Talk:Papuan languages

structural characteristics?
This article seems to be only about genetic classification. What about typological characteristics like Foley outlined in 1986? After all, the Austronesian page attempts this with a "structural" section. BlakeALee (talk) 02:59, 3 April 2017 (UTC)

TNG family
Kwamikagami, in your recent changes you appear to be saying that Ross argues that there is no Trans-New Guinea family and that it's actually an areal group. I can't see this in the paper you link to, and in fact he clearly says that his analysis of pronouns supports the idea of the TNG as a large family, although in a somewhat different form from that proposed by Wurm. Dougg 01:21, 22 January 2006 (UTC)
 * Don't worry. That was a very quick and sloppy edit. I'm still reading his paper, and working out all the subclassifications. I'll update the article accordingly. Maybe not tonight, though. kwami 08:22, 22 January 2006 (UTC)

Official Papuan languages?
Re the statement that...


 * The only Papuan languages with official recognition are those of East Timor.

I don't think this is correct. While the indigenous languages of East Timor do get official recognition in the constitution it is my understanding that there are presently two official languages, Tetum and Portuguese, neither of which is Papuan. Unless perhaps I misunderstand 'official language'? Dougg 14:19, 15 April 2006 (UTC)


 * We're both using the same phrase, "official recognition". AFAIK, none of them are offician languages. kwami 04:52, 16 April 2006 (UTC)

Oops, sorry, I meant to write 'special recognition'. I had thought of 'official recognition' as equalling 'official language' and looking at what the East Timor constitution says about the other indigenous languages...


 * 1. Tetum and Portuguese shall be the official languages in the Democratic Republic of East Timor.


 * 2. Tetum and the other national languages shall be valued and developed by the State.

I had thought of that as less than 'official recognition', but I guess it could be. Dougg 07:32, 16 April 2006 (UTC)
 * We could change it to 'national languages' or something if you prefer. I just don't know if that phrase implies official recognition. kwami 12:26, 16 April 2006 (UTC)

Yeah, I don't know either, and I don't particularly have a preference. I had a quick look at the PNG constitution and it makes a similar comment to the one in the East Timor constitution, to the effect that indigenous languages should be valued, etc. But I note that the East Timor government page states that their indig langs have 'official recognition', so maybe we should just leave things as they are, at least for the moment? Dougg 01:30, 17 April 2006 (UTC)


 * Does the ET constitution mention them by name? If not, we should probably delete the comment altogether. kwami 02:43, 17 April 2006 (UTC)

No, it just says as I quote above (it's here if you want to have a look, Section 13). Dougg 02:50, 17 April 2006 (UTC)

If you check the PNG Government educational guidelines, they say that everyone has a right to be educated in their own tok ples (village language). This includes Papuan as well as Austronesian languages. MarcusCole12 22:50, 1 May 2007 (UTC)

Isolates
I have removed Pyu from the listing of implied isolates, since Ross does not imply it is one. Refer discussion on the page for Left May-Kwomtari languages.MarcusCole12 08:29, 9 May 2007 (UTC)


 * Ross never addressed it at all, so it should probly go under unclassified. kwami (talk) 09:54, 9 December 2008 (UTC)

Fair enough. There is certainly no obvious resemblance to any data I've seen from the putative Kwomtari family. MarcusCole12 (talk) 07:17, 15 December 2008 (UTC)

papuan languages and geopolitic
For biological species, there is a status about how endangered they are, and specific articles may mention what endanger them. In the present article, the negative classification (neither australian, nor austronesian) hints at people that don't fit in the current geopolitic world. As it is the article is purely linguistic and does not fit in a more general context. I am certainly not competent to fill the blanks here :( 82.67.232.89 (talk) —Preceding undated comment was added at 16:14, 8 December 2008 (UTC).

Huh??!! MarcusCole12 (talk) 08:45, 9 December 2008 (UTC)

External relations
I added some sourced suggestions of external links (which I don't necessarily agree or disagree with, more research is needed) but as they have been proposed I feel they need a mention with a clear disclaimer they aren't commonly accepted. However, one person made a pov edit because he didn't agree with them and took them off, so can we leave them and only edit it to word it better rather than trying to censor it to our personal views.86.152.221.121 (talk) 16:25, 11 January 2010 (UTC)


 * There are all kinds of proposals. I'd like to be convinced that this isn't WP:Undo weight. Anything Wurm claimed is of course relevant, since he remains perhaps the principal name in classifying Papuan languages. Foley & Ross are also relevant. Greenberg is the world's most famous long ranger. But for anything else, such as connections to imaginative proposals such as "Borean", I feel we're giving undo credit to phantoms. Let's at least talk over what's appropriate before we start saying English is related to Chibchan. kwami (talk) 00:17, 12 January 2010 (UTC)


 * Murray Gell-Mann and George Starostin are notable scientists and the latter is a notable long ranger. That said, it would be better to cite academic sources like Foley's book than newspaper articles quoting him. --JWB (talk) 07:34, 12 January 2010 (UTC)


 * Added link to Foley's book and added an extra quote for George Starostin. If Borean is notable enough for its own page it is notable here and if the Kusunda link to Papuan languages is notable there it is notable here. Likewise talking about external relationships of Papuan languages as a group with no mention they may have separate external relations is pov as it's not proven they're all directly related. I think the Trans-New Guinea link with (all or just some of?) the Australian languages makes the most sense as the TNG languages spread out fast when Sahul was one island and the distribution of languages in Australia clearly shows signs of a recent spread out from the north near New Guinea to the rest of Australia so would make sense, much more so than just including the Tasmanian and not Australian languages. As for the rest I don't know anything about Kusunda but some people from within that area seem to have suggested the link and it is theoretically possible (only if the Great Andamanaese link were true IMO as that's somewhere in the middle), just as speaking Bornean languages on Madagascar is the reality. Any links would surely be very old and deep though (perhaps linked to the spread of haplogroup K to Sahul, which originated in India 30000-40000 years ago). If the proposed Borean languages can spread from the Maghreb and British Isles to Tierra del Fuego then they could easily have made the then short hop to Sahul (and if Wurm's suggestion that they are an outside language group coming in that would be a likely scenario due to the timing, but very hard to tell exactly). However, the other possibility is that the TNG languages were one regional Papuan language that spread out through expansion of farming. Not sure which is true til there's more research.86.152.221.121 (talk) 09:50, 12 January 2010 (UTC)


 * I don't buy that argument. There are many generally unaccepted hypotheses like Borean which warrant their own page, but that doesn't mean they're notable enough to appear on other, more mainstream pages. Likewise, the alleged Kusunda link to Papuan languages is not notable there. It's one thing to have articles on Borean, DC, etc. linking to each other, and another to give them the weight that proposals with actual supporting evidence receive. I could see a link to a few such articles in the See also section of Kusunda, but even then I don't see "Papuan" (which isn't even a proposed long-range language phylum) on the list. kwami (talk) 10:11, 12 January 2010 (UTC)


 * By definition an external relations section of Papuan languages is not mainstream and none of that section would be considered mainstream, even suggesting the lanaguages are all related as there's no proof of that. If you want a mainstream only page then the whole section would have to go as there are no mainstream accepted external links of Papuan languages (the Australian suggestion by Foley is the closest). We're not talking about putting this in the classification section, but a section of the article solely for non-mainstream views and by only including some suggestions and not others by major academics in the higher level then it becomes pov. Either all or just the Foley Australia mention and a link to Indo-Pacific and other similar pages. Kusunda mentions links to West Papuan languages but has no see also section. However, I have edited the section to avoid undue weight to these theories and instead provided a link to Indo-Pacific and one line for Borean, none for Kusunda (it's on Indo-Pacific page) 86.152.221.121 (talk) 10:30, 12 January 2010 (UTC)


 * Do I need to have the page protected? Edit warring to get your way is not a constructive way to edit an article.
 * Your refs include some that are not acceptable, such as an anonymous paper posted by who knows who at Dartmouth. The people who have actually worked on Papuan languages, Wurm, Foley, and Ross, should be included. Starostin, perhaps; lets talk about it. It might be better to say "see Borean for proposals for where Papuan may fit in the world's languages" or some such. Whitehorse, Bengston (I'm sure he has s.t. to say), & anonymous postings, no.
 * BTW, I don't have as much of a problem with making these claims at IP, because that's generally seen as a crock, and they therefore fit in nicely. kwami (talk) 12:27, 12 January 2010 (UTC)


 * Can you add one line in to mention Borean then, in your way? Or maybe one sentence saying something like: 'It has been suggested that the Trans-New Guinea languages have arrived more recently than the other Papuan languages (source:Wurm) but suggested links to the Borean languages are currently too scarce to establish firm conclusions (source:Starostin)'


 * Australian and Andamanese are the most likely suggestions and they are mentioned and the suggestion of TNG coming in from the west is needed as it's by Wurm but that's in a different section now, does that need a line in external relations as well or just in the top? That line suggested Papuan may not be one group so it might not be right to talk about a shared external relationship so seemed a reasonable disclaimer. Indo-Pacific as Greenberg suggested including Tasmanian and excluding Australian languages clearly isn't right I agree. More details are not well known now, until we know conclusively whether Papuan is a valid group or not it is difficult.86.152.221.121 (talk) 15:56, 12 January 2010 (UTC)


 * Nobody thinks Papuan is a language family; the scope of this article is negatively defined. "Suggested links to the Borean languages are currently too scarce to establish firm conclusions" suggests that there are "Borean languages" to link to, and so gives undue weight to that extreme-minority concept. None of the alleged members of Borean mentions it, so why should "Papuan", which isn't a member? I'd say once Borean is accepted enough to be added to the Indo-European, Afro-Asiatic, Sino-Tibetan, and Austronesian articles, then we could think of adding it here. We don't even have enough data to say with much confidence which Papuan families are related to each other; the outlines of TNG are tentative and most certainly at least to some extent wrong. I agree that Australia is the best place to look, and we say that, but so far there haven't been any positive results. (10,000 yrs might be recent enough to recover, but probly only after we have reconstructed protolanguages, and we're a long long way from that.) Anyway, we mention both Andamanese and Australian, since those have been discussed by people who actually know s.t. of Papuan languages. Andamanese looks quite promising, but we have the problem of what geneticists call "long-range attraction", where illusions in the statistics suggest distant isolates are closely related. I don't know if that's the reason that people keep trying to stick isolates like Kusunda in established families, or if that's just a psychological need to clean things up, but a major family is as likely to be related to some Papuan family or another as all the isolates that people keep proposing. kwami (talk) 20:25, 12 January 2010 (UTC)
 * Of course there is Borean languages to link to. It is not an extreme-minority concept, it's a mainstream concept in a field (long-range comparison) that some practitioners in a related field (reconstruction of lower-level families, particularly of North America) are vocally opposed to. The problem I see with the Borean quote here is more that it is a negative and vague result that Gell-Mann, Peiros, and Starostin do not further explain or attribute. Papuan languages is certainly not the article for full discussion of this paper, but on the other hand I don't see any problem with having a passing mention to point interested readers elsewhere. --JWB (talk) 23:58, 12 January 2010 (UTC)
 * Long-range comparison is mainstream, but most of the proposals are not. Borean is certainly not mainstream: it's composed of phyla which themselves are speculative and not mainstream. Most historical linguists would argue that it's simply trying to extract order from noise. And there is no "vocal opposition" to long-range comparison in North America, just vocal disappointment that Greenberg produced garbage, and entirely unnecessary garbage, when everyone was so hopeful of him bringing some sanity to the chaos of American language families. Plenty of Americanists who were disappointed with Greenberg have nonetheless suggested all sorts of long-range connections of American families, though admitting that there's not enough evidence to publish (until Dene-Yeniseian, that is). But for people who think the patterns they're seeing are real, the best explanation they can think of for the failure of others to accept them is some sort of in-built bias against even looking. kwami (talk) 00:50, 13 January 2010 (UTC)
 * Yes, it's a theory which is built on other theories, which is routine in every other field of science without howls that it's illegitimate to even discuss. Extracting information from a noisy background is also basic in natural sciences, but this is ignored by the splitters who think it's sufficient to shoot down individual data points.
 * I assumed there must be something to the conventional wisdom that Greenberg was too far out in front before reading Genetic Linguistics where he argues convincingly that global comparison is the essential first step preceding tree reconstruction, that the arguments used by Campbell, Poser, etc. would equally invalidate Indo-European and Uralic, that their non-genetic alternative explanations for the similarities are weak, that Amerind and Eurasiatic are similar in time depth and evidence to Afroasiatic, and that the difference in his reception by Africanists and Americanists is in those two communities' norms and the passage of time since his publication, rather than any difference in his method. --JWB (talk) 02:22, 13 January 2010 (UTC)


 * In Africa, Greenberg proposed two new families: Khoisan and Nilo-Saharan. Khoisan is now almost universally rejected, and NS, though still considered promising (sans a branch or two such as Songhai) is still considered tentative. His novel proposals for the internal stuctures of NC and AA have been largely abandoned. What he did for Africa was to settle the debate on classifying languages per linguistics rather than per race or economics, not much on actually classifying the languages (again, with the notable exception of NS, which would be quite a feat if it turns out to be correct).
 * In America, people were initially excited that he was going to fix things up the way he did in Africa. However, when they took a look at what he did with the families they were working on, they saw that he'd gotten most of it wrong: spurious data, dated data (sometimes by a century or more), misparsing, incorrect phonological interpretation, incorrect semantics, etc etc. Such problems invalidated half of the etyma, and when they were removed, what remained failed to support his claim. So they called up colleagues and asked if it were any better with their families, and no, it was the same everywhere. Now, there were people who would have happily reviewed Greenberg's data for him so that he could weed out such errors, or to suggest up-to-date lexicons, but he didn't bother to do the most basic fact checking. Garbage in, garbage out. And these are people who admire Greenberg as a typologist, and who regularly cover his work in their courses, and who want to know how their families fit into the larger picture.
 * In New Guinea, well, not even Ruhlen follows that.
 * Yes, a global perspective is important. We don't want to propose a Galician-Bengali language family, and binary comparisons are still far too common. Greenberg was right about that. Too bad he lost all credibility as a historical linguist, through no-one's fault but his own. Starostin I think remains much more credible. His work would seem to be more rigorous. But he was still trying to demonstrate Altaic, Nostratic is very tentative, and anything beyond that is pretty much just speculation at this point. Johanna Nichols is looking at ways of tracing migrations etc. at time depths for which the lexical evidence has been washed out, and there are some interesting results, but again rather tentative. Besides the global perspective, we need to know what it is we're comparing, which means good data on individual languages, and lower-level reconstructions such as proto-IE that get us part-way there. And even if you take the global perspective to be the first step, it can't be conclusive until corroborated, and that hasn't been done. kwami (talk) 03:13, 13 January 2010 (UTC)
 * Conclusive was not the question here - we were debating notability.
 * Regarding data, read Greenberg's response in Genetic Linguistics. He goes over several of the prominent critics' alleged lists of mistakes and finds the majority to be exaggerated or plain wrong. Others were due to progress since he made his notes, and he did not have time at the end of his life to recheck data for thousands of languages. He argues that the breadth of comparison overwhelms the problems with scanty and imperfect data, giving many examples of accepted results in Indo-European linguistics. Don't take my word for it, read the book yourself and then make your own judgement. --JWB (talk) 05:48, 13 January 2010 (UTC)


 * Oh, I know that's his take on it. And it may very well be true that his published critics have exaggerated his errors; I haven't bothered to read them. I'm considering what people I respect and who haven't published have said in private: they have no axe to grind, are not trying to get press coverage, and are careful researchers that don't make negative judgements of people just because they don't accept their claims. It's not a matter of scanty or even imperfect data, but of often spurious data. As for his claim that he didn't have time to recheck, that's nonsense: much good data that he never used was published well before he published, and in any case he could have sent sections of his charts to interested specialists in the individual families in a large number of cases. We're talking hundreds of languages here, not thousands. He could simply have removed the spurious data points they found from his comparisons: if it were true that the spurious data was not sending a false signal, then it wouldn't make any difference if it were removed. But it would appear that he's often picking the signal primarily from the spurious data; if the signal were real, the spurious data should just add noise to obscure that signal, not be half the reason for claiming it. kwami (talk) 06:15, 13 January 2010 (UTC)
 * It's not surprising that not consulting people leads to their being offended. The question is whether correcting the data in question would give a different result. May I ask what are the results contradicting Greenberg's that specialists have told you result from correcting some of his data?
 * Any attempt at large-scale organization is likely to annoy the people who work on smaller pieces and are able to point out detail errors in their own area; however, they may not be able to see the forest for the trees. This is as true in business or politics as it is in any scientific field. It doesn't mean that either the big-picture or small-picture people are disreputable; it is an inherent conflict of points of view. --JWB (talk) 06:43, 13 January 2010 (UTC)
 * When you try to read a noisy signal, you first try to remove the noise, and then try to determine what isn't noise. You don't just ignore it as irrelevant, or add noise that you don't need to.
 * They weren't offended at not being consulted, since they work on families with several other specialists and had no reason to think that they should be the one to go over it. What surprised them was that he didn't bother to run any of his data by anyone. That's just sloppy science. And it wasn't just that the details were wrong, but that correcting the data removed half or more of the cognates from the language families they knew: in one etymon he saw a connection in what was actually a suffix that should've been parsed out, in another it was a loan, in another the word was mistranslated and meant nothing like the reconstructed root, in another the language itself was misidentified from colonial records and wasn't even in the same family, in another reconstruction showed that the form was innovative and came from something originally quite different, etc. And what was left tended to be the weakest cognates, not the strongest. In the end, what looked like a reasonably convincing case to a naive reader that family X was Amerind, turned out to have very little substance, maybe even on the level that you can use to prove English is Chibchan. I can't say what they would have said if he'd only produced the much reduced case for Amerind based on valid data, whether it would've been enough, or if they were just turned off because the rest was so sloppy and they figured the whole thing was a crock. Some of these same people, BTW, were quite enthusiastic about Vajda's Dene-Yeneseian, even several years before he was able to demonstrate it to most people's satisfaction. And even though that relationship isn't all that ancient, it took Vajda years, and he doesn't see it as a precursor to DC (though, who knows, it might be). I'd love to be able to work out the relationships of all the world's languages, and think it's important that we document isolates for that very reason, but much of this long-lange stuff (with a few exceptions) seems like trying to take the easy way out. Someday, hopefully, we'll have enough groundwork laid out to do it. kwami (talk) 09:02, 13 January 2010 (UTC)
 * Of course you try to remove noise as much as practical, but you are mistaken if you think quantitative analysis depends on classifying all data points as noise or signal. Instead you make statistical arguments that hold up even if some unknown part of the data is noise. As for adding in noise, people do it all the time as a statistical test of whether it changes results.
 * They accepted Vajda's work because it was a binary comparison and reconstruction similar to their own work on lower-level families. It does not mean that they understand mass comparison enough to be able to judge it. As you say, some bad data points occurring in someone's own field can be enough of a turnoff so that they dismiss it immediately; I've seen this happen myself.
 * Indo-European was established by comparison before PIE was reconstructed, in fact before the comparative method was even formulated. Neither is it necessary to reconstruct a full family tree to establish the existence of the family; subgrouping is still murky for the best-known families like IE, Sino-Tibetan, Afroasiatic. --JWB (talk) 21:33, 13 January 2010 (UTC)
 * An old canard: linguists reject mass comparison because they don't understand it. This (and conspiracy, which I've heard plenty of times from long-rangers too) are the arguments used by advocates of every fringe idea out there. (Also for religion: you don't believe what I believe because you can't see the truth.) But mass comparison has AFAIK not resulted in a single verifiable proposal. That's the test of any scientific idea. Greenberg and others have stated this explicitly: you use mass comp to get an idea of where you are, which langs look like they may be related, then standard comparative linguistics to verify and work out their connections. Perhaps linguists don't understand it, but if it were to produce results, they'd learn!
 * As for IE, that's because it was easy. Most of the branches are transparently cognate to anyone with a passing knowledge of them. But there was debate on Celtic for some time, and it was only the use of reconstruction that demonstrated that Armenian was IE. At the time depths we're looking at, there aren't going to be any easy comparisons like IE or Austronesian. And even in New Guinea and South America, where we might expect mass comparison to uncover low-level families that could be easily verified with standard hist. ling. once they were ID'd, AFAIK Greenberg didn't find anything new. kwami (talk) 22:26, 13 January 2010 (UTC)
 * IE is easy in hindsight and because mass comparison was used. Simply looking at two IE languages at a time shows much weaker similarity and less ability to distinguish genetic relationship from borrowing.
 * Greenberg states explicitly that comparison is a precondition for reconstruction, but not that reconstruction is the proof of existence of the family. Reconstruction is something you do after it is already clear the family exists.
 * The most accepted date for settlement of the Americas is no farther back than the date for Proto-Afroasiatic.
 * Greenberg does in fact list lower-level classifications he made in South America that are generally accepted, p. 306 of Genetic Linguistics.
 * I'm winding up trying to repeat his arguments here; his own presentation is much more detailed and convincing. I recommend reading Genetic Linguistics. To say you're not even interested in reading it because your mind is already made up is taking it into the realm of religious belief. --JWB (talk) 05:12, 14 January 2010 (UTC)
 * Yes, I've read some of his stuff, and it's convincing as long as there are no counter-arguments. (When did I ever say I wasn't interested in reading it?) IE isn't just easy in hindsight; it was noticed several times over several centuries before the concept caught on in Europe in the 19th. And it wasn't "mass comparison", it was two languages already assumed to be related, Latin & Greek, in a binary comparison with Indic. I don't know that anyone believes AA is 15ky+, and in any case we have no good way of dating such things. Could you mention some of the accepted low-level classifications in South America that he found? (I took a look online just now, but those pages are missing. I have Ruhlen, but that doesn't provide anything useful.) Oh, and when guys like Ribeira work on large-scale connections in South America, they don't used Greenberg or Ruhlen's evidence, I would presume because they find that it doesn't clarify anything. kwami (talk) 05:32, 14 January 2010 (UTC)
 * Afroasiatic Urheimat has Christopher Ehret giving 15k, Militarev 10k. Clovis culture is about 13k. Can't find anything whatsoever on Ribeira and South America with either Google Scholar or web search. Not sure what you mean by "don't used Greenberg or Ruhlen's evidence" - if evidence, he should use the best he has, if you mean Greenberg's proposed subgrouping within Amerind, I'm not sure of the significance unless he is getting different results. Greenberg's critics also looked somewhat convincing as long as I hadn't seen counterarguments; I'd like to see responses to Genetic Linguistics but haven't so far. Will quote from p. 306 later when home. --JWB (talk) 19:07, 14 January 2010 (UTC)

[outdent] I don't know about Ehret's AA stuff, but his Nilo-Saharan work has been widely dismissed as essentially worthless, for many of the same reasons as Greenberg's Amerind: failure to use existing research and reconstructions of the constituent families, and also failure to even cite existing research--you'd think he'd invented the field himself. We have sites in southern Chile predating Clovis, so that's not the first migration into America. Assuming Militarev is correct, we have the colonization of America predating AA by at least 50%, and that could have involved several migrations of different language families. Ribeira at Chicago & I forget where in Brazil is the primary historical linguist today attempting to connect the large families of South America. I should have said he largely ignores G & R's claimed evidence; he works on the languages themselves, finds little lexical connection between them, but does find shared irregular morphology which contradicts G & R's conclusions, for example linking Carib with Tupian and not Panoan. In other words, I don't know of a single proposal for the Americas by Greenberg that has proven to have any merit in the subsequent half century, other than Na-Dene postdating "Amerind" (though I don't know that that idea was original to Greenberg). [I take that back: variants of Dene-Caucasian were proposed as far back as Sapir, whereas Amerind is ca. 1960. In other words, Greenberg just linked all the families that had not been linked to the old world by people like Sapir.]

But that's all really beside the point for this article. kwami (talk) 21:26, 14 January 2010 (UTC)
 * I would love to see Ribeira's work - are you perhaps misspelling his name? Or do you have some links?
 * Again, not sure what you mean by G & R's evidence - he was not a specialist or primary researcher on each language. If Ribeira simply used exactly the same evidence presumably he'd get the same conclusions, and any new work would involve developing new evidence.
 * Not disagreeing with criticism of Ehret, but it does seem like you dismiss various examples when they contradict your sweeping assertions and then claim most linguists support you, when this isn't so far evident from my reading.
 * Divergence of Amerind would be sometime after arrival in America, not on starting out from Siberia. Clovis is known to be a fast population explosion immediately after the end of the last ice age; at a minimum, it is likely to have spread a language over much of North America. Pre-Clovis settlement is still controversial; as Greenberg notes, adventurous archeology is ironically allied with conservative historical linguistics and vice versa.
 * Greenberg (1956) (p. 60 of GL) says large-scale subgrouping in S. America is the less certain part, and there and elsewhere he says the unity of Amerind is the main conclusion and subgrouping within it may remain unclear as it does in IE etc.


 * You say "you'd think he'd invented the field himself" - well, who else had done comparison across all American languages? --JWB (talk) 04:22, 15 January 2010 (UTC)


 * "you'd think he'd invented the field himself" - I meant that snark for Ehret, not for G. Most people I know respect G as a typologist, just not as a historical linguist.
 * Yes, sorry, it's Ribeiro, with an "o". The only thing I have for him offhand is demonstrating that Kariri is Macro-Ge,
 * Ribeiro, Eduardo. (2002) "O marcador de posse alienavel em Kariri: um morfema macro-je revisitado". Revista Liames, 2: 31-48.
 * And I exaggerated (ahem!) when I implied he was the only one. Here's a relevant article, a bit dated maybe, by an I believe senior colleague of his,
 * Rodrigues A. D., 2000, "'Ge-Pano-Carib' x 'Jê-Tupí-Karib': sobre relaciones lingüísticas prehistóricas en Sudamérica", in L. Miranda (ed.), Actas del I Congreso de Lenguas Indígenas de Sudamérica, Tome I, Lima, Universidad Ricardo Palma, Facultad de lenguas modernas, p. 95-104.
 * "If Ribeira simply used exactly the same evidence presumably he'd get the same conclusions". No, not if he checked his facts. If he found that the evidence was faulty, he wouldn't reach the same conclusions, just as North American specialists look at G's evidence for their families and reject his conclusions.
 * I'm not saying people are bad linguists because they disagree with me. I'm reporting the opinions of linguists I respect, and you just happened to cite another one who makes people roll their eyes. I have yet to find anyone (among linguists) who praises Ehret's NS work, but have found several who criticize it as sloppy, uninformed, bad on the evidence, self-congratulatory, etc., and I know of no-one who follows it. Okay, that's not saying a lot, but it's what I have to go on. I don't know either NS or AA, so I can't judge myself, but that is what peer review is all about.
 * Yes, G's Amerind would have diverged after arriving in America. But that only means a single massive migration if we first accept Amerind. Archeology would suggest otherwise. And there's now plenty of evidence for pre-Clovis settlement; the only question is how long before. kwami (talk) 06:06, 15 January 2010 (UTC)
 * Quotes from Ribeiro's paper NIMUENDAJÚ WAS RIGHT: "Nimuendajú (2000 [1935]) and Greenberg (1987) classified the languages as belonging to the Macro-Jê stock on the basis of even smaller word lists. In the present article we will provide sound evidence, on the basis of extensive and reliable new data from all language components, for the classification of the Jabutí languages as belonging to the Macro-Jê language stock." ... "The very fact that Greenberg (1987) adopted this classification may have been a reason that it did not become broadly accepted among Americanists." --JWB (talk) 22:36, 18 January 2010 (UTC)
 * Interesting. But as usual, when Greenberg's right, it seems to be because he used someone else's work. There's nothing wrong with that; intelligently synthesizing a diverse and sometimes inaccessible literature, picking out the wheat from the chaff, is a valuable contribution to the field, and is necessary for any credible classification. But AFAIK Greenberg's original suggestions, Amerind and its primary branches, also Indo-Pacific and its primary branches, have not been demonstrated after half a century, and only NS is retained as a promising proposal. And since NS even if right may have been accidental (it's all the African languages that were left over once the two established large families and the "click languages" were removed), he doesn't have a good track record. kwami (talk) 01:11, 19 January 2010 (UTC)
 * As you say further down this page, lack of fieldwork since then is a major reason they haven't been confirmed or falsified, especially to the high standard of requiring full reconstruction. The inadequacy of subsequent fieldwork can hardly be blamed on Greenberg. --JWB (talk) 23:15, 19 January 2010 (UTC)
 * The latest evidence seems to suggest that Nilo-Saharan is only valid as a grouping if it included Niger-Congo (and possibly excludes some of the fringe languages, but they might be part of it, just an earlier split) and Central Sudanic, Eastern Sudanic and Niger-Congo form a clade within it.
 * Also some people in Russia are working on a reduced Amerind (i.e. a TNG for the Americas) that includes most but not all of Amerind. Almosan would be a later arrival (possibly distantly related to Nivkh and the substratum of Chukotko-Kamchatkan) and some parts would be excluded as presumably predating Amerind. Clovis may represent the Almosans or may represent the reduced Amerind (more likely the former). Explicitly excluded seem to be Otomanguean, Gulf and Keresiouan (though not sure the latter is valid, in that case the component languages) but not sure which South American languages (if any) would be excluded, possibly some remote ones or in the Andes, but Hokan, Penutian (the smaller version, not Greenberg's), Macro-Mayan, Arawakan and Panoan would be some of the included ones, but also others are that I'm as yet unaware of.
 * However, as I say this research is in early stages so I can't state how reliable it will turn out to be but would disprove both Greenberg and the extremist splitters.
 * Some people are working on reducing the Papuan languages to about 5 groups but again too easrly to tell how well it will work. Which of these are related I'm unsure as either they're all related or some (most likely TNG) are later arrivals.
 * As to Greenberg's 2nd/3rd level: North Amerind (very invalid), Almosan-Keresiouan (invalid; Almosan is valid, unsure about Keresiouan), Hokan (seems valid), Penutian (partially valid, Gulf is definitely separate and Macro-Mayan probably should be, but is related), Central Amerind (invalid, should be split into Aztec-Tanoan and Otomanguean), Andean (no idea), Chibchan-Paezan (again unsure, maybe a reduced form or two separate), Equatorial-Tucanoan (invalid, Tupi should be in Je-Tupi-Carib, unsure of rest), Ge-Pano-Carib (invalid should be split between Je-Tupi-Carib and Macro-Panoan). Not that good a record (and then a lot of the correct parts Sapir got right long before, although he too made mistakes like Hokan-Siouan), but the loony splitters aren't good either. I guess Greenberg as you say as a synthesis of ideas is useful as a base for future improvement, but Ruhlen's copying him uncritically is a bad idea, another problem with him though is when he tried to claim Amerind as a family on a par with Indo-European or Eskimo-Aleut rather than as a superfamily like Nostratic. The work from Russia (Starostin etc.) is much better quality but unfortunately gets lumped in with Greenberg etc in the West.62.49.42.210 (talk) 13:01, 19 January 2010 (UTC)


 * Greenberg claimed Amerind on a par with his Eurasiatic, for example LIA p. 55. --JWB (talk) 22:34, 19 January 2010 (UTC)
 * Yeah I meant Ruhlen used to claim this, think he's changed his position of this as he seems to have reduced his 17 families (which included Indo-European on a par with Amerind or Indo-Pacific) to 12 including Eurasiatic.


 * This is a misrepresentation of "splitters". All they're doing is going on what we actually know, what can be demonstrated and independently verified. That's how science works, and linguistics doesn't get a free pass just because we're frustrated by the lack of progress. Anyway, much of the lack of progress can be pinned on lack of data (just take a look at Jabutian above, apparently demonstrably Macro-Ge once someone bothers to do a bit of field work), and the responsible response to that is to document the remaining languages before they go extinct, not to criticize people for being unwilling to accept wishful speculation. South America and New Guinea both—and many of the branches of Nilo-Saharan, for that matter—suffer from a grotesque lack of giving a shit, or from Chomsky's ridiculous (if it weren't so unfunny) delusion that all languages are really English. Thank god Brazilian ling depts have finally decided to start studying s.t. other than Portuguese. A couple decades ago Australia made a similar decision too, and just in the nick of time. kwami (talk) 22:05, 19 January 2010 (UTC)
 * If you are doing fieldwork in South America, PNG or Sudan, kudos to you. While I would like to help, most people including me are not placed to do the work. (Neither are we in Haiti helping earthquake victims, to take the current favorite thing to tell people to do instead of whatever you don't like.) Nevertheless vocal splitters like User:Billposer himself find time off from fieldwork to come to Wikipedia and insert character assassination of Greenberg into every language family article hiding behind the WP:Weaselly "most reputable linguists" while citations, if any, always seem to be from Campbell and a few others. --JWB (talk) 22:49, 19 January 2010 (UTC)
 * Yeah more research in South America and New Guinea is very much needed, I suspect the number of families would come down if we knew more. The problem is the people that insist on splitting up long established families (Hokan, Penutian etc. in the smaller non-Greenberg form) into about 10 unless there's 100% proof of them being related (which isn't applied elsewhere and is probably impossible with substratum/superstratum/borrowing effects). Clearly we don't need to follow Greenberg into groups with little evidence (he once even put everything in South America that didn't fit into Ge-Pano-Carib or Chibchan-Paezan into an Equatorial-Andean family but later realised how unrealistic it was and split it, but not enough). In some areas they put percentage likelihoods on classifications so would make sense to go down that route as if something has, say, an 80% or even 70% likelihood it makes sense to use it as a working hypothesis til we know more. Greenberg's methods of mass comparison might be useful for preliminary groupings prior to more detailed research (you have to know what's worth more careful comparisons), but clearly isn't rigid enough to prove things and the methods used in Russia are much more reliable. If standards like these were applied in other disciplines then most of Physics would be labelled pseudoscience (in fact the evidence for many things there said by top physicists has very little proof, we don't even know how gravity works).86.152.221.121 (talk) 14:19, 20 January 2010 (UTC)
 * Yeah more research in South America and New Guinea is very much needed, I suspect the number of families would come down if we knew more.
 * Most definitely.
 * The problem is the people that insist on splitting up long established families (Hokan, Penutian etc.)
 * But these aren't "long-established" families! Certain branches of Penutian have been demonstrated to be related, but at best half of the more conservative versions of the hypothesis. And Hokan is even worse.
 * In some areas they put percentage likelihoods on classifications so would make sense to go down that route as if something has, say, an 80% or even 70% likelihood it makes sense to use it as a working hypothesis til we know more.
 * That's what Campbell does. But that doesn't count, because he's a "splitter".
 * If standards like these were applied in other disciplines then most of Physics would be labelled pseudoscience
 * Not even close. Linguistics is extremely lax as a science. How else do you explain Chomsky?
 * If you apply any rigor to historical linguistics, you end up siding with the splitters. (Though not with all of them, by any means.) Of course, in science you want to move forward, so you work with the lumpers. You suggest a hypothesis, people test it, it comes up short, you retrench and try to get it to work. That's how science works. You don't suggest a hypothesis, and then when people don't buy it because you haven't provided any convincing evidence and they can't replicate your results, accuse them of being ignorant, biased, or somehow psychically incompatible with the Truth. That is pseudoscience. Penutian is slowly making its way into a respectable position, but it may never achieve what it was first supposed to. Hokan perhaps never will either. If they don't, then they don't. You can't expect a hypothesis to work out just because it would make things easier if it did. kwami (talk) 11:10, 21 January 2010 (UTC)


 * That's what Campbell does. Pointers welcome.
 * You don't suggest a hypothesis, and then when people don't buy it because you haven't provided any convincing evidence and they can't replicate your results, accuse them of being ignorant, biased, or somehow psychically incompatible with the Truth. Just who are you accusing of this? Greenberg had no problem with changes to his picture by later work, for example moving Ongota out of AA. He made explicit statements that groupings like Ge-Pano-Carib were preliminary and likely to be revised by future work. He did finally respond to Campbell and Poser, going over the sets of examples they claimed contained many errors, and noting they did not even try to replicate results or put forward alternative hypotheses, but simply took potshots at individual examples and then made a leap of faith equating these with statistical validity. --JWB (talk) 22:50, 21 January 2010 (UTC)


 * Added a few more Wurm quotes as he seems to be the most coherent model of immigration of Papuan languages attempted (I don't know enough to establish the validity of the theory but it is notable due to the proposer.86.152.221.121 (talk) 21:57, 25 January 2010 (UTC)

Papuan languages on Sumba and Flores?
Several languages of Flores, Sumba, and other islands of eastern Indonesia, most notably the Savu languages, are classified as Austronesian but have large numbers of non-Austronesian words in their basic vocabulary and non-Austronesian grammatical features. It has been suggested that these may have originally been non-Austronesian languages that have borrowed nearly all of their vocabulary from neighboring Austronesian languages, but no connection with the Papuan languages of Timor has been found.

This statement does not reflect any known linguists point of view; and cannot be supported by adequate references. It is simply and plainly wrong. The languages on Sumba definitely have no Papuan grammatical traits nor any Papuan lexicon (check e.g. Onvlee's dictionary of Kambera, and Klamer 1998, A grammar of Kambera). In fact the languages of Sumba are among the most conservative Austronesian languages in eastern Indonesia! There is NO evidence that Savunese has any Papuan features. Of SOME of the languages spoken in Flores (e.g. Lamaholot) there is grammatical evidence suggesting a prehistoric Papuan substrate.

The suggestion that all these languages are Papuan relexified with Austronesian is wrong, and there is no-one who ever claimed this, for the simple reason that it cannot be substantiated by any known evidence. — Preceding unsigned comment added by 132.229.235.16 (talk) 13:36, 15 August 2011 (UTC)

Names in Ruhlen 1987
Here are the names listed in Ruhlen, to verify that we at least have redirects for them. — kwami (talk) 08:32, 27 September 2012 (UTC)


 * Okay, all linked. — kwami (talk) 02:20, 28 September 2012 (UTC)

Oksapmin, Morwap, Molof, Usku, Tofamna,

Dem, Mor,

Kovai, Kate, Mape, Dedua, Sene, Momare, Migabac, Kube, Kosorong, Ono, Sialum, Nomu, Kinalakna, Kumokio, Selepet, Timbe, Komba, Tobo, Yaknge, Burum, Mesem, Nabak,

Abaga, Nuk, Nek, Nakama, Munkip, Numanggang, Sauk, Gusan, Finungwa, Nimi, Urii, Mamaa, Irumu, Yagawak, Bam, Wantoat, Ufim, Nahu, Rawa, Nekgini, Neko, Ngaing, Gira, Dahating, Bulgebi, Guiarak, Morafa, Forak, Degenan, Yagomi, Asat, Mebu, Nankina, Gabutamon, Domung, Bonkiman, Wandabong, Nokopo, Kewieng, Isan, Som, Sakam, Yau, Komutu, Weleki,

Wiru, Kenati,

Gants, Kalam, Kobon,

Owena, Gadsup, Auyana, Awa, Tairora, Binumarien, Waffa,

Gende, Kamano, Siane, Yabiyufa, Gahuku, Benabena, Fore, Gimi,

Medlpa, Chimbu, Chuave, Nomane, Wahgi, Nii, Narak, Maring, Ganja,

Huli, Enga, Katinja, Nete, Lembena, Ipili, Angal, Kewa, Sau,

Foe, Fiwaga, Fasu, Some, Namumi,

Somahai, Kamula, Iria, Asienara, Kamoro, Sempan, Central Asmat, Casuarina Coast Asmat, Citak Asmat, North Asmat,

Sawuy, Kotogut, Mapi, Ederah, Kia, Upper Digul, Upper Kaeme, [Mapi, Kia, Digul, Kaeme are the names of language surveys, not languages] Siagha, Pisa, Aghu, Airo, Kaeti, Wambon, Wanggom,

Southern Kati, Northern Kati, Yonggom, Ninggirum, Iwur, Telefol, Tifal, Kauwol, Faiwol, Setaman, Bimin, Mianmin, Ngalum, Awin, Pa, Nomad, Agala, Konai, Beami, Onabasulu, Kaluli, Kasua, Kware, Tomu, Bainapi, Sonia, Duna, Pogaya, Mombum, Koneraw,

Angaataha, Simbari, Baruya, Safeyoka, Kawatsa, Kamasa, Yagwoia, Ankave, Ivori, Lohiki, Menya, Hamtai,

Suki, Gogodala, Waruna,

Boazi, Zimakani, Marind, Bian Marind, Yaqay, Warkay, Kaugat, Kaygir, Tamagario,

Demta, Sentani, Nafri, Tanah Merah,

Wano, Western Dani, Grand Valley Dani, North Ngalik, South Ngalik, Nduga,

Saberi, Samarokena, Kwerba, Airoran, Sasawa,

Uhunduni, Ekagi, Wodani, Moni,

Tanah Merah, Mairasi, Northeastern Mairasi, Semimi,

Karas, Iha, Baham,

Guhu-Semane, Suena, Yekora, Zia, Binandere, Ambasi, Aeka, Orokaiva, Hunjara, Notu, Yega, Gaina, Baruga, Mawae, Dogoro, Korafe, Biangai, Weri, Kunimaipa, Tauade, Fuyuge,

Koita, Koiari, Mountain Koiari, Managalasi, Barai, Ömie, Humene, Kwale, Mulaha, Doromu, Maria, Bariji, Yareba, Doriri, Sirio, Abia, Magi, Domu, Morawa, Binahari, Bauwaki, Laua, Onjob, Maiwa, Jimajima, Daga, Mapena, Turaka, Gwedena, Ginuman, Sona,

Sinsauru, Asas, Sausi, Kesawai, Dumpu, Arawum, Kolom, Siroi, Lemio, Pulabu, Yabong, Ganglau, Dumun, Saep, Usino, Sumau, Urigina, Danaru, Usu, Erima, Duduela, Kwato, Rerau, Jilim, Yangulam, Bom, Male, Bongu, Songum,

Kare, Girawa, Munit, Bemal, Sihan, Gumalu, Isebe, Amele, Bau, Panim, Rapting, Wamas, Samosa, Murupi, Saruga, Nake, Mosimo, Garus, Yoidik, Rempi, Bagupi, Silopi, Utu, Mawan, Baimak, Matepi, Gal, Garuh, Kamba,

Mugil,

Amaimon, Wasembo, Pay, Pila, Saki, Tani, Ulingan, Bepour, Moere, Kowaki, Mawak, Hinihon, Musar, Wanambre, Koguman, Abasakur, Wanuma, Yaben, Yarawata, Bilakura, Parawen, Ukuriguma,

Dimir, Korak, Waskia, Malas, Bunabun,

Osum, Wadaginam, Sileibi, Katiati, Pondoma, Ikundun, Moresada,

Paynamar, Atemple, Angaua, Emerum, Musak, Isabi, Biyom, Tauya, Faita,

Pawaian, Dadibi, Podopa,

Kairi, Omati, Ikobi, Mena,

Ipiko, Minanibai, Tao, Karami, Mahigi,

Purari, Tate, Toaripi, Uaripi, Opao, Keuro, Orokolo,

Yelmek, Maklew,

Southern Kiwai, Wabuda, Bamu Kiwai, Morigi, Kerewo, Northeastern Kiwai, Arigibi, Tirio, Aturu, Lewada, Mutum, Bine, Gidra, Gizra, Meriam, Agöb, Idi, Waia,

Yey, Moraori, Nambu, Dorro, Upper Morehead, Lower Morehead, Tonda, Kanum,

Kosarek, Nipsan, Nalca, Korapun, Goliath, Eipo, Ketengban, Sirkai, Kinome, Anggor, Dera,

Yafi, Emumu, Dubu, Towei,

Waris, Manem, Senggi, Waina, Daonda, Simog, Amanab, Awyi, Taikat, Pagi, Kilmeri, Ninggera,

Turu, Mawes, Uria, Baburiwa, Taogwe, Taori-Kei, Tori, Tori Aikwakai, Papasena, Weretai, Taori-So, Taworta, Dabra, Foau, Berik, Bonerif, Mander, Itik, Kwesten, Maremgi, Wares, Mekwei, Kemtuk, Nimboran,

Sause, Kapori, Kaure, Narau, Kosare,

Barau, Arandai, Tarof, Kasuweri, Puragi, Kampong Baru, Inanwatan, Duriankere, Konda, Yahadian, Kimaghana, Riantana, Ndom,

Oirata, Lovaea, Fataluku, Kairui, Bunak, Kolana, Tanglapui,

Makasai, Kui, Woisika, Abui, Kelon, Kafoa, Kabola, Blagar, Tewa, Nedebang, Lamma,

Amberbaken,

Karon Pantai, Madik, Karon Dori, Brat, Kuwani, Tehit, Kalabra, Seget, Moi, Moraid, Borai, Hattam,

Ternate, Tidore, Galela, Tobelo, Loda, Ibu, Sahu, Modole, Tabaru, Pagu, West Makian,

Mantion, Meax, Meningo,

Yava, Turunggare, Baropasi, Bauzi, Bapu,

Sko, Sangke, Wutung, Vanimo, Krisa, Rawo, Puari, Warapu,

Pyu, Kwomtari, Fas, Baibai, Biaka, Rocky Peak, Iteri, Bo, Ama, Nimo, Owiniga, Amto, Musian,

Urim, One, Seta, Seti,

Urat, Yis, Yau, Olo, Elkei, Au, Yil, Alu, Ningil, Gnau, Galu, Yapunda, Valman, Nambi, Agi, Aruop, Kayik, Aiku, Bragat, Aru,

Laeko, Beli, Wiaki, Siliput, Yahang, Heyo,

Eitiep, Lou, Kombio, Yambes, Aruek, Wom, Mountain Arapesh, Southern Arapesh, Bumbita, Bungain, Mandi, Muniwara, Urimo, Kamasau, Elepi, Buna, Monumbo, Lilau,

Gapun,

Biksi,

Abau, Iwam, Amal, Wogamusin, Chenapian, Karawa, Bouye, Autu, Kalou, Pasi, Pahi, Mehek, Mayo, Namie, Ak, Awun,

Yerakai, Kwoma, Kwanga, Ngala, Manambu, Kaunga, Abelam, Boiken, Sawos, Iatmul, Kwasengen,

Sanio, Paka, Gabiano, Piame, Bikaru, Hewa, Bitara, Bahinemo, Mari, Bisis, Watakataui, Kapriman, Sumariup, Kaningara, Alamblak,

Walio, Pai, Yabio, Tuwari, Papi, Duranmin,

Murik, Kopar, Chambri, Yimas, Karawari, Angoram,

Langam, Mongol, Yaul,

Maramba, Changriwa, Mekmek, Miyak, Biwat, Bun, Waibuk, Aramo, Pinai, Wapi,

Banaro, Kambot, Aion, Adjora, Gorovu, Alfendio, Meakambut,

Rao, Anor, Aiome,

Watam, Gamei, Kaian, Bosman, Awar, Giri, Sepen, Mikarew,

Andarum, Igom, Tangu, Tanguat, Romkun, Breri, Kominimung, Igana, Akrukay, Itutang, Midsivindi,

Yele, Kazukuru, Guliguli, Dororo, Bilua, Baniata, Lavukaleve, Savosavo,

Sulka, Kol, Wasi, Anem, Panaras, Baining, Taulil, Butam,

Nasioi, Nagovisi, Buin, Siwai,

Konua, Keriaka, Rotokas, Eivo,

Aiwo, Santa Cruz, Nanggu

Warenbori, Taurap, Yuri, Busa, Nagatman, Porome, Pauwi, Massep

Language statistics by census
Where I can get Language statistics by census?--Kaiyr (talk) 08:49, 5 November 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Papuan languages. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20101011125015/http://www.elpr.bun.kyoto-u.ac.jp/essay/sakiyama.htm to http://www.elpr.bun.kyoto-u.ac.jp/essay/sakiyama.htm

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 14:59, 10 December 2017 (UTC)