Proceedings of the Annual Meeting of the Cognitive Science Society

Research on morphological marking of number across the world's language points to a potential asymmetry: while many languages maintain a distinct form for the singular (as distinct from the plural), fewer languages maintain a distinct form for the dual, and very few distinguish trial or paucal. The presence of these patterns and their asymmetries could indicate that morphological systems are structured around a small set of primitives, including \textsc{one} and \textsc{two}, which are weighted differently, i.e., one is prioritized over the other. But how these primitives might relate to conceptual representations of numerosity, where small numbers have also been shown to have a privileged status, remains unclear. Further, there has been little work exploring whether representations of small numerosities might differ in their relative weight or importance in numerical categorization tasks. Here, we report the results from two experiments which aim to provide direct evidence for the role of primitives for \textsc{one} and \textsc{two} in non-linguistic and linguistic categorization. We test participants whose native languages differ in how they express number morphology (English, Japanese, Slovenian). We find strong evidence for a preference to treat 1 distinctly across all populations, and in both task types. The evidence for categorizing 2 as distinct only appears in linguistic tasks, and is generally less strong, even in a population whose language morphologically encodes the dual.