2022-12-21 Wed 13:05 PM ![[moses_tiny_robots_happily_build_a_large_edifice_in_their_worksh_99f9ada0-bbf1-43ea-810a-ea9e5b7885cb.png]] Might imagine writing as a two-stage process. In the first stage, you're developing the conceptual edifice. Building it, elaborating it, modifying it, refactoring it, etc. You do this via reading, writing, sketching, pondering, discussing, mind wandering, sleeping, editing, tinkering, etc. This stage tends to produce a lot of throw-away artifacts (but you might as well keep them if you have an easy way to store them). In the second stage, you [[Trust the langauge generator]] and let it write words that describe the conceptual edifice as it currently stands in a way that's appropriate for the writing context. These stages might get repeated many times before the thing is "done". And this happens in big loops and small loops. You might have the outlines of the macro edifice pretty well sketched out, but then struggle to phrase a certain sentence. You might need to go into a thinking phase to figure something out. If you're confounded while trying to generate language, that's a sign you have some learning to do. [[Relish confounded writing]]. This separation of stages can be helpful for combatting a tendency I notice in myself, which is to get attached to certain language formulations, thinking that I need to save them and use them, as if language generation is hard and rare. But this attachment strangles the process. Better, I think, to recognize that the edifice development phase produces lots of disposable content, and to trust that when it's time to write, the [[Writing team_matt|writing team]] will be able to handle it. > [!-cf-]+ [[Related notes]] > - [[Hub. Experimental writing process techniques (list)]].