Proto Slavic borrowings

Slavic and Iranian

According to Matasović (2008), "solving the problem of Iranian loanwords in Slavic, their distribution and relative chronology, is one of the most important tasks of modern Slavic studies". Slavs in the era of the Proto-Slavic language came into contact with various Iranian tribes, namely Scythians, Sarmatians, and Alans, which were present in vast regions of eastern and southeastern Europe in the first centuries CE. The names of two large rivers in the centre of Slavic expansion, Dnieper and Dniester, are of Iranian origin, and Iranian toponyms are found as far west as modern day Romania.

For a long time there have been investigators who believe that the number of loanwords from Iranian languages in Proto-Slavic is substantial. For example, Gołąb (1992) maintains that all Slavic words with unexplained initial *x- are in fact Iranianisms. However, there have been other Slavists who have claimed that confirmed Iranianisms in Slavic are in fact surprisingly few, and Matasović has raised broad objections to the body of past Iranianist research. Meillet and Vaillant cite the Slavic word *taparu, axe (Russ. topór, Pol. topór, Sr-Cr. tȍpor), which came from Iranian *tapara- (cf. Persian tabar). (Gołąb, noting the etymological connection with Slavic *tepǫ, I hit, holds that this word is in fact a loan from Slavic into Iranian.) Meillet and Vaillant explain the alleged lack of Iranianisms in Slavic with the argument: "the civilization of warrior and partially nomadic tribes, like Scythian and Sarmatian, could have exerted only a cursory influence on the patriarchal civilization of Slavs."

Matasović criticizes Gołąb's approach as "methodologically unacceptable", emphasizing that initial *x- in Slavic has several sources, some of which have been ascertained (like PIE *#ks-), and others which have not. Matasović recommends that instances of initial *x- in Slavic should first be explained by recourse to regular Slavic sound laws, and that Iranian should be proposed as a source if and only if the etymon has been attested in Iranian and if and only if there is additional phonetic evidence to support the proposal.

Gołąb and Matasović concur that Reczek (1985) and Bernštejn (1961–74) compiled excessively large numbers of alleged Iranianisms by neglecting to thoroughly check the candidates against accepted sound changes in the various major descendants of Proto-Indo-European.

Two likely Iranian loanwords in Common Slavic are

There are Iranianisms which have not been attested in every one of the three first order subdivisions of the Slavic languages (East, West, South); hence the evidence is lacking that these borrowings date back to Common Slavic. Examples:

Meillet and Vaillant considered that the semasiological development of the Proto-Slavic word for god was an Iranianism. In both Slavic and Indo-Iranian, the root that denotes deity also denotes wealth, share (Proto-Slavic *bagu > Common Slavic *bogъ) and Indo-Iranian (Old Persian baga, Sanskrit bhága). However, they did not argue that the Proto-Slavic root itself was a borrowing, despite its similarity to the Old Persian and Sanskrit roots.

One of the Iranian-Slavic lexical isoglosses is a lone adposition: Old Persian rādiy, OCS radi.

Matasović notes typological coincidences between Slavic and Ossetian, an Iranian language whose ancestor was Alanic. In both modern Ossetian and the Slavic group, verbs are conjugated for perfective and imperfective aspects; prefixation is a prominent means of deriving perfective verbs from imperfective verbs; there are certain syntactic behaviors of pronominal clitics in common; both sporadically mark direct objects with the genitive. It remains to be determined, however, whether those correspondences are a result of prehistoric contacts between Slavic and Alanic tribes, or just a case of accidental parallel development.

Slavic and Germanic

It is not certain when speakers of Proto-Slavic first came into contact with Germanic tribes: among Common Balto-Slavic words that have Centum reflexes, there is not a single one that has typical Germanic sound features. As for the Baltic languages, all their prehistoric Germanic loanwords are either mediated through Slavic or are borrowed from Old Norse or Proto-Norse; i.e., borrowed during a period well after Slavic prehistory (which ended c. 600 CE). The conclusion is that the speakers of Germanic must have lived far from the area of the subsequent spread of speakers of Proto-Balto-Slavic.

The Goths are the first Germanic people who can be proved to have had intensive contacts with speakers of Proto-Slavic. Goths are believed to have reached the western shores of the Black Sea between the Danube and Dniester Rivers (an area covered by present day Ukraine and Romania) about 230 CE, so contact between them and Proto-Slavic speakers might have started in the preceding century. Contact between these two groups continued all the way to the early historical period. It is believed that the Gothic language was the intermediary for the entry of many terms of Greco-Roman cultural provenience into Proto-Slavic, for example

Here is a list of words which are generally held to be Germanic loanwords in Proto-Slavic:

This set of loanwords covers diverse semantic fields, fields from which languages readily borrow words: buildings (*xūzu, *mastu, *tūnu); terrain features (*xulmu, *bergu); social interaction and societal structure (*pulku, *želdān, *kōpītej, *činda); animals and cattle (*asilu, *gansi, *skatu).

Interpretation of the Germanic material

Many Germanic loanwords entered Slavic languages well after Proto-Slavic; i.e. in the Common Slavic period. After 600 CE most of them were borrowed from Old High German.

Stender-Petersen (1927) assumes two layers of Germanic loanwords in Slavic (whether into Proto-Slavic or into Common Slavic): words from Proto-Germanic and words from Gothic. Kiparsky suggests a more refined chronological layering:

1. from Proto-Germanic, or Proto-East-Germanic;
2. from Gothic, which have spread to all Slavic languages;
3. from Balkan Gothic, which were confined only to the Slavic South;
4. from Old High German.

All investigators agree that Germanicisms entered Proto-Slavic and Common Slavic longly and subsequently underwent numerous sound changes. *činda, *želdān, *xelmu would predate the Slavic first palatalization; *lōku, *kōsitej, *kōpitej would predate the change, PSl. *aw > PSl. */ō/. Of special interest are certain Proto-Slavic accentual developments that can be observed in Germanic loanwords:

1. Many Germanic words entered Proto-Slavic before the operation of Illič-Svityč's law, and accordingly most Germanic thematic neuters became masculine in Proto-Slavic.
2. Most Germanic loanwords entered Proto-Slavic before the operation of Dybo's law, and accordingly all Germanic loanwords with short initial syllable become Proto-Slavic and Common Slavic oxytones. This fact dates Dybo's law rather late, as almost all Germanic loanwords were affected by it except for those very late ones which entered Common Slavic from Balkan Gothic or Old High German.

Slavic and Celtic

By the time Slavs start to appear in historical records, Celtic languages were already limited to the British Isles and modern-day France. However, during Classical Antiquity Celts populated the regions of Central Europe in which Slavs spread in the 6th and the 7th century, there may have been speakers of Celtic languages in the regions of Slavic expansion. In any case, it seems very possible that Celtic words might have entered Slavic indirectly instead, through Vulgar Latin/early Romance dialects, since Celtic, being but marginally spoken, was probably a low prestige idiom spoken by the lower classes of society. Two likely examples of direct borrowings from Celtic are

1. PSl. *karwā ‘cow’ (Pol. krowa, Russ. koróva, SCr. krȁva), postulated by some to be a feminine derivative of a lost masculine noun supposedly borrowed from Proto-Celtic *karwos ‘deer’ (Welsh carw, Breton karv, Cornish karow), which would in turn be a regular Celtic centum reflex of PIE *ḱr̥(h₂)-u̯o-. Moreover, Lithuanian kárvė, whose accentuation matches that of the Slavic etymons, points to prehistorical Balto-Slavic borrowing. However, this hypothesis fails to take into account Old Prussian curwis ‘ox’, and the Slavic protoform is usually reconstructed as PSl. *kòrva, inherited with incomplete satemization from an o-grade PIE variant *ḱorh₂-u(e)h₂-.
2. PSl. *krawu ‘roof’ (OCS krovъ, Czech/Russ. krov) is traced by some to Germanic etymons with the same meaning (OE hrōf, ON hróf etc.); if Celtic mediation is assumed, from dialectal PIE *ḱrōpo- > Proto-Celtic *krāfo- (cf. MIr cró ‘enclosure’, Welsh crau ‘hovel, pigsty’). However, the Slavic preform is usually reconstructed as PSl. *kròvъ and considered a derivative of *krỳti ‘to cover, hide’. Furthermore, the Celtic words are unrelated to Germanic (< *hrōfa- < *ḱrōpo-), stemming instead from Proto-Celtic *krewo- ~ kruwo-, presumably from PIE *kreu̯h₁- ‘to hide’.

Other Indo-European languages

Speculations as to contacts between Proto-Slavic speakers and other Indo-European languages have abounded in the literature on Slavic historical linguistics. Proposals have included Italic, Illyrian, Thracian, Venetic, and even Armenian.

Loanwords from non-Indo-European languages

There are words that undoubtedly were borrowed from non-Indo-European languages. Most of these words came from Turkic languages, chiefly Bulgar and European Avar. Among commonly cited examples of non-IE loanwords are