I saw this on Archaeology a while ago and though I used to deal with linguistics a little bit for my job, I am still very surprised by linguistic reconstructionists. I subscribe to the idea of monogenesis, but it still baffles me as to how these reconstructions are arrived at. I understand that a great deal of research and cross-referencing is done, but I can't help but wonder how close these reconstructions really are and it bums me out that we can never know empirically. One of the many banes of a language teacher is explaining to students that sounds outside of the phonetic palette of their mother tongue, exist and I imagine that perhaps some of those pronunciations of PIE words might be inaccurate due to a lack of knowledge of the full extent of the PIE phonetic palette.
Another thing I've often wondered about is how sounds were selected as the basis of words and how that changed over time. In most languages, words of love are more musical and pleasing to the ear than words of anger or threat, but even so there are languages out there that sound quite ugly to me and seem like they do not lend themselves to easy learning, especially to children who have yet to fully develop their soft palate. Still, the ability children children have for mimicking sounds and cultivating an ear for the subtleties of language is pretty amazing.