In both everyday talk and artistic storytelling, speakers rarely put their attitudes or evaluations into words in any direct way. Bringing together ideas from expectancy expectation-basedsociopragmatics, and evaluative meaning research, the study builds a three-part model: Expectation Monitoring, then Cross-Modal Signal Coordination, and finally Stance Emergence. What emerges is that irony, a classic case of an expectational stance strategy, does subtle interpersonal work in the film. Rather than fraying social bonds, it tightens closeness, protects face, and stirs empathy. It acts as a bonding agent instead of a disruptor. The paper also draws attention to the multimodal evaluative meaning set out by Bin Dang ^[1]. In that view, evaluative meaning is not a fixed property parked in one channel but forms piece by piece in time, through cross-channel evaluative tension and synergy. This insight turns out to be important for grasping how stance actually materializes in multimodal settings. Fresh experimental work backs up this cross-modal predictive framework: gestural discourse markers can raise discourse-level expectations even without words ^[2], and iconic co-speech gestures are actively recruited to predict what is coming next, a pattern confirmed by both behavioral and electrophysiological measures ^[3]. Meanwhile, computational studies show that expectancy violation itself, treated as a “surprise” signal, works as an efficient device for steering communication, matching the performance of more complex Theory of Mind approaches without demanding heavy mentalizing ^[4].