László Marácz*
The book on the Ural-Altaic language classification ‘Towards Eurasian Linguistic Isoglosses: the Case of Turkic and Hungarian’ confronts two approaches to linguistic correspondences in the Eur-asian space. The first one is the postulation of genetic relations and the framing of relations on the basis of the classical binary branch-ing family tree theory developed in the work of August Schleich-er (1821-1868) in reference to the Indo-European languages. The alternative interpretation of linguistic correspondences in terms of the Sprachbund or linguistic area theory has been developed by the Prague School of which the two most important representatives were the Russian émigré linguists Nikolai Sergeyevich Trubetzkoy and Roman Jakobson (1896-1982). Before the linguistics affinities between the Uralic and Altaic languages are discussed let us first have a closer look at the hypothesis of the Indo-European genetic language family [Morpurgo-Davies 1998: 285].
The language family tree of Schleicher and his ideas on the In-do-European proto-language, hierarchies and naturalistic evolution [Schleicher 1850] were most visibly challenged in the twenties of the twentieth century by Nikolai Sergeyevich Trubetzkoy. According to the latter, these correspondences were the result of linguistic contact instead of linguistic genealogical relations. In order to accommodate this concept Trubetzkoy developed his theory of the Sprachbund, i.e. ‘linguistic area’. The theory of linguistic area conceptualizes the diffusion of structural features across language and language fam-ily boundaries within a geographical area. Neighboring languages in a geographical space display lexical and grammatical parallels, because of a permanent, intensive contact between them and these parallels may cross genetic boundaries. The key words for areal lin-guistics were not ‘genetic’, and ‘tree’ but ‘net’ and ‘chain’.
Contrary to Schleicher, Trubetzkoy argued that in order to account for the origin of the parallels between the Indo-Europe-an languages it is not necessary to assume an Indo-European Pro-to-language. These correspondences can also be due to permanent, intensive contact, mutual influence, borrowings or language mixing. The borrowings, i.e. the results of linguistic contact take place ac-cording to integrative rules as well. Trubetzkoy concluded that com-mon structural features instead of lexical and morphological ones are the most important for establishing “genetic” relationships. A language family located between two other families can be inter-preted as a link between the two, if it shares some structural features with its neighbors. TELI adopts Sprachbund, linguistic or diffu-sion area in the sense of Campbell [Campbell 2006: 330-331] that includes not only defining structural features but also lexical and morphological characteristics: “The term linguistic area refers to a geographical area in which, due to borrowing and language contact, languages of a region come to share certain structural features — not only borrowed words, but also shared elements of phonological, morphological or syntactic structure.” In terms of a theory of lin-guistic areas the shared features are referred to as isoglosses. Hence, isoglosses are the geographical boundary of certain linguistic fea-tures [Campbell 2006: 75, 251].
The idea to postulate a tree diagram of descent for language family groups was originally put forward in the work of August Schleicher*2. Already in his early work on the lexical and morpho-logical connections between Indo-European languages the idea of ‘language hierarchy’ played an important role. Note that this idea was widely shared in his age. In Schleicher’s Die Sprachen Europas in Systematischer Uebersicht [Schleicher 1850: 1, 8-9] he clearly adhered to the idea of ranking languages according to the linguistic type. Schleicher [Schleicher 1850: 1, 8-9] assumed that languages with inflection were superior to agglutinative languages and these languages were in their turn superior to monosyllabic languages [Toman 1995: 197]. Schleicher and his followers argued that related languages that were determined on the basis of lexical and morpho-logical correspondences in which the ‘sound laws’ in the sense of Jacob Grimm (1785-1863) could be attested were linked by these sound laws via intermediate stages; and finally all the languages and intermediate forms were derived from the so-called Ursprache, the original language, i.e. the proto-language. The application of Schle-icher’s theory to the Uralic/Finno-Ugric linguistic family is rejected in TELI and a theory for Eurasian linguistic relationships between Uralic and Altaic languages, especially concentrating on Hungarian and Turkic is developed.
Ural-Altaic: language family groups and linguistic categories In the seventies of the nineteenth century the Uralic/Finno-Ug-ric language family was still considered to be related to the Altaic languages, especially to Turkic and Mongolic. All these groups and languages were originally included under the “supercategory” of Al-taic. Only in the last quarter of the nineteenth century was there an attempt to establish the Finno-Ugric language family with the help of scientific argumentation. The German linguist Josef Budenz start-ed the work on finding regular sound laws in order to reconstruct the intermediate levels of the Finno-Ugric language family and de-rive these from the Finno-Ugric original language, the Finno-Ugric ‘Ursprache’. Finno-Ugric linguistics closely followed methods that were developed in the nineteenth century by the German School of historical-comparative linguistics. Not only the assumption of a genealogical classification in the domain of the Uralic/Finno-Ugric languages — and I hasten to add Indo-European languages — is con-troversial but the assumption of a common Altaic language family is controversial as well. Campbell [Campbell 2006] also takes notice of this debate about the make-up of the Altaic family group. His classification of the Altaic language family closely follows the so-called ‘micro-Altaic’ family: Turkic, Tungusic, Manchu and Mon-golian, to which some proposals also add Ainu, Japanese, Korean and others [Campbell 2006: 345]. However, the genetic relations and the hierarchical relations among these languages are not gen-erally considered as proved [compare Sinor 1982: 125]. The Polish Turkologist Jankowski [Jankowski 2012: 445] summarizes the long discussion on the Altaic theory as follows: “… we may say that there are three basic opinions. One group of scholars share the tradition-al view of the genetic affinity of Altaic languages, represented by Ramstedt, Aalto, Poppe, Starostin, and so on. At present this tradi-tional line of study has few supporters. Another group of scholars refute the Altaic theory using the arguments formulated by Doerfer, Scherbak, and so on. There are still some scholars who share this view. It seems that most scholars active in Altaic follow… Altaic unity based on language contacts.” In sum, this unity can have a kinship basis or not. Throughout TELI the practice will be followed that Altaic linguistics is simply the comparative research of Turkic, Mongolian and Tungusic.
It has been observed by a number of scholars that there are ‘deep’ Ural-Altaic linguistic correspondences. What to do with lin-guistic convergences between seemingly, unrelated language fami-lies? The research of the correspondences between Uralic and Altaic leads Campbell [Campbell 2006: 346] to claim that Ural-Altaic is even a case of distant genetic relationship. Are these affinities in-stances of cognates that point at a genetic relationship; instances of copies that point at language borrowing, or instances of look-alikes that are due to mere chance?
Sinor [Sinor 1982: 30-31] also argues against genetically de-termined Uralic and Altaic language families in similar vein as Trubetzkoy did against the postulation of a genetically determined Indo-European family. Sinor points out that common features alone are not suited to prove genetic relationships. They can be isolated phenomena and parallel phenomena does not have to originate from genetic relations [Sinor 1982: 126-127]. In both cases, the Uralic and Altaic genetic language families are only hypotheses that await further proof. A research strategy that would first clarify the internal problems of the individual Uralic and Altaic language families is also treated with skepticism by Sinor.
The research of the correspondences between Uralic and Altaic, interpreted by Campbell (2006, 346) as a case of distant genetic rela-tionship, is actually doing comparative research between the follow-ing groups: Samoyedic, Finnic, Ob-Ugric, Hungarian, Turkic, Mon-golian and Tungusic. Following Sinor (1982), TELI assumes that these language groups display a clear pattern of internal cohesion and generally are considered categories that have been studied sep-arately from each other. The idea of separate groupings within Ural/ Finno-Ugric and Altaic is of course also reflected in the different layers and branching of the family tree model. Hence, the research pursued in TELI will study Ural-Altaic affinities in a framework of comparative research relying on the above linguistic categories. The following diagram displays the full taxonomy TELI operates with:
Language categories of the Uralic and Altaic language groups Finnic, including Zyrian (Komi), Votyak (Udmurt), Lapp (Sami), Mordvin (Erzya, Moksha), Cheremis (Mari), Finnish, and
Estonian;
Ob-Ugric, including Vogul (Mansi),and Ostyak (Khanty); Hungarian;
Turkic, including Turkish, Ottoman Turkish, Azerbaijani, Turk-men, Karachay-Balkar, Crimean Tatar, Cuman, Karaim, Tatar, Ka-zan Tatar, Mishar Tatar, Sagai Tatar, Lebed Tatar, Tobol Tatar, Bash-kir, Kazakh, Kyrgyz, Uzbek, Sart, Eastern Turkic, Uyghur, Taranchi, Western Yugur (Yellow Uyghur), Old Turkic, Chagatay, Middle Tur-kic, Sakha (Yakut), Tuvan (Soyot), Tuba (Sojon), Khakas, Shor, Al-taic Turkic, and Teleut;
Mongolic, including Middle Mongol, Classical Mongolian, Buryat, Mongolian proper (Khalka), Oirat, and Kalmyk;
Tungusic, including Even (Lamut), Evenki, Solon, Manchu, Anai (Gold), Orok and Ulch;
Samoyedic, including Enets, Nenets, Nganasan, Selkup (Ko-jbal), and Sayan;
Yukagir;
Japonic, including Old Japanese and Japanese; Koreanic.
Firstly, TELI considers the alternative groupings of Uralic and Altaic in seven subgroups (and Japonic and Koreanic) outlined in the above diagram as the relevant levels of comparison (see espe-cially chapter six and ten) and secondly I will check whether different, related lexical and morphological phenomena can be cap-tured in terms of isoglosses. Convergences between languages in a geographical area are then not due to genetic principles a priori but rather are to be dealt with in terms of language convergences to be accounted for in a linguistic area framework that involves copy-ing and borrowing of lexical and structural features as well. These convergences may or may not neutralize a theory of genetic rela-tions and cognates. It is well-known from historical records that the languages of the so-called Uralic and Altaic language groups have been spoken in the Eurasian/Central Asian spaces. These territories are language contact area’s par excellence [Shirokogoroff 1970]. In TELI, I will arrive at four isoglosses involving the different Uralic, Altaic, and Indo-European language families. These isoglosses will define the ancestor of Hungarian, the so-called ‘Magyar-Ugor Pro-to-language’ to be elaborated in part four of TELI.
Vámbéry reinterpreted
The consequences of the Uralic/Finno-Ugric theory have been far-reaching, however. The Ural-Altaic relations and within this Eurasian “belt” the Hungarian-Turkic relations have been reduced to a case of borrowings, more or less accidental contacts in the course of history. That these relations are of a “deep” character on the level of the root and appear in lexical elementary domains, like kinship terms, body part terms and numerals have been neglected, suppressed or marginalized for “theoretical” reasons. TELI demon-strates that Hungarian-Turkic lexical and grammatical affinities had always a secondary position, as opposed to Hungarian-Finno-Ugric linguistic contacts. I argue in part two and three of TELI that this theoretical premise has been unfounded, though. This side effect of the primacy of the Uralic/Finno-Ugric contacts has actually led to a series of incorrect hypotheses in the domain of Hungarian-Tur-kic contacts. Apart from neglecting the “deep” linguistic connec-tions between Hungarian and Turkic also the hypothesis that the Volga-Bulgar and Hungarian linguistic contacts were mainly in the agricultural field of the lexicon are incorrect. Of course, it remains a vital question to be answered why the agricultural domain of the lexicon should not be a primary domain of the lexicon, including for example the lexical references to animals, like Hungarian ökör ‘ox’, Chuvash văkăr, and Common Turkic (including Kazakh) öküz. But even if this question is put aside that has been done in mainstream Uralic/Finno-Ugric it is easy to check from the Chuvash-Hungari-an lexical correspondences that among this set of words there are a number of elementary basic roots, root words and roots that have nothing to do with agriculture, like Hungarian gyűrű ‘ring’, Chu-vash śĕrĕ, and Common Turkic (including Kazakh) yüzük. But even more striking, the sound correspondences affecting the Turkic lan-guage family, that has been referred to in the literature as the ‘Lir’ and ‘Shaz’ alternation and are connected on the one hand to the Chuvash-type of Turkic languages matching with Hungarian sound patterns in parallel lexical items, including the —r in both languages, and are connected on the other hand to the Common Turkic languag-es respectively have never been considered as empirical evidence for Grimmian/Schleicherian divergence. Hence, it forced the setting up of a “straw man” called “West Old Turkic” leading to all sorts of new asymmetries and speculations that have never been proved actually, like the claim that Hungarian is always the recipient — but never the donor-language -, i.e. under all circumstances Hungarian has only borrowed from Chuvash (or one of its ancestors) and vice versa has been excluded.
The idea that Eurasia includes a trans-Eurasian Ural-Altaic typological belt has been put forward by the Finnish Altaist Juha Janhunen and the Italian linguist Angela Marcantonio specializing in Uralic/Finno-Ugric languages, although Janhunen and Marcanto-nio differ on the position of Hungarian in this belt*3. Both scholars support the idea of a large trans-Eurasian typological belt of lan-guages known as “Ural-Altaic” which compromises next to “Altaic” also Korean (Koreanic) and Japanese (Japonic). This is a position that TELI subscribes to. To conclude, let us consider again the “Ug-ric-Turkish War” and the position of Vámbéry in this discussion (see chapter one and four of TELI). Recall that the Ugric-Turkish War started in 1870, when Ármin Vámbéry published his book entitled Magyar és török-tatár szóegyezések (Hungarian and Turkish-Tatar Cognates). This study on the Hungarian and Turkish-Tatar cognates started a series of polemical papers between on the one side mem-bers of the Hungarian Academy of Sciences Paul Hunsdorfer (Hun-falvy Pál) and Josef Budenz and on the other side Vámbéry himself. Although Vámbéry did not deny the mixed “Finnish” and “Turkish” character of the Hungarian language he argued however that “in the amalgamation not the Finnish-Ugrian but the Turko-Tatar element predominated” [Vámbéry 1905: 480]. Until his dead, Vámbéry re-mained convinced of the fact that Hungarian was a “mixed” language that was above all related to Turko-Tatar. Just before his dead in 1913 Vámbéry repeated his point of view in a book that appeared after his dead A magyarság bölcsőjénél (At the Cradle of the Hun-garians) in 1914. Vámbéry’s view is supported by the percentages of the words of Hungarian basic stock of Finno-Ugric origin only, of Turkic origin only, and of Turkic and Finno-Ugric origin only that are as follows: 33.1 percent, 32.7 percent, and 34.2 percent re-spectively, compare his book A magyarok eredete (The Origin of the Hungarians) [Vámbéry 1882]. Although, Vámbéry’s collection of lexical Hungarian-Turkic parallels is more solid than the Budenz’s one-sided, selective investigation of Hungarian and Finno-Ugric parallels both the Hunsdorfer/Budenz camp and the Vámbéry camp have in common that they assume that the Hungarian forms are de-rived from the Finnish or Turkic ones respectively, that is Hungarian is in all cases the recipient-language. Hence, this yields Vámbéry’s concept of Hungarian being a “mix” Uralic/Finno-Ugric and Turkic language. In the light of the root system outlined in chapter three of TELI this would imply that there would be in fact no Hungarian roots, or to put it otherwise Hungarian would have no roots of its own. A conclusion extremely unlikely in the light of the organic root system uncovered in chapter three. The “mixing”-effect can also be explained by reversing the direction of borrowing, as is suggested in chapter eleven of TELI. If the ancestor of Hungarian, i.e. the Pro-to-Hungarian Ugor-Magyar language is the donor in both directions the lexical and other correspondences with Finnish and Turkic cre-ate the perception of a mixed language but in reality it is not.