26.1 Introduction

26.2 Variation with Respect to the Language Standardization Model

Variation with respect to the language standardization model implies the existence of levels of sociolinguistic variation which can be presented in terms of their ‘peripheral’ relation to the language standard. In other words, the nature of these levels relies on their linguistic deviation from the ‘standard language’, the ‘standard’ being understood as a set of criteria for selecting the correct language expression in the society. In many Slavic languages, definitely in the two languages described here, these so-called standard varieties are privileged over the other regional and social varieties. Having in mind that the notion of the language standard does not present one homogeneous whole, it is understandable why this type of variation correlates to social and language stratification. The most common types of variation in Slavic languages are: (1) regional (division according to various forms of regional and/or vernacular language use); (2) social (which implies that different social levels of speakers share common linguistic markers); (3) temporal (division according to the time of language use); and (4) idiolectal (which puts different kinds of ‘incorrect’ language use related to the sense of identity of the speaker(s)). However, this division should be taken conditionally, given that, linguistically, these types of variation do not form isolated wholes, but rather overflow into each other.

26.2.3 Temporal Variation

26.2.3.1 Descriptive Studies

In relation to the vertical, time-related dimension of the ‘standard language’, there are layers which can be characterized as ‘archaic’ or ‘obsolete’ and ‘new’.

The ‘archaic’ layer: lexis which belonged to the ‘higher’ styles in the history of Serbian literary language develops an anachronistic relation to the language standard (in Serbian, usually from the Slaveno-Serbian period: otvečanije ‘response’, vinodelije ‘winemaking’); today it can be used with highly expressive function.

The ‘obsolete’ layer: lexis, idioms, and meanings obsolete in relation to better choices (plav (adj. ‘blue’) in the meaning of koji ima svetliju nijansu, svetao (‘lighter’): ‘yellow can be more or less blue’, etc.); it can be mixed with the dialect (in obsolete dialectal words).

The ‘new’ layer can be attributed both to lexis and its meanings (e.g. new lexical borrowings (peč (patch), strimovati (to stream) …); in the standardizational sense, it signifies that the word is not yet accepted.

Variation with respect to the time of language use (diachronic level of sociolinguistic variation): there is also a model, to which this use is related. The model represents the ‘modern’ or ‘present-day’ usage of the standard language. In relation to this model, language can be ‘archaic’ or ‘new’ (with further subclassification in both levels, e.g. ‘archaic’ and ‘recently obsolete’). Differences can be established on the levels of word formation, word origin, syntax, and so on.

26.2.3.2 Corpus Evidence

The primary source of corpus evidence comes either from historical corpora, such as the Russian National Corpus (Reference Sharoff, Archer, Wilson and RaysonSharoff 2005) or from social media corpora annotated with author profiles, such as GICR (Reference Belikov, Kopylov, Selegey and SharoffBelikov et al. 2014). Historical corpora are suitable for detecting variation over large time intervals, while at the same time they are limited by the availability of their sources; in particular, fiction and legal texts are the main sources in the historical part of the RNC, as the language of formal writing can be preserved much better, but it reflects only a small portion of the total language use. This language is also very much influenced by the gatekeepers at a specific time.

On the other hand, social media have made it possible to capture the language of spontaneous interaction in everyday life. At the same time they are limited with respect to the relatively recent time period. Any sizeable social media collections are available from 2000s and their authors are mostly aged from 18 to 60; see the distribution of the number of blog posts for the authors who have explicitly provided their year of birth in Figure 26.1. A number of data points in a study of this kind can be expanded further via automatic age prediction for the authors who have not provided their age (Reference Nguyen, Doğruöz, Rosé and de JongNguyen et al. 2016), but this has not been attempted for Slavic languages yet.Footnote ¹

Figure 26.1 The number of blog authors in GICR with respect to the year of their birth

Another limitation of relying on corpora is that they often exhibit topical biases leading to predictable prevalence of specific topics, such as dating and education for the authors younger than 22 or history and illnesses for those older than 55. Nevertheless, corpora allow detection of more interesting patterns, in particular through associating grammatical properties with age differences. This can be done by building a statistical model predicting the author’s age on the basis of selected lexicogrammatical features, such as those from (Reference BiberBiber 1988). For example, a linear Support Vector Regression model (Reference Cherkassky and MaCherkassky & Ma 2004) can predict the author’s age in GICR using the grammatical features with the mean absolute error of 8.5 years. This model on the basis of modern GICR data for Russian associates higher positive weights (leading to positive correlation with age) for such features as the rate of third person pronouns, hedges, amplifiers and nominalizations, while higher negative weights (leading to negative correlation with age) are exhibited by such features as the rate of the first person pronouns, relative subordinate clauses, and omissions of čto (‘that’) after verbs of reporting, such as soglasit’sja ‘agree’, utverždat’ ‘claim’. This suggests greater acceptance of more personal, less formal communication in the blog posts of younger writers.

26.3 Variation with Respect to Communicative Functions

26.4 Variation with Respect to Communication Media

There is also an aspect of variation which is not often addressed in traditional accounts of sociolinguistic variation. Digital communication, in addition to providing a better window into sociolinguistic research (such as the GICR corpus discussed above), has also resulted in changes in the use of language. The relevant technological innovations on the Web include all forms of user-generated content, such as Wikipedia or review-sharing websites, as well as social media. They all offer more opportunities for individual contributions, with a relatively low threshold for participation in comparison to traditional communication and distribution channels. At the same time, the new affordances are first developed and then accepted by a small proportion of the population, primarily within the younger, more affluent, and better educated groups. In the case of Russian this has led (among other things) to the much greater use of both borrowing and calques from English, for example, sajt (‘website’), klik (‘click’), and partnerki (‘partnerships’). A little later some of the innovations gain enough popularity to lead to their standardization, for example, sajt belongs to the top 500 words in the general corpus of ruWac, as this word can be used in the most formal discourse contexts: Materialy konferencii budut razmeščeny na oficial’nom sajte konferencii v vide èlektronnogo sbornika i razmeščeny v RINC (‘The conference materials will be published on the official website of the conference as an electronic publication and will be included in the Russian Science Citation Index’).

Because of the difference in the alphabet between English and Russian, expression of some borrowings can be made easier by combining the Cyrillic and Latin characters when the more difficult combinations of Latin characters are copied, thus creating a mix of spellings, for example, call-центр (the call center), web-сайт (the website), and Android’ом (with Android).

The new ways of using language also lead to the emergence of new cultural phenomena, which can start their relatively unhindered development in the absence of gatekeepers (at least at the initial stage of their development). For example, many orthographic features have been borrowed into Russian social media from English, such as an extensive use of smileys, punctuation combining exclamation and question marks (!?!?), repeated characters (Daaaa!!, ‘yeeees!!’), and texts written in the upper case or with strikethroughs (Reference Piperski and SominPiperski & Somin 2013).

On the Russian Internet, there is also a case of creative spelling in the subculture known as padonki (lit. ‘scoundrels’ with the first vowel misspelled, o → a), also known as Olbanskij yazyk (Olbanian language, this time misspelling A → O). This is related to, though not directly borrowed from, the cacography tradition of making deliberate spelling or writing mistakes, usually with the purpose of mockery. What is specific to the Russian subculture case is a very extensive development of rules for violating the expected norms of Russian spelling, such as spelling a as o and vice versa, turning v to ff (for example, podonkov → padonkoff, ‘scoundrels’ in plural genitive), etc., often with the aim of expressing a phonetically similar string in a way which is maximally different from the accepted orthographic norms (Reference KrongauzKrongauz 2013).

27.1 Introduction and Definitions

The term false cognate refers to pairs of words in two languages or language varieties (such as dialects) perceived as having similar form and non-identical meaning. If more than two languages are included, one can talk about sets of words. We use this term rather than false friend, given that it is a more specialized linguistic term.

This concept comprises cases such as English gift vs. German Gift ‘poison’, or BCS palac ‘thumb’ vs. Rus. palec ‘finger’, where words in two languages share form while exhibiting different meanings. Inasmuch as the problem of false cognates arises primarily in translation, most designations of the concept refer to ‘false friends of the translator’, as in French faux amis du traducteur, German falsche Freunde des Übersetzers, Cze. falešní přátelé překladatele, Pol. fałszywi przyjaciele tłumacza, and Rus. ložnye druz’ja perevodčika (found also in shorter forms: faux amis, falsche Freunde, ložnye druz’ja). The fact that such terms can generate mistakes is underlined in many phrases including the German term irreführende Fremdwörter, lit. ‘misleading foreign words’, the Cze. mezijazykové falsiekvivalenty, lit. ‘interlingual false equivalents’, zrádné slovo, lit. ‘deceptive word’, and the Pol. złudny odpowiednik, lit. ‘deceptive equivalent,’ odpowiednik pozorny, lit. ‘misleading equivalent’.

All the aforementioned terms approach the concept from a psychological and applied linguistic perspective – they encompass all cross-linguistic pairs with a potential of generating false equivalence. A different perspective is observable in the Pol. term tautonim, the Cze. mezijazyková homonymie, Rus. mež‘’jazykovaja omonimija lit. ‘interlingual homonymy’, and the BCS međujezički paronim, lit. ‘interlingual paronym’, međujezički homonim, lit. ‘interlingual homonym’. They refer to a contrastive linguistic category, regardless of their translational and/or psychological functioning.

This points to the following two overlapping phenomena:

1. (1) false cognates – a relation between two words, each from its own language (or language variety), which can cause cross-linguistic false equivalence – a psycholinguistic and applied linguistic category;

2. (2) cross-linguistic paronyms – a relation between two words, each from its own language (or language variety), with a similar form and different meaning – a contrastive linguistic category.

The languages or varieties considered here need to have something in common. In Slavic languages, the link that is the sine qua non for the existence of false cognates is inherited Proto-Slavic vocabulary and the so-called internationalisms (the vocabulary from the shared Greco-Roman cultural circle). The first ground for the relationship is dominant in inter-Slavic false cognates, the second in Slavic-to-non-Slavic false cognates.

There are those who see false cognates as a subset of lexical parallels, that is, the words that coincide in form and may or may not coincide in meaning. More information about this approach can be found in Reference Dubičinskij and RojterDubičyns’kij & Reuter (2011, Reference Dubičinskij and Rojter2015, Reference Dubičinskij and Rojter2020), as well as in Reference Kozdra and DubichynskyiKozdra & Dubichynskyi (2019).

Being perceived as having similar form in Slavic languages presupposes phonological correspondences between the two words in question. The forms of the two words need to include one of the following or a combination of them:

1. (a) sounds occupying approximately the same place in the phonological systems of their languages, which means that suprasegmental and phonotactic features, such as positional palatalization and final devoicing, may be different;

2. (b) reflexes of the same Proto-Slavic sound; or

3. (c) standard phonological and morphological adaptation of a borrowed word.

To exemplify the first category, (a) above, let us consider LSo. sad ‘fruit’ and Pol. or Rus. sad ‘orchard’, where all sounds occupy approximately the same places in the phonological systems of these languages. Within this category there are examples such as BCS bólnica ‘hospital’ vs. Sln. bolníca ‘female patient’, where suprasegmental features (such as stress placement) are different. Different phonotactics can be exemplified by Cze. mrav [mraf] ‘custom, deportment, manners’ vs. BCS mrav [mrav] ‘ant’, with final devoicing in Cze. and a lack of it in BCS. The second situation, see (b), can be seen in the fact that Macedonian danok ‘tax’ corresponds with BCS danak ‘levy’, and Slovene dánek ‘day (diminutive)’ given the correspondences of the development of jers (o:a:e), established in the chapters on vowels and consonants, although the BCS and Macedonian lexemes are etymologically related to PSL *dânь ‘tribute, tax’, while the Slovenian diminutive is linked with PSL *dьnь. Finally, English scan in general and technical senses, and Pol. skanować ‘scan, in a technical sense only’ exemplify the third category, as a is the standard adaptation of the English æ sound in Polish, and the -ować suffix is one way to adapt borrowed English words in Polish. Phonological development is covered in this volume in Chapters 2, 3, and 5; phonological adaptations in borrowing are covered in Chapter 25 in this volume.

The importance of false cognates in Slavic languages is that they represent results of semantic development from Proto-Slavic to present-day Slavic languages, covered in this volume in Chapter 24, and lexical borrowing, covered in Chapter 25. The topic of false cognates in Slavic languages has commanded considerable interest in theoretical and practical research. The list of references at the end of the chapter lists various Slavic-to-Slavic and Slavic-to-non-Slavic dictionaries, papers about Slavic false cognates, and even papers about false cognates in Slavic dialects.

27.2 Types of False Cognates

Slavic languages feature large areas of commonality in their vocabularies (body parts, measures, kinship terms, flora, fauna, etc., which is covered in Chapter 23 in this volume). Reference Golubović and GooskensGolubović & Gooskens (2015) explored mutual intelligibility between the six Slavic languages in the European Union and obtained a variety of scores in different tasks intended to measure understanding of texts in another Slavic language, ranging from 9.52 percent to 96.52 percent. The ability to understand another Slavic language is limited by various factors. For example, some languages are closer to one another than others, understanding of certain types of texts is easier than with other types, etc. False cognates are definitely one of the factors that hamper inter-Slavic communication.

The lexical relationship in a pair of false cognates can take different configurations. First, there is a difference between full and partial false cognates. In some pairs, the two words do not share any meanings; their entire semantic structure is in the relationship of false cognates. For example, Pol. angielski ‘English’, while Rus. angel’skij ‘angelic, cherubic’ are full false cognates. On the other hand, Pol. broda ‘beard, chin’ and standard contemporary Russian boroda ‘beard’ exemplify partial false cognates, given that only a part of the semantic structure (in this case the sense of ‘chin’) engages in the relationship of false cognates. Second, some false cognates are monodirectional, others are bidirectional. The aforementioned pair broda : boroda is monodirectional. Only when one translates from Polish into Russian does the relation of false cognates exist (with the meaning ‘chin’). Translating from Russian into Polish will not involve false cognates, as the two languages share the sense of ‘beard’. The example angielski : angel’skij exemplifies the bidirectional type. Translating from Polish into Russian and from Russian into Polish alike will have to pay attention to false cognates.

The pairs of false cognates also differ as to the underlying cause of the relationship. The three main mechanisms behind the emergence of false cognates in Slavic languages are the split of an inherited Proto-Slavic root, lexical borrowing, and word formation. The split, the first aforementioned mechanism can be seen in the development of Proto-Slavic words *batę (meaning ‘father, brother, relative, uncle’, depending on the language), and *životъ (meaning ‘life’, ‘stomach’, ‘body’, depending on the language). Borrowing can be seen in BCS fleka ‘stain’ vs. Pol. flek ‘heel tip’, where both members of the pair are German loanwords, but their meanings differ. Finally, Sln. mostišče ‘pile dwelling, bridgehead’ vs. Rus. mostišče ’bridge (augmentative)’ exemplifies word-formation differences, in this particular case two different functions of the suffix -išče, referring to the place in Slovene and denoting the augmentative in Russian. Concrete pairs of false cognates can also include a combination of these factors. Word-formation differences can also emerge when the two words are derived from different stems, for example Rus. xiščnik ‘vulture, predator’, derived from -xitit’ ‘to plunder’, coinciding with Slovene hišnik ‘janitor’, related to Slovene hišen ‘residential’ and hiša ‘house’. The aforementioned factors can be concurrently present in a pair of randomly connected false cognates; thus BCS dah ‘breath’ vs. Pol. dach ‘roof’ is a product of the development of a PSL root in BCS and borrowing from German in Polish.

There is a scale of difference in the meaning of false cognates, ranging from subtle differences in the same general subject-matter field, as in BCS penzija, Rus. pensija ‘retirement money’ vs. Pol. pensja ‘salary’, to loosely related meanings, as in Cze. život ‘life’ vs. Rus. život ‘stomach’, to completely unrelated meanings, as in the aforementioned example dah : dach. Subtle differences also exist in the stylistic sphere, when the two languages in question share their basic meaning. For example, both BCS plesati and Pol. pląsać mean ‘to dance’, yet the BCS word is neutral and its Pol. equivalent archaic.

27.3 History

False cognates in various lexical fields can be traced back even in the development of early German and Roman loanwords. In some cases, semantic divergence can be related to the polysemy of the respective Proto-Slavic lexical items. For instance, PSL *duma ‘advice; thought; opinion’ (< Proto-Germanic *dōma ‘judgment’) has reflexes with divergent meanings, for example Rus. duma ‘thought; council’ and Bul. duma ‘word; conversation; thought’. In other cases, the diversification of the meanings of early loanwords has to do with idiosyncratic semantic changes that occurred in a specific language or a group of Slavic languages. For example, in Western Slavic languages PSL *kъ̏nęgъ ‘prince, ruler’ (< North-West Germanic *kuningaz ‘ruler’) develops a new sense of ‘priest’ (Pol. ksiądz, Cze. kněz, Slk. kňaz), while in Polabian t‘ėnądz, there was also an additional sense of ‘moon’.

The diversification of meaning that provoked emergence of Slavic false cognates can involve different degrees of semantic change from the original meaning. That is the case of the polysemy of the diverse adaptations of the Roman word *comes in Slavic languages. Its original meaning ‘companion’ most probably designated the official that represented the Roman power in late imperial provinces (Reference BočekBoček 2010). In feudalism, PSL *kъmetъ/*kъmetь, an early Roman loanword, underwent various semantic modifications, yielding Old Rus. ‘knight, warrior, member of a feudal lord’s retinue’ and Old Cze. ‘head of a family, village leader, representative of a village’. The traces of the former meaning, related to the high hierarchy position, are found in Bul. kmet ‘elected foreman; village headman’, Mac. kmet ‘village community leader’, and arch. Ser. kmet ‘land judge, village leader’. The designation of the person having a high position can be metaphorically linked with Cze. kmet and Slk. kmet‘, both of them marked as archaic and literary, denoting ‘wise and experienced old man’.

However, in the history of Slavic languages the lexeme *kъmetъ/*kъmetь underwent semantic devalorization: one of the semantic components of kmet implied subservient relationships between the village leader and his ruler. The impact of that component was that the word, influenced by the functional metaphor, got the meaning BCS ‘serf, one who doesn’t have his own land’, and with the evolution of feudal relationships BCS ‘peasant who has to pay rent for his land’. Sln. kmet ‘peasant’ and Ukr. kmet ‘village dweller’ are nothing but results of further semantic widening after the word lost its feudal connotations. In contemporary Sln. kmet shows two semantic drifts: by metonymic shift (‘peasant’ → ‘his land and surroundings’), the non-nominative plural forms (na kmeteh, s kmetov) acquire the meaning ‘in the countryside/from the countryside’; metaphorically, Sln. kmet denotes also ‘pawn’ as the least important chess piece.

Another semantic shift, triggered by a metaphor ‘peasant’ → ‘someone who behaves like a peasant’, occurred in some Slavic languages where the word obtained pejorative connotations. Thus Pol. kmiot ‘vulgar person, boor’ and Sln. kmet ‘awkward, inelegant person’ were the furthest departures from the original meaning.

After having considered two cases of loanwords that had developed different meanings in various Slavic languages, we will pass to inter-Slavic examples of false cognates, resulting from diverse semantic mechanisms. The word *sъklèpъ is a deverbative noun from PSL *sъklepa̋ti ‘to join, to unite’ (Reference SnojSnoj 2016: 683), and its original meaning denotes ‘a spot where two elements are united’. As the result of semantic narrowing in Sln. sklep denotes ‘joint, connection between bones’. In West Slavic languages the starting point of its semantic development was Old Pol. sklep ‘vault’. By metonymic shift, the word could express also a place with a vault: Old Pol. sklep, Cze. sklep ‘basement, cellar, treasury’. In Rus. sklep, a Pol. loanword adopted via Ukrainian, narrowed its meaning – it did not mean an underground place, but also acquired the meaning ‘burial vault’. The semantic development in contemporary Pol. was different than in Russian – motivated by metonymic shift, Pol. sklep became the general word for shops and stores, as they were often located in underground places (Reference BoryśBoryś 2008: 551).

In contrast to all the аbove-listed cases, where sklep denotes concrete objects, Sln. sklep takes on an additional abstract meaning ‘final meaning, finding, conclusion’. That semantic shift was triggered by the following metaphor: in the eighteenth century, the verb sklepati ‘to join, to unite’ got the meaning ‘to combine thoughts’.

As has been demonstrated, in the history of Slavic languages, one can distinguish the following mechanisms of semantic development in the development of inherited Proto-Slavic roots and words borrowed from other languages:

1. (a) various types of transfers, such as metaphorical and metonymical shifts, drift from concrete to abstract meanings;

2. (b) widening and narrowing.

This can be seen in Pol. żywot ‘hagiography, the life of a saint’, which is narrowed down compared to Cze. život ‘life’, which maintains the inherited Proto-Slavic sense (for more information about this root, see Reference SaenkoSaenko 2021). In the homographic pair of Sln. grad ‘castle’ vs. BCS grad ‘city’, the latter language features a broader semantic widening than the former. In the example of Rus. list ‘leaf, sheet’ vs. Pol. list ‘letter’, Polish features metaphorical transfer while Russian retains the original sense.

In addition, maintenance and loss have played a considerable role in the emergence of false cognates. This can be exemplified with the pair Ukr. ložyty ‘lay, put’ vs. BCS ložiti ‘heat, make a fire’, where the former maintains the inherited Proto-Slavic sense, while the latter loses it and features a further semantic development. This is not the case of the pair Rus. vrednyj vs. Mac. vreden ‘valuable, worthy, diligent’, as these units refer to two different Proto-Slavic roots: the former is related to PSL *vrědьnъ ‘harmful’, while the latter is linked with the PSL *verdьnъ ‘worthy’. The derivatives of *verdьnъ ‘worthy’ were in some Slavic languages exposed to additional semantic shifts, for example the comparative adverbial form urẹ:dnẹ in the Slovenian dialect of Upper Savinja Valley has acquired the meaning ‘cheaper’ and not ‘worthier’ (Reference WeissWeiss 1990: 102).

For more information about semantic development see Chapter 24 in this volume.

False cognates are occasionally generated by paronymic attraction, that is, the change of form caused by the similarity with another word. For example, the Pol. word dorożka ‘droshky, a horse carriage’ is a false cognate with Rus. dorožka ‘narrow road’ owing to the fact that the Rus. borrowing drożki ‘droshky, a horse carriage’ has been adapted in Pol. into dorożka in consequence of paronymic attraction with Rus. dorožka.

27.4 Inter-Slavic vs. Slavic and Non-Slavic

The distribution of the aforementioned mechanisms in the emergence of the Slavic-to-Slavic false cognates is substantially different from those involved in the emergence of Slavic-to-non-Slavic false cognates. Reference ŠipkaŠipka (2015) compared false cognates between BCS and Polish with those between BCS and English. It turns out that in the former category (Slavic-to-Slavic false cognates) the split of an inherited Proto-Slavic root accounts for 91 percent of the cases, borrowing for 7 percent, and word formation for 1 percent. In the latter group (Slavic-to-non-Slavic false cognates), Borrowing accounts for 97 percent of the cases and word formation for only 3 percent. Most commonly, the borrowing is from Latin in both languages (51 percent of the cases); English borrowings in BCS comprise 22 percent of the cases, and borrowings from other languages 24 percent. This disproportion has consequences in applied linguistics. False cognates will be an issue very early in the process of teaching Slavic languages to Slavs (given that inherited roots are found mostly in frequent vocabulary items). In teaching Slavic languages to non-Slavs, false cognates will become frequent much later, given that most of the borrowings are in the sphere of abstract vocabulary.

27.5 Research Tradition and Dictionaries

The topic of false cognates in Slavic languages has aroused considerable interest in theoretical and practical research. There are dictionaries and glossaries of the Slavic-to-Slavic false cognates, such as Reference BelinʹkajaBelin’kaja 2015, Reference BirbrajerBirbrajer 1987, Reference BunčićBunčić 2020, Reference ČemerikićČemerikić et al. 1988, Reference GrabčikovGrabčikov 1980, Reference Ivanova and MarijanaIvanova & Aleksić 2007, Reference Kononenko and SpivakKononenko & Spivak 2008, Reference LewisLewis 2016, Reference LewisLewis 2002, Reference LotkoLotko 1992, Reference LytovLytov 2020, Reference Szalek and NečasSzalek & Nečas 1993, Reference OrłośOrłoś 2006, Reference SedakovaSedakova 2005, Reference SedakovaSedakova 2008, Reference StojkovićStojković 2004, Reference ŠušarinaŠušarina 2015, Reference TokarzTokarz et al. 1994, Reference TokarzTokarz 1998, Reference Tokarz1999, Reference UryadovUryadov 2015, Reference VlčekVlček 1966, Reference VyxotaVyxota 2004, Reference WijasWijas 2014, Reference Žuravlev and ZaxarovŽuravlev Zaxarov 1977). There are studies of Slavic-to-Slavic false cognates (Reference FranekFranek 1998, Reference BunčićBunčić 2000, Reference Dubičinskij and RojterDubičinskij & Rojter 2015, Reference IlievskaIlievska 2006, Reference Kalenić and HadžiKalenić 2001, Reference KarpaczewaKarpaczewa 1987, Reference KonickajaKonickaja 2011, Reference Lewis, Příhoda and VaňkováLewis 2008, Reference LewisLewis 2015, Reference LobkovskajaLobkovskaja 2012, Reference Opačić, Djigunović and PintarićOpačić 1995, Reference Peti-Stantić and TivardarPeti-Stantić 2014, Reference Popović and TrostinskaPopović & Trostinska 1988, Reference Popović and Trostinska1989, Reference SoglasnovaSoglasnova 2018, Reference XucišviliХucišvili 2006, Reference XucišviliXucišvili 2010). There are furthermore papers on Slavic inter-dialectal false cognates (Reference Blažeka and BlažetinBlažeka 2013, Reference Blažeka and VendiBlažeka & Franc 2017). There are dictionaries of Slavic-to-non-Slavic false cognates (Reference AkulenkoAkulenko 1969, Reference BorisovaBorisova 2002, Reference Dubičinskij and RojterDubičinskij & Rojter 2011, Reference Gotlib and GenrixGotlib 1985, Reference KanoničKanonič 2001, Reference KovačevićKovačević 2009, Reference KrasnovKrasnov 2004, Reference PaxotinPaxotin 2006, Reference VyxotaVyxota 2002). Finally, there are papers about Slavic-to-non-Slavic false cognates (Reference IvirIvir 1968, Reference LipczukLipczuk 1988, Reference ŠipkaŠipka 1991, Reference Šipka2015).

27.6 Conclusion

The study of false cognates in Slavic languages has its applied and theoretical importance. In applied linguistics, false cognates are a formidable hurdle in inter-Slavic second language acquisition, as they are encountered very early in the process of instruction, unlike Slavic to non-Slavic language instruction, when most of them appear much later in the process. Theoretically, the study of false cognates is important as they represent the final point of various processes in the history of the lexicons of individual Slavic languages: semantic developments, lexical borrowing, and word-formation processes. As such, they encapsulate not only the aforementioned processes but also the dynamics of interaction between them.

While false cognates seem to be well documented between most major Slavic languages, the need for further research still exists in the following three areas. First, the gaps where there is a conspicuous absence of Slavic-to-Slavic and Slavic-to-non-Slavic dictionaries of false cognates should be filled. Second, the applied linguistic ramifications of false cognates should be elucidated. Finally, theoretical work on linking false cognates with semantic development in the lexicon of Slavic languages and on their typological classifications should be conducted.

28.1 The Concept of Dialectal Fragmentation

As a metaphor for the description of language variation over time, the notion of dialectal fragmentation presumes that there is a relatively homogeneous whole (a language) which in the course of time undergoes separation into pieces (dialects) due to a variety of historical circumstances. In this respect, the notion is analogous to the Stammbaum (branching diagram) conception of linguistic development, which starts from an artificially reconstructed point of departure (here Proto-Slavic [PSL], ca. fifth century AD), then gradually branches out into increasingly diverse, yet genetically related dialects (here East, West, and South Slavic, whose representatives are attested, aside from earlier toponyms, directly and variously from the tenth to twelfth centuries AD).

This conceptualization is of course rather schematic. Since all attested human languages exhibit some degree of variation, we must assume that protolanguages conform to this pattern. The generalization applies to PSL (particularly in view of its numerous tribal sub-affiliations, including at least one, whose name *Xъrvat- ‘Croat’, is ultimately of Iranian origin, thus dating back to the prolonged period of close Slavic-Sarmatian contact north of the Black Sea, ca. 700 BC– AD 300). Unambiguous supporting evidence for this assertion can only be traced back to the early period of the Slavic Migrations (seventh to eighth centuries) from the probable Slavic homeland, which was originally situated to the (north)east of the Carpathians, although some resettlement to the mid-Danubian basin had occurred by the time of the Avar Khaghanate in the seventh century (Reference Nichols, Maguire and TimberlakeNichols 1993). These migrations were part of the great movement of peoples which attended the collapse of the Roman Empire, marking the transition from late Antiquity to the early Middle Ages. By the end of this period, the Slavic speech continuum extended from the Elbe River and the Baltic Sea (also NW Russia) in the north to the Peloponnesus in the south, and as far east as the Dniepr. This territorial expansion (noted in the mid-sixth century AD by the Gothic historian Jordanes and the Byzantine historian Procopius, with reference to the Sklavenoi and Antai, as well as by John of Ephesus, who describes the Slavs as present in “all of Greece, Thessaly, all of Thrace” in AD 581–584) not only intensified prior contact with Baltic and Germanic in the northeast and northwest, but also brought the Slavs into contact for the first time with a number of other languages, including Finnic in the north, as well as Balkan Romance (chiefly Romanian and Aromanian), Albanian, and Greek in the south (not to mention possible surviving remnants of Paleo-Balkan languages other than Proto-Albanian). This geographical dispersion facilitated rapid and territorially differentiated linguistic change in the Slavic speech area. The effects of dispersion were augmented by those of actual division, when in the last centuries of the first millennium AD the South Slavs were gradually separated from their northern kinfolk as the Bavarian kingdom extended farther eastward (mid-eighth century), and the Magyar (Hungarian) kingdom incorporated most of Great Moravia in the early tenth century.

Some developments which we will consider partially link West and South Slavic (e.g. the ‘jugoslavisms’ in Central Slovak or the lenition of *g > γ in parts of northern West South Slavic), others point to an earlier separation of West Slavic from the other branches (as in particular treatments of the Second Velar Palatalization and *tl, *dl consonant clusters), still others oppose South Slavic to both West and East Slavic (sometimes referred to as ‘North Slavic’), as in the conservative retention of the reflexes of nasal vowels in certain desinences. At times the residue of earlier changes manifests itself in the form of discontinuous isoglosses (as in Carpathian-Balkan lexical parallels) or more fragmented, tessellated formations (as in the various endings for the third person singular and first person plural in the present tense). Finally, interaction with different non-Slavic languages (Finnic, Germanic, Venetian, and the many languages of the Balkans) has contributed to a mixture of divergent and parallel innovations in the Slavic languages which participate in the process. As a result, we find perfect past tense formations with ‘have’ not only in Balkan Slavic, but also in colloquial Czech, or the merger of palato-alveolar voiceless affricates with pre-palatals or dentals not only in northern Russia, but also on the Dalmatian coast.

The reader is advised that although the first sections of this chapter focus more on early dialect differences, the chapter is not intended to serve as a historical outline of the development from Proto-Slavic to the modern daughter languages (for this, see e.g. Reference Schenker, Comrie and CorbettSchenker 1993, Reference Greenberg and KapovićGreenberg 2017). Instead, it seeks merely to illustrate and provide a minimal historical context for some of the kinds of dialectal fragmentation which have been documented in Slavic languages. The topics covered include the following: PSL palatalizations; other PSL dialect differences in phonology, prosody, and morphology; pre-triadic variation (morphology); early East Slavic tribal dialects; lexical isoglosses across Slavic; inter-Slavic areal features; prosodic continua in West and South Slavic; the role of external factors (economic, political); the effects of contact with non-Slavic languages (Finnic, Germanic, Venetian, the Balkan Sprachbund); and finally, the impact of sociolinguistic factors (religious confession, gender). While most of the topics and illustrative material concern phonology and prosody, I have also made some reference to morphology, syntax, and the lexicon. Under the assumption that the reader has no prior familiarity with the historical phonology (or grammar) of Slavic, I have also provided a modicum of explanatory background where it seemed necessary.

28.2 Notes on the Transcription of Proto-Slavic and Old Church Slavonic

When citing PSL reconstructed forms, I follow current common practice, striking a balance between a precise phonetic reconstruction and typographic considerations, which are intended to minimize or avoid the use of special IPA symbols. This principally concerns the PSL ‘back low’ vowel, whose phonetic reconstruction would be [ɔ], but for which we use *a (as in Reference Greenberg and KapovićGreenberg 2017, Reference VermeerVermeer 2014); correspondingly, the PSL ‘front low’ *e probably resembled [ε] or [ä]. Further, in representing PSL vowels, we use the macron to indicate length /ā, ē, ū, ī/ and leave the low short vowels unmarked /a, e/, while marking the short high vowels (ŭ, ĭ), as a concession to tradition and a reminder that these are the ancestors of OCS ъ, ь, respectively.

For Old Church Slavonic (hereafter OCS), we retain the letters ъ (back jer, pronounced roughly as the vowel in English ‘but’) and ь (the front jer, pronounced roughly as in ‘bit’), but transliterate the jat’ vowel as ě (phonetically probably [^jæ], with a sub-phonemic palatal on-glide), and the jery <ы> as y (phonetically [ɨ], Rus. ы). Unless otherwise indicated, when referring to the hushing consonants, č = [tʃ], š = [ʃ], ž = [ʒ] (pronounced with varying degrees of secondary palatalization or retroflexion), I will use the term ‘palatal’ here in a broader traditional sense (as in Reference VermeerVermeer 2014), rather than only to refer to IPA [ɕ, ʝ, c, ɟ]; c = [ts]. The affricates ć, dź are pre-palatal, as in BCS and Polish. Czech <ě> usually indicates [jε], for example běs [bjεs] ‘demon’, except after /m/, where město = [mɲεsto], and after dentals, where it indicates an IPA palatal stop, for example děd [ɟet] ‘grandfather’.

Generic geographical abbreviations: N = north, NE = northeast, NW = northwest, S = south, SE = southeast, SW = southwest.

28.3 Early Differences in the Proto-Slavic Palatalizations

The clearest early evidence for PSL dialectal differentiation at the time of the Slavic Migration is provided by the so-called Second Velar Palatalization and the Jot Palatalization of PSL *dj, *tj (Reference Greenberg and KapovićGreenberg 2017, Reference Schenker, Comrie and CorbettSchenker 1993, Reference ShevelovShevelov 1965). These changes are best understood in the overall context of PSL palatalizations.

At a stage in the development of PSL which had been reached in the first half of the first millennium AD, but prior to the Slavic migrations, the existing velar obstruents (the stops *g, *k, and the voiceless fricative *x) could occur in front of both back and front vowels (*gadŭ ‘year’: *genā ‘woman’, *kala ‘wheel’: *kela ‘forehead’, *xaldŭ ‘cold’: *xestŭ ‘sixth’). Over the period of approximately the fifth to eighth centuries, this distribution was fundamentally realigned due to a series of palatalizations, by which velars adapted their pronunciation to following and (more rarely) preceding front vowels (compare the different [k] sounds in English ‘cool, cold’ vs. ‘kill’, or the even more emphatic alteration of historical *k in ‘chilly’, a cognate of ‘cool’, as well as the difference between the velar and palatal fricatives in German Ach [ax] ‘Ah!’ and ich [iɕ] ‘I’). These changes occurred in two historically distinct stages, known as the First and Second Velar Palatalizations (hereafter, 1st and 2nd Vel Pal). The two stages were formally separated by the Monophthongization of all inherited diphthongs (hereafter, Mono), a change which provided a fresh stock of roots and endings with long *ē. According to the current consensus, an integral part of the 2nd Vel Pal also includes what has traditionally been referred to as the Third Velar Palatalization (3rd Vel Pal). The 3rd Vel Pal yields the same reflexes as 2nd Vel Pal, but is caused by a preceding (high) front vowel under various conditions (including the value of the following vowel).

The outcomes of the velar palatalizations in OCS are in most respects typical of the majority of Slavic dialects (see Table 28.1).

The relative chronology of the 1st Vel Pal, Mono, and 2nd Vel Pal can be established by comparing the behavior of the voiceless velar *k- at four reconstructed stages in its PSL development in position before the back vowel *-a [ɔ] and the front vowel *-e [ε] in the following combinations: *ka-, *ke-, *kei-, *kai- (see Table 28.2). Of particular interest here is the different behavior of the Indo-European e-grade and o-grade cognate diphthongs in PSL *keist- > *čeist- (by 1st Vel Pal) > *čīst- ‘clean’ (by Mono) vs. *kaist- > *kēst- (by Mono) > *cěsta (by 2nd Vel Pal, thus not *čěsta > časta) ‘road’ (i.e. ‘a cleared path’, e.g. Cze. cesta, Sln. césta).

This chronology is consistent with the fact that later loanwords associated with the adoption of Christianity by the Slavs exhibit 2nd Vel Pal rather than 1st Vel Pal before front vowels, for example OCS crьky, Rus. cerkov’ < PSL *kьrky ‘church’, Gen. *kьrkъve < West Germanic *kirikō Fem. < Greek kurikon Adj. ‘the Lord’s’. The PSL forms here probably date to the mid-eighth century, when Carantanian Slovenes could have borrowed the word in its West Germanic form (see Reference Pronk-TiethoffPronk-Tiethoff 2013).

As shown in Table 28.3, the 1st Vel Pal (*g > *ž, *k > č, *x > *š) and Mono (more specifically, *ai- > *ě₂) yield uniform reflexes across all of Slavic. Note that the sub-critic in *ě₂ is merely a linguistic convention used to indicate the diphthongal origin of the vowel *ě in certain lexemes and morphemes. This vowel is identical phonetically to *ě of monophthongal origin, although this identity is partially obscured by a late PSL dissimilation, whereby *čě- (from 1st Vel Pal) > č’a, etc. in almost all of Slavic (e.g. *krīkētei ‘to scream’ > *kričěti > Rus. kričát’ vs. *kesŭ ‘time’ > OCS (Zographensis) čěsъ). The same change occurred in *cě₂-, etc., but only in Bul.-Mac. dialectal cal, pl. cali (*cě₂ l-) ‘whole’ (vs. standard Bul. c’al, celi), as well as dzalo (*dzě₂lo) ‘very’, etc. in a few medieval SW Bul.-Mac. texts (Reference Scatton and MagnerScatton 1976). The identity of the sound in question is confirmed by the fact that the OCS alphabet uses the same letter (called jat‘) to designate both *ě and *ě₂. Early Slavic loanwords from the period of the Slavic migration in Greece (primarily toponyms) show no surviving velars in the position for the 1st Vel Pal, nor any PSL diphthongs (see further below).

In contrast, the 2nd Vel Pal (again Table 28.3) already shows several different kinds of reflexation in the Slavic dialects. The differences involve various limitations on the occurrence of the 2nd Vel Pal (in West Slavic and dialectal East Slavic), as well as a particular treatment of *x- (again in West Slavic).

1. Type 1: *g-, *gv- > (d)z, (d)zv, *k-, *kv- > c, cv, *x- > s (OCS, South Slavic, and most of East Slavic). This most common type articulates all reflexes as dentals/alveolars. It also treats *kv-, *gv- the same as *k-, *g-. It is probable that an intervening ‘pre-palatal’ stage (dź, ć, ś) preceded the alveolar outcome (see Reference VermeerVermeer 2014).

2. Type 2: *g > (d)z, *k > c (as in type 1), but *x > š (i.e. a palatal rather than a dental) and the velars remain in *kv-, *gv- (West Slavic).

3. Type 3: This rare type does not undergo 2nd Vel Pal, but retains the velars before the reflex of *ě₂ as [k’, g’, x‘] in all positions (NW Russian), for example Old Novgorod-Pskov kěle ‘whole’ (Rus. cel), xěrь ‘grey cloth’ Fem. (Rus. séryj ‘grey’, OCze. šěř ‘grey cloth’), Xědovo (toponym) (Rus. sedój ‘grey’), lęgi Impv. ‘lie down, go to bed’ (-i₂ in the Impv. Sg. is a particular development of Indo-European *-ōi, rather than *-ēi), gvězda ‘star’, květ ‘flower’. The modern dialect evidence indicates that the velars here underwent palatalization, but did not shift their fundamental place of articulation, thus k’, g’, x‘ (Reference ZaliznjakZaliznjak 2004).

4. Type 4: This particular Ukrainian dialect type is the same as Type 1, except that *kv- and *gv- exhibit diffuse regional variation between preservation and standard East Slavic palatalization of the velars for this position, as is reflected in a small number of examples, such as (1) Ukr. kvit ‘flower’ ~ cvit ‘blossom; flower’ (kvit is more common in Kiev-Poles’e dialects), and (2) dialectal zvizda ~ hvizda (the lexeme itself is better attested in SW Ukrainian, mainly as zvizda, which occurs in Church Slavonic and folklore, but is generally replaced in most of Ukrainian by zorja, attested since the thirteenth century). Note that the distinction in the plural which Russian expresses with cvetý ‘flowers’ vs. cvetá ‘colors’ is rendered by Klymentij Zinovijiv (a late seventeenth-century Ukr. poet) through květy vs. cvěty (Reference ShevelovShevelov 1979: 56–58, as well as Map 28.2 for approximate isoglosses). The semantic complexity of the problem can be seen in the co-occurrence of the two types in the Carpathian Ukrainian dialect of Torunj (Reference Nikolaev and TolstajaNikolaev & Tolstaja 2001), where c’v’it, Gen. c’v‘ítu, pl. c’v’itɷ́ ‘color of foliage on a tree; blossoming; color (in general)’, and the plural signifies both ‘flowers’ and ‘colors’, whereas the plurale tantum kv‘ítɷ, Instr. -amɪ means only ‘flowers’, for exmaple v’inók … s kv‘ítamɪ ‘a wreath of/with flowers’, and the corresponding sg form is the diminutive Fem. kv‘ítka ‘a flower’. Meanwhile, the underlying verb shows only the 2nd Vel Pal (f. Inf. cvɪstɪ́, 1p.pres cvɪtú, 3p -é).

For still other East Slavic dialect types, some of which even differentiate the reflexes of 2nd and 3rd Vel Pal from one another, see Reference Nikolaev and IvanovNikolaev (1988).

Due to limitations of space, we will not discuss dialectal differences in the treatment of *skě₂-clusters (e.g. *voskъ ‘wax’, *na voskě ‘on the wax’ > OCze. voščě ‘wax’ Loc., OCS voscě; see e.g. Reference ShevelovShevelov 1979, Reference VermeerVermeer 2014).

To conclude this section, we will consider a different type of PSL palatalization (of somewhat later date than those described above), one which contributed perhaps even more significantly to early Slavic dialectal fragmentation. I refer to the so-called ‘Jot Palatalization’ of the dentals in the combinations *tj, *dj, as in Late PSL *světja ‘candle’ and *medja ‘boundary’, respectively (from *svaitiā, *mediā by glide formation [^j] of the derivational morpheme {i} in position between a stop and a vowel, a typologically common development).

In general, the reflexes of *tj, *dj exhibit four different kinds of reflexes (see Table 28.4):

1. Type 1 (East Slavic): Merger with the result of the 1st Vel Pal (thus, č, ž in ESl, except for č, dž with archaic retention of an initial stop, in parallel with the voiceless reflex, in SW Ukr.; less commonly in South Slavic – č, dž in Torlak, whereas č, but j [a palatal glide] in Slovene);

2. Type 2 (West Slavic): Merger with the 2nd Vel Pal (thus, WSL c, dz in Slovak and Polish, but c, z in Czech and Sorbian, although the merger is incomplete in Slovak, due to lenition in the reflex of the 3rd Vel Pal; contrast Slk. *dj > medza vs. *-g > peniaze by 3rd Vel Pal);

3. Type 3 (Western South Slavic): A distinct pre-palatal reflex, which arguably continues an earlier phase in the PSL development (ć, dź [spelled đ] in BCS, t́, d́ in some Čakavian and (rarely) Torlak dialects, which also attest ḱ, ǵ);

4. Type 4 (Eastern South Slavic): Clusters of the type št, šč, (rarely šk’) < *tj, *dj > žd, ždž, (rarely žg’), which bear witness to an earlier stage of gemination (ćć, dźdź), with subsequent dissimilation of the initial affricate. More specifically, OCS: certainly št, perhaps earlier šč; Bulgarian: nearly always št, žd, rarely šk’, žg’, Macedonian: šč/št, žđž/žd (in the Southwest and far East), but more commonly ḱ, ǵ (in the North, most of the Central West, and therefore also standard Macedonian). The last-mentioned reflexes in the standard language are due to the influence of Serbian ć, đ (perhaps even ḱ, ǵ or t́, d́, due to the proximity of Southeastern Torlak), which increased in the fourteenth century and particularly after Skopje became the Serbian capital, whereas the older forms št, šč, ždž, žd are more consistently maintained in peripheral Southwestern and many Eastern dialects. The two sets of reflexes exhibit a north–south gradation and a lexicalized distribution, for example Central West synonyms snošti ~ sinok’a ~ ‘last night’ (*sь-nokti), but also differentiated derivatives, such as praštilo ‘drawstring’ ~ praḱa ‘slingshot’ (both from *prati- ‘send, throw’). Unusual for Macedonian is the pairing of *tj > š‘č’ (guréš‘č’o ‘hot’) and *dj > ž (with lenition, e.g. méža ‘boundary hedge’) in the remote SW village of Vrnik in eastern Albania (Reference SchallertSchallert 2005).

As is evident from the above, lenition (spirantization) of the initial voiced stop in the reflex of *dj is widely attested in all three branches of Slavic: (1) [ž] in nearly all of ESl and (quite rarely) in southern dialects of Macedonian, (2) [z] in Czech and Sorbian, and (3) [j] (with complete lenition to palatal glide) in Slovene, as well as in adjacent Čakavian and (to varying degrees) some of West Štokavian.

The reflexes of *tj are joined by those of the cluster *kt’ (including *gt’ > *kt’ by early devoicing assimilation).

28.4 Other PSL Dialect Differences

Among other notable late PSL dialect differences (phonological, prosodic, morphological), we may cite the following.

Archaic retention of the PSL clusters *tl, *dl is typical of West Slavic (Cze. pletla ‘she knitted’, mýdlo ‘soap’), whereas in East and South Slavic the initial dental stop is lost (Rus. plelá, mýlo, BCS plèla, mȉlo, Sln. molíti ‘to pray’). Exceptions here are NW Russian, in which the stop is not lost, but dissimilates to a velar (*tlešč- > kleščь ‘bream’) and NW Slovene (mọ̑dli̥m ‘pray’1p sg., kridwo ‘wing’), whereas Central Slovak is congruent with South Slavic plela, šilo ‘awl’ (*ši-dlo), one of several ‘jugoslavisms’ which typify this zone of Slovak.

Innovations in the treatment of closed syllables ending in ‘liquid (r, l) + obstruent’ (the so-called *TORT and *ORT groups) fall into two categories, viz. root structural (for *TORT) and prosodic-vocalic (for *ORT). For the first type, *golva ‘head’ > East Slavic pleophony, as in Rus. golová (for complications in the development of Ukrainian, see Reference GardeGarde 1974; for further variation, including Rus. dial. polymja ‘flame’, see Reference ZaliznjakZaliznjak 1985, Reference Zaliznjak2004) ~ West and South Slavic share metathesis, but with different vocalism, even within West Slavic, see Pol. głowa ~ Cze. hlava, Bul. glavá (for Lechitic a preliminary stage *gǝlowa has been proposed). For the second type, we find that the treatment of *ORT depends on the PSL tone: PSL acute tone *őrdlo ‘plow’ > BCS short falling pitch rȁlo, North (= East + West) Slavic /a/ Pol. radło, ORus. ralo vs. PSL circumflex tone *ȏlkъtь masc. ‘elbow’ > BCS long falling pitch lȃkat ~ North Slavic /o/ Pol. łokieć, Slk. loket’, Rus. lókot’ (but cf. Central Slovak laket’, again as in South Slavic). Since /a/ is the late PSL reflex of PSL long *ā, whereas /o/ reflects short *ă, East Slavic in the course of metathesis appears to have converted the PSL acute in this position to length, but the PSL circumflex to brevity. This is the opposite of the South Slavic quantitative reflexes of these same two PSL tones (at least in disyllabic word forms). Nor is the observance of such a prosodic difference in early East Slavic at all surprising in light of differences such as koróva ‘cow’ from late PSL *kőrva (Acute) vs. górod ‘town’ from late PSL *gȏrdъ (circumflex).

There are also a small number of North–South dichotomies, two of which involve the treatment of late PSL nasal vowels in auslaut in specific morphological environments (Pres. Act. Participle: OCS nesy ‘carrying’ ~ ORus. nesa, gen. sg and nom-acc. pl in ja-stems: OCS zemlę ‘earth, land’ ~ ORus. zemlě, OPo ziemie, where the absence of a nasal vowel in OPol. is strongly diagnostic) and one which concerns the replacement of *-a (the anticipated o-stem masc.instr.sg, as probably in the adverb *vьčera̍ ‘yesterday’) either with the ŭ-stem desinence (as in North Slavic, cf. ORus. bog-ъ-mь ‘god’, Pol. bratem ‘brother’, for the Polish reflex of *ъ, cf. sen ‘sleep, dream’ < *sъnъ) or a freshly minted o-stem equivalent (as in South Slavic, cf. OCS bog-o-mь, BCS bratom). There are a very limited number of exceptions with the reflex of *-ъmь in South Slavic (attested in a tiny number of peripheral southern Štokavian dialects, see Reference IvićIvić 1994).

28.5 Pre-triadic Variation

Other candidates for evidence of PSL dialect fragmentation include some whose areal distribution differs at times markedly from the triadic type.

One such piece of evidence can be found in the idiosyncratic areal distribution of the two competing suffixes (original *-nū- and secondary *-nūn-, as per Reference AndersenAndersen 1999) for Slavic verbs formed with the n-suffix (cf. respectively Slovene sah-ni-ti ‘dry (intrans.)’ from -ny- < *nū, vs. BCS sah-nu-ti from *-nǫ- < *-nūn). The common element *-nū- in each form represents a lengthened grade variant of *-nu-, itself a zero grade of the IE *neu- (for the lengthened grade, cf. Greek deik-nū-mi ‘I show’ vs. zero-grade pl. deik-nu-men ‘we show’, for *-neu̯, cf. the Slavic past passive participle formation type mi-nov-en ‘past’). The archaic *nū-type occurs only on the western periphery of the West and South Slavic speech territories (cf. West Slavic: Polabian, Upper Sorbian, Silesian; South Slavic: Slovene, Čakavian, some Kajkavian, West Štokavian), which suggests that it was brought there by an early wave of settlers from the east and south, before the newer *nǫ-reflex began to spread to the rest of Slavic (for the pathways of migration determined by recent archaeology, see Reference BaranBaran 1990, whose map is redeployed in Reference AndersenAndersen 1999).

A second example of apparent pre-triadic variation is the case of three distinct endings for the third person singular present tense (thematic *-tь, pronominal *-tъ, and originally injunctive [?] *-t > zero), all of which occur in our earliest written sources (tenth to twelfth centuries), exhibiting an areal distribution in several respects incongruent with the threefold East–West–South classification (thus, in the Postscript to the Old Russian ‘Ostromir Gospel’, the Novgorod (?) scribe Grigorij writes da iže gorazněe sego napiše ‘but if anyone can write better than this’, using an East Slavic 3p.sg.pres dialect form napiše, even though throughout the main body of the Gospel text he has consistently used the ‘standard’ ORus. thematic ending -tь, which survives widely in modern dialects, as well as in i-verbs in standard Ukrainian, rather than Old Church Slavonic -tъ; see Reference MillerMiller (1988) for the linguo-geographical distribution and history of the question, as well as Reference OlanderOlander 2015). Of particular note is the retention of -t in Western Macedonian (as in OCS) down to the present day, despite zero in the rest of South Slavic.

A third example of a scattered distribution, also inconsistent with the modern triadic grouping, is evident in at least three of the four attested first person plural endings, see *-mъ (OCS, Bul. [-im, -em], Rus.); *-mo (BCS, Slovene; Central Slovak; Ukrainian, since at least the fourteenth century); *-me (Czech, Slovak, Bulgarian [-ame], Macedonian, Ukrainian West Galician dialect); -*my (Polish, Sorbian, and as a variant in some of OCS and medieval Bulgarian) appears to be secondary (for PSL reconstructions, see Reference VaillantVaillant 1966 Vol. 3, Reference OlanderOlander 2015). Two factors which initiated or supported modern vocalic endings are (1) the merging of 1sg -mь and 1pl -mъ in the numerically small, but highly frequent athematic verbs (whence -my in early Bulgarian, by analogy to the first plural pronoun), and (2) the analogical spread of the athematic 1sg -m(ь) (cf. dām ‘give’, ěm ‘eat’, věm ‘know’, esm̥ ‘am’, imām ‘have’) in South and West Slavic to some or all of the remaining verb classes, which would have increased the motivation for endings with a vowel following -m (e.g. BCS prȍsīm : prȍsīmo ‘ask’, čìtām : čìtāmo ‘read’; Mac. kradam : krademe ‘steal’ vs. Bul. kradá : kradém).

28.6 Early East Slavic Tribal Dialects

The Old Russian Novgorod birchbark documents (twelfth to fifteenth centuries) attest an interesting array of distinctive archaisms, including (1) nom.sg.masc -e (vs. -ъ in all the rest of Slavic; see Reference VermeerVermeer 1994, Reference ZaliznjakZaliznjak 2004), (2) the absence of the 2nd Vel Pal (for examples, see above), and (3) the retention of PSL *tl-, *dl- clusters (in the dissimilated form kl-, gl-, e.g. Old Novgorod kleščь ‘bream’, jogla ‘fir tree’, vs. Rus. leščь, jelь, but with a structural resemblance to Polish jodła and segmental identity with East Baltic cognates, such as Lithuanian ẽglė, Latvian egle). These traits set the Old Novgorod dialect apart from the rest of East Slavic and in the case of the nom.sg and the 2nd Vel Pal from the rest of Slavic as a whole. According to one theory, these archaisms indicate that the East Slavic Kriviči tribe, who composed the Slavic population in the western part of the Pskov-Novgorod regions in the tenth century, had separated from the main body of Proto-East Slavic speakers during the beginning of the Migration period, then re-established contact with other East Slavic tribes as the latter moved northward. Yet another Northwestern Russian dialect trait, which is already attested in the earliest Novgorod documents, is tsokan’je (the pronunciation of *č of diverse PSL origin as [c’] or [c], e.g. c’isto ‘clean’, skac’eš ‘you leap’ ~ Rus. čisto, skačešь), attributed by some scholars to assimilation of a Finno-Ugric substratum. The widespread occurrence of this innovation primarily in northernmost Russian dialects overlaps considerably with the domain of the Novgorod Republic (twelfth to fifteenth centuries) with its extensive colonial outreach. Other innovations contributed further to the differentiation of parts of Northwestern Russian, such as the formation of the so-called ‘Second pleophony’ (s’arép for Russian serp ‘sickle’) and a low-mid, non-palatalizing front-to-central value for the back jer *ъ when it occurred before the palatal glide (as in the nom.sg.masc.adj desinence *-ъjь, e.g. molodέj/-ə́j < *mold-ъjь ‘young’, Rus. molodój, or in the root vowel, e.g. mέju/mə́ju < *mъjǫ ‘I wash’, Rus. móju) or before palatalized sonorants (as in the residual lexeme odέn’je ‘stave at bottom of barrel; bale of hay’ < *o-dъn-ьje, where *dъn- is the same etymon as in Rus. dno, pl. dón’ja ‘bottom’ <*dъn-o). For further discussion of these and other archaic traits, including some (such as *květъ > t’v’et (sic), and ‘exotic’ oxytonesis in tučá ‘rain cloud’, bur‘á ‘storm’, vol‘á ‘will’ vs. Rus. túča, búr’a, vól’a) which arguably can be related to the formation of the Vjatiči tribal dialect (situated non-cotangentially to the southwest of the Kriviči in the eleventh to thirteenth centuries), see Reference NikolaevNikolaev (1994).

28.7 Early Lexical Differences across and within Slavic Languages

Here we offer a small sampling of lexical differences which exhibit significant linguo-geographical patterning within Slavic.

East, South ~ West (Reference BräuerBräuer 1961, Reference SkokSkok 1971–1973): (1) ‘feast, wedding feast, (patron) saint’s day feast’, Rus., OCS, BCS pȋr, pȉjer (Bosnia), Sln. pȋr ~ Pol. gody, uczta; Cze. svátek, hostina, (2) ‘watch, look at’, Rus. smotret’, OCS sъmotriti, BCS mòtriti ~ *patrěti > Pol. patrzeć, Cze. patřit ‘to belong, to be fit for’ (cf. hledět, dívat se ‘look at’), for the semantic shift ‘look at’ > ‘watch out’ > ‘belong to, pertain’, see Reference SkokSkok 1971–1973, and (3) ‘wedding’ Rus. svad’ba, BCS, Bul., Mac. svadba, ~ Pol. ślub (but Cze. svatba).

South ~ East, West (Reference SławskiSławski 1962): (1) *žica ‘thread’ ~ Rus. nit’, Pol. nić, (2) *slana ‘hoar-frost, frost’ ~ Pol. szron, Rus. inej, (3) *loza ‘grape-vine’ (whereas Pol. łoza ‘osier, Salix cinerea’), (4) *kaniti ‘invite > intend’, (5) *brojь ‘number’ ~ Rus. čislo, Cze. číslo, Pol. liczba, and (6) *gaziti ‘wade (at a ford), tread’ ~ *bresti, *broditi ‘wander’ > *brodъ ‘ford’ > Rus. brod, Pol. bród.

Within South Slavic, lexical differentiation allows us to distinguish an innovating Center (BCS, especially Štokavian) from a more conservative Periphery (Slovene and to some degree Čakavian in the northwest, as well as Bulgarian-Macedonian in the southeast; Reference TolstojTolstoj 1974/1997, Reference PopovićPopović 1960): (1) *vatra : *ognь ‘fire’, (2) *kyša : *dъzdjь ‘rain’, (3) *pravъ : *desnъ ‘right (hand)’, (4) *znojь : potъ ‘sweat’ (but cf. also S Bul. znoj), (5) *čadъ/*čad’ъ : *sadja ‘soot’, (6) *ručati : *obědati ‘dine (eat the main meal)’, and (7) *govoriti : *gъlčěti ‘to speak, talk’ (the latter is restricted to the Prekmurje dialect in easternmost Slovene and various Bulgarian dialects, such as Mizija in the northeast, the central Rhodopes, and some of the Stara Planina, including dialects adjacent to Torlak, where it is again attested in eastern Torlak; for *gьlčěti and other verba dicendi, see Reference Schallert and GreenbergSchallert & Greenberg 2007).

Certain Eastern Bulgarian innovations, which are also reflected in standard Bulgarian, are not generally found in the rest of South Slavic, including Western Bulgarian, Macedonian, and sometimes Southern Bulgarian (Reference StojkovStojkov 1993): (1) ‘leg’ *korkъ > Bul. krak ~ *noga pan-Slavic, including Mac., W Bul., S Bul., and (2) ‘to pluck, pick’ *kǫsati (PSL ‘bite’) > Bul. kǝ́sa ~ *rъvati ‘pluck; tear’ > Rus. rvat’, Pol. rwać, SW Mac. (Kostur) ǝ́rve 3sg ~ *ky(d)nǫti (PSL ‘throw’) > BCS, Mac. kine, W Bul. kinem (note that *ky(d)nǫti is itself an innovation, when viewed in the general Slavic perspective; see Reference Schallert, Friedman and DyerSchallert 2001, Reference Kočev, Vakarelska-Čobanska, Kostova, Kjaeva and Tetovska-TroevaBDA-OT 2001, Lexicon). There are an even larger number of lexical oppositions of smaller geographical scope, which tend to distinguish Eastern Bulgarian from the West and sometimes also the South (Reference StojkovStojkov 1993), although the landscape is sometimes complicated by additional terms (Reference Kočev, Vakarelska-Čobanska, Kostova, Kjaeva and Tetovska-TroevaBDA-OT 2001), for example (1) ‘hot’ goréšto : žéžko, (2) ‘loom’ stan (also some NW): razboj (also SE Thrace), (3) ‘do not (prohibitive imperative)’ nedéj (some NW): nemój (also the South), both attested, though usually in different contexts, in OCS: nedějь ‘do not’, ne mo[d]zi < *mogi ‘can’t’, and (4) ‘I’ az : jas, ja.

Northern ~ Southern Russian (Reference KuraszkiewiczKuraszkiewicz 1963): (1) ‘plow a field’ > orat’ ~ *paxat’, (2) ‘horse’ koń, komoń ~ lošad’, (3) ‘wolf’ volk ~ b’ir’uk, (4) ‘squirrel’ vékša ~ bélka, (5) ‘rooster’ petúx (as well as pevún [center-north], petún Pskov-Novgorod) ~ kóčet (although mainly limited to the SE, this form nonetheless extends north as far as Kostroma, whereas the areal of petux continues from the North into the center and southwest; see Reference KasatkinKasatkin et al. 2012: map 6), and (6) ‘ashes’ pópel ~ zola (note that variation between popel and pepel is also widely attested in Slavic, see Reference SaenkoSaenko 2017: especially figure 1).

There are a number of lexical isoglosses which tend to link Carpathian Ukrainian with ‘Balkan Slavic’ in sensu largo and which tend to overlap with very old accentual developments (see Reference NikolaevNikolaev 1994 for a basic list, as well as the relevant literature, especially Reference BernštejnBernštejn 1963; the ‘Balkan’ areal here basically encompasses W Bul., Mac., S, E, and N Štokavian, but not W Štok., Čakavian, Slovene, nor Kajkavian): *bričь ‘razor’, žirъ ‘acorn, beech nut’ (especially for feeding pigs); note that žir in Macedonian is dialectal, whereas the standard form is želad, as in Bul. žǝ́lǝd, Rus. ž‘ólud’, which etymologically appears to be the older of the two forms; *těrati ‘seek, chase after, drive’, *jarьmъ ‘halter, harness’, *osojь ‘sunny side of hill’, *tǫča ‘hail’, *sъnokti ‘yesterday evening, last night’.

28.8 Inter-Slavic Areal Features

Some phonological (and prosodic) developments in Slavic dialects exhibit an areal distribution over adjacent portions of different branches of Slavic. A well-known example is the lenition of voiced (lax) *g > γ (voiced velar fricative) [> h] in Upper Sorbian-Czech-Slovak-Ukraine-Western Belarusian-Southern Russian, as well as some of Slovene and Northwest Čakavian in Croatia (i.e. those South Slavic dialects which were geographically contiguous pre-historically in Pannonia with southernmost West Slavic). The dialects affected by lenition occupy a central position in the Slavic linguistic continuum, in contradistinction to dialects of the conservative periphery, where *g does not lenite, namely NW West Slavic (Polish, Lower Sorbian, Kashubian, and extinct Polabian), most of South Slavic, and the rest of East Slavic. The lenited forms are attested in Slovene toponyms by the end of the tenth century, but since the fifteenth century have receded in the eastern and central regions (Reference GreenbergGreenberg 2000).

However, an analysis of the phonological environments and areal distribution of the *g > γ change indicates an important difference in one of the positions of its occurrence (see Reference AndersenAndersen 1969, whose treatment of the question we follow here, while leaving aside differences in the phonetic details of the lenited *g, cf. γ vs. h). Whereas the outcome is generally uniform in (1) intervocalic position (*bogatъ > *boγat) and (2) following a so-called ‘weak jer’ (*sъgoda > *zγoda), in (3) the position following fricative *z, we find two distinct reflexes, namely (1) *mozgъ ‘brain’ > *mozg- (Resia [NW Sn.], E Moravia, Slk., Ukr., most of Bel., which form a contiguous ‘center within the center’) vs. (2) *mozγ-/*mozh- (NW-SSl other than Resia, Cze., N Bel., S Rus., where Czech represents the western ‘periphery of the center’, while N Belarusian and S Russian form the eastern edge thereof). The retention of -zg- is somewhat unexpected in the geographical center of the lenition, where one would anticipate -*zγ-.

Here we must bear in mind that for much of the first millennium AD the PSL inventory for permissible consonant clusters (such as -zg-) was limited to ‘fricative + stop’ (e.g. -zd-, -zg-, -st-, -sk-; an exception is the sequence of voiceless fricatives -sx-, as in *iz-xod- ‘go out’ > OCS isxoditi, isxod ‘exit’ and the like). This situation changed dramatically toward the end of the millennium when a large number of new clusters entered the language as a result of the so-called ‘loss of weak jers’, whereby the mid-high lax vowels ъ and ь were dropped in ‘weak’ position, that is, word-finally (*sъnъ ‘sleep, dream’ > Rus. son, Cze. sen, etc.) or (as pertinent here) when followed by syllables containing full vowels (*sъ-goda ‘early’ > Sln. zgodaj, *vъzgorьje ‘small hill’ > Rus. vzgor’je, *tъkati ‘to weave’ > Rus. tkat’, *bьdenьje ‘vigilance’ > Rus. bdenie). It was the creation of the new consonant clusters issuing from the fall of weak jers that removed the prior constraint upon combinations other than ‘fricative + stop’, thus opening the gate for *zγ- and providing us with a crucial diagnostic for the relative chronology of the lenition.

One can then apply the concept of variable rule-ordering (in our case, the two sound changes ‘lenition’ and ‘jer drop’) to the three positions of relevance for the treatment of *g in leniting dialects as described above (see Table 28.5). Assuming that lenition preceded the loss of weak jers in the ‘centermost’ dialect group (Chronology I), *g > γ would have occurred only in the first two positions, with the relative chronology for the second position being *sъgoda > *sъγoda > *sγoda > *zγoda (by voicing assimilation between fricatives), since /γ/ had been created before coming into contact with the preceding fricative /s/. Such is not the case in the third position, where the ‘fricative + stop’ rule blocked the change *mozgъ > *mozγъ.

Conversely, if lenition occurred after the loss of jers (Chronology II, which applies to the two peripheries, i.e. Czech in the west, and Northern Belarusian, Southern Russian in the east), then nothing would prevent the occurrence of sъgoda > *sgoda > *zgoda (by voicing assimilation) in the first stage, after which the ensuing phase of lenition (stage two) could occur without positional constraints, resulting in *boγat-, *zγoda, *mozγ-. This also indicates that the leniting impulse occurred as a wave, which passed from the center of the change to the periphery, rather than constantly recycling within the center.

A more circumscribed, but still powerful sphere of influence (this time in the domain of prosody) has been exercised by Polish, arguably upon Eastern Slovak (which, like Polish, has lost phonemic length in its vowel systems and generally stresses the penultimate syllable), but most certainly upon Lemko (from the idiosyncratic lexeme lem ‘only’) dialects of Ukrainian (spoken along the Polish-eastern Slovak border until 1947), where we usually find stress on the penult (po našómu ‘in our language’ [JS-conversation with Lemko emigree in Texas, 1990]; starósta ‘elder’, mołóda ‘young’ fem, každómu ‘to each one’ masc, na hołówu ‘onto (his) head’ (Reference KuraszkiewiczKuraszkiewicz 1963); compare Russian (which is generally conservative in matters of accent) nášemu, stárosta, molodá, káždomu, na gólovu/ná golovu.

Similarly, but more specifically induced by southern (Małopolska) Polish dialects, along both sides of the Carpathians as far as the San and Už river basins (and including some of the Hucul Ukrainian dialects), we find a particular phonotactic voicing rule, whereby voiceless obstruents become voiced and voiced obstruents fail to devoice in word-final position before words beginning with a vowel, or with a sonorant (l, r, m, n) or the palatal glide (j), for example (for voiceless obstruents) Hucul pładž-abo-śmix (*plač a[l‘]bo směx) ‘(you don’t know) whether to laugh or cry’, niz-mérzne (*nos mrzne) ‘(my) nose is freezing’, kid-utík (*kot utek) ‘the cat ran away’, tag-jeg- u̯ uyn (*tak jak on) ‘just like him’ (Reference KuraszkiewiczKuraszkiewicz 1963).

28.9 Prosodic Continua

The Slavic speech area is notable for its rich inventories of prosodic phenomena (including various combinations of pitch, quantity, and free word stress, as well as fixed or restricted stress). It is therefore not surprising that in some cases prosodic continua have formed across neighboring dialects within a single branch of Slavic or even within one language. Here we will consider two examples, both of which exhibit a longitudinal (here west–east) gradation of prosodic systems.

Our first example is drawn from West Slavic; we can diagram it as follows:

• Czech > Central Moravian > Western Slovak > Central Slovak > Eastern Slovak > Sotak

Czech has fixed initial stress with unrestricted distribution of long and short vowels, but with a generally higher concentration of long root vowels in some morphological and prosodic classes in SW Czech (see Reference VoráčVoráč 1955; note that in the examples cited below, Czech and Slovak long vowels are indicated with the acute diacritic). Central Moravian also has fixed initial stress, but with some shortening of high vowels (kúň/kuň ‘horse’, mlíko/mliko ‘milk’) and even of /a/ under the old acute tone (rak ‘crab’, krava ‘cow’, blato ‘mud’ vs. Cze. rák, kráva, bláto). West Slovak has initial stress, with regional variation in the quantitative reflex of the old acute (kráva ~ krava). Central Slovak (as well as Standard Slovak) combines initial stress with a ‘rhythmic law’, which tends to prohibit two successive syllables with long vowels by shortening the second of the two. Thus, for example, in Central Slovak even though the fem.def. adj.desinence is etymologically long due to vowel contraction (*-aja > *-ā), the manifestation of this length depends on the quantity of the preceding root syllable (cf. nová ‘the new (one)’ : krásna ‘the beautiful (one)’). East Slovak and Lacha Moravian show no long vowels and fixed penultimate stress, as in Polish. Finally, the peripheral (now mainly extinct) Sotak dialect has a locally variant mixture of free stress (as in Ukrainian, but with a different etymological distribution) and quantity. The Sotak dialect, which derives its name from the typical substitution of a fricative for an affricate (as in the interrogative pronoun [so] < *tso ‘what’), is little known except among Slovak dialectologists, despite its intrinsic historical and typological interest. Let us consider then the two main Sotak prosodic subtypes, both of which were attested in the Cirochou river valley in the mid-twentieth century.

A clear example of the ‘free stress’ Sotak type is the Humenné dialect of Modra nad Cirochou. Stress (marked here with the vertical ictus ̍) on the final open syllable is attested in PSL circumflex stems with original long root vowels: ucho̍ ‘ear’, meso̍ ‘meat’, hlava̍ ‘head’ nom., *hlavu̍ acc., z duba̍ ‘from the oak’, but not in na vo̍dźä ‘on the water’ (with short circumflex), or žä̍na ‘woman/wife’ (with short PSL oxytone) (see Reference LiškaLiška 1968, but also with reference to other research published in the 1940s).

As an example of the Sotak type with phonemic quantity under the accent, compare the neighboring dialect of Dlhé nad Cirochou, where long vowels /ā, ǟ, ọ̄, ẹ̄/ are pronounced more closed than their short counterparts, but also about one-and-a-half times longer. (This is reminiscent of Polish ‘pochylenie’, in which the reflexes of some of the medieval Polish long vowels are typically qualitatively distinct from their short vowel counterparts, but without any quantitative distinction.) Stress, though generally unrestricted, does not occur on final open syllables (perhaps an early phase in the development of a paroxytonic system). For examples, see Table 28.6, again drawn from data in Reference LiškaLiška 1968, who based his findings on spectrographic, oscillographic, and roentgenographic analysis of samples from the data he collected. For a derivation from PSL prosodemes, see Reference SchallertSchallert 2011.Footnote ¹

Our second example of a prosodic continuum is taken from the Macedonian speech territory in the southern Balkans, diagrammatically (Reference BaermanBaerman 1999):

• Western Mac (fixed) > Central Western > < Central < Easternmost Mac (free).

(Note that the arrows in the diagram indicate the basic directions of change.)

Fixed stress in the West is historically innovating and manifests itself in three varieties, falling on the penult in Korča in Albania (vodeníčar ‘miller’, def. vodeničáro ‘the miller’), on the initial syllable in Oščima near Lake Prespa in Greece (*vódeničar, *vódeničaro), and elsewhere on the antepenult as in Standard Macedonian (vodéničar, vodeníčarot), although the latter generalization holds mainly only for dialects spoken in the Republic of North Macedonia, as well as some immediately adjacent dialects within and to the north of the Šar Planina mountain range (in the extreme SW of Kosovo and a strip of Albania adjacent thereto) and to the south, extending several kilometers into NW Greece. This is to say that the rest of the Macedonian dialects in NW Greece belong to the Central-Western accentual system (see below).

The other typologically and historically conservative pole is represented by those Eastern Macedonian dialects situated more closely to the Bulgarian border, where dialects exhibit a free, lexically determined stress system comparable in most respects to Western Bulgarian (and, except for fixed root stress in the present tense of i-verbs, to Russian). These two typological extremes are separated by two intermediate zones, Central and Central-Western, in which phonological tendencies towards antepenultimate or penultimate stress interact with trends toward uniform stress in morphological categories or classes. The picture is further complicated by a tendency towards proclisis (initial stress) on verbal prefixes (usually just in the singular, but sometimes also in the plural) in some dialects in the aorist or (more commonly) the imperative.

The Central zone stretches in parallel with the Easternmost zone in a fairly broad band from southernmost Torlak in Serbia through Kumanovo in the north of Macedonia all the way to within a few kilometers of Thessaloniki in Greece in the south. Here lexicalized stress has been virtually eliminated, as root stress has been generalized in the present tense (péčeš ‘bake’ 2sg, cf. Rus. pečóš) in contradistinction to desinential stress in the aorist (pečé ‘baked’ 2-3sg, cf. BCS pȅče), while the accent has been retracted in PSL oxytonic nouns (rebró > rébro ‘rib’, žená > žéna ‘woman, wife’), although in districts adjoining the Easternmost zone, end-stress remains in the neuter type momčé ‘young man’, as well as in Turkisms such as džadé ‘road’, and polysyllabic suffixed feminine nouns, such as visočiná ‘height’ and slobodá ‘freedom’. In the Central zone, the imperative (with its simple paradigm and communicatively marked status) is the most volatile form in terms of a tendency to generalize one stress pattern; see root stress in the North (naprávi, naprávite), but initial (including prefixal) stress in the mid-Central region of Štip (nápravi, nápravite) and with oxytonic plural in most of the south (nápravi, napravéjte, with -ej- by analogy with the productive a-verb form -ajte).

The Central-Western zone (the second transitional type) runs next to the Western zone in a narrow strip across Southwestern Kosovo, then enters Macedonia to the north of Skopje, continuing down the Vardar river valley to Veles, while expanding southwestward to include the Tikveš and Mariovo regions, then turning fairly sharply to the west a few kilometers south of the border with Greece, before ending in the Kostur dialect of Vrnik in eastern Albania. In the CW zone (represented here by the Kajlar dialect in northern Greece; Reference DrvošanovDrvošanov 1993), although stress in the present remains on the root (nósam ‘carry’ 1sg, nósiš 2sg, nósime 1pl), it is generalized on the suffix in the highly productive -uva derivative class (dunisúvam ‘I bring’, dunisúvăme ‘we bring’). While the accent is retracted from the ending in the singular forms of the aorist (réku ‘said’ 1sg, réče 2-3sg vs. plural rikóme, -óte, -óa), in the imperfect, stress is advanced from the root onto the penult in all forms (vikáše 2-3sg ‘called’, vikáme 1pl, etc.), including the 1sg viká, which is leveled with the rest of the paradigm (cf. 2-3sg vikáše) and which may have ended in a closed syllable (*vikáx) at the time of the retraction from final open syllables. These changes in the preterite paradigms reflect an underlying trend towards the generalization of penultimate stress. Note that some of these preterite forms once had (and in some dialects still do have) long vowels due to compensatory lengthening attendant upon the loss of /x/ (rikó:me, rikó:te from *rekóxme, *rekóxte or imperfect vikā́me < *vikáaše, -áa[x]me). In some CW dialects the degree of lexical variation is quite high, in both the noun and the verb, as competing stress patterns overlap, suggesting subsystems still in the process of formation.

Early evidence for fixed stress systems in Macedonian is quite rare, but paroxytonic tendencies in the Central-West zone can be dated to the sixteenth century, as attested in a Macedonian-Greek lexicon and brief phrase book based on the Kostur dialect in southwest Aegean Macedonia (Reference Giannelli and VaillantGiannelli & Vaillant 1958). Here we find forms such as pl. golóbi ‘pigeons’ (sg *gólob); grében ‘comb’ : pl grebéni, as well as the stress shift caused in the imperative singular by the addition of an enclitic Postelí mi posteláta ‘Make my bed!’ (from *pósteli + mi, where initial stress in the imperative singular as the starting point in our derivation is suggested by other examples such as óstavi ni da spime ‘Leave us alone to sleep’). Such competing patterns may have contributed to the development of the so-called ‘double stress’, which is so widespread in parts of Aegean Macedonian and some western Rhodope dialects (e.g. Suxo [to the east of Saloniki] v’éč́er ‘evening, pl. v’éč́eróvi) and which has parallels in Greek (e.g. o δáskalos ‘the teacher’ : o δáskalós mas ‘our teacher’).

The origins of fixed antepenult and penult stress have not yet been determined with any certainty, although scholars have sought rough parallels in the framework offered by neighboring languages such as Greek, where stress is restricted to one of the last three syllables of the phonological word (more precisely to one of the last three morae or units of duration, since Ancient Greek had both short and long vowels). Of course, in the Greek noun at least, this ‘tri-moraic’ law allows for the existence of lexically specified end-stressed oxytona (ModGk γerós ‘robust’), penult-stressed paroxytona (ModGk γéros ‘old man’), and antepenult-stressed proparoxytona (ánthropos), whereas in the verb, so-called ‘recessive’ stress on the antepenultimate mora is generally consistent (Reference MackridgeMackridge 1985). An Ancient Greek exception is the Aeolian dialect of Lesbos, where stress is also recessive in the noun (e.g. pótamos ‘river’ vs. Ionic potamós; Reference Goodwin and GulickGoodwin & Gulick 1930/1958).

28.10 The Role of External Factors (Economic and Political)

The spread of dialect features is often driven by economic and/or political factors. An example of economic motivation is the quest for more arable land which attracted speakers of the northeastern Mazovian Polish dialect to the south (Lesser Poland/Małopolska), resulting in the spread of the phonological feature known as mazurzenie (the merger of alveolar /š ž č dž/ with dental /s z c dz/, e.g. szósty > sósty ‘sixth’, etc.) (Reference StieberStieber 1973). The structural-linguistic factor at play here is the pressure exerted by the recently formed pre-palatal /ś/ (siadam ‘I sit down’, siano ‘hay’) upon the phonological space previously occupied by postalveolar /š/. Since this factor was latent in all Polish dialects at the time, the change was readily adapted even by speakers of other neighboring dialects. Less productive was the northernmost Mazovian (e.g. Kurpie) assibilating palatalization of labial stops, for example ṕśęć < pięć ‘five’ (a unique development in the Slavic speech area, but one with some parallels in Romance, cf. French savoir : sache [saʃ] < *sapere : sapia-, rouge [ruʒ] < Rubia-]), which remained confined for the most part to Mazovia (Reference FriedrichFriedrich 1955). It is nonetheless notable that the twentieth-century western isogloss for mazurzenie conforms generally with the political boundaries which existed prior to the thirteenth century between Great Poland (Wielkopolska) (in the west) and most of the remaining regions of medieval Poland (Reference StieberStieber 1973).

A historical sequence of economic and political factors fostered the expansion of the Serbian Eastern Herzegovinian dialect, which eventually came to serve as the basis for one of the varieties of literary Serbian, as first encoded in Vuk Karadžić’s Serbian dictionary with German and Latin glosses (published in Vienna in 1818, but here cited from the 1852 edition, as per the reprint of 1935). One of the most notable features of Vuk’s dialect was the so-called jekavian pronunciation of the reflex of the Common Slavic jat’ vowel (vjȅra ‘faith’ < *vě̋ra, brȉjeg ‘hill, bank’ < *brě̑gъ < *bȇrgъ), although Vuk’s dictionary also records ekavian (vȅra, brȇg) and ikavian (vȉra, brȋg) variants from other dialects. Another prominent Eastern Herzegovinian trait that one finds systematically indicated by Vuk is the so-called Neo-Štokavian accent retraction, which also generated secondary short and long rising pitches on the newly stressed syllables (e.g. ženȁ ‘wife’ > žèna, rūkȁ ‘hand’ > rúka). Difficult economic conditions in the highlands of the Dinaric mountain range had always tended to stimulate gradual resettlement to the lowlands to the east and the Dalmatian coast to the west, but this trend was given a powerful fresh impulse by the vicissitudes attendant upon the Turkish conquest of the Balkans, which imparted a northwestern vector to the spread of the Eastern Herzegovinian dialect. These events culminated in the great Serbian migrations (seobe), which followed the unsuccessful outcomes of the two Habsburg-Ottoman Wars (1690, 1737–1739).

28.11 Fragmentation Induced by Contact with Non-Slavic Languages

Historical change in the Slavic dialects has also been shaped by prolonged contact with non-Slavic languages. As an example of contact phenomena on the northern periphery of East Slavic (in this case generally attributed by scholars to a Finnic substratum in Karelia), we may cite the case of so-called ‘ljapan’je’, which covers a range of prosodic and phonological features in Zaonežje North Russian dialects (Reference MixajlovaMixajlova 2019); see the widespread stress shift from the final syllable to the initial syllable, leading in some cases to the accentuation of original pretonic vowels which had undergone jakan’e vowel reduction at an earlier stage, for example Pjáškom s mjáškom do Medvjášky ‘On foot with a sack to Medveška’ (standard Rus. Peškóm s meškóm do Medvéški). The phenomenon was first observed by early Russian folklorists, such as Rybnikov and Gil’ferding (see Reference Ter-Avanesova, Bulatova, Dybo, Nikolaeva and ZaliznjakTer-Avanesova 1989, which also provides a detailed diachronic analysis of ljapan’je, including typological parallels with similar stress retractions in BCS dialects in Slavonija and the Drava river basin). This stress type also occurs in phrasal constructs with proclitics, for example n‘á bojus’ ‘I’m not afraid’. Perhaps more unusual in the Slavic context is the rise of a prothetic /g/ before initial mid-front vowel in this same kind of dialect, for example gésli ‘if’ (jesli), as in one type of Veps dialect (where the continuum is järva > därva ~ gärva ‘lake’).

Several other features of Russian or its northern dialects have been attributed (albeit with varying degrees of probability) to a Finnic substratum. These features include: (1) the formation of the so-called ‘second (partitive) genitive’ and ‘second (locative) prepositional’ cases, whose desinences have been recycled from the original u-stem endings (as in standard Russian kiló sáxaru ‘a kilo of sugar’ vs. vkus sáxara ‘the taste of sugar’, and iskál v snegú ‘he searched in the snow’ vs. govoril o snége ‘he talked about (the) snow’); (2) the use of a locative phrase, rather than the verb ‘have’, to indicate possession, again as in standard Russian u menja est’ dom ‘I have a home/house’, as in Finnish Minulla on talo (*iměti ‘to have’ in the rest of Slavic, including Ukrainian and Belarusian); (3) in Northern dialects, the use of the nominative case as a direct object (Kosi travá poka rosá! lit. ‘Mow grass while the dew’s still fresh’/‘The early bird catches the worm’) and the nominative object of infinitive, as well as (4) the use of resultative perfects such as U volkov zdes’ xoženo ‘Wolves have been walking/prowling here’ and On privykši ‘He is used to it’ (the two types overlap only in Karelia), both in contradistinction to standard Russian concordant L-preterites (guljali, xodili, brodili : privyk < privyklъ); and (5) the use of a postponed demonstrative pronoun (*t-) in a fashion which bears some resemblance to the development of the definite article in the history of Bulgarian and Macedonian (domo-t ‘the house’, domu-tomu ‘to the house’), particularly in the influential and widely popular writings of the mid-seventeenth-century priest Avvakum, who grew up in a region which included a Mordvinian-speaking population (Reference KiparskyKiparsky 1968).

In South Slavic, interaction with Albanian in a context of extensive bilingualism along the border with Montenegro may have supported retention of /ä/, which can be reconstructed as a plausible intermediate stage between /ə/ and /a/ in an archaic reflex of the jer vowels (dä̑n ‘day’ < *dьnь, sä̏n ‘sleep’ < *sъnъ), but this same interaction also promoted innovations such as the devoicing of consonants in auslaut (grop < grob ‘grave’, noš < nož ‘knife’) and the shift of the palatal lateral continuant to a palatalized variant (ljúljati ‘rock’ [ʎuʎati] > l’ul’at; Reference IvićIvić 1994). Similarly, it has been suggested that peripheral southwestern Macedonian dialects of the Kostur region retained the nasal feature in forms such as réndo ‘the row’, zə́mbi, zɔ́mbi, zǻmbi ‘teeth’ through decomposition of *ę, *ǫ into ‘vowel + nasal stop’ at a stage prior to the general denasalization which affected the rest of the Macedonian dialects, and that this ‘archaic innovation’ transpired due to centuries of contact with speakers of neighboring dialects of Aromanian, Albanian, and Greek, all of which have undergone prenasalization of voiced stops (in the case of Greek, from as early as the fourth century AD, when -mp- > -mb-, -nt- > -nd-, etc., for example seventh-century papyrus <Pondikón> ‘mouse’ < Pontikon, thirteenth-century <loumpardoí> < ‘Lombards’; Reference Arvaniti and JosephArvaniti & Joseph 2000). The hypothesis is supported by a correlation between the timbre of the root vowel in Macedonian and in the geographically relevant contact language. Thus, central and northern Kostur Macedonian, Albanian, and Aromanian all have a schwa-like vowel in this context (e.g. Mac. Northern Kostur zə́mbi ~ Alb. Tosk dhëmp ‘tooth’, dhëmbi ‘the tooth’, by chance an IE cognate of PSL *zǫbъ; Arm. Farsheroţi plăɲg^ụ ‘I cry’), whereas southeastern Kostur Macedonian (in this context) and Greek (altogether) lack such a vowel (e.g. Kostur Mac. zɔ́mbi ~ Gk. /a/ lambo ‘I shine’, WMac. Gk. /o/ [đond] ‘tooth’ [Reference NewtonNewton 1972: 149]) (Reference Friedman and BethinFriedman 2018, Reference LindstedtLindstedt 2016).

This same region of SW Macedonia is also the epicenter for a set of radical innovations in the verbal system, one of which is the development of a new series of perfect tenses formed with the verb ima (‘have’) and sum (‘am’, etc.) in combination with past passive participles, for example Imam bideno ‘I have been’, Imam imano ‘I have had’, Sum dojden ‘I have arrived’, and even Veḱe sum jaden ‘I have (literally ‘am’) already eaten’, all forms with differing degrees of productivity in Western Macedonian. Resultative perfects formed with ‘have’ as the auxiliary are typical of Aromanian, Albanian, and Greek, and their particular prosperity in Western Macedonia is generally attributed to intense contact with these non-Slavic languages in the context of the Balkan Sprachbund (linguistic league). Macedonian, in its turn, has arguably contributed to the formation of perfects with ‘be’ as the alternative auxiliary in Aromanian (Reference GołąbGołąb 1984), although the presence of both ‘have’ and ‘be’ perfects with lexically determined diathetic functions is typical of Romance languages in general.

Another well-known bilingual contact zone in South Slavic is the eastern Adriatic coast and adjacent islands, where centuries-long interaction with the politically and culturally dominant Venetian dialect of Italian has played a significant role in the formation of the Cakavian variant of the Čakavian dialect of Croatian. Cakavian is attested from as far north as the Labin district in the southeast of the peninsula of Istria to as far south as the island of Vis, occurring principally, but not exclusively, in the towns, which is where the Venetian administrators and colonists were concentrated (Reference MałeckiMałecki 1929, Reference MogušMoguš 1977: 66–79). The influence of Venetian was particularly strong during the fifteenth to eighteenth centuries and is manifested on a variety of linguistic levels, including the lexicon and morphology (see below for examples), as well as syntax and verbal semantics (Reference KalsbeekKalsbeek 2011 e.g. cites (1) clause-initial clitics, ću te kumpanjit ‘I will accompany you’, (2) the habitual preterite suffix -eva-, smo čuvievȁle skupa ‘We used to tend our flocks together’, (3) a simplified lexical system for verbs of motion, where voditi ‘lead/guide by hand’ and voziti ‘convey in vehicle’ are replaced by peljati ‘drive, impel’ (cf. Ital. condurre ‘lead; drive’), (4) substitution of adverbs for verbal prefixes denoting direction, da ne grȅ nuõtre ‘that (s)he not go inside/enter’, (5) purpose clauses with za ‘for’ + inf as in Italian per + inf, although this construction is more widespread). It is nonetheless a prominent phonological trait which lends its name to Cakavian, as reflected in the iconically typical pronunciation of the interrogative pronoun ca [tsa] ‘what’ (with variants co, ce, more rarely će [tśe]), corresponding to the geographically and demographically more widespread Čakavian ča [tša]). The historical substitution of a dental affricate /c/ for the palato-alveolar affricate /č/ that we see in this example is part of a broader tendency, which also affected the palato-alveolar fricatives /ž, š/, shifting their articulation to the pre-palatal /ź, ś/ or dental /z, s/ region, depending on the local dialect, for example žena ‘woman, wife’ > źenȁ, zena, široko ‘wide’ > śirõko, siroko (for phonetic details, see Reference MogušMoguš 1977). In those Cakavian dialects where /ź/ rather than /z/ replaced /ž/, the shift to pre-palatal articulation also often affected the dental sibilants, e.g. selo ‘village’ > śelo, zelen ‘green’ > źelẽn. In such dialects we find words in which both an original dental and an original palato-alveolar have been changed, resulting in extensive merger with /ś, ź/, for example šest ‘6’ > śȇś, železo ‘iron’ > źeleźo. An excellent example of a consistent Cakavian dialect of the last-mentioned type was recently recorded in the speech of an older female informant in the village of Štrmac (near Labin in southeast Istria; Reference NežićNežić 2013: 419–420): *č > ocȗva ‘keeps’, docȅkat ‘wait until’, *c > ocȁ ‘father’ gen., *ć > kȕća ‘house’, ćȅmo ‘we will’, *š > źnȏś ‘you know’, śȇś ‘6’, ślȁ ‘she went’, *s > śo ‘they are’, śe ‘self ’, śadȁ ‘now’, *ž > oźenȉli ‘they married’, źȉvi ‘alive’ pl., *z > źa ‘for’, źȋt ‘wall’. Cakavian since the early twentieth century has generally been in retreat before local Čakavian and literary Štokavian, while at times exhibiting nearly free variation between the possible outcomes of the fricatives (as documented, e.g. for Vis [Reference Moskovljevič and BulatovaMoskovljevič 1972, based on fieldwork from 1928] and Labin [Reference NežićNežić 2013]).

This elimination of the ‘hushing’ series /ž, š/ brought the Čakavian system into closer alignment with that of Venetian, which never possessed or developed /ž, š/. On the other hand, the introduction of /ś, ź/ in some Čakavian dialects seems to reflect a different kind of adaptation, since the corresponding Venetian sibilants, though commonly designated /s, z/ at the phonemic level, have often been described in phonetic terms as situated in their articulation and acoustic impression between [s, z] and [š, ž] (Reference BidwellBidwell 1967, Reference PiccioPiccio 1928, Reference MałeckiMałecki 1929). This is consistent with Reference Canepari and CortelazzoCanepari (1979), who identifies a type of Modern Venetian pronunciation of /s, z/ in similar, but more precise terms as ‘apico-alveolar’ sibilants articulated farther back than the dentals [s, z], that is, as alveolar retracted (not ‘retroflex’!) sibilants. Since the IPA alphabet lacks separate symbols for this articulation, one might have recourse to [s̠, z̠] or even ‘ṣ, ẓ’ (following Małecki and the earlier Italian linguists). It is conceivable that this pronunciation is not a recent innovation, as suggested by typological parallels elsewhere in Romance (e.g. Old French pousser > English push) and by the phonology of Venetian loanwords in Cakavian, for example Northern Cres (Reference VelčićVelčić 2003) śalvḁ̑n ‘saved’ (< Ven *salvo, cf. Ital salvato with morphological substitution of the Slavic past passive participle suffix /-an-/ for Romance /-at-/), and žvȇlt/źvȇlt ‘fast, quick’ < Ven zvelt (cf. Ital svelto).

The linguistic impact of long-term contact with German is felt most profoundly in West Slavic (especially Czech and Sorbian) and northern West South Slavic (chiefly Slovene and Kajkavian). Depending on local circumstances, this impact is reflected in various components of the languages involved (phonological, morphosyntactic, aspectual, and lexical). Some of the influence was reciprocal (as in Carinthia, between local Bavarian dialects and the northernmost Slovene dialects). Aside from possible prehistoric Iron Age Germanic-Slavic contact/mixture in the west in the context of the Przeworsk culture of the late La Tène to Roman times (third century BC to fifth century AD) (Reference Nichols, Maguire and TimberlakeNichols 1993), this contact was initiated or maintained by the Slavs in the course of the Slavic Migration, but by the ninth to tenth centuries the demographic and political tide had turned, such that at a later date, German colonies and concentration in urban centers such as Prague, Ljubljana, and Zagreb/Agram, as well as the smaller towns, served as points of radiation for the politically and culturally dominant language.

To begin with a West Slavic example, we may consider colloquial Prague speech (Reference TownsendTownsend 1990), where we find the use of a perfect tense formation involving ‘have’ (German haben) + past passive participles in two types of construction, the first with the short form of the participle in neuter gender, for example Máme tady rozsvíceno ‘We have the lights turned on’, Mám tady uklizeno ‘I’ve tidied/cleaned up here’, the second with the long form of the participle marking concord with topicalized (sentence-initial) direct object, for example Referát už mám napsanej ‘The paper/report I’ve already written’ (note that in standard Czech this last example would have a short form predicate participle (napsán), just as in the first type, whereas the long form would more often be attributive; in German there is no such distinction marked directly on the participle itself, a factor which may have played a role in the erosion of the formal difference in Prague Czech). In the noun phrase, demonstratives in colloquial Prague Czech can be used in ways that resemble colloquial and even standard usage of the article in German, for example Ten Honza je už dlouho pryč = Der Honza ist schon lange weg, or to replace the third person pronoun in emphatic (?) topicalized position, for example Toho jsem ale neviděl ‘But him I didn’t see’ = Den habe ich aber nicht gesehen. In the Czech Doudleby dialect, the usage of the demonstrative at times resembles a definite article, for example ti [those-Masc.-nom.-pl] tahouni se používali po sekaní obilí … ‘the draft animals were used after harvesting the crops …’ (where the reference to draft animals is generic, since there has been no prior allusion to them). More broadly in the Czech lexicon, loanwords and phraseological calques from German are evident at different stylistic levels: colloquial Cze. kšeft < Geschäft ‘business’ or Mám smůlu < Ich habe Pech ‘I’ve got pitch (tar)’ that is, ‘I’ve got bad luck’, learned calques (obrození < Wiedergeburt ‘Rebirth, Renaissance’, výlet < Ausflug ‘excursion, outing’), and high-register borrowings (as in the use of říše neut. ‘empire, kingdom, realm’ < das Reich, which dates back to Old Czech and is one of the rare Germanisms to be found in Mácha’s “Máj” (1836), cf. line 3 of canto 2, Klesla hvězda s nebes výše … padá v neskončené říše … ‘A star fell from the heavens on high … It falls into the endless kingdom …’).

Turning to Slovene, we begin with phonetic innovations which are typical of the contact zone between Bavarian and Slovene as spoken in Carinthia (Austria), but which are not found in other Bavarian or Slovene dialects (Reference GreenbergGreenberg 2000). These include shared innovations, such as (1) uvular trilled [R] and (2) g, k pronounced as laryngeals [h, ʔ], but also those which Bavarian has adopted due to contact with Slovene, such as (1) merger of /h/ and /x/, (2) lengthening of short stressed root vowels (similar to the well-known Slovene type bràta > bráta), and (3) diphthongal pronunciation of short e and o (Carint. Sln. smréaʔa ‘juniper’ ~ Carint. Bav. schean ‘schön’, Carint. Sln. ʔoaža ‘hide, leather’ ~ Carint. Bav. roat ‘rot’). In morphosyntax, a characteristic feature of several prominent sixteenth-century Slovene Protestant Bible translators is the use of a fully inflected, morphologically demonstrative definite article (*t-) in a fashion which closely resembles that of the German article, for example from Primož Trubar’s preface to his translation of the New Testament (1557): Tiga [masc.gen] Noviga testamenta ena [fem.Nom.] dolga predguvor ‘An extended preface to the New Testament’ (with topicalization of the genitive noun phrase and in addition the use of an ‘indefinite’ article, based on the numeral ‘one’), Ta [masc.nom] pervi deil te [fem.gen] slovenske dolge predguvori ‘The first part/Part one of the extended Slovene preface’, and Na to [fem.acc] pervo nedelo tiga [masc.gen] adventa ‘On the first week of Advent’ (contrary to English usage, Advent takes a definite article in German). This usage was somewhat pejoratively characterized by Kopitar in his 1808 grammar of Slovene as typical of German-influenced town speech (in contrast with the more authentic Slovene spoken in the countryside) and even Slovene Protestant writers of the next generation, such as Dalmatin, made less use of the form. Although the definite article has long since been banished from the standard language, an interesting and highly idiosyncratic reflex of the masc.nom.sg /ta/ is widely used in modern colloquial Slovene to modify adjectives (such as zelen ‘green’) and ordinal adverbs (such as prvič ‘for the first time’), but never the bare noun (*ta svinčnik ‘the pencil’), which proves that it is not a true definite article. Though derived from the demonstrative *t-, uninflected /ta/ is readily distinguished from the latter in minimal pairs, such as pr [*pri] ta lepem ‘with the nice one’ vs. pr tem lepem ‘with that nice one’ (example provided by Marta Pirnat). The form /ta/ can also co-occur with demonstratives, for example téga (gen) ta zelenega svinčnika ‘of this green pencil’ and in other unexpected combinations (see Reference Marušič and ŽaucerMarušič & Žaucer 2008).

The mutual interactions of Slavic (chiefly Czech, but also the rest of West Slavic, and Slovene, but also Kajkavian and Croatian) and Germanic (most intensely the Bavarian dialect) have also been viewed by scholars within the construct of a Central European (‘Danubian’ or ‘Carpathian Basin’) Sprachbund, which includes other languages of the region, such as Hungarian and Yiddish (see Reference ThomasThomas 2008 for a thorough survey and methodological study). Among many other features, both phonological and morphological, which have been attributed either to long-term Germanic influence or to participation in the Sprachbund (as underlying or contributing factors), we may note the development of different kinds of fixed stress in West Slavic (see above), the reinforcement of conservative traits in verbal aspect such as the use of the imperfective aspect in sequences of events (a trait which serves to distinguish West from East Slavic; see Reference DickeyDickey 2011), and the use of a ‘double perfect’ to form the pluperfect, for example Cze. byl jsem videl ‘I had seen’, Sln. Padel sem bil ‘I had fallen’, and Kajkavian ja sam ti bil pisal ‘I had written to you’ (cf. Yiddish, and Bavarian dialects, Er hat gǝhat gǝtrunkǝn ‘He had drunk’).

28.12 Sociolinguistic Factors

One sociolinguistic correlate of dialect variation which is reinforced by language contact is that of religious confession. Examples of this phenomenon are particularly common in South Slavic with its long-standing confluence of different creeds. Thus, according to Reference IvićIvić (1994), in the consonant systems of Štokavian dialects in Bosnia and Herzegovina, one archaism (the retention of *x, sometimes as /h/) and one innovation (the loss of a distinction between pre-palatal /ć, dź/ and postalveolar /č, dž/) are both far more common in villages with a predominantly Muslim Slavic population than in those whose population is Orthodox or Catholic, a circumstance which has probably been shaped in part by more intense exposure to Turkish (which has /h/, but lacks pre-palatals) and perhaps even Arabic (as mediated via Islamic practices of daily occurrence). In Mostar, an important administrative center with a significant Muslim population, even the Orthodox population has retained /x/.

A particularly striking example of dialect preservation and innovation in the context of a strongly insular religious community is that of the Pavlikjan Bulgarian dialect, spoken by descendants of Armenian and Syrian adherents of the Paulician Christian heretical sect, some of whose members were exiled by the Byzantine authorities in the eight to tenth centuries to the Balkans, where they settled in the south near the city of Plovdiv (where their descendants are known to linguists as the Southern Pavlikjans; see Reference MiletičMiletič 1900). In the course of time (but beginning perhaps as early as the fifteenth century), some of the Pavlikjans migrated due north through the central Stara Planina (‘Old Mountains’) to the Danubian plain near Svištov and Nikopol (regarding these Northern Pavlikjans, see Reference NedelčevNedelčev 1994). Due to the successful efforts of Franciscan missionaries, the Pavlikjans (‘Palk’éne’, in their native dialect) converted en masse to Catholicism in the mid-seventeenth century, a decision which subsequently exposed villages of the Northern branch to Ottoman Turkish reprisals following the unsuccessful Čiprovci rebellion of 1688 (undertaken principally by Bulgarian Catholics in the Northwest). As a result, many of the Northern Pavlikjan Catholics fled across the Danube to Romania, some settling in the east near Bucharest (Popešt-Leorden; see Reference Petrovič and VrabiePetrovič & Vrabie 1963, 1965; Reference MladenovMladenov 1993), but with the permission of the Austro-Hungarian authorities mostly in the west in the province of Banat in 1738–1741 (Bešenov, Vinga; see Reference MiletičMiletič 1900, Reference StojkovStojkov 1967).

Remarkable features encountered in Pavlikjan vocalic systems include (1) the raising of the PSL mid vowels (*o > u, e.g. *kon’ > kun ‘horse’, *e > i, mit ‘honey’ in Popešt-Leorden and in Miletič’s description of Southern Pavlikjan) and (2) preservation of PSL *y in forms similar to Ru. ы, Ukr. <и> =[ɩ], and [ə]. These vowels also tend to serve as common reflexes of *i after depalatalized consonants (pɪsmo ‘letter’, gŏdǝ́na ‘year’). Particularly notable is the treatment of *y after labial consonants in Banat and Southern Pavlikjan in isolated examples such as [vu] ‘you’ pl. < *vy (not *vie, as in most Bul. and Mac.) and [muja] ‘I wash’ trans. < *myja. This apparent relic is otherwise attested in South Slavic in the Old Slovene Freising Fragments (tenth century) <mui> *my ‘we’, and the SE Mac village of Visoka (Reference MałeckiMałecki 1936) múja, mújiš́. Reference BrochBroch (1899) also describes a similar reflex (high-mid rounded /ω/) in the Eastern Slovak-Rusyn dialect of Koroml’a in the Carpathians (e.g. mωt’ ‘to wash’, bώu̯at’ ‘to live, dwell’, gríbω ‘mushrooms’, cf. Russian myt’, byvat‘), although one must bear in mind that /ω/ in the Koroml’a dialect also originates from other sources, such as (probably lengthened) *o in rωch ‘horn’.

To conclude this chapter, we find it is not surprising that another sociolinguistic variable which impacts dialect performance is gender. Thus, according to Reference VidoeskiVidoeski (2005: 105), the pronunciation of the mid-vowels /e, o/ as low-mid [ε, ɔ] (bέlo ‘white’ neut., kɔ́sa ‘hair’) is a general tendency in Northern Macedonian dialects (such as Kumanovo and Kriva Palanka). Although unfamiliar with Vidoeski’s observation at the time, in the course of fieldwork conducted in 1983 in the village of Luke (formerly Drží kon’, literally ‘Hold the horse!’, immediately to the south of the border with Serbia in northeastern Macedonia), I observed that this type of vowel was used consistently in the speech of one of two young informants (male), but not at all in the speech of the other (female), whose mid vowels were pronounced slightly high-mid [e, o]. Both speakers had full command of other distinctive local dialect traits in morphology and word stress, which differ markedly from those of Standard Macedonian. Similar kinds of observations on gender-correlated phonetic differences in the pronunciation of vowels can be found in the commentaries to the various volumes of the Bulgarian Dialect Atlas. As an example, according to the commentary to BDA-SE (Map 9), in the town of Zlatograd in the SE Rhodopes, instead of the usual dialect (and literary) pronunciation žə́na ‘I reap’, older Christian female informants say žė́na (where ė is identified as a quite narrow vowel, yet pronounced with medium ‘lip rounding’).

The contact history of Slavic is probably as old as the Slavic language family itself, running back to prehistoric times when Slavic split off from Balto-Slavic (or directly from Proto-Indo-European, if one doubts the existence of Balto-Slavic).Footnote ¹ On the basis of lexical loans and a few structural features, we can reconstruct prehistoric contacts with Uralic, Germanic, and Iranian. Recently Reference GvozdanovićGvozdanović (2009) has argued for prehistoric contacts with Celtic, but these were rejected by most critics (Reference DaiberDaiber 2012, Reference VathVath 2013). More promising is Reference HolzerHolzer’s (1989) reconstruction of a prehistoric layer of loans in Slavic – and to a lesser extent in Baltic – from an unidentified Indo-European language, which Holzer calls Temematic and attempts to link to the Cimmerians mentioned in early Greek sources.

When the Slavs enter history, the social preconditions of language contacts become more visible. Multiethnic and multilingual empires rise all over Europe, and many if not most Slavic languages became involved in complex speech communities, giving rise to language contact areas in the Balkans and possibly also in Central Europe. Whereas in Southeast and Central Europe Slavic languages were impacted by dominant imperial languages (German, Turkish, Greek), as seen in lexical borrowing and ephemeral forms of mixed learner varieties, Russian became itself an imperial language, assimilated speakers of various Finnic languages and endangered many indigenous languages. Russian is also the only Slavic language that served as a lexifier for pidgin languages. With the rise of literacy in Slavic, lexical loans as well as structural patterns from major languages of learning (Greek, Latin, later replaced by Western European languages) were borrowed massively; mutual exchanges ensued among the Slavic languages.

29.1 Prehistoric Contacts

29.2 Slavic within the Multilingual Empires of Europe

29.3 Written Language Contacts

Language contact is a key feature of the very beginnings of Slavic literature, which was almost exclusively translations of Byzantine Greek clerical texts into (Old) Church Slavonic. Medieval translators applied more or less literal translation leading to syntactic copies of the Greek model (Reference BirnbaumBirnbaum 1996) and to frequent loan translations of Greek compound nouns (Reference MolnárMolnár 1985). However, the rigid barrier between sacred and profane, which on the linguistic level took the form of diglossia (Reference UspenskijUspenskij 1983, Reference LuntLunt 1987), counteracted the spread of Greek loan features into registers of wider communication. Diglossia gradually dissolved, so that in the Polish-Lithuanian Commonwealth a lexically mixed register (Ruthenian, Polish, Church Slavonic) for confessional polemics and catechetic literature, known as prosta mova, emerged from the late sixteenth century onwards (Reference MoserMoser 2002).

The modern age saw an increasing westwards orientation of the cultural elites of Eastern European Slavic-speaking countries, most markedly in Russia. Access to refined Western culture was initially sought via the Slavic-speaking immediate neighbors to the West. Before the Petrine reforms paved the way for accessing Western culture through Western languages directly, a steady influx of lexical items from Polish into Russian texts in writing is observed throughout the seventeenth century (Reference KochmannKochmann 1967, Reference WitkowskiWitkowski 1999), as are effects on syntax and style (Reference MoserMoser 1998).

With the Petrine reforms, direct loans of mostly technical terminology from Dutch (Reference Van der Meulenvan der Meulen 1909, Reference Van der Meulen1959), German (Reference BondBond 1974, Reference OttenOtten 1985), and to a lesser extent English (Reference WójtowiczWójtowicz 1993) entered Russian through translations, whereas the role of a refined model language for the higher classes was taken over by French from the mid-eighteenth century. The strong presence of French in the social life of Russian elites throughout the eighteenth and nineteenth centuries is overt in many Russian literary works of that period and can be traced in the syntax (Reference Hüttl-FolterHüttl-Folter 1996) as well as in the semantic structures and phraseology (Reference SmithSmith 2006, Reference Offord, Offord, Ryazanova-Clarke, Rjéoutski and ArgentOfford 2015) of modern literary Russian.

The Age of Nationalism also inspired large-scale attempts at language engineering in order to reslavicize those Slavic languages which had been subject to long-standing cultural and linguistic domination. Turkisms (Reference HenningerHenninger 1990) and Germanisms (Reference ThomasThomas 1996) became major targets of linguistic purification, and Russian as the one language that never underwent cultural domination came to serve as the major source of reslavicization (Reference Giger, Kosta and WeissGiger 2008: 126). Thus, Kollár’s idea of balanced Slavic mutuality (Reference Kollár1844) involuntarily turned into a policy of replacing one model of cultural-linguistic domination by another one, which was believed to restore lost authenticity on the basis of genetic proximity. Slavicizing purification efforts focused primarily on the lexicon, but also included phraseology and grammar.

29.4 Outlook: Topics for the Age of Globalization

The age of globalization is marked by high international mobility, migrations, and resettlement of workforces and refugees alongside the establishment of English as language of the globalizing world. In one way or another all these aspects of globalization have found reflection in recent research on Slavic linguistic contacts, but the overall amount of research done is still modest. The opening of borders in the post-socialist space has given rise to vibrant local cross-border economies, mostly within the informal sector, but only Russian-Chinese border traffic has so far attracted the attention of contact linguists (Reference Fedorova, Bruns and MiggelbrinkFedorova 2012, Reference NamsaraevaNamsaraeva 2014, Reference OgleznevaOglezneva 2007, Reference Oglezneva2014, Reference SternStern 2015, Reference Stern, Kamusella, Nomachi and Gibson2016). Soviet and post-Soviet emigration to the West, especially where it entailed permanent resettlement of larger groups of Russian speakers, has attracted research on language maintenance and attrition (Reference ZemskajaZemskaja 2001; see also in this volume Chapter 31, on Slavic heritage languages). Perhaps the best-documented and researched case of a Slavic-speaking immigrant community is that of the Russian Germans or Spätaussiedler, who adopted a bilingual in-group practice of conversational code-switching alongside an emergent mixed code with a Russian matrix (Reference BerendBerend 1998, Reference Brehmer and AnstattBrehmer 2007). Finally, the impact of Global English on Slavic languages has brought numerous studies addressing recent changes of language use and style that are underway in subcultures and specialized professions under the impact of English (Reference Pfandl, Muhr and KettemannPfandl 2002, Reference Rathmayr, Muhr and KettemannRathmayr 2002, Reference MečkovskajaMečkovskaja 2007). It is here that we perceive that while past contact is accepted as a fascinating ingredient of cultural history, present contact remains even for professional linguists a deplorable sign of cultural threat and decay (Reference GriffinGriffin 1997, Reference Griffin2001, Reference GorhamGorham 2000).

30.1 Definition of the Slavic Literary Micro-Languages (SLMLs)

In addition to the major Slavic languages that are treated in any handbook on Slavic languages, there are additional Slavic languages – usually used by Slavic minority groups – with (to some extent) established traditions of literacy. In the past, these languages were often dealt with under dialectology, with a special note that a given dialect has a written form. The status of those minority languages is often disputed, in the sense that there are discussions – often politically oriented – on whether to give them the status of languages of their own, or of dialects of a language X, or some status between the aforementioned two (with or without a written form). A turning point in studying these phenomena was due to the Tartu-based linguist, Aleksandr Duličenko (1941–), who introduced into Slavic studies the terminology Slavjanskie literaturnye mikrojazyki ‘Slavic literary micro-languages (SLMLs)’ in his seminal work with the same title published in 1981, and advanced the theoretical frame of SLML. Here the terminology ‘literary language’, traditionally used in Slavic linguistics, does not mean exclusively ‘a language of literature’, but rather means a ‘standard language’ with polyfunctional usage in a given linguistic community, including an aesthetic function for writing literary works (see Duličenko’s characteristics of SLMLs below). The idea of SLML can be found exclusively in Slavic studies, not in theoretical sociolinguistic literature (Reference Stern, Stern, Nomachi and BelićStern 2018: 20). It has not been supported by all Slavists as a fact (see Section 30.7), but since the appearance of Duličenko’s above-mentioned monograph and his works in this field, the terminology has become more or less stabilized and accepted in some handbooks written by specialists (see Reference SuprunSuprun 1989, Reference RehderRehder et al. 1993, Reference MečkovskajaMečkovskaja 2000, Reference Okuka and KrennOkuka & Krenn 2002, Reference PiperPiper 2008, Reference VečerkaVečerka 2008, Reference SiatkowskaSiatkowska 2004, and many others).

According to Reference DuličenkoDuličenko (2015: 572–573), an SLML is a language that has the following features: (1) It is one of the forms in which language exists – similarly to any ‘ethnic’ literary language (and unlike such linguistic forms as dialects or substandard vernaculars); (2) a micro-language is a written form of language and, as such, implements certain orthographic principles; (3) it is based on a specific dialect – like a majority of ‘ethnic’ literary languages; (4) it is characterized by normalization tendencies, that is, literary norms are developed at the phonological, grammatical [i.e. morpho-syntactic], and lexical levels; (5) [literary norms] may become stable and be further codified (i.e. established by way of normative grammars and normative dictionaries); (6) which, on the whole, is a consequence of the functioning of such a literary linguistic form in various spheres of life; (7) within the framework of an organized and socially maintained literary linguistic process (for a critique of this definition, see Section 30.6).

Reference DuličenkoDuličenko (2015: 574) considers that the existence of the above-mentioned features distinguishes SLML from so-called literary dialects in which no process of standardization can be observed and which often remain personal linguistic experiments. This characterization of SLMLs is rather close to that of major Slavic standard languages, though the overall quality is different. Another important fact is that, according to Reference DuličenkoDuličenko (1981: 9), an SLML always co-exists with a major national language and the SLML hierarchically occupies a lower position in two respects: language-internally, the degree of standardization is less strict in SLMLs, while language-externally, the functional polyvalency is narrower in SLMLs.

In addition to the term SLMLs, in his newer publications Duličenko applies the terminology Malye slavjanskie literaturnye jazyki ‘Small Slavic Literary Languages’ (cf. Reference Duličenko, Moldovan, Skorvid, Kibrik, Rogova, Jakuškina, Žuravlev and TolstajaDuličenko 2017) as a synonym for SLML. Other scholars who work on these languages, but without paying attention to the notion of a literary language (such as Reference Breu, Duličenko and NomachiBreu 2018), often use the terminology without ‘literary’, that is, Slavic Micro-Languages (SML). The term SLML or SML is widespread nowadays not only among scholars, but also among language activists who often play the role of codifiers of an SLML. For instance, codifiers of the Bunjevac literary language and the Podlachian literary language call their languages ‘micro-languages’ in addition to the usual terminology ‘language’ (Reference Popov and Kujundžić-OstojićPopov & Kujundžić Ostojić 2019: 3, Reference Maksymiuk, Stern, Nomachi and BelićMaksymiuk 2018).Footnote ¹

30.2 Changing and Varying List of SLMLs

The number of SLMLs varies, depending on time and scholars who count these languages, because, first, SLMLs not only appear but also disappear for various reasons. For instance, when Duličenko published his first monograph in 1981, it was inconceivable to consider Silesian as a literary micro-language, but Reference Duličenko, Duličenko and NomachiDuličenko (2018: 4) now regards it as an SLML. On the other hand, West Polesian, which had a certain success in the late 1980s and early 1990s in the Soviet Union’s Belarus and Ukraine, seems to have disappeared since the leader of the movement Mykola Šyljahovyč stopped his activity. In addition, the disappearance of any SLML does not always negate the possibility of reappearance of the same or a similar SLML. For instance, Óndra Łysohorsky’s literary Lachian is counted as one of the specific SLMLs, but since Łysohorsky’s death in 1989, nobody has published anything in Lachian. However, today, after 30 years (as of 2020), Eva Tvrda, a Silesian poet, inspired by Łysohorsky, publishes her works in Lachian. Nevertheless, at this stage it seems to be impossible to find any features of SLMLs that could be applied to her Lachian. In 1981, Reference DuličenkoDuličenko (1981: 10–11) counted 12 SLMLs with two unsuccessful attempts (Table 30.1). The 12 SLMLs’ approximate geographical distribution is as shown in Figure 30.1.

Figure 30.1 Approximate geographical distributions of SLMSs in Reference DuličenkoDuličenko (1981)

Venetian Slovene and Aegean Macedonian are not included in Figure 30.1.

In his encyclopedic handbook, Reference DuličenkoDuličenko (2003–2004) proposed the following 20 languages, including attempts. Here let us include only those which were added to Reference DuličenkoDuličenko (1981) (Table 30.2).

Among these, Pomak, West Polesian, Vič, and Halšan SLMLs are phenomena that appeared after the publication of Reference DuličenkoDuličenko (1981). The difference is that the two varieties of Sorbian have been added to the list, which is disputable. As will be mentioned later in Section 30.6, the linguistic stability with continuous tradition of a written form and its establishment, and the sociolinguistic status of both varieties of Sorbian, are clear enough, therefore this particular idea of Duličenko is not supported even among those scholars who recognize micro-literary languages as linguistic phenomena.Footnote ²

In his latest publication on this topic, Reference Duličenko, Duličenko and NomachiDuličenko (2018) proposes 20 SLMLs. The difference between the previous and current lists of languages consists in that Vič and Halšan are not mentioned any more, while Bunjevac (Northern Serbia) and Silesian (Southern Poland) are added. Reference Duličenko, Duličenko and NomachiDuličenko (2018) does not seem to cover all SLMLs and attempts. Among such, one could mention various attempts to (re-)codify literary Moravian (West Slavic) in the Czech Republic (Reference Šustek and DuličenkoŠustek 1998, Reference Osowski, Stern, Nomachi and BelićOsowski 2018), and literary Podlachian (East Slavic) (Reference Maksymiuk, Stern, Nomachi and BelićMaksymiuk 2018, Reference Mladenova2021). In the Balkans, there was an attempt to codify literary Gorani (South Slavic) that even gained an official status in Kosovo (Reference Nomachi, Duličenko and NomachiNomachi 2018, Reference Długosz, Greenberg and GrenobleDługosz 2020).

Only new attempts are mapped on Figure 30.2.

Figure 30.2 Approximate geographical distributions of SLMSs in Reference DuličenkoDuličenko (2003–2004) and later publications

If we include some Internet-based projects, one could mention the project of the Siberian literary language in the early twenty-first century. It was based on the northern Russian dialect spoken in Siberia and even some kind of a normative grammar was produced, but the project itself disappeared.Footnote ³ Similar phenomena such as Katsian in Northern Russia could also be mentioned, but so far there does not seem to be any notable success in establishing a written form (cf. Reference BaranovaBaranova 2014).

In contrast to the above-mentioned recently emerging, rather unstable SLMLs and attempts, some SLMLs such as Burgenland Croatian or Vojvodina Rusyn are much more stable and have a longer history that can be comparable to the major Slavic languages.

It is important to note that there are attempts to create a literary dialect with an established norm, but without an intention to establish a language: Mazurian (West Slavic) (Reference SzatkowskiSzatkowski 2020) and Kurpian (West Slavic) (Reference RubachRubach 2009, Reference Rubach2016, Reference Rubach2017, Reference Rubach2019). The codifiers of those dialects do not claim any official status or any linguistic separateness from Polish. However, particularly with regard to the latter, the Kurpian dialect has an already well-established normative grammar, dictionaries, textbooks, and other publications. Also, there is a local cultural society, Związek Kurpiów ‘The Union of Kurpians’, that has been promoting activity in the dialect.

Reference VečerkaVečerka (2008) offers a somewhat different list of SLMLs from Duličenko and his followers. In addition to those SLMLs in Duličenko’s open-ended list, as SLMLs Večerka counts the following seven defunct varieties: the Biblical Czech language (West Slavic) that was used among Slovaks in the eighteenth and nineteenth centuries; Jesuit Slovak/Trnava Slovak (West Slavic), in which some orthography manuals and textbooks were printed in the eighteenth century; Camaldolese Slovak (also West Slavic, named for an eighteenth-century monastery), which was the first attempt at a codified Slovak language based on the western dialect in the eighteenth century; Bernolakian Slovak (West Slavic), which was the first codified Slovak based on the western dialect in the eighteenth century; Slaveno-Serbian (South Slavic), which was a mixture of Russian, the Russian and Serbian recensions of Church Slavonic, and the Serbian vernacular (the Vojvodina dialect) in the eighteenth century; Transylvanian Bulgarian (South Slavic), which existed between the sixteenth and eighteenth centuries in present-day Romania; and Ruthenian/Prosta mova ‘Simple Speech’ (East Slavic), which was an official language of the Polish-Lithuanian Commonwealth in the sixteenth and seventeenth centuries. Although there is no universally accepted definition of SLMLs, many of Večerka’s additions are usually dealt with within the history of each major Slavic literary language (Slovak, Serbian, both Ukrainian and Belarusian for Ruthenian, respectively). In addition, not all these cases show features characteristic of the SLMLs discussed in Section 30.1; they will not be discussed in this chapter.

30.3 Classifications of SLMLs

In Section 30.2, the genetic affiliation of SLMLs was mentioned. Generally speaking, there is no difficulty in determining the genetic affiliation on the macro level, except for Vojvodina Rusyn, whose affiliation is disputed among scholars (West Slavic or East Slavic). Indeed, Reference DuličenkoDuličenko (2011: 333, Reference Duličenko, Moldovan, Skorvid, Kibrik, Rogova, Jakuškina, Žuravlev and Tolstaja2017: 630) offers a classification of SLMLs based on the traditional tripartition of the Slavic languages. However, the discussion about the detailed status of each SLML is not always obvious. When an SLML is formed based on a transitional dialect between language X and language Y, its linguistic affiliation often can be disputed. For instance, the literary Lachian language was formed based on a transitional dialect between Czech and Polish. In this case there can easily appear three (even four) camps among scholars and activists with regard to the status of Lachian: Lachian is not a language but a dialect of Czech, or a dialect of Polish, or neither Czech nor Polish, or both Czech and Polish. The last option is implausible, but at least Óndra Łysohorsky, the founder and the codifier of the Lachian literary language, took the third position (neither Polish nor Czech); he intentionally set up an orthographic system which is a mixture of Czech and Polish to show the transitional status between Czech and Polish (Reference ŁysohorskyŁysohorsky 1988: 816). There are many SLMLs that are difficult to uniquely classify genetically on the micro level.

In Slavic linguistics, there is another type of classification of SLMLs. The other pattern is also due to Duličenko and is based on the combination of an areal-geographical principle and an ethnolinguistic principle (Reference DuličenkoDuličenko 2003–2004: 6, Reference Duličenko2011: 334). In later works, Duličenko offers a different ordering of each group, but the concept remains the same. They are (1) Autonomous, (2) Insular, (3) Peripheral-Insular, and (4) Peripheral (or regional). According to Reference DuličenkoDuličenko (2011: 323–324), each group is defined as follows.

• Autonomous SLMLs: (a) they are autonomous both on the geographical and on the ethnolinguistic and genetic levels; (b) they have all possibilities needed for cultivating a norm of a literary language; (c) and for widening their functional spectrum (scope of application).

• Insular SLMLs: (a) they emerge as a result of migration to different areas in other Slavic or non-Slavic lands at different periods of time; (b) geographically, they are detached from their own ethnolinguistic root; (c) on a genetic level, they retain fairly clear links to their original linguistic root; (d) as for the form of the literary language, they look like the Autonomous SLMLs, that is, norms are formed or being formed; (e) attempts to try to keep or widen their functional spectrum are observed.

• Peripheral-Insular SLMLs: (a) these are languages which have moved beyond the boundary of the main area and found themselves partially in a different ethnolinguistic environment, and only state borders make them, relatively speaking, islands; (b) on a genetic level, just like Insular SLMLs, this group of languages keep clear connection with their ethnolinguistic root; (c) on the level of literary languages, work on normalizing them is underway in the same manner as Insular SLMLs.

• Peripheral (or regional) SLMLs: (a) geographically, they are adjacent to their original ethnolinguistic array, that is, to the corresponding majority Slavic people and their language; (b) they retain clear genetic ties with it, but their separateness appears on the cultural and linguistic level (the local culture performed in a local dialect or dialects in line with the process of making a single literary language); (c) being used mainly in literary and artistic works and thereby forming a literary language, this group of literary languages is least standardized, and if we bear in mind that writers often write relying mainly on their own speeches and dialects, then these can be called polycentric literary languages. But even so, the functioning process of forming a literary language is directed here towards the formation of a koine.

Based on the four typological features, SLMLs can be classified as follows (Reference Duličenko, Duličenko and NomachiDuličenko 2018: 4–6):

• Autonomous SLMLs: Upper Sorbian, Lower Sorbian, Kashubian

• Insular SLMLs: Vojvodina Rusyn, Burgenland Croatian, Molise Slavic, Resian, Banat Bulgarian

• Peripheral-Insular SLMLs: Carpatho-Rusyn, Aegean Macedonian, Pomak, Venetian Slovene, Bunjevac

• Peripheral (or regional) SLMLs: Čakavian, Kajkavian, Prekmurje Slovene, Lachian, East Slovak, West Polesian, Silesian.

As has been mentioned in Section 30.2, all these categories are open-ended, and the classification cannot be absolute. Some attempts at creating SLMLs could be added into one of those categories. In addition, in some cases, an SLML does not always fit with a given category. For instance, West Polesian is linguistically much closer to Standard Ukrainian than to Standard Belarusian, though West Polesian can be dealt with as a distinct dialect of Belarusian. Therefore, following the linguistic facts, West Polesian could be classified as a Peripheral-Insular SLML. Kashubian perhaps should not be treated absolutely as an Autonomous SLML, because it could be seen as a Peripheral SLML considering the traditional placement of Kashubian in Polish linguistics.

30.4 Conditions and Factors Particularly Pertinent for Creating SLMLs

There are no universal conditions whose fulfillment automatically produces SLMLs, because in the emergence of SLMLs, as is the case with major literary languages, each case is rather specific and has experienced different political, administrative, and ethnolinguistic changes and different dialectal patterning with different cultures. However, there are some characteristics shared by all SLMLs. According to Reference DuličenkoDuličenko (2015: 582–583), there are two kinds of features – obligatory and optional – which he regarded as conditions for creating SLMLs. The primary features are as follows:

• the existence of people in a given society who recognize their own ethnic specificity, which leads them to their common ethnonym (ethnic endonym)

• the existence of a linguistic separateness, that is, awareness of material difference between one’s own and other languages and awareness of homogeneity (uniformity) of one’s own language, based on which later on a literary language will be formed and a linguonym (i.e. self-name of a language) will appear

• the social cohesion of the ethnolinguistic community, that is, community members’ awareness of their own affiliation to the same ethno-social organism

• territorial isolation of the ethno-linguo-social community and in this connection a compactness of the environment.

These four features are essential and obligatory. The absence of one of them makes it difficult to create an SLML, according to Reference DuličenkoDuličenko (2015).Footnote ⁴ In addition to them, Reference DuličenkoDuličenko (2015) recognizes secondary features which are not obligatory:

• the existence of a pre-tradition of a literary language (more broadly, a cultural-linguistic pre-tradition) based on their own mother tongue and non-mother tongue

• the existence of cultural-religious or cultural-national movements that acted, for example, in the form of Protestantism (the sixteenth century) and national revival (the second half of the eighteenth and mid-nineteenth centuries)

• the existence of a subjective factor, which is usually expressed in the form of awakeners.

In this context, one could also mention political changes as a supplement to two of Duličenko’s secondary conditions. For instance, appearance of the Lachian LML was closely associated with the emergence of socialism, particularly that in the Soviet Union, as Łysohorsky saw an ideal future in the advent of socialism including the rise of proletarians and self-determination and that is why he started his linguo-cultural activities. Other cases are the most recent SLMLs and similar attempts. They became possible only after the political changes in the Eastern Bloc or during the collapse of the Bloc when democratization started to infiltrate into the (former) Bloc. In addition, it is important that the rise of nationalism in those countries stimulated this tendency. The attempts of Vič and Halšan in Lithuania, and of West Polesian in Soviet Belarus and Ukraine should be placed in this context. In connection with this, a further step was a political Europeanization of the former Eastern Bloc. The former Bloc countries saw it as almost imperative to become members of the European Union and cooperate with the Council of Europe, and therefore they had to accept the multilingualism that Europe has been promoting. In this context, Kashubian received official status in Poland in 2005, which influenced and activated the Silesian autonomy movement to pursue the same status in one way or another.

In addition to these conditions, one could also add the development of new technologies and their application to SLMLs for widening the sphere of usage. Most successful societies that have been promoting their own SLMLs need to possess not only traditional media such as books, journals, and language handbooks, but also multiple mass media such as newspapers and TV and radio broadcasting in SLMLs. Today, SLMLs without presence on the Internet are not even imaginable; the Internet is vital for their activities including the propagation of an SLML to non-members of the society related to it (cf. Reference Stern, Stern, Nomachi and BelićStern 2018: 85–123). For instance, most important information about the Podlachian LML can be found exclusively on the website Svoja.org. The use of the Internet implies the existence of language users and some successful language transmission from older generations to newer.

30.5 SLMLs as Minority and Endangered Languages

In most cases, SLMLs are lesser-used languages and endangered, though the degree of endangerment varies from one SLML to another. The UNESCO Atlas of the World’s Languages in Danger mentions only Upper and Lower Sorbian, Kashubian, Carpatho-Rusyn, Vojvodina Rusyn, and Polesian (Reference Salminen and BrenzingerSalminen 2008: 36), while the electronic version of the Atlas includes Banat Bulgarian, Burgenland Croatian, Resian, and Molise Croatian (=Slavic), as of September 2021.Footnote ⁵ This project is important for minority languages but has no binding legal force. According to the list of languages covered by the European Charter for Regional or Minority Languages (ECRML), the following seven SLMLs are included: Bunjevac (Serbia), Burgenland Croatian (Austria),Footnote ⁶ Kashubian (Poland), Lemko (= one of the varieties of Carpatho-Rusyn) (Poland), Lower Sorbian (Germany), Ruthenian (= Rusyn) (Bosnia and Herzegovina, Croatia, Hungary, Romania, Serbia, Slovakia, Ukraine), and Upper Sorbian (Germany), as of May 2015.Footnote ⁷ Since each European state that has ratified the Charter has to choose at least some measures to protect any minority language that it itself has recognized, the above-mentioned SLMLs are officially included.

This does not always mean that other SLMLs are not protected, but some of them are not treated separately as SLMLs.Footnote ⁸ For instance, Bulgarian is protected in Romania and in Serbia, while Banat Bulgarian is not treated as a separate entity. Indeed, in the communist period, Bulgarian was already protected and taught to Bulgarian minorities in Romania, but the language taught in the whole country was the standard Bulgarian language. Speakers of Banat Bulgarian, though its standardized form equipped with a Latin script-based orthography is significantly different from Standard Bulgarian (cf. Reference MladenovaMladenova 2021), had to study the standard variety in schools; therefore in the long run it is unclear whether the official status of minority language in Romania helps Banat Bulgarians to preserve and develop their linguistic heritage. The same is true for Resian and Molise Slavic in Italy, as they are protected simply as Slovene and Croatian. While in Serbia today, Bunjevac is treated as a separate entity, not as a variety/dialect of Croatian or Serbian, outside of Serbia, Bunjevac is mainly treated as a dialect of Croatian. These different treatments of the language–dialect dichotomy are based on different language policies of each state and, in some cases, on different linguistic identities of speakers, too.

In this context, one should note that Kajkavian and Čakavian LMLs are usually treated as dialects of Croatian, therefore they are not treated as endangered languages, though there are various local initiatives to preserve the dialectal diversity in Croatia. The same can be said about Kashubian spoken in Renfrew County in Ontario, Canada. But in this particular case, the problem is that speakers identify themselves as Poles and their language as Polish as well.Footnote ⁹

In all the above-mentioned cases, naturally, almost personal projects or attempts at creating SLMLs with less-visible success, dialects without speakers’ awareness of linguistic separateness are not subject to being mentioned or protected by laws. It is even impossible to specify how many speakers there are in these cases, as the national census in any country does not usually collect information about the number of dialect speakers.

In the context of protection and development of SLMLs as minority languages, contemporary Vojvodina Rusyn turns out to be a specific case. When Rusyns and their language were officially recognized in 1919 in the Kingdom of Serbs, Croats, and Slovenes, the Rusyns’ cultural organization Rusyn Popular Education Society was immediately set up and the Society published newspapers, journals, books, textbooks including a normative grammar which was taught in primary schools. Thus, in the very early phase of its codification, the polyfunctionality of newly codified literary Rusyn started its development. After World War II, Rusyns could take advantage of better social and legal conditions. The first Rusyn high school opened in 1945, the first radio and TV broadcasting in 1949 and 1975, respectively. In 1977 one high school in Ruski Kerestur was launched in which all subjects in three classes have been taught in Rusyn since then. In 1981, the department of Rusyn language and literature was established within the Faculty of Philosophy at the University of Novi Sad and it has existed until today (cf. Reference FejsaFejsa 2010). This rather straight development is exceptional among SLMLs. As of 2021, there are only three SLMLs (Vojvodina Rusyn, Carpatho-Rusyn in Slovakia, and Sorbian) that have their own departments/institutes in higher education.Footnote ¹⁰ However, this does not mean that the future of Vojvodina Rusyn or Sorbian is secured, particularly as the Lower Sorbian case clearly shows.

Establishing (more stable) standardized forms and their implementation in the educational system is an important strategy for the preservation and development of endangered languages, including Slavic minority languages. However, this strategy does not work well enough when a given language shows a high dialectal diversity. For example, Kashubian is being standardized based on the Central dialect, which is significantly different from Southern and Northern dialects. This means that Kashubs from Southern and Northern regions will eventually have to learn a new variety that is rather foreign to them.

30.6 SLMLs Pro et Contra: Critical Assessments of SLMLs as a Category and Alternative Approaches to Slavic Micro-Languages

Since the appearance of Reference DuličenkoDuličenko (1981), there have been various kinds of discussion over SLMLs. Critiques of each SLML are based on a language–dialect dichotomy. For instance, the Bulgarian scholar Ivan Reference KočevKočev (1984) attacked Duličenko because Duličenko divided Banat Bulgarian from the corpus of the Bulgarian language. A more scholarly critique was made by Hanna Reference Popowska-TaborskaPopowska-Taborska (1988), who claimed that Kashubian is nevertheless one of the Polish dialects, purely considering linguistic facts. Konstantin Reference Lifanov, Stern, Nomachi and BelićLifanov (2018) considers that there has been no East Slovak LML, because each sort of writing in Eastern Slovakia that emerged in different periods of time has been individual not only in terms of its linguistic features, but also of its mechanism of emergence and sphere of functions. In addition, it seems that Kajkavian and Čakavian dialects of Croatian may not be justified as LMLs at least today, as there are no clear or observable standardizing tendencies, though there were indeed rich traditions of written forms which can be labeled as literary languages as Croatian scholars described them in the history of Croatian.

There have been critiques about the theoretical aspects of SLMLs as well. For instance, Reference RehderPeter Rehder (1984–1985) criticized Duličenko for the claim that macro-languages and micro-languages are different not in their quantitative respects, but exclusively in their qualitative respects.

By criticizing the vagueness of the definition of SLMLs that could be also applied to a regular literary language, Dieter Reference Stern, Stern, Nomachi and BelićStern (2018) provides more concrete features to define SLMLs. Stern summarizes that LMLs have six features by which they can be distinguished from literary languages (Table 30.3).

According to Stern (ibid), it is feature (2) that makes scholars include Sorbian in SLMLs.Footnote ¹¹ In addition, Stern warns that the SLMLs defined by him can work only within the frame of a modern nation-state. Therefore, Renaissance poets’ works in Čakavian do not justify the existence of a Čakavian SLML now.Footnote ¹²

Reference Skorvid and SkorvidSkorvid (2017) criticizes Duličenko and offers a different paradigm for SLMLs with different terminologies. On the one hand, Skorvid points out that Duličenko’s terminology includes quite different linguistic entities under the same terms ‘SLMLs’ or ‘SMLs’, which are vague and are given totally different interpretations by one scholar and by another. On the other hand, Skorvid views as problematic the inconsistency of Duličenko’s usage of terminology even in his own publications, which has been pointed out and criticized by Reference HenzelmannHenzelmann (2016), too.Footnote ¹³

The most problematic issue is that since 2003, Upper and Lower Sorbian have been included in Duličenko’s list of SLMLs. However, in fact they could not be placed on the same level and treated together with West Polesian or Lachian, which Skorvid calls literary idiolects. Skorvid considers that Sorbian, Vojvodina Rusyn, and Lemko in Poland should be called simply minority languages, while he suggested that Kashubian, Čakavian, and Kajkavian could be classified as regional languages.

Along similar lines to Skorvid, Vladislav Reference Knoll and SkorvidKnoll (2017) criticizes the general vagueness of Duličenko’s paradigm of SLMLs and offers a different and detailed scheme to classify the cultural-written idioms that Duličenko has been calling SLMLs. Knoll differentiates five stages and defines them as follows:

• Unwritten idiom (UI): a linguistic form that functions only in an oral form (including the sphere of an oral culture/literature)

• Literary idiolect (LI): a linguistic form which a certain person uses, alone or with a group of confederates (in an oral and/or written form)

• Literary dialect (LD): a linguistic form which possesses a written artistic literature (genres are typically limited), but is absent in the social sphere in a written form; ways of writing it are often not unified

• Regional literary language (RL): a linguistic form that has a writing culture, which is limitedly used in artistic literature and in social life, but is characterized by its absence in some functional spheres

• National literary language (LL): a language form is present in all available functional spheres.

Considering these features, Knoll summarizes his re-classifications of SLMLs as in Table 30.4.

Knoll neither classified all SLMLs that Duličenko listed nor put them into his scheme, and historical development of each SLML could be placed in a different cell. According to Knoll, however, Lachian is LI, while Sorbian, Burgenland Croatian, Vojvodina Rusyn, Kashubian, and Carpatho-Rusyn in Slovakia are characterized as RL. Čakavian, Kajkavian, Prekmurje Slovene, and East Slovak could be classified as LD.

31.1 Introduction

The complex political, economic, and social forces of the last few decades have impelled massive waves of migration from regions of Eastern Europe and the post-Soviet space, prompting a renewed spike in the continuous spread of Slavic languages to new locales documented throughout their history. Along the traditional path of three-generational language shift characteristic of immigrant settings, these varieties have been passed down as heritage languages, that is, languages acquired naturalistically, albeit to varying degrees, by bilingual children who most typically attain higher proficiency in the societally dominant language by adulthood.

This chapter presents a broad overview of the current research on the formal properties of Slavic languages developing in heritage language (HL) settings. Considering the scarcity of directly comparable data across the Slavic varieties, the chapter will take stock of and draw generalizations from a relatively narrow set of the most representative and amply researched structural changes in heritage Slavic grammatical systems, analyzed against the backdrop of the corresponding baseline systems (the latter notion encompasses both the homeland and first-generation diaspora varieties). Starting with morphosyntactic phenomena in the nominal and verbal domains, I will focus on the restructuring of the case and gender systems and shifts in the grammatical encoding of temporal distinctions through aspect and tense morphology (Section 31.2). At the levels of sentence organization and discourse structure, I will discuss word order change in fixed and variable syntactic configurations (Section 31.3). The concluding discussion in Section 31.4 will generalize over the recurrent themes of the surveyed body of work and identify directions for future studies in heritage Slavic linguistics.

31.2 Morphosyntax

As highly inflected fusional systems, Slavic languages are characterized by complex nominal and verbal paradigms and elaborate agreement marking. Considering that inflectional morphology often represents the primary locus of change attested in HL settings (Reference PolinskyPolinsky 2018), it does not come as a surprise that the majority of existing HL scholarship to date has been concerned with their core morphosyntactic categories: case, gender, and tense-aspect marking.

31.3 Word Order

A typologically salient feature of Slavic languages is their flexible word order. A number of researchers have attested changes in word order principles operating in HL settings, both in constructions in which the order of elements is syntactically fixed and those where constituent placement serves to mark their discourse function. The majority of existing heritage Slavic studies have discussed changes in both of these domains primarily in relation to dominant language influence.

A good example of the former phenomenon is the placement of clitics in the South Slavic languages (see also Chapter 17). Drawing on sociolinguistic interviews with first- and second-generation speakers of Serbian in Australia, Reference Dimitrijević-SavićDimitrijević-Savić (2008: 67–74) identified differences in the use and location of clitics in the speech of heritage bilinguals, compared to standard grammatical conventions. While clitics in Serbian typically occupy the so-called second position (i.e. occur after the first clausal constituent or after the first word within that constituent), heritage speakers often placed dative and accusative pronominal clitics after the verb (i.e. in positions where the direct and indirect objects are realized in English). A similar pattern of non-canonical clitic placement was documented in Reference Ivanova-SullivanIvanova-Sullivan’s (2019) longitudinal case study of a Bulgarian-English child observed between the ages of 2;0 and 4;0. Example (8) illustrates one of the earliest occurrences of a Bulgarian object clitic in the child’s production, appearing post-verbally instead of its expected preverbal position.

1. (8)

   Iskam	da vidja	go.
   want	to see	it-cl.m
   ‘I want to see it.’ (Reference Ivanova-SullivanIvanova-Sullivan 2019: 22)

Both studies also document instances of clitic omission and their replacement with full pronominal forms, suggesting that differences between the heritage and baseline grammars, when apparent, lie not only in the syntax of clitic placement, but also involve their formal and semantic properties. However, as demonstrated by the analysis of direct object production by Polish-German and Portuguese-German school-aged children in Reference Rinke, Flores and SopataRinke et al. (2019), contact with a cliticless language does not appear to affect the acquisition of object clitics in the HL, modulated instead by the accumulated amount of contact with the HL.

Apart from the ordering of clitics, word order in Slavic is traditionally described as free, with various configurations of main clausal constituents deemed acceptable under the appropriate information-structural conditions (see also Chapter 17). To anticipate the discussion below, most published studies on syntactic change in heritage Slavic point to word order rigidification and strengthening of the predominant SVO pattern. However, since the majority of existing work has been conducted in an English-dominant setting (but see e.g. Reference LaskowskiLaskowski 2014 for similar patterns in heritage Polish in Sweden), it is not yet possible to decisively determine the underlying factors driving this change.

Several studies have employed oral narratives to quantify the occurrence of canonical and non-canonical word order patterns in the naturalistic data. Looking specifically at subject–verb inversion (see example (9a)), Reference Isurin and Ivanova-SullivanIsurin & Ivanova-Sullivan (2008) and Reference Ivanova-SullivanIvanova-Sullivan (2014: 105–110) have documented a significant reduction in the use of VS orders by heritage Russian speakers, with no detectable effect of proficiency on the frequency of their use. In a different study, Reference Laleko, Dubinina, Bauckus and KresinLaleko & Dubinina (2018) detected similar patterns of diminished use of VS orders (9 percent in the heritage group vs. 14 percent in the monolingual group) alongside a higher occurrence of canonical SV(O) orders in the heritage group (76 percent vs. 67 percent). Additionally, Laleko and Dubinina found that heritage and baseline speakers adhered to different principles in the use of dislocation constructions, which involve the movement of objects to preverbal positions as illustrated in (9b).

1. (9)

   Word order in Russian (Reference Laleko, Dubinina, Bauckus and KresinLaleko & Dubinina 2018: 194)
   a.	Ljagušku	poceloval	princ.					OVS
   	frog-acc	kissed	prince-nom
   	‘A prince kissed the frog.’
   b.	Ljagušku	princ		poceloval.			OSV
   	frog-acc	prince-nom		kissed
   	‘A prince kissed the frog.’

Dislocations occurred more commonly with pronominal and otherwise short (i.e. grammatically light) constituents in the HL, whereas for monolinguals the effect of grammatical weight was more likely to be overridden by information-structural requirements on the placement of constituents in accordance with their role in discourse (i.e. new vs. old information).

The effects of information structure on word order variation in heritage Russian in the US were investigated in Reference LalekoLaleko (2022), who employed acceptability judgment tests to measure the effects of subject focus on the speakers’ ratings of SV(O) and (O)VS sentences. In line with the previously attested trends, heritage speakers over-accepted the canonical orders and under-accepted the inverted orders, a trend manifested most categorically in low-proficiency speakers. Additionally, while the monolinguals were equally sensitive to focus in intransitive and transitive conditions, only in the latter contexts did focus serve as a predictor of subject placement for high-proficiency heritage speakers. As discussed in Laleko, these results suggest that HL speakers do not experience a global difficulty with focus marking through word order, but possibly find it more taxing to utilize information-structural cues in contexts where other cues (such as the lexico-syntactic properties of the verb in the case of intransitive predicates) must also be integrated.

While most of the above studies account for the observed decrease in HL word order variability in terms of dominant language transfer, work conducted outside of the English-dominant environment suggests that influence from the contact language may affect the shape of the HL word order system without loss of word order flexibility. Looking at a set of written narratives from heritage Russian adolescents in Germany, Reference Brehmer, Usanova and PeukertBrehmer & Usanova (2015) did not find a lower rate of non-canonical constructions in the HL corpus, compared to monolingual data. However, the study did report a more frequent use of verb-final orders, a property of German subordinate clauses, in the HL essays. In fact, the heritage Russian writers applied this restricted German-like pattern more liberally by extending it to main declarative clauses, resulting in a novel, less pragmatically marked structure adopted into the HL.

In another study involving German as the contact language, Reference Brehmer and SopataBrehmer & Sopata (2021) examined the placement of auxiliaries and infinitive complements in complex verb phrases in heritage Polish, as illustrated in (10).

1. (10)

   Word order in complex predicates in Polish (Reference Brehmer and SopataBrehmer & Sopata 2021)
   a.	On	chc-e	złapać		ptak-a.
   	he-nom	wants	catch-inf.pf		bird-acc
   	‘He wants to catch a/the bird.’
   b.	On	chc-e	go	złapać.
   	he-nom	wants	he-acc.cl	catch-inf.pf
   	‘He wants to catch him/it.’

In Polish, word order in these constructions varies depending on the pragmatic context, with both adjacent (10a) and discontinuous (10b) structures licensed by the grammar, whereas German only allows for the discontinuous placement of these elements. Using an apparent-time approach, the study examined data from elicited oral production and acceptability judgments in several age groups of HL speakers and documented transfer effects from German in both simultaneous and sequential bilinguals, manifested as an overuse and over-acceptance of the discontinuous pattern. The developmental trajectories in the acquisition of the relevant structures were further shown to depend on a complex interplay of factors, with a joint effect of delayed acquisition in childhood and attrition in adulthood documented for simultaneous bilinguals.

In summary, while syntactic convergence with the dominant language emerges as a prime direction of HL word order change based on the reviewed studies, more work is necessary to account for the mechanisms and outcomes of such change in a wider set of linguistic environments. The fact that HL grammars, like the grammars of all natural languages, are subject to language-internal pressures associated with diachronic change predicts the possibility of non-contact-induced developments in HLs that may produce novel patterns not fully mapping onto the dominant language template. Analyses of HL corpora, particularly those based on naturalistic observation, and experimental studies probing into the speakers’ judgments and processing would be essential for such investigations.

31.4 Conclusions and Future Directions

The chapter has offered a concise but comprehensive review of current research on grammatical properties of Slavic languages as HLs. Notwithstanding the expectedly high degree of variability in heritage speaker competencies observed across the surveyed studies, the many recurring motifs in this literature make it clear that the trajectory of change in heritage grammars undergoing restructuring is modulated by a set of common principles, many of which are already familiar from research on intergenerational language loss (Reference Seliger and VagoSeliger & Vago 1991). Two such trends seem particularly robust in relation to the data discussed here. On the one hand, the inflectionally rich and highly synthetic Slavic languages represent a very clear illustration of the effects of morphological leveling in heritage grammar formation, a process manifested as reduction and regularization of the nominal and verbal inflectional paradigms and decrease or loss of allomorphic variation. The resulting streamlining of morphophonological distinctions often goes hand in hand with the development of analytic or periphrastic constructions to replace synthetic forms, innovations in the marking of grammatical relations, and reorganization of word order principles. At first glance, some of these changes seem to operate in the direction of convergence between the typological structures of the heritage and dominant languages, especially considering that the lion’s share of existing empirical studies have examined the formation of heritage grammars in close contact with the inflectionally rigid and predominantly analytic grammatical system of English. Yet, several studies outside the English-dominant environment have demonstrated that greater morphological richness of the ambient language does not suffice to avert paradigm leveling in the heritage system, and that the internal principles of change are still operative in such systems alongside external, transfer-induced changes, resulting in a conglomeration of properties that require careful evaluation in relation to the structures of the relevant baseline and dominant languages as well as to the general principles of diachronic change.

On a larger scale, the predominantly fusional nature of Slavic morphology creates ideal conditions for observing linguistic changes motivated by the interactions between the principle of economy, which favors the development of grammatical systems with the least possible number of forms (this encompasses the leveling processes discussed above), and the principle of transparency, which serves to establish the most predictable and unilateral links between forms and meanings within a linguistic system (i.e. one meaning–one form). Effects of the latter principle seem particularly evident in processes manifested as elimination (or delays in the acquisition of) opaque forms, reduction in the range of their meanings, and difficulties with processing underspecified, ambiguously marked, or optionally occurring elements, as demonstrated by numerous studies reviewed in this chapter. In situations of sustained societal bilingualism, these global pressures may surface in the heritage varieties as no more than quantitative tendencies affecting the distribution of forms across different contexts or variation in accessibility during online processing. In the more extreme cases, these processes could be accelerated to the level of linguistic representation and culminate in the elimination of certain grammatical categories from the HL or their full or partial adaptation to the structural template of the societally dominant language.

HLs can offer much to the study of language change and historical research more generally. As the brief survey presented in this chapter hopes to have demonstrated, the quickly expanding field of heritage linguistics is positioned perfectly at the intersection of distinct linguistic traditions and can fruitfully build on and inform a range of methodological and conceptual approaches to language study. While much progress has been made in recent years toward the development of the empirical and theoretical foundations of heritage Slavic linguistics, a major drawback of the existing body of knowledge is that it remains limited to only a few most well-represented Slavic languages, examined in a relatively narrow set of language contact situations. Some notable gaps concern HL development in Slavic-to-Slavic contact settings (e.g. Ukrainian in Poland or Serbo-Croatian in Slovenia), that is, in situations where the grammatical systems of the heritage and dominant languages are maximally similar, and contexts in which typologically-related non-dominant HLs develop simultaneously (e.g. early Russian-Ukrainian bilinguals in the US) or undergo koineization under prolonged inter-linguistic and inter-dialectal contact in established immigrant communities. Going forward, along with the advancement in the range of research methods employed for the study of HLs at various points on the continuum of language change, a key imperative is to expand the typological diversity of language dyads under investigation in order to draw broader conclusions about the overarching principles of grammatical development in Slavic languages as HLs in different socio-demographic niches and across a wider range of linguistic locales.

32.1 Introduction

A script is a set of characters that have a certain relation to linguistic units. Every script belongs to a certain script type (Reference Daniels, Olson and TorranceDaniels 2009; called a ‘writing system’ by Reference CoulmasCoulmas 1999) based on its main level of representation: In logographic scripts (e.g. Chinese, Sumerian cuneiform, Mayan), a single character represents a word or a morpheme; in syllabaries (e.g. Hiragana, Cypriot, Cherokee), a single character represents a syllable; and in alphabets (e.g. Cyrillic, Latin, Greek), a single character represents a phoneme. Special cases are consonantal alphabets (or abjads, e.g. Arabic, Hebrew, Phoenician), in which vowels are not normally written, and syllabic alphabets (or alphasyllabaries or abugidas, e.g. Nagari, Ethiopic, Cree), in which vowels are represented as systematic modifications of consonant signs. However, writing systems as adaptations of scripts to a certain language (called ‘spelling systems’ by Reference Coulmas, Günther and LudwigCoulmas 1996: 1381) never belong to a ‘pure’ type. Thus, for example, all existing logographic systems are actually logosyllabic (making use of the rebus principle to also write words independently of their meaning), and alphabetic texts often contain logograms like 〈&〉, 〈€〉, 〈%〉, 〈°C〉, etc. All scripts that have been used for Slavic languages are either alphabets or consonantal alphabets.

All types of writing systems exploit links to phonology. However, no existing writing system is a simple ‘transcription’ of the phonology of a language. This means that, although most of the languages of the world are never written down, in a language that happens to have a written form the writing system is an independent subsystem of that language. We call this subsystem the graphematics of a language. It works in a similar way to the phonological subsystem. Consequently, we need to distinguish between the systematic aspects of the system, which are referred to as graphemics, and the material aspects, which are referred to as graphetics. Graphemic material is conventionally enclosed in angular brackets 〈…〉, and graphetic material (following Reference Fuhrhop and BuchmannFuhrhop & Buchmann 2009) in vertical bars |…|.

This chapter only treats general issues of the use and development of scripts for writing Slavic languages. The concrete spelling systems of the Slavic languages are the subject of Chapter 33.

32.2 Scripts Used for Slavic Languages

Six different scripts have so far been used for writing Slavic languages. The main ones are the Glagolitic, Cyrillic, and Latin alphabets. Apart from that, Slavic texts have also been written in the Greek alphabet and the Arabic and Hebrew consonantal alphabets.

32.2.2 The Glagolitic Script

The heatedly debated question of the nineteenth century about whether the Glagolitic or the Cyrillic alphabet is older and which of them was invented by Saint Cyril (Constantine) has been answered unequivocally: the oldest Slavic script is the Glagolitic alphabet (see Table 32.1), and it was invented by Cyril. Among the main arguments in favor of this are the fact that in palimpsests and glossed documents containing both alphabets, it is always the original text that is in Glagolitic while the newer text or the glosses are in Cyrillic, the tendency for the most archaic forms of Old Church Slavonic to be conveyed in Glagolitic documents, and the fact that it was the Glagolitic script that was originally called kurilovica ‘Cyril’s script’ (e.g. in Upyrʹ Lixyj’s colophon of 1047).

Table 32.1 Glagolitic and Cyrillic scripts

The debate about the ‘origin’ of the Glagolitic letters, however, is still ongoing. Several exogenic (‘natural’, ‘palaeographic’; Reference Trunte, Matulić and LupisTrunte 2019) theories have claimed an evolutionary development of the Glagolitic script from some other existing script, for example Greek (minuscule), Armenian, or some Oriental script, whereas endogenic (‘artificial’, ‘ideographic’; Reference Trunte, Matulić and LupisTrunte 2019) approaches have tried to understand the alphabet as created according to some underlying design principle, for example by combining the three Christian symbols cross, circle, and triangle (Reference TschernochvostoffTschernochvostoff [1947] 1995) or by connecting lines on a common grid (Jončev & Jončeva 1982). None of these approaches is convincing in its entirety. Therefore, most probably Cyril freely invented the Glagolitic letters, sometimes being inspired by theological ideas (e.g. when using the cross as the first letter of the alphabet or when designing the letters and , both of which show one part of the Holy Trinity enclosed by divine eternity, to form the abbreviation for ‘Jesus’, which can be read as descent from heaven and ascension) and sometimes using associations with other scripts he knew (e.g. ← Greek , ← Latin 〈P〉, ← Armenian , or ← Hebrew ).

Invented by Cyril, the Glagolitic script was first used for Cyril and Methodius’s Moravian mission, which began in 863 and continued after Cyril’s death in 869 until Methodius’s death in 885, after which their disciples were expelled from Moravia by Wiching, who was reinstated as bishop of Nitra by the Frankish adversaries of the Slavic liturgy. Cyril & Methodius’s followers mainly fled to modern Bulgaria and North Macedonia, maybe also directly to Dalmatia and Istria (cf. Reference BirnbaumBirnbaum 1996); continuity with the Moravian mission has also been claimed for the Sázava Monastery near Prague (where the Prague Glagolitic Fragments were written in the eleventh century; cf. Reference ČajkaČajka 2011: 41–45). From Bulgaria and North Macedonia, the script spread further to the region of modern-day Serbia, Bosnia and Herzegovina, and Montenegro (which, however, is only attested from Štokavian influences in Old Church Slavonic manuscripts like the Codex Marianus, the Gršković Apostle, or the Mihanović Homiliarium, as well as from individual Glagolitic letters in Serbian Cyrillic manuscripts, not from inscriptions), and to Kievan Rusʹ (cf. e.g. the Glagolitic inscriptions in Novgorod; Reference Gippius and MixeevGippius & Mixeev 2022). The oldest manuscripts and inscriptions we have were thus written about a generation after the Moravian mission, though some very early texts (especially the Kiev Folia) show West Slavic and Catholic influences. Consequently, the ‘classical’ version of Glagolitic we know from these texts must already have been the result of a development from the first version of Glagolitic originally invented by Cyril (cf. Reference Trunte, Dürrigl, Mihaljević and VelčićTrunte 2004). For example, the four ‘classical’ nasal vowel letters show clear signs of modifications (, which is a combination of , which does not exist separately, and , which might originally have been either the letter for or a general nasal vowel letter or diacritic); in the oldest texts we find a second /x/ letter next to canonical , the so-called ‘spidery x’ , whose original function is unclear (maybe it represented the archaic pronunciation [kʰ] of Greek , maybe [x] vs. for [ç] in Greek loanwords, maybe German [h], maybe it was even the original letter for , cf. Reference Fuchsbauer, Fuchsbauer, Stadler and ZinkFuchsbauer 2021: 107f.); there is evidence that, unlike Cyrillic letters, which have numerical values only for units, tens, and hundreds, in Glagolitic there originally was also a fourth row of letters representing thousands (e.g. = 1000, = 9000), but this row has been preserved only fragmentarily.

Between the end of the ninth and the end of the twelfth century, the Glagolitic alphabet was gradually superseded by Cyrillic, first in eastern Bulgaria (Preslav) and later in North Macedonia (Ohrid). During this time, both scripts existed side by side, as, for example, Cyrillic glosses in Glagolitic manuscripts show. Obviously, in some places people already used the Cyrillic script for most purposes while holding on to Glagolitic for liturgical texts. When Glagolitic text production had ceased, Glagolitic texts continued to be copied into Cyrillic, which of course still required reading them. Moreover, some developments within Cyrillic seem to have had a ‘backward’ influence on Glagolitic, for example the differentiation of and and the reuse of (from Greek ) as a letter for the reflex of *tj (i.e. št in Bulgaria) might have led to the subsequent analogous differentiation of Glagolitic and and the reassignment of Glagolitic .

After the twelfth century, Glagolitic continued to be used by the glagoljaši (‘Glagolites’, Catholic priests reading mass in Croatian Church Slavonic) along the Dalmatian coast, on the islands, and in Istria, after whom the script received its modern name. In this region, Glagolitic developed a different shape, the so-called angular Glagolitic (see Section 32.3.1). In the fourteenth century, the Croatian Glagolitic tradition was also carried to Prague, where Charles IV founded the Emaus Monastery with Croatian monks in 1347, and from there to Oleśnica (Lower Silesia) in 1380 and to Kleparz near (nowadays in) Cracow in 1390, where Glagolitic texts were written for several decades (Reference Żurek, Barciak and IwańczakŻurek 2006). In Istria, Dalmatia, and especially on the Croatian islands, the Glagolitic alphabet continued to be written, also for mundane purposes and for the vernacular language (mostly Čakavian), in some places up to the nineteenth century. In the twentieth century, the script finally fell out of use, being supplanted by the Latin alphabet.

32.3 Graphetics: Development and Variation of Scripts Used by the Slavs

Just like languages, scripts can exhibit diachronic and diatopic variation. Consequently, within a script we can distinguish several glyphic variants (Reference BunčićBunčić et al. 2016: 22–24).

32.3.1 Glagolitic Letters

Since no manuscripts from the Moravian mission are preserved, the original shapes of the Glagolitic letters are a matter of debate. Arguably the oldest preserved manuscript are the Kiev Folia (probably from the early tenth century; Figure 32.1). They differ from the ‘classical’ Glagolitic in the large Old Church Slavonic manuscripts in several respects. Most notably, the letters, which have markedly different heights (the shortest being , the tallest and ), do not sit on a baseline but appear to be hanging down from a line. The letters and are not symmetrical yet, the is simply a cross, the lowest line of is not connected to the upper part, the left part of is rectangular rather than rounded, and the right part of and is significantly smaller than the left part.

Figure 32.1 Oldest form of Glagolitic

(Kiev Folia, tenth century, fol. 4r) (https://lccn.loc.gov/2021667731)

In all the Old Church Slavonic manuscripts the letters have a lot of curves and loops (see Figure 32.2). This version, which seems to have been optimized for writing with ink (cf. the additional loops and curves in Latin and Cyrillic cursives: for , for , for , for , etc.) is therefore commonly called round Glagolitic. In Croatia, the Glagolitic alphabet was often incised in stone, and for these inscriptions it was more practical to have straight lines and angular shapes. The resulting angular Glagolitic script was also used in manuscripts from Croatia (see Figure 32.3). One of its characteristics is an abundance of ligatures that reduce the number of largely redundant vertical lines (e.g. round > angular → ligature , > → , > → ). The angular variant was also used in prints from the late fifteenth to the twentieth century. For everyday use, a cursive form, called brzopis ‘quick writing’ in Croatian, developed and was used up to the nineteenth century (see Figure 32.4).

Figure 32.2 Round Glagolitic

(Codex Marianus, eleventh century, fol. 36r) (https://w.wiki/8DwH)

Figure 32.3 Angular Glagolitic

(Vrbnik Statute, sixteenth century, p. 2) (http://urn.nsk.hr/urn:nbn:hr:238:355082)

Figure 32.4 Cursive Glagolitic

(Sermon on Love for Enemies, 1790, fol. 3v) (http://urn.nsk.hr/urn:nbn:hr:238:201216)

32.3.2 Cyrillic Letters

The oldest Cyrillic liturgical texts are written in a variant called ustav, which is a solemn two-line script, in which all the letters have the same height, and ascenders and descenders only occur in the form of hairlines (Figure 32.5). A more casual variant was the poluustav, a four-line script in which ascenders and descenders are more pronounced and straight lines can more often be bent or tilted (Figure 32.6). Gradually, ustav was completely superseded by poluustav, which was also the variant used in print, for Church Slavonic to this day. Among the Catholics in Bosnia and Croatia, however, poluustav was replaced by a Western variant of Cyrillic, often referred to as Bosančica (and in Croatian as hrvatska ćirilica ‘Croatian Cyrillic’; Figure 32.7). Diatopic (and diachronic) variation was most pronounced in the multiplicity of cursives: The letter shapes in a text written in skoropisʹ can be used to date and locate manuscripts (Figure 32.8) and, in the case of Western Cyrillic, even to determine the religion of the writer (Reference BunčićBunčić et al. 2016: 198–200).

Figure 32.5 Ustav

(Codex Suprasliensis, eleventh century, fol. 28r) (https://w.wiki/8DwE)

Figure 32.6 Poluustav

(Cologne Manuscript of Isaac the Syrian, fifteenth century, fol. 2r) (Isaak 2015: 1)

Figure 32.7 Western Cyrillic cursive

(Poljica Statute, 1665, fol. 1r) (www.croatianhistory.net/etf/et04.html)

Figure 32.8 Ruthenian Skoropisʹ

(Debt bond, 1569; Russian National Library, Zinčenko collection No. 46) (https://expositions.nlr.ru/rusautograph/pismo/ustav/albom.php)

A major turning point in the development of Cyrillic was tsar Peter I’s alphabet reform of 1708. By introducing the more frugal letter shapes of his graždanskij šrift (or graždanskaja pečat‘) ‘civil type’ (Figure 32.9), which were modeled on contemporaneous roman type (cf. |а| → |a|, |е| → |e|, |ѧ| → |я|, etc.), and omitting all the diacritics, suspended letters, titla, and synonymous letters characteristic of church prints, he made printing for mundane purposes simpler and therefore cheaper (cf. the title of a news report from 1711: “Релѧ́цїа. ѡ҆ поведенїи бы́вшемъ въ а҆́рмеѣ Е҆гѡ Црⷭ҇кагѡ Вели́чества · Ма́їа съ трїдесѧ́тагѡ числа̀” in Old Cyrillic and “Реляція. О поведенїи бывшемъ въ Армее Его Царского Велїчества · маїя съ 30 числа” in civil type; see the facsimiles in Reference BunčićBunčić et al. 2016: 112). A fact that is often forgotten is that this reform also applied to handwriting, replacing the old skoropisʹ with graždanskoe pisʹmo ‘civil cursive’. Subsequently, the civil script was also adopted by the other peoples using the Cyrillic alphabet. Consequently, the former diatopic variety has been replaced by an almost uniform Cyrillic alphabet in print and relatively little variation between the national cursives. Exceptions are a few special letter shapes in Serbian, Montenegrin, and Macedonian Cyrillic (|б| and italic |г|, |д|, |п|, |т| vs. |б|, |г|, |д|, |п|, |т| for 〈б〉, 〈г〉, 〈д〉, 〈п〉, 〈т〉). Apart from that, a Bulgarian variant of Cyrillic with more independent small-letter shapes and more ascenders and descenders (e.g. |ϐ|, |ꙅ|, |ɡ|, |u|, |k|, |ʌ|, |n|, |m| for 〈в〉, 〈г〉, 〈д〉, 〈и〉, 〈к〉, 〈л〉, 〈п〉, 〈т〉) has been developing for several decades (Reference Kempgen, Tomelleri and KempgenKempgen 2015); however, while this variant is ubiquitous on Bulgarian posters, signposts, book covers, etc., it has only recently come into use as body text font.

Figure 32.9 Civil type

(Sanktpeterburgskie vědomosti, June 1711, p. 8) (https://w.wiki/8DwB)

32.3.3 Latin Letters

In contrast to Glagolitic and Cyrillic, the Latin alphabet has a long pre-Slavic history. However, Slavic texts only start to be written after the introduction of the Carolingian minuscule (in which e.g. the Old Slovenian Freising Manuscripts are written; see Figure 32.10). The development of the Latin alphabet during the Late Middle Ages is also reflected in Slavic manuscripts. Thus, Slavic manuscripts of the Gothic period are written in the textualis style (a kind of blackletter), for example the Polish Sankt Florian Psalter (Psałterz floriański) or the Slovenian Rateče Manuscript (Rateški rokopis; also Celovški rokopis ‘Klagenfurt Manuscript’; Figure 32.11), both from the late fourteenth century. In Dalmatia, meanwhile, the rotunda style that is also widespread in Italy can be seen for example in the Croatian Red i zakon of 1345 or the Vatican Prayer Book from around 1400 (Figure 32.12). After the introduction of letterpress printing, Slavia Latina was divided into a northern part (i.e. the West Slavic languages) using blackletter and a southern part (i.e. the South Slavic languages) using roman type (or antiqua; Figure 32.13) since the sixteenth century (after the first Croatian books – e.g. the Lectionary of Bernadine of Split, Venice 1495 – were printed in a rotunda typeface, and the two very first Slovenian books were printed in 1550 in Tübingen in Schwabacher). In the north, the two main variants of blackletter are Schwabacher (Figure 32.14) and Fraktur (Figure 32.15), the former having slightly rounder, simpler, and wider letters than the latter (compare |d|, |g|, |m|, |o|, |s| in Schwabacher with |d|, |g|, |m|, |o|, |s| in Fraktur). While German texts were printed in Schwabacher only at the beginning but then mainly in Fraktur from the latter half of the sixteenth century to 1941, Polish texts remained in Schwabacher until blackletter was replaced by roman type around the turn of the eighteenth century. Czech was also mainly printed in Schwabacher (although a few Czech texts were in Fraktur), and the abolition of blackletter only happened in the 1820s–1830s (Reference BunčićBunčić et al. 2016: 300–303). Sorbian texts, by contrast, were printed mainly in Fraktur (just like German texts). Roman type was only introduced in 1847 for Upper Sorbian and 1862 for Lower Sorbian, but Fraktur remained in use for Upper Sorbian until the Nazi period, whereas for Lower Sorbian blackletter was even revived after 1989 (Reference BunčićBunčić et al. 2016: 303–305).

Figure 32.10 Carolingian minuscule

(Freising Manuscripts, Slovenian part, late tenth century, fol. 160v) (https://w.wiki/8D$a)

Figure 32.11 Textualis

(Slovenian Rateče Manuscript, late fourteenth century) (https://w.wiki/8D$m)

Figure 32.12 Rotunda

(Vatican Croatian Prayerbook, around 1400, fol. 22v) (Reference LupićLupić 2019: 59)

Figure 32.13 Roman type

(Marko Marulić, Judita, Venice 1586, fol. 3v) (http://urn.nsk.hr/urn:nbn:hr:238:371054)

Figure 32.14 Schwabacher

(Czech Grammar of Náměšť, 1533, p. 10) (https://doi.org/10.5282/ubm/digi.1032)

Figure 32.15 Fraktur

(Upper Sorbian New Testament translated by Michał Frencel, 1706, Matthew 1:1) (https://w.wiki/8E22)

The use of blackletter and roman type in print correlates with the use of a German cursive (Kurrent) or a roman cursive in handwriting. Although nowadays the German cursive is not taught in schools anymore, every country teaches a different cursive variant in their schools, so that people writing a foreign language usually have an ‘accent in writing’ (Schriftakzent; Reference Meyer, Meyer and ReinkowskiMeyer 2018) because the letter shapes they have learned at school are not the same that native speakers of the language have learned. This also applies across scripts (e.g. Russian schoolchildren learn a specific cursive variant of the Latin alphabet in their foreign language classes).

32.4 Social, Cultural, Religious, and Political Influences

Being a part of language planning, the adoption or change of a writing system for a language is an action undertaken by people and as such subject to all kinds of extralinguistic influences. The phenomenon that ‘script follows religion’ (Reference SampsonSampson 1985: 16), though vastly overestimated, is responsible for the fact that the use of the Latin and Cyrillic alphabet almost exactly coincides with Slavia Latina and Slavia Orthodoxa. The Serbian speech community, which belongs to Slavia Orthodoxa but uses both alphabets, writes texts about Orthodox religion exclusively in the Cyrillic script.

Furthermore, we can see areal effects of power relations between scripts (cf. Reference CalvetCalvet 1999: 94–98). For example, the Bosnian Muslims have never relied exclusively on the Arabic alphabet but used the Cyrillic script (just like the Catholics and Orthodox in Bosnia) until the end of the Ottoman era and mainly the Latin alphabet since 1878. This change is part of a Latinization process in the Balkans (which might be seen in the context of a Europeanization or Westernization), which also led to the obsolescence of Glagolitic and Cyrillic for Croatian and the introduction and increasing use of the Latin alphabet to Serbian (Reference BunčićBunčić et al. 2016: 321–324). In the Russian sphere of influence, by contrast, it was the Cyrillic alphabet that ousted the Latin alphabet that had been in use for Belarusian and (to a lesser extent) Ukrainian until the beginning of the twentieth century, and that was also exported to a great number of languages of the Soviet Union, Mongolia, and even China. The replacement of blackletter with roman type in Polish, Czech, and Upper and Lower Sorbian can be seen as a Slavization that gained momentum during the period of Slavic Revival, but also as a deliberate alienation from the German sphere of influence.

32.5 Biscriptality

In the course of history, many Slavic speech communities have employed more than one writing system at the same time. This phenomenon, which can be called biscriptality, can be broken down into 3 × 3 major types (Reference BunčićBunčić et al. 2016: 67): On the one hand, we have to distinguish between scripts, script variants, and orthographies. (Orthographic biscriptality is treated in, Section 33.6.) On the other hand, the writing systems in parallel use can be distributed in three different ways: in an equipollent opposition, a privative opposition, or a diasituative distribution. (Most of the following examples are taken from Reference BunčićBunčić et al. 2016, where further references can be found.)

32.6 Script Reforms

The writing system as the most ‘visual’ subsystem of language, and a subsystem that has to be consciously learned (nowadays usually in school), is much more accessible to language planning than most other subsystems of language. Consequently, apart from slow, evolutionary developments that are characteristic of language change in general, writing systems are often characterized by abrupt, revolutionary changes on the basis of writing reforms.

All Slavic languages have been the object of such writing reforms. However, changes of script have always occurred gradually, for example the replacement of Glagolitic and Cyrillic with the Latin alphabet for Croatian, or the introduction of the Latin alphabet to Serbian. When in Belarusian the 1933 Narkamaŭka orthography, in contrast to the 1918 Taraškevica orthography, included only the Cyrillic alphabet, this was only the formal recognition of a process that had already begun in 1912 when Naša Niva discontinued its Latin-alphabet edition. All attempts at changing the script of a Slavic language by reform have been unsuccessful: three proposals to latinize the Ukrainian language (in 1834, 1859, and 1923; cf. Reference BunčićBunčić et al. 2016: 280), the 1930 proposals for a Latin alphabet for Russian (cf. Reference Alpatov, Tomelleri and KempgenAlpatov 2015), an attempted Cyrillization of Polish by the Russian authorities after the January Uprising of 1863/64, and even the declaration in the Serbian constitution of 2006 that only the Cyrillic alphabet is official, which did not change anything about the status of the Latin alphabet in the everyday life of the Serbs.

This is different with script variants. The most well-known Slavic example of a reform directed at glyphic variants is Peter I’s 1708 introduction of civil type for Russian, which was successful not only in Russia but eventually also in all the other languages written with Cyrillic letters. The reform that introduced roman type to Upper and Lower Sorbian in 1847 and 1862 was equally successful, although blackletter continued to be used for a long time in Upper Sorbian and is even used to this day in Lower Sorbian. In the other West Slavic languages, blackletter was phased out more slowly.

32.7 Conclusion

The Slavic scripts roughly align with the cultural division between Slavia Orthodoxa and Slavia Latina. Although the Cyrillo-Methodian tradition of the Glagolitic script connected to the Old Church Slavonic language was originally not confined to either of the two areas and the Glagolitic script was used longest in Catholic Croatia, it is nowadays continued in the form of the Cyrillic script, whereas the Latin alphabet in Slavia Latina is based on a completely different, ‘Western’ tradition. However, mutual influences abound and can be seen in script changes and various instances of biscriptality as well as in the introduction of roman type in the West in the sixteenth century, the adoption of its design principles in the ‘civil type’ in the East in the eighteenth century and the gradual replacement of both blackletter and Old Cyrillic, which was (almost) completed only in the twentieth century.

33.1 Introduction

The writing system of a language consists of a script (or several scripts), which provides a set of characters and their general relation to the respective linguistic units, and an orthography, which provides a norm for the concrete adaptation of the script to the target language. In Reference CoseriuCoseriu’s (1952) sense, the script is the sistema, whereas the orthography is the norma, which makes choices from the options the script provides (e.g. whether /j/ is written as 〈j〉, 〈i〉, or 〈y〉 in an orthography based on the Latin script) but often also deviates from the system in arbitrary ways. The term orthography is sometimes restricted to such norms that are formally standardized; here, however, it also encompasses orthographic usage (in the sense of German Gebrauchsnorm ‘norm by use’) that might otherwise be referred to by terms like spelling or Pol. grafia.

33.2 Orthographic Principles