto those items.
Any time you’re using heuristics (which most good thinkers are) its’ important to go up and model the meta-heuristics that allow you to choose how much a given heuristic effects a given situation.
Yeah, I guess I wasn’t separating these things. A belief like “capitalists take X% of the value created by workers” can feel important both for its moral urgency and for its explanatory power—in politics that’s pretty typical.
Depends on the value of X.
Just wanted to quickly assert strongly that I wouldn’t characterize my post cited above as being only about value disagreements (value disagreements might even be a minority of applicable cases).
Consider Alice and Bob who are aligned on the value of not dying. They are arguing heatedly over whether to stay where they are vs run into the forest.
Same value, still a rock and hard place situation.
Similarly, we might both agree on the meta-heuristics in a specific situation, but I have models that apply a heuristic to 50x the situations that you do, so even though you agree that the heuristic is true, you disagree on how important it is because you don’t have the models to apply it to all the situations that I can.
If you make it explicit like “X is important” vs “X is not important” I have hard time to use the word “disagree” on it. Like if A and B emphasis and have signaling as similarly central in their worldviews saying “we agree on signaling” sounds wrong. Also saying stuff like “I disagree with racism” sounds like a funky way to get that point across.
I think disagree is not semantically accurate for the thing I’m trying to point at, but it still feels internally often like “We have a fundamental disagreement about how to view this situation”, it make more sense to talk about “our models being in agreement” than us being in agreement.