This study examines how listener position shapes spatial demonstrative choice in Korean and Chinese and whether advanced L1 Chinese learners of Korean show an L1-consistent pattern. Thirty-nine native speakers of Chinese, 24 native speakers of Korean, and 26 learners completed forced-choice tasks involving 16 referent positions in Opposite and L-side speaker– listener configurations. Bayesian mixed-effects analyses showed that native Chinese speakers maintained a broadly speaker-centered zhe–na division, with listener position producing only local boundary modulation. Native Korean speakers displayed an i–geu– jeo organization in the Opposite condition, but the L-side configuration produced a broader reorganization toward an i–jeo opposition, with geu restricted to transitional positions near the listener. Learners reproduced the broad native reorganization, whereas the internal distribution of geu remained more distance-anchored, which constituted an L1-consistent pattern. Their Chinese demonstrative system showed no comparable Korean-like reorganization. These findings indicate that the availability of a geu-dominant region is conditional and gradient, and that sensitivity to listener position does not necessarily entail listener-centered category organization.