To explain the origin of any transition, it is necessary to identify some phenotypic change that brings about a new fitness benefit.That change may be the result of a mutation, but it may equally be a developmental change - a protein folded this way rather than that during a hot spell, for example, or two cells failing to separate completely after mitosis - which subsequently comes under selection for heritability. Or there might be some change in gene expression, caused by features of the environment such as nutrient gradients or local cell-signalling, which allows a cell to adopt a new function. Perhaps the division of labour observed in multicellulars is contingent on the way they develop - closely packed together, competing for nutrients and *ahem* shitting in one another's back yard, to use Paul Rainey's marvellous phrasing.
Anyone who has ever lived in a house share where several people try to make themselves their own dinner at oncewill know that you can't all get in the microwave at once. The solutions are two: cook one meal for everyone, or organise so that one of you is chopping while another monopolises the tap and a third uses the microwave. Its not clear, in the evolution of apoptosis (cell death) during multicellular development (Ratcliff & Travisano 2014) whether the dead cells are altruistically sacrificing themselves for the greater good, or whether they just lost out in the scramble for nutrients. The only cells that survive are those that work out some way to divide the labour so that they all get to eat.
In their forthcoming review of Sterelny et al's 'Cooperation and its Evolution' Uller & Helanterä praise the contributors for illuminating neglected aspects of the evolution of cooperation: its hows as well as its whys. In contrast to many cases of cooperation between living things, human cooperation mostly takes place between non-relatives. This means that we need to identify direct fitness benefits, rather than relying on kin selection theory to explain the stability of cooperation against free riders. For example, psychological mechanisms may play a prominent role in scaffolding human cooperation, i.e. making to robust to decay when selfish individuals seek to avoid its costs. Central to Uller & Helanterä's perspective is the claim that "there is much to be gained from shifting the perspective away from explanations that rely on selective advantage (direct or inclusive fitness) and towards a more mechanistic study of how cooperation arises, spreads, and is maintained."
Human behaviours don't emerge from a simple linear process of natural selection on mutations. There are various ways in which our culture elicits particular phenotypes from the underlying reaction norms of human genotypes and drives runaway feedback processes in which novel traits emerge that are far away from what the genotype would have given rise to had the environment stayed fixed.
One issue is the supply of variation in human cooperative behaviours. Not only is it unlikely to be smooth, unbiased and abundant (isotropic, in Scholl and Pigliucci's terms) but we need to consider carefully the ways in which the supply is affected by things like social learning and the construction of the educational environment. Our mental capacity for reason and communication means that we can plan stuff: we have foresight where natural selection has none. So the supply of behavioural variation is very much biased, directed specifically towards solving problems, overcoming obstacles. We can jump ditches where genetically selected trajectories would get stuck and generally respond much faster and more efficiently to new challenges than if we were sitting around hoping for the right mutation to crop up. Some phenotypic changes might be easier to come by than others, for developmental and ecological reasons. The frequency of the corresponding traits will not correlate with their fitness or adaptive value, in these cases. For example, it has been argued that fraternal transitions are more numerous than egalitarian transitions, because the higher relatedness between cooperators makes the fitness gains bigger. But it could be that the higher prevalence of fraternal multicellulars is due only to the initial first steps being much easier to take, developmentally-speaking.
Another issue is the heritability of cooperative behaviours. Educational scaffolding can guarantee the persistence and heritability of adaptive novelties, regardless of the effect of those novelties on genetic fitness. Material overlap, or 'ecological inheritance' is important in insect colonies. Leaf cutter ant queens, for example, take a packet of leaf-digesting fungus along with them when they found a new colony. Early humans are thought to have relied on careful maintenance and transmission of fire, long before they acquired the ability to reliably create it anew. Humans pass on all kinds of material artefacts, from stone cities, precious jewellery and written manuscripts, to pollution, resource degradation and parasite virulence.
Finally, culture influences genetic evolution by creating the context against which the fitness of our genes is determined. Many adaptations only make sense within the path-dependent context of earlier steps. Lactose tolerance in adulthood only increases fitness once humans have learnt how to control cows. So in so far as we want to answer why have some humans evolved the ability to digest milk in adulthood- to say how adult milk digestion increases human fitness, in other words - we have to mention the appearance of animal husbandry. We might defend the proximate ultimate distinction by responding that animal husbandry must itself have some genetic cause but this is not necessarily true. Animal husbandry probably spread through European populations by cultural, not genetic evolution. To spell that out - the behaviour likely originated in behavioural plasticity, and was then copied by other humans who understood its benefits. If this is right then the behaviour could be taken up by the whole population without any change in the genetic structure of the population. Subsequent genetic change did take place because the new sheep-farming environment exerted strong selection on pre-existing genetic variance in the adult metabolism of lactase. But the initial driver of the process - the ultimate cause, if we take 'ultimate' to imply first - was not genetic.
It is in this sense that Uller & Helanterä champion the status of social, psychological and developmental mechanisms as transcending any putative proximate/ultimate distinction, and belonging firmly within an explanation of the evolution of cooperation. Associative learning, imitation, punishment - they are not mere proximate causes, something to blackbox when thinking about why human cooperation evolved, because they are intricately entangled in the causal web within which any putative fitness effects of human behaviours can be identified.
As well as considering these ways in which culture influences the natural selection of cooperation, by affecting its variation and heredity, and fitness effects, we need to consider that cultural selection, a parallel process, operates alongside and in interaction with natural selection. Human culture can itself be understood as an inheritance system - a mechanism for transmitting information both within and between generations - that supplements and interacts with genetic inheritance. We use language to coordinate group tasks, to draw up social contracts, to advertise ourselves as social partners, to share information about cheats and also to pass on our complex and adaptive cultural legacy. But the impact of much simpler inheritance systems - the inheritance that goes along with sub-personally driven imitation, for example, or the material inheritance each of us acquires in virtue of when and where we happen to be born - may be even greater.