Eliezer, one qualm: You consistently bring up mirror neurons and consider it to be obvious prima facie that they are used for action understanding in humans. Unfortunately, most contemporary neuroscientists in the field agree that there is no consistent evidence of this:

http://www.cognitionandculture.net/index.php?option=com_content&view=article&id=223%3Ado-we-have-mirror-neurons-at-all&catid=32%3Aoliviers-blog&Itemid=34

http://talkingbrains.blogspot.com/2008/08/eight-problems-for-mirror-neuron-theory.html

That is not to say that humans don't understand other people's actions or that we do not have adequate theory of minds! But it does mean that there is no reason to suspect that those complicated cognitive events can be reduced to simply a group of "mirror" neurons. Ramachandran often mentions them as well, which irks me slightly as well.