Whilst there are various theories on why cooperation emerges in nature, these all remain to varying degrees unsatisfactory and certainly not capable of full generalisation, as per Darwin's theory for competition.
Has there been any historic consideration of the quantum nature of food?
As a generality food comes in quanta in the wild - individual fruits, clumps of grass, individual prey, etc.
When population pressure increases and competition becomes intense for a limited supply of food, could this quantised nature of food be a generalised reason for the emergence of cooperation?
Consider scenario. 100 monkeys descend on a tree with 90 fruit. Only 90 monkeys eat that day. 10 go without and will be severely weakened before getting to the next fruiting tree. But if some monkeys learnt to share fruit, then they'd get, say, 9/10ths of a fruit each and thereby keep going another day.
Cooperation in this context may have little benefit on a one-off basis. But if food was scarce for a long period, then this evolutionary strategy would have a clear benefit. The first step of cooperation, then, would be a generalised measure to enable the smoothing of food supply from one day to the next. It would even apply to those herds on the savannah, forced to cooperate when the rains don't come and the otherwise abundant grass recedes to a few, dispersed clumps of grass and pools of water.