This wall was built in the days of topoterites Tzoula-beg, son of Poletas, in the year 6503.

 

Latyshev's edition is based on the transcription made by Skubetov, who had made several mistakes: in line 2 he did not see either a superscript mark over tau or a sigma inside omicron; in line 4 he mistook tau for iota, and in line 5 he mistook a gamma for tau. Because of these mistakes, the builder was thought to have been a hypoterites Tzoula-bey rather than a topoterites Tzoula-beg. Bely and Solomonik who had rediscovered the inscription corrected all these mistakes (see their publication for the history of studies).

1. Building formula with the verb ἐκτίσθη is attested in the Middle Byzantine period (RECAM 98 (894 C.E.)).

2. For the colloquial form τοῦτος we may compare V 219.

4. In the Xth century, a topoterites was a lieutenant for the commander of the thema or of the tagma, but over the course of the XIth century, by 1100, the title comes to signify commandant of a fortress (ODB, 2095–2096) — in our case both meanings are possible. The former makes better sense with the expression "in the days of" (unique formula ὑπὸ ἡμερο͂ν has one parallel in the inscription from Laki, 1364 C.E. (V 144)), where it refers to a Khan's representative - Kutluk-Temir. The second possibility, in turn, fits well with the local nature of the construction project (judging by the absence of either an emperor's name or that of the thema's strategos). There is no specification as to the regional association of the topoterites Tzoula-beg who served in the capacity of deputy strategos, but this must have been implied since in 994–995 at Mangup only a topoterites of Cherson's strategos could have been in charge. Thus we may conclude that at the end of the Xth century a Byzantine military garrison had a base at Mangup. The wall, mentioned in the text, was apparently built in connection with the introduction of the garrison and signifies active Byzantine presence in Mountainous Crimea.

5. The name of the builder represents a considerable problem. The most logical interpretation of the last part -βήγη is to see in it the Turkic term "Beg." This word, however, would have been pronounced by Byzantines as "vig," and in contrast to the terms πεχ and μπεκ, the variant βήγη is not attested in Moravcsik (1958, 250). At the same time, the possibility of conveying the Turkic term Beg as βήγη is bolstered by the analysis of the title of Bulgarian khans Omurtag and Malamir who were called κανα συβήγι: some scholars interpret the second term as sü-begi «Lord of the army» (see Beševliev 1963, № 251). We also encounter the same term in one inscription from Cherson dated to the IX–Xth centuries: ...]υβήγις (V 103). What is unclear is whether we should be restoring the nominative of the name of our builder as Τζουλα-βήγη or as Τζουλα-βήγης. On the history of the clan of Tzoula and its identification with the Tzoula of our inscription, see Vinogradov 2009a. One other possibility is the "p" of Poletas corresponds to the Turcic "b" (as in πεχ for bek), in which case Poletas might be an approximation of some such Turcic name as Boldy, Boltoy, Bulat, Bulatay, Bultash or alike.

6. The patronymic of our builder from Mangup is also puzzling: Πολέτας. It is not Turkic because there is no sound "p" in Turkic languages. At the same time, the only Greek example of this name is quite late: it is known on Corfu in 1391 C.E. as a family name — Theodoris Poletas (PLP, № 23472). Thus, it might have been a name of non-Greek origin which later became a family name, as did many others, including Tzoula.

7. The gravest, and as it turns out, fatal mistake was commited by Skubetov (who was a draftsman, without any knowledge of Greek) in line 7: he represented the first letter in the year as alpha, and for this reason the inscription was dated to 1503. This late date was accepted both by Bely and Solomonik, however, several considerations demonstrate that such dating is impossible. Firstly, dating by anno domini is known in Crimea prior to the XVIIIth century only once: in the inscription from the Genoese Sudak, 1412 (?) C.E. (V 258). Secondly, palaeographic characteristics are telling: letter shapes show no complex features typical of Crimean inscriptions of the XIV-XVth centuries. Thirdly, the term 'topoterites' is attested in the XVIth century only as a clerical position appointed by mitropolites, and such a functionary could not have been the builder of the Mangup wall. Fourthly, all representatives of the Tzoula family are attested in Crimea only at the end of the Xth - beginning of the XIth centuries. Finally, a closer inspection of the stone shows an S-shaped stigma (it is also visible on the photos published by Bely and Solomonik 1984, fig. 2) with a slanting stroke on the left, typical of inscriptions dated to the IX-Xth centuries. Therefore, there can be no doubt that our inscription is dated to year 6503 since Creation, that is, to the period between the 1st of September 994 and 31st of August 995 C.E. (on the correspondence of years since Creation and since the Birth of Christ, see Introduction IV. 4. D).

Despite indications that the letter-cutter was familiar with literary culture (cf., e.g., the abbreviation υ(ἱο)ῦ (see Avi-Yonah 1940, 107)), the quality of cutting is not very high: the lines are quite uneven. In addition, the cutter misjudged the amount of available space and could not fit the text inside the inset area, crossing over onto the border. The letters are wider, simpler and less ornate than in the inscriptions from the turn of the century (e.g., there is no beta with decorative horizontal stroke at the bottom). Other features, such as the pointy top loop of beta and kappa with free-standing vertical, which are typical of Crimean inscriptions of the Xth century, are also present in our text.

The Xth-century date of our inscription does not necessarily mean that the present wall at Tabana-Dere had been built at that time: the stone could have been taken from an earlier dismantled wall and re-used. If this were the case, however, it would be surprising that the later builders would have taken so much care to display an old inscription in a very prominent place (facing anyone ascending along the wall), while neglecting an opportunity to set up their own commemoration. Perhaps the wall itself should also be redated (Gertsen 1990)? A decisive argument in favour of the latter suggestion would seem to be the radiocarbon dating of a wooden beam from the wall, conducted by L.V. Firsov (1976, 170): it showed that the tree used for the beam had been cut down in 865–925 C.E. The bark could have survived till 994–995 C.E., but not until the early XVIth century.

The new date of the inscription raises questions both about the date of the wall into which it was incorporated and the date of the entire complex of the cave monastery, since the wall's northeast end leans against it: in the top level of the cave, the date is 1220–1221 C.E. (V 196), and in the church - 1224–1225 (V 174). With the original assigned date of 1503, it was incomprehensible that the wall should run up against a hermit's cell in the top level of the cave, and also that there should have been an unprotected passage made in the wall next to the monastery. With the chronology corrected, it becomes clear that both the monastery and the passageway were constructed 200 years after the construction of the wall, at the time when the latter might have already lost its defensive function (either as a result of the fall of the Byzantine empire in 1204 C.E., or due to internal causes, for example, raids of barbarians).