I've been reading about linguistics and have read that most linguists are harshly critical of proposals of genetic relationships between primary language families, and that the predominant theory of origins is polygenesis. But this raises a question: if all the primary language families don't share common origins like all human populations are thought to, then how did they form?

Many of these primary language families descend from proto-languages that must have been spoken by populations that were quite closely genetically related. And we know that language, especially before the modern era, is usually transmitted from parent to child (see e.g. Father Tongue theory). It would be very unusual for an ancient person's first language to not be inherited from his parents.

Take the speakers of Proto-Uralic and Proto-Indo-European. These people must have been very closely related, probably sharing a last common ancestor after the Last Glacial Period. So how could these language families not be related? For us to believe they're not related, wouldn't we have to assume either:

  1. The common ancestral population was pre-lingual, communicating only with... grunts, etc., such that all its various descendants had to invent a new language from scratch?
  2. Or the founders of one of these language populations at some point decided to stop speaking their common inherited language, and invent a new language, for some mysterious reason?

I'm not saying I can't believe either of these conjectures, but it's just not the most intuitive thing. How could language be such a late innovation for the first scenario to happen? Conversely, why would any population decide to invent a new language? We know the fate of constructed languages from our own time. I can understand enthusiasts inventing languages, but normally when a population changes its language, it's because individuals from one culture gradually adopt the language of a neighboring (or superstrate) culture, as the language offers some major social advantage to its speakers. But that can't be the case if the language in question is a novel construction with no other speakers.

Of course that's an extreme example. The speakers of Proto-Uralic and PIE were probably particularly closely related, as far as (allegedly) primary language families go. I can believe that there was no language at the time that Eurasians became reproductively isolated from Sub-Saharan Africans, Australians, etc. But most populations are more closely related than that.

A related question, if you'll indulge me: how many truly "primary" language families can there really be, given that people pass on languages the same way they pass on genetic ancestry? I know that some language families may be related but we'll never know it. Yet there are still people who claim to believe in the polygenesis theory. So clearly there are some people who don't merely believe that we can't reconstruct a common ancestor for Sino-Tibetan and Kra-Dai languages, but also believe that not all language families are genetically related.

Thank you~

  • 1
    Not sure genetics is part of it. Wikipedia: In the field of linguistics, polygenesis is the view that human languages evolved as several lineages independent of one another. It is contrasted with monogenesis, which is the view that human languages all go back to a single common ancestor.[1] genesis and genetics are not related here. Historically, when one people conquered another, the conquered ones would take on the language of the conquerors and genetics had little to do with it.
    – Lambie
    Mar 11 at 15:34
  • 6
    I'm not sure how accurate it is to say that polygenesis is the predominant theory of the origins of language. Afaict, the predominant view seems to be agnostic - that there simply isn't enough evidence for us to ever fully determine the relationships that likely exist between attested language families, or to fully distinguish between polygenesis and monogenesis
    – Tristan
    Mar 11 at 16:04
  • 1
    Also note that even many who don’t believe in complete monogenesis do accept that some language families are quite likely to be related to each other. Proto-Indo-European and Proto-Uralic are a good example of this. I think most Indo-Europeanists and Uralicists accept that these two families have enough in common that they are most likely related, but it’s too far back for us to be able to say much really useful about it. But those same people will likely be very skeptical about claims of Aboriginal languages and the various (internally unrelated) Khoisan languages being related. Mar 11 at 20:31
  • 1
    @Lambie This use of ‘genetic’ is completely standard in comparative linguistics. It does not refer to human genomes. Mar 12 at 22:17
  • 1
    The orthodox view seems to be that, while there must be kinships among at least some of the recognized families, similarities due to common ancestry cannot now be distinguished from chance or borrowing; because of the amount of drift over time, the evidence is too noisy. Perhaps monogenesis is true (my own ignorant instinct says it is) but, without new techniques (such as time travel), it cannot be confirmed. … But I guess I'm paraphrasing things you already said. Mar 13 at 22:12


Your Answer

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