A brief recap:
The human brain has two separate decision-making systems.
One system is intuitive, fast, and subconscious. It's designed to recognize patterns and react automatically to them. It holds your memory of motor skills and habits. We're calling that system our Animal Brain.
The other is rational, slow, conscious. It's designed to follow step-by-step instructions. It holds your memories of facts and events. We're calling that system our Android Brain.
Animal Brain tends to dominate our behavior. It broadcasts all kinds of information to Android Brain. But Android Brain has no easy way to communicate back to Animal Brain.
Now we're ready for my guess as to how art works. Remember that this is just an attempt at reverse engineering: to make something that behaves the same way the original does. The internal workings of the brain may be quite different from this. If so, that's OK. I'm really only concerned that the outputs are similar.
I propose that art is simply a way to communicate with our Animal Brain. We do that by taking advantage of Animal Brain's own nature. Animal Brain is constantly on the lookout for unusual patterns in its environment, so that is what we give it with art.
So here's my hypothesis, using my terminology:
The purpose of art is to provide a way for Android Brain to communicate with Animal Brain.
The definition of art is anything made with the intention of communicating with our Animal Brain.
The unit of measurement in art is a single new memory in our Animal Brain.
The quality of an artwork is the rate at which the artwork creates Animal Brain memories.
Now for the same thing, using scientific jargon:
The purpose of art is to enable the declarative memory system to communicate to the nondeclarative memory system. Or, to give System 2 a way to talk to System 1.
The definition of art is anything artificially constructed to stimulate the formation of nondeclarative memories.
The unit of measurement in art is the formation or enhancement of a single nondeclarative memory. Or, a chemical signal resulting from it.
The quality of an artwork is the rate at which nondeclarative memories are formed. Or, the cumulative strength of the resulting chemical signals.
If my hypothesis is correct, all we need to measure the quality of art is some kind of nondeclarometer, which can count the appropriate chemical signals from the Animal Brain's nondeclarative memories as they are created.
New memories send out strong chemical signals. Habituated memories release weaker chemical signals. These chemical signals tell Animal Brain what to pay attention to and what to ignore. I'm hypothesizing that the strength of these chemical signals are what we are measuring when we judge the quality of a work of art.
Drat! I just Googled "nondeclarometer" and got zero hits.
Neural scanners are still pretty crude, but I imagine someday it might be possible to measure memory creation fairly accurately. But for now, measuring art is possible only in theory, not in practice.
The brain is a complex organic machine. I'm sure this simple hypothesis is just that, too simple. But if our goal is usefulness rather than accuracy, simple is probably better, anyway. A hypothesis is a beginning, not an ending. We can test our hypothesis against observable phenomena, and adjust it as we learn more. So let's go use the hypothesis to explain some common phenomena we observe about art.