According to a recent article in the New Scientist:
One of a Neanderthal baby's first words was probably "papa", concludes one of the most comprehensive attempts to date to make out what the first human language was like.The article is a report on a paper delivered by Pierre Bancel and Alain Matthey de l'Etang of the Association for the Study of Linguistics and Prehistoric Anthropology in Paris at a conference on the Origins of Language and Psychosis held at Oxford in July. According to the reports in the New Scientist and The Telegraph Bancel and de l'Etang surveyed 1000 languages for which they were able to obtain detailed information on kinship terms and found that 700 of them contained the word "papa" with the meaning "father" or "male relative on the father's side".
"There is only one explanation for the consistent meaning of the word 'papa': a common ancestry," Bancel says.To be precise, we can break their conclusions down into four claims. One is that all human languages are descended from a common ancestor, which I'll call Proto-World. The second is that in Proto-World there was a word meaning "father" whose sound was something like [papa]. The third is that Proto-World was the first language spoken by human beings, which can call Proto-Human. The fourth is that Proto-Human was also spoken by Neanderthals. All four claims are dubious.
The observation that words like mama and papa are widespread in human languages is not new. In fact, it was made in the 1950s by the anthropologist George P. Murdoch. In 1959, in response to an appeal for an explanation by Murdoch, Roman Jakobson (my academic "grandfather") published a paper entitled "Why 'mama' and `papa'?", in which he offered the explanation that the mama and papa words come about through the wishful thinking of parents. Before babies start to speak, they go through a period of what linguists call babbling in which they experiment with their vocal tracts and make lots of meaningless noises. Parents don't realize this, though, and eager to hear their child speak, attempt to interpret their baby's vocalizations as words. Naturally, they are keen on the idea that the baby is addressing them, so they assign the meanings "mother" and "father" to the baby's first "words". It happens that certain consonants, such as [p],[t],[b],[d],[m],and [n] are among the sounds that babies produce frequently in the early stages of babbling, as are vowels like [a], so the early "words" perceived by parents are things like [papa], [mama], and [dada]. They aren't actually words, but the parents perceive them as such and assign them the meanings "father" and "mother".
I won't go into this in further detail because the late Larry Trask wrote a very clear and readable essay on this topic entitled "Where do mama/papa words come from?" which can be downloaded here [pdf document]. He explains Jakobson's proposal in more detail and shows how it is far superior to alternatives.
Jakobson's proposal is generally accepted by linguists. That doesn't guarantee that it is correct, but as far as one can tell from the news reports, Bancel and de l'Etang have not attempted to refute it or to address the additional arguments made by Trask. Indeed, the news reports cite my colleague Don Ringe, who told them of Jakobson's explanation, but mention no rebuttal by Bancel and de l'Etang.
Jakobson's proposal isn't just a plausible alternative to common descent
as an explanation for the frequency of words like papa with meanings
like "father"; it is much
Turning to the second point, even if the mama and papa words established that all of the languages known are genetically related, it wouldn't establish that these words were present in that form in Proto-World. We'd have to reconstruct the forms of these words to make any reasonable claim about what they sounded like, and the authors haven't even attempted a reconstruction. They aren't in a position to because they haven't established sound correspondances among the languages they are working with, and without them there is no basis for reconstruction.
The third claim, that the ancestor of all currently known human languages, Proto-World, was the language first spoken by human beings, Proto-Human, assumes that no top-level branches are unknown to us. We can only reconstruct to the lowest common ancestor of the languages for which we have data. Suppose, for example, that the only Indo-European languages known to us were Germanic languages. The best we could do would be to reconstruct Proto-Germanic. To reconstruct anything above the level of Germanic, we have to have data from languages outside of Germanic. If any languages branched off from Proto-Human before the lowest ancestor of the languages we know and became extinct, as is quite possible, Proto-World would be a language separated, possibly by thousands of years, from Proto-Human.
The last claim, that Proto-Human was spoken by Neanderthals, is just plain weird. There is no basis for inferring anything about Neanderthal language from historical inferences about human language. Neanderthals were a different species. We aren't sure whether they had language, much less whether it was genetically related to the language or languages of modern humans.
Ironically, the earliest human beings to have a language probably did have words like "mama" and "papa", but what tells us this is not historical linguistics but Jakobson's argument. If Neanderthals had language, and if their articulatory apparatus and cognitive systems were sufficiently similar to ours, they very likely also had words like "mama" and "papa" for "mother" and "father", for the same reason.