Slavic liquid metathesis and pleophony
Slavic liquid metathesis refers to the historical phenomenon of metathesis of liquid consonants occurring in Common Slavic period in South Slavic and Czecho-Slovak area. Onomastics evidence indicates that it seems to have occurred sometimes in the latter half of the eighth century. Closely related corresponding phenomenon of pleophony occurred parallelly in East Slavic languages.
Law of open syllables 
During the Common Slavic period, a tendency known as the law of open syllables created the series of changes that completely eliminated closed syllables. This was evident in OCS, which had no closed syllables at all. Some of these changes include the monophthongization of diphthongs, loss of word-final consonants (e.g. 3rd person singular aorist OCS reče < *reket, OCS N sg of o-stems -ъ < *-us etc.), simplification of some medial consonant clusters (e.g. OCS tonǫti < *topnǫti etc.) and the formation of the nasal vowels *ǫ and *ę from *am/*an and *em/*en respectively.
Another change involved a liquid consonants (R) *l or *r in closed-syllable *eRC and *aRC clusters, which had to be eliminated. The application of the law of open syllables for such clusters differed amongst the already differentiated Slavic dialects, but characteristically manifested in a number of dialects as the metathesis of liquid consonants, and is therefore called the liquid metathesis.
Historical evidence 
There are glosses of Slavic words that show no effect of liquid metathesis in foreign-language sources, such as when the late 8th century Greek chronicler Theophanes writes Slavic names as Άρδάγαστος and Δαργαμηρός (OCS versions of these names would be Radogostъ and Dragoměrъ). Liquid metathesis can be seen in various borrowings preserved in toponymics; Latin Arba > Serbo-Croatian Rȃb, Latin Albōna > Serbo-Croatian Làbīn, Latin Scardōna > Serbo-Croatian Skràdīn etc. Very illustrative of the process is the name of the King of the Franks Carolus Magnus, died 814, which was borrowed initially as *karlju to become the proper word for king in all Slavic languages (Bulgarian and Macedonian крал, Serbo-Croatian and Slovenian kralj, Polish król, Czech král and Russian король).
Reflexes in Slavic languages 
In South Slavic the vowel and the liquid metathesize, and as a side-effect the vowel lengthens (*e > *ē > ě, *a > *ā > a):
- *al > la
- *ar > ra
- *el > lě
- *er > rě
- PSl. *gardu > OCS gradъ ('settlement')
- PSl. *walti > Serbo-Croatian vlȃt ('blade, stalk')
- PSl. *wertmēn > OCS vrěmę ('time')
- PSl. *melka > OCS mlěko ('milk)
Word-initially, metathesis with lengthening occurred always only in south and central Slovak dialects (i.e. just like in South Slavic), and in the rest of West and East Slavic languages only when the syllable is under acute (rising) accent; compare: Proto-Slavic *a̋rdla ('plough'):
If the syllable was not acuted, metathesis in West and East Slavic occurs without the lengthening so EPSl. *a retains short quantity and yields /o/; compare EPSl. *ȃlkuti ('elbow') > Serbo-Croatian lȃkat, but Czech loket.
Word-medially, on the other hand, the following occurs: in Polish and Sorbian languages metathesis without lengthening occurs; compare Polish brzeg, mleko, groch, młot as opposed to OCS brěgъ, mlěko, Slovene gràh, OCS mlatъ. In North-West Lechitic (northern Kashubian, Slovincian, Pomeranian and Polabian) *CalC and *CelC yield ClŭC (Polabian glåvă ‘head’, å < ъ), *CerC > CreC (without lengthening, as in Polish), while in *CarC, ar becomes ŭr, just like word-initially under acute (Polabian råmą ‘arm’ < *rъmę < *armę), but does not undergo metathesis. Compare Polabian porsą to Slovene prasè and Pomerian gard (often in toponimics, e.g. Białogard and similar) to OCS gradъ (note that unchanged ar in *gardŭ would have given or in Pomeranian). In Czecho-Slovak word-medial metathesis occurs with the lengthening, just as in South Slavic; compare Czech mlat, hrách to Polish młot, groch with /o/ inside. East Slavic languages manifest so-called pleophony (also called polnoglasie or full vocalization) - *CarC > CoroC, *CerC > CereC, and *CalC/*CelC > ColoC; compare Russian górod, béreg, mólot, molokó. Here the closed syllable problem is resolved by inserting another vowel after the liquid consonant. In North-West Lechitic the reflexes of *CelC and *CalC are the same.
Complete and incomplete, first and the second metathesis 
If one considers the liquid metathesis complete only under the condition that it occurs with the corresponding vowel lengthening, then the complete metathesis occurs only in South Slavic, partially Slovak and in non-word-initial position in the whole Czecho-Slovak area. The complete metathesis has been operational in all Slavic languages under the acuted syllable. Under the word-initial non-acuted syllable there was no lengthening except in South Slavic and partially Slovak. As it was mentioned, word-medially the complete metathesis occurred, besides in South Slavic, in Czecho-Slovak group; in Polish and Sorbian it operated without lengthening, and in North-West Lechitic it didn't operate even in the case of *CarC syllables (otherwise the incomplete metathesis occurred - without the lengthening). In East Slavic languages pleophony yielded *CVRC > CVRVC. The reflex of *l in North-West Lechitic and East Slavic is always "hard".
Since the reflexes of acuted word-initial *ar- and *al- have been the same in all Slavic dialects, the change of acuted *ar-, *al- must have preceded the change of other syllables closed by a liquid, where the reflexes are different. So one can distinguish the first and the second metathesis of liquids.
As it was stated, the liquid metathesis occurred in the Common Slavic era, at the end of the 8th century. That can be shown by the fact that the name of Frankish ruler Charlemagne underwent this change, so:
- Old High German Karl[note 1] > PSl. *karlju[note 2] > Common Slavic *korljь > Russian koról’, Polish król, Slovak kráľ, Serbo-Croatian krȃlj
It has been suggested that East Slavic preserved the actual state of affairs, i.e. that the vowel was inserted in Common Slavic period, and only subsequently it was lost in all dialects except in East Slavic, in a position preceding the liquid. So the exact development would be, e.g. in case of Serbo-Croatian:
- PSl. *bardā 'beard' > *Common Slavic *boradā > *baradā > SCr. bráda
See also 
- Note that karl – which is not only a proper name, but also a common noun with the meaning "adult male" – is a normalized form based on the Old East Franconian dialect. The common noun is also attested as charal in Old Upper German, i. e., with an epenthetic vowel, for example in the Old Alemannic rendering of the Rule of Saint Benedict; this may be the origin of the intrusive vowel in Carolus, the Latinised version of the name, but the Slavic process seems to be unconnected.
- The rendering of OHG l as PSl. *lj is regular; the reason is that PSl. plain *l was dark, while OHG l was a normal [l], which was identified with the PSl. palatalised liquid *lj.
- Holzer 2007:55-56
- Matasović 2008:150
Further reading 
- Kapović, Mate (2008). Uvod u indoeuropsku lingvistiku (in Croatian). Zagreb: Matica hrvatska. ISBN 978-953-150-847-6.
- Matasović, Ranko (2008). Poredbenopovijesna gramatika hrvatskoga jezika (in Croatian). Zagreb: Matica hrvatska. ISBN 978-953-150-840-7.
- Holzer, Georg (2007). Historische Grammatik des Kroatischen. Einleitung und Lautgeschichte der Standardsprache (in German). Frankfurt am Main: Peter Lang. ISBN 978-3-631-56119-5.