To me, this goes back to the issue of time. At what point in time are we performing our analysis of play and the products of play?I see the argument. It's the "but you could have scripted the exact same thing (in an infinite universe of scripts)." The parenthetical is the part that's often left unspoken. That a thing can happen does not mean it must or should happen, and the idea that a game that is created in the moment could have been scripted, given the perfect attempt in a sea of infinites, is not a strong or even useful argument.
After the event, any even halfway-successful RPG session produced a fiction that consists of a series of (in-fiction) events. And in the real world there were various moments of play that occurred that are at least loosely correlated with those in-fiction events.
But the previous paragraph doesn't get us any further than a RPG was played. It doesn't tell us how the fiction was created, nor what the moments of play consisted in (eg who got to make what decisions subject to what constraints?). It is the answers to those questions - how?, what? - that determine the actual play experience, as a participant in a process that begins at (say) 2 pm on a Sunday afternoon and finishes around 6.
Adventure Path play yields one range of answers to the how and what questions - a huge part of the how answer is We (or perhaps I, the GM) read it from the book, and inevitably therefore a significant part of the what answer is The GM decided what happened and told everyone else.
This is very different from answers like the fiction was created by allowing the winner of an opposed check to establish the outcome and play consisted in rolling dice to see who wins and hence whose desire for the fiction gets realised. This latter doesn't become a version of the former simply because the GM, in narrating a consequence or in framing a fictional context the leads to an opposed check, had regard to an evinced PC motivation.