So, to add a slightly different perspective from outside games as food for thought....
In mathematics and theoretical physics, there's a word for much of what seems to be intended by the "meta-aesthetics". For us, the word is "elegance".
As you may imagine, a lot of the math in these areas can be really, really hairy and complicated. But, it turns out that most of the time, the general statements that turn out to be true are... elegant. Simple to state. They have symmetries, are smooth, and often fit in just a few lines. Einstein's Special and General theories of Relativity. Newton's equations of motion. Maxwell's Equations. Schrodinger's equation. The Laws of Thermodynamics. The basic statements of all these things are painfully simple. They only get hairy when you then layer them into the real world, where there are sharp corners and small details that make math hard. But the concepts? Elegant.
Elegance is part of the sniff-test for being true. It isn't sufficient, but inelegance is a thing that suggests that you don't have it all quite right yet.
For us in games, of course there's going to be a very reasonable press towards practicality. Maxwell's equations, and Einstein's, are elegant, but aren't practical for day-to-day use at home. However, there's perhaps a thing to be said for looking at the elegance of a design, looking at that meta-aesthetic, and seeing if maybe it is an indication of some true thing about play that you can then look to preserve as you find a practical implementation.
And if I were saying that elegance were
inherently bad, this would be a great argument against my position.
But I'm not saying that. I have
repeatedly and explicitly said that I'm not saying that. I have repeatedly said that "elegance" (using my own terms, of course) is a perfectly valid consideration and one that shouldn't be dismissed out of hand.
What I am saying is that "elegance" has been elevated above
empirical adequacy (as van Fraassen would put it). That equations which have nice properties like smoothness or absence of singularities or infinite differentiability, but which
sacrifice some amount of ability to accurately and precisely predict empirical data, are being prioritized over equations which
do accurately and precisely predict the empirical data. And, specifically, that this priority is being assigned
because the former equations are more elegant than the latter, and treating any challenge to that as self-evidently ridiculous.
Consider, for instance, the theories of light that held before Michelson–Morley. Based on common sense and comparison to all other waves of which we had knowledge (all mechanical waves) required a medium. It was, originally, shockingly
inelegant to suggest that there was some second, "new" kind of wave that didn't need a medium (but could still pass through one), hence the long-term sticking power of the aether models despite continual concerns and problems. The adherence to the "luminiferous aether" until all possible avenues of escape had been eliminated was
not driven by the empirical adequacy of such theories, which stumbled repeatedly and required multiple ad-hoc modifications to fix, but rather by the desire to avoid the loss of elegance. It resulted in quite a lot of wasted time and, frankly, dodgy science--something that even its late-stage proponents (like Sir Oliver Lodge) had to admit looked bad.
That is what I am arguing against. Stridently advocating for a structure
regardless of any practical design consequences it has, on the basis that its aesthetics,
whatever those aesthetics may be, are superior. My concerns would apply just as much to a stridently pro-keyword approach, if anyone were actually advocating that, as it does to a stridently pro-centralization approach, which is commonly advocated both here and elsewhere. (Seriously,
every single thread anyone
ever starts about making a 4e heartbreaker, you're gonna get a third or more of the responses advocating for either all powers from a given source being condensed into a single pile, or all powers
in the whole game condensed into a single pile, not because that would be more effective to play nor because it would be easier to design, but because single piles are presumed to be self-evidently "better" than separate piles. When one naturally points out that build differences would become extremely difficult to implement in such a system--e.g. every power could easily have 4 or more build-specific clauses in it, making them incredibly bloated--these concerns are flatly dismissed as unimportant compared to the importance of collecting all plausibly-similar options together in a single list.)
Elegance may be a useful heuristic--or it may be a stumbling block to admitting that we really DO have more to learn about a subject. Pretending that it is
always a useful heuristic--or, as I'm asserting people are doing, pretending that it is not only useful but
more important than any other considerations unless they are unequivocally overwhelming--is not better than pretending it is totally useless and should always be ignored.
Elegance is good to have. All else being equal, it should be pursued.
We should never forget that all else is often NOT equal.