Me, personally, I always thought G, N, and S were BEST achieved when they were all achieved together. Conversely, a narrative, for example, that lacks internal consistency falls flat.
But the nice thing about GNS making up new terms is, whenever you present an argument that clearly debunks the theory, they can say "well, that's not what we meant by narrativist" or "that's not what we meant by creative agenda." It's a game of three-card monty, you can't win.
Take a look at the wikipedia version of GNS, for example. Narrativism seems to have little to do with narratives or stories. It's about how character motivations affect play. So for some reason, under this scheme, following through on what logically happens is simulationist when it involves the game world but narrativist if it involves a select few characters from that game world.
The difference between Gamism and Narrativism is a little bit more reasonable, but still falls apart under close scrutiny. It seems like narrativism concerns itself with the goals of characters (I know it didn't say the word "goals" but motives always imply goals). But gamism is about the goals of players. Now what happens if a player's goal is to put himself into his character's shoes and drive that character towards that character's goals? In other words, I'm asking what happens if he's playing a roleplaying game. It would seem the distinction between G and N would disappear just as surely as N and S.
Showing G and S as one is almost too easy. In the early days of the theory, it took a lot of work to try and distinguish the two. But I will say again, suppose, as is the case in a roleplaying game, that the rules, conditions, and obstacles of the game were that of a consistent game world? And why not just throw highly motivated characters in as well.