Balto–Slavic: what Meillet was Thinking, or, what “was” Meillet “Thinking”?!

40 Acta Linguistica Lithuanica LXXVI BRIAN D. JOSEPH The Ohio State University Fields of research: Balto-Slavic controversy, history of linguistics, Meillet biography. BALTO-SLAVIC: WHAT MEILLET WAS THINKING, OR, WHAT WAS MEILLET THINKING?! Baltų ir slavų prokalbė: ką manė Meillet, arba ką Meillet spėjo?! ANNOTATION Antoine Meillet was as serious an Indo-Europeanist as there ever was, and yet not everything he wrote is uncontroversial. His take on Balto-Slavic, from Les dialectes indo-européens (1908, 2nd edn. 1922), is one such case – see Szemerényi 1957 – and specifically Meillet’s claim that there is no compelling evidence for a Balto-Slavic subgroup within Indo-European. I explore here just what Meillet meant by “‘dialect’ of Indo-European” in relation to Balto-Slavic, e.g. what gave rise to the 10 (or so) branches (branches as “dialects”) within the Indo-European family, or dialect variation within Proto-Indo-European itself. Further, in the 1922 “avant-prôpos”, Meillet refers to the Indo-European unity as “national” in nature, raising the question of the relevance of Meillet’s sense of the relationship between language and nation (Moret 2013) to the issue of a possible Balto-Slavic unity. KEYWORDS: Antoine Meillet, Balto-Slavic, dialect, history of linguistics, Indo-European, Italo-Celtic, language and nation, ANOTACIJA Antoine’as Meillet buvo vienas iš svarbiausių indoeuropeistų, tačiau kai kurie jo veikalai yra prieštaringi. Pavyzdžiui, jo nuomonė apie baltų ir slavų prokalbę knygoje Les dialectes indo-européens (1908, antrasis leidimas 1922) – žr. Szemerényi 1957 – ir ypač Meillet įsitikinimas, jog trūksta įtikinamų įrodymų, patvirtinančių baltų ir slavų prokalbės egzistavimą indoeuropiečių kalbų šeimoje. Šiame straipsnyje aiškinamasi, ką Meillet turėjo omenyje kalbėdamas apie „indoeuropiečių kalbų dialektą“baltų ir slavų prokalbės kontekste, pavyzdžiui, kas lėmė beveik dešimties šakų (šakas suvokiant kaip dialektus) atsiradimą Straipsniai / Articles 41 Balto-Slavic: what Meillet was Thinking, or, what was Meillet Thinking?! indoeuropiečių kalbų šeimoje, arba dialektų variacijas pačioje indoeuropiečių prokalbėje. Vėliau, 1922 m. leidimo pratarmėje, Meillet vadina indoeuropiečių kalbų vienovę tautine, pabrėždamas savo suvokimo apie kalbos ir tautos santykį (Moret 2013) svarbumą, nagrinėjant galimą bendros baltų ir slavų prokalbės egzistavimo klausimą. ESMINIAI ŽODŽIAI: Antoine’as Meillet, baltų ir slavų prokalbė, dialektas, lingvistikos istorija, indoeuropiečių, italų ir keltų prokalbė, kalba ir tauta. 1. INTRODUCTION Antoine Meillet (1866–1936), the great French historical linguist, is quite rightly remembered as one of the giants of Indo-European linguistics; he was a student of Ferdinand de Saussure and teacher of Émile Benveniste, so he is part of tradition of scholarship that is among the most important in this field, both in the past and into the present day.1 He was a prolific scholar, authoring some 50 or more books (some in multiple editions and translated into various other languages) and literally hundreds and hundreds of articles. He produced many standard, foundational works on various ancient Indo-European languages and language branches that have been taken seriously by Indo-Europeanists for decades and are still relevant, including the following, given with the dates of first editions: (1) a. Aperçu d’une histoire de la langue grecque (1913) b. Altarmenisches Elementarbuch (1913) c. Caractères généraux des langues germaniques (1917) d. Le slav commun (1924) e. Esquisse d’une histoire de la langue latine (1928) f. Dictionnaire étymologique de la langue latine (1932, with A. Ernout) g. La méthode comparative en linguistique historique (1925) h. Introduction  l’étude comparative des langues indo-européennes (1903) A further such work, the focus of the present study, is Les dialectes indo-européens (1908; 2nd edn. 1922), with an English translation, The Indo-European Dialects, produced by Samuel Rosenberg (1967). This last-mentioned work is especially important for showing that not everything Meillet wrote is uncontroversial and taken to be correct by most scholars, due to his position on Balto-Slavic. Balto-Slavic is a possible 1 On a personal note, because one of my mentors, Calvert Watkins, studied with Benveniste during an extended period in the 1950s when he was living in Paris, I am honored to be a part myself of this tradition. BRIAN D. JOSEPH 42 Acta Linguistica Lithuanica LXXVI subgrouping within Indo-European that has been controversial as to whether Baltic and Slavic constitute a legitimate subgroup or instead represent two distinct branches within the larger family. Meillet’s position on Balto-Slavic can be summarized as follows: (2) a. there is no compelling evidence for a Balto-Slavic subgroup within Indo-European, that is, for a Balto-Slavic “dialect” within the family b. all of the similarities between Baltic and Slavic that look like they might be significant shared innovations in fact are either inherited features from Proto-Indo-European or represent “deux développements parallèles, mais depuis longtemps autonomes” (“two parallel but long autonomous developments”). Several scholars, most notably Alfred Senn, have embraced Meillet’s position; see especially Senn 1966. Still, Meillet’s particular claim about the lack of evidence for a Balto-Slavic subgroup has come in for more than its share of criticism. Especially devastating is the critique to be found in Szemerényi 1957, where there is a careful assessment of each of Meillet’s points. Moreover, Szemerényi adds some points of his own, ultimately concluding that there are several nontrivial shared innovations that point to a prehistoric Balto-Slavic unity. Others have noted additional such innovations (e.g. Olander 2009). Most current scholarship seems to side with Szemerényi in accepting Balto-Slavic unity, though there are some scholars who follow Senn. While there are numerous interesting questions to consider that are purely linguistic in nature with regard to the Balto-Slavic controversy, my concern here is to explore, to the extent possible, the thinking that lay behind Meillet’s position on Balto-Slavic with regard to two key notions: (3) a. the sense of “‘dialect’ of Indo-European” that informed Meillet’s position on Balto-Slavic b. Meillet’s views on “nation” and how they may have intersected with his take on Balto-Slavic. Exploring the mind of a scholar is a tricky enterprise, but Szemerényi (1957: 103) opened the door to such an undertaking: A comparison with the criteria deemed sufficient to establish Italo-Celtic (: l’unité italo-celtique n’est pas douteuse, p. 12 of the 2nd ed. of Dial.) would throw interesting light on the scholar’s mind at work; the impression is strong that the admission of community in the one case, denial of it in the other, is not really based on the evidence, but some more remote (psychological) grounds. Straipsniai / Articles 43 Balto-Slavic: what Meillet was Thinking, or, what was Meillet Thinking?! I thus attempt to offer some insight into just what these “more remote (psychological) grounds” might be. I consider here just what notions of “‘dialect’ of Indo-European” – and of Proto-Indo-European itself – and of “nation” were at work in Meillet’s position on Balto-Slavic. In particular, it appears that there are some inconsistencies that could lead one to wonder just what he was thinking when he wrote what he did about Balto-Slavic. It may end up that there are more questions than answers but, as is often the case, just asking the questions in itself would seem to advance our understanding somewhat. 2. ON THE NOTION OF “INDO-EUROPEAN DIALECT” With regard to “dialect”, Meillet appears to use several different possible interpretations for the notion “dialecte indo-européen”/“Indo-European dialect”. First, there is the sense whereby it refers to the development of speech-forms/ speech-communities that led to the emergence of the 10 (or more or less depending on Italo-Celtic and Balto-Slavic) subgroups/branches within the Indo-European family; this view identifies branches as “dialects”, referring to divisions within Proto-Indo-European that ended up being recognizable entities in later times, after the breakup of the proto-language into the individual branches. Second, there is the sense whereby this notion refers to the diffusionary or “Wave-theory”-like spread of features, after that break-up of Proto-Indo-European, in ways that cut across the speech communities that led to the recognizable branches; this particular notion is highly relevant in that understanding these similarities contributes to an understanding of Indo-European dialectology, perhaps best by separating out “borrowings” from “shared innovations”. Finally, there is the sense whereby it refers to variation within the Proto-Indo-European language itself that is not linked to the groups giving ultimately identifiable subgroups/branches, but rather is more like the sort of variation associated today with quantitative variationist sociolinguistic studies of the Labovian type.2 2 Admittedly, a further question is how to reconcile innovation and diffusion with traditional notions of shared innovations. Meillet was not really in a position to say much regarding diffusion, as advances in our understanding of that area (e.g. à la Labov) have come only well after Meillet’s time, but he nonetheless had an idea about how diffusion works; still, there is inconsistency in that while he recognizes that no speech community is entirely uniform, it is also the case that innovations have to spread in order to be part of some sub-sector, i.e. “dialect”, within a speech BRIAN D. JOSEPH 44 Acta Linguistica Lithuanica LXXVI As noted, Meillet seems to utilize all three interpretations in his study on Indo-European dialects. That is, with regard to the first interpretation, the book itself includes discussion of different branches, e.g. at the level of Germanic or Indo-Iranian, and includes subgrouping that goes beyond the obvious, e.g. Italo-Celtic, and Meillet calls these “dialectes” (1928: 3). However, there is also discussion of features whose presence or absence crosscuts the 10 (or so) branches, such as the *o/*a merger, which he refers to as “lignes d’isoglosses” (“lines of isoglosses”), a term associated with the identification of dialects, or “lignes dialectales” (“dialectal lines”), a usage which associates with the second interpretation. Finally, at least in the case of the development of the syllabic sonants (see §3 below), which he calls “un fait dialectal de date indo-européenne” (“a dialectal fact of IE date”), the third interpretation would seem to be at issue. 3. MEILLET ON BALTO-SLAVIC Thus Meillet had three senses of “dialect” at play in his treatment of Indo-European dialects. It is fair to ask whether he can have it all three ways, whether all three are valid. The answer would seem to be affirmative, given what we know about the phenomena embodied in these senses of “Indo-European dialect”. For instance, Meillet talks about Indo-Iranian, Italo-Celtic, and Balto-Slavic; are they all of the same ilk? In a sense they are, in that if such subgroups are justified, each represents an intermediate node between Proto-Indo-European and the relevant attested languages. However, it must be admitted that in terms of evidence, they are qualitatively different. In particular, for Indo-Iranian, besides the linguistic evidence, there are numerous aspects of shared culture between Indic and Iranian, for instance commonalities in religion and mythology in the form of shared names of divine figures – e.g. the Vedic Sanskrit twin gods known as the Nsaty have a direct parallel in the Avestan demon3 name community. Thus one has to ask what, for Meillet, constituted a “common/shared innovation” as opposed to a feature that spreads (in a “contact”-like, diffusionary, way). 3 That the Vedic entities here are gods whereas the Avestan entity is a demon reflects the inversion of good and bad characteristic of the religious upheaval embodied in Zoroastrianism, the religion that Avestan is associated with. But the form of names like this and other religious terminology shows the common cultural heritage of Indic and Iranian. Straipsniai / Articles 45 Balto-Slavic: what Meillet was Thinking, or, what was Meillet Thinking?! Nŋhaiθya – and shared mythology.4 Such evidence of shared culture is lacking for Italic and Celtic, on the one hand, and for Baltic and Slavic, on the other. Moreover, there are numerous diffusionary developments within Indo-European that must be recognised. This is definitely the case in more recent times, as shown by the convergence of the Indo-European languages in the Balkan Sprachbund (cf. Sandfeld 1930), but it was also almost certainly so in ancient/ prehistoric times as well; the treatment of the centum/satəm split given in Hock & Joseph (1996/2009: 338–9) would be a case in point, with the diffusionary spread of satəm assibilation of the Proto-Indo-European palatal stop series *k’ *g’ *g’h. Finally, if we believe that proto-languages were real languages spoken by real people, a position most if not all practicing historical linguists would take, then we have to reckon with Labovian structured variation in these constructs; see Joseph 2006, 2013 for some relevant discussion. Meillet himself was well aware of this fact about language, as he wrote (1908: 1) “On ne rencontre nulle part l’unité linguistique complete” (“Nowhere is complete linguistic unity to be found”). Still, we can ask further, even if it is clear that Balto-Slavic, whatever it is, as a linguistic construct is different in some respects from Indo-Iranian, whether these senses of “dialect” are applied evenly. Relevant here is Meillet’s treatment of the syllabic sonants, where he goes to great lengths to deny that there is any sort of shared innovation between Baltic and Slavic, and in fact brings in proto-language variation. The issue is that the outcomes of syllabic sonants in the Indo-European branches are quite varied as to the quality of the vocalism they show, but with parallel vocalism in Baltic and in Slavic; for instance, Proto-Indo-European *ṛ ends up as ǝrǝ in Indo-Iranian, αρ in Greek, or in Latin, ur in Germanic, and ir in Baltic and ir in Slavic. Meillet’s position here is essentially to claim that there was a range of vocalism in Proto-Indo-European, saying (1908: 41–42) that “il semble que le timbre de la voyelle accessoire qui se joint à la sonante voyelle ait été fixé dès l’indo-européen, et qu’il y ait là un fait dialectal de date indo-européenne” (“it seems that the timbre of the prop vowel that attaches itself to the syllabic sonant had been fixed since (Proto-)Indo-European, and that in this case it was a dialectal fact of Indo-European age”). Meillet thus appears to be invoking what is tantamount to proto-language variation. Therefore, even though no other branches have exactly the Baltic and the Slavic outcome, a fact which especially in the light of the differences in vocalism seen in the outcomes in the other languages looks very much like a shared 4 See, for instance, such studies as Benveniste 1969, Watkins 1995, Mallory and Adams 1997, Frame 2009, among others. BRIAN D. JOSEPH 46 Acta Linguistica Lithuanica LXXVI innovation, Meillet locates the variable quality of the vocalism in variation that he felt was present in the Proto-Indo-European speech community. However, in the case of Italic and Celtic, Meillet takes seriously a particular outcome involving labiovelars that each of these branches shows. The specific development is the distant assimilation of *p…kw to kw … kw , seen in a handful of examples, most notably *penkwe > Latin quinque (= [kwinkwe]), Welsh pimp, Irish coic. Meillet takes this to be a shared innovation, and thus significant evidence in favor of Italo-Celtic unity, largely because it is unique and not found in any other Indo-European language.5 Moreover, he takes this view even though assimilations are natural6 and even though labiovelars are inherently unstable and are highly marked sounds that in almost all languages in the family except Anatolian end up being involved in some sort of changes.7 It therefore seems that something more was at work in Meillet’s assessment of Balto-Slavic. This matter is explored in the section that follows. 4. LANGUAGE AND NATION AND ITS RELEVANCE HERE It is at this point that Meillet’s ideas about language and nation may be relevant. In the “avant-prôpos” to the 1922 edition, Meillet refers to the Indo-European unity as “national” in nature (“l’unité indo-européenne était une unité nationale”), raising the question of the relevance of Meillet’s sense of the relationship between language and nation (on which see Caussat 1988, Moret 2013) to the issue of a possible Balto-Slavic unity.8 For the most part, Meillet’s works dealing with “nation” and language were written in the era around World War I – see Meillet 1918a, 1918b, 1919 – and dealt with contemporary matters. However, it is fair to assume that even if he did not confront them in writing until this era, these were matters that he had thought about prior to that time. He was a supreme student of language after all, and his interest in Armenian, which showed up in his writings on nation 5 Note that Germanic dealt with ‘five’ differently, with *penkwe > *pempe (as shown by Gothic fimf). 6 One can even view the assimilation here as a way of lessening the markedness of the sounds as singleton segments in a word, thereby enhancing the naturalness of this change. 7 It may well be that had Anatolian languages survived past the first millennium BC, their labiovelars might have undergone some change or other. Mycenaean Greek attests labiovelars, but in Greek of the first millennium BC, labiovelars underwent significant changes. 8 I thank Michiel de Vaan for bringing the Moret article to my attention. Straipsniai / Articles 47 Balto-Slavic: what Meillet was Thinking, or, what was Meillet Thinking?! and language, predated the war years by at least a decade. Moreover, one might extrapolate from his ideas about “nation” in what for him was the present and apply them to his use of similar terms in the distant past. In that regard, this quote from Meillet (1918a: 8) is particularly telling: Il y a nation là ou un ensemble d’hommes a le sentiment et la volonté de former un groupe à part, ayant ses traditions, ses usages et ses aspirations d’avenir. Rien ne marque plus nettement l’existence d’une nation que la possession d’une langue qui lui soit propre. One has a nation where a collection of people has the feeling and the desire to form a separate group, having its (own) traditions, its (own) uses, and its (own) future aspirations. Nothing marks the existence of a nation more neatly than possessing a language which is proper to it. Further, it seems to be especially significant that the quote about Indo-European unity being “national” in nature comes in the second edition of Les dialectes indo-européennes, which came out in 1922, that is after the turmoil of World War One and the reconfiguration of the map of Europe that emerged in its wake and in the wake of the dismantling of two empires, the Austro-Hungarian Empire and the Ottoman Empire. Clearly, the notion of “nation” was on Meillet’s mind in that post-war period, as his writings show, and thus it is not surprising that a statement of that sort would work its way into his commentary (via his Preface) to his book on how language individuation took place within Indo-European. The question to ask now is how these considerations apply to Balto-Slavic and to statements made a decade and a half earlier in the first edition of the book. Can it be assumed that Meillet had thought about the notions of nation and language well before he wrote about them overtly? If so, then whatever uneasiness he might have felt about Balto-Slavic on linguistic grounds – an uneasiness which, while perhaps not fully justified, would nonetheless have been enhanced by the fact that Baltic and Slavic are not as obviously connected as Indic and Iranian are9 – most likely resonated with the absence of any sort of shared culture that might be taken as collective “traditions … usages … aspirations”. And, the difference between Meillet vis-à-vis Balto-Slavic and Meillet vis-à-vis Italo-Celtic, even though there are also no signs of shared culture in this latter group, is that Meillet began as a Classicist and had an especially deep 9 Note in this regard Szemerényi (1959: 103): “Meillet’s thesis [was] that the Aryan unity was obvious but the BS not”. BRIAN D. JOSEPH 48 Acta Linguistica Lithuanica LXXVI knowledge of Latin and Italic (note his 1928 and 1932 works on Latin in the list in (1)), enabling him to see a connection with Celtic at a rather deep chronological level. 5. CONCLUSION To conclude, a close reading of the way in which Meillet talks about Balto-Slavic together with indications from other writings of his offers a basis for some reasonable inferences as to the thinking behind his interpretation of the facts concerning similar features found in both Baltic and Slavic. Balto-Slavic thus seems to represent an area of Meillet’s scholarship that is far from the clearest of the many otherwise very sharp and acute set of insights into Indo-European that he put forth. And that is why I wonder, as in my title, just what was Meillet thinking, letting nonlinguistic sentiments get in the way of his linguistic assessments. REFERENCES Benveniste Émile 1969: Le vocabulaire des institutions indo-européennes. Paris: Les Éditions de Minuit. Caussat Pierre 1988: Langue et nation. – Histoire Épistémologie Langage 10 (2),19 5 –2 0 4 . Frame Douglas 2009: Hippota Nestor (Hellenic Studies Series 37). Washington, DC: Center for Hellenic Studies (http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Frame. Hippota_Nestor.2009). Hock Hans Henrich, Brian D. Joseph 19961/20092: Language Change, Language History, and Language Relationship. An Introduction to Historical Linguistics. Berlin: Mouton de Gruyter. Joseph Brian D. 2006: On Projecting Variation Back into a Proto-Language, with Particular Attention to Germanic Evidence and Some Thoughts on ‘Drift’. – Variation and Reconstruction, ed. by Thomas Cravens. Amsterdam: Benjamins, 103–118. Joseph Brian D. 2013: Demystifying ‘Drift’ – A Variationist Account. – Shared Grammaticalization, with Special Focus on the Transeurasian Languages, ed. by Lars Johanson, Martine Robbeets, and Hubert Cuyckens. Amsterdam: John Benjamins, 43–66. Mallory James P., Adams Douglas Q. 1997: Encyclopedia of Indo-European Culture. London: Fitzroy Dearborn Publishers.