RÁCZ, ANITA Ethnic groups and settlement names in Hungary1 Ethnic group names as elements of the Hungarian onomasticon are present from the earliest period of the Hungarian language. According to my re- search there are 39 ethnic group names that took part in coining settlement names. The analysis of the toponyms containing ethnonyms shows us that they can be clearly described morphologically, and they belong to differ- ent morphological types. Between these types we can detected significant chronological differences. In the earliest period on the one hand we find toponyms formed without any name-formant, while some names were coined by the formant -i, less frequently by -d, but there are other suffixes as well which were used rarelier. Later such structures appeared that con- tain an ethnonym and another lexeme (e.g. other toponyms, common geo- graphical name). In my paper I show the latest findings on this important layer of the Hungarian onomasticon. Keywords: ethnonyms, typological analysis, topoformants, interdisciplinal. 1. The Hungarian language was the first of the Uralic languages to be recorded in writing. The earliest written records of the language, dating from the 11th century, are found in foreign language texts. These are essential for research in historical linguistics, since from them we can learn a number of valuable lessons about the contemporary state of the Hungarian language, and the subsequent change processes it underwent. However, these names are not only of interest to those who concern themselves with the linguistic sciences. Since place names reflect the everyday life of the name-bestowing community from many aspects, representatives from the fields of history, ethnology and geography consider it vital to use such name data in accordance with the goals of their own research. In this respect historians and linguists are fully in agreement that historical re- search into toponyms provides such important source material on the early his- tory and language history of Hungary that it cannot be ignored. The verifiable name layers from the early period of the conquest were sub- jected to study in Hungary from the 1930s and 1940s, resulting in early histori- cal place-name typologies (Moór 1936: 110–117, Kniezsa 1938, 1943, 1944, 1960, Kertész 1939: 33–39, 67–77, Kristó 1976), the results of which are still to this day largely accepted by the research community without reservation. These 1 This work was carried out as part of Research Group on Hungarian Language History and Toponomastics (University of Debrecen – Hungarian Academy of Science). Nyelvtudományi Közlemények 109. 255–266. 256 RÁCZ, ANITA pioneering studies established the semantic and morphological characteristics of the typical components encountered in the Hungarian place name system. In ad- dition this research can be linked to historical onomastics and to the generally accepted proposition that there are so-called ’era-defining’ settlement names. By this we understand that certain types of name were characteristic of particular periods of Hungarian history, which of course does not mean that the presence of that name type was exclusively indicative of that age, but rather that its ap- pearance is strikingly characteristic. Accordingly, we can use the designation of „old or early settlement name type”, amongst which we can identify personal names, names of tribes, ethnonyms and occupational names, either without any formant or with an affixational morpheme, which have given rise to the creation of settlement names (see Kristó 1976 and L. Kiss 1997). Thus we can see that, from the very beginning, the names of ethnic groups were present in the naming system of Hungarian place names, and that these impacted on the name layer and on any general statements that can be made about early place-name types. In the first half of the 20th century István Kniezsa systematized Hungarian place names based on the names of peoples into two major groups: simple and complex type names. The first type includes a subset of ethnonyms without any name-forming element (Horvát, Tót, Orosz etc.); the other subtype has the pos- sessive suffix -i attached to the names (as in Németi, Csehi, Tóti etc.). The com- plex type place names are those where the name of a people has some common noun used in topographical description attached to it (Tótfalu, Oroszfalu, Oláh- telek, Olasztelek etc.) (Kniezsa 1943: 124). Roughly two decades later, this basic division into three typological groups was accepted by Gyula Kristó (1976: 58– 65), and such an organizational framework must still be kept in mind by all re- searchers who engage with the topic. Since the research methods, principles and basic concepts established in the early 20th century have been passed on to the present day essentially unchanged, I decided that it was high time to undertake, after looking back through the dec- ades-old typologies, a comprehensive review of settlement names based on eth- nonyms using a large database of name material, with a view to refining and adding to the store of knowledge. I actually started work on this theme in 2011 with the publication of a monograph entitled „Data concerning early settlement names derived from ethnonyms”, which deals with name data from the conquest until 1526, gathered from various written sources, collected and published. For 945 settlements a total of 1355 name-formants and something in the order of several thousand items of name data can be found, amongst which can be highlighted 39 lexemes related to the names of ethnic groups. The frequency order for these items is as follows: tót, német, magyar, oláh, orosz, besenyő, szász, cseh, olasz, székely, horvát, maróc ~ marót ~ morva, lengyel, tatár, böszörmény, kun, nán- dor, káliz, úz, zsidó, kazár, román, várkony, jász, török, cigány, rác, bolgár, Ethnic groups and settlement names in Hungary 257 szerecseny, polyák ~ polány, komán, bajor, korontál, görög, kölpény, sváb, szerb, avar, örmény. Here I only want to show in outline what kind of overall picture we can draw at the present time of the role played by name-forming eth- nonyms in that layer of names belonging to the early name stock of Hungary. 2. There are significant chronological differences between the various types of place names constructed on the basis of ethnonyms. Those which either don’t use a name formant or simply use a suffix added to the ethnonym are much older than those which are formed from ethnonym + other lexeme (e.g. common nouns or adjectives referring to a geographical locality). In studying the relationship between ethnonyms and toponyms, linguists es- tablished from the very beginning that settlement names can be constructed on the basis of singular nominative forms of ethnonyms without any difficulty. In- deed, these lexemes (together with the names of tribes and occupations) have a collective meaning even in the singular form (Kertész 1939: 37, Kniezsa 1943: 124), being so-called social group names. Early typologies considered this name group as the first type of construction, and in my taxonomy I have also taken this name structure into consideration, indeed I have slightly expanded the traditional framework. While previous studies only considered name elements of the type Besenyő, Német, Székely type of name forms were recorded here, I considered that name forms name forms such as Bogorbesenyő, Felnémet, Püspökszékely etc. also belong here, as the basic form of the salient ethnonym often occurs in an attributive structure, playing a role in the second part of the name. All such names are, however, the result of a secondary development, the primary form of ethnonym-based place names being the nominative singular. In fact, the result of the emergence of tag complementation can be taken into consideration even in cases where there is no concrete evidence for a specific form of the name in ex- tant written records. For this purpose the name taxonomy itself provides enough proof. Just how natural this name pattern is in Hungarian is shown by the fact that, out of 39 ethnonyms which I considered in my samples, 36 were turned into settlement names on the basis of this very pattern. Furthermore, when we look back to the first appearance of all the settlement names based on ethnonyms, roughly one third are traceable in origin to name forms of this type. Regarding the chronology, several researchers agree that the use of settlement names referring to peoples without recourse to a name formant was probably established during the 10th century. László Makkai listed them together with place names derived from tribal names and discussed them as one unit, restrict- ing their development exclusively to the 10th century (Makkai 1947: 112), a con- clusion echoed by György Györffy (1958: 60–87) as well as the authorial team András Mező and Péter Németh (1972: 116). Typologically it is clear that these names can be linked to place names based on personal names without 258 RÁCZ, ANITA name-formant elements, as is done in the work of of István Kniezsa and Géza Bárczi. Nonetheless the chronology they establish is the same as stated by the researchers given above. The final date for the development of this type of name is considered by all of these researchers to be no later than the 13th century. Gyula Kristó, by contrast, states that the earliest formation of place names character- ized by this structural pattern can be dated back to turn of the 10th–11th centuries, and may have occurred up until the end of the 14th century (Kristó 1976: 59, 61, see also L. Kiss 1997: 180). The examination of the chronological particuliarities of the place names which I collected, based on a much larger sample compared to those of previous studies, seems to confirm the arguments cited above and the analyses proposed by Gyula Kristó and Lajos Kiss. The early period is in general is poorly docu- mented and the database contains a relatively insignificant number of names: from the 11th century just 10 citable items of data, and from the 12th century only 16 names. Naturally, this may be attributable to record-keeping practices which had not yet been fully established and, in consequence, the scant number of sur- viving documents, as well as the „newness” of the foundation of permanent set- tlements. The number of settlement names based on ethnonyms without name- formant elements is seems to jump significantly at the beginning of the 13th century, and this naming method most clearly in evidence during the second half of this century, and is also well-represented during the first half of the 14th cen- tury, but there is no question of it disappearing during the historical period fol- lowing that currently under scrutiny. Metonymic name formation is present throughout, but as we move forward in time, it may be possible to state that it plays a less important part in the creation of new names. For the vast majority of name forms representing this name type (63%), the earliest name form is synonymous with any possible variant names of the set- tlement. In other words, on the basis of the available documentary evidence only a small proportion of structural type variations have been discerned, the names in question rather retaining their primary formations. Amongst the secondary formations the following structural type is represented in large numbers: affixed initial name element + ethnonym-based main element without name for- mants. For these acts of naming the vast majority are characterized by comple- mentation through suffixation, which is explainable for onomastic systemic rea- sons as a common phenomenon. In my opinion the following conclusion is sup- ported by the name data: the type of names formed by affix + ethnonym without name formant came into being with a slight chronological „phase delay” in comparison with the structure ethnonym without name formant. The former started to appear in writing when the latter forms were still in use and the chosen modes of complementation were influenced by the desire to reflect extra- linguistic factors from the environment. Ethnic groups and settlement names in Hungary 259 3. The second group in the traditional typology includes those instances where a topographical affixational morpheme is added to the ethnonym, the most typical being -i, followed by the less common-d, to create settlement names. According to my most recently collected name material, the suffix -i was used to created the first attested formation of 79.1% of such names, with the -d suffix accounting for roughly one-fifth of that number at 16.5% of the occurring designations. The presence of other affixational morphemes observed (-y, -j, -ka/-ke, -ny, -s) is in- deed negligible at just 4.3%. Some of our researchers noted very early on that, when analysing names con- structed using topographical suffixes, certain ethnonyms were characteristically associated with one or the other name formant and, conversely, that there are ethnonyms with which this type of name structure never occurs. In propounding an explanation for this phenomenon chronological criteria were mostly considered: that amongst certain early settlement names the name components *Oláh(i), *Rác(i), *Tatár(i), *Török(i), *Cigány(i) do not occur, and these ethnonyms only appear in two-part names as the first signifying element (Oláhbáród, Cigányfal- va). The responsibility for this fact occurring on a regular basis is placed on the late emergence of these settlements in a period when settlements names could no longer be formed on the basis on bare ethnonyms or ethnonym + the suffix -i (c.f. L. Kiss in 1996: 447). Loránd Benkő analyzed the specific behaviour of settlement names based on ethnonyms in several of his studies, but his explana- tion is quite different from that of researchers who preceded him. He sees the reason in the connected intersection of lexical, root and phonological dependencies and certain chronological aspects (Benkő 1998a: 71, 1998b: 119). Although this idea was used by its originator without entering into very detailed analysis to justify it, by drawing on my own name sizeable corpus, I was in some cases able to specify the morpho-phonetic conditions connected to certain name formants (2008). 3.1. It seems clear from my previous research experience that the most com- mon and the most natural way of compounding ethnonyms into toponyms between the 10th–16th centuries was through the addition of the place name suffix -i. An ex- planation could perhaps be sought from the same direction as the semantic nature of ethnonyms, and the close relation to this of the suffix’s origin and original „mean- ing” and function: from the possessive marker é, from which parallel forms and meaning cleavages were separated out, becoming a place-name suffix in its own right (Szegfű 1991: 254, cf. Makkai 1947: 113, Kázmér 1970: 57). When considering the chronology of settlement names formed with the -i suf- fix it is important to note that, since it was in use for a relatively short period of time, it is usually considered as having a role to play in determining the limits of an era. Kniezsa’s position is that the vast majority of representatives of this name type came into being before the middle of the 13th century (1949: 100, 107, 1960: 20, c.f. also Bárczi 1958: 149, 157, 160). Gyula Kristó placed the first formation of place names employing the suffix -i rather earlier, at the very 260 RÁCZ, ANITA beginning of the 11th century, or even at the end of the 10th century, with the most productive period using this formation being the 13th–14th centuries and, in his opinion, new coinages were appearing evan as late as the 15th century (1976: 51–52, 77). The most recent monograph to deal exhaustively with place-name formation is by Ágnes Bényei, who writes that the place-name forming suffix -i appeared in place names at the end of the 13th century or in the 14th cen- tury. Then it was at its most productive, and the effect of contemporary fashion was such that it was responsible not only for the formation of primary names but, by analogy, existing names were added to using this name formant (Bényei 2012: 84). And while from her findings Bényei considers the general application, features and chronology of the place name suffix -i, on the basis of my own name material I can confirm that place names constructed on the basis of ethnonyms plus this morpheme blend into the overall picture. The ethnonym + -i struc- ture was applied most notably in the second half of the 13th and the first half of the 14th century. However, this structure also played a part in the creation of new names in the subsequent period of language history. In my name material almost half of all settlements using the -i name formant first appeared as names in their primary form between the 12th century and the end of the 15th century. Among secondary names two major groups can be distinguished, that is to say that within the type shift two specific forms can be seen: one is the expansion of the ethnonym (as toponym) without formant through the addition of the -i suffix (Cseh > Csehi, Horvát > Horváti, Orosz > Oroszi). The second group are the names in which the type shift occurs by the addition of a first part to the already affixated eth- nonym (Németi > Szatmárnémeti, Oroszi > Füzesoroszi, Tóti > Kistóti). – From my investigations I conclude that the systematic lack of the -i formant with some ethnonyms can be explained by a form of suffix blocking, which would prevent it from being added to the final vowel ending of the given ethnic signifier. 3.2. The second most common place-name affix -d was originally a diminu- tive, pet-form, and from this has evolved or developed in parallel a meaning ap- proximating to ’supplied with something’. Relatively early on it became a typi- cal place-name affix, and its use in connection with place names is attested from the 10th century (cf. Szegfű 1991: 253). It is linked for the most part to lexical items for plants or animals, and when connected with the basic has the function of indicating the wealth of the named plants or animals in the surrounding coun- tryside. In addition, though somewhat rare, it also occurs in connection with eth- nonymic lexemes. From our point of view it is significant to note that this suffix is often added as a pet form to a personal name (Inánd, Jánosd, Kaszád etc.), and thence may be created a settlement name on the basis of a nickname without any problem thereby enriching the number of -d ending place-names. These desig- nations further contributed to the spread of the group of settlement names using the d-formation and, to some extent, encouraged the development of this affix as a name formant for settlements. Ethnic groups and settlement names in Hungary 261 The basic position taken in earlier literature was that the -d formation, like the -i suffix had been previously, was at its most productive during the 13th–14th centuries, and that place names were still being created on this pattern until the end of the 14th century, with later coinages of this type only occurring through analogy (Kniezsa 1943: 127, Bárczi 1958: 155). Gyula Kristó examined and re- viewed the time boundaries for this name type. In his opinion, the spread of place names using the suffix -d was uneven (as is the case for various place- name types), differing in popularity from one geographical area to another (1976: 86). He also maintains that the name formant’s vitality may have per- sisted down to the 15th century. It is his opinion that those names which turn up with a -d ending based on analogy only appeared after the 15th century (cf. 1976: 88). According to my own research, the most intensive period for the use of this name formant coincided with the period when the -i suffix also most character- istically appeared, i.e. the second half of the 13th century to the first half of the 14th century, albeit in much smaller numbers, with the tokens of the former type being roughly one-fifth of the latter. The reason for this might be, on the one hand, that name-givers may have got more used to collocating the -i suffix with ethnonyms and in this semantic category felt the -d morpheme to be less appro- priate as a place-name formant, though its use was not ruled out. On the other hand, for some ethnonyms (in addition to the random lack of attestation) there may be morpho-phonetic reasons for their absence, since the data I have a col- lected data show that names ending in -t and -h never take the -d suffix, encour- aging one to believe that a form of suffix-blocking is in operation. 3.3. Settlement names based on ethnoyms rarely turn up with the name for- mants -é ~ -j ~ -aj/-ej, -ka/-ke, -ny, -s. These have hardly been dealt with by ear- lier typologies, which is understandable if you consider that, in my substantial corpus, a total of only 16 such settlement names appear using any of these com- ponents. In addition, for some of these the etymology is uncertain, leading us to refrain from entering into more detailed discussion of them. 4. The third structural type of settlement names of ethnonymic origin according to classic typologies are those names structured on the pattern of a first element, which is the marker of the ethnic group, followed by the main element, which is a common geographical name signifying some type of settlement (-ház(a), -te- lek(e), -lak(a) etc.). The propagation of the earlier established typologies appears to have begun during the 13th century. According to Elemér Moór, the names of settlements using the -falva suffix started to take root during the second half of the 12th or the first half of the 13th century (1936: 117). István Kniezsa’s view is that the ge- nesis of earlier place names based on personal names without a formant [or more generally any constructed without name formants – A. R.] were replaced by the new type of compound place name at almost lightning speed during the 13th 262 RÁCZ, ANITA century, and that they functioned next to each other in parallel for less than half a century (cf. Kniezsa 1943: 128, Bárczi 1958: 160, Szabó 1966: 136). The issue of the chronology of early Hungarian place-name types was subjected to detailed scrutiny by these authorities, and their findings have been complemented and clarified by Gyula Kristó who, however, comes to quite different conclu- sions. According to him, there was no sudden change of type and the period of parallel use started earlier and ended later than Kniezsa and his followers claim: spanning from the 11th to the middle or end of the 14th century. In support of his claim he points out that the type of compound settlement name reached its peak of popularity in the 15th century when the type of place names without for- mant was much less lively (Kristó 1976: 92). My research results are more in line with the findings of Kristó than of any others. The appearance of this name type as a lexical group can be located in the second half of the 12th century, even though the highest proportion occurred from the first half of 14th century up to and including the first half of the 15th century. During the second half of the 15th the century the number of instances of this name type falls to less than half of the tokens found for the preceding half century. The traditional typology did not consider the question as to how and accord- ing to what structure common geographical names were connected to eth- nonyms. I have been able to conclude on the basis of my investigations that in the case of geographical common nouns linked to ethnonyms the tendency is for the morphology of the generic suffixes to be grammatically unmarked. Although some such place name formants do use personal suffixes implying the posses- sive, a much higher proportion of the names inspected employ the term in its base form, this phenomenon being clearly visible in the endings -falu, -telek and -vár. However, for some other formants such as -földe, -háza, -laka, the opposite tendency is true. Some common geographical names, while being of only scant overall incidence, exclusively appear in their base forms, these being -egyház, -sok, -szállás, -ülés, -város, -vég and the special case of -falud. One the other hand, the toponymic terms -lakosa, -monostora and -váralja are only found in posses- sive structures. It is noteworthy that if we look into the chronology of the name formations that appear in documents, we can see that many of the forms bearing a personal marker are due to a change of type: they were established by the ad- dition of a personal suffix to ethnonym-based settlement names without name formants. This means that the earlier or earliest name was an unmarked struc- ture, and the subsequent addition of a possessive marker created the grammati- cally marked forms. 5. In addition to taking into account the three traditional name types presented according to established morphological criteria, in my monograph I also distin- guish two types of name group. One of them is in some ways related to the cate- gory described above, since the ethnonym occupies the initial position as a pre- Ethnic groups and settlement names in Hungary 263 fix, but the principal element is a common geographical term which does not belong to the regular list of place-name formants (Besenyőfő, Tótvölgy, Zsidóha- vasa etc.). In terms of their structure, such place names with ethnonymic content are the least frequently encountered in the time-frame of the 10th–16th centu- ries. On the one hand their creation can be explained by metonymy: a place name took on a narrower or a broader meaning, but was after all still linked to an ethnic group, and the new naming was motivated by the possibility of expressing its characteristics through the use of an ethnonym, the resulting micro-names for emerging settlements in the vicinity becoming eponymous and thus, in a me- tonymic way, turning into settlement names. Generic geographical terms that appear in a settlement name thus indicate primary denotata in the naming, it be- ing named as a particular type of location. On the other hand, this kind of micro- name > settlement name metonymic shift is not the sole means by which names can evolve. Katalin J. Soltész has written a monograph whose goal is to answer this question: „what do we know about proper names?” In it she states that „for certain name types, the characteristic word elements, prefixes, suffixes, deriva- tional affixes and, in fact, non-functional endings act as name formants, on the basis of which people use the existing model to create new names [...]” (1979: 19). Thus, their existence confirms the status of proper names as linguistic ele- ments, and when name users want to create a new proper name, they draw upon this stock of name formants. And if you have a proliferation of similarly struc- tured settlement names, the name makers get accustomed to the fact that these morphemes may be used in creating settlement names, new village names can be created analogically along the lines of the many existing similar names in the name system, and these start being used directly to form settlement names, then we no longer need to take into consideration the metonymic process. Examining the chronological development of the properties of the basic name forms we can see that the model was established for these name entities in the second half of the 14th century and came to the fore during the first half of the following century. However, due to the small amount of data, I don’t think that very far-reaching conclusions can be drawn. 6. Another name type with which I concerned myself are those acts of naming in which the ethnonym serves as the modifier in the first part of the name. The names considered in this group (23.5%) exhibit either a syntagmatic structure, or so-called name differentiation, in which the modifying prefix serves to bring about the result of complementation. In the latter case, structural changes to an ex- isting settlement name have created new name formations. One characteristic onomastic feature of medieval Hungary is that quite a number of identical names can be found in the name materials of several coun- ties, and there is even the example of one basic name appearing several times 264 RÁCZ, ANITA within a single county. The case of villages with the same name and lacking any distinctive markers suggests that their populations used some distinguishing name elements in their spoken language but that these were not recorded in offi- cial documents. However, the result of using of these expanded forms was that, over time and as they became embedded in language use, sooner or later they started to appear in records as well, finally taking their place as fully fledged names in official documentation. At the same time we could asseverate about this name type that settlements with the same base name which are geographi- cally near each other may be derived from a common source. We need to bear in mind that very often we encounter names, the sources of which suggest we are dealing with the existence of a separate settlement, whereas in reality it may not cover a discrete village at all. Such nominally duplicated settlements only tem- porarily bore marked name elements, and later (as also originally designated) they continued their existence under a single base-name form. Three sources of motivation can be identified for the creation of names of this type: 1. The dis- tinctive marking prefix is used to express and clearly identify a hitherto non- existent, newly-formed settlement; 2. The new name element is used to express some change (e.g. of ownership) to some already existing settlement; 3. It is re- lated to some unique feature of the settlement, of a trait being expressed unambigu- ously through the name to distinguish it from another settlement of the same name, possibly expressing the salient difference of the inhabitants. This latter motiva- tion may come into play when a society’s name knowledge – their actual scope of movement, their system of connections – expands and widens, and other, more distant settlements with the same name enter within the name users’ fields of knowledge (cf. Inczefi 1965: 75, Szabó 1966: 135, Bölcskei 2010: 239–248). Through the examination of the chronological characteristics of this name type we can ascertain that its popularity began to rise sharply during the first half of the 14th century, with numbers continuing to grow throughout the examined period and reaching a golden age in the second half of the 15th century. This name structure, its upward trend and zenith essentially coincide chronologically with the appearance of this previously discussed settlement names with a geo- graphical common noun suffix. The new fashion for two-part names brought about the creation of this name structure. 7. One possible research direction for linguists and historians who concern themselves with place names (settlement and water names, micro-names) is the attempt to answer the question as to how useful these names are for the deter- mining the ethnic composition of any given area. The settling of the Hungarians in the Carpathian Basin is still a matter which occupies archaeologists and histo- rians are still employed, and even after all the available resources have been processed a number of issues still remain unclear. However, the theme was ap-