In this presentation, I will show that various multimodal resources—such as utterance, prosody, rhythm, schematic images, and bodily reactions—may integratively contribute to the holistic achievement of poeticity. By incorporating the ideas from “ethnopoetics” (Hymes 1981, 1996) and “gesture studies” (McNeill 1992, 2005), I will present a plurimodal analysis of naturally occurring interactions by highlighting the interplay among the verbal, nonverbal, and corporeal representations. With those observations, I confirm that poeticity is not a distinctive quality restricted to constructed poetry or “high” culture, but rather an endowment to any kind of natural discourse that is co-constructed by various semiotic resources. My claim specifically concerns a renewed interest in an ethnopoetic kata ‘form/ shape/ style/ model’ embraced as performative “habitus” among Japanese speakers (Kataoka 2012). Kata, in its broader sense, is stable as well as versatile, often serving as an organizational “template” for performance, which at opportune moments may change its shape and trajectory according to ongoing developments. In other words, preferred structures are not confined to an emergent management of performance, but should also incorporate culturally embedded practices with immediate (re)actions. In order to promote this claim, I explore a case in which mutually coordinated performance is extensively pursued for sharing sympathy and camaraderie. Such a kata-driven construction was typically observed in a highly involved, interactional interview about the Great East Japan Earthquake, in which both interviewer and interviewee were recursively oriented and attuned to the same rhythmic and organizational pattern consisting of an odd-number of kata. Based on these observations, I argue that indigenous principles of organizing discourse are as crucial as the mechanisms of conversational organization, with the higher-order, macro cultural preferences inevitably infiltrating into the micromanagement of spontaneous talk.