Austronesian is usually regarded as a separate family, not related to any other. It is never groupped into Eurasiatic or Nostratic. Yet it seems to me that it may be related to PIE. I wonder whether it is just a coincidence.

Usually the most close relative to PIE among other Eurasiatic languages is considered Chukchi-Kamchadal family.

First of all, lets look at numerals. It is accepted that Eurasiatic languages do not have common numerals because counting emerged after the proto-family split.

one PIE: sem "one united", PA: sa

two PIE: du̯oe̯, PA: dusa

three PIE: trei̯es, PA: telu

four PIE: q̆etu̯ores, PA: apat, sepat

alternations r/l and q̆/p are common in IE family and in world languages. Compare also PIE q̆eta̯ "pair", Proto-Uralic ket-ka "two", Yukaghir ikit "two", Itelmen (Chukchi-Kamchadal family) katxan "two"

It seems numerals greater than four do not coincide.


I PIE: eghom PA: i-ka-u Chukchi: e-ɣә-mi

We PA: i-ka-mi

Thou Chukchi: e-ɣә-tu PA: i-ka-su

It should be noted that Chukchi-Kamchadal family is usually considered the closest to PIE because it is the only family that has a cognate to PIE's first person singular pronoun eghom rather than more ancient mi(n)-based. It is hypothesized that eghom is a compound of something like "e̯e + ghe + mi" which meant "it is me". It seems, Proto-Austronesian had the same pattern.

  • 7
    What does Chukchi have to do with this? It's not widely accepted as related to either PIE or PA. – brass tacks May 19 '15 at 17:03
  • 6
    OK, so this guy "considers" it to be related to PIE, but not all linguists accept Eurasiatic as a valid grouping in the first place. Anyway, if Austronesian is actually related to PIE, it should be a simple matter to reconstruct a proto-language that is an ancestor of both PIE and PA. If nobody's been able to do this, that's a good sign that apparent similarities like these are coincidences. – brass tacks May 19 '15 at 17:23
  • 1
    I'm voting to close this question as off-topic. I don't think this SE site is an appropriate place for original research on speculated language family relationships, because I don't think the expertise is here so that any answers will be equally speculative, and to properly deal with the question would require answers that would be too long. When scholars tackle these relationships they do so with full books after all. – curiousdannii May 20 '15 at 7:58
  • 14
    It is just a coincidence. Chukchi-Kamchatkan exists, PIE exists, and Austronesian exists. However, there's no evidence that they are related. Any two languages will normally have half-a-dozen words that sound and mean the same as words in some other language (e.g, Yucatec /ho:l/ and English hole), just by simple probability. When dealing with reconstructions (which subsume hundreds of words), the odds go way up, because the "resemblances" get very vague, semantically and phonologically. Consequently one needs hundreds of regular correspondences and regular sound laws to establish relations. – jlawler May 20 '15 at 13:41
  • 3

Usually the most close relative to PIE among other Eurasiatic languages is considered Chukchi-Kamchadal family.

You probably know this already, but the idea of a "Eurasiatic" language family isn't widely accepted. Nor is the idea of Indo-European being related to Austronesian, or Afro-Asiatic, or really anything else.

The problem is, the comparative method's predictions get weaker and weaker the farther back in time you go. Were there languages spoken before Proto-Indo-European, with relatives and families and heritages of their own? Undoubtedly. But Proto-Indo-European is right up against the limits of the comparative method. Anything before that is lost to the mists of time.

So the question of a "closest relative" to PIE—or any relatives at all, for that matter—simply can't be answered with any confidence using current methods. Until we discover some startling new evidence, or come up with a brilliant new method that can extrapolate farther back with confidence, I'm afraid that's the best answer modern linguistics can give.

| improve this answer | |

I came up with a similar idea and made a blog-post about it here:


In short, I think that Proto-Indo-European *d corresponds regularly to Proto-Austronesian *d, that Proto-Indo-European *r corresponds regularly to Proto-Austronesian *l, and that Proto-Indo-European *s corresponds regularly to Proto-Austronesian *q.

| improve this answer | |
  • 2
    I upvote your answer but the examples following the link are not convincing. The sound correspondences are approximate, the semantic meaning seems a stretch and if we look more precisely, there are plain errors, for instance, PIE had a rule that no root start from r-. – Anixx Feb 28 at 17:23

It is thought [by whom?] that *eghom was a late-PIE variant of *egh-. Therefore the tag proto-world and its implication of a common ancestor to PIE and PA seems dubious.

The claim for a closest relative does seem to be too strong for what its worth, and does not even make sense in a wave model (at least not unambiguously), nor in a topological sense if the IE-homeland and dispersal question is still open.

Your reasoning is sound and convincing, as far as I can tell, but incomplete.

You are giving the official answer in your own question, so this answer should be an easy reputation grab: "Austronesian is usually regarded as a separate family, not related to any other."

The implied question for a closer look is to remain implicit for the time being, as arm-chair arguments (e.g. Buddhist missionaries teaching grammar, reading and math to a native population that soon flees the imposing forces) are discouraged here.

Update:, I stumbled over a statistical Swadesh list analysis treating separately both families, though I would not vouche for it: Networking phylogeny for indo-european and austronesian languages (2009) by Blanchard, Petroni, Serva, and Volchenkov.

In particular, we support the Anatolian theory of Indo-European origin and the ‘express train’model of Austronesian expansion from South-East Asia, with an essential role for the Batanes islands located between the Philippines and Taiwan.

The Anatolian Urheimat theory is less popular than the Steppes', but I would give them leeway because Urheimat is a weak definition with error margings covering at least one millenium, so I wouldn't draw systematic inferences from this result to the other, pertinant half of the results.

Update: I stumbled upon a 2009 comparison of PA and PIE. I can't access it, didn't read the preview; Dumped here for later reading:

Sound-correspondence laws of word-initial consonants between proto-Indo-European and Austronesian languages

Ohnishi, K. Artif Life Robotics (2009) 14: 567. https://doi.org/10.1007/s10015-009-0747-1

The conclusion subscribes to "Austronesian origin of Indo-European language family"; Feel free to ignore that part. The paper should be valuable instead for its application of methods from statistical biology.

Update: the question about the Similarities seen between the Rapa Nui and Indus Valley Script might be informative, though nothing certain so far. That's Easter Island vs ancient, unidentified kingdoms in south-south-east-central Asia. The comments there are, of course, destructive.

| improve this answer | |
  • I'm confused, why do you say the paper's methods are valuable when you haven't even looked at the preview? – Draconis Sep 19 '19 at 17:03
  • I say "should", hypothetical, because I cannot actually read it, just collecting the resources. – vectory Aug 22 at 11:05

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