300 y.a., chairs were thought to have a very different fundamental nature than we think that they have now.
Yet, we don’t say that therefore chairs aren’t real, or that people who talked about chairs 300 y.a. were deluded. Nor would we say that we are engaging in delusion whenever we talk about chairs nowadays, even though we probably have the fundamental nature of chairs wrong.
In coding, there is an approach called object-oriented programming (OOP). OOP collects data and functions in an ‘object’. When working with the object, one can work not directly with the internal data and functions, but with what’s known as an interface. This interface encapsulates the ‘private’ data and functions of the object – it makes it so that you as a coder can work with the object without worrying about its fundamental implementation. Therefore, the private implementation can be changed while the interface remains the same, which is useful.
Similarly, the concept ‘chair’ is centrally about an everyday interface with chair objects, where the fundamental ‘implementation’ of a chair (it’s fundamental nature) isn’t that relevant. The interface works, regardless of how we change our concept of what chairs fundamentally are, because the interface is about the everyday macroscopic interactions people have with these things we call ‘chairs’.
So, why do we say that some things that we used to say exist don’t exist anymore, even though both today’s and the historical concepts refer (at least in part) to the same (real) things? I think the answer is that, for whatever reasons, the relevant interfaces for the two concepts are significantly different.
Take the idea of dragons and dinosaurs. We say that dragons don’t exist, even though they referred to fossil evidence that was shared by some (what we now call) dinosaurs. Understood in terms of reference, these were the same (sub-)set of things, and so dragons did exist, just as dinosaurs did exist. Yet, we don’t say that. Why not? Since the primary way that we interact with dinosaurs is in a paleontological setting, the concept is determined by the relevant paleontological interface. Yet, ‘dragons’ don’t have a primarily paleontological interface, but rather are interfaced through stories, where they have various attributes that are incompatible with dinosaurs (such as talking, casting spells, living contemporaneously with humans, and so on).
Consider the idea that certain illnesses were caused by demons, that now we believe are caused by bacteria. Why didn’t we identify demons with these bacteria? Both are invisible, enter into the body, and cause us to become sick. One answer is that the relevant interface for dealing with the illness changed, i.e., to deal with demons, you don’t apply antibiotics. In order for an identity to have occurred, those using the concept ‘demon’ would have had to make a massive change to various parts of it, so it fit with what we were learning about these certain types of bacteria. That didn’t occur, and so we now say that demons don’t cause these certain illnesses, rather certain kinds of bacteria do, even though the concept ‘demon’ used to refer (in part, it turns out) to those certain kinds of bacteria.