[In answer to Q on "origin of language families"]

Creation vs. Evolution : Reviewing a paragraph from If You Believed Moses · Φιλολoγικά/Philologica : [In answer to Q on "origin of language families"]

I'll start with a quote:

It is seldom emphasized that similarities between language families are themselves susceptible to the same three explanations. That we so seldom see mention of this corollary principle is largely because twentieth-century historical linguistics has been laboring under the delusion that language families like Indo-European share no cognates with other families, thus offering nothing to compare. At this level, it is alleged, similarities simply do not exist.

What is striking is that this position—for which considerable evidence to the contrary existed already at the start of this century (Trombetti 1905) and which on a priori grounds seems most unlikely (Ruhlen 1988a)—came to be almost universally accepted by linguists, most of whom have never investigated the question themselves.

From: 14 Global Etymologies
John D. Bengtson and Merritt Ruhlen
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.476.9508&rep=rep1&type=pdf

The three explanations he is talking of are : common origin, borrowing, convergence.

Now, there is a big difference in how diverse linguists classify human languages, if you go down below the signature, you will first in figure 1 see 244 language families on which wikipedians have no doubt of common origin.

In figure 2. you will partly see groups contentiously pretending common origin for more than one of them, and partly groups that consist of one or two single languages, which could be isolates. This lands us with probably at least anything from 208 to 273 language families.

In quote 3, you will find out how Merritt Ruhlen and John D. Bengtson classify only 32 language families, of which they consider some related to each other. Amerind and Khoisan are two families not found in the wiki fact space on "primary language families" which you can usually access from each of them, like from "Algic" - a family comprising Wiyot † (extinct), Yurok and Algonquian. In Ruhlen-Bengtson, you don't find Algic or Arawakan, they are both Amerind subfamilies. Pre-Columbian America has Eskimo-Aleut, Na-Dene and Amerind and that's it. And in Central and South America, only Amerind.

By Khoisan, Ruhlen-Bengtson - I'll add, already Greenberg, whom they are folllowing - dispense with Hadza, Sandawe, Khoe, Tuu and Kx'a. All of which are found separately in figure 1.

I submit as my hunch, and that it is a hunch, that the 32 groups are valid. I also submit, since they are fewer than 72, which is the number of original languages after Babel, that some groups have converged through mutual borrowing. That is, they are not families in the strict sense.

Note well, there is one explanation which Ruhlen and Bengtson left out. Miraculous "instant family" structures. Suppose that two groups of men came to adopt Quenya and Sindarin of Tolkien. In his "legendarium" or "mythology" these are related by a common origin some millennia before they emerge as clearly different languages. Proto-Eldarin cannot be used as someone's adopted language, Tolkien only sketched it out to have sth to derive Quenya and Sindarin from. This means that both languages in actual history were constructed as artistic constructed languages between, say, 1915 and 1950 (I don't think Tolkien made any revisions after that, major ones, since he used both as background enhancers in Lord of the Rings and any even slight revision after that would still involve the text scraps in Lord of the Rings). So, the two groups would be adopting languages with a "family origin" merely sketched out and an "instant family" arising between 1915 and 1950.

What Tolkien could do with fictitious races, what two groups of Tolkien fans could do, God also can have done.

This means, some languages arising immediately at Babel could have had a common proto-language never spoken on Earth, but extant only in God's imagination.

However, this is far from the only way in which Indo-European or Semitic, even Afro-Asiatic, languages could have become language groups.

Common origin is possible, but for Afro-Asiatic, it staggers belief that Mizraim, Assur, Joctan, Kush, would all adopt the language of Peleg and their descendants have time to diversify them as much as ... Old Egyptian, Akkadian, Arabic, Ethiopian, all the while Canaaneans and Syrians adopted Hebrew and Aramaic along with descendants of Peleg. The 72 languages were extant in 2562 BC (if my reconstruction of chronology within Roman Martyrology context is correct for Peleg), which is carbon dated as 8600 BC (if my identification with Göbekli Tepe as Babel is correct) and we have Old Egyptian and Akkadian texts carbon dated (directly or indirectly, like organic material from same shelves or same tombs) to 2600 BC, which would arguably be the times of Joseph, around 1700 BC. Do 700 years really explain such a diversity?

So, this leaves two more possibilities : convergences through mutual borrowing and auxiliary languages.

In a scenario of auxiliary languages, I'd consider Sumerian was the first one. Note, it has been classified diversely as "isolate", as "Uralic" and - by Ruhlen-Bengtson - as Dené-Caucasian, if I recall correctly, related to Athapascan Apache and to Basque and to Chinese and Tibetan and to Circassian, but not to Georgian. I'd go one further, it can have been the first auxiliary language, and been part template for other ones, both Ural-Altaic and North Caucasian. And even for Bantu. Or Dravidian. Depending on whether the "blackheads" that Sargon chased away, the indigenous speakers of Sumerian, went to Africa or India.

Of course, a natural language can also serve as an auxiliary language, as English does for me, since I am not a native English speaker.

I'd venture that following auxiliary languages leading up to Indo-European were Hattic and Tyrsenian - which Alinei classifies as Old Hungarian. Then Indo-European, whether as a natural language (like Hittite) or as an articificial auxiliary language, tried to bridge Uralic and Afro-Asiatic traits. As Merritt Ruhlen and John D. Bengtson wrote, Indo-European can have outside affinities. Personal verb endings are close to Uralic, specifically Finnish.

eleipon		lähden
eleipes		lähdet
eleipe		lähdee
eleipomen		lähdemme
eleipete		lähdette
eleipon		lähdevät

Ablaut system and some aspects of how tense stems are formed from a common verb root seem loosely based on Afro-Asiatic, perhaps Hebrew, but note well, loosely. Semitic have basic short vowels a, i and u, along with absence of vowel, while Indo-European has (in Greek or even more reconstructed proto-form) e and o and null. Both languages insert or take away diverse extra consonants in forming diverse stems, like present stems with an extra -n- or -yod or aorist stems with an extra -s, taking the Indo-European side, which I know.

And in tenses, "classic" Indo-European (Italic, Greek, Sanskrit) combine the Uralic opposition of present and past with the Semitic opposition of ongoing or finished. In sense these overlap, and therefore a language can get along with only one of them, but Indo-European has both.

In nouns, feminines in -a (originally according to one theory -e+laryngeal 2, -eH₂) seem close to Semitic feminines in -at, which in some positions is -ah. The Arabic and Akkadian case ending vowels seem echoed in at least some Indo-European forms, like -u in Latin and Greek -us/-os, -i in Old Irish and Latin Genitive -i, in Latin and Greek Dative -i and perhaps even -a in Greek third declinsion accusative -a. And if you reply this -a is really -m, well, the Arabic case endings also come with Nunation -un, -in, -an. Again, very loosely based, not at all a clear identity.

And on the other side, some similarities in the Afro-Asiatic group - notice, I don't say family, since I don't believe in a common origin - would have been there due to Hebrews being a bit everywhere, in their attempst to shirk the building of Babel - and successful ones, since God didn't punish them by language loss and new imposed language. Both in the way that some took up Hebrew instead of their own language (notably in Canaan), but also in the way that some used Hebrew as lingua franca with other neighbours and they both started adopting Hebrew features, like feminines in -at/-ah.

The other possibility for Indo-European is, languages started out on a fairly equal footing as isolates around Aegean, Iavan's language in Greece, Lud's language in Troy (more Luwian than Lydian or Phrygian, possibly), Gomerite further East in Cappadokia, notably Kanesh, whence Nesili (the indigenous term for Hittite), while they were less in mutual contact with the more or less Uralic Hattic, probable language of Canaan's son Heth. Unless, as mentioned, based on Sumerian, with some looseness. If then a certain French linguist was right and plucks up his courage again to reclassify the unclassified Cypro-Minoan language of the Caphthorim as an Aryan one, this is a fairly big hunch of Indo-European languages in a fairly small area.

Now, I have a test to do on vocabulary before I have what I consider as scientific confidence in this model, in this hunch. So long as the test is not done, it is a hunch. Nevertheless, even a hunch is a valid alternative to a theory which is based on ignoring that hunch.

But for language families or language groups outside those mentioned, I do not even have a hunch. Except mathematics, 244 "language families" are clearly more than the original ones and 32 are clearly fewer than them. I was very tempted to make an angry reply like the Genie in Aladdin on being asked for the egg of the bird Roc - that's about what you can expect from a linguist if asked to outline (as fact, not as alternative hunches) how the language families arose - unless they are of Ruhlen-Bengtson's school and trace all 32 back to "proto-world" through 27 roots. One of them could well be onomatopoic since M and GLGL sound like a baby suckling and swallowing "milk" (Indo-European meaning), from his mother's "breast" (meaning in some other language family) into his "gorge" (meaning in yet another family). While this does not rule out a common origin, it certainly poses a problem in proving Merritt Ruhlen's and John D. Bengtson's theory of proto-world by that root, since onomatopoeia is one of the areas where "convergence" rather than common origin is highly featured as a possibility. Also, Biblically, I don't believe in "common origin" of this type.

Edenics? I haven't read the paper, do not take this as a review of it, I do not want the paper to be blocked because I'd give an adverse review, I don't believe in peer review before publishing, you do with your conscience as to that paper, but if God used Hebrew as ultimate template for languages, it is possible in some cases He used as roundabout coding for it as "hala" for "penyo" or "penyo" for "hala". Perhaps one could loook up Jalaa and Plateau Penutian and see if they have words glossing Hebrew ones in the jala=peño way. I don't expect languages to have only a slight divergence from Hebrew, and I don't think any divergence from Hebrew is linguistically speaking garbling. The "confusion" or "garbling" was on the level of personal linguistic competence. Not of resulting one. Each one who had his language changed would have been fine with learning the new one if he had kept Hebrew too - except, then they would have executed Nimrod's orders, and probably trying to fuel a three step rocket with Uranium would have resulted more in like a mushroom cloud than in man walking on the Moon. So, God made cooperation impossible in the last moment.

Hans Georg Lundahl
Nanterre UL
Vigil of St. Thomas
20.XII.2018

(Typo corrected on Boxing Day.)

Also on Boxing Day : Tolkien tinkered on with all aspects of Quenya and Sindarin up to his death, except where some aspect was already fixated through Lord of the Rings, H/T to Stephen St. John for this correction./HGL

Figure 1. Generally acknowledged families.
Abinomn · Afro-Asiatic · Ainu · Alacalufan · Algic · Alsea · Arai–Samaia · Arafundi–Piawi · Arawakan · Arauan · Araucanian · Arnhem/Macro-Gunwinyguan · Arutani–Sape · Austroasiatic · Austronesian · Aymaran A 16 Baining · Bangime · Banzsl · Barbacoan · Basque · Binanderean–Goilalan · Boran · Border · Borôroan · Bulaka River · Bunuban · Burushaski B 12 28 Caddoan · Cahuapanan · Camsá · Candoshi · Cariban · Catacaoan · Central Solomons · Chapacuran · Charruan · Chibchan · Chimakuan · Chimane · Chimariko · Chimbu–Wahgi · Chinookan · Chiquitano · Choco · Chonan · Chukotko-Kamchatkan · Chumashan · Comecrudan · Coosan · Cuitlatec C 23 51 Darwin River · Doso–Turumsa · Dravidian D 3 54 East Geelvink Bay · East Strickland · Eastern Daly · Eastern Tasmanian · Eskimo–Aleut · Elamite · Eleman · Engan E 8 62 Fas · Fulniô F 2 64 Garawan · Giimbiyu · Great Andamanese · Guaicuruan · Guajiboan · Guató G 6 70 Hadza · Haida · Harákmbut–Katukinan · Hattic · Hmong–Mien · Hodï/Joti · Huaorani/Waorani · Huave · Hurro-Urartian H 9 79 Indo-European · Iroquoian · Itonama · Iwaidjan I 4 83 Jalaa · Japonic · Jarrakan · Jê/Gê · Jicaquean · Jirajaran · Jivaroan J 7 90 Kalapuyan · Karajá · Kariri · Kartvelian · Karuk · Katembri–Taruma · Kaure–Kosare · Keresan · Khoe–Kwadi · Kiwaian · Kol · Koreanic · Kra–Dai · Krenak · Kunza · Kuot · Kusunda · Kutenai · Kutubuan · Kwomtari · Kx'a K 21 111 Lakes Plain · Leco · Lencan · Lower Mamberamo · Lower Sepik L 5 116		Madang · Maiduan · Mairasi · Maku-Auari of Roraima · Malak-Malak · Marrgu · Mascoian · Matacoan · Maxakalian · Mayan · Meemul · Mirndi · Misumalpan · Mixe–Zoque · Mongolic · Movima · Mura-Pirahã · Muskogean M 18 134 Na-Dene · Nadahup · Nambikwaran · Niger–Congo · Nihali · Nivkh · North Bougainville · Northeast Caucasian · Northeastern Tasmanian · Northern Tasmanian · Northwest Caucasian · Nyulnyulan N 12 146 Ofayé · Ongan · Oto-Manguean · Otomákoan O 4 150 Palaihnihan · Pama–Nyungan · Pano-Tacanan · Pauwasi · Pawaia · Peba–Yaguan · Plateau Penutian · Piaroa–Saliban · Pomoan · Porome · Puinave · Purian P 12 162 Quechuan Q 1 163 Ramu R 1 164 Salishan · Sandawe · Senagi · Sentani · Sepik · Seri · Shastan · Shuwa gozoku · Shǒuyǔ · Sino-Tibetan · Siouan · Siuslaw · Skou · South Bougainville · Southern Daly · Sumerian S 16 180 Takelma · Tambora · Tangkic · Tanoan · Tarascan/Purépecha · Teberan · Tequistlatecan · Ticuna–Yuri · Timotean · Timucua · Tiniguan · Tiwi · Tor–Kwerba–Nimboran · Torricelli · Totonacan · Trans-Fly · Trans–New Guinea · Trumai · Tsimshianic · Tucanoan · Tungusic · Tupian · Turama–Kikorian · Turkic · Tyrsenian · Tziij T 26 206 Uralic · Urarina · Uru–Chipaya · Utian · Uto-Aztecan U 5 211 Wagaydyic · Wagiman · Waikuri · Wakashan · Warao · Washo · Wintuan · Wiru · Witotoan · West Papuan · Western Daly · Western Tasmanian · Worrorran W 13 224 Xincan X 1 225 Yabutian · Yam · Yamana · Yana · Yangmanic (Wardaman) · Yanomaman · Yawa · Yeniseian · Yokutsan · Yuat · Yuchi · Yukaghir · Yukian · Yuman–Cochimí · Yuracaré Y 15 240 Zamucoan · Zaparoan · Zhōngguó · Zuni Z 4 244

The ones marked with question marks would include both language isolates, adding one family, and disputed superfamilies, ridding of families. Some are families other than the ones above, and so add. "Ijaw" is considered a distinct group within "Niger Congo", so, Ijaw either is its own family or a Niger Congo group. I am considering Ijaw as + 1 (adding one group to above) and a compulsory one unless you agree with Niger Congo. Similarily, Songhay is either its own family or you agree with Nilo-Saharan. Vasconic? = Basque + extinct languages like Aquitanian and Iberic. Northeast New Guinea languages, I could not study, has no article.

Figure 2. To add or detract.
Anêm? + 1 · Aikanã? + 1 · Andoque? + 1 · Ata? + 1 (+4) Chimuan? + 1 · Cofán? + 1 (+6) Digaro? + 1 (+7) Esmeralda–Yaruro? + 1 (+8) Hibito–Cholón? + 1 · Hrusish? + 1 (+10) Ijaw? + 1 · Irantxe? + 1 (+12) Kadu? + 1 · Kho-Bwa? + 1 (+14) Laal? + 1 · Lule–Vilela? + 1 (+16) Miju? + 1 (+17) Nilo-Saharan? + 1 · Nukak? + 1 (+19) Shabo? + 1 · Siangic? + 1 · Songhay? + 1 · Sulka? + 1 (+23) Taiap? + 1 · Tequiraca–Canichana? + 1 to + 2 · Totozoquean? + 1 · Tuu Mande? + 1 (+27 / +28) Ubangian? + 1 (+28 / +29)		Altaic? - 3 to - 5 · Austronesian–Ongan? - 1 · Austro-Tai? - 1 (-5 / -7) Bora–Witoto languages? - 1 (-6 / -8) Dené–Yeniseian? - 1 (-7 / -9) Hokan? - 7 (at least) (-14 / -16) Macro-Jê? - 5 (-19 / -21) North Papuan? - 4 (-23 / -25) Papuan Gulf? - 4 · Penutian? - 7 (-34 / -36) Sino-Austronesian? - 2 (-36 / -38)

244 + 2 = 246
244 + 29 = 273
246 - 38 = 208
273 - 36 = 237

3. Quote Bengtson Ruhlen:

Specifically, we will be comparing items in the following 32 taxa, each of which we believe is a genetically valid group at some level of the classification: Khoisan, Niger-Congo, Kordofanian, Nilo-Saharan, Afro-Asiatic, Kartvelian, Indo-European, Uralic, Dravidian, Turkic, Mongolian, Tungus, Korean, Japanese-Ryukyuan, Ainu, Gilyak, Chukchi-Kamchatkan, Eskimo-Aleut, Caucasian, Basque, Burushaski, Yeniseian, Sino-Tibetan, Na-Dene, Indo-Pacific, Australian, Nahali, Austroasiatic, Miao-Yao, Daic (= Kadai), Austronesian, and Amerind.

14 Global Etymologies
John D. Bengtson and Merritt Ruhlen