I have asked about reconstructing pre-PIE from PIE and possibly using daughter languages for help and got no response. What I remembered is a book about Proto-Afroasiatic that I stumbled upon. First, I want to know what type of method the author used or may have used to create Pre-Proto-Semitic so I can recreate Pre-PIE and Pre-PU. To clarify, what I am asking is what the established linguistic method to reconstruct pre-proto languages is.

Note: Despite the fact that this branch only has two daughter branches, I believe that it was spoken approximately 10,000 years ago. Proto-Afroasiatic, according to Wikipedia was spoken twelve to eighteen thousand years ago. This is something that I might do later work on in the future if my future research finds enough backing for it.


https://books.google.com/books/about/Reconstructing_Proto_Afroasiatic_Proto_A.html?id=cN9yUe3AgQIC&printsec=frontcover&source=kp_read_button#v=onepage&q&f=false (Reconstructing Proto-Afroasiatic by Christopher Ehret, see appendix 1)

Could Proto-Indo-Uralic be reconstructed?

  • The methodology is exactly the same as for establishing the existence of and recreating any proto language.
    – curiousdannii
    Jan 28, 2020 at 4:01
  • The method usually used to attempt reconstruction of a 'pre-proto-' language, where there is of course only the Proto-language available to provide evidence, is internal reconstruction. Jan 28, 2020 at 5:42
  • Is there a pre-proto language for PIE? Jan 28, 2020 at 11:56

2 Answers 2


The comparative method is the standard way to reconstruct a proto-language: take all the descendants, look at correspondences between them, and postulate ways those correspondences might have evolved from a single earlier form.

That's how "Pre-Proto-Semitic" can be reconstructed: Proto-Semitic is established by comparing Hebrew, Arabic, Amharic, etc, and Proto-Afro-Asiatic is established by comparing all of the Semitic languages against Egyptian, Berber, etc (which are sometimes called "Hamitic" languages in older books, but nobody uses that word any more).

For "Pre-Proto-Indo-European", the problem is, we have no other descendants to compare against! Proto-Indo-European is as far back as we can go with the comparative method, unless someone discovers a new Indo-European language that branched off earlier than the others.

So when people talk about "Pre-PIE", what they usually mean is, "the standard reconstruction of PIE has some features that are very very weird, typologically, but they become a bit less weird if we assume they evolved from an earlier system like so". It's not using the comparative method, it's speculation on how certain weirdnesses could have come from a typologically-less-weird predecessor.


Your question about PIE and Pre-PIE is quite complex. To some extent, the issue is also about what PIE is in the first place.
For example, PIE v1.0 was reconstructed without laryngeals in a very vocalic format. Afterwards, with Saussure and Hittite, it was proposed a laryngeal-based PIE v2.0. Should we consider that laryngeal PIE v2.0 is the Pre-PIE of vocalic PIE v1.0 or is it the same PIE language in two different disguises?
As a rule, there are constant interactions between the comparative method and its raw output on the one hand and the examination of said raw output as regards plausibility, internal consistency, typology, distributional holes, patterns, etc. This process often results in a rearrangement of the raw output of the comparative method into something less "naive" and immediate.
Besides, you have a number of classical issues about (Pre-)PIE:
Q1. Do the e-grade and o-grade of PIE stem from only one vowel at a deeper level?
There's some indication that e/o-grades and pitch are to a large extent correlated.
Q2. How should we explain that a number of combinations of phonemes like *tegh are possible but not attested in the reconstructable vocabulary?
Q3. The so-called thematic nominals (type dominus), with a stable columnal stem *domino-, appear to be "younger" than heteroclitic neuters like wodr, (oblique stem) uden-. The so-called thematic verbal forms are also "younger" than basic athematic verbs like *H1es- "to be".
So we can see that our reconstruction of PIE has historical layers. Not everything has the same age.
Q4. There's some indication that case-markers derive from adpositions, so that PIE probably originates in a language that had close to no case at all.
Another issue is that PIE as currently is mostly works for non-Anatolian languages, there's in fact no real reconstruction of PIE that includes Anatolian. The verbal stem of Hittite for example does not fit in the mold of orthodox "PIE". To a large extent, people make do as if the current model of PIE worked for the Anatolian branch, but this is objectively false.

  • Did people already say “v2.0” back then? Mar 14, 2022 at 2:16

Your Answer

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

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