1 Introduction

In this paper we will present two eleventh-century phrases that are inscribed many times on the walls of the St. Sophia Cathedral in Novgorod: коуни рони and парехъ мари. The first one has been known since 1978, when two instances were published by A. A. Medynceva without any further interpretation.Footnote 1 Since then, a plausible explanation for its provenance has not yet been provided, although the non-Slavic nature of the text seemed quite obvious.

In recent years, the epigraphic data in the St. Sophia Cathedral have been thoroughly reexamined by A. A. Gippius and S. M. Mikheev.Footnote 2 In the course of this work more than forty other cases of коуни рони were discovered. Another puzzling graffito—парехъ мари—was identified by comparing its two fragmented attestations. In the summer of 2011, both inscriptions were brought to the attention of J. Schaeken who, together with H. Gzella, proposed a possible Semitic provenance and a religious nature of these expressions.

In Sect. 2 below, we will first provide the linguistic arguments which support the hypothesis of a Hebrew source for коуни рони and a Syriac one for парехъ мари. Then, in Sect. 3, we will discuss the possible historical pragmatic context of both exclamations. In the case of коуни рони, A. A. Gippius put forward a most plausible hypothesis that the many attestations of this phrase—which is, in fact, the oldest direct proof of contact with Hebrew in early Rus’—should be connected with the seizure of Novgorod and the plundering of the St. Sophia Cathedral in the year 1066. As for парехъ мари, S. M. Mikheev was able to identify the handwriting of the two instances of this phrase with another inscription in the Cathedral, which reads ‘Efrem the Syrian’.

2 The inscriptions

2.1 Коуни рони

The first inscription to be discussed is коуни рони. It is attested more than forty times on two adjacent compact surfaces, on the northern wall of the Diaconicon and on the western wall of the passageway from the altar to the Diaconicon (see walls A and B on the plan in Fig. 1). Paleographic research has shown that most of the instances of коуни рони are inscribed in different hands. In some cases, the word is accompanied by a personal name, like, for instance, коуни рони Хот нъ Носъ or коуни рони До{д}брата ().

Fig. 1
figure 1

Plan of the south-eastern part of the St. Sophia Cathedral, indicating walls A and B where the inscriptions коуни рони and парехъ мари are located

Full size image

Stratigraphically, the коуни рони inscriptions can be dated before the year 1109 (cf. Gippius and Mikheev 2011, 37f.). Thus, they belong to the oldest period of the St. Sophia Cathedral, which was built between 1045 and 1050. It should also be noted that in a number of cases the inscription was crossed out afterwards (see below, Sect. 3.1).

We propose that коуни рони is a Slavic adaptation of the Hebrew expression qūmī ronnī () which occurs in the Hebrew Bible, verse 2:19 of the Book of Lamentations (Kniga Plač Ieremii). Hebrew qūmī ronnī consists of two verbal forms, both imperatives, feminine singular, of the unmarked stem. Its meaning is ‘Arise, cry out’: “Arise, cry out in the night, at the beginning of the watches! Pour out your heart like water before the presence of the Lord! Lift your hands to him for the lives of your children, who faint for hunger at the head of every street”.Footnote 3 As is well known, the Book of Lamentations relates to the Babylonian conquest and mourns the destruction of the city of Jerusalem in the sixth century BC. More particularly, Chap. 2 describes the horrors of the siege, the sins of the nations and the wrath of God.Footnote 4

According to the extensive French patristic online database Biblindex,Footnote 5 there are but very few references by Greek and Latin Church Fathers to verse 2:19 of Lamentations. The database provides us with only three names: Origen (Adamantius), Theodoret of Cyrrhus, and St. Ambrose of Milan. This may point to a rather limited social environment in which the exclamation qūmī ronnī  was known and used: among Jewish people, of course, and others who were in close contact with the Jewish community.Footnote 6

The inscription коуни рони is written in three different ways: (a) (approximately ten times), (b) (at least four times), and (c) (more than fifteen times) (see Figs. 2a, 2b). In a number of cases the first vowel letter cannot be identified with certainty, including some of the instances where the inscription was crossed out. The variation in the spelling of the first syllable () can be explained by the realization of etymological /ū/ as [i] or, less frequently, as [ü] in the customary pronunciation of Hebrew in Ashkenazi communities in the Ukraine, Poland, Western Hungary, and other places, just as in Southern Eastern Yiddish (cf. Morag 1972, 1135, and, more extensively, Katz 1993, 61–65). The Hebrew velar /q/, by contrast, which was originally pronounced as an unvoiced uvular plosive [q] (cf. Khan 1997, 89), was realized as [k] in Ashkenazi Hebrew (Morag 1972, 1133). In Slavic, qū-/kū- can be rendered either in an approximate way by means of the ‘foreign’ sequence or in the slavicized form . The third variant, , also reflects the foreign nature of the initial consonant-vowel combination by using the Slavic grapheme which was introduced to reflect Greek υ. In fact, the variation , , for Hebrew qū-/kū- (according to Ashkenazi pronunciation) finds a perfect parallel in the rendition of Greek κυ in Old Slavic (see Diels 1963, 27ff., for examples from Old Church Slavonic, who attributes the use of these coexisting spellings to a pronunciation of Greek υ that vacillated between [i], [u], and perhaps also [ü]; see also Krivko 2004).

Fig. 2a
figure 2

Photograph of attestations of коуни рони on wall A

Full size image
Fig. 2b
figure 3

Drawings of attestations of коуни рони on wall A

Full size image

Hebrew qūmī is consistently written with the consonant letter -н- in коуни рони; there are no attestations which show the expected consonant -м-. However, the spelling is well attested elsewhere in Slavic because of its occurrence in the New Testament, where in Mark 5:41 we read: “He took her by the hand and said to her, ‘Talitha cum’, which means, ‘Little girl, get up!’ ”. In Old Church Slavonic, in the Codices Zographensis and Marianus, the quote is translated as , and in the Codex Suprasliensis as (cf. Cejtlin et al. 1994, s.v., 689).Footnote 7

Since the phonetic value of Hebrew m thus does not prevent in any way its rendition in Slavic (Cyrillic) by м, there must be some other reason for the consistent writing of н instead of м on the walls of the St. Sophia Cathedral. Perhaps the scribes only knew the Hebrew expression qūmī ronnī by heart, without actually being able to understand Hebrew. They may have heard it and misinterpreted it by assimilating the m of qūmī with the n of the following ronnī.Footnote 8 This tentative, hearer-based explanation is understandable in view of the phonetic nature of the two nasal consonants involved as well as the parallel syllabic structure of the two words.

2.2 Парехъ мари

In addition to our explanation of коуни рони as a biblical quotation of Hebrew provenance, we will now address another Cyrillic inscription, which in our opinion has a similar Semitic background: парехъ мари. It is found twice in the near vicinity of the коуни рони inscriptions (once on the northern wall of the Diaconicon and once on the western wall of the passageway from the altar to the Diaconicon; see Fig. 1) and must also be dated before 1109. The two attestations of парехъ мари are most probably written in the same hand, which does not seem to coincide with the hands which inscribed коуни рони. Although in both cases the text is partially damaged, a paleographic comparison between them leaves no doubt concerning the sequence of letters (see Figs. 3a, 3b and 4a, 4b).

Fig. 3a
figure 4

Photograph of парехъ мари on wall B

Full size image
Fig. 3b
figure 5

Drawing of парехъ мари on wall B

Full size image
Fig. 4a
figure 6

Photograph of парехъ мари on wall A

Full size image
Fig. 4b
figure 7

Drawing of парехъ мари on wall A

Full size image

We propose that парехъ мари could be a Slavic adaptation of Classical Syriac /barrek mār/ ‘Bless (imperative, masculine singular, factitive stem), O Lord’. We are dealing with a very popular liturgical formula which is used before and during prayers.Footnote 9 Other Semitic languages can confidently be excluded as a source for парехъ мари in view of the fact that mār ‘Lord’ is an Aramaic isogloss. Also, the liturgical expression // is widely spread in Syrian Christianity.Footnote 10

As for the rendition of // by Cyrillic парехъ мари, we can observe that the Syriac expression was obviously perceived as two separate words. This explains the insertion of the jer-letter ъ at the end of the first word, which is in accordance with Slavic writing conventions in the period when this phrase was inscribed on the walls. The preceding velar is spirantized in Syriac after a vowel and pronounced as [x] (see, e.g., Daniels 1997, 135f.), hence spelled as x in the Cyrillic inscription.

The alternation between initial Syriac b- (i.e. a voiced bilabial stop; see Daniels 1997, 132) in // and Slavic p- in парехъ has, of course, many typological phonetic parallels elsewhere and does not prevent us from identifying the Syriac form as the source of the Slavic one.Footnote 11 Word-initial devoicing of /b/ may be facilitated by a following voiceless obstruent, as is the case here.Footnote 12 It is obvious that both п- (instead of б-) and -х- (instead of -к-) in do not reflect the Syriac model in its usual written form; the writer of the two inscriptions must have noted them down in the way they were perceived by ear. With all due caution, one could possibly imagine that the devoicing of initial /b/, trivial though it is, may have been caused by a vernacular pronunciation of the Syriac phrase (for examples of a shift /b/ > /p/ in Neo-Aramaic, albeit in distinct environments, see Nöldeke 1868, 51, and cf. p. 400). As the linguistic background of the scribe is unclear, however, this remains speculation, and other explanations should not be excluded.

The identification of Syriac /mār/ with мари, by contrast, is less easy to account for and needs two comments. Firstly, in /mār/ the long (stressed) vowel -ā- would point to an East Syriac reflex of etymological */ā/. In West Syriac, the vowel was pronounced as [ō] (see Daniels 1997, 134f.). The distinction may at face value suggest that the origin of Slavic мари can be narrowed down to the Eastern variety of Syriac, which is traditionally the language of the Nestorian Church. However, this cannot be substantiated within our specific context, because in the Old Novgorodian dialect the vowel letter -а- in мари can reflect both foreign [a] as well as [o]; cf., e.g., in the birchbark corpus the spelling of personal names of Greek origin such as , , , and (Zaliznjak 2004, 205).

Secondly, the final vowel letter -и in мари is unexpected. Although in Syriac /mār/ is written mry in the (usually unvocalized) consonantal text, the spelling of final -y for the first person singular possessive suffix /-ī/ is of a purely historical nature; in the eleventh century—the time of our inscription—it had not been pronounced for already many centuries (see Beyer 1984, 122–125). Against the background of our observations on // = парехъ (see above), -и in мари as a purely graphic analogy patterned after the historical spelling of /mār/ in Classical Syriac seems to be at odds with the evidence for an oral transmission of the phrase in question. Instead of мар и we would expect a hearer-based rendition of /mār/ as ∗ or ∗ , including a final jer-letter in accordance with Slavic spelling conventions.

There are two possible explanations—admittedly, both of a speculative nature—for the occurrence of a final -и in мари. The first one is the simple assumption that the spelling -и was triggered by the final -и in the many occurrences of коуни рони in the near vicinity of парехъ мари. This hypothesis presupposes that коуни рони was inscribed first on the church walls, prior to the two instances of парехъ мари. The analogical spelling ( ∼ мари) may well have been prompted by the symmetrical structure of both expressions, each of them consisting of two words with two syllables.Footnote 13

The second explanation is based on the assumption that мари reflects an unattested substandard pronunciation of the Syriac form, presumably influenced by a contemporary Aramaic vernacular, in which a phonetic distinction is made between forms written with a possessive suffix /-ī/ and without. Thus, for instance, /mār/ means ‘Lord’ and /mārī/ ‘Lord-poss’. Such a phonetic distinction is common for Neo-Aramaic languages, but for standard Syriac it is of a purely graphical nature. The possibility that мари reflects a ‘vulgar’ Syriac pronunciation [mārī] is not inconceivable in view of the coexistence of Classical Syriac as a liturgical language and the forerunners of Neo-Aramaic varieties as idioms that were actually spoken in daily life. Although the respective language situation requires much more in-depth research, it is quite feasible to assume that Classical Syriac was often used by speakers of Aramaic vernaculars as a kind of lingua franca or restricted to liturgical contexts. Hence, one can entertain the possibility of a hybrid Syriac / vernacular Aramaic form.

3 The historical pragmatic context

The possible identification of коуни рони as a Slavic adaptation of a Hebrew biblical quotation (‘Arise, cry out’) and парехъ мари as a Slavic adaptation of a Syriac liturgical formula (‘Bless, O Lord’) forms the basis for the following reconstruction of the historical pragmatic context of the inscriptions. In both cases, it is important to stress their specific location in the St. Sophia Cathedral, i.e. the Diaconicon and the passageway from the altar to the Diaconicon. This part of the church was usually only accessible for the clergy, not the average churchgoer.

3.1 Коуни рони

The phrase коуни рони is written in the Cyrillic alphabet and in a hearer-oriented spelling. More in particular, the consistent long-distance assimilation of the consonant letters -m- -n- > -n- -n- in this expression points to writers with little or no active command of Hebrew. We might therefore be dealing with Slavs, probably Novgorodian clergymen, who must have had a special reason for inscribing many times one and the same phrase on the walls of the Cathedral. The fact that all коуни рони inscriptions are compactly distributed over a relatively small surface areas suggests that they were all written within a short period of time. It is also reasonable to assume that those who wrote the phrase on the walls, though not being able to reproduce it correctly, were still aware of its meaning and reference to the conquest of Jerusalem.Footnote 14

On what kind of special occasion might verse 2:19 of the Book of Lamentations have been of topical interest in the history of Novgorod? Within the timeframe 1045–1109 there is only one major event in the history of the city which parallels the Babylonian conquest of Jerusalem, namely the seizure of Novgorod by Vseslav Brjačislavič of Polotsk. This tragedy is vividly reported in the First Novgorod Chronicle in the passage that deals with the year 1066: “Приде Всславъ и възя Новъгородъ, съ женами и съ дтми; и колоколы съима у святыя Софие. О, велика бяше бда въ час тыи; и понекадила съима” (Nasonov 1950, 17).Footnote 15 The chronicler specifically mentions the plundering of the St. Sophia Cathedral, including the removal of its bells and chandeliers: ‘Oh great was the misery at that time!’ From the Hypatian Chronicle we learn that Vseslav took away more from the Cathedral than only its bells and chandeliers. In the year 1178, Mstislav Rostislavič (‘The Brave’) “поиде на Полтьскъ на зятя на своего на Всеслава, ходилъ бо бяше ддъ его на Новъгородъ и взялъ ер(у)с(а)л(и)мъ ц(е)рк(о)вныи и сосоуды слоужебны … Мьстиславъ же все то хотя оправити Новгородьскоую волость и обиду” (Šaxmatov 2001[1908], 608).Footnote 16 Thus, “more than a century later the Novgorod Prince still wants to avenge this robbery” (Jakobson and Szeftel 1966, 341, fn. 111) and the stealing of the ‘Church Jerusalem’ and the ‘liturgical vessels’ are recalled. With the former a Eucharistic vessel is meant which is also known as the Zion or church tabernacle. Both types of precious tabernacles (the ‘Small’ and the ‘Great’) from medieval Novgorod have been preserved and are commented on in detail by I. A. Sterligova in the preface of the 1996 catalogue Dekorativno-prikladnoe iskusstvo velikogo Novgoroda.Footnote 17 Sterligova stresses the liturgical symbolism of these Zions as images of Holy Jerusalem and the extraordinary significance of the theme of Jerusalem for the St. Sophia Cathedral in Novgorod: “Для Новгoрода XI в., расположенного на краю православного мира, идея прямой, непосредственной связи с Иерусалимским храмом, была одной из важнейших” (1996, 36). Against this cultural-historical background it seems most plausible to connect the коуни рони inscriptions—originally referring to the Babylonian Siege of Jerusalem—with the dramatic events in Novgorod and, more particularly, in the St. Sophia Cathedral in the year 1066.

Of special interest for our discussion is the fact that, according to the Bible, Nebuchadnezzar II, after having captured Jerusalem for the first time (in 597 BC), took away the sacred vessels from Solomon’s Temple and put them in his temple in Babylon. This event is recalled several times in the Bible (see, e.g., 2 Kings 24:13; 2 Chron. 36:7; Daniel 1:2; Jeremiah 28:3). Vseslav acted in a very similar way in Novgorod and may have been perceived by the local clergymen as the ‘new Nebuchadnezzar’, which makes the reference to the Fall of Jerusalem highly relevant.

At first sight it seems rather enigmatic how this allusion could have taken the shape of a citation from the Book of Lamentations of the Hebrew Bible on the walls of the St. Sophia Cathedral. In the Eastern Christian Church the text of Lamentations did not have any liturgical function; the earliest attestations of the Slavonic version only appear in the fifteenth century (cf., e.g., Thomson 1998, 850–865). However, in medieval Jewish communities, at least from the seventh century onwards, reading from the Book of Lamentations was an essential part of synagogal worship on Tisha b’Ab, a major fast day commemorating the destruction of the Temple. Descriptions of the pattern for this midnight service stress the specific liturgical role of verse 2:19 (“Arise, cry out in the night […]”) (see Brady 1999, 114f., 246; 2003, 110–114). Familiarity with this tradition thus seems to be the most plausible factor underlying the emergence of dozens of коуни рони inscriptions in the St. Sophia Cathedral.Footnote 18

How was the knowledge of this tradition transmitted in Novgorod? From whom did the коуни рони writers—at least the first one—learn the expression? Was it a baptized Jew? Or a Jewish merchant dwelling in Novgorod at the time that Vseslav seized the city? It is beyond the scope of this article to discuss the huge body of literature on the small and scattered pieces of evidence of the Jewish community in Rus’ and the use and knowledge of Hebrew in that area. As S. Franklin concludes, after a brief rehearsal of the limited amount of data:

The intensity of Jewish settlement is unknown, and there may be no direct links between the various identifiable groups—Khazars, Radhanites, and later traders—but the evidence for a Jewish presence over several centuries, whether sporadically or continuously, is adequate. Hebrew was, presumably, a feature of the graphic environment for those who cared to look.

But did anybody look? and if so, how intently? The Rus came into contact with Jews and Hebrew from two quite separate sources: the real, and the abstract. The real Jews bringing real Hebrew were the Khazars and Radhanites and other merchants. The abstract Jews and the abstract Hebrew are images from Church Slavonic Christian writings: Hebrew as a sacred language, one of the three inscribed on Christ’s Cross; Jews in sacred history. (Franklin 2002, 118)Footnote 19

In the case of the коуни рони inscriptions, both the contact with the Jews and the contact with Hebrew seem to be involved. On the one hand, the association of the seizure of Novgorod by Vseslav in 1066 with the Fall of Jerusalem, which, according to our hypothesis, was the reason for writing the biblical expression on the walls of the St. Sophia Cathedral, may serve as an example of how ‘abstract’ knowledge of Jewish history participated in the formation of an early Rus’ identity. On the other hand, the inscriptions must be regarded as the oldest tangible proof of direct contacts with Jews and Hebrew, as early as the mid-eleventh century.

As far as we know, there is only one other example from pre-Mongol times which seems to come close to the коуни рони case, i.e. a Hebrew word in Cyrillic which may point to a hearer-based orientation of the writer: the word машиаакъ for the Jewish messiah. It is attested twice in an early-thirteenth century manuscript from Novgorod (GPB Q.p.I.18; see the edition by Wątróbska 1986, 194), which includes a dialogue between a Christian and a Jew called “Addresses to a Jew on the Incarnation of the Son of God”.Footnote 20 The use of the consonant letter ш in машиаакъ is difficult to explain as part of a transcription from Greek (which had no /š/) . It rather suggests a direct rendering of the corresponding sound in the underlying Hebrew word . In addition, the /a/ in the first syllable is distinctively Hebrew, the /e/ in Greek clearly resulting from Aramaic (Beyer 1984, 115f.). Pereswetoff-Morath rightfully concludes: “This is proof enough that the translator or a later editor / copyist knew (some) Hebrew or had had (some) live contact with Jews conversant in Hebrew” (2002, 140). Alekseev already pointed out that the attestation машиаакъ must be regarded as a “факт прямого контакта между еврейской и славянской средою в древнюю эпоху” (1993a, 71; see also 1993b, 241). The коуни рони inscriptions, which can be dated some 150 years earlier than the instance of машиаакъ, provide new evidence for Alekseev’s statement.

Finally, it should be noted that quite a few instances of коуни рони have been crossed out. This gives rise to the assumption that a direct citation from the Hebrew Bible might have been considered by others to be inappropriate in an Orthodox church.Footnote 21 It may be no coincidence that nearby the коуни рони inscriptions on the northern wall of the Diaconicon another inscription is carved which reads: “O Lord, have mercy on the Christians and curse the heretics”.Footnote 22

3.2 Парехъ мари

Like in the case of коуни рони, the Cyrillic hearer-oriented spelling of парехъ мари also presupposes that the scribe used a Slavic idiom as his pragmatically dominant language; a Novgorodian citizen or at least someone who was integrated in local society. In fact, the Syriac provenance of парехъ мари can be directly related to another inscription in the St. Sophia Cathedral, namely , consisting of the monogram Ефремъ and the word сиринъ.Footnote 23 The handwriting of this inscription is most probably the same as the two instances of парехъ мари (see Fig. 5).

Fig. 5
figure 8

Paleographic comparison between

Full size image

Since all other monograms discovered so far in the Cathedral are autographs of the bearers of inscribed names,Footnote 24 it is unlikely that Ефремъ сиринъ refers to the famous theologian and liturgical poet St. Ephrem the Syrian. It is more plausible to assume that Efrem was a local citizen, probably a clergyman, who carried the epithet ‘the Syrian’, either as a nickname or because he was a Syrian by descent.Footnote 25 As for the first possibility, a parallel can be drawn with the Novgorodian church painter Olisej Grečin, who was of local provenance; he was called ‘the Greek’ merely because of his biographical ties with Byzantium (see Gippius 2005). The second possibility is also conceivable since the presence of Syrians in pre-Mongol Rus’ is well established; the most famous Syrian we know was Petr, the physician of Prince Svjatoslav Davydovič of Černigov, whose story is told in the Kievan Cave Patericon (see Ol’ševskaja and Travnikov 1999, 29ff., 132–134). This does not necessarily entail that Efrem was a native speaker of (Classical) Syriac, since various languages (including Aramaic vernaculars, Greek, and, after the spread of Islam, also Arabic) were spoken in the area that was frequently, though often imprecisely, referred to as ‘Syria’ in Antiquity and the Middle Ages, cf. Nöldeke (1871).