Well, to use your x + 1 example, we can say that one is simpler than the other, even if we can't define "simple" in exact terms, or even measure it precisely.

(By the way, I just noticed that your first sentence is perfect for this topic. I suspect that it was accidental, but it's beautiful.)

So I think that simplicity is perhaps somewhat like encoding in Shannon's information theory. Even if you can calculate the minimum number of bits required to represent something, you can't necessarily devise an encoding that will represent it in that number of bits. The analogy isn't perfect, but it has the same feel of "pushing toward something that you can't actually reach". (I guess you said the same thing with your newton's laws analogy, so I'm kind of repeating your point here.)

But I disagree with your last paragraph. Even if we can't measure "simple", we can still kind of feel when complexity is there. If the user needs to be able to do 50 things, then you either need a program that lets the user do 50 things, or you need 50 programs (or something in between). If your program is going to let the user do 50 things, then there has to be a way for the user to specify which of those 50 things they want to do, and the code has to be able to do all 50 things. You might be able to come up with some commonalities, a unified framework to simplify things somewhat, but that can only take you so far. You're still going to have to have 50 buttons on the GUI, or 50 command-line parameters that you accept, or 50 different URLs that you interpret, or whatever.

Or, you can have 50 separate programs, and each of them can be much simpler. But then you have 50 programs, and the user has to figure out which one to use for each thing, and remember them all.

So I claim that you can kind of tell that it's there, even if you can't measure it, and that's good enough for the statement to be a useful thing to keep in mind, even if it's not something we can measure or calculate.