>this 'Singing Neanderthals' by the Pinker sceptic Steven Mithen looks interesting

It is.

It's really an extension of Prehistory of the Mind, or depending on how you look at it, the precursor to the ideas in Prehistory of the Mind. It's got similar bits of 'Neurology 101' as the earlier book, but still done in a digestible but non-condescending manner. There's some good succinct accounts of the various theories that his arguments are based on, with accounts of the arguments put forward against some of them (though you never know which counter arguments have been omitted).

The general gist of it is: Bipedalism causes more complex vocalisations in early hominids, these vocalisations are holistic in that they are complete sequences of sounds more akin to those made by other primates, but longer and with a greater degree of variation in terms of pitch, rhythm etc as a result of physiological changes resulting from an upright posture. Mithen dubs it the 'Hmmmmm' mode of communicaton, with the 'm's being: Manipulative, Multi-Modal, Musical and Memetic. So there have to be 5 'm's, no more, no less, else it would be silly.

Mithen makes a good argument that Neanderthals stuck with Hmmmmm, but developed it to a much greater degree than those common ancestors who drifted off to become Homosapiens. This maps onto his earlier theories about cognitive fluidity vs. domain specific ways of thinking. He reckons that the Neanderthals stayed domain specific, with tool making and social cognitive modules staying firmly delineated. He offers evidence to support this from the archaeological record, as he does to back up the idea that Homosapiens began to segment the big chunks of non-representational vocalisations, to isolate common elements which were then taken to refer to actual objects/events/concepts.

He thinks that music is the surviving form of Hmmmmm in Homosapiens, as the segmentation of 'holistic utterances' is not efficient at communicating more abstract or emotional experience.

There's also a whole chunk on infant directed speech and how this may have played a role in the development of Hmmmmm and our current syntactic, grammatical way of vocalising. He even manages to throw in a snippet of genetic pathology at the end to suggest that there need only be a small mutation in a particular gene in order to cause the ability to start segmenting and reconstructing Hmmmmm-speak. Which, according to the Mithen view of human evolution, means that the development of language presgaed the development of cognitive fluidity, and therefore was largely responsible for humans first starting to make complex multi-part tools and the subsequent domination of the globe. I'm sure I've missed bits out, like the chapter on neuropathological states can shed light on the ways humans experience music, and what this might imply about stuff (not to mention things).

It's a good un. Really got me thinking about music and why it does what it does to humans.