We have over 5000 language on Earth as of now, some extant and others not. These all came from what we now call proto languages, but do scientists believe that all proto languages came from one "mono-language" or several different language origins. Ergo are there multiple "language trees" so to speak?

  • 4
    we don't know much at all about the origin of language, including whether it was "invented" or developed some other way Sep 24, 2017 at 22:41
  • 6
    Language wasn't invented at all -- like any other species-specific trait, it evolved. How many times were human arms invented, or eyes?
    – Greg Lee
    Sep 24, 2017 at 23:17
  • @GregLee I disagree here. Communication and language are not the same thing as you are suggesting. Communication certainly evolved from grunting to elocution, but language could have been less gradual. Language in the way we know it (sophisticated sounds, vowels, consonants etc) may well have been discovered and implemented by one individual who taught others how to do it.
    – Charlie
    Sep 25, 2017 at 15:01
  • 2
    @Charlie, It's too bad that genius of old didn't keep a diary, since linguists still haven't figured out how language actually does work. There are no explicit working grammars yet. We haven't discovered them.
    – Greg Lee
    Sep 25, 2017 at 17:55
  • 2
    You may underestimate the age of language, from genetics we have some information that it is at least 300'000 years old, see also linguistics.stackexchange.com/questions/11128/… Dec 18, 2017 at 18:57

2 Answers 2


The question raises three terminological issues: what is "language", what is "invented" and what is "once"? It does presuppose that there was a prior state without language, and a later state with it (no controversy about that). To answer the question, we need to understand exactly what is being asked.

Language is one aspect of a more general set of cognitive tools unique to humans. Some parts are clearly learned, and others are clearly genetically inherited (those who deny a language-specific genetic component nevertheless do not deny that human cognitive abilities are genetically transmitted). "Language" is a myriad of evolutionary developments such as the physical enhancements which allow us to make speech articulations; the ability to retain an open-ended and replicable set of articulations (we don't just learn 12 calls, and we can easily reproduce the hundred and thousands of calls that we make – not by memorizing acoustic patterns of whole words); the ability to combine sequences of units into higher-level units; the ability to do this recursively; the ability to treat order of elements as distinctive... and many facts that are plainly about language, such as "grammatical agreement", "movement", "palatalization".

One can always conjecture that there was a single evolutionary event, via a single linguistic Eve, which catapulted mankind from chimpanzee-like linguistic abilities to having full grammars, but that is more a sci-fi movie storyline. The various genetic abilities that resulted in our current or pre-current state must have happened over the course of many generations. And it is highly improbable that humans had these latent abilities – to create phonemes, words, sentences, and unbounded dependencies – without actually ever doing this, until one day somebody got the idea "I should invent language". Forming and fleshing out the idea "I should invent language" cannot be done without language, in some form.

It's not clear what "once" would mean, in an evolutionary sense. If we take a hereditary genetic mutation to be the causal spark for the development of a new trait, "once" can be limited to a part of a human lifespan: there may be a single mutation that brings about a change, transmitted to a being's offspring. As I said, it is extremely unlikely that there was just one mutation that created the full underpinnings of language. However, you could look at "once" from a grander perspective, summing up over hundreds of thousands of years of human development. You can, effectively, answer the question in the affirmative simply by adopting a broad definition of when "once" is, as opposed to "twice".

Since language involves both genetic and learned facts, and in light of the fact that the question is about language being invented (not developed), we could narrow the question by focusing on the learned aspects, presupposing the prior existence of requisite genetic abilities. An example of a learned aspect of language is the set of sounds used to communicate something. Somehow, particular sound sequences got associated with real-world situations like "water; food; bear; tiger; rabbit; sex". The learned ability to create words (stable correlations of signal and referent) must have preceded the learned ability to create sentences out of multiple words, which must have preceded the learned ability to distinguish event by ordering words in a different fashion. I would say that it is inescapable that different learned aspects of language developed at different times (thus, "more than once").

There is a related question about "single common language", which can't be answered. We can reasonably suppose that in the way-olden days, there were many groups of humans with developed languages. It is then possible that at some point, all of those languages were obliterated (perhaps all their speakers died off), and thus every human language is the result of linguistic change from that single surviving language. Thus every existing language could be, accidentally, descended from one language.

There is a theoretical experiment (which will never get IRB approval), where infants are completely and absolutely deprived of all forms of language, with the question being whether the infants could develop full-blown languages. Poto and Cabengo-like situations differ from this imaginary experiment because in the former case, the stimulus is impoverished, but not totally lacking. Even performing the experiment would not tell you if language was invented twice, but it would be evidence that it could have been.

  • 3
    It is said that Derek Bickerton applied to the NSF for a grant to isolate some pre-lingual children on an island in the Pacific, then go back after a few years to see what sort of communication system the kids had come up with. He didn't get the grant.
    – Greg Lee
    Sep 25, 2017 at 18:01
  • Great answer. But if the experiment happens, then technically the answer is >= 2, assuming that that language survives. Sep 25, 2017 at 18:11
  • I heard about the Bickerton idea in the day, and that inspired my experimental suggestion. I always assumed the story was apocryphal.
    – user6726
    Sep 25, 2017 at 19:04
  • 2
    @user6726, Derek told me with a straight face that he did it. That's all I know.
    – Greg Lee
    Sep 25, 2017 at 23:48

Well, it happens even nowadays that languages are created anew from fragments of other languages.

There are creoles like Papiamento or Tok Pisin that are created from fragments of different and unrelated languages and have no clear antecessor. There are also constructed languages like Loglan or Lojban that did not evolve from any natural language.

I also remember (but I could not retrieve a quotable reference in short time) a story about twins raised in circumstances of language deprivation who developed their own language allegedly not related to the language of the country they lived in.

Having said all this, the main reason for having unrelated language families is time depth: Any traces of relatedness are completely wiped out by linguistic changes over time.

  • 1
    It could have been Poto and Cabengo (en.wikipedia.org/wiki/Poto_and_Cabengo#Language). Sep 25, 2017 at 13:08
  • 2
    In addition to your examples, there are many full-fledged sign languages that come into existence ex nihilo recently, as well as pre-language communication systems know as "kitchen sign". Sep 25, 2017 at 13:12
  • While Loglan and Lojban didn't evolve from other languages it still builds on existing fragments from existing language when building its vocabulary.
    – Christian
    Sep 29, 2017 at 9:38

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.