The goal of the current study was to statistically evaluate the reliable scalability of a set of tasks designed to assess Theory of Mind (ToM) without language as a confounding variable. This tool might be useful to study ToM in populations where language is impaired or to study links between language and ToM. Low verbal versions of the ToM tasks proposed by Wellman and Liu (2004) for their scale were tested in 234 children (2.5 years to 11.9 years). Results showed that 5 of the tasks formed a scale according to both Guttman and Rasch models whereas all 6 tasks could form a scale according to the Rasch model only. The main difference from the original scale was that the Explicit False Belief task could be included whereas the Knowledge Access (KA) task could not. The authors argue that the more verbal version of the KA task administered in previous studies could have measured language understanding rather than ToM.