We all know confidently incorrect people. People displaying dunning-kruger. The majority of those people have low education and without someone giving them objectively true feedback on their opinions through their developmental years, they start to believe everything they think is true even without evidence.
Memorizing facts, dates, and formulas aren’t what necessarily makes someone intelligent. It’s the ability to second guess yourself and have an appropriate amount of confidence relative to your knowledge that is a sign of intelligence.
I could be wrong though.
I am a flight instructor. I had to study the fundamentals of instruction to earn that title, so I believe I can speak with some authority on this subject.
When discussing facts, figures and such, we consider four levels of learning. The easiest, fastest and most useless is rote memorization. Rote memorization is the ability to simply parrot a learned phrase. This is fast and easy to achieve, and fast and easy to test for, so it’s what schools are highly geared toward doing.
An example from flight school: A small child, a parrot, and some Barbie dolls could be taught that “convective” means thunderstorms. When a meteorologist says the word “convective” it’s basically a euphemism for thunderstorms. You’ve probably already memorized this by rote. You would correctly answer this question on the knowledge test:
Which weather phenomenon is a result of convective activity?
A. Upslope Fog
B. Thunderstorms
C. Stratus Clouds
Okay, what should a pilot do about thunderstorms? Are they bad? What about a thunderstorm is bad? A student who can answer those questions, who can explain that thunderstorms contain strong turbulence and winds that can break the airplane or throw it out of control have reached the Understanding level.
Problem: Sitting in the classroom talking about something is NOT flying a plane. I’ve had students who can explain why thunderstorms are dangerous fly right toward an anvil-shaped cloud without a care in the world, because they didn’t recognize a thunderstorm when they saw one. Living in a forest, people around here don’t get a good look at them from the side; the sky just turns grey and it rains a lot and there’s bright flashes and booming noises. If you can get a good look at one, it’s a tremendously tall cloud that flattens out way up high and tends to have a bit that sticks out like the horn on an anvil. Even in the clear air under that horn you’ll get severe turbulence. A student that can identify a thunderstorm and steers to avoid it can Apply their knowledge, and have thus reached the Application level.
It’s a sign that you’re ready for your checkride if, upon getting a weather briefing that includes convective activity, the student makes wise command decisions to either reschedule the flight for a day of safer weather, or for isolated storms plots a route that steers to the safe side of the weather and plans for contingencies such as turning back or diverting to alternates. A student that alters his navigational choices based on weather forecasts has reached the correlation level.
It’s difficult to go beyond the understanding level in a classroom with textbooks and paper tests, which is too much of what K-12 and college is like.
Love this comment. If anyone knows anything about machine learning or brains, this resembles modal limitations in learning.
A lot of our intelligence is shaped around our sensory experience, because we build tools for thinking via the tools we’ve already built, ever since baby motorbabbling to figure how our limbs work. Why Hellen Keller had such trouble learning, but once she got an interface she could engage with for communication, things took off.
We always use different tools, but some people don’t see colour. This doesn’t mean they are stupid when they answer differently in describing a rainbow.
Also why llms struggle with visual/physical concepts if the logic requires information that doesn’t translate through text well. Etc.
Point being, on top of how shitty memorization is as the be all end all, learning and properly framing issues will have similar blindspots like not recognizing the anvil cloud.
This is also why people in informational bubbles can confirm their own model from ‘learning’ over people’s lives experiences.
Like most issues, it doesn’t mean throwing the baby out with the bathwater, but epistemic humility is important, and it is important not to ignore the possibility of blindspots, even when confidence is high.
Always in context of the robustness of the framing around it, with the same rules applied at that level. Why “nothing about us without us” is important.
But also we gotta stop people giving high confidence to high dissonance problems, and socializing it into law. We should be past the “mmr causes autism” debate by now, but I’m hearing it from the head of health in the USA.
Ok, but you never finished the example with the child, parrot and barbie dolls. What is the punchline there?
Children can be taught to repeat something even if they don’t understand it. So can many species of parrot, they famously mimic sounds they hear including human speech without understanding the meaning behind the sounds. And I seem to remember a model of Barbie doll that had a little sound recorder built in so she can “really talk.” These things can repeat something they’ve "learned’ without any deeper understanding.