I guess another way of asking this is: are cultures with written languages more likely to exhibit diglossia?

I'm listening to John McWhorter's course "The Story of Human" language on audible and he's going over the concept of diglossia. I can see how in modern societies how the 'standard' language of government, media and the courts would engender a situation where regional, social and ethnic local dialects flourish. The languages "on the street" would be living, dynamic phenomena while the standard would be relatively slow moving due to historical documents, media...etc.

So, do we know if diglossia was a feature of pre literate cultures? Did the speakers of PIE, for instance, have two different ways of speaking? Is diglossia endemic to all cultures?

  • 2
    My guess is your question is unanswerable, since there are no written records of societies before writing began. Maybe someone here knows any different. Feb 5, 2019 at 16:00
  • Typo. All cultures. I imagine it could be unanswerable. Maybe there’s something can can be inferred from the study of current non-literate languages? Feb 5, 2019 at 16:03
  • 1
    Maybe! I'm interested to see what you get. Feb 5, 2019 at 16:05
  • 1
    @Wilson We have written records of the situation in places like New Guinea. Personally I do mix languages in which I am not functionally literate. It seems like we agree that writing increases stability and uniformity, so a given diglossia can last longer and cover a wider area, but that doesn't prove that there is more diglossia net than in a less stable scenario with lots of local diglossia. Feb 7, 2019 at 10:57
  • 1
    @Wilson Adding to what Adam said, long before we started having linguists go out in the field and record pre-contact languages, we had neighboring literate cultures (Greek, Persian, Latin, Assyrian, Hebrew, Egyptian, Chinese, etc.) who described their societies to at least some extent.
    – abarnert
    Feb 8, 2019 at 22:24

2 Answers 2


Written languages tend to be more stable and uniform than their spoken counterparts. Besides, written languages cannot be separated from the issue of who is in power. The standard language is usually the sociolect of the upper layer of society. As regards PIE, we may expect a difference between the most learned and the lowest component of the society (probably foreign-born slaves in fact ...). Besides, it is not infrequent that males and females do not have the same language in traditional societies. This may have been the case in PIE, even if we cannot prove it.

  • 3
    I don't know whether we can ascribe the stability of a written language to the fact that it is written. David Nettle postulates that the speed of language evolution depends on the size of the speech community, and obviously larger speech communities are more likely to have a written representation of their language. See linguistics.stackexchange.com/questions/20971/… Feb 7, 2019 at 11:35
  • 3
    Is there any evidence as to whether PIE speakers owned slaves?
    – TKR
    Feb 8, 2019 at 0:12
  • @TKR That's a good question. The earliest writings describing slaves (which are some of the earliest writings we have) all seem to use metaphorical extensions of words for "enemy" (Sanskrit "dāsa"), "booty" (Greek "dôlos"; possibly the proto-Italic root of "servus"), etc. Also, while semi-nomadic people might have something like thralldom, but commercial chattel slavery seems unlikely—but capturing slaves and trading them to the more settled people to the south seems a lot less unlikely, so maybe they borrowed Semitic, etc. terms (as many IE languages later did separately did)?
    – abarnert
    Feb 8, 2019 at 22:38
  • @TKR there are several possible words for "slave". For example Latin servus can be compared with Hittite saru "plundered booty (usually cattle)". The other word is Latin sclavus "slave" which can be compared with Akkadian kalû "captive, slave" (the underlying root means to capture). Anyway, PIE speakers were not especially peaceful and prisoners of war were probably turned into laboring slaves, or prostitutes if female.
    – user23769
    Feb 9, 2019 at 9:42
  • Interesting, thanks. Servus : šaru seems to be a controversial equation (neither de Vaan nor Kleokhorst mention it), but it doesn't seem implausible. Sclavus on the other hand is medieval, and is thought to come from the ethnonym Slav via Greek.
    – TKR
    Feb 9, 2019 at 17:58

PIE had a three way division of society, farmers, warriors, priests, if not more (cp. the Indic cast system). Arguably, these might be accompanied with respective sociolects, figures of speech and surely overlap through common metaphors. PIE was likely a dialect continuum so exchange across isoglosses, later branches and then isolation of new languages would be possible.

  • You're referring to Dumezil's Tripartite ideology. Note that it's an ideology, not a direct description of society.
    – user23769
    Feb 9, 2019 at 18:45
  • @ArnaudFournet I'm affraid I don't understand the difference. The nature of the division isn't well understood of course, so it's just a tangent. I kind of missed the question, so far.
    – vectory
    Feb 10, 2019 at 10:36

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.