This paper reflects on 'late development', or more precisely, 'late industrialization': the idea that the industrialization of one part of the world altered the possibilities for industrialization elsewhere, such that later industrializations would take different forms from the original industrial revolution. For half a century, following Alexander Gerschenkron's account of late Tsarist Russia, the leading role of the state has been seen in the literature as the major characteristic of late industrialization in Latin America, Asia and Africa. The paper endorses the emphasis on the importance of the role of the state, but argues that a gap has emerged between the economic historiography of late industrialization before the First World War, and the political economy literature on the more recent industrializations, written by Alice Amsden and others. A number of issues are identified for further research, in the direction of a unified analysis of the history of late industrialization. The paper goes on to argue that the study of 'late industrialization' needs to be given a thematically wider and chronologically deeper frame. This possibility arises from the approach to very long-term economic development developed over the last decade or so by Kaoru Sugihara. In his view, Asian development was distinguished from Western development not just by the greater role of the state but by greater labour (increasingly accompanied by skill) – intensity. The paper explores the significance and problems of Sugihara's distinction of plural 'paths' of very long-term economic development, defined by characteristic technical and institutional responses to relative factor endowments. Questions arise, for example, in the context of the different experiences of Latin America and Africa, as well as from variations across Asia. It is suggested that a research agenda emerges that may allow us to enrich the analysis of 'late industrialization', and set it firmly in the context of longer-term economic development.