I was generally taught to carry significant figures further than strictly necessary to avoid introducing rounding errors.

Which is why I also discussed error propagation, which compounds.

Propagation of uncertainty is not a reason to drop qualifiers. It's a reason to use them.

I can only say that you are reading the metaphor too literally given the examples I've given thus far.

If I want an answer to three significant figures, I do not begin my reasoning by rounding to two sigfigs, then trying to add in the last sigfig later.

Of course!!! This isn't applicable to dialogue, however, as it has the opposite problem: the degree of cognitive burden to retain the informational value of a statement increases with the increased complexity. There is a limit on how much of this can be done in a given conversation.

Increasing complexity of statements to increase their accuracy can cause the ability to comprehend a statement to be reduced.

If the reader does NOT share the assumptions, then relieving them of the cognitive burden of being aware of disagreement is not a service.

This statement carries a specific assumption of depth of dialogue which may or may not be valid.