How Old Is Language?

On time machines, talking neanderthals, and the long(ish) past of language.

Posted Feb 19, 2015

Until relatively recently, it was not considered good form to speculate on the origin of language.  And, in part, this was a sound position to adopt.  By the time that Charles Darwin published On the Origin of Species in 1859, and his subsequent publication, The Descent of Man in 1871, all manner of exotic theories had been put forward to explain the origin of language, replete with some equally ludicrous names.  My personal favourite is the ‘ta-ta’ theory of Sir Richard Paget.  Paget, who was influenced by Darwin, proposed that language may have arisen as an unconscious, vocal imitation of specific body movements, for instance, the way my tongue sticks out when I attempt a task that leaves my thumbs-for-fingers confused, for instance trying to thread a needle, or play the guitar. 

Indeed, Darwin himself was not immune, speculating that spoken language may have arisen from our ability to produce song—but it seems unlikely that singing gave rise to language.  Recent research, for instance, demonstrates that ‘amusia’—an impairment of the brain that results in the inability to process musical cadences and recognise music, and ‘aphasia’—the loss of language—are unrelated.  A person can lose the ability to use or comprehend language, but still retain, perhaps counterintuitively, the ability to recognise music.  Indeed, the nature and organisation of music, and spoken language appear to diverge in a host of ways--although some scientists do, nevertheless, see value in Darwin’s musical origins of language hypothesis.
When the Linguistic Society of Paris was founded, in 1865, it famously included a prohibition against speculating on the origin of language in its constitution:   “Article 11: The Society will accept no communication…dealing with the origin of language”.  This stance, at the time, made good sense, being later adopted by The Philological Society of London.  In 1873, the President of the Philological Society, Alexander Ellis, declared that “We shall do more by tracing the development of one work-a-day tongue, than by filling waste-paper baskets by reams of paper covered with speculations on the origins of all tongues.”           

Speculation on the evolutionary basis of language remained off-limits for over a century.  But, by the 1970s, scientists began to again speculate on the origin of language.  Today, the scientific literature is brimming at the seams with a whole host of theories about how language may have come into being.  The cacophonous din of different accounts of language origins, range from the sublime—language evolved to enable us to engage in gossip, proposed by Professor Robin Dunbar—to the ridiculous—the aptly named Bow Wow Theory, attributed to the German philosopher Johann Gottfried Herder.  This contends that language may have arisen from sounds used to identify animals, perhaps based on onomatopoeia, which eventually became their names.  But, in the last decade or two, new findings, have begun to shed light on this most thorny of issues. 

In my previous blog post I looked at the possible pre-cursor for language, what I dubbed cooperative intelligence.  But in this post I address a slightly different issue: how old is language?  Put another way: what is the time-depth of language: language, more or less, of the kind we recognise today?  And in so doing, I’ll consider what we now know about the linguistic capacities of our now extinct sister-species: Homo neanderthalensis.

Oh, for a time machine!
Aficionados of the BBC TV science fiction series, Doctor Who, will be familiar with its novel approach to the concept of time travel: the eponymous Doctor is a Time Lord from the now extinct planet Galifrey.  The Doctor travels through time in his Tardis—a time-travelling spaceship, which is famously larger on the inside, a feat made possible by time-lord technology.  But until relatively recently, in lieu of a Tardis, it was almost impossible to say anything meaningful about the evolution of language, as I observed at the outset of the chapter.  But in the last decade or so, with more recent fossil finds, and with advances in the genetic dating of ancient DNA, the picture has begun to change.
Until recently, it had been fairly widely assumed that human-like language was a very recent evolutionary development. One important reason for thinking this was that it had been assumed that language was absent in Homo neanderthalensis (‘Neanderthal man’).  And if Neanderthals lacked language, then the presumed, last common ancestor of humans and Neanderthals, Homo heidelbergensis, must also have lacked language.  Consequently, the emergence of language must have been a uniquely human innovation.

Genetic evidence for early modern humans points to our species being around 200,000 years old.  Moreover, from around 50,000 years ago—the period referred to by archaeologists as the Upper Paleolithic—an unprecedented cultural explosion began to manifest itself in human communities.  This resulted in art work, sophisticated jewelry, advanced stone tool technology, evidence of complex ritual systems and social structures, fishing and boat-building, the manufacture of projectile javelins for hunting, and other trappings of a relatively sophisticated material culture.  The conclusion was clear: language must have emerged sometime after 200,000 years ago and prior to this cultural ‘big bang’, some 50,000 years ago.

Dumb Neanderthal Cavemen?
There were three main reasons for thinking that Neanderthals lacked a (spoken) language capacity.  For one thing, Neanderthals were robust creatures, which implied great physical strength, but at the expense of smarts.  Moreover, early reconstructions of the fossil vocal tract of Neanderthal specimens, in the 1960s and early 1970s, seemed to suggest that they lacked speech capacities, implying a lack of language, full-stop.  And finally, there appeared to be a large gap between the cultural products and capabilities of Neanderthals, compared with Upper Paleolithic humans—again implying reduced mental acuity, and hence, a lack of language. 

But as the evidence has come in, it is now becoming clear that  Neanderthals may, in fact, have had a spoken, human-like, language capacity; and from that, it also follows that the common ancestor of humans and Neanderthals also had language: spoken language may be older than previously suspected.  Moreover, this conclusion fits with the proposal of the comparative psychologist, Michael Tomasello, that Homo heidelbergensis was the first species of Homo to have a fully-optimised joint intentionality cognitive strategy, as I discussed in my previous post: Cooperative Intelligence: The Pre-cursor for Language?
Recognisable Neanderthal fossils have been found in Europe from around 400,000 years ago.  Their physique was shorter and stockier than the more gracile, later arriving Homo sapiens, most likely an adaptation to the severe glacial European environment of the last ice age. They occupied regions, at various points in their existence, ranging from as far north as sub-arctic regions of Europe, as far east as Siberia, and as far south as the Middle East.

In contrast, the earliest modern looking human fossils have been found in Omo from around 200,000 years ago, and in Herto, both in modern-day Ethiopia, dating to around 160,000 years ago.  Human fossils have been discovered in the Middle East from around 100,000 years ago, suggesting the beginning of an out-of-Africa dispersal.  And by around 70,000 years ago, humans began to disperse around the Old World, reaching glacial Europe sometime before 40-50,000 years ago.
Scientists once thought that Neanderthals and Homo sapiens were completely different species.  But with advances in genetic testing techniques, and the mapping of the Neanderthal genetic sequence based on ancient DNA samples, we now know that the picture is far more complicated.   Moreover, it’s now becoming clear that Neanderthals may have been, in many ways, the cognitive and linguistic equals of the new influx of early Homo sapiens, entering Europe for the first time. 
It appears that early modern humans regularly interbred with the prior existing communities of Neanderthals they came across in Europe.  Evidence for this comes from genetic testing of modern humans and comparing it to Neanderthal DNA.  It turns out that non-African living humans have a higher admixture of characteristically Neanderthal genetic material, consistent with an inter-breeding pattern for those early humans who had left Africa.  On average, non-African adults share between 1 and 4% of their DNA with Neanderthal DNA, with an admixture of 6.4% in European adults, and an even higher admixture ratio of nearly 10% in Asian adults. Moreover, different Neanderthal genes are found, to varying extents, in different individual adult humans today.  According to one prominent expert, what this means is that “the number of [breeding] contacts was not very small—more like low thousands or high hundreds than dozens”.
Other evidence for an inter-breeding scenario comes from the fossil record.  In one famous find, the fossils of a child’s burial were discovered in Arbrigo de Lagar Velho, in Portugal.  This child exhibited features intermediate between humans and Neanderthals, suggesting that it shared one human and one Neanderthal parent.  Moreover, the nature of its burial, suggests that it was a fully integrated member of its community: there appears to have been no evidence of stigma attached to this ‘mixed race’ child.  This indicates that human/Neanderthal interbreeding was common-place. 

In terms of the widely accepted definition of a biological species, introduced by the evolutionary biologist Ernst Mayr, a species is one that can only successfully breed amongst its own members.  On this definition, the horse and the donkey are distinct species as their offspring, a mule, is typically infertile and incapable of reproduction.  But in contrast, genetic and fossil evidence of successful interbreeding between humans and Neanderthals strongly suggests that these two groups were not distinct species, from a biological perspective.  Indeed today, many archaeologists and biologists treat them as related sub-species, rather than distinct species.  The up-shot, of course, is that if humans and Neanderthals successfully interbred, and otherwise interacted to facilitate breeding, then the neurobiological differences are likely to have been less significant than was at one time suspected.

The second line of evidence indicating that Neanderthals had modern, speech-like capabilities comes from new evidence relating to their articulatory and auditory capacities.  Species with sophisticated communication systems—as in the case of modern humans—have broadcast capacities (e.g., speech) that match reception capacities (e.g., hearing).  For instance, human auditory capabilities are particularly attuned to the sound range for speech production, with unrivalled sensitivity, amongst other primates, in the 1-6 KHz range.  In recent reconstructions of the outer and middle ear of five fossil Homo heidelbergensis specimens, it was found that heidelbergensis, the presumed common ancestor of both humans and Neanderthals, would have exhibited an auditory capacity in the modern human range.  Moreover, analysis of fossil human and Neanderthal ear ossicles, found in Qafzeh and Amud, in the Middle East, and around 50-100 thousand years old, are essentially identical to those of humans.  This implies that Neanderthals had a modern hearing capacity.

In the early 1970s, reconstruction of the Neanderthal vocal tract led scientists to believe Neanderthals to have been incapable of spoken language.  The argument was based, in part, on the reconstruction of the tiny hyoid bone, essential for controlling and coordinating tongue movements in human-like speech.  In Neanderthals, it was believed the hyoid bone hadn’t yet descended to its human-like resting place, making human-like speech impossible. 

However, since then it has been established that the assumptions underpinning this  reconstruction were flawed. In fact, a hyoid bone located higher in the skull is associated with an earlier evolutionary feature exhibited by ancestral humans, namely air sacs connected to the vocal tract.  The presence of air sacs reduces the ability to produce articulate speech.  This was likely to have been a feature of the anatomy of the much earlier Homo erectus (1.8. million years ago), but had most likely disappeared by the time of Homo heidelbergensis.  This reveals that while Homo erectus had at best a limited speech capacity, later species of Homo, including Homo heidelbergensis, may have already developed some speech-like capacity.

A further line of evidence comes from the size of the thoraic vertebral canal.  Spoken language requires voluntary control of the lungs and breathing, enabling, in modern humans, the expulsion of air from the lungs to produce speech sounds.  But while breathing is under involuntary control in most species, humans have a larger thoracic vertebral canal.  And this is a direct result of additional nerve control of the intercostal muscles, and diaphragm, enabling voluntary control and hence speech.  A well-preserved Homo erectus specimen, the so-called Nariohotome boy lacks an enlarged vertebral canal suggesting lack of the voluntary control required for the articulatory apparatus.  However, Neanderthal fossils, like modern humans, provide evidence of an enlarged canal, suggesting that Neanderthals were also capable of the voluntary control of breathing, a pre-requisite for speech
Finally, recent archaeological finds cast considerable doubt on the received view that Neanderthals produced a meagre material culture, lacking in sophistication compared to that of Homo sapiens.  For one thing, there is emerging archaeological evidence, based on analysis of finds in Border Cave in southern Africa, that some elements of the later Upper Paleolithic culture in Eurasia were already present in Africa 75,000 years ago, prior to Homo sapiens entering Europe.  This includes evidence for pigment use, beads, engravings, and sophisticated stone and bone tools.  Moreover, the material culture that produced such artefacts disappears from the archeological record around 60,000 years ago, before reappearing later.  This suggests a non-linear development of a rich material culture.  Moreover, it also means the assumption that the apparent cultural explosion in Eurasia around 50,000 years ago may not have been a unique event, nor necessarily unique to modern humans. 

As one team of experts have observed, it is becoming increasingly clear that Neanderthals “exhibited many complex behaviors (pigment use, funerary practices, complex hafting techniques, wood-working, personal ornamentation, and bone tool manufacture) before or at the very moment of contact with modern humans.”  They possessed, and made use of, a sophisticated stone-tool technology.  The manufacture of stone tools of this type involved up to 50 distinct actions.  And in modern training experiments, it takes adult humans several months of training to master the techniques that would have been deployed by Neanderthals to fashion their stone-age tools.   

Neanderthals manufactured clothing by sewing animal skin, as well as footwear. They buried their dead, and may well have left grave offerings. They manufactured jewellery by painting shells with red ochre which they then perforated to be worn.  Moreover, decorated pendants found in sites occupied by a Neanderthal community, and dating to 50,000 years ago, have been discovered in Spain, long before Neanderthals had contact with early humans in that region of Europe.  Neanderthals appeared to live in small, social communities, with married couples living with the husband’s parents.  They built huts, with complex foundations, they used pitch to haft their tools, which they extracted by fire, and even mined materials to manufacture their tools, of up to two metres in depth.

There is also recent evidence of cultural borrowing by Neanderthals from the influx of Homo sapiens.  In the late Neanderthal period, the Neanderthal Mousterian stone-age industry was giving way to a more complex technology, dubbed the Châtelperronian, from the location in France where these more advanced artefacts were first discovered.  This particular technology was a blend of the older Neanderthal technology and the more advanced techniques that were arising in the human populations at that time, dubbed the Aurignacian industry.  None of this could, presumably, have been possible, without some linguistic basis. 

So, how old is language?
So, if Neanderthals had a modern-like (spoken) language-like capacity, then it stands to reason this derived from the common ancestor they shared with humans: Homo heidelbergensis.  Language of the sort we recognise today would seem, on this evidence, to be much older than previously assumed.  Rather than language being a very recent innovation—no more than perhaps 100,000 years old, on some oft-cited accounts—in fact, language may have a greater vintage, perhaps going back half a million years or so, possibly further.  And of course, the emergence of language, as I proposed in my previous blog post, was most likely emerging for much of the 2.5 million year lineage of ancestral humans, the lineage Homo.  It increasingly seems that modern-like language, and possibly also speech, has a much longer past than previously suspected or imagined.