The Origin of Language

I recently came across a video by NativLang that brought back the memory of reading about a fringe linguistic theory - the theory that we can trace language families back to superfamilies and eventually a proto-World that was the common ancestor of all spoken languages.

Back when I first came across that theory as a teen, I thought it sounded very plausible, in fact pretty much obvious. Of course there would be a common ancestor of all spoken languages! The only question was whether we could figure out any of the features of that language with any degree of accuracy.

But now, considering the concept again, I have a lot more questions and uncertainty about it.

First, imagine if we tried this exercise with signed languages? Imagine we pretend to know absolutely nothing about the history of any of these languages, and we tried to infer the common ancestors between sign language families, then superfamilies, and then the original sign language? Would we find a signed proto-World?

If we did, that would cast serious doubts on the whole theory, because unlike spoken languages, most signed languages are young enough that we know exactly how they started. And we know for a fact that several signed languages have appeared independently among populations that were not exposed to any other signed language. One of the best studied of these examples of a brand new language emerging independently is Nicaraguan Sign Language (ISN). ISN had its birth when the first Nicaraguan school for Deaf children opened in 1977. Education in this school, and in the vocational school for Deaf teens that opened soon after, focused on lipreading and spoken Spanish, with some use of fingerspelling.

However, young Deaf children who grow up without exposure to sign language often make up “home signs” to communicate with their families. These Nicaraguan Deaf children used idiosyncratic home sign systems with their families before they went to school, and then taught each other their home signs and began developing a common vocabulary. Initially, this was a simplified pidgin-like sign language, but as a batch of new students came each year, those children learned and innovated upon the developing language, until it was a language similar in complexity and usefulness to spoken languages and the many older sign languages around the world.

Similarly, many of the older, more established sign languages have known histories. In 1771, Charles-Michel de l'Épée, a French philanthropist, established one of the first schools for the Deaf, and in the process, founded French Sign Language (LSF). Around the same time, British Sign Language may have been having its birth in England - or it may be descended from pockets of signing communities dating back to the 1400s. In the 1700s, a genetic bottleneck in Martha’s Vineyard led to a high rate of hereditary deafness (autosomal recessive), which led to the invention of Martha’s Vineyard Sign Language. MVSL is now extinct, but it and LSF both contributed to the formation of American Sign Language.

The example of sign languages proves that anatomically modern humans have the potential to invent new languages without prior language exposure. But when and how did this ability emerge? Our closest relatives, chimpanzees and bonobos, do not seem to generate their own languages spontaneously, but are capable of learning human languages at a proficiency level similar to a 2 year old human child. Most likely, the transition from non-linguistic to linguistic vocalizations in hominids was gradual rather than abrupt, and would have involved individual variability, with the same community including individuals with varying genetic predisposition to language.

Like a deaf child surrounded by non-signing hearing people, a single individual with more innate language predisposition than their community would not reach their full linguistic potential. But imagine they have children - their children would likely inherit some of their parent’s innate language potential, and have exposure to an adult who has used that potential to develop a spoken equivalent to home signs. Like Deaf children exposed to older Deaf home signers, these children would take their parent’s communication and innovate and expand upon it.

So, imagine that separate language predisposition genes arise as new mutations in different communities (several candidate genes have been identified, others we might never find). Each of these genes leads to slightly more complexity in vocal communication in the communities where they’re common. And then, in communities that intersect these mutations, we suddenly get individuals who combine the benefits of multiple language predisposing mutations, and as they interact with relatives who share these mutation combos, a true language begins to emerge.

Of course, home signs likely draw upon gestural systems that hearing non-signers use as an adjunct to speech. Similarly, there were probably regional variations in prelinguistic vocal communication patterns that ended up shaping language families. But many of the features of a true language would have been independently invented in each community that happened to cluster more language-predisposed individuals.

And having a community language would bring advantages. Two people from separate groups stumble upon a dead mammoth, killed by a cave lion who ate their fill and left. There is still most of the mammoth left, good meat that must be eaten quickly before it rots. Both run home to fetch members of their tribe to share the meal.

One person runs home and gabbles out something in a quasi-linguistic communication system and manages to convey “I want you to follow me” but can’t explain why they should follow. This person isn’t very high-ranking, so only their closest family and friends decide to check out what has them so excited, while most of their group instead follows their leader to go on a hunt that may or may not be successful.

The other person comes home to a community with a true, albeit still relatively simple language, and is able to convey “dead mammoth, much food, follow me”. This message gives a clear and compelling reason, convincing the leader of their group to abandon the planned hunt for a less dangerous and more sure meal.

When the whole group comes back to the mammoth, they find a handful of people from the other group. Whether they drive them off or let them share is entirely up to their leader. Even if they let those few share, the rest of the group fails in their hunt and goes hungry. Over time, experiences like these cause the linguistic community to grow until they fracture into multiple communities that gradually develop different dialects, while their quasilinguistic neighbors die out or get absorbed into them over time.

If this is how language developed, there may not be one proto-World. There probably would be a superfamily that accounts for many of the languages of the Americas, since they were only colonized by humans roughly 13,000 years ago, well after the point that paleontologists think human language first emerged. If homo sapiens were the only hominid to have a true language, then all non-African languages should descend from the language spoken by the roughly 1,000 people who left Africa around 30,000 years ago and managed to spread across the world. However, the descendants of those people met other hominids in many of the lands they settled, and if those people, too, had languages, the picture could be even more complex. Even if they merely had quasilinguistic communication systems, those could have influenced the languages of their homo sapiens neighbors in ways that made their language unique compared to the languages of those humans’ long-lost cousins who went in a different direction and met different hominids.

And, of course, another challenge to finding a potential proto-World. If we find linguistic universals, are they remnants of proto-World, or are they constraints of how human brains do language?