Purpose - The article extends the distinction of semantic from syntactic labour to comprehend all forms of mental labour. It answers a critique from de Fremery and Buckland, which required envisaging mental labour as a differentiated spectrum. Design/methodology/approach - The paper adopts a discursive approach. It first reviews the significance and extensive diffusion of the distinction of semantic from syntactic labour. Second, it integrates semantic and syntactic labour along a vertical dimension within mental labour, indicating analogies in principle with, and differences in application from, the inherited distinction of intellectual from clerical labour. Third, it develops semantic labour to the very highest level, on a consistent principle of differentiation from syntactic labour. Finally, it reintegrates the understanding developed of semantic labour with syntactic labour, confirming that they can fully and informatively occupy mental labour. Findings - The article further validates the distinction of semantic from syntactic labour. It enables to address Norbert Wiener's classic challenge of appropriately distributing activity between human and computer. Research limitations/implications - The article transforms work in progress into knowledge for diffusion. Practical implications - It has practical implications for determining what tasks to delegate to computational technology. Social implications - The paper has social implications for the understanding of appropriate human and machine computational tasks and our own distinctive humanness. Originality/value - The paper is highly original. Although based on preceding research, from the late 20th century, it is the first separately published full account of semantic and syntactic labour.