[bbor]

I love this long detailed conversation with many people jumping in, and 0 references to philosophers of the mind… gee guys, I wonder how we could crack this code? Even the paper itself cites one cognitive psychologist then moves on! A bit of relatable intellectual arrogance from us SWEs/mathematicians, I think — we are “on top of the world” right now.

FWIW I think you in particular are exactly right. I always think of Schopenhauer’s quote, and I think any software engineer might appreciate it: human memory isn’t storing items received from the world in a database, it’s more like folding creases into a napkin so that it naturally tends to fall into that shape in the future. In other words: remembering an event is equivalent to developing the skill of imagining a scene/dataframe that relates to that event.

In specific math terms: math is a collection of intellectual tools building on one another. You can certainly practice the ability to apply tools in new situations, but if you don’t also practice the ability to recall the tools themselves, it’s useless.

[fallingknife]

But is that actually what human memory is like? AFAIK nobody actually understands the internals. The "philosophers of the mind" who claim to know are the ones guilty of arrogance, not those who don't cite them.

[bbor]

Well, we should collect some evidence and write a book! If we did, it would be filed into the philosophy of mind section, I believe ;)

We don’t know everything, but we have more evidence than “it’s a black box” - in fact, that’s basically the scholastic / Aristotlean view that was conquered by our friends Bacon, Hume and Kant a few hundred years ago.