I've seen speculation that the relatively rapid expansion of the language centers of the brain within our ancestors, had an increase in survivability not just because of the ability for individuals to communicate with each other, but for internal communication as well.

The ability to use discreet symbol structures gave the ability to eventually utilize abstract concepts such as "past", "future", and so forth, enabling a higher efficiency of action.

In terms of group/kin selection, populations that would be better able to establish commonality of symbol structure and default behaviors would have a distinct edge, both in terms of dealing with the external environment, and in internal modulation of inter-social actions.

Demonstrating heredity for a function such as this would be difficult specifically because of the somewhat ethereal nature of information that would be primarily encoded within oral traditions.

It might potentially be possible to test for minor variations in genetically driven neurological development, as correlated with archeological and historical data, but the minimal record of cultural mechanisms prior to a few tens of thousands of years ago would make it difficult.

It might be more useful to look at the question in terms of capacity for shared paradigmic transference.

To that degree, even the general range of unspoken assumptions that operate within new social environments (such as computer driven virtual forums) could be considered as deriving from the same force of selection as religion/culture.

It'd be interesting to see a timeline for the genetic development of 'Mirror Neurons', and see if there is any aspect of their functioning, that would interact on a lingual/world view level, as opposed to the bio-motor/emotional learning role that they are normally ascribed to.