Search Abstracts | Symposia | Slide Sessions | Poster Sessions
The Orthographic Mind: Orthographic Imagery is Associated with Semantic Geometry
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Huichao Yang1, Xiaosha Wang2,3, Yanchao Bi2,3; 1Hebei Normal University, 2Peking University, 3Beijing Normal University
Human semantic representations integrate information abstracted from multiple representational systems. Evidence from sensory deprivation (Wang et al., 2020) and language deprivation (Wang et al., 2023) studies suggests that sensory-derived and language-derived systems contribute overlapping yet partially distinct structural constraints on semantic organization (Fu et al., 2023). Although semantic theories typically assume that semantic structure is broadly shared across individuals, habitual recruitment of internal representational modes (e.g., high reliance on visual imagery or internal verbalization) may subtly modulate semantic geometry. Here we directly tested whether individual differences in internal representational propensities systematically relate to semantic organization. Across nine semantic-rating experiments spanning concrete, emotional, and abstract domains (N = 1109), participants further completed the Internal Representation Questionnaire (IRQ; Roebuck and Lupyan, 2020), assessing orthographic imagery (i.e., the mental simulation of written word forms), internal verbalization, and visual imagery propensities. We first conducted exploratory data-driven analyses using relevance vector regression (RVR) analyses to test whether participants’ overall semantic association patterns predicted IRQ factor scores. Across experiments, overall semantic structure significantly predicted orthographic imagery propensity (Fisher Z-transformed r = .093, 95% CI [.036, .150], p = .015), whereas verbalization and visual imagery were not reliably predicted (ps > .113). We next examined specific semantic structure measures, including clustering-based indices of between-cluster association strength and within-cluster association strength, as well as semantic conformity to group-level structure. Meta-analytic integration across experiments revealed that stronger orthographic imagery was consistently associated with greater between-cluster semantic association strength (Fisher Z-transformed r = .098, 95% CI [.057, .140], p = .006) and reduced semantic conformity (r = −.065, 95% CI [−.121, −.009], p = .035). Stronger internal verbalization was associated with greater within-cluster association strength (r = .072, 95% CI [.031, .113], p = .012), whereas visual imagery showed no reliable associations with semantic geometry. Follow-up analyses indicated that the reduced semantic conformity associated with orthographic imagery was primarily driven by increased cross-category semantic associations. Together, the present research demonstrates that internal representational propensities—particularly orthographic imagery—are associated with systematic differences in semantic organization. These findings suggest that internal representational modes are not merely subjective experiences but correspond to stable weighting biases within distributed semantic systems.
Topic Areas: Meaning: Lexical Semantics,