The Romanian language is a Romance language, derived from Latin, introduced during the Roman occupation of the Balkans. The recruitment of Thracians, Illyrians and Celts into the Roman army gradually spread the use of Latin through the Balkan regions. Many sources have defined a line for the divide of Greek and Latin usage at this time which runs across modern Macedonia and Bulgaria below the Balkan mountains.
Romanian language contains features reflecting continued contact with Romans until the decline of the Roman Empire and influx of Slavic peoples in the 6th century. From this time separate Northern and Southern dialects formed. The two principle northern dialects are Daco-Romania spoken in modern Romania and the Republic of Moldova, and Istro-Romanian spoken in a few villages in the Istria peninsular of Croatia. The two principle southern dialects are Arumanian or Macedo-Romanian spoken in Macedonia, Albania and Greece, and Megleno-Romanian spoken in a few villages in Greece. Romanian is also still the first language of many Romanian Jews who now live in Israel.
Romanian has some 100 words in common with Albanian and peculiar to only Romanian and Albanian. Albanian is thought to be the only surviving language thought to be related to the languages spoken by the Thracians, Dacians and Illyrians. Hence, it is thought that these words have remained from an earlier language, possibly Dacian, and that Romanian and Albanian have a similar route in the old Balkan languages.
There are also many words of Slavic, Greek and Turkish origin, however, in the 19th century words from the other Romance languages were introduced to re-Romanize the vocabulary. The Cyrillic alphabet was used until it was replaced by the Latin alphabet by the end of the 19th century.
Two language trees are shown, the first is the traditional Indo-European language tree showing current and past European languages. The second tree is based on a statistical analysis, analysed in a similar way to that of genetic DNA information. This allows the splitting of the languages to be derived with some idea of statistical reliability. The Albanian (then possibly the Illyrian, Thracian and Dacian languages if they related) and Armenian are have a clearly separate earlier separation so may represent the first wave of neolithic farmers, the later branches are possibly from later migration, some postulate from the Kurgan area. The separation of Greek is less certain from the statistics, but somewhere between.