A bridge too far – A Uralic perspective on Volga Bulgarian

A bridge too far – A Uralic perspective on Volga Bulgarian Agyagási, Klára. 2019. Chuvash historical phonetics: An areal linguistic study. With an appendix on the role of Proto-Mari in the history of Chuvash vocalism (Turcologica 17). Wiesbaden: Harrassowitz. XII + 334 pp. 1. Introduction Klára Agyagási’s monograph is a multifaceted contribution to the historical phonology of the Chuvash language as well as to the ethnohistory of the Chuvash and neighbouring peoples of the Volga-Kama region. Agyagási is a specialist in Turkic and Slavic historical linguistics, working in the Slavic department of the University of Debrecen, and since the end of 1970s she has published widely on contacts between Turkic, Uralic and Slavic languages, and especially on the areal linguistics of the Volga-Kama area. She was also one of the co-editors of Etymologisches Wörterbuch des Tscheremissischen (Mari) (Bereczki et al. 2013). As is stated in the preface of Chuvash historical phonetics, the book is a culmination of many decades of research work. The present monograph surveys the reconstructed development of FUF 65: 172–186 (2020) the Chuvash sound system from Proto-Turkic to modern Chuvash, taking into account contacts between Chuvash and its neighbouring languages, Mari, Permic, Kipchak Turkic and Old Russian, mainly in the form of loanwords. Uralic data (such as Turkic loanwords in Hungarian, Permic and Mari) is discussed throughout the book. Due to its scope and abundant content, the book will certainly spark a great deal of discussion among Turkologists, Uralicists and other specialists in historical linguistics. In this review, we present a short overview of the book and then proceed to discuss some of its claims in further detail. The book starts with a chapter entitled The predecessors of Chuvash in the Volga region (pp. 1–34), which presents the most important sources of Chuvash historical linguistics: the Proto-Turkic, Bulgar Turkic and Chuvash loanword layers in various Uralic languages (and vice versa) as well as in Slavic, Tocharian loans in Proto-Turkic, early loans in Middle Mongolian and later loans between Kipchak and Bulgar Turkic in the Volga region, along with possible substrate words. It also presents the relevant written sources, such as 172 https://doi.org/10.33339/fuf.89967 A bridge too far – A Uralic perspective on Volga Bulgarian Volga Bulgar glosses and epitaphs, Khazar sources, and later, modern era written sources of Chuvash. This chapter also includes a discussion of methodology. This first chapter is informative and gives the reader a good overview of the research situation in Chuvash and Turkic historical linguistics, but it would have been even more informative if examples from all of the loanword layers had been provided. Without any examples, it is difficult to assess the accuracy or relevance of the information at a glance. Especially regarding such important loanword layers as various Iranian loanwords into Turkic or loanwords into and from Mongolian, some examples would have been illustrative. The second chapter is called Oppositions in the Oguric consonant system (pp. 35–91), and it discusses the history of sound changes leading from Proto-Turkic to Volga Bulgarian, listing the main developments that set the Oguric languages apart from the rest of the Turkic languages, such as rhotacism and lambdacism. The discussion also addresses Volga Bulgarian loanwords into Uralic, especially into Hungarian. The third chapter, Oppositions in the WOT/VB vowel system (pp. 97–183), deals with the development of Volga Bulgar and Chuvash vocalism. Agyagási argues in favour of different Volga Bulgar dialects, backing up her claims with evidence from Volga Bulgar epitaphs and loanwords into Old Russian, Hungarian, Permic and Mari. Some of the views expressed in this chapter are explored in more detail below. The discussion of vowel developments in the mediaeval and early modern period continues in the last main chapter of the book, Changes in the Middle Chuvash period (pp. 185–243). This is followed by a two-page summary listing the main findings of the monograph. The summary is rather short, and it largely repeats the conclusions made earlier in the book. For such a lengthy book with a great deal of details and data from different languages, a longer and more comprehensive summary of the results, or rather a chapter of conclusions, would have served the reader better. In the end of the book, a lengthy Appendix (pp. 247–298) discusses the role of Mari evidence in the vowel history of Chuvash. The Appendix explores the various suggestions for Proto-Mari vowel reconstructions, loanwords and ethnohistory of the Volga region, arguing that the Cheremis people mentioned in the mediaeval sources are not the same people as the modern Mari. Agyagási also attempts to 173 Sampsa Holopainen & Niklas Metsäranta reconstruct the language of this lost ethnic group, which she calls Low Cheremis, and argues that linguistic traces of this language can be found in the vocabulary of Mari and Chuvash. New West Baltic etymologies for Mari and Chuvash words are also presented, and it is argued that they were borrowed through this Low Cheremis language. 2. General remarks about the book Some general remarks regarding the ways the data is presented in the work are in order here. The bibliographical entries after each chapter make it easier for the reader to go back to the original sources. However, it would have been a great help to have tables summarizing different phonological developments; in particular, a side-by-side comparison of the various Volga Bulgarian dialects and Chuvash would have probably served the reader well. It is an interesting approach to combine purely linguistic data with philological evidence, to try to trace the Chuvash phonological developments through known historical sources and to track down the movements of the Chuvash (and Mari) speakers during the tumultuous Middle Ages. That said, it seems that sometimes Agyagási mixes up linguistic and ethnic evidence. Her remarks about the incompatibility of the linguistic family tree model (p. 92) with the prehistorical movements of the Turkic-speaking populations are unintelligible. As is too often done in linguistics, it seems that here the usability of the family tree as a theoretical model is rejected too hastily. Nor is it clear what exactly the author means by “pedigree theory” (p. 201) as this, to our knowledge, is not a standard term employed in historical linguistics. There are some inconsistencies in the way the book refers to Mari dialects. Hill Mari forms are referred to as ‘Mountain Cheremis’ in one sentence and as the ‘mountain dialect of Mari’ in the next. This is not so much a problem for scholars in the field, but it may be misleading for those unfamiliar with these languages. This terminological ambiguity is especially troubling as Agyagási also discusses the Low Cheremis language, which she assumes is completely unrelated to Mari. Subscript numbers are used widely. Referring to the three different Volga Bulgarian dialects as VB₁, VB₂ and VB₃ may be justified, but it is difficult to see why Late Proto-Mari is referred to as PM₂ when there is no PM₁. Applying subscript numbers when there is neither a dialectal nor chronological distinction to be made seems unnecessarily confusing. 174 A bridge too far – A Uralic perspective on Volga Bulgarian Regarding Hungarian etymology, Agyagási follows the most upto-date views on Turkic loanwords, which are found in the work of Róna-Tas & Berta (2011). Here we would only like to remark that using the Ancient Hungarian reconstructions and not modern Hungarian words makes it a bit difficult for anyone not familiar with these reconstructions to follow. Using modern Hungarian forms (maybe alongside the reconstructed forms) would have been a reader-friendly choice. Agyagási is clearly familiar with the most important sources and research results on the ethnic history and archaeology of the Volga region, citing recent sources such as Zimonyi (2014), but some recent sources are missing here. For the history of the Uralic peoples of the region, Rahkonen’s (2013) results about the substrates in the languages of the Volga region might have also provided interesting insights into the problems that Agyagási discusses. 3. General notes on the Uralic material As Turkic-Uralic contacts play a significant role in Agyagási’s argumentation, some remarks about her use of the Uralic data are in order. While Agyagási refers to several key sources on Uralic historical linguistics and knows the material rather well, there are some unfortunate gaps that have consequences for the results of the book. Agyagási’s views on Uralic vocalism are based on a limited selection of sources, and many important details from recent works have been ignored. Agyagási states that the Proto-Uralic/Proto-Finno-Ugric vowel reconstruction in UEW is the widely accepted one in Uralic linguistics, but this statement is not entirely correct and may be a bit misleading for someone not keeping up with the developments in Uralic historical linguistics. Agyagási presents the two possible vowel reconstructions found in UEW (which differ mostly with regard to vowel length) but ignores notable developments in the field, such as Janhunen (1981), Sammallahti (1988) or Aikio (2012, 2015). While Proto-Uralic vocalism is obviously not the main concern of this book, taking into account the modern views on Proto-Uralic vocalism would have probably resulted in a more balanced view. Unfortunately, similar problems can be seen in Agyagási’s views on the taxonomy of the Uralic language family. Here again, references to relevant modern sources are missing, and the single reference to Honti’s (2010–2011) very traditional view of 175 Sampsa Holopainen & Niklas Metsäranta the taxonomy on page 289 gives the reader a distorted picture about the state of the art of the field. References to the alternative views of Salminen (2002) and Häkkinen (2009) might have been in order, even if one does not completely agree with them. Especially as the position of Mari among the Uralic languages is uncertain, it would have been desirable to pay more attention to this problem: Agyagási does comment on Mari’s position and the problem of the Volgaic node on pages 248 and 252–256, but it remains uncertain to the reader what conclusion she reaches about this. It would be impossible to go through every claim made by Agyagási in the book, but some more detailed remarks on the use of Uralic evidence and on the conclusions derived from it seem to be in order and are thus discussed below. 4. Volga Bulgarian dialects based on Permic Agyagási weaves an intricate web of Volga Bulgarian dialects. The different Volga Bulgarian dialects and Agyagási’s evidence for them are discussed in Section 3.2 of the book (p. 160 onwards). The first Volga Bulgarian dialect (VB₁), is postulated based on loanwords in Old Russian, Permic (both ProtoPermic and “Ancient Votyak”) and Proto-Mari (Late Proto-Mari and Proto-West-Mari). Interdialectal borrowing constitutes the second source of evidence. According to Agyagási, around 20 words were borrowed from an extinct Volga Bulgarian dialect (VB₁) into a dialect that would eventually become Chuvash (Early Middle Chuvash, abbreviated as MČ₁). Agyagási’s evidence for Volga Bulgarian loanwords in Permic relies mainly on two articles by Károly Rédei and András Róna-Tas (1972, 1983). In these articles, Volga Bulgarian loanwords are divided into two layers, Proto-Permic and ProtoUdmurt (Proto-Votyak). According to Agyagási, this chronological division is unfounded (pp. 110–112) and both the Proto-Permic and Proto-Udmurt loans originate from a specific Volga Bulgarian dialect (VB₁) that corresponds phonologically to the Late Old Bulgarian of Rédei and Róna-Tas. For example, Proto-Udmurt (Proto-Votyak/Ancient Votyak) *olma ‘apple’ (> ulmo), which is thought to be a borrowing from either Late Old Bulgarian *ålma or Middle Bulgarian *olma (Rédei & Róna-Tas 1983: 31), is in Agyagási’s view a loan from VB₁ *ålma (p. 165). Phonologically, there is no obvious reason to prefer LOB/ VB₁ *ålma over Middle Bulgarian (Proto-Chuvash) *olma as the source for Proto-Udmurt *olma. 176 A bridge too far – A Uralic perspective on Volga Bulgarian Agyagási postulates four Volga Bulgarian dialects based on how Proto-Turkic *o is reflected in Mari, Permic and Old Russian loanwords of Volga Bulgarian origin (pp. 122–126). In one of these supposed dialects, PT *o closed to VB *u, with examples including WOT *komdï ‘basket made of bark’ > VB *χundï → PP *kundi ‘id.’, WOT *bora ‘homemade beer’ > VB *bura → ProtoEast-Mari *pura ‘beer, homemade beer’ > E1 pura, Proto-West-Mari *pura > NW pŭra, VB *bŭraγ ‘domestic beer’ → Old Russian bъraga. The evidence is very fragmentary and there is a significant chronological difference between the postulated recipient languages, which makes it hard to believe that these examples constitute a chronologically uniform loanword layer. Even by conservative estimates, ProtoPermic probably predates ProtoMari by several centuries. As the number of Volga Bulgarian loans is lower in Komi-Zyrian than in Komi-Permyak and Udmurt, it is quite clear that Proto-Permic had begun to disperse or had already significantly dispersed geographically by the time of these contacts around the 9th and 10th century AD (Rédei & Róna-Tas 1983: 3–4). Phonologically speaking, Volga Bulgarian loanwords were adopted into a level of Proto-Permic that one might call Late Proto-Permic, as it had already undergone all of the major sound changes typical of the Permic languages. For some of the loans, even parallel borrowing into Proto-Komi and Proto-Udmurt is a possibility. Intensive contacts between Mari and Volga Bulgars cannot have occurred earlier than the 13th century (Bereczki 1994: 16). As some of the earlier Volga Bulgarian/Chuvash loanwords display sound changes common to all Mari dialects, it can be assumed with some certainty that they were borrowed into a unified proto-language at some point after the 13th century. Interpreting these words as reflecting VB *u also ignores the fact that Mari and perhaps also Proto-Permic were also subject to the *o > *u change (in some reconstructions of Proto-Permic vowel correspondence, Udm. u and Komi u are reconstructed as Proto-Permic *o instead of *u 1. For the most part, Agyagási’s abbreviations for the different Mari dialects are congruent with those used by Beke (1997–2001). This, although faithful to the original source, makes for laborious reading. For this reason, this article uses a simplified system of abbreviations that corresponds to the abbreviations used by Agyagási as follows: E = East (proper) = P B BJ BJp. M MK MM UP US USj.; M = Meadow/Central = CÜ UJ; NW = Northwest = JO V; Vo = Volga subdialect = CK Č ČN; W = West = K. 177 Sampsa Holopainen & Niklas Metsäranta (Zhivlov 2014: 123–124)). The closing of *o to *u could then easily be explained as an internal change in both of these languages. The preferred modus operandi throughout the book is to attribute phonological variation of any kind to different Volga Bulgarian dialects. This is not to say that some variation could not be interpreted as having resulted from dialectal variation within Volga Bulgarian, but competing ideas are not generally entertained. 5. No first-syllable reduced vowels in Proto-Mari? There are different views concerning the reconstruction of the Proto-Mari vowel system. The most relevant point of contention for Agyagási is whether or not reduced vowels can be reconstructed for Proto-Mari. There are those who argue in favour of reconstructing an opposition between first-syllable full (*i, *ü, and *u) and reduced close vowels (*ĭ, *ü̆ , *ŭ) (Itkonen 1954; Aikio 2014). No such opposition between full and reduced vowels is assumed by Bereczki (1994), who argues that first-syllable reduced vowels are a later, contact-induced phenomenon. Agyagási follows Bereczki in not reconstructing firstsyllable reduced vowels for ProtoMari. According to her, reduced labial vowels only emerged in the Volga, North-Western and JoškarOla dialect as a result of contacts with Middle Chuvash (pp. 293–298). There are several reasons for reconstructing an opposition between full and reduced close vowels in Proto-Mari. First of all, minimal pairs or semi-minimal pairs indicate that there was such an opposition: E M Vo NW W šur (< PM *šur ‘horn’) vs. E M šur, Vo šŭr, NW šŏr, W šə̑r (< PM *šŭr ‘shit’), W tul (< PM *tul ‘storm, stormwind’) vs. E M tul, Vo tŭl, NW tŏl, W tə̑l (< PM *tŭl ‘fire’) (Aikio 2014: 126–127). Second, the Proto-Mari full and reduced vowels have different origins: PM *ŭ is a reflex of PU *u (although the opposition between PU *u and *o has been neutralized adjacent to labial consonants), whereas PM *u is a reflex of PU *o (Aikio 2014: 130), PM *tŭl ‘fire’ < PU *tulə ‘id.’ Vs. PM *tul ‘storm, stormwind’ < PU *towlə ‘wind’. If one reconstructs invariably *u for both PM *u and *ŭ, one should have a good explanation for how the reduction process has systematically managed to avoid those instances of PM *u that reflect PU *o even if one does consider reduction to be a secondary phenomenon. The examples here are of PM *u and *ŭ, but mutatis mutandis the same is true for other close vowels. Agyagási does not address these shortcomings of reconstructing 178 A bridge too far – A Uralic perspective on Volga Bulgarian only full close vowels in detail, but she does seek to demonstrate that Proto-Mari lacked reduced vowels in the first syllable by listing a number of East Mari words of Volga Bulgarian origin where no reduced vowels appear in the first syllable (pp. 202–203). Examples include: VB₃ *kürük ‘fur’ → PM *kürük ~ *kürə̑k ‘id.’, VB₃ *külčün ‘loan’ → PM *küśün ~ küśə̑n ‘id.’, VB₃ *puruś ‘pepper’ → PM *puruś ~ purə̑ś ‘id.’, VB₃ *pus ‘misty, foggy’ → PM *pus ‘id.’, VB₃ *śul- ‘ransom, buy out’ → PM *sul- ‘id.’, VB₃ *sür- ‘sift, filter’ → PM *sür- ‘id.’, VB₃ *tuluχ ‘orphan; widow’ → PM *tuluk ~ *tulə̑k ‘id.’, VB₃ *χir ‘plain’ → PM *ir ‘id.’, VB₃ *χis- ‘press’ → PM *is- ‘id.’, VB₃ *χuntur ‘beaver’ → PM *unδur ~ *unδə̑r ~ umδə̑r ‘id.’ It is not clear what Agyagási’s criteria are for Proto-Mari. There are phonological reasons to assume that many of the examples were borrowed only after the dispersal of Proto-Mari. The Proto-Mari full front vowels *i and *ü were lowered to the mid-vowels *e and *ö before *r (Aikio 2014: 135–136), e.g. PU *närə > PM *nir ‘nose’ > E W ner. On top of this, there has been a tendency for Proto-Mari *i to change to e before sonorants in the eastern Mari varieties (Itkonen 1954: 219–221). As East Mari ir ‘wild (terrain); steppe, unforested area’ lacks both the ProtoMari and eastern Mari lowering, it cannot reflect Proto-Mari *ir and must have been borrowed into Mari only after these changes. The words that Agyagási reconstructs as PM *kürək ‘fur’ and *sür- ‘filter’ also lack lowering. If East Mari forms such as kürə̑k or šüre- actually reflected PM *ü, one would expect to find **körə̑k and **šöre- instead. Proto-Mari had two sibilants: *š (< PU *ś and *š) and *s (< PU *s). Proto-Mari *s changed to *š in all other dialects except for a number of eastern dialects, which seems to suggest that the sound change was and is still ongoing (Beke 1934: 90–92). The opposition between PM *š and *s is observed most consistently in the Malmyž area, where PM *s is reflected as s (adjacent to back vowels) and ś (adjacent to front vowels). The lack of PM *s > š (ž intervocalically) in most eastern varieties, however, is indicative of post-Proto-Mari origin. For this reason, words such as Birsk küśün ‘loan’ or sule- ‘ransom, buy out’ could have been borrowed only after PM *s > š, and the idea of this borrowing having taken place after the dispersal of Proto-Mari is all the more obvious when forms from other Mari varieties are included, i.e. Vo küsön, W küsən ‘loan’ (TschWb: 311), Vo sŭle-, NW sŏle-, W sə̑le- ‘ransom, buy out’ (TschWb: 642). A Proto-Mari form can be reconstructed for a few of the words, but based on the vowel correspond- 179 Sampsa Holopainen & Niklas Metsäranta ences there is no reason to reconstruct a Proto-Mari full vowel. Examples of such words are PM *pŭs > E M puš, Malmyž pus, Vo pŭš, NW pŏš, W pə̑š ‘steam, vapour’ (TschWb: 564), PM *sü̆ re- ‘sift, filter’ E M šüre-, Malmyž śüre-, Vo šü̆ re-, NW W šəre- (TschWb: 751) and PM *tŭlək ‘orphan; widow’ > E M tulə̑k, Vo tŭlŭk, NW tŏlŏk, W tə̑lə̑k (TschWb: 825). How eastern Mari varieties that lack first-syllable reduced labial vowels altogether constitutes evidence against ProtoMari reduced vowels in the first place is also unclear. All and all, the evidence is rather unconvincing as most of the examples can be shown to have been borrowed only after the dispersal of Proto-Mari. It seems that the main reason Agyagási is so keen to reject ProtoMari first-syllable reduced vowels, and maintains that the prominent syllabic structure in Proto-Mari was a first-syllable full vowel followed by a reduced vowel in the second syllable (V-V̆ ), is that she seeks to explain the appearance of second-syllable vowel reduction in Chuvash as partial code-copying from Mari (pp. 203–205). As this premise was shown to be untenable, the conclusions derived from it are untenable as well, and thus codecopying does not provide a solution for Chuvash vowel reduction in the way Agyagási envisioned. 6. West Baltic loans in Mari and Chuvash and the role of the Low Cheremis language? One of the most interesting and thought-provoking parts of the book is the treatment of possible Baltic loanwords in Chuvash and Mari (pp. 265–288). Agyagási suggests several new etymologies that she considers loanwords from a form of West Baltic, the branch that included Old Prussian and probably the poorly attested languages of Yotvingian, Galindian, whose existence is known from tribal names in Russian chronicles and hydronyms, and Curonian, which was spoken in Northern Latvia. Yotvingian and Galindian are thought to have been spoken in Central parts of European Russia. It is these languages that Agyagási considers the source of a group of words in Mari and Chuvash, and this is the reason she is determined that the loanwords are from West Baltic. However, Derksen (2015: 2–3) has noted that the knowledge of these only fragmentarily attested Baltic languages is very poor, as it is based on tribal names and loanwords alone, and that their classification as West Baltic is only tentative. Moreover, it seems that Agyagási’s etymologies do not show any actual West Baltic features, and her reconstructions are mostly based on Lithuanian, 180 A bridge too far – A Uralic perspective on Volga Bulgarian as she herself admits (p. 268). The loanwords from Mari and Chuvash thus offer no further evidence for the classification of Yotvingian and Galindian as West Baltic. Agyagási assumes that the words were borrowed into Mari and Chuvash through the unattested Low Cheremis language, with the borrowing taking place quite late (not earlier than the 16th century). Some forms in Mari and Chuvash contain elements that Agyagási regards as Low Cheremis derivational suffixes. As this alleged loanword layer is rather significant for Agyagási’s ethnohistorical claims, it is important to comment on it here in some detail. Unfortunately, many of Agyagási’s etymological suggestions are very complicated and involve various problems. Agyagási postulates some sound substitutions that are difficult to accept: in the Baltic etymologies *lẽk- ‘fly’ (Lithuanian lẽkti) → Mari E lǝ̑γe, liγe etc., Chuvash lĕkĕ (pp. 276–278) and *lẽpš‘wither’ (Lithuanian lẻpti) → Mari lǝ̑wǝ̑žγem, liwǝ̑žγem, etc., Chuvash lĕpešken- (pp. 278–279), Agyagási assumes that the *i she reconstructs for the Mari and Chuvash forms is a substitution of Baltic *e, but she does not account for why this sound substitution was used. Note that the Mari words reflect Proto-Mari *ĭ in Aikio’s (2014) reconstruction. This vowel could reflect Pre-Mari *e, so if the borrowing into Mari was old enough, the vowel substitution could be explained. But this does not fit with Agyagási’s ideas of the very late West Baltic influence in the Volga region and the late borrowing of the words into Mari through the hypothetical Low Cheremis idiom. A similar problem is the relation of Baltic *ṹ in *pṹčio‘blow up’ (Lithuanian pṹčioti) with *i (in Agyagási’s reconstruction) in Mari E pič́ , M pit ́ , Vo pǝ̑č́ , NW W pǝc ‘thick, dark; airless, stuffy, stifling’ and Chuvash pǎčǎ ‘stuffy’ (pp. 280–283). Agyagási notes that because of Baltic accentuation, the realization of the vowels might have led to these substitutions, but more substance would be needed to validate this argument. The semantic side of the etymologies is in some cases rather unconvincing: the connection between *lẽk- ‘fly’ with Mari lǝ̑γe, etc. and the meaning ‘dandruff, membrane’ is difficult to grasp (Agyagási assumes that the semantic connection is that dandruff flies off easily), Baltic *pṹčio- ‘blow up’ would have produced ‘stuffiness, stuffy, dark’ on the Mari and Chuvash side (pp. 280–283), and the Baltic verb *su-stó- ‘stop’ would have given rise to the nouns sustǝ̑k, śüstük, etc. ‘stammering’ in Mari (eastern dialects), and sŏstŏk 181 Sampsa Holopainen & Niklas Metsäranta ‘one whose speech is incoherent’ in Chuvash (Viryal) (pp. 283–285). If one wishes to argue in favour of these etymologies, semantic parallels should be provided. At present, the semantic developments remain quite hypothetical. Agyagási’s etymologies also feature non-existent derivational suffixes; it is easy to explain something as a loan if part of the donor or recipient form is explained ad hoc as a derivational suffix in an unattested language. According to Agyagási, Baltic *dub- (Lithuanian dub-ti ‘grow, become hollow, sunken, sink’) yields a derivative dubka- in the Low Cheremis language, which then is borrowed onwards into Mari as M tupka ‘a loose sheaf of hemp or flax’, W tǝ̑pka, etc. ‘combed wool; human hair’ and Chuvash as tăpka, topka ‘tuft, shred, splinter [пучок, клочок]’ (pp. 272–273; note that here the semantic difference is once again quite problematic). The borrowing from the Baltic verb *juos‘gird’ (> Lith. juósti) to Mari → E üštö, Vo üštǚ, W əštə etc. ‘belt’ proceeds through a Low Cheremis form where, according to Agyagási (p. 274), -t- is a derivational suffix. In this case, would a better source for -t- not be Baltic *juosta > Lithuanian júosta ‘woven sash; tape, band’? This form is also listed by Agyagási as one of the reflexes of the verbal root *juos- (the word is an old verbal adjective; see Derksen 2015 s.v. juosta). Another example of the dubious use of obscure suffixes is *kum̃ p- ‘bend’ (Lithuanian kum̃ pti) → Mari M kuptǝ̑rγe, W kǝ̑ptǝ̑rγe ‘(walk) slowly and crookedly [langsam und gekrümmt (gehen)]’ kǝ̑ptǝ̑rγem ‘become bent, bowed’, etc. and Chuvash kǎptǎrka‘grow old and weak’ (pp. 275–276), where only the stem *kup- would have been borrowed and the suffix *-tur- could be a possible Low Cheremis suffix. (Note that we do not consider the last etymology impossible, but the suffixal elements would require further investigation before the etymology could be accepted.) Some of the etymologies also involve various phonological difficulties. Baltic *juos- ‘gird’ → Mari üštö ‘Gürtel’ involves an ad hoc loss of *j- in Mari. In her treatment of the etymology *kump- → Mari kǝ̑ptǝ̑rγe, Chuvash kǎptǎrka (pp. 275–276), Agyagási discusses Mari denasalization and notes that the loss of m must have happened in the hypothetic intermediary language, as Mari retains clusters of a nasal and a stop. This is only partly correct: Aikio (2014: 83) has noted that Uralic/Pre-Mari *mp regularly develops into *w, and Metsäranta (2018: 123, footnote 3) further argues that in Uralic *-i-stems (= e-stems in UEW’s reconstruction), the 182 A bridge too far – A Uralic perspective on Volga Bulgarian regular reflex of *mp is Mari p at least in word-final position. The etymology of Baltic *popliaũ- ‘chat, gossip’ (Lithuanian pliaũkšti) → Mari W popaš ‘speak’, E popǝ̑lδatem ‘chat, gossip’, Vo popem ‘babble; speak [plappern; sprechen]’, Chuvash puple-, (Viryal) pople- ‘talk’ (pp. 281–283) seems unconvincing to us because it presumes that only the first part of the Baltic word (including the prefix *po-) was borrowed in an arbitrary way. The donor form is also entirely hypothetical, as a prefixed form *popliaũ- is not attested anywhere in Baltic. It is also difficult to understand why many modern sources of Baltic etymology are not referenced, as the recent years have seen the publication of several etymological dictionaries of the Baltic languages (Smoczyński 2005, 2007; AlEW; Derksen 2015), as well as Indo-European etymological dictionaries that would have been useful here (especially LIV). In some cases, these works would have supported Agyagási’s arguments. Although the etymologies cannot be treated here in more detail, we hope this illustrates that Agyagási’s conclusions about a West Baltic loanword layer in Mari and Chuvash are far from certain. While the existence of Baltic loanwords in Mari (and possibly in Chuvash too) remains an openended question, more research is certainly needed. It also seems that Agyagási’s remarks about the taxonomy of the Baltic languages and their relationship with Slavic (p. 268) do not reflect the most recent findings of research: Agyagási notes that Baltic separated from the “Old European language”, which is a misleading term, as “Old European” is usually used as the name of the hypothetical substrate language that produced various hydronyms in Europe (see Krahe 1963; Schmid 1968). Agyagási also argues that Proto-Slavic diverged from West Baltic, which is also a statement that does not reflect the communis opinio in Baltic historical linguistics (for recent discussions of the taxonomy of Baltic and its relationship to Slavic, see Petit (2010: 3–51) and Hill (2017)). Since there are various problems and uncertainties in the Baltic loanwords, it goes without saying that Agyagási’s (p. 287) reconstruction of Low Cheremis derivational morphology rests on shaky grounds. She claims to have reconstructed several derivational suffixes based on evidence from loanwords, but these results remain highly inconclusive at this point, as does the entire existence of the Low Cheremis language, or at least the argument that Low Cheremis was 183 Sampsa Holopainen & Niklas Metsäranta the meditator of Baltic loanwords. (Note that we do not wish to take a stance on whether Agyagási’s conclusions regarding the differences in Mari and the Cheremis of mediaeval chronicles are correct, but we leave this to specialists in the mediaeval history of the region to judge). 7. Concluding remarks One might consider this amount of criticism unfair, but we are not saying that there are no good sides to Agyagási’s book. As already mentioned, the author discusses the problems of Chuvash historical phonology from many points of view in a cross-scientific perspective, and she presents the sources clearly. Agyagási’s extra-linguistic ideas about the Mari ethnogenesis will also certainly give future researchers of Central Russian history a great deal of food for thought. As noted above, the book has shortcomings that make the conclusions uncertain. Nevertheless, the book is an interesting addition to Chuvash historical linguistics and the ethnic history of the Volga-Kama region, a field where modern, comprehensive contributions are few and far between. Sampsa Holopainen & Niklas Metsäranta References AlEW = Hock, Wolfgang & Fecht, Rainer & Feulner, Anna Helene & Hill, Eugen & Wodtko, Dagmar S. 2015. Altlitauisches Etymologisches Wörterbuch (Studien zur historisch-vergleichenden Sprachwissenschaft 7). Hamburg: Baar. Aikio 2012 = Luobbal Sámmol Sámmol Ánte (Ante Aikio). 2012. On Finnic long vowels, Samoyed vowel sequences, and Proto-Uralic *x. In Hyytiäinen, Tiina & Jalava, Lotta & Saarikivi, Janne & Sandman, Erika (eds.), Per Urales ad Orientem: Iter polyphonicum multilingue. Festskrift tillägnad Juha Janhunen på hans sextioårsdag den 12 februari 2012 (Mémoires de la Société FinnoOugrienne 264), 227–250. Helsinki: Société Finno-Ougrienne. Aikio 2014 = Luobbal Sámmol Sámmol Ánte (Ante Aikio). 2014. On the reconstruction of Proto-Mari vocalism. Journal of Language Relationship / Вопросы языкового родства 11. 125−157. Aikio 2015 = Luobbal Sámmol Sámmol Ánte (Ante Aikio). 2015. The Finnic ‘secondary e-stems’ and Proto-Uralic vocalism. Journal de la Société Finno-Ougrienne 95. 26–66. Beke, Ödön. 1934. Zur geschichte der finnisch-ugrischen s-laute. FinnischUgrische Forschungen 22. 90–122. Beke, Ödön. 1997–2001. Mari nyelvjárási szótár 1–9. Unter Mitarbeit von Zsófia M. Velenyák und József Erdődi, neu redigiert von Gábor Bereczki, bearbeitet von Margarita Kuznecova, herausgegeben von János Pusztay (Bibliotheca Ceremissica 4). Szombathely: Berzsenyi Dániel Tanárképző Főiskola. 184 A bridge too far – A Uralic perspective on Volga Bulgarian Bereczki, Gábor. 1994. Grundzüge der tscheremissischen Sprachgeschichte I (Studia Uralo-Altaica 35). Szeged: József Attila Tudományegyetem. Bereczki, Gábor. 2013. Etymologisches Wörterbuch des Tscheremissischen (Mari): Der einheimische Wortschatz. Nach dem Tode des Verfassers herausgegeben von Klára Agyagási und Eberhart Winkler (Veröffentlichungen des Societas Uralo-Altaica 86). Wiesbaden: Harrassowitz Verlag. Derksen, Rick. 2015. Etymological dictionary of the Baltic inherited lexicon (Leiden Indo-European Etymological Dictionary Series, vol. 13). Leiden & Boston: Brill. Häkkinen, Jaakko. 2009. Kantauralin ajoitus ja paikannus: perustelut puntarissa. Journal de la Société Finno-Ougrienne 92. 9–56. Hill, Eugen. 2017. Phonological evidence for a Proto-Baltic stage in the evolution of East and West Baltic. International Journal of Diachronic Linguistics and Linguistic Reconstruction 105. 205–232. Honti, László. 2010–2011. Personae ingratissimae? A 2. személyek jelölése az uráliban. Nyelvtudományi közlemények 107. 7–57. Itkonen, Erkki. 1954. Zur Geschichte des Vokalismus der ersten Silbe im Tscheremissischen und in den permischen Sprachen. Finnisch-Ugrische Forschungen 31(3). 149–345. Janhunen, Juha. 1981. Uralilaisen kantakielen sanastosta. Journal de la Société Finno-Ougrienne 77. 219–274. Krahe, Hans. 1963. Die Struktur der alteuropäischen Hydronymie. Wiesbaden: Steiner. LIV = Rix, Helmut. 2001. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstammbildungen. Bearbeitet von Martin Kümmel, Thomas Zehnder, Reiner Lipp und Brigitte Schirmer. Zweite, erweiterte und verbesserte Auflage, bearbeitet von Martin Kümmel und Helmut Rix. Wiesbaden: Dr. Ludvig Reichelt Verlag. Metsäranta, Niklas. 2018. Marin ja permiläisten kielten jakamista äänteenmuutoksista – aitoa konvergenssia vai kielikunnan sisäistä lainautumista? In Holopainen, Sampsa & Saarikivi, Janne (eds.), Perì orthótētos etýmōn: Uusiutuva uralilainen etymologia (Uralica Helsingiensia 11), 91–134. Helsinki: Société Finno-Ougrienne. Petit, Daniel. 2010. Untersuchungen zu der baltischen Sprachen. Leiden: Brill. Rahkonen, Pauli. 2013. The SouthEastern contact area of Finnic languages in the light of onomastics. University of Helsinki. (Doctoral dissertation.) Rédei, Károly & Róna-Tas, András. 1972. A permi nyelvek őspermi kori bolgár-török jövevényszavai. Nyelvtudományi közlemények 74. 281−298. Rédei, Károly & Róna-Tas, András. 1983. Early Bulgarian loanwords in the Permian languages. Acta Orientalia Academiae Scientiarum Hungaricae 37. 3−41. Róna-Tas, András & Berta, Árpád (with the assistance of Károly, László). 2011. West Old Turkic: Turkic loanwords in Hungarian. Wiesbaden: Harrassowitz. 185 Sampsa Holopainen & Niklas Metsäranta Salminen, Tapani. 2002. Problems in the taxonomy of the Uralic languages in the light of modern comparative studies. In Кузнецова, Ариадна Ивановна & Кибрик, Александр Евгеньевич & Агранат, Татьяна Борисовна & Казакевич, Ольга Анатольевна (eds.), Лингвистический беспредел: Сборник статей к 70-летию А. И. Кузнецовой, 44–55. Москва: Издательство Московского университета. Sammallahti, Pekka. 1988. Historical phonology of the Uralic languages with special reference to Samoyed, Ugric and Permic. In Sinor, Denis (ed.), The Uralic languages: Description, history and foreign influences (Handbuch der Orientalistik 8: Handbook of Uralic studies 1), 478–554. Leiden: E. J. Brill. Schmid, Wolfgang P. 1968. Alteuropäisch und indogermanisch. Wiesbaden: Steiner. Smoczyński, Wojciech. 2005. Lexikon der altpreussischen Verben. Innsbruck: Institut für Sprachen und Literaturen. Smoczyński, Wojciech. 2007. Słownik etymologiczny języka litewskiego. Wilno: Uniwersytet Wileński, Wydział Filologiczny. TschWb = Moisio, Arto & Saarinen, Sirkka. 2008. Tscheremissisches Wörterbuch. Aufgezeichnet von Volmari Porkka, Arvid Genetz, Yrjö Wichmann, Martti Räsänen, T. E. Uotila und Erkki Itkonen. (Lexica Societatis Fenno-Ugricae XXXII). Helsinki: Sociéte FinnoOugrienne & Kotimaisten kielten tutkimuskeskus. UEW = Redei, Károly. 1988. Uralisches Etymologisches Wörterbuch. Band I–II: Unter Mitarbeit von Marianne Bakró-Nagy, Sándor Csúcs, István Erdélyi †, László Honti, Éva Korenchy †, Éva K. Sal und Edit Vértes, Band III: Register Zusammengestellt von Attila Dobó und Éva Fancsaly. Wiesbaden: Harrassowitz. Zhivlov, Mikhail. 2014. Studies in Uralic vocalism III. Journal of Language Relationship / Вопросы яызыкого родства 12. 113−148. Zimonyi, István. 2014. A magyarság korai történetének sarokpontjai: Elméletek az újabb irodalom tükrében. Budapest: Balassi Kiadó. 186