There are significant scale-disparities in Darwin's theory. Viewed from the micro-scale of demographic events, for example, species and genera are long-lived, almost static structures. Species, when they change, do so on the meso-scale, where laws of large numbers apply and Darwin's principle of non facit saltum ensures that they will do so in a continuous, almost linear way. Natural selection is manifest at this meso-scale as a selective field - comparable to a magnetic or gravitational field - that shapes and re-shapes the destinies of populations. These selective fields go out of focus as one zooms in on the micro-perspective of demographic events, because laws of large numbers collapse and time-asymmetry becomes significant. Sometimes individuals get lucky against the odds. Sometimes their luck slides the other way. Those small-scale deviations between observed and expected behaviours are often damped out by synergetic 'friction' (i.e. conflict and competition). Occasionally, however, they may create co-operative opportunities that are amplified synergetically.
These three perspectives form a recurrent theme in evolutionary anthropology and we need some terms to describe them. Historians of the Annales school, following Fernand Braudel (1980), refer to them as event-time (histoire événementielle); conjuncture and deep time (longue durée). Archaeologists interested in the longue durée often favour a 'culture history' approach that uses the methods of typology, seriation and description to describe system ontology. The processual focus would shift attention to the meso-scale or conjuncture (Binford's 'middle range') because this is the level at which laws of large numbers kick in, processes stabilise and prediction is locally meaningful. Events form a small-scale narrative chain of the sort that often appeals to post-processualists. Some of these events do little more than add human interest to the work, but others can trigger cascades of synergetic reorganisation that sweep established processes and typologies aside.
In a metastable system, symmetry-breaking events bounce dynamic systems from one pre-existing attractor to another and all that changes is the processual conjuncture. In an innovative system, however, conceptual taxonomies change too and this becomes manifest as a behavioural and cultural 'revolution' that sweeps old categories away and changes the system's deep-time structure. These innovations can be explained, ex post, with the wisdom of hindsight, but could not possibly have been predicted, ex ante.
Darwin's theory required, rather implausibly, that changes at all three levels (demographic event, conjuncture and deep-time) be smooth and continuous. By the time Darwin was writing Descent of Man, he and Alfred Russel Wallace were beginning to disagree about human evolution. Wallace believed evolution was driven by selective winnowing that eliminated failures and adapted the population to local circumstances. He also believed in the existence of a universal intelligence that exempted some humans and higher life-forms from the struggle for existence (Wallace 1914). Wallace drew a sharp line between natural and artificial selection. Natural dynamics were shaped by the heroic struggle for existence and the selective constraints that punished failure. Synergetic dynamics, including human agency, were supernatural phenomena.
Wallace's universal intelligence was no divine patriarch, but a necessary corollary of his ideas about human intelligence and the constraining nature of natural laws. If humans were agents, then that agency required a supernatural explanation. Henri Bergson (1907), reacting against Darwinian atheism, developed a similar argument in his book Creative Evolution. Descent was Darwin's principal contribution to anthropology. In it, he felt obliged to back away from some of his earlier arguments. He acknowledged, for example, that too much emphasis had been placed on natural selection. By now he had adopted Spencer's phrase 'survival of the fittest' as a punchy alternative to formulations like 'descent with modification under natural selection' and wrote:
'…I now admit … that in the earlier editions of my Origin of Species I probably attributed too much to the action of natural selection or the survival of the fittest. I have altered the fifth edition of the Origin so as to confine my remarks to adaptive changes of structure. I had not formerly sufficiently considered the existence of many structures which appear to be, as far as we can judge, neither beneficial nor injurious; and this I believe to be one of the greatest oversights as yet detected in my work.' (Darwin 1871a, Vol. 1, 152).
This passage is significant because it strengthened Darwin's thesis that the wellsprings of heritable diversity would never dry up. He no longer had to argue that any variability, however small, would be heritable. If some traits were not selected, there would be an untapped pool of heritable variation that could underpin new ways of becoming fit at some time in the future. The word exaptation is sometimes used to describe this phenomenon (after Gould and Vrba 1982). Some of these non-selected traits may have exapted organisms to life-ways that only came into being after patterns of plesionic interaction had activated synergetic multipliers.
Where Wallace had solved the problem of human agency by making a Cartesian distinction between natural and supernatural domains, Darwin (1871b, Vol. 1, 163-65) solved it by writing about natural selection operating at the level of the 'tribe', and about peer-pressure - particularly sensitivity to 'praise and blame' within the community. Darwin did not need nebulous god-agents to explain higher cognition, co-operation and agency. Humans that sacrificed immediate gain for the well-being of the tribe, though they appeared to have lost fitness at an individual level, would have gained compensating benefits by being embedded in a more cohesive and viable tribe. (For a 21st century take on these ideas, see Nesse 2007 and Tognetti et al. 2012.)