The Balkan sprachbund or linguistic area is the ensemble of areal features—similarity in grammar, syntax, vocabulary and phonology—among languages of the Balkans, which belong to various branches of Indo-European, such as Slavic, Greek, Romance and Albanian. While they share little vocabulary, their grammars also have similarities; for example they have similar case systems and have all become more analytic, although to differing degrees.

History

The earliest scholar to notice the similarities between Balkan languages belonging to different families was the Slovenian scholar Jernej Kopitar in 1829. August Schleicher (1850) more explicitly developed the concept of areal relationships as opposed to genetic ones, and Franc Miklošič (1861)studied the relationships of Balkan Slavic and Romance more extensively.

Nikolai Trubetzkoy (1923), Kristian Sandfeld-Jensen (1930), and Gustav Weigand developed the theory in the 1920s and 1930s.

In the 1930s, the Romanian linguist Alexandru Graur criticized the notion of “Balkan linguistics,” saying that one can talk about “relationships of borrowings, of influences, but not about Balkan linguistics”.

The term "Balkan linguistic union" was coined by the Romanian linguist Alexandru Rosetti in 1958, when he claimed that the shared features conferred the Balkan languages a special similarity. Theodor Capidan went further, claiming that the structure of Balkan languages could be reduced to a standard language. Many of the earliest reports on this theory were in German, hence the term "Balkansprachbund" is often used as well.

Languages

The languages that share these similarities belong to five distinct branches of the Indo-European languages:

• Albanian
• Greek
• Indo-Aryan (Arli Romany/Gypsy)
• Romance languages (Romanian, Aromanian, Megleno-Romanian and Istro-Romanian)
• Slavic languages (Bulgarian, Macedonian, Serbian — especially the Torlakian dialect which is transitional between Macedonian, Bulgarian and Serbian)

However, not all of these languages have the same number of features shared. That is why they are divided into three groups:

1. Albanian, Romanian, Macedonian, Aromanian and Bulgarian have the most properties in common
2. Serbian language (especially transitional Torlak dialect) and Greek share with the others a lower number of properties
3. Turkish - shares mainly vocabulary and replacement of infinitive with subjunctive.

The Finnish linguist Jouko Lindstedt computed in 2000 a "Balkanization factor" which gives each Balkan language a score proportional with the number of features shared in the Balkan linguistic union. The results were:

Language	Score
Macedonian	12
Balkan Slavic	11.5
Albanian	10.5
Greek, Balkan Romance	9.5
Romani (Gypsy)	7.5

Another language that may have been influenced by the Balkan language union is the Judeo-Spanish variant that used to be spoken by Sephardi Jews living in the Balkans. The grammatical features shared (especially regarding the tense system) were most likely borrowed from Greek.

Origins

The source of these features as well as the directions have long been debated, and various theories were suggested.

Thracian, Dacian or Illyrian

Since most of these features cannot be found in languages related to those that belong to the linguistic union (such as other Slavic or Romance languages), early researchers, including Kopitar, believed they must be inherited from the Paleo-Balkan languages (Thracian, Dacian and Illyrian) which formed the substrate for the modern Balkan languages. But since very little is known about any of these languages, it cannot be determined whether the features were present.

Greek

Another theory, advanced by Kristian Sandfeld in 1930, was that these features were an entirely Greek influence, under the presumption that since Greece "always had a superior civilization compared to its neighbours", Greek could not have borrowed its linguistic features from them. However, no ancient dialects of Greek possessed Balkanisms, so that the features shared with other regional languages appear to be post-classical innovations. Also, Greek appears to be only peripheral to the Balkan linguistic union, lacking some important features, such as the postposed article.

Latin and Romance

The Roman Empire ruled all the Balkans, and local variation of Latin may have left its mark on all languages there, which were later the substrate to Slavic newcomers. This was proposed by Georg Solta. The weak point of this theory is that other Romance languages have few of the features, and there is no proof that the Balkan Romans were isolated for enough time to develop them.

An argument for this would be the structural borrowings or "linguistic calques" into Macedonian from Aromanian, which could be explained by Aromanian being a substrate of Macedonian, but this still does not explain the origin of these innovations in Aromanian.

Multiple sources

The most commonly accepted theory, advanced by Polish scholar Zbigniew Gołąb, is that the innovations came from different sources and the languages influenced each other: some features can be traced from Latin, Slavic or Greek languages, while others, particularly features that are shared only by Romanian, Albanian, Macedonian and Bulgarian, could be explained by the substratum kept after Romanization (in the case of Romanian) or Slavicization (in the case of Bulgarian). Albanian was influenced by both Latin and Slavic, but it kept many of its original characteristics.

Several arguments favour this theory. First, throughout the turbulent history of the Balkans, many groups of people moved to another place, inhabited by people of another ethnicity. These small groups were usually assimilated quickly and sometimes left marks in the new language they acquired. Second, the use of more than one language was common in the Balkans before the modern age, and a drift in one language would quickly spread to other languages. Third, the dialects that have the most "balkanisms" are those in regions where people had contact with people of many other languages.

Timeline of contacts

(under development)

Most likely the earliest contact was between the Proto-Romanians and Proto-Albanians, (1st century - 5th century AD) this theory being supported by the Albanian vocabulary borrowed from Balkan Latin, as well as the Romanian substrate, which has words cognate to Albanian words.

The exact area where contact occurred is under debate, ranging from Northern Albania to Transylvania. For more, see Origin of Romanians and Origin of Albanians. All Romanian varieties (from the Republic of Moldova to the Vlachs of Serbia) are part of the sprachbund, which shows that the contact happened before they diverged.

The invasion of the Slavs led to a period of migrations throughout the Balkans which created multi-ethnic communities and this led to the sprachbund beginning around the 8th century; most features were present by the 12th century, but in some parts it continued until the 17th century.

Serbian was perhaps the last language to join, as shown by the low number of features, most of which were only in the Torlak dialect, a dialect intermediary to Bulgarian which emerged rather late, after most features were settled in the sprachbund.

Features

Grammatical features

Case system

The number of cases is reduced, several cases being replaced with prepositions, the only exception being Serbian. In Bulgarian, on the other hand, this development has actually led to the loss of all cases except the vocative.

A common case system of a Balkan language is:

• Nominative
• Accusative
• Dative / Genitive (merged)
• Vocative

Syncretism of genitive and dative

In the Balkan languages, the genitive and dative cases (or corresponding prepositional constructions) undergo syncretism.

Example:

Language	Dative	Genitive
English	I gave the book to Maria.	It is Maria's book.
Albanian	Ja dhashë librin Marisë.	Është libri i Marisë.
Aromanian	U-ded vivliapi Maria.	Easte vivlia aliMarie.
Bulgarian	Дадох книгата на Мария [dadoh knigata na marija]	Книгата е на Мария [knigata e na Marija]
Romanian	I-am dat cartea Mariei. colloq. for fem. (oblig. for masc.): I-am dat cartea lui Marian.	Este cartea Mariei. colloq. for fem. (oblig. for masc.): Este cartea lui Marian.
Macedonian	ѝ ја дадов книгата на Марија. [ì ja dadov knigata na Marija]	Книгата е на Марија. [knigata e na Marija]
Greek	Έδωσα το βιβλίο στην Μαρία. [édōsa to biblío stīn maría]      or Έδωσα το βιβλίο της Μαρίας. [édōsa to biblío tīs marías]	Είναι το βιβλίο της Μαρίας. [eínai to biblío tīs marías]
	Της το έδωσα [tīs to édōsa] 'I gave it to her.'	Είναι το βιβλίο της. [eínai to biblío tīs] 'It is her book.'
Serbian	Dao sam knjigu Mariji*.	To je Marijina* knjiga.

(* Marija in the Nominative case)

Syncretism of locative and directional expressions

language	"in Greece"	"into Greece"
Albanian	në Greqi	në Greqi
Aromania	tu Gãrtsia	tu Gãrtsia
Bulgarian	в Гърция (V Gărtsiya)	в Гърция (V Gărtsiya)
Greek	στην Ελλάδα (stīn Helláda)	στην Ελλάδα (stīn Helláda)
Macedonian	Во Грција (vo Grcija)	Во Грција (vo Grcija)
Romanian	în Grecia	în Grecia
Serbian	u Grčkoj*	u Grčku*

(* Grčka in the Nominative case)

Verb tenses

Future tense

The future tense is formed in an analytic way using an auxiliary verb or particle with the meaning "will, want", referred to as de-volitive, similar to the way the future is formed in English and originating with Greek innovation around the 1st century AD. This feature is present to varying degrees in each language. Decategoralization is less advanced in Romanian voi and in Serbian ću, ćeš, će, where the future marker is still an inflected auxiliary. In Modern Greek, Bulgarian, Macedonian, and Albanian, decategoralization and erosion have given rise to an uninflected tense form, where the frozen 3rd person singular of the verb has turned into an invariable particle followed by the main verb inflected for person.

Language	Variant	Formation	Example: "I'll see"
Albanian	Tosk	"do" (invariant) + subjunctive	Do të shikoj
	Gheg	"kam" (conjugated) + me + verbal noun	Kam me shikue
Aromanian		"va" (invariant) + subjunctive	Va s-ved
Greek		"θα" (invariant) + subjunctive	Θα δω / βλέπω (tha dō / blépō); "I'll see / be seeing"
Bulgarian		"ще" (invariant) + present tense	Ще видя (shte vidya)
Macedonian		"ќе" (invariant) + present tense	Ќе видам (kje vidam)
Serbian	(literary standard)	"хтети/hteti" (conjugated) + infinitive	Ја ћу видети (видећу) (ja ću videti [videću])
	(colloquial)	"хтети/hteti" (conjugated) + subjunctive	Ја ћу да видим (ja ću da vidim)
Romanian	(literary standard)	"a voi" (conjugated) + infinitive	Voi vedea/vedeare
	(colloquial)	"o" (invariant) + subjunctive	O să văd
	(colloquial alternate)	"a avea" (conjugated) + subjunctive	Am să văd
	(archaic)	"va" (invariant) + subjunctive	Va să văd
Romani	(Erli)	"ka" (invariant) + subjunctive	Ka dikhav

Analytic perfect tense

The analytic perfect tense is formed in the Balkan languages with the verb "to have". The origin of this language feature could be Latin. However, this does not apply to Bulgarian and Serbian, where the analytic perfect is formed with the verb "to be" and the past active participle: обещал - "who has promised" (past active participle); съм (Bul.); сам (Ser.) - "I am"; обещал съм; обећао сам (Ser.) - "I have promised" (lit. "I am one who has promised"), perfect tense. Constructions using the verb to have are characteristic of Macedonian language (Имам ветено./Imam veteno. = I have promised.).

Avoidance of infinitive

The use of the infinitive (common in other languages related to some of the Balkan languages, such as Romance and Slavic) is generally replaced with subjunctive constructions, following early Greek innovation.

• in Bulgarian, Macedonian and Tosk Albanian, the loss of the infinitive is complete
• in demotic Greek, the loss of the infinitive was complete, whereas in literary Greek was not; the natural fusion of the demotic (vernacular) form with the literary (archaic) one resulted in the creation of the contemporary common Greek (Koine Neohellenic), where the infinitive is principally used as noun (e.g. λέγειν "speaking, fluency, eloquence", γράφειν "writing", είναι "being", etc.)
• in Aromanian and Southern Serbian dialects, it is almost complete
• in Gheg Albanian and Megleno-Romanian, it is used only in a limited number of expressions
• in standard Romanian and Serbian, the infinitive shares many of its functions with the subjunctive. In these two languages, the infinitive will always be found in dictionaries and language textbooks. In Romanian, the long infinitives, which are identical to the Italian ones (-are, -ere, and -ire) can also be used in both formal and informal conversation.
• Turkish as spoken in Sliven and Šumen has also almost completely lost the infinitive, clearly due to the influence of the Balkan Sprachbund.

For example, "I want to write" in several Balkan languages:

Language	Example	Notes
Albanian	"Dua të shkruaj"	as opposed to Gheg me fjet "to sleep" or me hangr "to eat"
Aromanian	"Voi sã-ngrapsescu"
Macedonian	"Сакам да пишувам"
Bulgarian	"Искам да пиша"
Modern Greek	"Θέλω να γράψω"	as opposed to Ancient Greek "βούλομαι γράψαι"
Romanian	"Vreau să scriu"	as opposed to "Vreau a scrie/scriere", which is also correct, but rarely used.
Serbian	"Želim da pišem"/"Желим да пишем	as opposed to the literary, more correct form: "Želim pisati"/"Желим пиcaти, where pisati/пиcaти is the infinitive. Both phrases are correct and do not create misunderstandings, although the colloquial one is more commonly used in daily conversation.
Bulgarian Turkish	"isterim yazayım"	In Standard Turkish in Turkey this is "yazmak istiyorum" where "yazmak" is the infinitive.

But here is an example of a relict form, preserved in Bulgarian:

Language	Without infinitive	With relict "infinitive"	Translation	Notes
Bulgarian	"Недей да пишеш."	"Недей писа."	Don't write.	The first part of the first three examples is the prohibitative element недей ("don't", composed of не, "not", and дей, "do" in the imperative). The second part of the examples, писа, я, зна and да, are relicts of what used to be an infinitive form (писати, ести, знати and дати respectively). This second syntactic construction is colloquial and more common in the eastern dialects. The forms usually coincide with the past aorist tense of the verb in the third person singular, as in the case of писа; those that don't coincide (as in the last three examples) are highly unusual today, but do occur, above all in older literature.
	"Недей да ядеш."	"Недей я."	Don't eat.
	"Недей да знаеш."	"Недей зна."	Don't know.
	"Можете ли да ми дадете?"	"Можете ли ми да?"	Can you give me?

Bare subjunctive constructions

Sentences which include only a subjunctive construction can be used to express a wish, a mild command, an intention or a suggestion.

This example translates in the Balkan languages the phrase "You should go!", using the subjunctive constructions.

Language	Example	Notes
Macedonian	Да (си) одиш!
Bulgarian	Да отидеш!
Serbian (mainly Torlak)	Да идеш!
Albanian	Të shkosh!
Modern Greek	Να πας!
Romany (Gypsy)	Te dža!
Romanian	Să te duci!	in Romanian, the "a se duce" (to go) requires a reflexive construction, literally "take yourself (to)"
Megleno-Romanian	S-ti duţ!
Aromanian	S-ti duts!

Morphology

Postposed article

With the exception of Greek and Romani, all languages in the union have their definite article attached to the end of the noun, instead of before it. None of the related languages (like other Romance languages or Slavic languages) shares this feature and it is thought to be either an innovation or Albanian borrowing spread in the Balkans.

However, each language created its own internal articles, so the Romanian articles are related to the articles (and demonstrative pronouns) in Italian, French, etc., while the Bulgarian articles are related to demonstrative pronouns in other Slavic languages.

Language	Feminine		Masculine
	without article	with article	without article	with article
Albanian	shtëpi	shtëpia	qiell	qielli
Aromanian	muljare	muljarea	bãrbat	bãrbatlu
Bulgarian	жена	жената	мъж	мъжът
Macedonian	жена	жената	маж	мажот
Romanian	muiere	muierea	bărbat	bărbatul
Serbian Torlak	жена	жената	муж	мужот

Number formation

The Slavic way of composing the numbers between 10 and 20, e.g. "one + on + ten" for eleven, called superessive, is widespread. Modern Greek does not follow this.

Language	The word "Eleven"	compounds
Albanian	"njëmbëdhjetë"	një + mbë + dhjetë
Aromanian	"unãsprã"	unã + sprã
Bulgarian	"единадесет"	един + (н)а + десет
Macedonian	"единаесет"	еде(и)н + (н)а + (д)есет
Romanian	"unsprezece" or, more commonly, "unşpe"	un + spre + zece < *unu + supre + dece; unu + spre; the latter is more commonly used, even in formal speech.
Serbian	"jedanaest/једанаест"	jedan+ (n)a+ (d)es(e)t/један + (н)а + (д)ес(е)т

Clitic pronouns

Direct and indirect objects are cross-referenced, or doubled, in the verb phrase by a clitic (weak) pronoun, agreeing with the object in gender, number, and case or case function. This can be found in Romanian, Greek, Bulgarian, Macedonian, and Albanian. In Albanian and Macedonian, this feature shows fully grammaticalized structures and is obligatory with indirect objects and to some extent with definite direct objects; in Bulgarian, however, it is optional and therefore based on discourse. In Greek, the construction contrasts with the clitic-less construction and marks the cross-referenced object as a topic. Southwest Macedonia appears to be the location of innovation.

For example, "I see George" in Balkan languages:

Language	Example
Albanian	"E shikoj Gjergjin"
Aromanian	"U- ved Yioryi"
Bulgarian	"Виждам го Георги." (colloquial form; see note)
Macedonian	"Гo гледам Ѓорѓи."
Greek	"Τον βλέπω τον Γιώργο"
Romanian	"Îl văd pe George."

Note: The neutral case in normal (SVO) word order is without a clitic: "Виждам Георги." However, the form with an additional clitic pronoun is also possible in colloquial speech: "Виждам го Георги." And the clitic is obligatory in the case of a topicalized object (with OVS-word order), which serves also as the common colloquial equivalent of a passive construction. "Георги го виждам."

Adjectives

The replacement of synthetic adjectival comparative forms with analytic ones by means of preposed markers is common. These markers are:

• Bulgarian: по-
• Albanian: më
• Romanian: mai
• Modern Greek: pió; though Greek has retained a few of the earlier synthetic forms.
• Aromanian: (ca)ma

Suffixes

Also, some common suffixes can be found in the linguistic area, such as the diminutive suffix of Slavic origin "-ica" that can be found in Albanian, Greek and Romanian.

Vocabulary

Loan words

Several hundred words are common to the Balkan union languages; the origin of most of them is either Greek or Turkish, as the Byzantine Empire and later the Ottoman Empire directly controlled the territory throughout most of its history, strongly influencing its culture and economics.

Albanian, Aromanian, Bulgarian, Greek, Romanian, Serbian and Macedonian also share a large number of words of various origins:

Source	Source word	Meaning	Albanian	Aromanian	Bulgarian	Greek	Romanian	Macedonian	Serbian	Turkish
Latin	mensa	table	menca (tavolinë)	masã	маса (masa)	-	masă	маса (masa)	-	masa
Thracian	rompea	spear	rrufë	-	roféja	ρομφαία (rhomphaía)	-	-	-	-
Byzantine Greek	λιβάδιον	meadow	livadh	livadã	ливада (livada)	λιβάδι	livadă	ливада (livada)	livada ливада (livada)	-
Byzantine Greek	διδάσκαλος	teacher	-	dascal	даскал (daskal) (colloquial)	δάσκαλος	dascăl	даскал (daskal) (colloquial)	-	-
Byzantine Greek	κουτίον	box	kuti	cutii	кутия	κουτί	cutie	кутија (kutija)	kutija кутија (kutija)	kutu
Turkish	boya	paint, color	bojë (but also ngjyrë)	boi	боя (boya)	μπογιά (boyá)	boia	boja (boja)	boja боја (boja)	boya

Calques

Apart from the direct loans, there are also many calques that were passed from one Balkan languages to another, most of them between Albanian, Macedonian, Bulgarian, Greek, Aromanian and Romanian.

For example, the word "ripen" (as in fruit) is constructed in Albanian, Romanian and (rarely) in Greek (piqem, a (se) coace, ψήνομαι) by a derivation from the word "to bake" (pjek, a coace, ψήνω).

Another example is the wish "(∅/to/for) many years":

Language		Expression	Transliteration
Greek	(medieval)	εις έτη πολλά	is eti polla
	(modern)	χρόνια πολλά	khronia polla
Latin		ad multos annos
Aromanian		ti multsã-anj
Romanian		la mulţi ani
Albanian		për shumë vjet
Bulgarian		за много години	za mnogo godini
Macedonian		за многу години	za mnogu godini

Idiomatic expressions for "whether one or not" are formed as "-not-".

Language	expression	meaning
Bulgarian	ще - не ще	"whether one wants or not"
Greek	θέλει δε θέλει	"whether one wants or not"
Romanian	vrea nu vrea	"whether one wants or not"
Turkish	ister istemez	"whether one wants or not"
Serbian	Hteo- ne hteo/хтео - не хтео	"whether one wants or not"
Albanian	deshti - nuk deshti	"whether one wants or not"
Macedonian	сакал - не сакал / неќел	"whether one wants or not"
Aromanian	i vrei - i nu vrei	"whether one wants or not"

Phonetics

The main phonological features consist of:

• the presence of an unrounded central vowel, either a mid-central schwa /ə/ or a high central vowel phoneme
• ë in Albanian; ъ in Bulgarian; ă in Romanian; ` in Macedonian; ã in Aromanian
• In Romanian and Albanian, the schwa is obtained via centralizing unstressed /a/
• Example: Latin camisia "shirt" > Romanian cămaşă /kə.ma.ʃə/, Albanian këmishë /kə.mi.ʃə/)
• The schwa phoneme occurs in most dialects of the Macedonian language, even in some cases in the western-central dialects, on which the standard is based (сл`нце, к`лбас, к`смет etc.)
• some kind of vowel harmony in stressed syllables with differing patterns depending on the language.
• Romanian: a mid-back vowel ends in a low glide before a nonhigh vowel in the following syllable
• Albanian and Bulgarian: back vowels are fronted before i in the following syllable.

This feature also occurs in Greek, but it is lacking in some of the other Balkan languages; the central vowel is found in Romanian, Bulgarian, some dialects of Albanian, Macedonian and Serbian but not in Greek or Standard Macedonian.

Less widespread features are confined largely to either Romanian or Albanian, or both:

• frequent loss of l before i in Albanian, Romanian, and some Romani dialects
• the alternation between n and r in Albanian and Romanian
• change from l to r in Romanian, Greek, Albanian, and very rarely in Bulgarian
• the raising of o to u in unstressed syllables in Bulgarian, Romanian, Albanian and Northern Greek dialects.
• change from ea to e before i in Bulgarian and Romanian.

See also

• Paleo-Balkan languages
• Balkan languages
• Greek grammar
• Bulgarian grammar
• Macedonian grammar

Notes

References

• Balkan Linguistic Union. In Encyclopædia Orbis Latini .
• Du Nay, André The Origins of the Rumanians: Balkan Linguistic Union. (1977). .
• Victor A. Friedman, "After 170 years of Balkan Linguistics: Whither the Millennium?", Mediterranean Language Review 12:1-15, 2000. PDF -- an excellent survey article
• Grey Thomason, Sarah Linguistic areas and language history. (1999). .
• Joseph, Brian D. Romanian and the Balkans: Some Comparative Perspectives. (1999). .
• Rosetti, Alexandru History of the Romanian language (Istoria limbii române). 2 vols., Bucharest:
• Russu, Ion (1967). The Language of the Thraco-Dacians (Limba Traco-Dacilor). Bucharest: Editura Ştiinţifică.
• (1999). Introducere în lingvistica balcanică (An Introduction to Balkan Linguistics). Iaşi: Editura Universităţii "Alexandru Ioan Cuza".
• Thomason, Sarah G. Language Contact: An Introduction. Washington, D.C.: Georgetown University Press, 2001, pp. 108-9.
• Tomić, Olga Mišeska The Balkan Sprachbund properties: An introduction to Topics in Balkan Syntax and Semantics. (2003). .
• Andrej N. Sobolev (Ed.) Malyi dialektologiceskii atlas balkanskikh iazykov. Muenchen: Biblion Verlag, 2003-