It might not be worth spending much ink here on it, though, so it might just go on a wishlist of “I’d like someone to discuss this in more detail sometime”. I think it would probably require a bigger focus on Sengers’s overall argument to separate what I see as these two critiques and how they influenced other work, and that’d have to be done with some thought, because I’m not sure Sengers would actually agree they were as separate as I (and maybe you) see them. From the cultural-theory standpoint she comes at the problem from, the critique of agents-as-autonomous-simulation isn’t just that it isn’t a useful perspective for authoring interactive narrative, but that it’s a fundamental misunderstanding of how the world and real people work, taking a stance against a cognitive-science-style analysis of individual humans as independent agents.

I’m not sure if she’d agree, but I tend to see this as a somewhat different criticism of traditional agent research that can stand separately from her more specific criticisms of atomization and schizophrenic agent behavior, especially in the context of interactive narrative. It could be seen as an agentsy variation of the “simulate authors, not characters” reaction that some people have had to Tale-Spin: that when you’re designing interactive experiences, the point of agents is to communicate something from the author to the player, not to simulate them accurately as an end in itself.