Consider moving recalculation "gating" into the core

In the discussion for #453, we noted that we might want to move this into the core at some point since it's getting a bit repetitive. I haven't thought too deeply about how to integrate this into the internal state, but seems plausible.

Start by reading the discussion in issue #453 and locating the repeated recalculation gating referenced there. Review how the core and its internal state currently handle recalculation, then establish whether a shared integration is appropriate and what behavior the completed refactor must preserve.

Written by the indexing model from the issue text.