I have read somewhere that if there ever was a world's first language*, that language must have had very much in common** with the Khoisan languages. Arguments in support of this hypothesis are:

  1. Khoisan languages have click consonants, a feature that has not developed independently in other languages. Bantu languages, for example, have borrowed them from Khoisan. So, this was probably some sort of an "early invention" that was not very "fit" and got abandoned along the way.
  2. If you look at a map depicting where Khoisan languages are spoken, you will see a major area over parts of Namibia (eastern part, mostly), Botswana and South Africa and two small pockets (one in central Tanzania and one in the western coast of Namibia). These small pockets suggest that Khoisan languages once dominated most parts of Africa and then got replaced by Bantu languages (which, in fact, sorround those pockets).
  3. Genetic and archeological evidence suggests that Homo Sapiens originated in that part of the world.

Are these arguments generally accepted among linguists? How much of this is "good science" and how much is just "pop science"?

* In the sense of a single language from which all other languages of the world came out.

** The first version of this question contained the expression "direct ancestor". But, as was pointed out in the answers and comments, this is not an appropriate term.

  • 9
    With what we know now about the change history of actual languages beyond 5000 years ago (very little) and human evolution and expansion patterns >60K years ago, I'd say it's all fun speculation. The genetic data don't rule out return to and expansion in Africa from outside.
    – Mitch
    Oct 3, 2011 at 0:40
  • 12
    If there is "a single language from which all other languages of the world came out", then all languages must be direct descendants of it, no?
    – Louis Rhys
    Oct 3, 2011 at 1:55
  • 2
    I'd be more convinced of the monogensis sceptics if they'd propose an alternative, such as ancient conlangers or wolf children that found new societies with new isolate languages. Oct 3, 2011 at 15:20
  • 4
    @MatthewMartin Isn't the alternative rather that Humanity spread beyond a single local region before language developed at all?
    – Random832
    Oct 3, 2011 at 18:21
  • 2
    Ah, good point. However, for that theory to work, the Out of Africa theory en.wikipedia.org/wiki/Recent_African_origin_of_modern_humans would have to be wrong, and language would have to have been created after the appearance of modern physical culture (atl-atls . Anyhow, not enough space here to discuss the archaeological record and what can be inferred from it. The other humans that did leave in waves were the Neanderthals & Homo Erectus are all extinct (Maybe they spoke, but hard to say, Neanderthals had a physical culture that hints at the possibility.) Oct 3, 2011 at 18:52

4 Answers 4


Let's assume what's sometimes called the "proto-World" hypothesis: language arose only once, in once community of speakers, and spread from there. If that's true, then — like Alek Storm says — every human language has proto-World as a direct ancestor. So the Khoisan languages wouldn't be unique in that regard.

When linguists talk about a connection between Khoisan and the hypothetical proto-World, what they're suggesting is that the Khoisan languages must be especially conservative — that they've preserved some features of proto-World more or less unchanged. One feature that usually comes up is the one you mention: the presence of clicks.

The implicit argument here is roughly: "Well, Khoisan languages are spoken in the birthplace of the human species. So it stands to reason that they would be the most conservative languages in the world."

But there's a problem with that argument. There is no clear connection between migration and language change. Sitting still won't prevent language change, and moving around won't make your language change any faster than normal. (For instance: British English is more conservative than American English in some ways, and less conservative in others. If migration and language change were connected, we'd expect British English to be much more conservative across the board.)

So in fact there's no particular reason to expect the Khoisan languages to be the most conservative, just because its speakers have stayed in one place. For all we know the most conservative language in the world is one spoken in South America, or Australia, or ... well, wherever. We just can't tell.

  • 2
    +1 for clarifying the issue of "direct descendant". Oct 3, 2011 at 12:53
  • 4
    +1 And even if staying at home correlated to linguistic conservatism to some degree, it is likely that Khoisan at least changed in some ways faster than some other language. More importantly, there must be many other factors that influence the rate of change a great deal, such as the size of a population, its culture, etc. etc. So it is really too much of a mess to speculate much.
    – Cerberus
    Nov 9, 2011 at 3:31
  • 4
    Good call. And in fact, languages spoken in small communities often change very quickly, because there's less "inertia" -- in a village of 20, if one guy adopts a sound change, it's already spread to 5% of the population! :) So on those grounds, you'd expect the Khoisan languages not to be especially conservative, since AFAIK at least some of their speakers have always been hunter-gatherers living in small bands. Nov 9, 2011 at 18:47
  • 2
    Actually, the tendency (if anything) is for peripheral regions to be conservative while central regions are innovative, so that's certainly not an argument for Khoisan languages to be conservative vis-a-vis a proto-World language.
    – Charles
    Dec 28, 2014 at 15:22
  • Supporting Dan Velleman's comment: I read - in one of David Crystal's books, I think, but I can't find it - that having a written form tends to slow down a language's rate of change. The author cited the example of an Australian language that changed its grammar radically over just one or two generations. Jan 21, 2015 at 14:10

Regarding point #1:

Clicks have arisen in at least one other language: Damin. It was a special form of the Lardil language in aboriginal Australia used only by men who had gone through their initiation rites. Damin had four different click sounds (and one spurt and one pulmonic ingressive). No other languages anywhere in the region use clicks.

Obviously clicks are very rare, but if they were innovated for Damin, they could have been innovated for Proto-Khoisan (if such a thing ever existed - the status of the language family is in doubt).

Furthermore, if click sounds were not fit for human language use, how have they persisted this long in the Khoisan languages, and even spread into neighboring unrelated ones?

  • 1
    Damin is nearly a constructed language. We should demand the same high standard of evidence as people would demand if you decided to say something general about language change by citing how Polish and other languages "changed" into Esperanto. Re: why don't they spread? The usual explanation is that clicks are easy to learn in childhood but but ridiculously hard for adults to acquire. So in contact language scenarios, they disappear permanently. Oct 3, 2011 at 15:24
  • 2
    @MatthewMartin, you're right about Damin being basically a constructed language, but it's still on target. Otavio's question posits the idea that clicks could not have originated another time in human language. The people who invented Damin innovated clicks without borrowing them from anywhere else, proving that clicks can be invented again, not only once in "Proto-World".
    – Joe
    Oct 3, 2011 at 15:40
  • 1
    I don't think anyone said clicks can't be invented and invented multiple times, or else how were they invented the first time around? The core argument is probabilistic, that it is unlikely to to happen and unlikely to be spread and retained. I'm not sure that SO is the best place to hash out the validity of controversial theories-- they have to be worded so carefully to avoid people seizing upon extreme words like "impossible" or "could not have" Oct 3, 2011 at 15:50
  • @MatthewMartin, spot on regarding extreme words.
    – Joe
    Oct 3, 2011 at 16:31
  • The innovation isn't inventing clicks, since they're commonly used for nonlinguistic stuff like geeing a horse and tut-tutting, the innovation is incorporating them into an actual language. Oct 3, 2011 at 19:23

I don't think there's enough data to say anything:

  • pick any language family, you'll find it has some common property within that doesn't appear in -any- other family, and if it does appear elsewhere (like any true scotsman!) can be explained by borrowing (like the clicks in Bantu).

  • pockets of one language type surrounded by another, could be explained by all sorts of different historical/cultural migrations.

  • how long was it from the last excursions from Africa and the earliest known reconstructed proto-language? 60K yoa (by genetic evidence (wikipedia) to 6K yoa (PIE is the only one I know of (again wikipedia)). That's a long time, where a lot of things can happen (population movement, language change).

  • "You'll find it has some common property within that doesn't appear in any other family": that's half of the definition of a taxon such as a family (the other half being that the common property be an innovation, not a retention).
    – Colin Fine
    Dec 5, 2011 at 15:42
  • @Colin: sure, uniqueness of a property will often define a taxon (the property will distinguish from other taxons, as you say, almost by definition. I was simply addressing the OPs main question of whether Khoisan is 'original', and that originality or 'age' has little to do with the fact that it has a distinct property from other language families (is a 'taxon'). (i.e. existence of a taxon or unique feature does not imply originality).
    – Mitch
    Dec 5, 2011 at 16:45

I agree, not enough data to say anything an too much time depth. But despite not being a historical linguist, I can say re question 1 that where a feature is found in only one language family (if Khoisan is a language family!) and there's no evidence of it in others (apart from borrowing), the default assumption is usually that it was an innovation of the ancestor of that language family. So it makes sense to say that clicks were probably an innovation of proto-Khoisan and therefore pretty recent, at least compared to 'proto-World'. The spread of the Bantu languages was even more recent, so they were able to borrow them. The fact that they were borrowed and have been stable in those languages for some time suggests that they aren't that 'unfit'.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.