The tripartition of the Proto-Slavic world is usually dated to the 9th century AD, when the process known as the Liquid Metathesis (LM) separated common Slavic into the 3 main branches: south, west and east.

http://en.wikipedia.org/wiki/Liquid_metathesis

LM involves the fate of the proto-slavic structures CoRC/CerC (C= any consonant, R = liquid resonant R/L).

1) West Slavic operated metathesis without lengthenment: CoRC> CRoC, CeRC>CReC
2) East Slavic operated polyphony: CoRC> CoRoC, CeRC> CeReC/CoRoC
3) South Slavic operated lengthened metathesis: CoRC> CRōC~ CRaC, CeRC > CRēC ~ CRěC

The south slavic cases show that when the LM was occuring proto-slavic had still the long vowels, because the new long vowels that were produced by the process followed the same evolution as the inherited ones.

Examples:

1) Evolution of long vowels inherited from PIE:

Root *g'enh3- "know" > zero grade *g'n.h3- ~ *g'nō- (cf. Greek γιγνώσκω, Latin cognōscō) > Slavic znati

http://en.wiktionary.org/wiki/Append...o-Slavic/znati

Root *deh3- ~ dō- "give" (cf. Greek δῶρον , Latin num) > Slavic dar

http://en.wiktionary.org/wiki/Append...avic/dar%D1%8A

2) Evolution of the new south slavic long vowels:

PIE root *gherdh- "to fence" (cf. English gird) > noun *ghordhos ("fenced/wall settlement") > Balto-Slavic *gardas > late proto-slavic gordŭ > South slavic grōdŭ > gradŭ

http://en.wiktionary.org/wiki/Append...vic/gord%D1%8A

South Slavic grad, Vlah
East Slavic gorod/horod, Voloh/Valah
West Slavic grod/hrod, Vloh (Vlochy = Italy in Polish)

Similarly, the root *h2melg- "to milk" gave proto-Germanic *meluks (the ancestor of english milk and german milch), which was borrowed into proto-Slavic as *melko:

http://en.wiktionary.org/wiki/Append...o-Slavic/melko

South Slavic: *melko > OCS mlěko > modern SSlv variants mleko,mlyako,mljeko etc.
West Slavic *melko > mleko/mloko etc.
East Slavic *melko > moloko/malako

In this way, a Bulgarian, for example, can translate germanic "gastarbeiter" in bulgarian using the same IE roots as "gost rabotnik".

Since the Old Church Slavonic (OCS) was codified in the second part of the 9th century AD and it shows the SSlv LM, most linguists agree that most of the LM process had been concluded during the period 750-850 AD.

The question is, for how much longer have the South Slavs operated the LM? That is which was the real terminus ante quem of this linguistic process?

http://en.wikipedia.org/wiki/Terminus_ante_quem

I have found 3 examples that show that LM was still operative after 1000 AD.

1) The Serbian name Raban for Arbanon:

The first reference to Albanians as historical people comes from an Old Bulgarian document written around 1010 AD possibly in Ohrid, where the "Arbanasi" are mentioned as one of the "half-believers" (Christians but not Orthodox):

http://www.albanianhistory.net/en/te...99/AH1000.html
It can be seen that there are various languages on earth. Of them, there are five Orthodox languages: Bulgarian, Greek, Syrian, Iberian (Georgian) and Russian. Three of these have Orthodox alphabets: Greek, Bulgarian and Iberian. There are twelve languages of half-believers: Alamanians, Franks, Magyars (Hungarians), Indians, Jacobites, Armenians, Saxons, Lechs (Poles), Arbanasi (Albanians), Croatians, Hizi, Germans.
This form is in agreement with the earliest Greek name «Αρβανίται» for the Albanians. In Greek, «Αρβανίτης» literally means "inhabitant of Arbanon", just like Latin Arbanensis/Albanensis > Italian Albanese.

Arbanon was the original Albanian nucleus from where the Albanians begun gradually to spread and is commonly equated with the tableland around the river Mat north of Kruja:

http://en.wikipedia.org/wiki/Arbanon#Arbanon

In the Serbian poetry of the Ottoman period the form Arbanasa = "Albanian" is attested as the ethnikon of the legendary thief Musa.

However, in a Serbian document dating around 1200 AD, the region Arbanon is called Raban showing south slavic metathesis.



2) The "ping-pong" Bulgarian-Tosk Albanian toponym Labėria:

http://en.wikipedia.org/wiki/Labs_(people)

The region Labėria is usually etymologized from Greek Αλβανία (Albania) where:
i) The Bulgarians have operated the LM Albania > Labania and the Tosk Albanians have operated the rhotacism of intervocalic -n->-r- Labania > Labėria as in Avlona > Vlonė > Vlorė.

http://en.wikipedia.org/wiki/Vlor%C3%AB#Etymology

So far so good. But the problem is that in order for the Greeks to call the place "Albania" in the first place, there must be Albanians living there and no historian admits that there were Albanians in what is modern south Albania before 1100 AD.

So the south slavic LM has occured sometime after 1100 AD.

3) The Macedonian word dlabok = "deep"

http://en.wikipedia.org/wiki/Dlaboka_Reka

The adjective dlabok = deep is related to serbo-croat dubok:

http://en.wiktionary.org/wiki/dubok

and derives from proto-slavic *dъlbokъ.

Now we do know that the yers were still present in OCS and that around 1000 AD most south Slavs were pronouncing the word as *dl.bokъ with a liquid resonant l. as in vl.kъ = "wolf" and pl.nъ = "full".

So the subsequent Serbo-Croat evolution was full vocalization and loss of the l (l.>u): dubok, vuk, pun, meanwhile the Macedonian evolution was l.>ol: *dolbok, volk, polno.

Both evolutions had certainly occured after 1000 AD.

But in order to go from *dolbok to dlabok we need the south slavic Liquid Metathesis. This means that the LM in this example has occured much later than 1000 AD.