Reasoning From First-Principles v. Inherited Wisdom
·
I’ve been struck in recent years by a push–most availably from the Elon Musk genre of person–to reason from first principles. This on the one hand strikes me as good because I immediately can list off a litany of problems with reasoning from \(n\)th principles where \(n \neq 1\).
Inherited conventional wisdom is hard to buck against because of availability biases and social pressures (how’s that for a conventional wisdom about conventional wisdoms?). But of course history is replete with examples of how conventional wisdom was wrong and it took some maverick thinker to unsheath our blinded eyes. Geoffrey Hinton said “The future depends on some graduate student who is deeply suspicious of everything I have said.”
Reasoning by analogy is another example of relying on what you’ve learned about one situation to reason about another situation where this learned logic might not apply. So if you reason about what to do with your neighbor who annoys you with their loud music by analogizing “well what would you do if a mosquito was buzzing by your ear?”, we can probably all agree that you’ll want to find a different analogy.
So we rely on the maverick thinker to understand arguments down to their first-principle constituent parts and to think through things as most appropriate for the situation, not ones that are in some dimension semantically similar and in some other crucial dimension categorically different (e.g., an annoying mosquito versus an annoying neighbor). This makes me think of a framework for making arguments for whom I’ve long harbored a crush.
Think about if you decomposed arguments into their atomic bits, bits that could be understood by basically any high-school-educated English-speaker.
Well, what if there were a framework where things were argued or explained in terms of other things? For example, you can argue that the world is 4.5 billion years old (and you probably do treat this as a perfectly unassailable fact). Okay, well, based on what?
Well, based on the fact that radiometric dating shows the age of a rock based on breakdown of radioactive elements. Isotopes are variations of an element based on number of neutrons in nuclei. The isotopes of unstable radioactive elements–parent isotopes–turn into daughter isotopes in a predictable amount of time called a half-life.
But how do we know about half-lives? And on and on and on until you get down to the level of atomicity such that the argument hinges not on isotopes but on “If you look really close, you can see the number of neurons.” at which point someone’s gonna go “ahhhh okay I get it” and it will feel like a solid argument to them.
So I acknowledge that I like this because maverick thinkers seem to require understanding of first-principles and so if you could get more understanding, which I think tech is perfectly able to facilitate, since we’re no longer constrained to rigid forms of arguments but we can see arguments expanded and collapsed and animated dynamically as a person sees fit because of things like javascript and a browser instead of a printed book.
And I also have to acknowledge that I think like this because I live in a democracy, and I want individuals to understand the decisions that they’re chipping in to make, and I don’t believe that right now they do.
