Tao, Oliver, and Venville's paper addresses a debate between two hypotheses of children's development of conceptual understandings of the Earth. The authors aim to investigate whether culture influences students' conceptions of the Earth. However, one questionable assumption shared among conception and conceptual change studies is that researchers can identify children's mental models, through their discourse, to explain their conceptions of scientific phenomena. In this commentary, I challenge this assumption by looking closely at various interview data and examining different dimensions of interview discourse about the Earth. Based on my findings, I suggest that instead of seeing science talk as representations of mental models, we look at children's science talk as discursive practices mediated by the immediate social situation and the broader social milieu.