Ancestral human populations were exapted to social learning, collective forgetting and patterns of interaction with the virtual agents we have been calling institutions. They must also have acquired the range of genetic dominance-relations and stable polymorphisms needed to protect rare deleterious genes from expression, at least in statistically stable gene pools. The most serious genetic vulnerability would have arisen when previously isolated populations came together and exchanged genes. The situation would have been genetically analogous to that which we observe in cultivation, where small, genetically diverse founder populations give rise to a range of morphological 'sports' and hybrids. Hugo De Vries (1917) called these sports 'mutations', but the modern synthesis limits the word to de novo copying errors in the genetic code. De Vries' 'mutations' were saltations - sports arising as a consequence of genetic bottlenecks, hybridisation, hot-spots in the genome where mutations tend to occur, and polymorphism.
Our cognitive skills, flexible learning and compassionate impulses arose as part of a package of coping strategies that enabled these 'sportive' populations to co-exist with genetic risk. There is no need to posit a gradual selective constraint that weeded out the less smart and less compassionate. In this genetically sportive community saltatory leaps could activate synergetic multipliers and come to predominate within a few generations. After a period of isolation and restabilisation, the next bottleneck or influx of immigrants could produce further emergents, driving cascades of rapid evolutionary change.
The model of early human evolution we have developed here could, with some justice, be described as the 'vulnerable ape' hypothesis. It suggests that many of our characteristically human traits arose as by-products of hereditary disability and genetic risk. The only reason our hairless bodies, short arms, straight feet, macrocephalous offspring with long periods of infantile dependence and strong predispositions to neurotic and depressive illness are not described as disabilities is that opinion-shaping institutions now present them as normal, or even heroically superior to the ancestral condition. Individuals who deviate from this ideal - infants who struggle to acquire language, have boundless reserves of energy, limited self-awareness and little interest in the mind-games of conventional education - are considered 'disabled'.
Our human ability to construct distinctions of this sort and to empower or exclude neighbours as circumstances change was not caused by genetic challenges in the extreme past any more than airborne dust-grains cause snowflakes. Institutional power-relations arose as emergent by-products of a complex package of cognitive and behavioural coping strategies that opened up a range of possible futures and closed some possibilities down.
The vulnerable ape hypothesis suggests that the early prehistory of our species would have produced a fossil record that cannot be resolved into a clear, divergent hierarchy of coherent demes and clades. Rather, we should expect a mosaic of trait combinations that generates a network of divergent, parallel and convergent lineages. The fossil evidence is indeed consistent with this interpretation. Moreover, the archaeological evidence suggests that disabled individuals do indeed occur in the archaeological record and that their close neighbours sometimes protected them from the effects of natural selection (reviewed in great detail by Spikins et al. 2010). We believe the archaeology of compassion and co-operation to be an important open area of research for 21st century archaeology (Spikins 2015) and see this article as a contribution to that research initiative.