This accurately describes how my brain works. My thought process is like a bunch of graph nodes, and when new information doesn't "fit", it puts tensions on the links, and I want to resolve that tension. I can...feel it happening inside my mind when I think, more or less? -- It's hard to describe
Resolving that tension may occur in several ways, in order of increasing significance:
- Rejecting the new information
- Refining the graph (splitting a node representing a concept into multiple sub-nodes representing sub-concepts with their own relationships)
- Making local modifications to the graph
- Making sweeping architectural changes to the graph as a whole
The author seems to imply that cognitive biases are an inherent qualitative problem that is fundamentally forced to arise from this graph structure. I personally respectfully disagree. In my view, cognitive biases are a quantitative problem, incorrectly setting the threshold at which a large reorganization should occur. "Extraordinary claims require extraordinary proof" is qualitatively a sound epistemological principle -- but to correctly apply it, you must quantitatively set a reasonable threshold for "extraordinary."
I feel like we need to get better at understanding the graph structure of people we disagree with. The best example I can think of is the abortion debate [1]: If you accept the premise "Life begins at conception" [2], the pro-life camp has an enormously strong case; the rest of the graph between that premise and "Abortion should be illegal" is very strong (it's mostly tremendously well-reinforced nodes in near-universal moral foundations, like "Do unto others" or "Murder should be illegal").
Arguments against abortion are frequently just bad when looked at from the graph point of view: They often don't directly confront the premise "Life begins at conception," nor do they attack the graph between the premise and conclusion. [3]
[1] I'm personally in the pro-choice camp; I do not accept the premise that a human fetus has the same moral status as a fully grown human.
[2] "Life" here is not in the technical biological sense, but something more akin to "The ethical standing of human-equivalent sentience." (Bacteria and protozoa and so on are biologically alive, but nobody moralizes about killing them en masse by, e.g., cooking your food.)
You're absolutely right about the abortion debate; the entire pro-life argument makes sense only if you deny that "life begins at conception". But I don't believe I've ever heard the pro-choice camp attack that axiom; instead, it's always about the mother having control over her body, which just ignores the pro-life axiom. (If life begins at conception, then abortion is not about the mother's body, it's about the pre-born's body.) No wonder the two sides don't communicate.
Resolving that tension may occur in several ways, in order of increasing significance:
- Rejecting the new information
- Refining the graph (splitting a node representing a concept into multiple sub-nodes representing sub-concepts with their own relationships)
- Making local modifications to the graph
- Making sweeping architectural changes to the graph as a whole
The author seems to imply that cognitive biases are an inherent qualitative problem that is fundamentally forced to arise from this graph structure. I personally respectfully disagree. In my view, cognitive biases are a quantitative problem, incorrectly setting the threshold at which a large reorganization should occur. "Extraordinary claims require extraordinary proof" is qualitatively a sound epistemological principle -- but to correctly apply it, you must quantitatively set a reasonable threshold for "extraordinary."
I feel like we need to get better at understanding the graph structure of people we disagree with. The best example I can think of is the abortion debate [1]: If you accept the premise "Life begins at conception" [2], the pro-life camp has an enormously strong case; the rest of the graph between that premise and "Abortion should be illegal" is very strong (it's mostly tremendously well-reinforced nodes in near-universal moral foundations, like "Do unto others" or "Murder should be illegal").
Arguments against abortion are frequently just bad when looked at from the graph point of view: They often don't directly confront the premise "Life begins at conception," nor do they attack the graph between the premise and conclusion. [3]
[1] I'm personally in the pro-choice camp; I do not accept the premise that a human fetus has the same moral status as a fully grown human.
[2] "Life" here is not in the technical biological sense, but something more akin to "The ethical standing of human-equivalent sentience." (Bacteria and protozoa and so on are biologically alive, but nobody moralizes about killing them en masse by, e.g., cooking your food.)
[3] If you're curious about my own views on this specific subject, I've talked about them here before: https://news.ycombinator.com/item?id=36255493#36270990