Szamado S and Szathmary E. 2006. Selective scenarios for the emergence of natural language. TRENDS in Ecology and Evolution 21: 555-561.

Honest signaling seems to be applicable in most modalities of communication, but we humans experience it particularly in reference to language. The evolutionary origins of language are difficult to explore, say authors Szamado and Szathmary in their 2006 article, because of the three confounding time elements surrounding it: phylogeny (ancient origins of organisms), ontogeny (developmental origins of organisms) and horizontal or cultural transmission of information through learning. We know a lot about the ancestors of humans, but little about the anatomy and genetics of the first humans that would have used language, and little about why linguistic communication is unique to humans. The authors speculate that some sort of transition in human development linked to either anatomy or social structure is at the crux of the origin of language.

Transitions can be either variation or selection limited. Variation limited transitions require long evolutionary time in order for proper alleles to be present and proliferate. By contrast, selection limited transitions merely require proper conditions for a pre-existing genetic condition to be selected for. The authors suggest that the evolution of human language may have been selection limited because our closest relatives, the great apes, are able to learn sign communication and their genomes are 99% equivalent to our own. This begs the question: what context could have been sufficient for language to develop in some ape descendants and not others? Some theories suggested by contemporaries are for gossip, group bonding, to coordinate hunting efforts, to calm children (“mother-ese”), or as a method of attracting a mate or explaining a pair bond or “social contract.”

The authors wondered if some of these possible contexts could have elicited evolution of dishonest communication. Producing speech and gestures seems to be a low-cost form of communication for humans, and both participants in the communication should benefit (or else it would be pointless) so this should favor using language in conflict free situations (as in hunting coordination or motherese), where honest signaling would be helpful to senders and receivers and dishonest signaling would be easily weeded out. Whatever the original context of language, though, there are certainly situations where communicators have conflicting interests (as in gossip and negotiating social contracts), making lying a strategy that could have some stability but would likely come at a cost—punishment by other group members if the lie is caught.

Based on these warring contexts, the first words spoken by humans could be anything from a call of “I’m here!” to an acknowledgement of a relationship, like “Mama” or “friend,” to a diagnostic of good game to hunt or plants to find. I found a lot of these first word guesses to be somewhat far-fetched. Language doesn’t seem to be absolutely necessary to tell someone where you are, and other social mammals are capable of coordinating group hunting without talking. The authors also recognize that presently the level of selective pressure necessary to develop language isn’t understood and modeling should take this into account.