This is a chair with two legs:
Passive_behaviours_by_Benoit_Malta_dezeen_sq.jpg
This is a chair without a seat:
MirandaBefore.jpg
I am aware that you set a criteria for a chair in order to prove a point. I am not taking it literally. I think this is a good definition of a chair. I gave an example about chairs to keep it simple. I gave these two examples of chair to point out that it is hard to set a strict criteria even for simple things, although, it doesn't prevent us from defining objects by using the same labels even they don't exactly fit to their given set of criteria.
We categorize and label everything according to our past observations. Logical equation can be proved easily without any observation, such as if p is True and q is True, then p v q is True. Other than that, every experimental testing or science experiment is based on past observations. If I am wrong, please give me an example. Personality models/structures are also going to be based on past observations.
I agree that "pieces of furniture are more easily defined as having static properties". Open systems or circuits also have a inner mechanism. We and even some machines to some degree can also understand the inner mechanism of open systems, therefore, we and some machines (to some degree) can predict the future outcomes of an open systems, you can find related articles online, I can find them if you couldn't. Once the inner mechanism is understood, the output of the system can be predicted by its given input.