Measurement in psychology is at the heart of a major debate in the academic literature. We aim to contribute to a critical discussion of this issue. We propose to reposition the object of this type of measure, namely a mental attribute as measured by mental tests. Mental attributes should be considered not as a true object independent of the knower, but as an emergent property of a person dependent on the social context. On the basis of this clarified ontology, we consider that an empirical approach to measuring a mental attribute is possible. This approach must be resolutely pragmatist and realist. In practical terms, this means that a test needs to be renegotiated relative to the context. The validation of quantitative measures requires verification of a certain number of criteria. Consequently, our work critically explores measures as they are usually implemented in the area of psychometrics.