Or maybe, just like all the symbolism, it was chosen at random because it looked cool. More work went into this analyses than the original font selection process.
That's sort of how art works, though. You hone your sensibility as an artist, and then you 'know it when you see it'. Anno et al looked at the fonts, and they saw it was good. Explaining why it's good is quite another thing altogether. Artists are always being surprised by how things 'click' or how the viewer notices patterns the artists weren't even (consciously) aware of putting in. (And to a certain degree, if you could adequately explain and convey it in words, why would you bother with the art in the first place?)
I've been using NNs to generate poems and Irish music lately based on pairwise comparison of poems/music (based on https://openai.com/blog/fine-tuning-gpt-2/ ), and it's definitely striking how fast I can make the evaluation of which one is better, but then how long it would take to reread or relisten to them and verbalize what exactly made one better than the other.
Apparently Anno is particularly passionate about Matisse EB. He uses it so many other places outside of NGE that it's become not just a "NGE font" but also a "Hideaki Anno font".
It probably wasn’t chose at _random_ even if there wasn’t a lot of thought that went into it. It was still an artistic decision.
When I used to DJ a lot, I would go to the record store and listen to dozens of records over the course of an hour or two and buy a handful of them. If you asked me in the store why I bought them, I might just say, because I liked it. Sometimes I had only listened to 30 seconds of it and that was enough.
But ‘I liked it’ hides a lot of inputs. I might have liked it because it reminded me of another song I liked, or because it had a key change that felt emotional or it had a beat that fit in with another record I liked, and if you asked me to dig into my reasoning for any of them at length, I could have probably come up with a few aspects that fed into why I bought it. In most cases they’re just post hoc rationalizations because when you’re digging through all the new releases you don’t have time to think of that stuff, but that doesn’t mean they’re not legitimate explanations.
Most difficult things work like this. The top performers in many complex tasks operate from pure intuition, and backtracking what their choices actually mean is a much more time intensive process.
Doing things quickly and just because the author feels like it's a good move does not necessarily mean the decision is not based on decades of experience and preliminary work.
