Africa has the oldest artefacts and evidence for fire. It is where Homo sapiens evolved and developed novel technologies before dispersing into the rest of the world some 70ka ago. There is, however, no reliable evidence in Africa for artificial shelters and dwellings older than 20ka. This paper sets out to understand why such basic architecture appears so late in a continent with great environmental variation and a deep history of innovation. The approach combines evidence from micro and macro scales of analysis. The micro scale uses ethnoarchaeological studies of Africa’s small circular houses to examine how and why gender separates their occupants both spatially and through their access to food stores. At the macro scale, the absence of food stores among Africa’s extant hunters and gatherers is predicted from environmental factors that apply to the whole continent. Without food storage there are no significant dwellings. I then turn to the archaeological evidence for the appearance of dwellings and storage from Africa and the Levant, a contiguous region where huts are known at 23ka. The evidence for dwellings in Europe is then considered. While dwellings are earlier here than in Africa and the Levant none are reliably older than 32ka. They are found with evidence for food storage. The paper explores the implications of this chronological framework for a major transition in hominin evolution that, before agriculture, involved intensification in subsistence combined with storage, and a novel architecture of gendered spaces now found worldwide.

Excavations of pre-pottery levels at Gua Talimbue and Gua Sambagowala in Southeast Sulawesi, Indonesia, have yielded nearly 4kg of baked-clay fragments, half of which exhibit intentional patterning. The fragments appear to derive from clay hearths. Here, the authors link the patterning on Early Holocene (c. 9900–8800 cal BP) fragments with the intention to enhance the appearance of the hearths’ rims. During the Mid/Late Holocene (c. 4500–2000 cal BP), patterning shifts to the interior surfaces. The effort and specialised skills required to impress patterns on these hearths is, to date, unique in the archaeology of pre-Neolithic Island Southeast Asia.

Presented here is the so far unique discovery and interpretation of an occupation area directly associated with Upper Palaeolithic cave paintings. The paintings, of red spots and hand stencils, overlook two hearths with selected flints. There were also fragments of stalactite, deduced by analysis and experiment to be waste products from the manufacture of beads. The authors deduce that the hearths and their assemblage complement the ritual nature of the paintings.

Analysis and dating of new Upper Palaeolithic sites suggest that microblades emerged in the Transbaikal after 18,000 years ago. These findings encourage review of earlier assertions that such technologies developed in northeast Asia prior to the last glacial maximum.