Did teaching drive the evolution of language?

Learning from each other is a key part of human success. Language helps with teaching now, but was it enough of a force in the past to influence evolution?

Teaching is something we’ve all experienced. We may not always have enjoyed it, but there’s no denying its importance in passing on information. Now, imagine trying to teach someone in silence. Could you still pass on that information effectively?

And no, you can’t learn from reading. That’s cheating.

Trying to teach without language seems like a tall order. As such, many have argued that the two go hand in hand. Language evolved to help us pass on information better, faster, harder, stronger.

However, just because it sounds right (pun intended) doesn’t mean it actually is. Language is also useful for a whole bunch of other stuff, like socialising. Or reading Filthy Monkey Men. So, how can we figure out which of the many uses of language is the primary one (or ones) that drove its evolution?

Teaching seems like an obvious benefit

First, we need to address the elephant in the room. Yes, the ability to pass on information that could aid in your survival does seem like an obvious way language could aid in your survival. So why wasn’t teaching considered the driving force of its evolution?

Well, despite how intuitive it sounds the evidence really wasn’t there. Many experiments have shown language is important in learning, but only in modern settings. When trying to teach someone how to use ancient stone tools, language seems to become a lot less important. Prior to 2017, there were three studies on these sorts of “authentic” examples of teaching. 2/3 indicated you could learn just as well without verbal language.

Like I said, just because something sounds right doesn’t make it so.

Studies on how modern people use language also seem to reinforce this. When we talk to each other, it’s most often aimed at socialising, bonding and so forth. We devote relatively little of our linguistic energy to teaching. Plus. when you look at our ape relatives we see a similar pattern; with most of their interactions reinforcing social bonds rather than teaching.

In fact, some have inferred when language evolved based on those similarities. Primates traditionally bond through grooming, but can only groom one ape at a time. This places a limit on how big their groups can be before they can’t groom enough people in the group to keep it together.

We live in groups larger than this limit, meaning we needed to develop something that could allow us to socialise a lot more efficiently. Enter language:

How much time we'd have to spend grooming if language didn't evolve. 20% is the cut off time for what is plausible

How much time we’d have to spend grooming if language didn’t evolve. 20% is the cut off time for what is plausible

Case closed?

So, it looks like language-based teaching wasn’t a big deal in the past but language-based socialising was. Is the matter settled? If you’d asked me 5 years ago I probably would have said yes. But there’s an interesting pattern in this research which suggests something else is going on.

Like I said, a few experiments have investigated how important language is for teaching people how to use stone tools. Now, there’s a lot of stone tools out there and each experiment picked a different one to study. Interestingly, those which examined newer, more complex tools found language was less important. People who learnt by watching were just as good at making Neanderthal technology as those taught verbally. Yet repeating the experiment with simpler Oldowan tools found language was better at teaching.

The Oldowan is mode 1 whilst Neanderthals made mode 3. Yet the importance of teaching doesn’t follow this same trend.

This seems completely counter-intuitive. Surely a more complex item would benefit from better teaching?

Well, it may be the case that the Neanderthal technology is just too hard to make. It can take years to learn how to make their tools. As such, even after a bit of teaching, both groups would still be rubbish. The impact of a small amount of better teaching is negligible. However, people can pick up Oldowan tools faster, so the impact of teaching becomes more pronounced so it can be detected in these studies.

In other words, the case isn’t quite closed on the role of teaching in language evolution.

So this new study…

Palaeanthropologists wanted to try and solve this issue once and for all. So in 2017, they came up with a new, super-rigorous study.

Crucially, they tried to take into account all of these confounding factors which might be confusing earlier work. For example, they made sure everyone started with the exact same raw materials. And they were teaching them a method they knew could be learnt in the time frame studied.

Besides these tight controls, the experiment followed the pattern seen before. They divided up a bunch of Palaeolithic newbies into groups. They then taught them how to make the stone tools our ancestors created; varying the amount of language, imitation etc. that was allowed per class. After having their lesson, each group went off to try and make tools by themselves.

The teaching methods used, arranged in increasing complexity from left to right

The teaching methods used, arranged in increasing complexity from left to right

The results are quite interesting. When they tried to make tools during the lesson, all the groups did very well. However, when they went off by themselves things began to change, but not all in the same way.

For instance, the non-teaching “imitation-emulation” group made more attempts. However, they were much less successful at it than those who had verbal or gestural teaching. As a result, the taught groups would up making twice as many tools. Notably, this was true for both teaching groups. They both produced a similar increase in knapping skill.

What’s more, the resulting tools were of a higher quality for the taught groups. This shows they also picked up an understanding of the method that the imitation group lacked.

Flakes from: (A) teacher, (B) imitation-emulation group, (C) gestural teaching, and (D) verbal teaching. Note how the taught groups (C & D) more closely match the size and the shape of the teacher (A).

So can we forget about sociality?

Based on all this, the researchers argue that this shows teaching was a big deal. As such, they conclude that these benefits are so significant that evolution wouldn’t have ignored them.

Sure, socialising may have kicked off the evolution of language (and likely continued to influence it) but soon after the benefits of teaching would also begin to drive its development. But it’s worth remembering there were likely many other influencers too. Evolution is a complex thing and it’s likely many factors drove and influenced the evolution of language. This research simply adds one more to the pile.

Crucially, this also shows that this drive would have existed fairly early in our technological evolution. As such, the evolution of language may have similarly early. Maybe even ape-like Australopithecus was taking steps towards conversation.

Although it does have an interesting implication: if teaching is driving the evolution of language, could our focus on old languages be holding it back?

Clearly, it’s time we took Shakespeare out of schools. Evolution demands it.


Dunbar, R. I. (2003). The social brain: mind, language, and society in evolutionary perspective. Annual Review of Anthropology, 163-181.

Lombao, D., Guardiola, M. and Mosquera, M., 2017. Teaching to make stone tools: new experimental evidence supporting a technological hypothesis for the origins of language. Scientific Reports7(1), p.14394.

Mesoudi, A., Whiten, A., & Dunbar, R. (2006). A bias for social information in human cultural transmission. British Journal of Psychology, 97(3), 405-423.

Morgan, T. J. H., Uomini, N. T., Rendell, L. E., Chouinard-Thuly, L., Street, S. E., Lewis, H. M., … & Laland, K. N. (2015). Experimental evidence for the co-evolution of hominin tool-making teaching and language. Nature communications, 6.

Related Articles

8 thoughts on “Did teaching drive the evolution of language?”

  1. Artem Kaznatcheev says:

    It is interesting that teaching is not considered a social benefit, especially with things like Rogers’ paradox bringing up the same dilemmas as more standard social interaction. This would be especially moot in settings where the learning is not about external-environment manipulating tools (as in this study) but about the norms of the tribe: “Don’t talk about Grok’s bald-spot or he’ll show you how fists are good for hurting”.

    A quick question about the verbal teaching condition in the study. Did they constrain the vocabulary that could be used? Presumably our set of communicable abstractions and metaphors is much richer than that the humans of the palaeolithic; this rich suite of language tools would have had to bootstrap on something. I would be interested to see a condition between e and f, where the teacher is allowed to use language but the student and teacher don’t speak the same (or similar) language.

  2. Jim Birch says:

    There are two functions or language that can be critical here: teaching and retention. In the case of flaking the transmitted knowledge becomes (mainly) procedural rather than semantic memory, but in the case of a remembered narrative that say gives survival instructions for a 20-year extreme weather event it is retained semantically as a narrative.

    We could see language as having begun as extended grooming, then adding procedure instruction capability, then adding narrative retention. It would be interesting to work out how important each function is and was. It seems to me that modern life is lived as at least much in a virtual universe of abstracted narratives – eg, money, law, culturally defined relationships, etc – than in a physical world. At some point in the past in must have been pretty much a case of dealing with the exigencies of physical reality.

  3. harikrv says:

    Language may have been used more for dividing people rather than uniting them or passing on knowledge. A large part of our talk in day to day life is, whether we like it or not, lying. It is more like I want to hide my real intentions by saying something entirely different from what I mean. There is also the phenomenon of enormous diversity in languages – languages evolve into local dialects which change sufficiently to emerge as new languages in course of time. This could be due to one group wanting to hide their thoughts and occurrences from all others and hence wanting to have their own unique words and syntax. There is no reason to believe that it was any different at the dawn of evolution.
    Most of our speaking is done to ourselves. Everyone has a running conversation in their heads, a voice that speaks non-stop in the language most comfortable to each one of us. This could point to the probability that language evolved initially to make sense of the world around us or in other words safeguard ourselves from external dangers. Probably speaking began when we used the same language running in our heads to threaten our fellow humans with violence in extreme conflict situations ( expletives and curses are, no doubt, the most spoken words by mankind everywhere)

  4. Andre Salzmann says:

    It is interesting to take note of what the primary communication in a troop of baboons is about. It is as interesting to look at studies done on how distinctly dogs communicate and how well humans understand there ” language”. The ” bushmen” communicate with click sounds.

  5. Leslie Randall says:

    I’d like people to remember that adults can’t acquire language. After the age of 5 it’s almost impossible to be fully linguistic. It was children who began talking, and their mothers who responded to their sounds. Early language probably developed through children’s play and social interactions, and gradually was incorporated into adult life, and put to new and more sophisticated uses, including teaching. New language abilities, and the expansion of simple syntax, was undoubtedly all accomplished by children.

    1. Adam Benton says:

      That is certainly a very interesting point; although you have to remember we have now spent (maybe) millions of years with language. How much can we apply those observations to the first speakers?

      1. Leslie Randall says:

        I don’t know, and there is no hard evidence. But given that technology moved so slowly to develop, I imagine that language similarly moved quite slowly at the beginning. 1.7 million years ago (mya) the first Acheulian toolkit consisted of large stone flakes, sometimes further shaped; hand axes – essentially sharpened rocks held in the hand. Given that this crude technology held sway until 400,000 – 250,000 years ago, it’s hard to imagine that language use developed much faster. Brain size would suggest that Homo erectus around 1 mya achieved a brain size of around 1000-1200 cc, well below modern humans. I think the signs of culture (burial, adornment) found among Neanderthals is a clearer indication that they had attained some cultural, social and linguistic sophistication. In my view, language likely puttered along, until it reached a tipping point, probably fairly recently, when (as with technology) innovations started to build on one another more quickly. But it’s just a guess…

  6. Brett Martin says:

    “Symbolism is the language of the Mysteries; in fact it is the language not only of mysticism
    and philosophy but of all Nature, for every law and power active in universal procedure is
    manifested to the limited sense perseptions of man through the medium of symbol…. By symbols men have ever sought to communicate to each other those thoughts which transend the limitations of language. Rejecting man-concived dialicts as inadequate and unworthy to perpetuate divine ideas, the Mysteries thus chose symbolism as a far more inginious and ideal method of preserving their trancendental knowlage. In a single figure a symbol may both reveal and conceal, for to the wise the subject of the symbol is obvious, while to the ignorant the figure remains inscrutable. Hence, he who seeks to unviel
    the secret doctrine of antiquity must search for that doctorine not upon the open pages of books which might fall into the hands of the unworthy but in the place where it was originaly concealed.”-The Secret Teachings of All Ages, Manly Palmer Hall.

    The Symbols are in the stone tools and on the rock cave wall’s.

    Onomatopoeia is also worth consideration.

Leave your filthy monkey comments here.

More in Ancient technology
First meat-eating humans were actually crocodiles

Damage to bones indicates our ancestors were eating meat more than 3 mya. However, it turns out crocodiles can cause similar damage, challenging this