Dunning-Kruger effect - about how to think

[sonia]

A young woman, 20-something, bright, said more people get cancer recently because evolution tends toward entropy. I said, uh, no, bad mutations get eliminated over time, and she said well, the Bible says we used to live so much longer, and I said, I like my woo and science kept separate, and also the Bible and science. The Bible is not historical! And she said we'll have to agree to disagree on that, and I was sputtering internally for a good half hour.

She attended public schools in a small town in Oregon, not religious school as I would have guessed. I've run into this wall with her before, where she thought she could assert something as fact because she felt like it, or change the meaning of a word, and now I see the underpinnings. I can see that I know so much more than she does (about how to *think*!), and she can't, and she's asserting her view as valid. Microcosm of our political situation. Reality bats last...

I did mention the dinosaur feathers in amber. (So cool!)

I wonder if in twenty years or so she'll be sadder and wiser, or if she'll continue to be hermetically sealed against actual thought. She can't *learn* with her current framework!

Comments

[tim]

I think being good at programming and debugging -- which is to say, symbol manipulation -- doesn't save you from having bad axioms.

[sonia]

Yes.

In the case of my friend, it's the "assert something as fact because she felt like it" and not being open to new information that I found jarring. Computer programs don't play that game.

[suncat]

Yup, yup.

Exactly what Tim and you have now said. My colleagues reasoning faculties were just fine.

But by rejecting some of the base premises and evidence of the sciences, because they didn't intersect their daily life, they were free to hare off in their own favored lines of thought.

[sonia]

What I'm trying to say is that rejecting evidence is not a useful practice when debugging computer programs. It's not all about reasoning. It's about noticing and questioning assumptions about what is and isn't happening, what does and doesn't matter.

I find that tracking down difficult bugs in complex software is classically scientific. Formulate a theory, run experiments, take notes, keep updating the theory according to the evidence. Also, come to think of it, doing web searches for other people's experiences, which parallels reading scientific research papers.

[suncat]

Ah, yes. I see your point.

I will say that in all of my years of debugging, I have yet to run across evidence in the code that I find personally or morally repugnant ^_^. So I haven't found need to reject it.