In conversation at the Hay Festival in Wales this May, the English poet Simon Armitage made an arresting observation. Discussing the nature of language and why it is so good at capturing the experience of being alive, he said: ‘My feeling is that a lot of the language that we use, and the best language for poetry, comes directly out of the land.’ Armitage was placing himself within the Romantic tradition’s understanding of the origins of language, which argues that words and grammar are not the arbitrary inventions of human brains and minds, but are rather suggested to human beings by nature and the cosmos itself. Language is an excellent way to understand the Universe, because language springs from the things it describes.
The English philosopher Owen Barfield, a member of the Oxford Inklings in the 1930s and ’40s, whose work as a philologist convinced him that the Romantic tradition was broadly right, put it succinctly. Words have soul, he said. They possess a vitality that mirrors the inner life of the world, and this connection is the source of their power. All forms of language implicitly deploy it. Poets are arguably more alert to it because they consciously seek it out.
It’s an insight with radical implications for theories about the origins of language, primarily because the dominant hypotheses in modern science regard words very differently, as soulless signs that act as labels for objects and symbols that facilitate cognitive agility. The English evolutionary psychologist Robin Dunbar summarises the main two approaches: ‘Historically, the consensus has been that language evolved to allow humans to exchange factual information about the physical world, but an alternative view is that language evolved, in modern humans at least, to facilitate social bonding.’ In short, language as we know it emerged because of what it could do for Homo sapiens, because of its utility. It increased our ancestors’ chances of survival by enabling them to hunt more successfully or to cooperate more extensively. Language meant that things could be explained and that plans and past experiences could be shared efficiently.
These benefits, however, have nothing to do with the soulfulness of words, and when words attribute personality or agency to phenomena, say by describing the wind as ‘listless’ or mountains as ‘wise’, they are not reflecting the inner life of the world but fabricating it. Such metaphors add colour to life but are essentially fanciful and inventive and, further, often lead to erroneous inferences about reality – say, that the wind is a spirit or the mountain a god. The Israeli historian Yuval Noah Harari is a prominent advocate of this view: ‘the truly unique feature of our language is … the ability to transmit information about things that do not exist at all,’ he writes in his bestseller Sapiens: A Brief History of Humankind (2011). ‘This ability to speak about fictions is the most unique feature of sapiens language.’ On this view, language doesn’t come out of the land. It comes out of (often misguided) human brains.
Dunbar and Harari advocate what might be called the ‘grunt theory’ of the origin of language. The working hypothesis goes something like this: early human beings, or perhaps our evolutionary cousins, began pointing at bananas or lions, and grunting. These grunts became signs, in a process akin to the signs for ‘shoe’ or ‘eat’ that some lab chimps can be gradually trained to recognise over the course of many years. Then, the hypothesis continues, there was a cognitive revolution or ‘great leap forward’, as the American anthropologist Jared Diamond calls it in Guns, Germs, and Steel (1997). Suddenly, the brain rewired. It allowed for a shift from the deployment of a few signs to the heady artfulness of speech and grammar. It facilitated language proper, which could now emerge along with its many solution-solving, survival-aiding, practical advantages. It was like giving the mind a Swiss Army knife, says another champion of these evolutionary explanations, the Canadian-American cognitive scientist Steven Pinker in The Language Instinct (1994).
After grammar came metaphor, in a period sometimes called the ‘metaphor phase’. Now the developing human brain began to make fanciful leaps from the words it had invented for physical objects. It began using them to name the less tangible feelings of its mind and, in so doing, to manufacture immaterial, spiritual realities. It’s as if a whole bunch of prehistoric poets got to work, generating ingenious memes that spread across the world.
The fancifulness of this phase is a key part of the grunt theory, which generally assumes that early humans were wildly superstitious and so made all sorts of mistakes about the nature of reality. They thought that winds were spirits, or mountains gods, and they are assumed to have been bonded together through these daydreams and delusions, which, often enough, helped them to survive. (Better to flee believing that the leaves in the grove are shuddering because of the presence of a demon than not flee and so be at risk when they’ve been moved by a predator.) This kind of assumption explains how, say, pneuma in Ancient Greek can mean ‘spirit’ but also ‘wind’, and sometimes ‘breath’ – a sharing of meaning that is found widely across ancient languages.
Finally – the theory continues – came literature, when words were redeployed solely to express the interior worlds of human beings. This is the most dramatic instance we know of that exceptional phenomenon, consciousness, which somehow emerged out of an otherwise nonconscious cosmos.
It should be said that many of the details of the pathway tracked by this kind of evolution of language are contested, with some who work in the area quietly admitting that, in truth, the origins of language remain as mysterious as the origins of life. But still, there’s a widespread assumption that words and language emerged along an evolutionary route at least a bit like this one. The grunt hypothesis, or something like it, rules.
The Romantic theory of language suggests an alternative pathway. Its (admittedly minority) proponents argue that the grunt hypothesis doesn’t add up theoretically, and that a lot of evidence mitigates against it. There is the obvious logical objection that if language were originally a means of speaking about fictions, who’s to say that what we talk about now, including science, isn’t itself an elaborate falsehood? Science would thus work not because it is true but because it provides a technological society with a common discourse that facilitates social bonding and survival-aiding byproducts. In short, if language deludes, science would be strewn with delusions, too. Or to put it another way, the grunt theory saws off the branch it sits on.
In his work as a philologist, Barfield discovered that words, so far as they can be traced back, never pointed to physical things alone or acted as arbitrary symbols. There was no ‘metaphor phase’ either because words never were just signs, the offspring of grunts. They always seem to have had both physical and inner meanings, and to have had an aboriginal poetic charge. ‘Early man did not observe nature in our detached way,’ wrote Barfield in The Rediscovery of Meaning (1977). ‘He participated mentally and physically in her [sic] inner and outer processes.’ He’s not alone in this observation. Thinkers as different as the American transcendentalist Ralph Waldo Emerson and the British utilitarian Jeremy Bentham took similar note. Bentham summed it up in this way, in his posthumous Essay on Language (1843): ‘Throughout the whole field of language, parallel to the line of what may be termed the material language … runs a line of what may be termed the immaterial language … [T]o every word that has an immaterial import, there belongs, or at least did belong, a material one.’
Pneuma is an example of an old word that originally had an immaterial and material meaning. But consider another: the very old word ‘heart’ – cardium in Latin, kardia in Greek, and its precursors. Today, it has two distinct meanings. It refers to a physical organ, akin to a pump, housed a little on the left side of the chest cavity. But it also denotes a feeling, as in ‘warm hearted’, ‘cold hearted’, ‘soft hearted’, ‘hard of heart’.
The grunt theory supposes that the second, inner meaning is figurative or metaphorical, which is to say that it’s a later addition to the physical, literal meaning of the word. Only, the evidence points in the opposite direction. The felt meaning is as old as the word itself because, in ancient use, the heart is not regarded as a pump but as the seat of the emotions. Today’s mechanical meaning is the late arrival because the heart was not thought of as a pump until the early 17th century, following the discovery by the English physician William Harvey of the circulation of blood (who also, incidentally, continued to regard the heart as a spiritual organ).
If metaphors were mere fancies of the individual brain, how could they have possibly caught on?
Plenty of evidence, therefore, implies that the evolution of language moves in the opposite direction to the grunt hypothesis. Inner meanings were there from the start. This makes sense in another way, because if you subscribe to the literal-signs-to-fanciful-metaphor history of language, you might expect to find fewer metaphors in Homer and more in, say, Virginia Woolf and Harold Pinter – but you don’t.
Now, such counterevidence can also be challenged. After all, it’s hard to go back more than a few thousand years to directly track the evolution of words. But a further reason for rejecting the grunt hypothesis can push back further. It arises from considering the nature of metaphor, and the obvious observation that there is no point in having a simile or metaphor that is suggestive only to you. If I were to compare my lover to, say, an aubergine, you’d suspect I was talking nonsense, which you wouldn’t if I compared my lover to a summer’s day. This is to say that if metaphors were, originally, mere fancies of the individual brain, and therefore chaotic and idiosyncratic, how could they have possibly caught on? And if they were merely products of the individual brain, then today we’d all be speaking our own private languages.
Therefore, metaphors must work because they raise inner life to conscious awareness by innovatively combining meanings that release the innate poetry of words, and the world. In Hamlet, Shakespeare writes of the morning as like a ‘russet mantle’ that ‘walks o’er the dew of yon high eastern hill’. His metaphor works if you know what the morning is and what a russet mantle is, and can mentally put the two together in a meaningful way. The combination prompts what was half-known but, previously, not fully thought. ‘The morning as a russet mantle?’ we wonder, before continuing: ‘Of course!’ Shakespeare doesn’t create a new sense of the morning. He reveals something about the inside of the morning that previously lurked in the shadows, undisclosed.
Or consider again the conflation of ‘wind’ and ‘spirit’, and the assumption that our distant ancestors must have first felt the wind and then, by inventive extension, used the same word to refer to spirits. That begs the question of why they didn’t already have a word for ‘spirit’, if they were so superstitious. More specifically, it raises the question of how the word for ‘wind’, which the grunt theory insists originally had a strictly literal meaning of ‘physical breeze’, could become a globally used metaphor for ‘immaterial spirits’. It is impossible to believe that there wasn’t a spontaneous affinity between the notion of ‘wind’ and the notion of ‘spirit’, and that this is why there’s a link. The two must have been born together.
If this is so, then it implies that the origins of language are not rooted in grunt and sign references to objects. Instead, early words always were loaded with the inner and outer meanings that our ancestors detected in nature and consciously articulated when they first started to speak. They are the tokens of a prehistoric communion. The poles of meaning only subsequently split apart, when our very much more recent ancestors, on the cusp of a modern, scientific consciousness, stopped experiencing the world in the natural way, and imposed the modern dualism.
An alternative theory of the evolution of language is required. As the English palaeobiologist Simon Conway Morris writes in The Runes of Evolution (2015), the dominant explanations for the origins of language are inadequate for the very reason that they are essentially utilitarian and materialistic. It would be better to assume what language itself tells us. It is innately meaningful because its poetry enables us to perceive deeper structures of reality. Words channel the vitality of the natural world. They have soul because nature does – for all that, these days, we struggle to feel it and are quite inclined to disbelieve it. It is a recognition of this loss that has led the English writer Robert Macfarlane on a project to rescue words. As he explains in Landmarks (2015): ‘[What] is lost along with this literacy is something precious: a kind of word magic, the power that certain terms possess to enchant our relations with nature and place.’
Narrower Darwinian explanations offer inadequate descriptions of the evolution of language because they are blind to what’s really driving its development, which is the way that words and grammar participate in and spring from the things described and the experiences shared. And thinkers who aren’t trapped in reductive traditions are putting alternative possibilities forward. Again, the details are unsettled and disputed. But overall, it probably looks something like what the American sociologist Robert Bellah in Religion in Human Evolution (2011) called the offline theory.
Our distant ancestors had to spend much of their time ‘online’. They had to forage, fight, flee and procreate. But they also had plenty of time ‘offline’. They played, as many other animals do. Or they engaged in rituals and making music, which all the evidence suggests they did prolifically. Discoveries of bone pipes go back tens of thousands of years. Beads and stones even further. Or take sleep, an activity that typically takes up about a third of any 24-hour period.
What’s interesting about offline activities is that they don’t readily fit in with orthodox Darwinian frameworks of understanding because they don’t seem to yield much adaptive advantage. Play, rituals, music and sleep appear to have little utility and can even compromise the survival of the creatures that indulge in them. Because of this adaptive inefficiency, evolutionary psychologists tend to describe their emergence as ‘not well understood’. Alternatively, it could be argued that these activities don’t themselves have any function (Pinker called music ‘auditory cheesecake’), but serve instead to stimulate other activities that do aid survival.
Language arose when humans engaged with life not to survive but to participate in the great rush of it
The Romantic school rejects these explanations as fudges. Instead, its champions recognise that language is an offline activity, too. It differs from screaming alarm at a predator, or yelping delight at finding food – which are online. As Bellah puts it, language is not primarily an intervention in the material world, because most speech does not refer to objects that are actually present. It is not a Swiss Army knife for the mind because, for the most part, it has little immediate practical use. Being offline, it is not adaptive, or maladaptive, or a spandrel byproduct of adaptive developments; or simply inexplicably there, as if it sprung into existence after a great leap forward following a chance rewiring of the brain. Rather, language is one of those phenomena that arose when human beings engaged with life not so as to survive but so as to participate in the great rush of it that they felt teeming around them. Human beings do not live on bread alone.
If this is right, then the route followed by the evolution of language is very different. First, there would have been not grunts and signs but music and ritual, akin to what is found in other social animals, from the displays of fish and reptiles to the songs of birds and whales. Further, these creatures appear not only to deploy music and ritual online, perhaps to attract a mate, but also offline, as though sexual selection also required aesthetic sensibility. It’s a possibility that Charles Darwin himself observed and speculated about. He became interested in the behaviour of female Argus pheasants, which seem to choose a mate by male display. However, Darwin’s observations led him to conclude that much more was going on than mere selection, because the male pheasants were clearly trying to charm their females. The birds were taking aesthetic pleasure in their dances. ‘Many will declare that it is utterly incredible that a female bird should be able to appreciate the fine shading and exquisite patterns. It is undoubtedly a marvellous fact that she should possess this almost human degree of taste,’ Darwin wrote in The Descent of Man (1871). He himself did not find it incredible. Careful observation made it obvious.
From this soil, music and ritual would have sprouted words and language. Presumably, this happened as the intelligence of early hominids, maybe only Homo sapiens, allowed for music and ritual to evolve into stories and myths, perhaps in a push to explore the wind-spirit that fills all things. Such language would have been driven not primarily by the need to survive but by a drive to share in the life of the natural world. Early people must have felt immersed in it – an experience that is momentarily recaptured for us in, say, rituals. Think of the power of lighting a candle or of tying a clootie – a rag or ribbon – to a sacred tree. Or of hearing a line of poetry: ‘But look, the morn, in russet mantle clad, Walks o’er the dew of yon high eastward hill.’ At such times, in such moments, our inner lives momentarily merge once more with the inside of the world.
After stories and myths came thought. Finally, as a very last stage, arrived literalness, when the words originally formed by partaking of the inner life of nature took on meanings more or less divorced from it. They were applied to denote the inner experiences of human beings, and eventually to the inner experience of human beings alone. ‘Wind’ ceased to mean ‘spirit’; ‘heart’ could become an organic pump.
You could say that the development is not from grunts to signs, to grammar, to metaphor, to literature. It is from music to myth, to thought, to literalness, to spiritual isolation. It’s that final twist that has led science now to assume that primitive pointing came before primordial poetry; meaningless signs before living symbols; practical needs before an awareness of relationship. But in truth, language sprang from an involvement with life. ‘The first metaphors were not artificial but natural,’ summarised Barfield.
If it’s a radical theory, it has radical ramifications. It suggests that words that have come to describe the inner life of human beings can have evolved only if the cosmos is full of spirit. The first humans were not inventive onlookers: they were intelligent participants in that wider consciousness. Their difference from other creatures lay only in a capacity self-consciously to communicate the meaning they felt pulsing around them. In History, Guilt and Habit (1979), Barfield concluded that the evolution of words ‘always points us back to a cultural period when there was a much closer interpenetration between thinking and perceiving than is the case with us today’. Or as Alexander von Humboldt, one of the key figures at the origins of Romantic science, realised: ‘Nature everywhere speaks to man in a voice familiar to his soul.’
Further, it is only because we are embedded in nature, materially and imaginatively, that science can be done. The idea of the scientist standing outside of nature and being able to understand it makes no sense. ‘I myself am identical with nature,’ said Humboldt. That’s why language arose. It does, indeed, come out of the land. That’s why it has its power, and works.