Two major governance aspects here are totally unrelated to each other in terms of execution; layer 1 and layer 2's.
Yes, that is (hopefully) a major takeaway for folks who read this post and the comments. Layer 2 has the freedom to be governed any way the owner/founder of a given Layer 2 solution sees fit because of Layer 1's protections against individual accounts being deplatformed. As such, the governance method for Layer 1 is critical, and thus makes the governance methods for Layer 2 solutions essentially inconsequential.
Stated another way, on Layer 1, it is the method by which decisions are made that is consequential (critical, even), while on Layer 2, it is the decisions themselves that are consequential (not the methods) because bad decisions by a given Layer 2 solution will merely send people to 'better' Layer 2 solutions.
The key point is, on layer 2, you no longer need to worry about censorship resistance of the base layer, IE your metadata (account, community list, etc.), and most importantly, your token balance.
This is sorta like the concept of federalism. If you have a federal government that guarantees your right to freely move from state to state, then the states themselves have to compete for residents. As such, the governance method at the federal level is paramount, making the governance method at the state level secondary, or even inconsequential. For example, in the U.S. if a given state chose to elect a dictator and operate as a dictatorship, that would not be a big deal because citizens of that state are free to move to a different state. However, dictatorial powers at the federal level severely diminish the rights of all the citizens, irrespective of which state they live in.