What might the first human language have sounded like?
Last year I was involved in a project with Maggie Tallerman and Coppe van Urk, part of which involved creating an oral language for a fictional early modern human (Homo sapiens) tribe. The project in the end didn’t pan out (I’ve learned that in the world of TV and Films, little ends up working like you expect it to!), but Maggie, Coppe and I spent some time talking about how we might design such a language, and I even spent an enjoyable flight from London to New York creating a language along our design decisions, and translating quite a lot of material into it (I know, nerdtastic or what?). Even though the project ended up not going anywhere, it was fun to think about the issue, and I thought I shouldn’t let the work go to waste. Hence this blog.
Modern humans, we know, first appeared in sub-Saharan Africa. Maggie, Coppe and I decided that there was no reason why their languages wouldn’t use the same basic building blocks as modern languages. Early modern humans were more or less identical to us anatomically, and likely no different to us cognitively. That meant we made design choices about the language on the basis of the idea that this language worked just like human languages of the present. But we decided to give it features that might have been characteristic of a young, newly developed, language.
When the three of us were thinking about interesting features to put into the language, one obvious candidate was click consonants. These exist almost nowhere else in the world except Africa (there is an interesting exception, the ritual Damin language of the men of the Lardil people in Australia, which needs a separate blog post really). By clicks, I mean true click consonants, where the back of the mouth is closed off, some area near the front is also closed and a vacuum is created between. The closure at the front is sharply released, and air is drawn into the resulting vacuum. Think of the disapproving sound made in English, often orthographically represented as tsk tsk, or tut tut. Or the sound one makes to encourage horses to gee-up. Or the kissing sound. These are clicks (technically, velaric ingressive obstruents). Linguists love clicks — other people not so much. I still recall quite clearly a dinner of fourteen linguists in Los Angeles. My partner and my friend Scott were the only non-linguists present. My partner is used to linguists, but the look of horror on Scott’s face, when virtually everyone at the table started showing off their clicking prowess, will never leave me. Since clicks are only found in Africa, and that’s where our fictional archaic Homo sapiens tribe lived, Coppe, Maggie and I decided that our constructed language needed some clicks.
Maggie also had an idea about nasality. A lot of the sounds that humans vocalise in situations of hesitation (though not all) are nasal (think of English erm, hmmm, mmm?). Further, the usual airflow in normal breathing is a nasal airflow, and nasal sounds appear early when babies babble. Most tellingly, deaf babies, when they babble orally, use more nasal sounds, because they don’t have evidence that the language around them is predominantly non-nasal. These facts suggested to us the idea that we’d have a lot of nasals in our language. This fitted well with the decision to have clicks. To make a click, you close off the back of your mouth, meaning that any airflow comes out through the nose. This makes clicks very easy to articulate with a nasal airflow, and this led us to decide to nasalise our clicks. For extra measure, we also prenasalized some of our other consonants.
We decided to have four distinct places in the mouth where consonants would be articulated: the lips (labial sounds), the ridge just behind the back of the teeth (alveolar sounds), the hard palate in the centre of the mouth (palatal sounds) and the velum, or soft palate, at the back of the mouth (velar sounds). This gave us a consonant inventory that looks like this:
|Prenasalized clicks||m! [mʘ]||n! [nǀ]||–||–|
|Prenasalized stops||mb||nd||nj [ɲɟ]||ngg [ŋg]|
I’ve written each of the sounds here as a letter or combination of letters from the Roman Alphabet. Where these letters correspond to sounds that aren’t their usual values in English, I’ve put the relevant International Phonetic Alphabet (IPA) symbols in square brackets afterwards. There are some great online resources where you can hear the various sounds being made.
We also included two semivowels (w and y [j]), and a very simple vowel series, which is just the vowels i, a and u (where i is the ee sound in feel, and a and u are ah and ooh). For variety, we decided these vowels could be long or short. This gives us a vowel system like the following:
|Front High unrounded||i, ii||y [j]|
|Mid Low unrounded||a, aa|
|Back, High Rounded||u, uu||w|
Our basic idea was that this language, (supposedly!) spoken by archaic modern humans, would have appeared recently. We might expect it, then, to have the hallmarks of other very young languages. There would have been little time for processes that age languages to take place. We were constructing an oral language, as opposed to a sign language, and most of the evidence for the structure of very young languages that we have comes from signed languages, so that was quite a challenge. But then we were creating a fictional language, so we just dived in. We decided that the language would have a very simple syllable structure, with a single consonant followed by a single vowel. These are the first syllables that babies make, and it’s not entirely implausible that early languages would rely on them, with syllables getting more complex as languages change over time. With all these decisions about the design of the language made, we now have the wherewithal to invent words for our early Homo sapiens speakers. Here are some:
I’ve recorded these, so you can just click on the word to hear it. Notice that the prenasalized sounds cluster together with the following vowel. This is because these sounds in the system (the phonology of the language) are just one consonant, so they open a new syllable.
If you fancy it, try to make up some words in Tan!aa Kawawa Ki (literally “the mouth at the people”, or the People’s Tongue).
But a language is more than just sounds and words. Our trio of conlangers had to decide on how these words are put together. What is the grammar of the language?
Though controversial, there is some evidence that Subject-Object-Verb was possibly the order of the earliest languages, possibly reflecting an underlying cognitive basis in humans to place agents of actions early in sentences, interacting with a second bias, to set up the participants in a situation before specifying the action. So we went with that order for our language, since that’s plausibly the order that is most natural for humans, and most likely in a young language. We also decided to avoid sentences where one sentence is placed inside another, which linguists call embedding. People often avoid complex embedding in conversation, even if their language allows it, and many languages avoid grammatical devices that allow embedding almost entirely.
With these choices in place, on my long plane flight to New York, I started to translate sentences into the new language. But the outcome will have to wait until the next blog!
— This is original content written by David Adger, Professor of Linguistics at Queen Mary University of London and President of the Linguistics Association of Great Britain.
Image Source: https://www.alamy.com/stock-photo/puebloan-rock-art.html