I read that in several language families the pronouns and/or the verb endings of the first and second person singular are "mi" and "ti" or something similar. There is a theory that these languages are all derived from a very ancient superfamily called "Mitian".

Is this a true phenomenon? Do these correspondences really exist? If so, can they be explained by coincidence? Or was there an ancient Mitian superfamily?

Assuming the corresponences exist, is there another possible explanation, namely that "mi" and "ti" are Wanderwörter, borrowed into numerous languages like the word "coffee"? Could it be that grammatical person (first, second or third person) was a relatively late innovation and that in very ancient languages verbs simply denoted actions, without specifying who or what was performing the action? One language then developed the idea of grammatical person and the morphemes it used to denote the first and second person then spread into other languages.

  • Well, it’s certainly true that marking the first and second persons singular with *m and *t, respectively, is something that exists in a number of languages from multiple language families (Indo-European and Uralic, for example). Whether that’s a Sprachbund feature, an indicator of more distant genetic relation, coincidence or something else entirely is a lot harder to answer. May 12 at 20:05
  • The Classic Maya language has them as n and t, should it be also included into Mitian?
    – Yellow Sky
    May 12 at 21:56

2 Answers 2


Pronouns (especially first and second person singular pronouns) are only very rarely borrowed, so the idea that they are Wanderwörter seems unlikely.

The idea that personal pronouns are a sufficiently recent innovation to justify them being shared Wanderwörter also does not seem plausible. As the amount of time human language has existed is generally believed to be many times longer than even the most daring of macrocomparativists extend the comparative method, any language we can find evidence of ought to be consistent with what we know of languages today. In this case, that is that all languages have some means of marking personal deixis (these may not necessarily be independent pronouns, but in these reconstructions the distinction between independent pronouns and person-marking affixes is often glossed over) and so we should not be able to detect any evidence of a stage where this was a new innovation and such marking could be easily borrowed as a Wanderwort.

Most likely it is some combination of coincidence and extremely distant genetic relationship (distant to such an extent that almost all other evidence has been eroded away and so the relationship is beyond the reach of the comparative method).

Note also that the "Mitian" proposal includes a very similar set of languages to Greenberg's proposed Eurasiatic macrofamily (for which he adduces several other points of evidence, notably the marking of dual and plural). This proposal has now generally been superseded amongst macrocomparativists by the Nostratic proposal (which includes several other families), with some positing Eurasiatic as a branch in its own right, and others not.

  • Why the words for "tongue" also related?
    – Anixx
    May 16 at 2:16
  • @Anixx I haven't looked at the specific data you're referring to, but any similarity would likely also be due to some combination of coincidence or genetic relationship too distant to properly establish following the comparative method
    – Tristan
    May 16 at 8:17
  • Thanks for the answer. I am not a professional linguist, but an interested layperson. I suggested that mi and ti could be Wanderwörter, because it seemed to me a way to avoid positing some vast superfamily, whose existence cannot be proved. Would it be correct to say that most scholars do not regard the Nostratic hypothesis as ridiculous or absurd, but in fact as quite plausible? However there is no way of verifying it, so it must receive the so-called "Scotch verdict" - not proven. Regarding grammatical person, is it a language universal, are there any languages which do not mark person?
    – Neandertal
    May 17 at 9:07
  • there are languages that don't mark person in specific ways (e.g. they may not mark it on verbs, or may not have specific independent pronouns), but some means of distinguishing the speaker, the listener, and other parties (which may then be subdivided into more than one third/fourth/fifth persons) seems to be universal
    – Tristan
    May 17 at 9:35
  • regarding the plausibility of Nostratic I'd say consensus is the negative side of agnostic. Most linguists accept that there are (unprovable) genetic relationships between the well-accepted families, but that is quite different from accepting any specific proposal as likely true but unproven because Nostratic/Eurasiatic/Mitian are specific choices of what those relationships are. As an analogy, whilst an agnostic grants the existence of a god or gods the Scotch verdict, they may be perfectly happy to say that any individual religion is, on the balance of probabilities, untrue
    – Tristan
    May 17 at 9:42

I would say that similar forms of personal pronouns indeed indicate the relatedness of the languages. The pronounces are the words of the most conservative properties and generally change the least.

On the other hand, I do not think that using only this personal pronoun feature one can reconstruct a valid genetic node.

It seems that besides the *m- starting first person pronouns, at the same time were used other pronouns with different structure but within related laguages and maybe even in the same languages.

Thus, there were at least two global pronoun patterns.

Reconstrction:        aini-gʷi 
Meaning:              I

* Mapucho:            in-če

* Mojave:             ʔiɲe-č
* Yahi:               aini-ci

* Proto-Korean:       neugi
* Chukchi:            e-ɣә-

* Akkadian:           anaku 
* Somali:             ani-ga

* Lalofa:             ne-ji
* Lagol:              eni-gw

* Beni Sheko          ingi
* Nyalgulgule         angi

It is possible that the *min was used in some different meaning, for instance, for dual or plural number or in different context.

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.