The problem is we don't have enough lexical tokens to build the grammar we want. Commas perform two distinct functions: they separate items on a list, and they separate grammatical clauses. It turns out that situations arise where you can't know which function a comma is performing without prior knowledge of the writer's intent.
Suppose I write, "I owe everything to Jocasta, my wife, and my mother." There's no telling whether I'm talking about two people or three. Even if you *know* I use the Oxford Comma to separate every single item in lists, there's no way to determine that's what I'm doing here; "Jocasta" and "my wife" may be the same person. The only way for you to figure out my meaning is to have prior knowledge of my wife's name.
Yet if I adopt the policy of *never* using the Oxford comma, that leads to equally ambiguous results. Try it. There is simply no way to fix this problem working with nothing but commas.
Now as a CS graduate, the solution is clear: I need distinct tokens to separate list items and appositive clauses. Suppose I use "+" for list items and em-dash to set off appositive clauses:
I owe everything to Jocasta -- my wife -- + and my mother.
I owe everything to Jocasta + my wife + and my mother.
Or if you're Oedipus Rex:
I owe everything to Jocasta -- my wife and my mother.
But until a major style guide adopts distinctive punctuation marks, we're stuck with ambiguity. That means you have no choice but to read any list or appositive clause you've written with a critical eye, and then rewrite the sentence if it can be misinterpreted. The result may be awkward and ungainly ("I owe everything to my wife, Jocasta, and also to my mother.") and the whole process is irritating, but you have no choice.