["Developmental Science, Volume 29, Issue 5, September 2026. ", "\nABSTRACT\n\nAround age four, children in urban, post‐industrial societies begin to understand the subjectivity of beliefs and other mental states: they can be true or false, and they are aspectual, representing situations under specific descriptions. These two insights have been found to emerge in tandem, suggesting a shift toward a metarepresentational Theory of Mind. However, as evidence for this shift comes primarily from formally educated, urban, post‐industrial societies, its broader generalizability remains speculative. Moreover, standard pretense‐heavy false belief tasks may bias conclusions drawn from existing cross‐cultural work. In a preregistered, multi‐method study, we assessed two versions of false belief tasks (change‐of‐location and aspectual) in children from two rural Namibian communities (Hai||om, N = 65; Khwe, N = 35, ages 3–9) and an urban German community (Leipzig, N = 66, ages 3–6). We compared two task formats that differed in their reliance on pretense (a real agent and objects vs. depictions thereof). Task format did not affect Leipzig or Khwe children's performance, but Hai||om children showed a better belief understanding in the realistic task format. Across cultures and task formats, change‐of‐location and aspectual false belief tasks were robustly linked at the individual level. These results show universality without uniformity: besides variation in overall performance and task format sensitivity, children across diverse cultures show closely aligned performance across belief tasks, consistent with accounts proposing a shared metarepresentational basis of Theory of Mind.\n\n\nSummary\n\nStandard change‐of‐location and aspectual false belief tasks are tightly linked at the individual level across rural Namibian (Hai||om and Khwe) and urban German (Leipzig) communities.\nThis developmental unity holds despite cross‐cultural differences in overt task performance.\nRural Hai||om children perform better in realistic (versus pretense) tasks, whereas Khwe and Leipzig children show no such task format advantage.\nThe findings support a universal but non‐uniform emergence of a metarepresentational foundation of belief understanding.\n\n\n"]