8 Nisan 2013 Pazartesi


1. Introduction 

The Turanian language family that plays a very important role in Hungarian linguistics as well as history, comprises the Finno-Ugric (or, together with the Samoyed languages, the Uralic) language family (FU) on the one side and the Altaic language family (A) on the other side. While the existence of an Altaic language family has always been doubtful, the existence of a Finno-Ugric or Uralic family has become controversial only very recently. 

This latter fact is the more astonishing, because Hungarian –besides Finnish the most prominent member of this little family – was originally not considered to have genetical relationships to Finnish and the languages close to Finnish, although a first attempt to group Hungarian and Finnish together goes back to the 18th century (Sajnovics). In this small contribution, I will show the language data necessary to prove that both FU and A cannot be considered to be linguistic families. At the end of this contribution will stand the question, if the existence of the Turanian language family is thus a phantom or not. 

2. The alleged Finn-Ugric (Uralic) language family 

Even in traditional Finno-Ugric or Uralic departments, the existence of a Finno-Ugric or Uralic language family has been doubted for a couple of years (cf. Marcantonio 2002). But while more and more scholars are convinced, that the former Uralic language family is nothing but a Sprachbund, most Finno-Ugrists defend their position as representatives of a language family. 

In this chapter, using the 100 words Swadesh-list and considering 10 Finno-Ugric and 2 Samoyed languages, it will be shown that neither the one nor the other assumption is justified. The politically motivated construction of a Finno-Ugric language family in the 18th century shows such a small basis of common words that would put back Proto-Finno-Ugric or Proto-Uralic long before 10’000 B.C. and therefore leads itself ad absurdum. On the other side, it will be shown that the theory of the Sumerian 
origin of Hungarian, commonly accepted before the invention of the Finno-Ugric and Uralic language families (cf. Érdy 1974), is acceptable also from a language-statistical point. 

The 12 Uralic as well as the Sumerian and Akkadian Swadesh lists were compiled from dictionaries. Unfortunately, the Ostyak dictionary of Karjalainen (1948) and the Mordwin dictionary of Paasonen (1990-96) were not available to me, because the Library of Congress does not borrow reference works. 

From the living languages only the Finnic and Estonian lists could be controlled by native speakers in the spring of 2003 in the Institute of Uralistics of the University of Szombathely (Hungary). The Hungarian list was compiled by the present author according to his native speaker’s proficiency. 

Since, as it is known, the Swadesh list was and is still discussed controversially, I would like to mention here only a few recent cases, in which the list could be applied successfully, i.e. where the calculations that follow from the list are matching with the chronological data of non-statistical linguistics: Elbert (1953) for Polynesian languages; Rabin (1975) for Semitic languages, Blažek for Sumerian (including 
Emesal), Akkadian, Elamitic, Kassitic, Hurrian, Urartian and Hattic; Forster, Tóth and Bandelt (1998) for 17 Retoromance/Ladinic dialects and recently Forster and Tóth (2003) for Celtic languages....

Hungarian = Sumerian: 91% 
Hungarian = Akkadian: 27% (2%) ...

3. The alleged Altaic language family 

Unlike the situation in the FU languages, in the A languages, most linguists assume that Turkish, Mongolian and Manchu are not genetically related, but have relationships due to early mutual borrowings. This is the standpoint of leading researchers like Sir Gerald Clauson and Gerhard Doerfer, while other capacities like Nicholas Poppe and Karl H. Menges not only accept a Urverwandtschaft, but consider also Korean and Japanese to be members of the alleged A family. However, in order to decide if Turkish, Mongolian and Manchu – the „hard-core“ A members – are related or not, we will use, like in chapter 2, the 100 words Swadesh list. Since this has already be done by Clauson (1969), with a slightly different list in the alphabetic order of the English words that does, however, not influence our results, we can trust his list that we reproduce in the following. Borrowings are italicized according to Clauson.

- Turkish and Mongolian have maximally 16 words in common, i.e. share 16% of the basic vocabulary 
- Turkish and Manchu have 0 words in common, i.e.are completely unrelated 
- Mongolian and Manchu share a number of common Chinese borrowing (Clauson 1969, p. 22) and maximally 22 common words, i.e. 22% of their basic vocabulary 

This goes with Doerfer’s statement: „Meines Erachtens nun haben diejenigen Forscher recht, die eine Urverwandtschaft der sog. altaischen Sprachen ablehnen“ („Now according to me, these researchers are right who denie an Urverwandtschaft of the so-called Altaic languages“) (Doerfer 1963, p. 51). Doerfer goes even further when he denies that any words between Turkish, Mongolian and Manchu are 
related – according to him, these are merely borrowings (Doerfer 1963, p. 53). 

Therefore, we would have 0% and total genetical unrelatedness between all three A languages. It is thus anyway superfluous to prove that the further languages, Korean and Japanese, that some researchers wanted to show to be genetically related to the A languages, cannot be related to these languages, since these are not even related to one another (cf. Doerfer 1974) – after all, even given the above percentages of 16% and 22% - these are – like amongst the FU languages – much to low in order to prove a genetical relationship. 

Our conclusion is simple: neither the FU nor the A languages form a language family, since they can in no way be genetically related. This thus, of course, not exclude mutual borrowings or the existence of Sprachbünden between these languages. 

A next question is, how the respective relationships look in other language families. Are there even real language families or not, and if yes, how do the respective percentages of their relationships look like? 

From an important article by Samuel H. Elbert, we get the following percentages for the MalayoPolynesian (MP) languages: 

Futunan vs. other MP languages: 63.9% 
Uvean vs. other MP languages: 61.6% 
Niue vs. other MP languages: 55.1% 
Tongan vs. other MP languages: 53.7% 
Tikopian vs. other MP languages: 66.9% 
Ellice Island vs. other MP languages: 64.9% 
Samoan vs. other MP languages: 58.3% 
Sikiana vs. other MP languages: 56.8% 
Fila vs. other MP languages: 55.8% 
Ongtong Java vs. other MP languages: 53.5 % 
Nukuoro vs. other MP languages: 49.2% 
Kapingamarangi vs. other MP languages: 49.5% 
Easter Island vs. other MP languages: 63.1% 
Marquesan vs. other MP languages: 71.2% 
Mangarevan vs. other MP languages: 69.2% 
Rarotongan vs. other MP languages: 82.5% 
Tuamotuan vs. other MP languages: 79.6% 
Maori vs. other MP languages: 72.0% 
Hawaiian vs. other MP languages: 76% 

One gets a total average of 63.3%, whereby the maximal value is 86% (Uvean = Tongan) and the minimal value is 44% (Nukuoro = Mangarevan). The MP languages, therefore, do form a language family – unlike FU and A. 

According to an earlier study of mine, English and German share 70.0% of their vocabulary and the Slavonic languages amongst themselves according to a study by Fodor (1961) even almost 90%. 

Moreover, Kroeber and Chrétien have shown (1937, 1938) that the percentage of relationship between the IE languages is higher than 70%, so that we can conclude that the IE languages, too, form a language family – unlike FU and A. 

4. Conclusions 

In his influenceful book „The science of language“, in 1891, F. Max Müller, comprised amongst the „Turanian language family“ the following language families: Samoyedic, Tungusic, Mongolic, Turkic, Finnic (incl. Hungarian), Taic, Malaic, Gangetic, Lohitic, Munda, Tamulic. The Turanian languages thus comprise, according to Müller, not only the FU and A languages, but a geographically vastly extended group of languages reaching, roughly speaking, from the Ice Sea via China to the South Seas. 

In my „Etymological Dictionary of Hungarian“ (EDH), I have shown that following the identity of at least 1042 basic Sumerian words with their respective Hungarian cognates according to Gostony (1975), all of Müller‘s language families are genetically related with one another – up to a certain degree that goes from 61% down to 3%: 

Hungarian (100%) > Chinese (61%) > IE (58.3%) > Turkish (55%) > Tibeto-Burman (ca. 50%) > Dravidian (36%) > Munda (33%) = Etruscan (ca. 33%) > “FU” languages (31.9%) > Japanese (23%) > Mayan (11%) > Bantu (8%) > Caucasian (7%) > Austronesian (incl. Mon Khmer, Australian and Tasmanian) (3%). 

In default of reliable historical dictionaries using strict sound-laws, I did not treat other members of Müller’s language families (f. ex. Mongolian and Manchu as representatives of the A languages) that form the “macro-family” of the Turanian languages. Although Müller’s method – the reconstruction of proto-languages – and my method used in EDH – the reconstruction of sound-laws given a common ancestor language – are not the same, both methods lead to the same languages and language “families”, but the number of languages and language families shown in EDH is even larger than Müller’s. It is thus important to state that the languages of such a vast area from Europe to Tasmania, as presented in EDH, are only related to one another because they share their common ancestor, Sumerian, up to a certain degree. 

However, the languages in the above downward hierarchy, starting with Japanese, can hardly be considered to be genetically related to Sumerian, but even their low percentages show that Sumerian did not just die out when it was replaced, in the 17th century B.C., by Akkadian, but influenced most other languages in the world in an extremely vast area. 

We can thus answer the question in our title: No, the Turanian language family is not a phantom, it exists – but only because all of it’s languages are related to their common ancestor language Sumerian. 

As shown in EDH, the Turanian language family is even much larger than it was assumed in the 19th century, when the Turanian concept was brought up. 

As we have seen in chapter 3, Turkish and Manchu have no common word in the 100 words Swadesh list. From that one concludes that these languages are not related in the Altaic language family. 

However, this does not exclude that they are related in a bigger macro-family. 

For example, Latin cuprum and Greek kúprion are related in the IE language family, i.e. both cuprum and kúprion have a common ancestor word in the IE proto-language, but German Kupfer and English copper don’t, since they are borrowings from Latin cuprum. Thus, Kupfer and copper are not related either with one another in the Germanic language family, but they are in the IE language family. 

The same is true for Latin avunculus on the one side and German Onkel, English uncle on the other side. Latin avunculus developped to French onlce, and from here the German and English words have been borrowed, i.e. Onkel and uncle are not related in the Germanic language family, but they are related in the bigger IE family It is therefore possible using the simple Swadesh list and classical historical linguistics to prove that the FU and the A languages do not form language families. But while it is impossible with simple reconstruction of proto-languages to learn how far an ancestor language like Sumerian spread out, it is possible by comparing this assumed ancestor language with possible successor languages, using strict sound-equations and the best available dictionaries. 

Exactly the same method was used by IndoEuropeanists. Our method, however, is much more strict, since we compared always three or more languages at the same time. Our results go together with the basically historical results of Károly Dombi: “Thus, from the evidence left by this process of colonization, it appears that the Sumerian citystates were able to exert a preponderant economic, cultural, linguistic and ethnic influence during several thousand years not only in Mesopotamia and the rest of the Near East, but also beyond, in the Mediterranean Basin, in the Danubian Basin, in the regions North of the Caucasus and of the Black 
Sea, the Caspian-Aral, Volga-Ural, and Altai regions, as well as in Iran and India. 

It seems therefore that the Sumerians and their civilization had a determining influence not only on later Near-Eastern civilizations, but also on the Mediterranean, Indian, and even Chinese civilizations, as well as on the formation of the various Eurasian ethno-linguistic groups” (Dombi 2001, p. 7). 





We know from von Torma (1894) , Bobula (1951) , Vlassa (1963) , Badiny (2001) and Toth (2007a) that the SUMERIANS originated in Transylvania before they migrated to Mesopotamia from where they spread out and a part of them went back in the Carpathian basin.

The Sumerian-Hungarian genetical-linguistic ralationship was proved in a long series of books and articles, last in EDH (Toth 2007). The Huns were already considered to be Mongolians by Deguignes, Pallas and Bergmann (cf.Doerfer 1973,p.13)

The genial German orientalist Julius von Klaproth (Berlin 1783-Paris 1835) identified them directly with the Hungarians (Klaproth 1831) cf. also Moravcsik (1983pp.231 ss).

Moreover ,the Avars as well as the Scythians were already identified with the Huns by Konstantinos Porphyrogenetos in the 10th century, who also first mentioned the Magyars as "sabartoi asphaloi" showing thus the genetical ralationschips of the Hungarians and the Mesopotamian Subaraens (Moravcsik 1983,pp53,279).

The Medes were stated as Scythians by the famous orientalist Friederich Wilhelm König who showed also that they originated in the Zagros mountains in Mesopotamia (König 1934).

Finally, that the Partians were a Scythian people, is accepted by everybody today, cf Colledge (1967) and was shown on Sumeriab background by Badiny (1998,1999).


Prof.Dr.ALFRED TOTH - pdf



Since I have stated here using linguistic, archaeological, historical and biological-genetic facts that the Rhaetians are not related with the Etruscans, but the Etruscans and the ancient inhabitants of Lemnos are related with the Hungarians, the Turks and the Sumerians, we may ask further if Etruscan is really, as asserted by Alinei (2003, 2005a), a form of Early Hungarian or if it belonged to another language closely related to Hungarian. This question is not superfluous, since I have recently proven that Hunnic was not – as generally assumed (e.g., Menges 1968, p. 17ss.) – a Turkic language, but a very close relative of Hungarian (Tóth 2007, forthcoming). 

Thus, in the following I will give a maximally complete Etruscan-Hungarian-Sumerian dictionary that comprises all up to now known Etruscan words. 

Further, I will base the word-list on my “Hungarian-Mesopotamian Dictionary” which does not only compare modern Hungarian forms with 3000 years old Etruscan and 5000 years old Sumerian forms, but also shows the proto-Ugric, proto-Finno-Ugric, proto-Uralic and proto-Altaic forms reconstructed by the adherents of the respective linguistic theories in order to bridge the time-gaps between the actually testified words.

Prof.Dr. Alfred Toth - pdf