Collingwood points out that our concept of this thing is pretty vague, and tied up with other concepts, so much so that there isn't even a name for it that is not equally naturally applied to many other concepts. So in specifying our vague notion, we will find that we uncover mistakes in our vague concept, and find that the concept we are interested in has some properties (and extensions and so on) we thought it didn't, and hasn't other properties we thought it did. It is like we have a patch of earth covered in various objects, and we are interested in capturing one in a translucent sheet, through which sheet, held in our hands, we survey the scene. The light from the objects is distorted by the sheet, so that as we lower the sheet onto our chosen object, we find that it is different to how it appeared from a distance. It is of course more sharply defined, but it is also a different shape; further, it is a similar colour to some of its neighbours, and from a distance we thought that some of the neighbouring objects were actually one and the same as our object.
The problem with this approach to clarifying a concept is this: How can we know that the object we now have hold of is in fact the one we were originally after? Could it not be that what we were originally after was in fact two objects (or half of the object), and that in covering only one, we have only captured half of what we were after? Or could it be that what we were after was something vague, and by covering something definite, we have not captured what we were after? To bring this back to concepts: We are after something best called art, which is not actually art, but which has a lot to do with art. When we have it in our hands, we see that it turns out not to include a lot of what we call art - e.g., muzak and its equivalents in the other arts - and to include a lot of what we would never call art - e.g., much of our everyday speech.
Now, for me, personally, the problem is not so acute. I had been told when being given my translucent sheet that the object has a certain magnetism, and that I would know it when I had it, and indeed I do. But this is of no concern to anyone else. I am trying to share Collingwood's insights with other people; to demonstrate that what he ends up with, although jagged and bright in ways few would have suspected, is in fact what he set out to capture. Or if it turns out that he has picked up the wrong object, I want to be able to see where he went wrong without abandoning the entire approach. So my question here is: who else has grappled with this problem? How would you deal with it?