Considering written languages that we know, restricting ourselves to alphabetical languages if helpful: Can we make any general statements or assert any constraints on how a language could have ever been spoken? (I think this would be of particular interest when considering a dead language like Latin.)

For example, can we place constraints on the number of syllables or phonemes that might have ever been used in established dialects for a particular word? Consider the French "hors d'oeuvres," in which (strangely, to this English speaker) the second word is commonly pronounced with a single syllable. I can imagine it being pronounced with up to four syllables, but no more than that. Or consider the English "one:" I've heard rural dialects in which it is spoken with two syllables ("wuh-uhn"), but I don't imagine it could ever have been spoken with more than two. In terms of phonemes: I would be surprised if the English word "one" were ever spoken with fricative phonemes. These constraints that I am imagining are intentionally very broad; I am wondering whether linguists have found any useful constraints, and if so whether they get significantly more narrow.

Or are there other constraints that can be articulated for spoken languages? For example, perhaps there are phonemes that we know have never been used in particular branches of particular languages?

  • It depends hugely on the specific language you're wondering about. We can't make generalized statements about "dead languages" as a whole, because dead languages aren't a single group. We assume, as a rule, that dead languages follow the same kinds of typological patterns and restrictions that living ones do; ex. if a language in 2021 CE can have a certain a phoneme, so could a language in 2021 BCE. Your question isn't answerable; you have to be much, much more specific about the specific dead language you'e wondering about.
    – Khove
    Commented Jun 16, 2021 at 23:44
  • 1
    You might be interested in the Wikipedia entries for "sound change" and "comparative method" and as a place to articulate a more specific question and understand how linguists approach the study of phonology across time: en.wikipedia.org/wiki/Sound_change, en.wikipedia.org/wiki/Comparative_method.
    – Khove
    Commented Jun 16, 2021 at 23:50

1 Answer 1


In conjecturing about "what is possible" in language, we make some basic assumptions. First, there isn't a substantial difference between the past vs. the future in terms of "what is possible", at least as long as you don't go 100,000 years back in the past. Second, there's no difference between written languages and un-written languages. Third, languages that die aren't fundamentally different in type from languages that aren't dead. Fifth, we avoid conjecturing about unknowable past events, e.g. we don't ask if there could have been a dialect of English with the phoneme /ʕ/, because there's no way to answer the question, until they invent the time machine.

A problem with any conjecture about what "could happen" in future English is that if you have enough time, anything is possible. "Four" used to be pronounced kʷetwóres, so in general the question is whether there is a plausible sequence of sound changes that can turn "iron" into "gold" (a parlor trick often performed in historical linguistics classes). It might take millenia; it might even require the collapse of civilization (big civilizations with centralized language standards tend to be more resistant to language change than small local languages with no writing system).

The word [wʌn] might change to be pronounced [ʍʌn] i.e. "whone", with a voiceless glide. This is actually a change that shows up in contemporary US English. But [ʍ] is acoustically very similar to the fricative [φ]. This is not the most likely sound change, but it's plausible enough that I would not actually be surprised if it happened somewhere (long after I'm dead). I would have said that "tour" (tʊr) would never be pronounced as [tɔr] in my lifetime, and yet this sound change has caught on like wildfire within my life.

When linguists talk about "constraints on language", we take that to refer to fundamental properties of human language: something that always has been true and always will be true, leaving aside science-fictiony far-future evolutionary changes. There is a ginormous debate over what such constraints would be (indeed whether "constraint" is even the right way to look at it: do we say that there is a "constraint" that humans don't breathe methane?).

  • Anyone trying to breathe methane would certainly feel constrained – for a short period of time, until they ceased to feel anything at all. Commented Jun 17, 2021 at 18:42

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.