Talk:Nostratic languages/Archive 1

General
"The list of etymologies of lexical words reconstructed by Dolgopolsky" is alphabetic from a thru k. Obviously, over half the list is missing. — Preceding unsigned comment added by 75.187.206.38 (talk) 13:15, 19 November 2011 (UTC) Yikes! Learn Tamil in 30 Days as a piece of evidence? Identifying Colin Renfrew and other professional linguists as teaching a 'liberal theory' is a little wacko, too, but that's a matter of content. Oh, well, I'm not participating in this entry, other than to point out to whoever just inserted all the pro-Nostratic parts that if you start a new line with a space on the left margin the formatting will be all messed up. Another pointer - titles in English are usually not printed in all capitals, but in italics, which wikipedia supports. Your lists of words would be better in indented lists. You might go look at How does one edit a page to see how things work. --MichaelTinkler

Well, I'm just going to put my money where my mouth was and see if this entry fixes itself again without my participation. A quick hint: the new additions just go to show that if you look hard enough and spread your definitions wide enough (i.e. thinking "barley" and "gravel" have similar enough meanings to count as cognates), you can find cognates for short syllables almost anywhere. Even Sir William Jones didn't rely on single words, and he did his thing more than 200 years ago when this stuff was pretty primitive. This is what I meant when I included the anecdote about English being pigeon-holed into a Central American Indian language family by using this technique, and why modern linguists have moved on to grammatical structures as a more reliable way to figure out what's related to what. -- Paul Drye


 * Paul, that's a great anecdote about the "demonstration" that English was a member of a Central American language family. But could you please give us more information. Ideal would be a citation to some journal in which this spoof demonstration appeared. Failing that, could you at least mention the name of the person(s) who did this. Thanks Interlingua 14:20, 3 March 2006 (UTC)


 * Yep, I thought barley/gravel was a beautiful example. Of course where I come from gravel is rather coarser than barley, but then I'm sure that's just cultural relativism.  The entry seems to have self-healed, except that in the reversion someone cut the talk.  I'm going to put it back.  --MichaelTinkler

Across some 15,000 years, the possibility that "barley" and "gravel" are related is not so big a jump. Don't forget that in English, we use the same word for a single piece of barley and a single piece of sand - viz. grain! One root in Proto-Northwest Caucasian has already differentiated into the meanings acid and sharp in Abkhaz and Ubykh respectively (and I'm not talking about sharp of taste, either, but sharp of knives, needles, et cetera). The point is that two words don't have to be perfect semantic or phonetic cognates in the present, but semantic and phonetic evolution from a common ancestor, with not too much semantic deviation, must exist. Don't forget that English wheel and Sanskrit cakra are cognates. But look at English and Georgian mama - the English means "mother", but the Georgian means "father". Also, Manx beg means "little"! Paul is correct - phonetic cognates aren't much good, but cognates on more than one level - eg. grammatical AND semantic, or semantic and phonetic - are quite useful. - thefamouseccles


 * Remember that the word gay once meant happy but now homosexual. "Nice" comes from Latin nescius, ignorant. For example "night", "to sleep", "dark", "death" and "black" are more closely related meanings: People sleep during night. Sleeping can be considered an euphemism for death. During night there is dark. Black is a dark colour. --85.156.128.119 18:53, 22 October 2006 (UTC)

Some obvious trangressions are merely individual(regional) perspectives. It takes about 200-400 years for night to change in time of action, if you get what i mean. I often see that development twist through the ages. That is why i define 'relations' as "associative". Any dinuminative aspect of language is associative. (people use the words they think they remember).77.248.56.242 11:39, 20 June 2007 (UTC)

Arabic/Hebrew and Japanese 'You' in Arabic/Hebrew is 'Anta'. In Japanese it is 'Anata'. Jondel | Talk


 * The above is a good example of the coincidental look-alikes you can find between any two languages. I have no idea what the ancestral Japonic 2SG pronoun was (if it had one: Japanese, like many East and Southeast Asian languages, doesn't have personal pronouns in the Indo-European sense), but it wasn't anata, which is a distal demonstrative (in a set with medial sonata and proximal konata) and historically a compound. --kwami 06:00, 18 Apr 2005 (UTC)

I've studied several Uralic, Indo-European, and Semitic languages, as well as Sumerian - and I could build an "Ur-language" out of any two of these well enough to convince laymen. Historical comparative theories are all very nice, but as they leave too much room for imagination, they're not very reliable - and definitely not the latest word in general linguistics. While it could be argued that the closeness of Uralic, Semitic and Indo-European languages is quite well-grounded, the others proposed members of Nostratic family are much more dubious, and even the exact nature of the relationship between these "core" languages is most likely much more complicated than a "language tree" - consider the diffusional effects, at least. Even in USSR, Nostratic theory was not the official version, but a (possibly attractive) alternative, advocated by Illich-Svitych etc. Still, while I suppose it often hovers on the edge of the "academical" science, the basic ideas behind Nostratism are considerable, at least, and I doubt we've seen the last of it yet. So, don't kick it too hard. --Oop 22:56, Oct 7, 2004 (UTC)

Proposal to add Austronesian languages to list
A number of books (some quite convincing, notably Bopp) have been written proposing a very close link between the Indo-European languages and the Austronesian languages. To wit, the similarities of such fundamental PIE/PAN words as *duaus/*duau, *egos/*aku(s) and others. Bopp wrote a 600+ page book on the subject, quite in depth. While Bopp's exuberance might have undermined his credibility, I think a lot of his evidence makes his proposal at least worth mention. I have, myself, found a striking number of apparent cognates (and I'm not talking about loanwords) to indicate a strong possibility of a relationship between Japanese and the Indo European languages, as well as between Japanese and the Austronesian languages. I can't testify as to what affect the apparent Austronesian similarities might be due to contact between the two groups/languages, since there are, to this day, three Austronesian languages still spoken in Formosa/Taiwan, which even from ancient times, was in contact with Japan.

My observations are not unique. While the primary groups included in Nostratic are invariably (despite what the article seems to indicate to the contrary) are Indo-European, Finno-Ugric, Kartvelian, and Afro-Asiatic, a great many linguists include the Manchu-Tungus (Turkic) family (with its loose cognates Korean and Japanese) and the Dravidian languages...some even including the Bantu languages of Africa (although I personally view this as an attempt to include w/in Nostratic the greatest number of languages as conceivably possible). The problem with the Austronesian languages is that they seem to be anomalous in every way, when it comes to categorization, since the language family seems to have at least three core areas...Thor Heyerdahl's work notwithstanding. TShilo12 09:05, 14 Feb 2005 (UTC)

Move
Moved from "Nostratic language" to "Nostratic languages" for consistency with practice for other language family articles, e.g. Indo-European languages --Tabor 19:02, 15 Jun 2005 (UTC)

Australia
'On the other hand, the comparative method has been successfully applied to Australian Aboriginal languages. Even though Australia has been inhabited for about 50 thousand years, and no significant technological changes occurred, aborigines living on seven eighths of Australia use languages belonging to relatively recent Pama-Nyungan language family (estimated to be about five thousand years old).'
 * I think this is wrong on one and possibly both counts. Australia did have an archeologically documented technological revolution around 6000 years ago. Also, I have seen Australian languages cited as an area where linguists divide languages geographically or featurally in acknowledgement that genetic relations are not always clear.--JWB 06:02, 24 October 2005 (UTC)
 * I don't think "Australian" is seriously held as a family anymore, but Pama-Nyungan certainly is. The only linguist I'm aware of that disputes its validity is Dixon. He makes some interesting points, but all other Australianists I'm aware of seem comfortable with it as a genetic node. There are speculations, though, that it may have been a cultural revolution with the moeity system, and the rituals that went with them, that spread Pama-Nyungan - or maybe that was caught up in the technological revolution. On the other hand, take a look at PNG, which was settled when Australia was (as the continent of Sahul). The subbranches of the East Papuan "family" have no demonstrable cognates! kwami 07:35, 24 October 2005 (UTC)
 * I wasn't taking a position on genetic unity of Pama-Nyungan, but just disputing the assertion that no significant technological changes occurred, and the suggestion that the comparative (reconstructive) method had been successfully applied to Australia as a whole.--JWB 16:21, 24 October 2005 (UTC)

Borean
New name of the language is Borean. The dictionary of Borean roots: --Nixer 12:34, 11 November 2005 (UTC)
 * That looks broader than Nostratic - it appears to cover all of Eurasia and Americas.--JWB 14:33, 11 November 2005 (UTC)

???

 * "Armenian was not added [to the Indo-European family] until the 1880s (until then, it had been thought to be an aberrant dialect of Iranian)"

Even if Armenian was a member of the Indo-Iranian languages, wouldn't that still make it Indo-European? Or were they not recognized as Indo-European by then? And what exactly is "Iranian" anyways? I've only ever heard the language of Iran called "Persian" or "Farsi".

One more pedantic note: it is not accurate to speak of a language being "added" to a language family. The language always was part of the family; we only recently recognized it as such. — Ливай 16:00, 24 January 2006 (UTC)


 * If you think of IE as being a hypothesis, then it would be correct to speak of adding languages to it. That's common phrasing, anyway. Iranian is the IE family that includes Persian, Kurdish, Ossete, Balochi, etc. The passage should probably say that Armenian was not recognized as a separate IE family until the 1880s. kwami 20:29, 24 January 2006 (UTC)
 * The phrase "aberrant dialect of Iranian" implies that Iranian is a single language with many dialects, rather than a group of languages each with their own dialects. In any case, I don't think it serves as a good analogy in this paragraph, where the question at hand is the very inclusion of language groups into a single family, not the particular classification of languages within the family. — Ливай [[Image:Anarchy_symbol_neat.png|15px]] 00:57, 25 January 2006 (UTC)

"Defense" of Nostratic theory
This section looks to me like it's a pro-Nostraticist apologia. The author of this takes the previous section on criticisms of Nostratic, and shoots it down point by point. This, and the slim amount of references provided, suggests to me that it's original research. Also, some of the arguments are specious at best - for example, "Had linguists followed this advice, they would not have succeeded in establishing the existence of Indo-European.". Indo-European is one of few families with long written history, so it's a lot easier to compare languages diachronically, and splits between languages of the family are often directly attested in the written record; this partially obviates the need to establish regular sound correspondences, because they're often established already. So, until such time as the author provides references, I'm removing this part to the talk page for discussion. Thefamouseccles 02:10, 25 May 2006 (UTC)


 * Defense of the Nostratic theory
 * Note: this section includes numbered replies to criticisms made above.
 * 1. "Almost all modern linguists express considerable skepticism of the data put forward to demonstrate interrelationships between the various language families under the Nostratic umbrella. The main criticism of Nostratic holds that the methodology used leads people to see patterns that actually result from coincidences. In reconstructing Nostratic, supporters do not use the techniques that linguists have established to prevent false positives, such as insisting on examining only regular sound shifts."
 * The phrase "almost all modern linguists" mainly applies to American linguistics. In Russia, Nostratic theory has more adherents.
 * 2. The demand that investigators admit "only regular sound shifts" ignores the fact that linguists typically identify sound laws long after the discipline has established relationships between languages through lexical and morphological correspondences. Had linguists followed this advice, they would not have succeeded in establishing the existence of Indo-European.
 * 3. "The more recently introduced technique of comparing grammatical structures (as opposed to words) has suggested to some that the Nostratic candidates lack interrelatedness."
 * However, recent work by Joseph Greenberg (and Allan R. Bomhard, forthcoming) has done a lot to dispel doubts in this area.
 * One cannot claim "comparing grammatical structures, as opposed to words" as a "recently introduced technique". Compare Sir William Jones's famous remarks of 1786 (bold emphasis added):
 * "The Sanscrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek; more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident."
 * Nor can one speak as if proponents of higher-order classification of languages relied on lexical as opposed to morphological data. For instance, much of Bomhard and Kerns (1994) attempts to reconstruct Nostratic morphology and syntax. Likewise, function-words and morphological elements form the subject matter of the first and longest volume of Joseph Greenberg's Indo-European and Its Closest Relatives (2000).
 * So, whatever valid arguments critics may launch against Nostratic, this is not one of them.
 * 4. "The proposed Nostratic language "super-family" hypothesis suggests links between many Eurasian language families. The precise nature of the links remains the subject of debate, however, as proponents have not agreed on the set of families to include."
 * In some respects, however, the situation somewhat resembles that which occurred within Indo-European studies in the early stages of research. At first, researchers did not definitively identify the Celtic languages as part of the Indo-European language family, while they did not admit Armenian until the 1880s (until then, expert opinion regarded it as an aberrant dialect of Iranian). Lycian and Lydian did not gain definitive recognition as Indo-European languages until the middle of the twentieth century. Even today, uncertainties linger about the subgrouping of the Finno-Ugric languages, not to mention Afro-Asiatic.
 * 5. "Many mainstream linguists have dismissed claims (by Aharon Dolgopolsky, among others) that the words reconstructed for Proto-Nostratic point to a pre-agricultural society in the Middle East (as one might expect for a language pre-dating Proto-Indo-European). Critics see the notion as wishful thinking exacerbated by that very expectation shaping the results."
 * However, one cannot lightly dismiss the possibility that Natufian and Zarzian Proto-Nostratic speakers helped spread the cultures involved in the post-glacial "broad spectrum revolution", using new bow-and-arrow hunting technologies and domesticating the dog.
 * 6. "Some linguists also object to the assumption that languages must ultimately all stem from one reconstructable root. Linguists know that unrelated languages in close geographical proximity can trade vocabulary, syntax, and other features, and some suggest that the present-day "family" structure of languages may simply exemplify an aberration."
 * Advancing technology might allow one language to rapidly expand in geographic scope, as the people speaking it conquered their neighbours. This would then allow that one language to evolve into a family (in fact, some have argued that Indo-European languages have spread as far as they have because of the war-making advantages that the domestication of the horse gave to one small group of Proto-Indo-European speakers).

Prospects for the Nostratic theory
I've put an original research tag on this section, since the content is unsourced and seems argumentative. Some of the points verge on the nonsensical, in any case. Citations would be welcome, and preferably a little more balance. On the other hand, it could be deleted without significantly damaging the article.--Chris 06:23, 24 June 2006 (UTC)


 * I'd say its faults are that it is very general and stating the obvious. It hardly even makes assertions that would require sources for backup. It does seem to try for balance and mention the various viewpoints though. --JWB 11:26, 24 June 2006 (UTC)
 * The whole article is rife with unsupported assertions; given the lack of footnotes, that's hardly surprising. The tone is also rather argumentative, that is, it seems to consciencely making a case against N., rather than simply presenting evidence. That's largely a matter of style. [See following comment]


 * Let me quote the whole section with comments.


 * As the foregoing arguments demonstrate, strongly-held opinions separate proponents and opponents of Nostratic studies. It seems unlikely the two sides will agree any time soon or even agree to disagree.


 * "Demonstrate" = argumentative = orginal research. "It seems unlikely..." = unascribed opinion.


 * Against the Nostratic theory, note that consensus opinion among professional linguists strongly opposes it, at least in the Anglo-Saxon world. In favor of Nostratics, note that scientific innovations usually begin with a few individuals and that the academic establishment resists until the case for them proves overwhelming (this argument can generically "validate" all pseudo-science).


 * "Note" = argumentative turn of phrase. "Anglo-Saxon world": dubious logic - are non-Anglo-Saxon linguists ipso facto incompetent? "In favour of Nostratics..." - without a citation this sentence amounts to original research. "Pseudo-science" - POV, and irrelevant to Nostratic, which, however wrong it may be, is not pseudo-science, given its espousal by serious linguists.


 * Again, one can point out the lack of intellectual credentials of some of the persons advocating the Nostratic hypothesis. Contrariwise, one could point out that some of the most distinguished linguists of the twentieth century (such as Holger Pedersen and Joseph Greenberg) favored Nostratic or similar theories.


 * "Again" = argumentative turn of phrase. "Lack of intellectual credentials..." - nonsensical argument: any theory may be supported by non-credentialed amateurs - the critical point is whether some credentialed professionals support it, which they do in the case of Nostratic, as the following sentence concedes. Amounts to a smear, ergo POV.


 * Only one comfort remains for those perplexed by the ongoing furor: that science has a way of eventually correcting itself. Time will tell whether the Nostratic theory resembles the grain of sand that produces the pearl, or just resembles a grain of sand.


 * Highly rhetorical, unencyclopediac. As a quotation from a Reliable Source, this might OK, but not from a Wikipedia author.


 * I admit that the shortcomings I've pointed out are not restricted to this section, but could serve as a basis for tightening up the whole article.None of the foregoing means, by the way, that I think Nostratic is a winner - I have no qualifications to judge that - just that if the facts are as clear as people say, they should be able to speak for themselves.--Chris 16:11, 24 June 2006 (UTC)


 * I think the statements are describing arguments of proponents and opponents of Nostratic (and reasonably accurate as such), not being stated as unqualified fact or as the author's opinion. Describing the various points of view and their arguments is completely appropriate to WP:NPOV, although the tone and phrasing could use improvement, and more specific and sourced statements would of course be better.. --JWB 18:00, 26 June 2006 (UTC)


 * I think we're close to agreement here. What I'm thinking of doing is getting hold of a couple general linguistics texts - the Cambridge Encyclopedia of Language comes to mind, but I'm sure there are introductory textbooks and so on that are suitable - and summarize their take on the controversy. Not that I will do this today. I might also want to change Prospects to Status, since discussing the prospects of a theory seems inherently speculative.--Chris 18:07, 26 June 2006 (UTC)


 * The truth about Nostratic is that hardly any linguists in the world accept it as valid. Now, whether it actually is valid or not it a different matter from what the overwhelming academic opinion of its veracity is. As Wikipedia is not meant to be proposing what is "right," but instead presenting the viewpoints on what it "right"--giving more emphasis on views held by the majority of specialists--the proper stance to take in this article, in my opinion, is one that makes it much clearer than the present version that hardly any linguists accept Nostratic or the methods used by its proponents, but give the pro-Nostratic side their fair say (it's obvious where my bias lies--but I will do my best to let them have their fair say). Anyway. What I'm getting at is, I don't at all see how this section is even marginally encyclopedic. All of your points, Chris, are good. This is pro-Nostratic arguments, with weird poetry-like rhetorical stuff mixed in, but it doesn't actually present any evidence at all in favor of the hypothesis, since it's trying to pretend it's neutral. Pro-Nostratic arguments should have substance to them--there are a number of valid defenses of the hypothesis. I can't see the point of this section. It adds nothing to the article, at least at this point. A rewritten version, which reasonably and accurately and fairly presents both sides of the debate, would of course be fantastic. I...hope I did't just rant. Crud. Well, whatever. Take care, both of you, --Red Newt 09:18, 28 June 2006 (UTC)

Follow-up - Reading the article over again, I find it's pretty good on the whole. The "Prospects" section is the only one I would criticise at all harshly.--Chris 18:22, 24 June 2006 (UTC)

Tik/pal challenge
Who added all those languages? Was it a serious linguist, or a Wikipedian? It doesn't seem too serious to me. 32.106.193.203 18:33, 2 July 2006 (UTC)


 * Tricky question, given the general esteem with which expertise is regarded here. The two etymologies here come from On the Origin of Languages by Merritt Ruhlen. Chapter 14, "Global Etymologies," by John D. Bengtson and Ruhlen. They're good examples of the technique of mass lexical comparison (a rough triangulation of phonological and semantic similarities without regular phonological correspondence rules, validated by massing examples against chance rather than by predictable form). Now, whether that amounts to the work of a "serious linguist" is precisely the point on which linguists themselves seem to be divided. 66.142.229.12

Thanks, but what's 'given the general esteem with which expertise is regarded here' supposed to mean? I hold expertise in high regard. It's just that I've just come up with about 45 examples for why *tik is two and *pal is one. I'll show them to my professors, and then get back to y'all. 32.106.193.203 21:21, 2 July 2006 (UTC)


 * For Wikipedians, the gold standard of expertise is peer-reviewed publications in recognised journals or from recognized academic presses. Pretty simple, really. Several pro-Nostratic publications meet this standard, so that makes the views expressed in them Wikipediable. Naturally, you can cite critical opinions from other experts. What you can't do is cite your own independent research.


 * I know that :). But perhaps I can get it published, or have them publish it.32.106.193.203 10:25, 3 July 2006 (UTC)


 * Now, on another topic, what's this supposed to mean?


 * "...but well-studied families show forms hardly recognizable as related to *tik (such as French /dwa/, doigt) (Trask 1996:394-5). "


 * As far as I know, French doigt 'finger' comes from Latin digitus, which, on the surface at least, seems a very plausible cognate of *tik. So what's the objection? If anti-Nostraticists are as querelouos as Trask seems to be here, maybe Nostratics deserves a break.--Chris 22:44, 2 July 2006 (UTC)


 * Right, that's his point, actually. Since we have written records of French's history, we know that /dwa/ was once digitus, which looks a heck of a lot more like *tik than /dwa/. If French and Latin had been unwritten until the last century or two (as many or most of the languages used in the *tik reconstruction were), we would have no way of knowing that /dwa/ "should" be regarded as a reflex of *tik. Salmons' (Trask is just summarizing his arguments) point is that in unwritten languages, we end up with all these forms virtually identical to *tik, or at least clearly recognizable as "cognates", but in well-studied languages with long written records or well-studied families with good reconstructions, we end up with all these forms that look very little like *tik, but we "know" they're cognates because the ancestral forms did look like *tik. This is unlikely to be some sort of weird coincidence, where all languages that have long written records suddenly underwent changes that obscured the form of *tik just after they were written down (or just after they split up into various daughter languages), while all unwritten languages never underwent such changes. What IS likely is that for families with known histories, proponents of Mass Comparison go back as far as can be reconstructed and THEN look for cognates--and the fact that it is in the oldest recoverable forms of well-studied languages that "cognates" look most like *tik, while for poorly-studied languages it's the present-day forms, suggests that what the method does is find words that LOOK similar to *tik in WHATEVER language it is applied to, rather than finding COGNATES of *tik. Which if true is a very serious problem, because though the method is designed to find words that look alike, this is supposed to be because they ARE cognates; if they're not, the mass comparison method falls apart. --Red Newt 04:20, 3 July 2006 (UTC)


 * Ah, that makes sense. But surely the article needs some expansion around that point.--Chris 05:48, 3 July 2006 (UTC)


 * The mass comparison argument is that even if in most cases cognates have changed enough to be unrecognizable, there will still be a few cases where they haven't, and that when you compare a large number of words and languages, you may be able to see statistically significant connections even if the good cases are only a few percent, especially if the nonrelated cases can be assumed be random.
 * Also, some mass comparison studies have used as little as the first consonant, or even just the point of articulation of the first consonant, as the data for comparison. For some of these loose criteria, /tik/ and /dwa/ would be considered a match. Relaxed criteria also increase the number of random matches, but again, if the noise signal is filtered out, it may still be possible to recover useful data. --JWB 20:13, 3 July 2006 (UTC)

POV
I've tried to revise what seemed to me some recently accumulated POV stuff. "Vast majority"? We'd need a headcount for that to be meaningful. The only actual reversion was to restore Ed Finnegan's take -- he's a linguist of stature, his formulation is simple and direct, a relevant remark by a competent authority. Adamdavis
 * Whoops, sorry about that. And looking it over, it was too POV. A number of serious linguists do agree with the Nostratic hypothesis, and many try to use the comparative method (although they tend to be somewhat laxer in its application than most). Thanks for the changes! Take care, --Red Newt 04:27, 3 July 2006 (UTC)
 * I've noticed that the late Larry Trask - a fierce critic of Ruhlen's World hypothesis and some of Greenberg's hypothetical families - is actually fairly respectful of Nostratic. Perhaps the article should mention some of these more positive evalutions.
 * I'm going to strike the reference to World, since it seems like guilt by association.--Chris 18:43, 4 July 2006 (UTC)
 * That's true. He mentions in his book on historical linguistics that although the Nostratic hypothesis certainly needs a lot more investigation, linguists should be open-minded about it. The whole *tik/*pal section, now that you mention it, doesn't belong here at all; it's a demonstration of mass comparison and Proto-World claims, so it could be referenced there, but Nostraticists generally try to use the comparative method, although it's often claimed they use somewhat laxer criteria than mainstream linguists. I'm going to remove it, if no one objects. --Red Newt 22:12, 4 July 2006 (UTC)

--
 * Nice to see that people are starting to recognize the difference between Nostratic and Proto-World. No they aren't the same thing despite a persistent popular view that they are on the same level. --Glengordon01 11:18, 8 July 2006 (UTC)

--

Phonology, IPA
I just added the reconstructed sound system, taken from the Kaiser & Shevoroshkin reference, and transcribed the poem into IPA. I hope I haven't misunderstood Illich-Svitych's transcription – it is similar but not identical to that Kaiser & Shevoroshkin use. *wince*

I will later insert sound correspondences between Proto-Nostratic, Proto-Indo-European, Proto-Kartvelian, Proto-Afro-Asiatic, Proto-Uralic, Proto-Altaic, Proto-Turkic, Proto-Mongolic, Proto-Tungusic, and Proto-Dravidian; they are explained in the same reference. Maybe I'll add a couple of etymologies; the paper contains a respectable amount.

David Marjanović | david.marjanovic_at_gmx.at | 1:41 CEST | 2006/9/20

Added sound correspondences between Proto-Nostratic plosives and their derivatives according to Kaiser & Shevoroshkin (1988).

David Marjanović | david.marjanovic_at_gmx.at | 0:40 CEST | 2006/10/7


 * Can you add the etymologies of the words in the poem? --85.156.128.119 11:45, 10 October 2006 (UTC)


 * At least for some of them. Thank you for the idea! I won't be able to do it soon, however; I'm ill, and an exam is approaching (I'm a university student... not of linguistics).
 * David Marjanović | david.marjanovic_at_gmx.at | 15:37 CEST | 2006/10/11

Finished the sound correspondences as far as possible. I'll try to plug a few holes by scouring through the etymologies, but help would be greatly appreciated! I also wonder what the reconstruction of Proto-Uralic Kaiser & Shevoroshkin used looked like... it contains at least and, while the one at Proto-Uralic language has only one "x"!

David Marjanović | david.marjanovic_at_gmx.at | 00:48 CEST | 2006/10/14
 * Update: I wondered if that reconstruction includes Yukaghir data. It clearly does not: Yukaghir has a, but also a /q/.
 * David Marjanović | david.marjanovic_at_gmx.at | 00:55 CEST | 2006/10/23

Added the vowels – for the most part from the etymologies, because Kaiser & Shevoroshkin only give a table that compares Proto-Nostratic, Uralic, Altaic, and Dravidian. Also fixed the supernumerary ejective in the poem. Etymologies next!

Most of the Criticisms section will have to be deleted (for being bullshit – follow that link!) or thoroughly revised.

David Marjanović | david.marjanovic_at_gmx.at | 00:51 CEST | 2006/10/18

Removed Proto-Turkic, -Mongolic and -Tungusic because Blažek has now published a complete correspondence table of phonemes within Altaic, including Korean and Japanese – see link at the bottom of the Altaic languages page. I suppose I'll have to copy that table to the Altaic languages page…

David Marjanović | david.marjanovic_at_gmx.at | 00:55 CEST | 2006/10/23


 * Spot on about the Criticisms section. There are 'criticisms' in that list which no linguists would ever make, such as that Nostratics - 'mainstream' Nostratics anyway - doesn't employ only regular sound correspondences! (Non-criticisms like this could only come from people who are under the misapprehension that Nostratic is one of Greenberg's mass-comparison speculations.) There used to be a section of replies to the criticisms but it got removed because a contributor thought it looked pro-Nostraticist (but it's not a matter of being pro- or anti-Nostratic; largely, it's a matter of not criticising what people aren't actually doing or claiming). Perhaps parts of this should be put back and integrated with the Criticisms section to give a balanced perspective on each of the valid points? I'll leave it to you 'cos I'm concentrating on the early sections and Urheimat when I get the chance. There's a fair bit still to do there.
 * Andrew Huckerby a.k.a. 172.141.153.188 19:30, 23 October 2006 (UTC)
 * Much of the replies will probably end up in the rewritten criticisms section… but anyway I'll complete the table first. The (new) link to the Starling database has a correspondence table of affricates, so I'll add that. If you want to see really ludicrous "criticisms", go to Altaic languages – I mean, did anyone ever really claim that the Altaic languages were related because they lack grammatical gender!?! Either it's a weird 19th-century phenomenon or the most embarrassing strawman ever.
 * David Marjanović | david.marjanovic_at_gmx.at | 21:45 CEST | 2006/10/23

Just added the affricates and voiceless fricatives from. For the time being, I've ignored the voiced fricatives and the clusters with /w/ because the "Nostratic" on that page may just be Eurasiatic. Someone who actually understands that matter should check that.

Etymologies next! :-)

David Marjanović | david.marjanovic_at_gmx.at | 19:15 CEST | 2006/10/24

Can you translate the poem to proto-Uralic and Indo-European? --Muhaha 17:19, 30 October 2006 (UTC)
 * Who, me? That would be interesting to do, but I can't do it.
 * BTW, the grammar of the poem is probably wrong (the word order, that is), and some etymologies can be found in the papers listed in the external links.
 * David Marjanović | david.marjanovic_at_gmx.at | 22:34 CET | 2006/10/31

Problems with the recent edits
It is good that there has been some extensive work on this article recently. Even so, I also perceive some problems in the way it has been edited, and I'd like to make some suggestions.

First, the extensive table of proposed sound correspondences does not serve a purpose in this article. It breaks the text and is only distracting. The article should present basic information about the Nostratic theory in a clear style; a good encyclopedia article should be a coherent and easy-to-follow text. Now the text almost gets buried under the numbered lists, huge tables with footnotes, etc. Therefore, I think that the table along with its explanations should moved to a separate article titled Nostratic sound correspondences or something similar, and only a brief description plus a link to it kept in the main article.

Second, the section on critcism should be converted into normal text instead of a numbered list. Moreover, some of the criticisms seem inexact or peripheral, while central issues concerning methodology and data should be explained in more detail. It should also be made more clear in the article that the majority of comparative linguists have not accepted the Nostratic hypothesis. As for references to the criticism section, I could try dig some up. --AAikio 14:46, 2 November 2006 (UTC)


 * I'm not going to register soon. If you want to make a separate article, maybe Proto-Nostratic language (as exists for Proto-Indo-European language and Proto-Semitic language), please go ahead! I agree that the long, unweildy table doesn't make for easy reading.
 * The criticism section needs to be entirely reworked. Many of the points in there are flat-out wrong, such as the claim that Nostraticists don't insist on regular sound correspondences.
 * References are, of course, always good.
 * Your friendly neighborhood armchair linguist :o)
 * David Marjanović | david.marjanovic_at_gmx.at | 16:09 CET | 2006/11/2

A couple of (comparatively) minor points to David Marjanovic: (1) There are a couple of problems with your addition on Elamite. First, it reads as if G. Starostin is the first to include Elamite in Nostratic and I'm certain that's not the case. Greenberg effectively included Elamite in Nostratic - although of course he didn't use the name - and he wasn't the first either. (I suspect it might have bveen Dolgopolsky; I'd say more, except most of my files and notes are at my other address.) Secondly, the AA/Elamite/Nostratic tripartition anticipates a separation between AA and 'Micro-Nostratic' that has barely been touched on at this stage in the article: it is not properly discussed until the following two paragraphs and seems out of place here. Moreover, the connection most frequently made (Blazek & Starostin dissenting) is still between Elamite & Dravidian, so if any subgrouping hypothesis deserves a brief mention in connection with Elamite it is that one. Your short insertion on G. Starostin's view is really not so much about Elamite as about the overall coherence of Nostratic & its internal subgrouping and as such, belongs naturally with the AA/Eurasiatic discussion. For all these reasons, I've taken the liberty of moving it and inserting a new sentence on Elamite; I hope youi don't mind. 217.23.229.250 14:37, 4 November 2006 (UTC)
 * I don't mind; you put it exactly in the right place! :-) I just stumbled over that pdf, found that it seems to do a good job of arguing against Elamo-Dravidian (a hypothesis which I had liked for geographical reasons), and tried to put into the article what it says about Nostratic, just like how I gathered the pronoun table, without caring about the overall layout or readability of the whole thing (which is still not breathtaking).
 * David Marjanović | david.marjanovic_at_gmx.at | 17:50 CET | 2006/11/4

Oh, and (2): the external links you've put in the text result in the printed version getting bloated out with long unwieldy internet addresses, several of which just duplicate the ones in References & External Links. Most of it may be moved into the proposed new article but the same problem will recur there. Personally I'd just state the author & date in the text where possible and leave the other details to References and Links.
 * Right. I'll see what I can do. (Even though I can hardly imagine people using the printed version. What for?)
 * David Marjanović | david.marjanovic_at_gmx.at | 17:52 CET | 2006/11/4


 * Not everybody has a laptop; more importantly, some of us still like stuff that we can read behind a newspaper in the pub.
 * 217.23.229.250 14:31, 7 November 2006 (UTC)

Having said all that - and despite the problem with the Correspondences as pointed out by Ante Aikio - I've a lot of respect for what you've been trying to do with this article. I wish we saw the same enthusiasm from some of the previous contributors who left it in a mess. 217.23.229.250 14:44, 4 November 2006 (UTC)
 * Even more so because at least some of the previous contributors were actual linguists who really had access to literature! Most of my information comes from that paper from 1988! I look at the Dravidian roots in the Starling database and am baffled because they contain /d/ and /g/!
 * Update: That's because newer reconstructions of Proto-Dravidian, mentioned and used by Starostin (1998), assume a voice constrast instead of a length contrast for plosives. This seems to make good sense, so I just changed the correspondence table. However, they also find a contrast between dental and alveolar plosives, and I don't know how to relate this to anything in Kaiser & Shevoroshkin. Are there any Dravidianists out there?
 * David Marjanović | david.marjanovic_at_gmx.at | 13:33 CET | 2006/11/11
 * The only reason I'm working on this article is that I find myself able to steal enough time, apparently unlike just about anyone else. Do you happen to know if Mother Tongue is available online anywhere (except the TOCs)?
 * David Marjanović | david.marjanovic_at_gmx.at | 17:58 CET | 2006/11/4


 * Don't I wish! All I know is that, apart from that issue 31, it seems to be unavailable anywhere that I've looked. There are MT articles on Borean and 'Afro-Eurasian' (i.e. a postulated grouping of Nostratic with Dene-Caucasian) that I've been stalking for years, although it's only fair to point out that my researches have also been stalled for years through personal circumstances, and I've not been online for long, so I wouldn't be surprised if I've missed something somewhere. Wish I could be of help. Maybe someone in the know, a friendly neighbourhood professional linguist, say, can point us amateurs in the right direction?
 * Andrew Huckerby 217.23.229.250 14:31, 7 November 2006 (UTC)

Added date & provenance of Elamite article. It is actually from Mother Tongue; my source supplies the date (2002) and volume number (VII) but not the edition, and I can't be certain that the page numbering belongs to the original MT article. Since the Bengtson & Norquest references are from 1998, MT 31 should be an edition from an earlier volume, so MT '7' AND '31' in the refs. are not on the same level. The numbering will have to be completed by somebody familiar with the journal. 217.23.229.250 16:09, 9 November 2006 (UTC)
 * Well, there's the journal, and there's the newsletter (later renamed "The Long Ranger"). The "volume 31" that is online belongs to the newsletter, not to the journal.
 * David Marjanović | david.marjanovic_at_gmx.at | 13:36 CET | 2006/11/11

Cognates
Compiled the pronoun table from the cited sources. There's much to explain about it, but it's too late for "today".

David Marjanović | david.marjanovic_at_gmx.at | 00:29 CET | 2006/11/3


 * Added most footnotes.
 * David Marjanović | david.marjanovic_at_gmx.at | 00:25 CET | 2006/11/4
 * Finished the footnotes and removed some mistakes from the table. Also added Etruscan.
 * David Marjanović | david.marjanovic_at_gmx.at | 17:34 CET | 2006/11/5

What part of the Nostratic hypothesis is considered the least controversial?--Muhaha 20:48, 6 November 2006 (UTC)
 * Probably none. What do you mean exactly? Strictly speaking the hypothesis only consists of the idea that the language families in question are related.
 * You might say that it's controversial whether Sumerian is Nostratic or Dené-Caucasian – but the long-range people don't quarrel much among themselves (sometimes the same people work on both, like the late S. Starostin), and there seems to be very, very little research on the affinities of Sumerian.
 * David Marjanović | david.marjanovic_at_gmx.at | 21:34 CET | 2006/11/9

The "Criticisms" section
I have removed the following paragraphs: That's simply wrong. There are reconstructions of regular sound correspondences and a reconstruction of the Proto-Nostratic language, as I hope the table I inserted makes clear. Nostratic is not a result of mass lexical comparison (scare words: Amerind, Greenberg, Ruhlen). Whoever inserted that paragraph must have confused ignorance with knowledge. I don't quite understand this, but it's probably a critique of mass lexical comparison and therefore does not apply. To me, what the Nostraticists do looks like the good old comparative method. Could someone clarify that? Yeah, so what? Was Indo-European "wrong" before everyone agreed to include the Anatolian languages? I consider it relevant that, to the contrary, everyone seems to agree on not including Basque, (North) Caucasian, Burushaski, Yeniseian, and Sino-Tibetan in Nostratic (see Bengtson 1998), showing that, whatever the merits of the Nostratic hypothesis, it does not consist of merely sweeping up everything into a geographical grouping. While I don't have an opinion about that, it simply doesn't matter. It is not a criticism of the Nostratic hypothesis, it's a criticism of Dolgopolsky's interpretation of his own particular reconstruction of Proto-Nostratic. Where and when that language was spoken has no bearing on the question whether it ever existed in the first place. Nobody claims Indo-European is fictitious just because the Kurgan folks and the Anatolian folks can't agree. This, too, is irrelevant. Monogenesis of all languages of the world is not a necessary assumption for the Nostratic hypothesis. Why should it be? Please. (That said, I don't understand why monogenesis isn't the default assumption and the burden of evidence on any dissenters. Or is it?) Clearly this is backwards. If we can reconstruct a language that was spoken before whatever arbitrary date happens to be put forth, then the hypothesis of that particular limit is disproven, not the other way around. – BTW, why should "rapid technological change" have anything to do with language evolution? Shouldn't it be assumed that borrowings are more likely to occur when there's more trade, and that there's more trade when there is more stuff and technology to trade? Isn't the most common case of lexical borrowing when a word is imported along with the thing it describes?
 * 1) Many modern linguists express considerable skepticism of the data put forward to demonstrate interrelationships between the various language families under the Nostratic umbrella. The main criticism of Nostratic holds that the methodology used leads people to see patterns that actually result from coincidences. In reconstructing Nostratic, supporters do not use the techniques that linguists have established to prevent false positives, such as insisting on examining only regular sound shifts.
 * 1) Most of the proposed phono-semantic sets appear much more speculative than those used to group languages into the accepted families — one technique used to support a similar super-family famously "demonstrated" in the 1960s that English belonged to a proposed Central-American language family.
 * 1) The proposed Nostratic language "super-family" hypothesis suggests links between many Eurasian language families. The precise nature of the links remains the subject of debate, however, as proponents have not agreed on the set of families to include.
 * 1) Many mainstream linguists have dismissed claims (by Aharon Dolgopolsky, among others) that the words reconstructed for Proto-Nostratic point to a pre-agricultural society in the Middle East (as one might expect for a language pre-dating Proto-Indo-European).
 * 1) Some linguists also object to the assumption that languages must ultimately all stem from one reconstructable root. Linguists know that unrelated languages in close geographical proximity can trade vocabulary, syntax, and other features, and some suggest that the present-day "family" structure of languages may simply exemplify an aberration.
 * 1) Certain linguists suggest that in the absence of rapid technological change (which did not occur prior to about the 8th millennium BC) the tendency for languages to trade features with each other would drown out the tendency of languages to evolve. In such circumstances, the axiom that languages change in a manner that can be reversed does not hold before a certain point in the past, and one thus cannot reconstruct older proto-languages (Nostratic or otherwise) using the techniques used to reconstruct the proto-languages of the accepted major language families (all of which, linguists believe, post-date the invention of agriculture).

I have kept the following paragraphs in: If true – and I'm definitely not in a position to judge that – it's a perfectly valid criticism. Does someone happen to have a pdf of Campbell (1998)? After all, most of the data I've used to expand the article is ten years older (Kaiser & Shevoroshkin 1988) and therefore runs a high risk of containing all those errors! I only added the Fact template. I don't quite understand the criticism, though. Several grammatical suffixes have been reconstructed for Proto-Nostratic (I hope to put a table into the article later today). More importantly, how can you ever show that two languages are not related? Can't one only show a negative here – that there's no evidence to consider them related?
 * 1) Certain critiques have pointed out that the data from individual, established language families that is cited in Nostratic comparisons often involves a high degree of errors; Campbell (1998) demonstrates this for Uralic data.
 * 1) The technique of comparing grammatical structures (as opposed to words) has suggested to some that the Nostratic candidates lack interrelatedness.

The criticism section is now very short, maybe shorter than it deserves. (After all, Nostratic is not as plain obvious as Indo-European or Algonquian or Athabaskan.) Therefore I encourage everyone to expand it – but please not by putting plain falsehoods and obvious logical fallacies in again.

David Marjanović | david.marjanovic_at_gmx.at | 12:10 CET | 2006/11/13

Everything else
I've tried to clean up the introduction, but I can't get much beyond that. The Background section still needs a cleanup. For example, does PIE really have so many words for flat, open landscapes? Gamq'relidze & Ivanov, at least, claim it has a lot of words for mountains, mountain lakes, and the like – even one for "snow leopard"!

David Marjanović | david.marjanovic_at_gmx.at | 13:05 CET | 2006/11/13

Added the morphology.

David Marjanović | david.marjanovic_at_gmx.at | 20:36 CET | 2006/11/17

Added more on grammar.

David Marjanović | david.marjanovic_at_gmx.at | 21:18 CET | 2006/11/17

Added more etymologies. Should I add yet more? Are there any mistakes in them (e. g. Proto-Uralic *sēxE with a long vowel in front of *x looks suspicious, right?)?

David Marjanović | david.marjanovic_at_gmx.at | 22:18 CET | 2006/11/17


 * To be frank, the Uralic data in this article is pretty much of the same kind that one commonly in Nostratic references. First, hardly any Uralist supports the reconstruction of long vowels in Proto-Uralic, most have explicitly rejected such a solution. E.g. the widely supported reconstruction by Janhunen and Sammallahti derives long vowels (*-VVC-) from Proto-Uralic *-VxC-. Hence e.g. *kuuli- 'to hear', if really a Proto-Uralic item, must derive from earlier *kuxli-. As for the word meaning 'to eat', this would be correctly *sexi- or *sewi- with a short vowel. As for the other words, Proto-Uralic *purV- 'storm' does not exist, but *purki does (the meanings of the reflexes oscillate between 'snow storm', 'smoke' and the like). The reconstructions *n'o/amV- 'grasp' and *(e)Nte- 'face' are supported by no acceptable data. Moreover, I can't see why the vowel in the last form is in brackets, as if a possible alternative was *Nte- (even though Uralic had originally no initial clusters).--AAikio 07:36, 18 November 2006 (UTC)


 * Very good – then emend or delete those examples, please. I don't have time currently, unfortunately.
 * David Marjanović | 22:43 CET | 2006/12/15
 * Update: I just did it: deleted *ńo/amo and *eNte*, and changed *purV to *purki and mentioned that the presence of the -/k/- is unexplained (it isn't an obvious suffix, right?). Kaiser & Shevoroshkin imply that the ū of *kūli- is not derived from a disappeared *x but from the *-/iw/- they reconstruct for the Proto-Nostratic root: */q'iwlV/- < **/kuwlV/- < *kūli-. That seems plausible enough. Except that I can't fathom why they assume *-/i/- in that word in the first place. The Proto-Chukchi-Kamchadal reconstruction fits it, but they didn't use that. The Proto-Dravidian reconstruction clearly hints at a front vowel, but why precisely */i/?
 * David Marjanović | 00:53 CET | 2006/12/29 —The preceding unsigned comment was added by 131.130.1.143 (talk) 23:53, 28 December 2006 (UTC).
 * If *iw-> "*u:" is regular and *u: always comes from *ux, can't we then say that proto-Nostratic *iw corresponds with proto-Uralic *ux? 85.156.246.82 17:24, 30 December 2006 (UTC)
 * Is there any more grammar correspondences than the ones currently listed? E.g. accusative: Uralic *-m, IE *-m, Altaic *-be.85.156.246.82 21:04, 30 December 2006 (UTC)
 * Ablative: Uralic *-ta, IE *-d (in pronouns, like *h₁med, *tued) 85.156.132.203 18:54, 1 January 2007 (UTC)
 * I have no idea if */iw/ > */u:/ is regular.
 * Uralists currently think that every *u: comes from *ux, and if only Uralic is taken into account, this is clearly the most parsimonious hypothesis. Taking all of Nostratic into account, however, Kaiser & Shevoroshkin imply that there's at least one exception…
 * I bet there's a lot more grammar correspondence, and your two examples look plausible. Are they published anywhere? Remember – WP:NOR.
 * David Marjanović | david.marjanovic_at_gmx.at | ignore my IP address, it's dynamic!


 * If the meaning of the words must be exactly same to be a valid reconstruction (i.e. grain/seed and grain/small piece of stone are too different meanings, even though a small piece of stone can be mistaken as a seed when not carefully looking), how "snowstorm, smoke and the like" is accepted?--Muhaha 12:08, 24 November 2006 (UTC)


 * I don't quite understand your point here - have I claimed that "the meaning of the words must be exactly same to be a valid reconstruction"? I never said such a thing. As for the cognate set of Uralic *purki, all the intermediate meanings are well attested in verbal forms with meanings such as 'to smoke, to pour, puff (of smoke)', 'to whirl in the air (of smoke, dust, leaves, snow, etc.)', etc. Moreover, there are semantic parallels which show both of these meanings in one and the same language, e.g. North Saami soica 'thick smoke; thick, drifitng snow'. What I tried to point out above is that the several of the Uralic forms are cited in this article (and in Nostratic studies in general) are unacceptable on phonological grounds, and based on outdated sources/reconstructions. --AAikio 14:53, 24 November 2006 (UTC)


 * Of course they don't need to be the same. The example of the meanings of "grain" in English demonstrated this very well.
 * David Marjanović | 22:44 CET | 2006/12/15

Footnotes to Personal pronouns section
I believe footnotes 6 and 7 were switched around, so I have fixed them. Also, footnote 1 cast doubt on Ruhlen's citation of Chukchi as being "Chukchi in the narrow sense" because of the latter's supposed lack of [ɣ]; I've taken this parenthetical out, since Chukchi does have [ɣ] at least according to []. I'm also changing "Kamchadal" to "Itelmen", since that seems to be the more current name; feel free to revert if I am wrong. I'm also changing the "Proto-Chukchi-Kamchadal" column heading to "Proto-Chukotko-Kamchatkan" for the same reason. 24.159.255.29 05:05, 23 November 2006 (UTC)

I forgot to mention that, according to [], the Chukchi words for "I" and "you" are /ɣəm/ and /ɣət/, respectively -- not /ɣem/ and /ɣet/. I'm not changing that though, because I don't know specifically what Ruhlen said the forms were. 24.159.255.29 05:11, 23 November 2006 (UTC)


 * I may have looked at another description of Chukchi, not necessarily the same dialect. As I don't have any evidence that you're wrong (and don't have time for checking anyway), I won't revert anything. "Itelmen" is of course fine, I only think "Kamchadal" is more widely known.
 * I'll try to check if I switched the footnotes. Of course it's also possible that Kaiser & Shevoroshkin switched them in their explanations; don't forget that most of my data comes from this one article which is no less than 18 years old and clearly outdated in some respects (Proto-Altaic, probably Proto-Uralic, let alone Proto-Afro-Asiatic).
 * I can't check currently what Ruhlen said, but who says Ruhlen can't have mixed it up? It seems to be a tradition in Eskimo-Aleut linguistics (and the orthography of Alutiiq!) to use e for /ə/ because it's so convenient – those languages have at most 4 vowels: /a/, /i/, /u/, /ə/. Maybe Chukchi is similar, what do I know.
 * David Marjanović | 22:51 CET | 2006/12/15

More issues with the recent edits
I can't see the point of transposing the two paragraphs after the list of representative families. The Sumerian/Etruscan/Elamite paragraph is really just an extension of the list (only a tad less representative) and in its present location it breaks the continuity between the two paragraphs which deal with the Eurasiatic/Afro-Asiatic question. I'm going to move it back to its original position, and also remove the reference to ChukKam (for the time being) because it's not accurate to say that it had not been considered before. Dolgopolsky & Golovastikov did consider it in 1972, and Collinder included it in his 'Eurasian' macrofamily (similar to Eurasiatic) in 1965. However, I will be inserting something soon about ChukKam and other 'Paleosiberian' groups, either just before or just after the paragraph on Sumerian etc. The sentence on Greenberg's Eurasiatic would still be correct since, although he left a few families out, he didn't really add much. In fact I'm not sure that we can strictly say that Eurasiatic 'overlaps' Nostratic at all: did Greenberg include anything which somebody or other hadn't added before?

I have other comments and suggestions but they will have to wait until another day. After that, I will be inserting comprehensive references, at least for the bits that I've added over the past few months. Previous contributors are of course more than welcome to add their references too: the more the better. 172.214.141.88 11:16, 1 December 2006 (UTC)


 * You are probably right. Go ahead.
 * David Marjanović | 22:53 CET | 2006/12/15

POV?
I smell POV in this:

"When this method is applied with strictly correct linguistic methodology, the evidence suggests that no such proto-language existed. However, several linguists have selectively misread the data to suggest that the theory has some factual basis. For this reason, the Nostratic hypothesis is viewed within mainstream historical linguistics as a fringe theory."

I don't know much how it is read but you don't say how it is misread?

If the word has to mean exactly the same thing to be considered related, many Indo-European etymologies are "fringe" theories. E.g. "shit" is suggested to come from a word that means "split, separate" and "hammer" from a word that means "stone". --85.156.240.76 12:25, 11 December 2006 (UTC)


 * Where do you get the idea from that anyone insists on identical meanings? The paragraph neither says nor even implies that. It seems to be a common claim by long-range fans against sceptical linguists that semantic drift is somehow a problem, but that's a straw man.
 * If anything, the problem with long-range comparisons is that after so many millennia, it is reasonable to expect that sound, formation and meaning can have changed so much that even the generous semantic leeway admitted by long-rangers. Consider all the changes that happened between attested Latin and even highly conservative modern Romance languages such as Sardinian, Italian, Spanish or Portuguese, within 2000 years; consider the staggering changes in every respect in between reconstructed Proto-Indo-European and even conservative modern Indo-European languages like Lithuanian, Old Prussian and Modern Greek, which is likely to span a period of not much more than 5000 years; and then consider how much can happen in additional 5,000 years (from Proto-Indo-European to something like Proto-Nostratic) and likely has!
 * The actual problem with long-range comparisons is that the compared reconstructions are too similar, rather than too dissimilar, for common inheritance to be a plausible explanation, if Nostratic is supposed to have a time-depth like that.
 * We of course expect some cognate words to retain wide distribution, similar or identical meanings, and/or highly similar shape (let's say 'two' du- and 'night' nVkt-, nVt- in Indo-European, or 'to go' men- and 'fish' kal-, kol- in Uralic), but more particularly, Nostraticists seem to find too many semantic and sound correspondences where barely anything has changed.
 * Following your Standard French and Standard German example below, tons of related words like beurre : Butter, bleu : blau, faux : falsch, nord : Nord can be easily found, but almost all immediately obvious cases will represent loans and not cognates. The data for actual etymological non-trivial correspondences like *dʰ- > f- : t- is going to be much harder to find. In retrospect we can identify a few cases like faire : tun or ficher : Teich, but we might be hard-pressed to claim them as clear counterparts, especially if we fail to establish the regular correspondences existing in them. -- Trɔpʏliʊm  • blah 11:01, 31 July 2015 (UTC)
 * Nobody, or almost nobody, is saying that something like Proto-Nostratic, a proto-language as far removed from Proto-Indo-European and its closer relatives than Proto-Indo-European itself is from modern languages like Lithuanian and Modern Greek cannot have existed in principle, in fact it seems likely that such a proto-language did exist; something like Indo-Uralic may be on the border of demonstrability, but the old contacts make it hard to distinguish old loans from possible common inheritance; the actually disputed problems are: Is the similar vocabulary inherited or borrowed (even across Eurasia, which is well possible)? Are the similar grammatical elements (morphemes), all very short, sufficient evidence of relatedness? (Irregular paradigms are far better evidence, and in fact the best evidence we can conceive of, because such paradigms are utterly unlikely to be borrowed, nor are they likely to be due to chance if they are distinctive enough and irregular in the same ways.) How much of the similarities can be due to chance? How much can be due to contact? How much of the data and reconstructed proto-forms (such as Proto-Uralic forms) used by Nostraticists is even correct, and how much is incorrect, outdated or spurious?
 * Yes (to address David's argument farther below), if our data for Indo-European were no older than 100 years, but they were rich and from numerous languages, and included dialects, we could probably figure out that Indo-European is a true language family, although whether we could achieve a reconstruction anywhere as good as we have now is doubtful. But that's not what we actually have! We have only fragments of often poor and dubious quality, not even from the same time, from a small number of proto-languages. We just do not have the equivalent of Bavarian data to check Standard German /tseː/ against; we have only the equivalent of fragmentary, poorly preserved scraps of Standard German and Standard French (plus a couple others) and no collateral dialects to help. That would make the whole job much, much harder. (For example, a collateral family related to both Indo-European and Uralic but long-isolated from both could be invaluable to help distinguish inherited from borrowed material, cognates from loans, and to bridge the gap between the enormously different sound systems.) There are in fact situations much like that, where a putative language family consists of a bunch of (not necessarily very closely related) long-extinct languages of which missionaries recorded a couple of short word-lists. The situation of big, deep language families in the Americas is nowhere like that, because there are usually several well-attested languages (such as Nahuatl in the case of Uto-Aztecan, with a ton of living dialects and a real lot of significantly older material) and small component families (such as Numic). And by the way, Proto-Algic has not actually been reconstructed. There is very little material. The Algic connection has barely been demonstrated to the satsfaction of specialists, but that's it. And in the case of Athabaskan–Eyak–Tlingit, what I've read implies that it's nowhere as old as Indo-European (only 2nd, not 4th millennium BC), more comparable to Iranian or Balto-Slavic perhaps. It all comes down to the quality of the data and structure of the family (Austronesian has an elaborate branching structure, which helps a lot), which has to be favourable. A couple of big successes in the Americas (and elsewhere) are offset by lots of failures (or dubious cases, such as Afro-Asiatic), where for various reasons recovering a deeper affiliation has been difficult or proved impossible, even though it may well be possible that several smaller families are indeed remotely, but not too remotely, related, and with more complete data, it would have been possible to demonstrate the connection, but the data isn't there (that's why in historical just like in typological research, even the most unimportant seeming language or dialect – spoken in the middle of nowhere by a poor, marginalised ethnic group – can matter, and in fact can matter deeply; to linguists, there are no inherently unimportant or uninteresting languages and dialects, because every language can harbour precious data for some purpose and rare or unique phenomena).
 * Personally, I would love to know where Proto-Indo-European comes from, but the work on Nostratic fails to convince me. --Florian Blaschke (talk) 02:23, 30 July 2015 (UTC)


 * As an aside, I neglected to mention that Uralic and Austroasiatic have in the meanwhile (since David's comment) been argued to have a "flatter", more comb-like tree structure, with Samoyedic and Munda respectively merely a branch (or even sub-branch) among several others, and both families not necessarily older than 4000 years. The families with the gretaest time-depth whose validity is really uncontroversial are probably Northeast Caucasian, Austronesian, Semitic, Indo-European, Sino-Tibetan, Trans–New-Guinea (core members), Oto-Manguean, Mayan, Dravidian, Uto-Aztecan, and Tupian, plus some subset of Niger–Congo (Atlantic–Congo?) and perhaps of Afro-Asiatic (Northern Afro-Asiatic?), although their precise ages are uncertain and none are necessarily older than 6000 years. Algic should probably be included as well (Algonquian is said by us to be to be only 2500–3000 years old, though) – Proulx' papers listed on Algic languages suggest that the evidence is better than I admitted. Most secure language families do not have a great time-depth, however, conspicuously in regions (outside Mesoamerica, Europe, southern regions of Asia and some few parts of Africa) where no ancient texts are available, so there cannot be any confidence that deep reconstructions are actually practicable. --Florian Blaschke (talk) 02:57, 4 August 2015 (UTC)

Problems with another recent edit
To the list of criticisms someone added the following:

"The theory seems to assume that languages (or language families) which possess a long written history have evolved away from the proto-language faster than languages which possess only relatively recent written records. For a given Nostratic etymology, modern French or German will as a rule fail to provide exact correspondences, whereas equally modern Eskimo seems to be able to do this quite easily. Given only the words "tze" (German "Zehe", toe) and "dwa" (French "doigt", finger), it is highly dubitable that one would be able to find even a common Indo-European root, nevermind a Nostratic one, whereas isolated words taken from modern Uralic or Eskimo languages time and time again seem to possess phonetic forms whose connections with the Nostratic roots are completely obvious."

Wait, wait, wait, wait. You can use this criticism against Proto-World, but you cannot use it against Proto-Nostratic. You see, Nostraticists use the good old comparative method. Except for the isolate Nivkh, Nostraticists do not use the modern languages, but the reconstructed proto-languages – Proto-Indo-European, Proto-Uralic, Proto-Altaic, and so on – as their input.

As for the claim about isolated words from modern Eskimo languages, I disagree. I mentioned my doubt behind my quoting of Bengtson's comparison of Proto-Eskimo /qat/ with Proto-Nostratic /k'Erd/ because I have yet to see evidence that a Proto-Nostratic velar can correspond to a Proto-Eskimo uvular (in the light of the fact that reconstructed P-N has both velars and uvulars).

Given only French and German, you would not find a common Indo-European root. But given a dozen modern IE languages, or two dozen, I think you would. So unless someone shows I'm wrong, I'll delete this "criticism" next week for being aimed at a strawman.

David Marjanović | david.marjanovic_at_gmx.at | 23:04 CET | 2006/12/15


 * Done that.
 * David Marjanović | david.marjanovic_at_gmx.at | 00:32 CET | 2006/12/29 —The preceding unsigned comment was added by 131.130.1.143 (talk) 23:32, 28 December 2006 (UTC).

Check out what's written under "sample Nostratic etymology" in this very article, genius. "Burunge", "Mehri", "Alagwa", "Kolami" - these are all ancient languages comparable to Latin and Old Slavonic, aren't they? NOT to modern German and French of course.


 * And at the beginning of the paragraph, what do you see? "Proto-Afro-Asiatic". I agree that there is currently no good reconstruction of PAA, but this one is clearly better than nothing.
 * Besides, this one reconstructed root is by Bomhard and Kerns. Neither are there just two Nostraticists in the world, nor do they agree on everything, nor is there just one such reconstructed word; if you would read the article, you would figure out these three points in no time. David Marjanović 22:31, 11 March 2007 (UTC)

The problem is that most of your "reconstructed" proto-languages are a joke in themselves. How can you claim to reconstruct a proto-language, which I assume must be several thousand years old, without having any written records older than a century (I am not speaking about Semitic here)? You would've NEVER been able to do that with Indo-european. (Bearing in mind that modern European languages use historical othography, which Eskimo languages clearly don't have.)


 * Excuse me? Proto-Athabaskan? Proto-Athabaskan-Eyak-Tlingit? Proto-Algic? Proto-Munda (not to mention Proto-Austro-Asiatic, which you can't reconstruct without Munda, no matter how much written Vietnamese you have)? Proto-Samoyed (not to mention Proto-Uralic, which you can't reconstruct without Samoyed, no matter how much written Finnish and Hungarian you have)? People do that all the time. Yes, it is more difficult, but it's nowhere near impossible. Few people doubt the existence of the Afro-Asiatic language family any more, so there must have been a protolanguage, and it must be possible to reconstruct it at least partially. PIE would have taken longer, but the result would have been very similar (think of the nominative -s preserved in Greek, Lithuanian, and Latvian, for example). David Marjanović 22:31, 11 March 2007 (UTC)

I was aiming at a strawman? You're just trying to cover up the weak points of your theory by refering to dubitable research carried out by THE SAME PEOPLE THAT ADVOCATE THE NOSTRATIC THEORY.


 * "My theory"? I haven't contributed at all. I just use what little literature I can get, because nobody else does it. But yes, you were aiming at strawmen, such as mass lexical comparison, as I have explained above. No, that the same people carry out the research and as a result of that research advocate the hypothesis is not a problem; by your logic e. g. evolution would be wrong! David Marjanović 22:31, 11 March 2007 (UTC)

P.S. Btw, going around and deleting criticisms in Wikipedia articles is REALLY going to help with achieving scientific acceptance.

Remember WP:NPOV. Muhaha 18:40, 19 February 2007 (UTC)


 * I didn't delete them, I moved them to the talk page and explain there why I removed them from the article. What more do you want? BTW, stop shouting. Talking of scientific acceptance. David Marjanović 22:31, 11 March 2007 (UTC) Edit: added links to "Criticisms" section David Marjanović 22:56, 11 March 2007 (UTC)

Besides… let's dwell on your example a little. The Standard German word for "toe" is indeed. Now compare this to the word of my dialect of German:. Suddenly it becomes apparent that the silent h in Zehe is etymological, like the silent g in French doigt. Then compare Dutch teen and English toe and maybe a few more. If we do that with enough words, we will figure out the High German consonant shift (even though we won't be able to reconstruct its development in time and space with the accuracy that Old High German texts allow us). That's the key: regularity. Yes, for the third time, Nostraticists use regular sound correspondences.

So we'll be able to reconstruct the Proto-Germanic form, and so on.

Hey, biologists don't need fossils for reconstructing phylogenies either. They just help (by removing degrees of freedom and by exposing convergence). David Marjanović 22:56, 11 March 2007 (UTC)

Actually, now that I think of it, there is a problem with Bomhard & Kern's reconstruction of Proto-Afro-Asiatic */bar/ (which you overlooked while your blood pressure was rising higher and higher): they can only find its derivatives in two of the many PAA subfamilies, Semitic and Cushitic.

However, this is not a big problem. This kind of phenomenon happens in IE all the time – for an amusing example, read the first post here.

I think the introduction of that "sample etymology" should be changed. Someone needs to establish that it is in fact representative, and we should remove the stupid paternalistic "we". I think we should put it below the pronouns and above the other vocabulary. David Marjanović 00:02, 12 March 2007 (UTC)

"No regular correspondences have been found"-claims
Why do the so-called mainstream linguists claim that there are no correspondences even if there are?

And why "everyone" thinks Nostratic hypothesis is result of mass-comparing modern words (e.g. http://en.wikipedia.org/wiki/Talk:Estonian_language#Nostratic_hypothesis )

I think this should be rewritten: http://en.wikipedia.org/wiki/Estonian_language#Vocabulary

--85.156.132.81 20:05, 21 December 2006 (UTC)

From the Criticisms section
Were one to collect all the words from the various known Indo-European languages and dialects which have at least one of these 4 meanings, one could easily form a list that would cover any conceivable combination of two consonants and a vowel (of which there are only about 20*20*5=2000).

I don't understand the point of this sentence. Would the author (Taharqa?) please explain it? David Marjanović 22:58, 21 March 2007 (UTC)

History of the Nostratic theory (Section "Origin")
Henry Sweet's book "The History of Language" (1900) discusses something very akin to the Nostratic theory in chapter VII "Affinities of Aryan", devoting some twenty pages (pp. 112–132) to assorted evidence of relationships between Aryan (Indo-European), (Finno-)Ugrian, Altaic and Sumerian. (Sweet expresses his conviction that these relationships are valid.)

On page 114 he mentions that (Nikolai) Anderson was the first to investigate the relationship between Aryan and Ugrian (Indo-European and Finno-Ugric) "in a scientific and impartial spirit" (in his bibliography, Sweet lists "Anderson, N. Studien zur Vergleichung der Indogermanischen und Finnisch-ugrischen Sprachen, Dorpat, 1879").

This info could be used to expand the "Origin" section a bit, showing that relationships between a number of these languages had been proposed and discussed prior to 1903.  Jayen 466 10:19, 19 September 2007 (UTC)


 * Nice one. I'll try to grab a copy but it'll take a while so why not go ahead and expand away. If you're interested in chasing them up, it seems that Delitzsch compared IE and Semitic in 1873 (Stud. üb. indogerm.-sem. Wurzelverwandtsch), and Abel compared IE, Semitic and Egyptian in 1887 (Einl. in ein ägypt-semt.-indoeurop. Wurzelwörterb.). There's also MacCurdy's Aryo-Semitic Speech (1881). I know nothing about any of them beyond a footnote in a journal for Biblical studies dated 1903 (in German). Gnostrat 22:45, 19 September 2007 (UTC)


 * Hi, thanks for the encouragement; I've had a go. Btw, I've just learnt that a facsimile version of Sweet's book was published quite recently (see ISBN in the footnote). It's also searchable to a limited degree in google books.  Jayen 466 18:24, 20 September 2007 (UTC)

More on Nostratic morphology
Hi, I'd love to read more on Nostratic grammar. Judging from the poem at the bottom of the article, Proto-Nostratic must already have been reconstructed to the level of declension and conjunction tables. What kind of language was it? Most language groups included are agglutinative, but Afro-Asiatic and Indo-European are inflecting. And was it absolutive or ergative. I am sure some Nostraticists at least have published guesses or suggestions. Steinbach (fka Caesarion) 11:08, 4 October 2007 (UTC)

Language family size: how to define it?
Gnostrat, I changed the first sentence because a "proposed language family" is by definition not accepted by the majority of historical linguists.

As to language family size, how do we define this? Do you have any information on this subject? Should we count the number of languages that belong to the family, or the number of families of which it is composed? Number of speakers doesn't seem like a very strong criterion. Time depth doesn't seem like an entirely secure criterion, since languages can differentiate into separate languages at different rates. Taxonomic depth could provide a criterion, but does this lead to radically different time depths, such that a language family at one taxonomic depth is recent, and another at the same taxonomic depth is in the Stone Age? Regards, VikSol 20:11, 10 October 2007 (UTC)


 * Crikey. Well, I'm not sure there's any one right answer here. I modified your edit on the assumption that what you meant by size was numbers of languages, since that seems to be the usual sense in the literature that I've read. If it's numbers of families, then which nodes at which taxonomic level/rank do we count? We could define numbers of families in terms of distinct, so far unaffiliated stocks which it is proposed to affiliate (i.e. what are usually termed phyla) but that is uncomfortably like relying on the quality of the linguists rather than on the quality of the evidence.


 * I think you're right about numbers of speakers, and in spite of glottochronology we still don't have a consensus method of estimating time depth. Fortunately, we probably don't need one for the present purpose. If we didn't want to define size by numbers of languages, it would make most sense to go for taxonomic depth, which is also time depth but expressed as relative age. Gyula Décsy and, later, Irén Hegedüs (see her "Principles for palaeolinguistic reconstruction" in Roger Blench & Matthew Spriggs, Archaeology and Language I, 1997: 65-73) proposed a rough scheme of ranks defined by a relative chronology, with broad and flexible time limits that allow for adjustment without affecting the relationship between the levels. Something along these lines (which I've modified a little to take account of current long-range hypotheses):


 * I don't know if that answers your question. My point was that family size &mdash; whether you define it in terms of language numbers or time depth &mdash; isn't really relevant. Nostratic isn't distinguished from well-established phyla by numbers of languages because Trans-New Guinea is about as big while Niger-Congo and Austronesian are much bigger. If you say Nostratic is distinguished by greater absolute time depth, that will be true in many cases but you might run into trouble with Niger-Congo, Nilo-Saharan and Australian (arguably as old if not older) and Khoisan (25-30 Ky?)


 * We need to be wary of the assumption that time depth is the factor which distinguishes securely established families from unvalidated ones like Nostratic. Those who argue that relationships cannot be established at higher time depths are making an unproven assertion. To quote Mary R. Key (Intercontinental Linguistic Connections, University of California, 1981: 18), "there are no criteria, in comparative linguistics, to indicate how far back one can go with the comparative method". I'm pretty confident that Nostratic will be validated in one shape or another, and I also suspect that the real reason why we cannot further connect up the 7 or 8 proposed superphyla, in spite of the Borean and Proto-World speculations, may have nothing to do with the time depth. It may simply be that the original out-of-Africa migrants came equipped with something that was on its way to becoming language but was not quite there yet, and true language evolved independently in several different population centres after the initial global migration. But I'm digressing and, since I'm only an interested amateur, I'd better shut up right there. Gnostrat 21:58, 12 October 2007 (UTC)

Order of languages, Sound correspondences re-visited
I have rearranged all tables so that the order is IE, Uralic, Altaic, Kartvelian, Dravidian, Afro-Asiatic, in line with more recent views of Nostraticists (cf. the Tower of Babel database). In some cases, the main "Euraltaic" core becomes more obvious. In some cases, though, it appears to me that Kartvelian is quite close to Proto-Nostratic and should be leftmost. This goes especially for the consonants, but much less so for the vowels. I might re-arrange again later.

The other thing I changed is that in the sound correspondences table, the usual transcription and the IPA values were written twice even if they were written with the same letter, so I've deleted about half of them. Quite frankly, the table was almost unreadable before to the less avid linguists. Now, it's much less frightening, especially the vowels.

Also, I have merged three pairs of consonants according to the phonetic table. The number of consonants is daunting enough as it is, and given that Nostratic is more probably over-reconstructed than under-reconstructed, there are still more than enough consonants.

I hope this is ok with the diligent editors who started this page. Chris cdblome ät yahoo dod komm —Preceding unsigned comment added by 91.7.28.57 (talk) 20:49, 3 May 2008 (UTC)


 * I believe I've commented on this before, but the Membership table was already in simultaneous geographical AND phylogenetic order, starting with the most divergent branch (AA) and finishing with Eskimo-Aleut (which separated most recently from Altaic according to authors like Starostin). That way, you get a good idea of the geographical sweep of the superphylum from west to east, as well as a coherent sequence of successive branchings which track the movements of Nostratic-speakers through time. It also conveys an idea of where the macrofamily originated by the fact that its oldest branchings cluster in western Asia. This will be harder to perceive if we are starting way down the tree and working outwards.


 * Also, because this is a representative list, not an exhaustive one, we don't need to clutter it up with a "maybe Yukaghir" here or a "not always included" tacked on to Eskimo-Aleut. (Actually, apart from IE and Uralic, there are not that many branches of which you can say that they are always included. Check out this site, for example, which uses Tower of Babel data to build a case for a Nostratic affiliation for Turkic, but a non-Nostratic origin for the rest of 'Altaic' &mdash; neatly combining the Nostratic with the Anti-Altaic hypothesis, although this is very much a fringe view.)


 * The phonetic tables are tidier, though, after the effort you've put in. I'm not being grumpy. Gnostrat (talk) 22:52, 3 May 2008 (UTC)

I guess an apology is in order here: Sorry! I've reverted my change of the representative grouping. The subdivisions are shown in the genetic classification box anyway. Oh, _that_ page. Interesting. I remember reading it and finding out that people from what is today Mongolia founded all the relevant societies in Eurasia, agriculture and all. Other than that, I'm waiting for a book on Nostratic I ordered, so I might add some information later. I'll try to write 'around' former contributions and I guess I should get myself a name as an editor. Cheers, Chris —Preceding unsigned comment added by 91.7.13.41 (talk) 16:02, 4 May 2008 (UTC)


 * No apology needed : ) The list was here before I was, and it's open to discussion like anything else. I thought Stetsyuk's method looks a little unorthodox, though it pins down the Urheimat to the Transcaucasus, same as Renfrew. But Mongolia?? If that suggestion has any significant Nostraticist support, it should probably be added to the Urheimat section. It would work for the Eurasiatic half, but I've yet to hear any archaeologically plausible scenario which would bring AA from Mongolia.


 * Registration's a good idea, though I'm not exactly sure how long it takes to make your name as an editor around here! (Where's my Barnstar, guys? Guys?) Gnostrat (talk) 00:43, 8 May 2008 (UTC)

Get myself a name, whatever. Barnstar? White man speak in riddles. English too complicated for me. By the way, I put the Bomhard book that is about to appear on my Amazon wish list. Just when the item is confirmed to be on the wish list, an expected price appears: 280€... —Preceding unsigned comment added by 83.236.164.131 (talk) 07:24, 8 May 2008 (UTC)

The price for the book is up to 325€, but I just found a version of Dolgopolsky's Nostratic Dictionary that is available online. It has 3124 pages so far. I doubt this belongs into the article, but I'd like to mention it: Nostratic is expensive and time-consuming, hence the lack of supporters.Physiognome (talk) 15:39, 21 May 2008 (UTC)

In-text citations
Removed a "no footnotes" tag that has been added, which asserts the article lacks "in-text citations", for the following reasons:

— Footnotes are not the only form of in-text citation authorized in Wikipedia (Citing sources). Wikipedia recommends only that one of the authorized citation forms be followed and that the citation form be consistent throughout the article. The form followed in this article is author-date referencing, in contrast to the use of footnotes. In my experience, author-date referencing is less cumbersome than footnotes to execute and works better with electronic material, where there is no "bottom of the page" to place the note on. It is the current trend in scholarly references.

— It is incorrect that this article lacks in-text citations. The article contains the following in-text citations, to be precise:

(Sweet 1900: vii, 112-132), (Szemerényi 1996:124),  Georgiy Starostin (2002),  ),  Ruhlen, 2001,  Greenberg ...(2002: 2),  Kent Flannery (1969),  (Kaiser & Shevoroshkin 1988),  Kaiser & Shevoroshkin (1988),  Starostin,  ,  (Yakubovich 1998 ),  Altaic Etymological Dictionary (2003; see Altaic languages article),  G. Starostin (1998),  Kaiser & Shevoroshkin (1988),  Kaiser & Shevoroshkin (1988),  Kaiser & Shevoroshkin (1988),  Kaiser & Shevoroshkin (1988, p. 314f.),  (from Bomhard and Kerns, The Nostratic Macrofamily, p. 219),  (from Gamkrelidze and Ivanov, p. 598),  Kaiser & Shevoroshkin (1988),  Bengtson (1998),  (cf. Bomhard, Indo-European and the Nostratic Hypothesis, Signum Publishers, 1996.), (e. g. Ruhlen, 1998), Ruhlen (1998),  Norquest (1998),  Kaiser & Shevoroshkin (1988),  Kaiser & Shevoroshkin (1988),  Bengtson (1998),  (Kaiser & Shevoroshkin 1988; see also */tV/ instead of */t̕V/ above),  Kaiser & Shevoroshkin (1988),  (Kaiser & Shevoroshkin 1988; the Proto-Eskimo form given by Bengtson [1998] may indicate that the vowel was /æ/ or not),  Campbell (1998),  (Ruhlen 1994),  Bomhard and Kerns (1994) – a total of 36 citations to 21 different works, mostly in author-date format, plus a few in-text external links.

— In addition, the following in-text citations are included in material quoted:

Pokorny 1959:111, Walde 1927-1932. II:134, Mann 1984-1987:66,  Watkins 1985:5-6,  Gamkrelidze-Ivanov 1984.II: 872-873,  Ehret 1980:338,  Burrow-Emeneau 1984:353, no. 3959,   (Būga 1958-1961:I.300) – a further 8 citations.

Grand total: 44 citations.

Nevertheless, the fact that an experienced administrator has perceived a deficiency in referencing in this article suggests that additional references may be called for. In my view, these would be especially helpful in the sections on "Background" and "Nostratic Urheimat", which at present are lightly referenced.

VikSol 02:47, 19 May 2008 (UTC)


 * The tag was definitely inappropriate but I'm finding it's all too common for editors to scan the article, find no footnotes, and slap a no-citations tag on (or even nominate it for deletion) without a further thought. You are also right that further references are called for and I can add in a fair number, but I've been reluctant out of uncertainty whether or not to switch the entire citation system to footnotes first. There are advantages and drawbacks either way. In-line superscript links to numbered footnotes create uneven line spacing and produce an untidy looking article where the separate paragraphs can sometimes barely be distinguished. Another result is an overlong Notes section if every sentence carries a citation or two. On the other hand, the superscript numbers do compensate somewhat through the economy of multiple inline links to a single footnote, whereas the author-date citations can be repetitive and would have to be used sparingly or they would crowd the text. Sometimes you can make a choice of systems based on what the article most badly needs, but the costs and benefits to this one aren't very clear. Gnostrat (talk) 18:15, 19 May 2008 (UTC)

Gnostrat,

I agree there is something to be said on each side of this issue. An advantage of footnotes (really endnotes in the Wikipedia format) is that the reference number is small and therefore, presumably, not disruptive of the reading experience. On the other hand, they disrupt the reading experience in another way, since you have to constantly flip back and forth to see what the note says, and whether it contains additional text or just a technical reference. I also notice that errors in punctuation tend to cluster around footnote numbers, producing a disconcerting appearance. The worst disadvantage of footnotes in the Wikipedia format, I think, is to editors, since the bracketed notes invade the text of the "edit" page, making them – as others have pointed out – almost unreadable, and therefore hard to edit. This may be why so many errors in punctuation cluster around them. In contrast, an author-date reference is a little longer, but you see right away all the information it gives, and the "edit" page remains entirely readable.

Wikipedia does recommend that whichever style is first used in an article be maintained subsequently – a good way of reducing constant switching back and forth. Quoted from Citing sources:


 * There are a number of citation styles and systems used in different fields, all including the same information, with different punctuation use, and with the order of appearance varying for the author's name, publication date, title, and page numbers. Any style or system is acceptable on Wikipedia so long as articles are internally consistent. You should follow the style already established in an article, if it has one; where there is disagreement, the style or system used by the first editor to use one should be respected.

So this would seem to solve the issue for the Nostratic article.

I think you could go ahead and add additional references in author-date format without unduly cluttering the text and in any case footnotes should be avoided in this article, since the other style is already established.

If a reference crops up constantly it can be abbreviated, e.g. Kaiser and Shevoroshkin 1988 could be abbreviated to KS 1988 or even just KS, provided the abbreviation is listed under References, for instance as (Cited as "KS".) Personally, I don’t think this is necessary here, but it would certainly be defensible.

The main technical differences between in-text numbers with notes and author-date citations, it seems to me, are that (1) notes can include additional text – the counter-argument is that if it’s important enough to include, it’s important enough to include in the main text – and (2) footnotes can include titles and publication details of works cited, whether or not a list of references is present, whereas the author-date method absolutely requires a list of references, which serves as a key to the abbreviated references used. As a list of references is useful in various ways, this seems to me to be a small price to pay.

Regards, VikSol 05:33, 21 May 2008 (UTC)


 * Thank you for taking the time to make what looks like a pretty comprehensive case for the author-date system! I might refer people to this talk page if there's a similar dilemma on another article. Though perhaps we shouldn't rule out a few footnotes for additional text or commentary, if necessary. This would not detract from the overall author-date system, and I have known cases where relevant and useful information (such as supporting quotations) would unduly disrupt the flow of the main text.


 * Incidentally, I noticed that you removed the Edward Finnegan quote, "there's too much there to be nothing, but not enough there to be something." It seems a pity, as it was such a good pithy summing-up. Do you, or does anybody, have an idea where a source for this might be found? Gnostrat (talk) 19:28, 21 May 2008 (UTC)

Vowels
I see a contradiction in the text. Of the vowels, it says "the phoneme /y/ does not occur in any reconstruction other than Bomhard's". Immediately after, the sound correspondences table is given with the remark that it comes from Kaiser and Shevoroshkin, and this table does include /y/ despite not being attributed to Bomhard. What's the matter? Steinbach (fka Caesarion) 17:14, 21 May 2008 (UTC)


 * Yes, that's the impression it gives, but the exact text (cut and pasted) is "This phoneme does not occur in all non-Bomhard reconstructions of Proto-Nostratic", leaving open the possibility that one or both of the two sources cited for the table does in fact have /y/. The text is not in contradiction with itself but it could perhaps be more explicit. VikSol 18:46, 21 May 2008 (UTC)


 * Thanks for your quick reply. Okay, perhaps I was a little fooled by my native language. I understood "this phoneme does not occur in any non-Bomhard reconstruction". I'm not sure if the sentence could be ambiguous in English. Anyway, I would be glad if it were put more clearly. Steinbach (fka Caesarion) 20:21, 21 May 2008 (UTC)
 * Has been done, I see. Thanks! Steinbach (fka Caesarion) 20:25, 21 May 2008 (UTC)

Edit to criticisms / responses section
The edit that puts together the criticisms and responses section has been reverted for the following reasons:

After years of being subject to editing wars, this article has finally achieved stability. Apparently, everybody feels that their point of view is sufficiently represented and that the article reports all views acceptably.

The reason the criticisms and responses section was split in the first place is that an anti-Nostratic reader felt that the previous arrangement allowed pro-Nostraticists to shoot down the objections one by one.

By dividing the sections, the anti-Nostratic side is free to build a case. The pro-Nostratic side has the disadvantage of not being able to structure the argument, but in turn gets the last word. It's rather like the balancing of prosecution and defense in legal proceedings.

Readers have to jump back and forth between these sections to get both sides of the issue, but the advantage is that it is easier for them to scan through the pro- and anti-Nostratic arguments as integral wholes. The overall amount of effort they have to make is about equal with either arrangement.

Thus, we have got an editing balance here that needs not to be disturbed.

While it seems more logical to put the sections together, experience has shown the value of keeping them apart. Sincerely, VikSol 18:52, 19 July 2008 (UTC)


 * Not realising this has come up before, I recently tried to do the same thing with this page- consolidate the separate paragraphs and sections into single sections that flows together, not realising it would be reverted. To me, the system of having separate sections for each point of view spoils the flow of the article and stands in contrast to the rest of Wikipedia, where there are many articles with editors who take different lines on an issue keep the issues in the same section, rather than separating them out.  Abortion, The Troubles and Homeopathy are examples.  The NPOV policy WP:NPOV concurs:


 * "Segregation" of text or other content into different regions or subsections, based solely on the apparent POV of the content itself, can result in an unencyclopedic structure, such as a back-and-forth dialogue between "proponents" and "opponents".[6] It may also create a hierarchy of fact: details in the main passage are "true" and "undisputed", whereas other material is "controversial" and therefore more likely to be false, an implication that may be inappropriate. A more neutral approach may result by folding debates into the narrative rather than "distilling" them into separate sections that ignore each other.


 * You argue that "The reason the criticisms and responses section was split in the first place is that an anti-Nostratic reader felt that the previous arrangement allowed pro-Nostraticists to shoot down the objections one by one. " yet, under the current system, the pro-Nostraticists are still shooting down all the objections one-by-one.  They're just doing it in a different section, mcuking up a reader's chain of thought and actually making the anti-Nostraticist's job harder because they would have to but in an additional section if they later want to put in counter-arguments to the objections.  (I myself don't know enough about the subject to be pro- or anti-, I suppose I'd like it if the theory was true, while at the same time not knowing if it is true).  Also, I think that some of the statements in that section could do with sources; so that instead of saying "Nostratic comparisons show excessive semantic latitude. Nostraticists should only compare words whose meanings are identical or close to identical."  it would say "Professor J Bloggs-Smith argued in his book 'Cognates etc' that Nostratic comparisons show excessive semantic latitude and that Nostraticists should only compare words whose meanings are identical or close to identical.  In response, Dr J Doe-Jones said ..."  This might help the problem of it looking like one group is constantly shooting down the other group, statements would not be placed bluntly on a page as an absolute truth ("Green-bearded people are more likely to be violent") but be expressed as the view of a specific person ("James Bligh wrote that Green-bearded people are more likely to be violent") Kaid100 (talk) 20:25, 9 January 2009 (UTC)

Please no letters ten feet tall
While I understand and to some extent sympathize with the sentiment behind it, I have reverted the recent edit, partly because the first sentence already contains two dubitatives, "proposed" and "according to its proponents", which clearly indicate that the family is not generally accepted, mostly because several contributors have been complaining about some who want "highly controversial" signs placed on disputed language families in letters ten feet tall. We need to report, fairly and accurately, that the Nostratic language family is controversial; we do not need to force the note.

The intro to the article contains numerous dubitatives in addition to those in the first sentence. These are highlighted in the version below.


 * The Nostratic languages constitute a proposed language family that, according to its proponents, includes a high proportion of the language families of Europe, Asia, Africa, and North America.


 * The hypothetical ancestral language of the Nostratic family is called Proto-Nostratic, following standard linguistic practice. Proto-Nostratic would necessarily have been spoken at an earlier time than the language families descended from it, which would place it toward the end of the Paleolithic period.


 * Nostratic is sometimes called a macrofamily or a superfamily, but these terms have no scientific signification: they simply denote a language family that groups two or more other language families and is not (or not yet) generally accepted by those linguists who have concerned themselves with the question.


 * In contrast to some other macrofamilies, most versions of the Nostratic hypothesis rely upon an application of the comparative method, involving systematic sound-and-meaning correspondences between the constituent families as well as systematic correspondences in their grammar. Notwithstanding this, the hypothesis is very controversial.


 * The Nostratic hypothesis has varying degrees of acceptance, depending in part on local academic traditions. In Russia, it is endorsed by a substantial minority of linguists working in relevant areas, e.g. Vladimir Dybo, but does not constitute a generally accepted theory.  In the English-speaking world, it is strongly condemned by a minority of linguists, e.g. Lyle Campbell; others take an agnostic view, e.g. Philip Baldi (2002); a few support similar but not identical classifications, e.g. Merritt Ruhlen; declared supporters of the Nostratic hypothesis, e.g. Allan Bomhard, are a small minority at the present time (2008).

Anyone who can come away from this with the impression that Nostratic is not a very controversial theory is best left to their own ignorance.

You want "highly controversial" posted on Nostratic because to you this is the most important fact about it. Put yourself in the position of Nostratic supporters: to them the family is interesting, fruitful, even revolutionary in its implications. I certainly wouldn't support putting such adjectives in the opening sentence of the article, though. We have to balance these competing claims without tipping too far to either side. You don't get the first word in the introduction, but you do get the last. That's the most powerful position in a court case.

Sincerely, VikSol 04:07, 6 August 2008 (UTC)


 * A few points, as a (so far) third party.
 * Actually, the lead paragraph (without something akin to the sentence that was added and reverted) says nothing at all to indicate that the hypothesis is controversial.  The tautologous phrase "according to its proponents" only serves to make the text more verbose, since the membership of the family is by definition what its proponents say it is, no matter how many or few those proponents are.  The word "proposed" isn't redundant, but it also doesn't indicate controversy; it's entirely possible for a hypothesis to be universally supported without being treated by its supporters as an established fact.
 * There is nothing about the sentence in question to suggest that the author of the sentence is not a supporter of the Nostratic hypothesis. I wouldn't be at all surprised if a supporter were to write just such a sentence, and in any case it doesn't matter whether the author takes any position at all, since Wikipedia editors strive to write neutrally even (or especially) about topics on which they are personally passionate.  If a reader came away from the article knowing only what is in the lead paragraph &mdash; and that is entirely possible, if they just happened by and the lead paragraph failed to peak their curiosity &mdash; the two things they ought to know about it are that the family consists of roughly X Y and Z, and that it's controversial.  Of course the sentence needs to be scrupulously neutral since it's in the lead paragraph and helps to set the reader's expectation for the level of neutrality in the article as a whole; but if it falls short in neutrality, the best way to help the article (and foster community spirit, along the way) is to improve the sentence, rather than delete it.  After all, the absence of an up-front acknowledgment of the controversy could actually cast doubt on the article's neutrality if the reader does read on and realizes that the hypothesis is quite controversial.  I admit to surprise that anyone found the sentence objectionable; when I saw the sentence I thought it quite self-evident and NPOV that the hypothesis is highly controversial, and making that point clear up front seemed like a great way (in addition to the other advantages I've already mentioned) to peak that passing reader's curiosity so they're more likely to keep reading.
 * With such a sentence in the lead paragraph, several later passages in the introduction would become partially or entirely redundant, and pruning these might help the whole introduction to flow more freely. Pi zero (talk) 17:40, 6 August 2008 (UTC)


 * After some hesitation (because I really don't want to blow this out of proportion), I'll also remark that the initial comments in this thread used a couple of phrases of the sort that are easy to write casually, but can later tilt everyone's thinking toward unproductive paths.
 * "Anyone who can come away from this with the impression that Nostratic is not a very controversial theory is best left to their own ignorance." As editors of the article, an important part of our job is to make the article as easy as possible to understand.  There is no point at which we get to say "that's clear enough, no need to look for ways to do better than that".
 * "That's the most powerful position in a court case." Best to avoid this comparison.  A court case is an argument; any Wikipedia article that has become an argument is failing in its mission.  A Wikipedia editor shouldn't be trying to balance competing claims, because that requires judgment between them, which fails WP:NPOV.  "Debates within topics are clearly described, represented and characterized, but not engaged in."  Pi zero (talk) 13:25, 8 August 2008 (UTC)

"Highly controversial" vs. "not widely accepted"
Hi.

I was wondering: is it better to call this hypothesis "highly controversial" or "not widely accepted"? Because the former seems to suggest there is a great deal of arguments, discussions, and debates going around about it (controversy), as opposed to just not accepting the hypothesis. Is there such a thing going on with it, or not? And if not, I'd suggest the "highly controversial" thing be changed to "not widely accepted" or something similar. What do you think? mike4ty4 (talk) 05:01, 11 September 2008 (UTC)


 * If you're talking about the lead paragraph, I don't think the "highly" part serves any useful purpose except to promote a POV. On the other hand, "controversial" seems to be a very accurate description, and is in fact already hiding buried further down in the lead (where it manages to get away with being "very").  Regarding the organization and content of the lead section of the article, see also my comments from August 6 in the immediately preceding thread.  Pi zero (talk) 07:50, 11 September 2008 (UTC)


 * From the view of our knowledge of historical human migrations (based on genetics), most of these trans-continental language theories make no sense. I'll bet that in the future they will be revealed as a salad of complete bullsh*t. Did anybody of the involved linguists think over, how Indo-European languages could be related to Altaic languages, when people speaking these languages diverged at least 50 000 years ago and haven't been in any contact until ca. 4000 years ago? Again, how could Uralic languages be related to Indo-European languages, when the ancestors of Uralic speakers dwelled in north-eastern China during the whole Ice Age, and got to eastern Europe as late as 5200 BC? These people should study genetics and not to waste time by creating cracky theories based on accidental resemblances of two or three vowels.
 * For example, we know that speakers of the language families of South-East Asia (Chinese-Tibetan, Hmong-Mien, Austro-Asiatic, Austronesian, Dai-Kadai) posess a high % of a certain lineage of Y-haplogroup O, and these language families are thus very probably distantly related. We also know that some language isolates of today's Siberia like the languages of Ket, Nivkhi, Yukaghir and Eskimo-Aleut, are bound together by the rarely high presence of Q+A lineages that were typical of Paleoindians inhabiting Siberia before the last Ice Age, and thus are very probably related to languages of American Indians (both Na-Dene and "Amerind"). The recent publication of Edward Vajda may be a pioneer work on this field deserving high attention.
 * Alternatively, some Native American languages may be rather related to Altaic languages, because American Indians are actually a mixture of Siberian Paleoindians (Q+A) and old Mongoloids (C3+C, Z, D, G). As Vajda's work itself showed, we can't deduce with certainty, which language these people spoke, because it was a variably mixed population (Kets are typical by high proportion of Y-haplogroup Q, yet their original maternal lineage A was almost completely erased from their gene pool; on the contrary, their alleged relatives from the Na-Dene group posess Altaic Y-haplogroup C3 and Paleoindian mtDNA haplogroup A. This suggests that the lineage of the Altaic man bearing C3 prevailed by chance over a largely Paleoindian population with Q+A).
 * Linguists should concentrate on the possible connection between Paleoindian languages and Burushaski, because Burusho may very well be remnants of an old Aurignacian group with Y-haplogroup R2, to which Q is distantly related. Naturally, Basque with Aurignacian/Cro-Magnon Y-haplogroup R1b may also belong to this group, as well as Indo-Europeans (if they really came from the Ukrainian steppe and bore R1a). But again, since the majority of Basque maternal lineages is of Gravettian origin (HV), the Basque language may be actually closer to some Near Eastern languages. I could quote many other examples, but I don't want to waste space on Wikipedia in a similar way like certain linguists waste paper in prestigious proceedings. 89.235.19.204 (talk) 10:29, 27 December 2008 (UTC)

In the phrase "hypothetical large-scale language family", near the very start, would one of {"conjectural", "conjectured", "hypothesized"} be better than "hypothetical"? Peter010101 (talk) 20:19, 19 October 2015 (UTC)

Inline Citations
This article has gneral references but not inline citations. It could be improved with inlines. Kind Regards SriMesh | talk  01:32, 10 October 2008 (UTC)

Finnish
Well, no language today obviously sounds too similar to this hypothetical reconstructed poem, but I have to say that Finnish would still win if you had to compare any language. I'm really surprised that no one's pointed this out yet. Just look at the poem:

Format: Nostratic = Finnish = English

K̥elHä = kieli = tongue wet̥ei = vuoden ("year's") = of-time kähla = kahluu = ford na = me (but that's already been demonstrated with proto-Uralic) = us wetä = vetää ("pull") = leads ʔälä = älä (imperative mood of the negative verb "ei") = not ja-k̥o = joka = which/who pele = pelkää = fears t̥uba = syvä (affrication, t -> s and b -> v) = deep wete = vettä = water

-- nlitement [talk]  22:27, 11 October 2008 (UTC)

How much is enough?
Pi zero,

As pointed out above in "In-text citations", this article contains 44 references to 29 different works. Nearly everything in the intro section is linked. These links are also references and lead to more information. Sincerely, VikSol 07:22, 25 October 2008 (UTC)


 * Citations does not mean links to other articles, it means that a particular statement is immediately followed by information telling the reader exactly where to go in source material to verify it. For example, the lead section of this article contains five paragraphs of material, yet it contains only two citations (broad ones, though in the particular case that seems reasonable), both in the last sentence &mdash; and those two don't cover the full content of that sentence.
 * This article has stalled out at B-class; it cannot advance to GA-class, let alone A- or FA-class, without being thoroughly sourced with detailed inline citations. If you look at the Featured Articles of the day on the main page &mdash; which are there partly as exemplars, in addition to the various other functions of providing them there &mdash; you'll find that, varying with details of content, it's not uncommon for them to have a footnote at the end of almost every sentence.  That's drastically different from print encyclopedias, but it's a key part of what makes Wikipedia work remarkably well as a radically open project.
 * This is largely independent of the citation style, though of course this particular article currently doesn't use footnotes (which is why, when I restored the nofootnotes template, I took your point and adjusted it to refimprove). It is true that footnotes are more felicitous for the intensive citation on high-quality Wikipedia articles, because very many in-text citations tend to interrupt the flow of the content text.
 * (BTW, you really needn't have addressed your comment specifically to me, especially since the issue you're addressing has been raised by multiple people. I'm just the one who adjusted the cleanup tag in accordance with your remark about it.)  Pi zero (talk) 13:33, 26 October 2008 (UTC)

Pi zero,

A lot of people don’t realize that Wikipedia accepts more than one style of references. As Gnostrat points out above, there are certain editors who breeze through Wikipedia looking for articles with no footnote numbers or even just no footnote numbers in the intro section and slap "no footnotes" labels on articles that in fact contain anywhere from a few to many in-line references. This is what happened here. You have no call saying I removed this tag "reflexively". It stated "this article lacks in-line references" when it has 44 in-line references.

I also cannot agree fully with the statement that "this article needs additional citations for verification". The material referenced is vast (Dolgopolksy’s entire Nostratic Dictionary online, for starters!). Have you actually tried to learn about Nostratic through this article? It leads the reader to a very substantial amount of information.

It would true to say (as I have said before) that two or three sections would benefit from additional references, but this should not be an excuse for mischaracterizing the entire article. If you want to put a tag on the "Nostratic Urheimat" section, for instance, I don’t think anybody would seriously object.

If you care about this article, why don’t you dig out some references for it yourself?

Sincerely, VikSol 06:28, 28 October 2008 (UTC)


 * I'm not certain we're talking about the same thing, but I will tentatively assume for these remarks that you have read my comments above. If you missed them, I commend them to your attention now.
 * Although it is, of course, always possible that the nofootnotes tag was merely someone not noticing the alternative citation format, my original supposition was that someone had observed the general state of the article and merely chose the most commonly used tag; the only point at which under-thinking comes in to that is right at the end with the choice of tag. I was already aware that this article needs just the sort of attention that I supposed was being suggested by the editor who put the tag on it.  Awareness that the tag was a technically inaccurate expression of the need is why I chose a different tag that does not imply footnote style.  Because the lead is one of the parts of the article that needs work of this kind, some tag at the top of the article is certainly needed, though of course I don't claim that refimprove is necessarily the best possible choice; there are an awful lot of different tags to choose from.
 * Figuring out how an article should next be improved &mdash; all articles need improvement, even occasionally FA-class articles &mdash; figuring out what's needed is a time-consuming and necessary task. Doing the work, once identified, is another time-consuming and necessary task, involving different kinds of activity.  (In this case, it is a task that would probably be much easier for some other members of the Wikipedian community than it would be for me; tags enable this division of labor.)  Pi zero (talk) 16:42, 28 October 2008 (UTC)


 * After another couple of days of consideration, I'm proposing an alternative template for the purpose; I'm hoping the internal name of the template won't do any harm after all since its obvious bias toward footnote style isn't reflected in the generated message, and the generated message does explicitly point out that there are inline citations already. Pi zero (talk) 14:44, 30 October 2008 (UTC)


 * I've added some inline notes and removed the tag. More could be added if a page number was specified. Some of the referenced books are multi-volume, spanning on > 1000 pages, and it might be non-trivial finding a particular claim in such forest of etymons and theories. So converting e.g. Georgiy Starostin (2002)-type refs to the format of  ::   would be much appreciated. --Ivan Štambuk (talk) 03:00, 8 December 2008 (UTC)


 * There seems to be a misapprehension about the meaning of the tag. I suspect this misapprehension may be rooted in the fundamental difference between the purpose of citations in academic work (which is meant to be verified by colleagues) versus in Wikipedia (which is meant to be verified, straightforwardly, by anyone who is willing to look up given citations of specific pages in specific reliable sources).  I'll try to illuminate the different Wikipedian approach to inline citations, and the meaning of the tag, with specific reference to this article.  None of this should be taken to in any way belittle the valuable things you've done and further that you've suggested; it's just that the problem being addressed by the tag is much bigger.


 * Most of what is said in this article is not specifically sourced &mdash; and it all should be specifically sourced. For the standard of Wikipedia, it is probably reasonable to omit sourcing of "the sky is blue", but no such exception should be made for "Nostratic is controversial", let alone "The three groups universally accepted among Nostraticists are Indo-European, Uralic, and Altaic".  If every individual statement that needs sourcing were marked with fact, I'd guesstimate more than fifty, and conceivably more than a hundred, of them.  For the sort of attitude that should be taken toward sourcing, see the Featured Article for 13 November of this year: Anti-tobacco movement in Nazi Germany.  Thirteen distinct footnote numbers, some of them used several times, in the lead alone &mdash; which is somewhat more than most subjects require, but Nostratic is controversial (within its sphere), and the Nostratic lead of similar length has two citations (even setting aside their lack of page numbers).  Most of the Nostratic sections that follow are as citation-poor, or even poorer.  The tag isn't indicating a small problem that can be cleaned up easily &mdash; it's indicating the next major phase in the evolution of the article, a phase that could take a year or even several years to move beyond.  (I'd love to see a demonstration that the phase can be completed in much less than a year.)  It's a sign that the article has reached a new level of maturity, where it's ready to begin the serious work of advancing toward GA (and beyond that, FA) status.  Given the controversial nature of the subject of this article, some barriers to GA or FA status will I suspect be content issues that will only be resolvable (and perhaps only be identifiable) once everything in the article is sourced.  (I really would like to someday see "Nostratic languages" appear on the English Wikipedia main page as Today's Featured Article; IMHO the subject is fascinating, and therefore deserves that kind of recognition, and deserves the extraordinary quality of article that will give it that recognition.)  Pi zero (talk) 07:01, 8 December 2008 (UTC)
 * I understand what you're trying to say but so far you've given no (0) evidence of what particular claims in this article need sourcing. Yes, the Nostratic theory itself is controversial, but the claims laid out in this article within that theory are not. Please understand this simple difference. Thus, It is not controversial to say that "The three groups universally accepted among Nostraticists are Indo-European, Uralic, and Altaic". Is there a group of Nostraticists that advocates non-inclusion of e.g. IE or Uralic in Nostratic family? No there is not.
 * FAs are often over-cited are not a good analogy here. Citations should be enforced for disputed and generally implicitly controversial statements, that are likely to be challenged. It is very time-consuming to dig out stupid citations for claims you yourself very much know are true, and should not be a matter of dispute with a fellow Wikipedian, but don't have the necessary books at hand, or don't wanna spend a few hours at the library finding the exact citations from verifiable sources. According to you it is impossible to get rid of the ugly template at the start of the article until the article becomes FA-class, which is ridiculous as there are plenty of articles dealing with controversial theories (not just in linguistics), but don't have this template and really don't need it as even though the theory itself they deal with is at large controversial and disputed, no one disputes the truthness of the claims written in the article within the framework of that theory. Please note the subtle difference between those two. There is really no need to insist on providing citation in e.g. [[Geocentric model] that in that theory the Sun revolves around the Earth (and genocentrism is much more disputed today [being completely wrong] than Nostraticism is in historical linguistics). 33 inline cites are quite enough at this point for this article (which is more than 99.9% WP articles have). Please remove the ugly template and use fact for particular claims that you think need sourcing within the Nostratic theory itself. --Ivan Štambuk (talk) 01:28, 9 December 2008 (UTC)

(outdent) My attempts to explain haven't been working very well; I seem, for example, to have led you down a wrong turn thorough my (by intent, incidental) allusions to controversy and the fact tag. I may be about to fail to explain well enough yet again, because I'm only going to say a few of the things I can think of to say to try to explain further &mdash; but if I said them all, this comment would probably be about as long as the entire preceding content of this thread, which would not be at all useful.


 * The morefootnotes tag here is pointing out the need for work that is prerequisite to achieving GA quality, therefore it is logically necessary that the tag will have disappeared by the time GA quality is achieved.


 * The fact tag is a way of testing specific vulnerable points by attacking them; it carries the implication that if a focused inline cite is not forthcoming in a relatively short time, the tagged passage should be deleted. The morefootnotes tag is a way of promoting better security without attacking.  Since I believe the truth content of the article to be of reasonable quality (but Wikipedia cares about verifiability, not truth), I want to see the article improved instead of decimated, and therefore I don't want to use fact.


 * It took me some time as a registered participant in Wikipedia before I finally understood what Featured Articles are really about. The FA mechanism is the beating heart of the organic structure that causes Wikipedia to succeed in defiance of the expectations of people accustomed to the central-authority model of print encyclopedias.  The whole incredibly complex dynamic of Wikipedia is tied in to this one vital organ.  It's scarcely possible to imagine any misunderstanding more profound than dismissing them as not a good model.


 * It might be objected that using FA articles as examples when discussing morefootnotes here is setting an unnecessarily high standard, since the template will be satisfied before GA status. As a less dizzying height to aspire to, take a look at the lead of The Tales of Beedle the Bard.  That's merely a good article, not featured.  It is about a distinctly uncontroversial topic.  And its lead is much shorter than that of Nostratic, but contains, at the most conservative count, five footnotes (or by a less conservative count, six, eight, or nine).  I would not, BTW, particularly expect that number to change for FAC.


 * Inline cites on Wikipedia are not about controversy, they're about verifiability; as you are clearly aware, controversial statements are to be described rather than asserted. As a rule of thumb (though I can't even guarantee that my own intuitive grasp of inline cites would always match this closely), provide an inline cite for a statement if a reader with some reasonable general knowledge and common sense, but no prior knowledge of the subject of this article, would wonder 'is that really true, or is this just inaccurate information?'  Such a reader wouldn't know from beans if the article said that the Nostratic family doesn't include any languages of North America but does include all the aboriginal languages of Australia, therefore there should be an inline cite on the constituency of the family (not necessarily in the lead, perhaps, unless someone complains, but certainly where it occurs later), so that this hypothetical reader could, if they so chose, look it up on the given pages of the given reliable source.  And sometimes they will look it up, for their own benefit or Wikipedia's.  BTW, that particular cite needs to verify not merely that one Nostraticist says such-and-such are included in the family, but that all Nostraticists say that.

It's good that you realize this is a huge task; it is. The morefootnotes tag doesn't imply a deadline (unlike the fact tag). Pi zero (talk) 00:36, 11 December 2008 (UTC)


 * Addendum, on one small point: It's just come to my attention that, although the guidelines used to say that things said in the lead don't have to be sourced there if they are repeated and sourced later in the article, that exception has been withdrawn; so the basic constituency of the Nostratic family probably should be sourced in both places, after all.  Pi zero (talk) 19:32, 11 December 2008 (UTC)

IE-Kartvelian sound correspondences
Nostratic sound correspondences in the Russian School's version [see Kaiser & Shevoroshkin (1988), Kaiser (1989) map glottalized (ejective) stops to PIE voiceless stops and voiceless stops to traditional PIE voiced ones. This view contradicts the PIE glottalic theory (defended by Bomhard), which regards traditional PIE voiced stops as being originary glottalized ones. To achieve the maximum coherence among these conflicting views, I've interchanged the correspondences of glottalized and voiceless stops in Proto-Nostratic in order to correlate them with the PIE ones. —Preceding unsigned comment added by Talskubilos (talk • contribs) 22:07, 31 March 2010 (UTC)

That means that in the Russian School's flavour, Kartvelian glottalized stops correspond to PIE voiceless ones and Kartvelian voiceless stops corresponds to PIE voiced ones (= glottalized in the PIE glottalic theory), while in Bomhard's view (also defended by Manaster Ramer) is just the opposite. Both views would agree on Kartvelian voiced stops corresponding to PIE voiced ones, but I'm affraid things aren't so simple. For example, Kartvelian *deg- 'to stand' corresponds to PIE *(s)teH2-. This implies that 1) voice constrast was neutralized after PIE mobile *s- and 2) Kartvelian *g is the reflex of an uvular. Talskubilos (talk) 13:28, 10 April 2010 (UTC)

Indo-Semitic theory
Somebody has added a link to [Indo-Semitic] in the references section. As this is a fringe (that is, no mainstream) theory and isn't contemplated on any standard version of Nostratic, I've chosen to remove it. Talskubilos (talk) 09:43, 6 June 2010 (UTC)


 * Talskubilos, the importance of Indo-Semitic to the Nostratic theory is historical:


 * It was, to all appearances, because of his friend and teacher Hermann Möller's theory of a genetic relationship between Indo-European and Semitic that Holger Pedersen included Semitic (and therefore, as it seemed at the time, Hamitic) in his Nostratic grouping, and that, following Pedersen, Vladislav Illich-Svitych did as well.


 * The most important American Nostraticist, Allan Bomhard, cut his teeth on the Indo-Semitic theory in the 1970s and '80s (see article Hermann Möller), before adding in other languages to the comparison, such as Dravidian and Kartvelian. I doubt very much that Bomhard views Indo-Semitic studies as "fringe". Bomhard's early work (and, via it, his later work) draws heavily on the theories of Albert Cuny, a major French linguist of the mid-20th century for whom "Nostratic" and "Indo-Semitic" were synonymous (Cuny recognized that other languages might be related to Indo-European and Semitic, but felt it was important to compare them directly).


 * The three key groupings that Holger Pedersen assembled the Nostratic theory from were Ural-Altaic, Indo-Uralic, and Indo-Semitic. (I am not implying that he necessarily accepted any of these as valid genetic nodes.) Each of these groupings had a long prehistory by the time Pedersen coined the term "Nostratic" in 1903. Readers deserve to be informed of their existence and nature.


 * Admittedly, these things remain to be explained clearly at some point in the article on Nostratic or maybe that on Indo-Semitic, but the subjects belong together.


 * VikSol (talk) 12:46, 7 June 2010 (UTC)

I'm affraid it isn't your opinion (I doubt very much ...) what it matters, but the actual relevancy of that theory on Nostratic studies, which nowadays is nearly zero. According to Wikipedia's policies, a separate article about "Indo-Semitic" would give undue weight to this subject, so I propose you merge your article into a separate section of the Nostratic article. Talskubilos (talk) 19:54, 7 June 2010 (UTC)


 * I haven't expressed my opinion but verifiable facts traceable to easily locatable sources, which you have not felt obliged to check before expressing yours. As I have already tried to get across to you, while an Indo-Semitic node might not have many partisans today, the Indo-Semitic comparison is of tremendous importance for the Nostratic hypothesis. Where do you think all those tables of Indo-European—Afroasiatic sound correspondences in the Nostratic article come from? Also please kindly note that Wikipedia is not just about current scientific theories, it's also about their history: how they came into being. So there is no "undue weight" in having an article about the Indo-Semitic hypothesis. Even assuming it was just a now-forgotten episode in the history of linguistics, the weight of the personalities involved (Johann Christoph Adelung, Karl Richard Lepsius, Rudolf von Raumer, Hermann Möller, etc., etc. would amply justify and indeed require an article on this subject.


 * I will, however, add some remarks to the Indo-Semitic article to clarify the issues and prevent any misunderstanding.
 * VikSol (talk) 00:34, 8 June 2010 (UTC)

I see you're also the author of the Indo-Uralic article. I think the use of the expression "X languages" referring to hypothetical high-level groupings can be misleading, so perhaps "Indo-Uralic hypothesis/theory" and "Indo-Semitic hypothesis" would be more fair expressions. I also see you've improved the Indo-Semitic article, so I think there's no objection in re-linking it to the Nostratic article.

I don't how close you got to Bomhard's thought (I suppose you aren't a sockpuppet), but I'm affraid he's underestimated the weight of Semitic loanwords (e.g. *bhars- 'barley') in IE-Semitic comparisons. It seems that much of the IE lexicon referring to agriculture and Neolithic technologies was borrowed, and one of the sources was precisely Semitic (you can consult a list of proposed Semitic loanwords here: http://paleoglot.blogspot.com/2008/08/list-of-possible-proto-semitic-loanword.html). When I personally communicated this to Bomhard (he had the kindness to send me a PDF copy of his book), he got furious.

A defect inherent to all Nostratic theories is having mistaken substrate loanwords and cross-borrowings among branches as true cognates. The above is an example of how in general Nostraticists are unable to differentiate them. Talskubilos (talk) 09:12, 8 June 2010 (UTC)

Joseph Greenberg
I see a pattern in the article related to Joseph Greenberg. Almost all the paragraphes are "judged" or "criticised" by Joseph Greenberg's opinions as if he is the ultime linguist. There is no argument in the article which is not "assessed" by Joseph Greenberg statements. Why every signle paragraph has a "special section" for the opinions of Joseph Greenberg? As if only after his opinions, tha argument will receive legitimacy. "but Joseph Greenberg says so!"...

I know Wikipedia is run by Jews but this "one-of-us" stance is not the best stance for Wikipedia.--98.199.22.63 (talk) 07:24, 23 January 2012 (UTC)

Core vocabulary
This section is alphabetic from a to k...it seems obvious that it is cut off in the middle. Could it be extended to the end? — Preceding unsigned comment added by 98.27.168.226 (talk • contribs)


 * Thank you for noticing this. It seems this list has been added by an unnamed editor beginning in May 2009, and was intended to be continued - it always contained a note for expansion until it was removed by another editor in February 2010, along with a change of the title from "Lexical words" to "Core vocabulary". I am not sure where the justification for calling it a core vocabulary comes from; the Nostratic Dictionary - Third Edition does not mention a "core vocabulary", and the lexicon at http://www.proto-nostratic.ru, which uses that term, does so apparently only because they copied the list from our article. It does not refer to any core vocabulary in its lexicon entries.


 * In conclusion, I would suggest cutting this list down to only a handful of examples, replacing the word "core" in the headline with "example", and referring for anyone who wants to read more to the Nostratic Dictionary. It will also be necessary to explain the notation, so we could maybe select those entries that use the most common symbols. &mdash; Sebastian 07:52, 19 November 2014 (UTC)

I remember looking at this wordlist almost 25 years ago and Im sure it will be easy to find on the Internet. But Im not sure we should be sharing the wordlist at all ... to start with, does anyone here know how to pronounce /∇/? Is it a particular vowel, or a symbol for an unknown vowel? There's also at least three rhotics and two laterals but we have no key to tell us which is which. I would rather just link to the wordlist, ideally to a copy of it that appears on the website of one of the researchers, but if not, a link like this will do. — Soap — 07:25, 25 April 2020 (UTC)
 * Though perhaps a better link can be found ... that PDF contains the wordlist but also a great deal else; it would be better to link to something smaller. Perhaps it is difficult to find an exact match for the wordlist because the copy we used is now at least 24 years old (and I think much older) and Bomhard has greatly expanded it since then. In which case, I say that's another reason for us not to post it .... but it seems he has expanded the list so much that it's now 400 pages long. I'm hoping somewhere there's just a list and not a list with a list of all the descendant forms for each reconstructed root, since it'd be likely of more interest to casual readers than the full thing. — Soap — 16:35, 25 April 2020 (UTC)
 * Tables of such extent do not belong in a WP article and often border on WP:copyvios, so I support your removal. A small, well-sourced illustrative selecion will always do, and readers who want to know more about it can get the whole straight from the source. –Austronesier (talk) 09:51, 27 April 2020 (UTC)
 * Thank you. Im still looking to find a suitable page we can link to, ... I managed to find an updated version of the dictionary, and it has a key explaining all of the weird symbols.  Only problem is, umm .... well, it's 3100 pages long, and a  lot of people, especially on mobile, are going to be literally unable to read that. In fact the key is on page 2693.  So Im still looking.  If I cant find anything more concise, though, the link is here and it seems to be direct download only, no option to read it in the browser.  — Soap — 22:31, 27 April 2020 (UTC)
 * Okay, the first link I found was just a blind link to the PDF. Now that I've found that there is in fact a web interface, I looked at our article again and noticed that the link is already there.  So Im going to say this is probably as good as we're going to get, but if I find something more concise, say, a list of roots *without* all the cognates, I will still add that link because I think it would be the most reader-friendly way to present the data.  — Soap — 22:39, 27 April 2020 (UTC)

Nostratica
I don't know if it could be useful but I found an archived copy of Nostratica on Internet Archive. https://web.archive.org/web/20120311211525/http://www.nostratic.ru/index.php?lang=en Seems then the English version is off but the Russian version is still online. --Sd-100 (talk) 19:35, 13 November 2015 (UTC)

Not science
Has to be said and if you are a researcher you should be angry at universities and professors who promoted this, not myself for pointing out the obvious. Arutun (talk) 15:08, 9 July 2016 (UTC)


 * @Arutun: And that statement by you is based on what reliable sources? —C.Fred (talk) 18:29, 9 July 2016 (UTC)

The table of sound correspondences - OR
Is it a common practice in Wikipedia to modify the presented data according to private views of Wikipedia authors? Isn’t such an act a violate of neutral point of view? Isn’t it an original research of the Wiki authors?

If Wikipedia presents facts, not personal invetions of its redactors, such a practice should be strongly prohibited. But what do we read in the article?

The following table is compiled from data given by Kaiser and Shevoroshkin (1988) and Starostin. - followed by a link to the source. It sounds great! But what does it follow? [...] correlate Nostratic voiceless and glottalized stops with PIE ones, so this is done in the table.

Either the table presents the view of Kaiser, Sheveroshkin and Starostin and is based on the cited sources, or it follows Bomhard’s view. Please decide! If it is based on the Russian authors, please STOP MODIFYING their views!!! This is ORIGINAL RESEARCH, and I have tagged it with OR.

Bomhard’s sound correspondences are NOT any “upgrade” of the Moscow School (by the way: “Muscovite School” is a derogatory term and should not be used at all, especially in encyclopaedias). It is just a different point of view, based on the glottalic hypothesis which is not proved, criticised, and which has little followers now. This is why Bomhard’s reconstuction is not “modern” - it is implausible. The cross-correspondence of Nostratic and Kartvelian ejectives is the most obvious proof (Nostratic tʼ = Kartvelian t, while Nostratic t = Kartvelian tʼ). In his newest publication Bomhard himself does not respond the strong critique of the glottalic hypothesis, and instead of answering to serious arguments he tries to discuss old and less important counterarguments as if he did not know the bibliography of the subject (including Barrack’s work). Having considering it, I do not think that presenting his glottalic view as the best one, the most modern one or as an upgrade to the classical view - is nothing but making an advertisement of his view, contrary to common practice among scholars. It should never happen in an encyclopaedia! It is a real example of OR.

Please remove all changes to Kaiser, Sheveroshkin and Starostin view, and present it as is, without all this glottalic stuff. Wikipedia should present scholars’ views and not private judgement of Bomhard’s fans. And if you love so much the glottalic hypothesis, why not to make another table, based on Bomhard’s implausible view?

You suggest to readers that Bomhard has made an improvement to Starostin and others. THIS IS NOT TRUE. Why do you present such a view? Is it based on some independent judgement of authors who do not follow either Starostin or Bomhard?

Please do not remove my OR tag as long as the table presents the true, original view of Kaiser, Sheveroshkin and Starostin, exactly as it is it presented in the cited work.

178.235.146.5 (talk) 08:55, 24 August 2017 (UTC)

"Controversial" and "widely rejected"
An "edit war" has been happening for years on this article regarding the terms "controversial" and "widely rejected". As I said in the edsum, most linguists are indeed agnostic because evidence are insufficient to conclude that these languages are either related or unrelated. If the Nostradic family is described as "controversial" and "widely rejected" on Wikipedia, then it should be the same for most language families. As Kallio & Koivulehto put it in "Beyond Proto-Indo-European" (2017): In general, Nostratic studies have failed to meet the same methodological standards as Indo-European studies, but then again so have most non-Indo-European studies. Alcaios (talk) 13:53, 26 December 2020 (UTC)

I have seen in the edsum that invoked WP:BLUESKY to write that "The hypothesis is controversial and has varying degrees of acceptance amongst linguists worldwide with most rejecting Nostratic". (an hypothesis proposing Middle Eastern origins for Tungusic, Mongolian, etc. is not accepted/endorsed by the mainstream). The whole issue is the definition of "Nostradic". If it refers to a slightly extended version of Indo-Uralic, then it's not "widely rejected", but if it's about including a bunch of languages based on a long-range phonetic comparison that fails basic scientific standards, then I can only concur with such a description. Alcaios (talk) 14:09, 26 December 2020 (UTC)


 * Hi . The latter is literally the definition of Nostratic. Neither Indo-Uralic nor its slightly extended version has to be relevant to Nostratic and Nostratic is certainly not limited to Indo-Uralic. See the relevant discussion here. Puduḫepa 14:59, 18 June 2021 (UTC)

Recent edits
There have been some interesting edits to this article over the last couple of months...  Tewdar  07:16, 19 April 2022 (UTC)
 * Yes, it is the ultimate sock/CIR fest. Let's restore the August 2021 version, what do you think? –Austronesier (talk) 08:04, 19 April 2022 (UTC)


 * Hello and thank you for pinging me and I'm very sorry for what I said and it made me feel sad and I won't do it again if I can help it and it's nice to talk to you again 😁 - yes, probably not a bad idea... also check out the Jomon people/period article and anything recently linking to Yana RHS...  Tewdar   08:17, 19 April 2022 (UTC)
 * Shouldn't I be the one, since the humor of my rebuttal was way nastier than yours 😁...? Anyway, yes please let's TNT the genetics section of Jomon people, it's a total clusterfuck, a mess and everything is cited wrong—just a big insult to our readers. The perpetrator(s) is/are here; the SPI history is more complex than the f4-statistics of East Eurasians. –Austronesier (talk) 08:37, 19 April 2022 (UTC)
 * Done for now. There is still a lot work to do, however. I find this page overly promotional: the sections with "evidence" are too detailed and visually dominate over the criticism by mainstream scholars. –Austronesier (talk) 08:44, 19 April 2022 (UTC)
 * Your description of my comment as an ancient shit was harsh but not entirely inaccurate 😳. The "sockpuppet reverting a sockpuppet who reverted a sockpuppet who..." on those articles are hilariously transparent.
 * Reversion looks better. Of the many remaining problems, the assertion Proto-Nostratic would have been spoken between 15,000 and 12,000 BCE, in the Epipaleolithic period definitely requires attribution...  Tewdar   08:50, 19 April 2022 (UTC)
 * And while we're using Stetson-Harrison methods, I reckon 40,000BP is a much more likely date for Proto-Nostratic.  Tewdar   09:11, 19 April 2022 (UTC)
 * As long as it is attributed to Bomhard, it is fine. Bomhard is a reliable source for everything that Bomhard says. –Austronesier (talk) 09:24, 19 April 2022 (UTC)