Often I encounter arguments that Armenian is in fact not an Indo-European language. The claims assert that the regular correspondences between Armenian and PIE are too unrealistic, too rare and too irregular. The basic vocabulary is quite different, including the numerals up to ten and words for relatives. The existing correspondences may be explained by Persian borrowings or by a distant genetic relation between the languages through a pre-Indo-European super-family (that is, Euroasiatic/Nostratic, or even earlier).

So, are there solid arguments to prove that Armenian is in fact Indo-European?

Some Armenian words compared to Lithuanian, Latin and Sanskrit:

  • Lith. and Skt. sūnus (son) Arm. ordi
  • Lith. and Skt. avis and Lat. ovis (sheep) Arm. ochxari
  • Lith. dūmas and Skt. dhūmas and Lat. fumus (smoke) Arm. c'owx
  • Lith. antras and Skt. antaras (second, the other) Arm. myows
  • Lith. vilkas and Skt. vṛkas (wolf) Arm. gayl
  • Lith. ratas and Lat. rota (wheel) and Skt. rathas (carriage). Arm. aniv
  • Lith. senis and Lat. senex (an old man) and Skt. sanas (old). Arm. c'erowk
  • Lith. vyras and Lat. vir (a man) and Skt. vīras (man, hero). Arm. tghamard
  • Lith. angis (a kind of snakes) and Lat. anguis (snake) Arm. o'd'
  • Lith. linas and Lat. linum (flax, compare with English 'linen') Arm. vowsh
  • Lith. jungiu and Lat. iungo (I join) Arm. harakic linel
  • Lith. gentys and Lat. gentes (tribes) and Skt. jánas (genus, race). Arm. cegh
  • Lith. mėnesis and Lat. mensis and Skt masa (month) Arm. amis
  • Lith. dantis and Lat. dentes and Skt dantas (teeth) Arm. atam
  • Lith. and Skt. naktis and Lat. noctes (night) Arm. gisher
  • Lith. sėdime and Lat. sedemus (we sit) and Skt. siedati (sits). Arm. nstel
  • 6
    There's a complement to this that apparently some people think that Georgian is in fact Indo-European too! Commented Jan 7, 2012 at 17:37
  • 1
    It's not really obvious which sounds are meant by the romanizations of Armenian here: ch, x, c, w, ow, y, o'. Commented Jan 8, 2012 at 4:52
  • 11
    Your problem is that you compare wrong words, for example, the Armenian word "gisher" should have Latin "vesper"(and Russian "vecher") as its cognates. That's why you fail to see Indo-European reflexes in Armenian.
    – Alex B.
    Commented Jan 10, 2012 at 13:47
  • 1
    And try to use Greek data; e.g., in the above-mentioned example, Armenian "gisher" and Greek "hesperos".
    – Alex B.
    Commented Jan 10, 2012 at 14:01
  • 1
    I did not make any analysis. I just asked for arguments why Armenian is IE. And the table is just for illustration.
    – Anixx
    Commented Jan 10, 2012 at 16:06

5 Answers 5


I have never seen a serious linguist claiming that Armenian didn't belong to the IE family (it is a satem language). Where did you read that?

I strongly recommend chapter 16, Armenian, in Fortson 2010 textbook, esp. parts "The Introduction" and "From PIE to Classical Armenian", pp. 382-393. It's a good idea to read it thoroughly!

Also, have a look at https://dictionaries.brillonline.com/armenian#introTab (Tab "Aspects of Historical Grammar"); this is a chapter from an excellent etymological dictionary of the Armenian lexicon inherited from the IE).

Some of the claims you made are in fact incorrect (e.g. words for relatives; also see Basic Words from Wikipedia, posted by Daniel Briggs):

Arm. mayr ‘mother’

Arm. hayr ‘father’

Arm. khoyr ‘sister’

Arm. eġbayr 'brother'

Arm. dowstr 'daughter'

Arm. taygr 'brother-in-law' (cf. Sanskrit devar-)

Arm. nu 'daughter-in-law' (cf. Greek nuos, Old Church Slavonic snuxa)

some other common examples are:

Arm. kov ‘cow’

Arm. tun ‘house’

Arm. em ‘I am’

Also, it's a good idea to compare Armenian with Greek:

Arm. get 'water' - Greek hydor

Arm. ayr 'man'- Greek aner

Re: numerals, your assumption is also wrong; here's a quote from Matasovic 2009 (Winter 1992 is of the same opinion):

"Here are the numerals from 1 to 10: mi, erku, erekc, čcorkc, hing, vecc, ewtcn, utc, inn, tasn. Although this is not obvious at first sight, their forms are inherited from PIE (*smi-yo-, *dwoh1, *treyes, *kwetwores, *penkwe, *(k's)wek's, *septm, *h3ek'toh1, *newn, *dek'm)."

  • water - IE *vodor, Eurasiatic *wetV, Borean *WVTV. No evidence that it could only come directly from PIE (for example, Nenets, Mansi *wit.)
    – Anixx
    Commented Jan 21, 2012 at 0:39
  • 7
    The reconstructed forms you mentioned (Eurasiastic, Borean) are highly questionable since they make certain highly questionable assumptions. And what use is the form with two unspecified (i.e. any) vowels, like in *WVTV? Also, a comment on methodology. If you dispute the IE origin of that particular word, the burden is on you, not on me. So far, I have seen nothing.
    – Alex B.
    Commented Jan 21, 2012 at 20:14
  • Yes. This is just another spelling for the same phoneme. Like in classical Latin they used v for both.
    – Anixx
    Commented Jan 21, 2012 at 23:11
  • 1
    Exactly, letter "v" was pronounced as [w] in Latin. Since we reconstruct phonetic forms for PIE and there are no written documents, that is why "v" is not used for PIE - the sound was an approximant (a glide), not a fricative.
    – Alex B.
    Commented Jan 22, 2012 at 18:06

When it comes to distant relatives, simply comparing words or grammar is of no use. What should be done instead is establishing regular sound correspondences between cognates.

Here's a table that shows something like that (though it doesn't show the underlying material)

  • 1
    This table does not include the most notorious or Armenian correspondencies such as dv->erk
    – Anixx
    Commented Jan 11, 2012 at 12:12
  • 3 PIE *dw‑ > Arm. ‑rk‑ or ‑k‑: Pedersen 1906: 176‑177, 178; AčaṙLiak 6, 1971: 402‑403; Grammont 1918: 251-252; Pisani 1934: 185; Dumézil 1938b: 51-52; Belardi 1950: 148; Schmitt 1972/74: 10‑11; J̌ahukyan 1982: 75; Ivanov 1983: 27‑29 (*dw‑ > *rkw‑> erk‑); Szemerényi 1985: 788‑795; Vennemann 1986: 33‑34, 41‑42; Kortlandt 2003: 2‑3, 7, 28, and especially 88‑95 (= 1989); Ravnæs 1991: 162-166; de Lamberterie 1992: 257; Bolognesi 1994: 34‑35; Harkness 1996; Olsen 1999: 270‑271; Beekes 2003: 199‑200, 209; Viredaz 2003.
    – Alex B.
    Commented Jan 11, 2012 at 16:45

I requested entries for all of the words listed by the OP on the English Wiktionary.

My Armenian friend in Yerevan who contributes Armenian and Old Armenian to Wiktionary and has an impressive collection of Armenian etymological dictionaries has now made (or added etymology to) the entries on every word plus provided these notes which I've copied verbatim thanks to both Vahagn and Wiktionary:

  • 1
    Thanks. Still the question remains as the number of cognates is not that big and the correspondencies are either too far or represented by only a few examples (there are some correspondencies which were attested on no more than 4 words).
    – Anixx
    Commented Jan 9, 2012 at 11:50
  • 1
    Basic words Commented Jan 11, 2012 at 14:02
  • 7
    Well, we don't have to convince Anixx that Armenian is an IE language. Professional linguists solved that problem a long time ago. There is so much written on Armenian that it will take a couple of years just to read it: mostly in German, some in Russian and English, and, of course, in Armenian.
    – Alex B.
    Commented Jan 11, 2012 at 16:41
  • 2
    @Anixx: If those 500 words form most of its core vocabulary, and if its grammar also follows, for the most part, Indo-European patterns, then the conclusion can be only one, namely that Armenian is indeed an Indo-European language.
    – Lucian
    Commented Dec 15, 2016 at 20:28
  • 3
    @Anixx English is a Germanic language. Only about a quarter of English vocabulary is Germanic, but they are the core words, as are the PIE words in Armenian. If English is any guide, a likely hypothesis is that Armenian had a lot of influence from non-IE languages, but that doesn't change the fact that Armenian is Indo-European at its core. People don't borrow words for close family relations, staple crops and the like; they borrow words for new technology and exotic items.
    – phoog
    Commented May 1, 2020 at 23:55

I consider Armenian to be a "saturated" language, because of Urartian (non-IE), Luwian and Iranian influence. Proto-Armenian was probably a bit different from how Armenian is today, but in its basis it is obviously Indo-European. It just needs a little bit more attention and things clear up. I usually make comparisons with Greek, Hittite and other Anatolian or Balkanic languages. See some examples below:

The order is English : Armenian - other languages

great: mec - Greek mega

bad: çhar - Greek kako

bite: xacanê - Greek dakno

bone: oskr - Greek oston, Hittite hastai

breathe: phçhê - Greek pnei

cloud: amp - Hittite alpas

dry: çhor - Greek xeros

fire: howr - Hittite pahhur

give: tay - Hittite pāi

hair: her

I: es - Greek/Latin ego

in: i - Proto-Germanic *in, Greek en and so on...

lake: liç - Hittite lulis, Greek limne

name: anown - Greek/Phrygian onoma

bring: berê - Phrygian ber-, Greek pher-

other: ayl - Greek allos, Lydian alev

salt: al - Greek alas

star: astl - Hittite hasterz, Greek aster

warm: jerm - Greek thermo

what: zi - Greek ti

wife: kin - Greek gyne, Lydian kana, Thracian *goni

bear: arj - Greek arktos, Hittite hartaggas

cow: kov

goat: ayc - Greek aiga

ship: naw - Greek nāus

Now you might be interested also on this speech by Charles de Lamberterie (Sorbonne). He speaks about the place of the Armenian language in the Indo-European language family.

The Armenian Language & its Place in the Indo-European Linguistic Family


Well, PIE is unattested, and the Urheimat was never found (and never will be). Indo-European is only a proposal, just like Austronesian for example. Where there can be hardly any doubt that each branch (Latin, Greek, Albanian, Armenian, Germanic, Celtic, Slavic, Iranian and Indic) constitutes a valid family, the link between each other is best explained by Spraschbund. This leaves Greek and Albanian as isolates, but closer to any said IE languages than outliers of the "family". Armenian is a riddle, as it appears to be the mix of two languages in equal proportions, This mixing seems to have first happen 3200 years ago ; of course it took time, but today we can consider it a language of its own. It is clearly an isolate and shows quite as much affinities with said IE languages than Caucasian languages. Sadly, Caucasian languages have been very poorly studied, that's why it is still (and will be) considered an IE languages. Nevertheless, if we exclude dead languages like Tokharian, Luwian and Hittite, it is by far the most divergent of the "family".

  • 6
    Do you have a source for "the link between each other is best explained by Sprachbund"? There's an enormous amount of vocabulary shared between branches.
    – Draconis
    Commented Jan 27, 2021 at 22:07
  • @Draconis, you yourself suggested to consider a dialect continuum many times. The inquisitive tone is absurd. "In short, late items of technology, plants or lifestock could have circulated across a linguistic continuum after expansion had already begun. If linguists find such a conclusion absurd, then they must reconsider their own chronologies." (Mallory 1995, The Indo-European Homeland Problem: a Matter of Time) Granted, this advocates for expansion from a single mother tongue, but the early chronology is still uncertain 26 years later.
    – vectory
    Commented Jan 29, 2021 at 21:48
  • @vectory There's a big difference between a dialect continuum and a Sprachbund. A Sprachbund is when languages in an area share grammatical features due to contact, but remain distinct; English, German, Spanish, etc all mark perfective aspect with an equivalent to "have", for example, but those languages aren't mutually intelligible. A dialect continuum is when you can't draw clear boundaries between different languages/dialects because the variation is continuous.
    – Draconis
    Commented Jan 29, 2021 at 22:10
  • Yes, you have for example 'the european linguistic area (SAE)". Some great linguists went as far as arguing PIE to be one the greatest lie of all. In fact, what you have to keep in mind is that those branches of said-IE do not share CORE VOCABULARY. Linguistically talking, it means those branches are families on their own.
    – Kewari
    Commented Jan 30, 2021 at 3:24
  • @Draconis, so what's the big difference? It's a false dichotomy needed to prune tree models. Romance as one of the biggest examples of a dialect continuum evolved not only from Latin as a layer over relative Italic languages, but under the influence of Greek, Etruscan and Punic (IMHO) to some degree, then it took over Celtic tongues which in the end appear so similar that it's not clear whether there was a common parent node or areal influence. Finally it formed another Sprachbund in Medieval Latin including languages as distant as Norman where it was soon not anymore the dominant language.
    – vectory
    Commented Jan 31, 2021 at 13:48

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.