Emergence is cool. It's what happens when things interact in new ways and produce unexpected results. But it ain't the only game in town. Cosmologist George Ellis explains in this lecture* how complex systems do a dance of give-and-take between higher level, top-down causation and bottom up emergence.
We all like for our world to make sense, so when it gets messy and unclear we look for how it can come back to order, and one of the best places to look for change is the surface -- the day to day stuff, like us humans interacting with our environment, the disorder itself. Emergence.
But Ellis explains that there's a ceiling to how much complexity you can get from bottom-up emergence before you once again encounter top-down causation. In other words there's a limit to how much sway emergence (including innovation) has on a system before built-in mechanisms kick in and reinforce the higher structures upon the lower.
This happens in even the most innovative cultures and the flattest workplaces. Eventually, a higher order such as your founder's directive, your budget, prevailing attitudes or regulations come to bear on what can happen next. And that's well and good because if no one ever drew boundaries anywhere to interact with what emerged, the whole works would fall prey to entropy.
Higher levels of order, which tend to be autonomous and more static, draw on the randomness of lower orders to create new possibility, but they aren't controlled by them. (Example: higher order - the organization's mission; lower order - the daily moods and inspiration levels of your content creators).
I think eureka moments occur when, as if by magic, selections from the lower and higher orders seem to arise mutually. This is where a complex system finds resilience, the dry land in an ocean of possibilities. In a dynamical system, resilience to small perturbations can lead to a kind of stability, or coherence. The hierarchies of top-down causality and bottom-up emergence that make up a complex system are working together to produce possibilities that we can begin to identify. In other words, stuff starts to make sense.
In complexity-savvy organizations, we hopefully build top-down hierarchies to be flexible, allowing random innovations to help us periodically revise the higher-level structures that steer the ship, but we don't become so enamored with emergence that we sail completely off course in search of sirens.