Chiasmus is a form of play.

Playfulness can evoke a mood of celebration, and celebrations evoke a mood of playfulness. (What? Yet another chiasmus – how did that happen?) An essay should culminate in a good conclusion, and a good conclusion will be the culmination of an essay. (Oh, no, it happened again.) How can I finish this argument, without this argument finishing me? (Help, someone, get me out of here!) And so another paradox emerges, which is that the playfulness of the chiasmus is closely related to its being a small prison house of language.
(Hariman, 2014, p. 59)

Part of chiasmus’s rhetorical power is its sense of play—both in the sense that chiasmus is playful, entertaining, an invitation to a kind of linguistic game, perhaps even a form intellectual trickery; but also play in the sense of “the space in or through which a thing can or does move,” like the slack in a fishing line that has hooked something in motion, the “opportunity, or room for action” (OED) that a chiastic phrase leaves open. That is, chiasmus invokes rhetorical possibility within constraints.

Like most games, chiasmus is at its core collaborative—it takes more than one person to play. There is a pleasure in “solving” the puzzle of a chiasmus, and this pleasure incentivizes cooperation, which is what makes chiasmus such a persuasive pattern. According to Lissner (2007):

In the course of apprehending the formal pattern, an audience becomes involved in a cooperator’s role. Cooperation amounts to a “measure of agreement” (Perelman and Olbrechts-Tyteca 494) whether or not audience members accede to the reasoning developed... This investing is not a matter of conscious choice, but rather of the compelling nature of the form itself.
(p. 248)

Regardless of what a chiasmus’s intended “direction” is, audiences find it difficult to avoid “buying into” the process of the chiasmus, if not the conclusion. An argument must be entertained simply to enjoy the game of chiastic rhetoric, and the explicit suggestion that audiences have agency in the game, even if they do not create the rules, makes the form all the more seductive.

Crossing rhetoric is interactive, democratic in the sense that everyone gets a chance to “play.” It persuades through what Bogost (2007) calls procedural rhetoric, similarly to what games do, by “authoring arguments through processes” and through the “construction of dynamic models” (Bogost, 2007, p. 29). It is chiasmus’s dynamism that attracts us, and the limiting but not deterministic set of rules that prompt audiences to do chiasmus and not simply read chiasmus.