Orion Montoya is a user on mastodon.xyz. You can follow them or interact with them if you have an account anywhere in the fediverse. If you don't, you can sign up here.
Orion Montoya @mdcclv

In this piece a few weeks ago Martha Nussbaum calls something a “a generative lie”. It is such a perfect formulation I had to google it to see if it was a Known Term of Critique. It’s not! nyti.ms/2yW8L5d

The idea: that there are lies that, if you believe them, not only make you wrong about that proposition, but also lead you to generate new beliefs that are also wrong.

Then my mom was telling me about a Very Bad Talk that she went to about raising children. She said that there were audience members asking really engaged questions about the speaker’s bullshit. I realized: “they’re trying to generate new knowledge based on the dude’s principles!”

I was introduced to the idea of “the generative test” through linguistics (syntax). When you’re trying to account for linguistic evidence, you build a model to describe how that evidence (utterance, sentence, etc.) may have been generated, and then you test if it can also generate stuff that is not grammatical.

It’s really cool to see people applying generative tests to ideas, both good and bad ones, to see if they lead to more knowledge or insightful propositions, or if they lead to false propositions.

It’s really scary to see people apply faulty chains of reasoning to generate new, bad ideas, which they generated through induction, but which have falsehoods that lead them to generate more and more shitty thinking.

A bunch of the things that are fucked up in our societies are entrenched because they are founded in generative lies that led to a bunch of other wrongnesses.

Oh shit am I one of those people now?

(I mean: have I naive-inductioned myself into a Rationalist? What are the pitfalls of a noob logician thinking that the solution is more rigorous logical models?)

@mdcclv the trick is probably to carefully back away from the generative properties of rough but interesting models once you've gleaned an insight or two.

instead of, i dunno, founding a wiki with an associated meetup or something.

@mdcclv i have watched this shit in action _so many times_.

(usually in the course of business meetings with trainers or consultants in attendance.)