The Gardener and the Architect
Two archetypes of building. Both are real. Neither is morally superior in the abstract -- the dichotomy is about which relationship with emergence you accept.
The Architect specifies. They know the end state and work backward to make the medium comply. The Gardener tends. They create conditions under which the medium produces what the Gardener could not have specified in advance.
Most of what is called "entrepreneurship" is Architect work. Most of what is called "art" is Gardener work. The interesting move is noticing when you’re using Architect methods on Gardener problems, or vice versa, and watching the results decay accordingly.
The hERG QSAR Research Program’s Paper II is the clean empirical case. An Architect approach to the paper would have engineered findings post-hoc to match the original predictive-theory scoping. The Gardener approach pre-committed the cascade discipline -- pre-commit 2, pre-commit 5, operator-level disposition slot at Phase 1c.D -- and let three reframes surface as the data well-conditioned them. The resulting (C) data-landscape paper, with its two-headed contribution structure, is psychologically a different work product from what scoping anticipated. The discipline is what made the reframe legible as a finding rather than a retreat.