The Balkan peninsula is home to a number of languages whose interrelationships present various elements of interest for questions of language contact, language change, areal linguistics, and sociolinguistics. Between 1975 and 2018, no fewer than thirteen book-length general treatments of the Balkan languages were written: Schaller 1975; Solta 1980; Banfi 1985; Feuillet 1986; Desnickaja & Tolstoj 1990, 1998; Reiter 1994; Steinke & Vraciu 1999; Hinrichs 1999a; Asenova 2002; Sh. Demiraj 2004; Feuillet 2012; and Fiedler 2018.Footnote ³ All of these works build in some way on the classic 1926 study by Kristian Sandfeld, better known in its 1930 French version, whose synthesis of the previous century’s research combined with his own analyses is considered the founding work of the modern field of Balkan linguistics – the subject of this book. By Balkan linguistics we mean the study of the special relationship that obtains among the languages of Southeast Europe that form the Balkan linguistic league.Footnote ⁴ It was the grammatical convergences that characterize the Balkan languages that led Trubetzkoy 1923, 1930 to theorize, for the first time in modern, explicit terms, the ways that linguistic structure can change not only owing to divergence but also as a result of convergence in multilingual contact, and that therefore in addition to the concept of language family – a group of languages whose resemblances result from descent from an earlier common form – there is also the linguistic league (in German, Sprachbund), i.e., a group of languages whose resemblances result from mutual influence.Footnote ⁵

Despite the debate that emerged after 1989 around the use and definition of the term Balkan (cf., e.g., Bakić-Hayden & Hayden 1992; Bakić-Hayden 1995; Todorova 1997; Aronson 2007; and cf. also Friedman 1997a: 36, 2000a: 142, 2011a; see Račeva 1996 for an interesting etymological note), there are four living Indo-European language groups which, having come into contact for centuries or even millennia in Southeastern Europe, are universally recognized as containing members that constitute the classic Balkan sprachbund: Balkan Albanic, Balkan Hellenic, Balkan Romance, and Balkan Slavic (see Chapter 1, passim). We use the adjective Balkan in each case, since there are members of each of these groups spoken outside the Balkans.Footnote ⁶ To these four can be added Balkan Turkic, which is generally treated as a contributor to but not a participant in the Balkan sprachbund (but cf. Friedman 1982c, 2003a: 50–83), and Balkan Indic, i.e., the Romani dialects of the Balkans, which, until Hinrichs 1999a, were never accorded more than passing mention in the general handbooks and were rarely treated in specialized studies (but see to the contrary, e.g., Gilliat-Smith 1915/1916; Kostov 1973, 1998; Uhlik 1973; Joseph 1983a: 252–253; Friedman 1985a, 2000b, 2000c; Matras 1994a; Boretzky 1995, 1998). We can also mention here Balkan Judezmo (sometimes called Ladino, Judeo-Spanish, etc.; here we accept the judgment of Bunis 2018, cf. also Harris 1982; Yenisoy 2015), which is Romance but generally not treated with Balkan Romance languages (but cf. Joseph 1983a: 252–253; Kowallik 1992/1993; Gabinskij 1992: 154–173, 1996, 1998; Desnickaja & Tolstoj 1998; Friedman 2006a, 2021; Friedman & Joseph 2014, see §1.2.3.3).Footnote ⁷ We discuss the specifics of our definition of language and of Balkan language in §1.2.

In the first chapter we also give a general survey of the relevant aspects of Balkan geopolitical, cultural, and linguistic history and discuss other extralinguistic factors relevant for the linguistic situation. In Chapter 2, we present a historical survey of the field of Balkan linguistics, and in Chapter 3 we discuss the key methodological and theoretical issues relevant for Balkan linguistics as a specific manifestation of complex language contact. Of particular significance here are studies of other proposed linguistic areas such as Amazonia (Aikhenvald 1999, 2002), Araxes-Iran (Stilo 2018, 2023), the Caucasus (Catford 1977; Klimov 1978; cf. also Friedman 1996a; Tuite 1999; Comrie 2008), Ethiopia (Ferguson 1970; Zaborski 2010 and references therein, but see also Tosco 2000), Mainland Southeast Asia (Enfield 2003, 2005, 2017), Meso-America (Campbell et al. 1986; Suárez 1983), the Northwest Coast of North America (Beck 2000), South Asia (Emeneau 1980; Gumperz & Wilson 1971; Masica 1976, 2001), parts of Australia (Dixon 2002: 668–689; Heath 1978: 140–143) and of Papua New Guinea (Ross 1996, 2001).Footnote ⁸

The Balkans represent not only the most studied such case but also the most studiable, in that of all the sprachbunds that have been discussed in the literature, the Balkans offer the greatest amount of and the longest time-depth for information on the linguistic history of the area, the social history of the peoples in the area, and relevant reconstructible linguistic prehistory. Regarding this last point, we note the considerable evidence for earlier linguistic stages that comes from the attestations in texts and inscriptions for the ancestral languages of four out of the five Indo-European groups (Ancient Greek for Hellenic, Latin for Romance, Old Church Slavonic for Slavic, and Sanskrit for Indic), leaving only the ancestor of Albanian as problematic and without any textual attestations above the level of the word (but cf. Hamp 1982). Further, the Comparative Method offers another point of entry for Balkan linguistic prehistory, working with evidence from related languages elsewhere in the families.

These three chapters together constitute the first part of the book. The second part, comprising Chapters 4 through 7, treats Balkan sprachbund phenomena in the lexicon, phonology, morphology and morphosyntax, and syntax, respectively. Pragmatics, and semantics are treated, where appropriate, throughout these chapters, and especially in Chapter 4 for lexical semantics. In all of these chapters, we cover both features widely discussed in the literature and some brought to light for the first time here or only given brief mention elsewhere. Although, as we noted previously, it was convergences in grammatical structure that led to the theorization of the Balkan sprachbund as a linguistic entity whose origins are due to contact-induced change, convergences at other levels are also relevant to a comprehensive study of the field. It is also worth noting here that divergences – and particularly superficial convergences that at a deeper level represent divergent structures (cf. Friedman 1978, 1983; Sobolev et al. 2018) – are equally relevant in a comprehensive treatment of a linguistic area such as the Balkans, and they are noted where relevant.

A major issue in understanding the Balkan sprachbund is causation. The causes of convergence and divergence in the Balkans are discussed throughout the presentation of particular features, regarding language contact and other factors, including sociolinguistic phenomena such as multilingualism (both productive and passive, i.e., receptive; see §3.2.1.1), codeswitching, the roles of politically and culturally dominant or prestigious and also subordinate languages, i.e., superstrate, substrate, and adstrate, also acrolectal, mesolectal, and basilectal factors and structural mechanisms (the role of prosody, selection among competing features, internal motivation versus borrowing). The final chapter, Chapter 8, provides a summation of the main issues pertaining to causation, and offers a careful assessment of the sort of construct the Balkan languages represent, by way of problematizing the notion of “sprachbund.”Footnote ⁹

Given the large number of surveys of Balkan linguistics cited earlier, the reader might question the justification for another work on this topic other than the fact no such work has ever been available in English. There are, however, a number of reasons why the present book proposes to contribute something new to the ongoing study of the Balkan languages.

To begin with, we can note that certain specific contact phenomena in the Balkan languages, commonly referred to as Balkanisms, have not been exhaustively investigated or have been described in scattered places. Thus, for example, observations by Joseph 2001b on negation and Friedman 2003c, 2014a on indefinite marking and the narrative imperative, as well as other data, have yet to find their way into general accounts of Balkan linguistics but are given voice here. We can also include here more local phenomena, e.g., the Albanian-based admirative construction in the Fărshălots (Frasheriote) Aromanian dialect of Bela di Suprã (Mac Gorna Belica, Friedman 1994b) or the Aromanian-influenced marking of direct objects with prepositions in southwestern Macedonian (Koneski et al. 1968; Koneski 1981), which are also generally absent from the handbooks (see §3.5 and §6.1.1.1.2).

Another reason for this volume is the emergence of contact linguistics, to which areal linguistics can be construed as a contributing field, as an increasingly popular focus of scholarly inquiry (of the several handbooks on these fields, Mufwene & Escobar 2022 appeared just as this book was going to press). From the time of Sandfeld and Trubetzkoy to the present, the majority of Balkan linguistic handbooks have remained within the traditional paradigm as it was first defined. This paradigm, which basically concentrates on a combination of specific morphosyntactic, lexical, and, rarely, phonological, features, lends itself to a kind of listing approach (cf. Lindstedt 2000; van der Auwera 1998; Campbell et al. 1986), which has the disadvantages of completely eliding the social mechanisms of linguistic change and failing to address how the relevant elements fit into their respective languages, whose systems often show underlying divergences masked by surface convergences (cf. Friedman 1978, 1983). This last point is itself an important problem in the study of the nature of language contact. Of all the handbooks that have appeared since Sandfeld’s important works of 1926 and 1930, Asenova 2002 represents the most substantive advance within the traditional paradigm. Those works that have attempted different approaches, while useful and interesting, have either been unintegrated collections of articles on specific languages or topics (Desnickaja & Tolstoj 1990, 1998; Hinrichs 1999a; Sobolev 2021c) or have basically abandoned the areal model (Reiter 1994).

Thomason & Kaufman 1988 took the areal model a step further in uniting various types of contact-induced language change (including creolization and pidginization) into a single model of contact linguistics in opposition to the genetic or genealogical model. Their explicit unwillingness to treat multilingual contact situations in general (Thomason & Kaufman 1988: 95), however, and the Balkans – which they call “the world’s most famous contact situation” – in particular, while understandable within the parameters of what they were trying to achieve, nevertheless leaves the phenomenon uninterrogated. Having observed that “Sprachbund situations are notoriously messy” (Thomason & Kaufman 1988: 95–96), they devote only a paragraph and two or three comments in passing to Balkan phenomena. In subsequent work on contact linguistics, the Balkans have been similarly cited as well known and then passed over. Typical in this respect is Myers-Scotton’s 2002: 174 paragraph on the Balkans that simply lists ten features culled from Hock 1988: 286–290, some diagnostically Balkan, some not.Footnote ¹⁰

The present study of the Balkan linguistic league, building as it does both on the achievements of traditional Balkan linguistics and on modern approaches to language contact in its social setting, aims at new answers to questions of how language contact works in complex multilingual situations in general and in the Balkans in particular. In addition to the importance of contact phenomena in shaping the structures of the Balkan languages, however, we also address the question of limitations, i.e., those portions of the grammar that are resistant to contact-induced change or where apparent surface resemblances mask differences at other levels (cf. Friedman 1978, 1983, 2001c). In so doing, unlike various scholars (e.g., Andriotis & Kourmoulis 1968; A. Belić 1936; Birnbaum 1965, 1966; Minčeva 1987: 139–149; S. Mladenov 1939; Reiter 1991, 1994; Steinke 1983), we are not attempting to minimize the importance of contact, but rather to contribute to the discussion of the nature of constraints (cf., e.g., Myers-Scotton 1993a,b, 2002, but to the contrary Thomason & Kaufman 1988; Friedman 2007a).

Beyond modern approaches to contact linguistics, which, as we have noted, have eschewed dealing with the complexities of Balkan multilingual contact, the Balkan languages and the interesting linguistic phenomena they present are mentioned in other works on general linguistics, but often with erroneous and even comical results, as with Hopper & Traugott’s 1993: 210 (also 2003: 213) reference to Sandberg [sic] 1930, or Croft’s 2003 reference to articles and definiteness in the Balkans that is stated in such a way that Greek, in the Balkans itself, is a counter-example to his generalization. Similarly, many general linguistic surveys cite Balkan examples, but by their very nature such surveys cannot engage in the type of detailed analysis that gives a nuanced picture; the result is a presentation of standard, often prescriptive examples that fail to capture the full range of complexity that the facts demand.Footnote ¹¹ And, there are general works as well that devote (necessarily) limited discussion to the Balkan sprachbund per se, but not always with satisfactory results; such works are commented upon, with corrections offered, where relevant. The present book therefore seeks, among other things, to bring correct versions of the facts about the Balkan languages to the general linguistic public by way of setting the record straight but also making available in as accurate a manner as possible the interesting facts from the Balkans that can inform and enlighten more general matters in linguistics.

The relevance of linguistic contact for the history of specific languages is often elided or minimalized in the linguistic equivalents of national histories (cf. K. Brown 2015).Footnote ¹² Such works attempt to ignore or explain away the possibility that multilingualism could be responsible for change in their respective national languages (cf., e.g., Kazazis 1977 on Greek denials of Turkisms in Greek). This approach is epitomized in Andriotis & Kourmoulis’ 1968: 30 conclusion to their plenary paper at the first international Balkan Congress in 1966:

Such ideologically determined approaches to language contact (cf. Friedman 1997a) are also connected with other – often interrelated – issues of linguistic social history such as language planning (e.g., Friedman 1989a, 1993b), language shift (e.g., Tsitsipis 1998), linguistic purism (e.g., Friedman 1989b; Heyd 1954; Lewis 2000), and linguistic nationalism (e.g., Friedman 1986b; Kazandžiev 1943; Joseph 1985a). These are issues that typically have not been addressed in the usual handbooks on Balkan linguistics but are of enormous relevance to a thorough account of the field.

Similarly, more detailed discussion of relevant languages other than those that are routinely treated in the standard handbooks, such as Romani, Judezmo, and Balkan Turkic, as noted above, is sorely needed.Footnote ¹³ Of the various handbooks cited above, only Desnickaja & Tolstoj 1998 treats Judezmo and only Hinrichs 1999a treats Romani and Balkan Turkish (but not Gagauz), but neither of these integrates these languages into a unified Balkan linguistic account.Footnote ¹⁴ Moreover, in the period since the emergence of the modern Balkan standard languages, their relation to dialectology and to competition among diasystems (cf. Ivić 1962) has only rarely been addressed. The overwhelming majority of publications in Balkan linguistics in the past fifty years have been based almost exclusively on standard languages (Sawicka 1997, 2014, 2023 are notable exceptions to the general trend). As Heřman 1968: 342 observes, however, “[a] historical study of Balkan convergences is possible only on the basis of both diachronic and synchronic dialectology.” We rectify this situation by integrating the under-studied and under-described languages of the sprachbund into our account, and by giving consideration to regional and other (e.g., social) dialects as needed.

After seven decades of calls for a Balkan linguistic atlas (Anagnostopoulos 1935: 273; Małecki 1935: 76; Deanović 1961; Hamp 1989a; Friedman 2000d), an international group of Balkanists attempted finally to address the task through a pilot project begun at the University of St. Petersburg in 1996, the Malyj dialektologičeskij atlas balkanskix jazykov ‘Small dialectological atlas of the Balkan languages’ (Domoselickaja 1997; Sobolev 1997, 1998b, 1999; Plotnikova 1996; Rusakov & Sobolev 2003). This project, however, pursued the more modest goal of working to produce an atlas that would cover selected, representative dialects in depth, rather than attempting the Herculean task of surveying all the dialects of all the languages of the region. For this purpose, a total of twelve points were chosen, seven Slavic points (three in Bulgaria, one each in Croatia, Montenegro, and Serbia, and North Macedonia), two Albanian, two Greek, and one Aromanian (in Greece).Footnote ¹⁵ Moreover, since the focus of the project is precisely areal, it examines the dialects in terms of those categories that are of greatest relevance to language contact: syntax and lexicon (see Sobolev 2001a). We can also mention here the Romani Morpho-Syntax Database at the University of Manchester (Matras & Elšík 2001–2005, 2008, hereafter and listed in the references under RMS) and the Romani Lexical Database at the University of Graz (Halwachs et al. 2006, hereafter and listed in the references under ROMLEX), which offer coverage in these same areas for Romani dialects.Footnote ¹⁶ Material from these projects also informs our work.

In addition to examining neglected speech forms, we also devote attention to subdisciplinary approaches that have not received adequate treatment in a Balkan context. For instance the study of the role of pragmatics as the entry point for structural change (cf. Mithun 1992; Prince 1988; Friedman 1994a) as well as more modern approaches to syntax are other areas addressed here for the first time in a general Balkan linguistic handbook.

Finally, the social setting of the Balkan languages and the relationship of sociolinguistic issues in general to the field of Balkan linguistics have only relatively recently begun to be explicitly addressed in the handbooks (Hinrichs 1999a), although Morozova & Rusakov 2018a, 2018b; Sobolev et al. 2018 and the articles in Sobolev 2021c are all relevant here. The present work is therefore more than a description of the structural convergences and divergences of the Balkan languages; it is also a history of the social implications of both changes and conservatisms.

We thus invite the reader to use this study as a vehicle for achieving the goal set out in the epigram at the head of this section, but also for entering into this fascinating – and instructive – part of the (linguistic) world.