This is my original response on such a comment under a completely unrelated topic on Duolingo’s course in Romanian for English speakers. The statement was that: There is no such group as “Balkan languages”.
Actually, there is: Balkan sprachbund. They do not form a single language family, but a language group nonetheless since all Balkan languages share some peculiar similarities despite being from different language families. Probably has to do either with Paleo-Balkan languages, although we still know too little about them, and/or the extensive cultural contact between the people from the region through history.
I have no idea how the others involved managed to make it political, but a sprachbund is also known as an area of linguistic convergence that usually constitutes of genetically unrelated languages that share similarities in grammar, syntax, vocabulary and phonology which are not found in other languages in the respective language family of each language, probably coming from the extensive language contact and the high number of multilingual people throughout the centuries (e.g. due to arbitrary border changes). Several common features are found across these languages though not all need apply to every single language.
So if there are the languages A-F and the features α-ζ and if, for instance, the following statements are true, there will still be a language group that is called a sprachbund by linguists:
- A, B, C, D, E, F – spoken in geographically close area
- A, B, C, D, E, F – from 6 different language families
- α, β, γ, δ, ε, ζ – usually not typical for the respective language family of the languages sharing them
- B, C, D, E, F share feature α; A – doesn’t
- C, D, E, F share feature β; A, B – don’t
- A, B, C, E, F share feature γ; D – doesn’t, but has become to in colloquial speech in recent decades
- A, C, D share feature δ; B, E, F – don’t in literary speech, but do in various dialects of each language
- A, B, C, D, F share feature ε; E – doesn’t
- A, B, C, E share feature ζ; D, F – don’t
- this happens for a lot more features as well but they haven’t been defined as precisely due to their nature (e.g. phonologically common verb conjugation endings in A, E that are more similar, but not exactly the same, to those of B, F as to those of their language families; similar degrees of how synthetic or analytic the languages are; syncretism of various noun cases with the pairs of converged cases the same in C, D, F; etc.).
On the Balkans, there are a lot of similarities like avoidance of the infinitive and future tense formation between genetically unrelated languages from the following language families: Romance, Slavic, Hellenic, Albanian, Turkic and Indo-Aryan in the case of various Romani languages. Quote from the Wikipedia article linked at the beginning of this paragraph (as of 2016-11-20):
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, whereas 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.
Finally, it is true that there could be many sprachbunds everywhere, but the thing is that languages tend to border similar ones (e.g. Western Romance, Germanic, West- and East-Slavic, etc.) or there is a clearly dominant language in the area (e.g. English on the British Isles, Russian in the Former Soviet Union, etc.) which defines language changes in other languages that are spoken there. The curious thing in areas like the Balkans is that this has changed multiple times and in an overly complex way throughout history and it’s a fascinating subject for linguists and hobbyists who want to trace the historical development of such languages without getting politics involved whatsoever.
It is, however, very sensitive topic since quite often the genetic affiliations are unclear and the sprachbund characteristics might give a false appearance of (genetic) relatedness, which you can see in the so-called folk linguistics where amateurs try to prove their point without the proper amount of scientific consistency. This often happens to reanalysis of word etymology (also known as pseudo-etymology).