As my colleague (and co-author of the imprinted brain theory) Bernard Crespi points out, "The origin of speech and language is arguably the most important transition in the evolution of modern humans." He adds that there is now good evidence that one gene in particular--FOXP2--is critical to both the mirror neuron system in humans and to articulate speech. Mirror neurons fire when a person sees someone else performing an act which larger-scale firing of motor neurons in the same region would produce in the observer. As such, they have been seen as part of the neural basis of empathy, and there are averagely more of them in women than in men, and fewest of all in autistics, whose deficits in empathy have been claimed to epitomize the disorder.
Of course, empathy is important in language--at the very least you need to understand what other people are trying to say. But Crespi goes on to note that evidence that FOXP2 is imprinted and predominantly expressed from the paternal chromosome reminds us that language is also much concerned with self-assertion--and in the case of young children, with demands on the parents.
According to the best theory of imprinting we have, paternal genes have a greater self-interest in such demands than the mother's. This is because, although both parents derive the same benefit from investment in the offspring, it is the mother who provides all the resources during gestation and lactation, which are from the father's point of view a free lunch for his genes present in the child. Crespi concludes that "By this hypothesis, articulate human speech evolved as it develops, predominantly in the context of mother-offspring interactions, which are permeated by a complex mix of cooperation and conflict."
Two syndromes reveal the conflict most dramatically. In Beckwith-Wiedemann syndrome both copies of the normally maternally-imprinted/paternally-active IGF2 growth-factor gene are expressed, and the result is pre- and post-natal overgrowth. In its opposite, under-growth condition (where neither parent's copy of IGF2 is expressed) Silver-Russell syndrome. Significantly, loss of FOXP2 is linked to expressive but not receptive verbal dyspraxia, just as the parental antagonism theory would predict.
For a superb account of language evolution from the point of view of the imprinted brain theory, see Bernard Crespi's "Language unbound: genomic conflict and pyschosis in the origin of modern humans" (chapter 13 of Sociobiology of Communication).