A Cognitive Invocation Model of Haiku: Six Modes of Reader-Side Rendering

This paper proposes a cognitive model of haiku, not as a poetic form that merely compresses emotion or atmosphere, but as a minimal linguistic structure that activates cognitive processes on the reader's side. In this model, the emotional effect of haiku is not directly stored inside the poem. Rather, it emerges as a rendering result generated by the reader from a small number of verbal variables. The paper analyzes six haiku: Takarai Kikaku's “Meigetsu ya / Tatami no ue ni / Matsu no kage,” Matsuo Basho's “Shizukasa ya / Iwa ni shimiiru / Semi no koe,” Yosa Buson's “Na no hana ya / Tsuki wa higashi ni / Hi wa nishi ni,” Kobayashi Issa's “Yasegaeru / Makeru na Issa / Kore ni ari,” Kaga no Chiyojo's “Asagao ni / Tsurube torarete / Morai mizu,” and Matsuo Basho's “Natsukusa ya / Tsuwamono-domo ga / Yume no ato.” These poems activate different cognitive modes: optical projection, acoustic saturation, spherical spatial rendering, subject-object isomorphism, character reconstruction from behavioral traces, and reconstruction of historical agency from temporal traces. The paper further argues that the 5-7-5 structure of haiku is not merely a conventional sound pattern, but an asymmetric cognitive architecture composed of an initial cognitive hook and a subsequent 7-5 developmental body. Through this structure, haiku minimizes linguistic input while maximizing reader-side rendering. The traditional Japanese notion of yojo is reinterpreted as residual cognitive processing, or an unresolved cognitive task that continues after the poem has ended. Haiku is therefore understood as a compressed cognitive invocation: the art of reconstructing a world inside the reader through the smallest possible amount of language.