• 1 This article is based on a paper presented at the international conference organised by Iwona Gajd (...)
• 2 The history of the mention of this site can be found through these references: Mordini (1944-1945: (...)
• 3 Bermudes adds that Tēwodros (elder son of king Dāwit, 1379/80-1412) would have given a district na (...)

1Ever since the early 1940s, Māryām Nāzrēt has regularly been mentioned in the secondary literature as an important archaeological site presenting two levels of remains.2 The older ones seem to suggest Aksumite architecture, while the more recent ones—quite misunderstood so far—present Islamic features (Henze 2007: 1159). Antonio Mordini (1944-1945: 150 n. 2), who visited the place in 1940 and 1942, reports conducting a survey involving digging in the surroundings of the church, from which he extracted Aksumite coins. Both Mordini and Henze also note the association of the site with the toponym of ʿAddi Abun (in reference to the Ethiopian title of the metropolitan bishop of Ethiopia, the Abuna). This toponym is not just a modern reference since it was mentioned in the sixteenth century by João Bermudes (2010: 81)3 while he traveled through the region. Further, a list of metropolitans bears witness to the burial in Nāzrēt of Egyptian bishops appointed to Ethiopia as heads of the Church (Ms. BNF Éthiopien 160, fol. 90r; Zotenberg 1877: 263). Lastly, mention is made in the colophon of the Kebra Nagaśt (a text used to reinforce the Ethiopian royal ideology) of the translation of that work from Coptic into Arabic, again at Nāzrēt (Bezold 1909 [1905]: 172, 138).

• 4 All our thanks go to Hailay Teklay, of the Tigrai Culture and Tourism Bureau in Mekelle, who infor (...)

2Mekelle University undertook a few soundings on the site of Nāzrēt in 2014, following the villagers’ intention to construct a new church in the same compound.4 The villagers had already dug a large pit, from which eight exploratory trenches were laid out. These tests brought to light a building, on one plane only, oriented northwest-southeast, at least ten metres wide, and surrounded by tombs. The excavations did not disturb assumed Aksumite levels lying some distance away.

• 5 Communication of M. Gaudiello at the 23th congress of the Society of Africanist Archaeologists, To (...)
• 6 On the site of Kwiḥa and its stelae, see Conti Rossini (1938: 402), Pansera (1945: 3-6), Schneider(...)

4Given these objectives, the site of Nāzrēt is an ideal candidate to work with. Situated south of Mekelle, it is found in a region that lies in the southern part of the kingdom of Aksum, according to what is presently known about its territory. This area may, therefore, appear as a front line of Christianisation, where it is interesting to observe at what point in time the first churches appear and transformations of funerary practices occur. The research activities conducted over the past years by Paul Yule and his team at Mifsas Baḥri (Bāḥri)—near Lake Ašangē, farther south still—confirm how relevant it is to pay attention to this region in order to understand the timeframe that stretches from the seventh to the thirteenth centuries.5 Moreover, it is in this same region that the site of K^wiḥa is found.6 Islamic funerary stelae as well as a fragment of the foundation inscription of a mosque discovered there are witness to the presence of a Muslim community in these parts between the tenth and twelfth centuries.

• 7 RIEth 219 pl. 156 (Bernand, Drewes & Schneider 1991). The pictures used by Bernand, Drewes and Sch (...)

10A stone basin displaying an inscription of a non-vocalized Ge’ez also seems to be related to this ancient edifice (fig. 4). Abraham Drewes and Roger Schneider edited this inscription in 1991, stating it was an “écriture ancienne”. The inscription runs on three of the four sides of the basin and is engraved in two lines,7 but the second line of face 2 was destroyed whenever the basin was damaged and face 4 is unreadable.

• 9 It too is inscribed with an unvocalised Ge’ez inscription (ዘለወ[ደ]).
• 10 He tells of two other basins discovered during the excavations at the Dungur palace. If this conce (...)
• 11 “Northern entrance to the Monastery Church with stone basin next to it; photo looking south from R (...)

11As Francis Anfray (1965: 7, pl. ivc)9 pointed out, other basins of the same type—although less ornate—have been retrieved: one near the rock-church of Abbā Maṭṭāʿ, south of Aksum, another in the church of Endā Māryām of Ham, and a third one in the church of Abbā Ṗanṭalēwon (Anfray 1970: 39 n. 6).10 The association between churches and basins would lead one to think that such objects were designed and used for liturgical purposes. It is unlikely that they were used for baptisms because the Ge’ez term of maqala/da (“basin”) hardly evokes such connotations (one would expect something like meṭmāq in this context). They may have been meant for the ablutions performed before entering a church. Such basins have also been observed near the northern door of later Nubian monastic churches (like at the monastery of Kom h in Old Dongola and Al-Ghazali), whereby the monks used to enter the church when coming from their dormitories.11 It could also be a water storage unit for various liturgical purposes such as washing ones’ hands, mixing water with wine, consuming water after receiving Holy Communion, or performing the ablutions of the sacred vessels at the end of the mass.

• 12 A part of the coin collection from the excavations of A. Mordini is now in the care of the Bibliot (...)
• 13 Idea already developed by Godet (2003: 50-53, 400-440). Munro-Hay and Juel-Jensen (1995: 51, 264-2 (...)

12The excavations undertaken near the building by Antonio Mordini (1944-1945: 150 n. 2)12 in the 1940s revealed Aksumite coins of Kings Hataz and Gersem, but no information was published on the exact whereabouts or the results obtained. Recent numismatic research places these kings among the last who minted money, likely in the 7^th century (Piovanelli 2014: 339). Wolfgang Hahn (2010: 10) proposed that King Hataz (Hethasas) was the last monarch who minted money, placing Gersem as his predecessor, before Iyoēl. Instead, Pierluigi Piovanelli considered Hataz to be a predecessor of Gersem, with Israēl having reigned in between, and Armaḥ being the last of the kings who minted currency.13 This is not the place to delve deeper into these chronological debates, for no matter what thesis is adopted, all point towards the end of the Aksumite kingdom between the sixth and seventh centuries.

• 14 A first description of the church was published in Fritsch (2016: 56-65). We are grateful to N. Wa (...)

18The orientation of the structure, including the existence and type of niches which reinforce this orientation as well as the west doors of the side rooms in the axis of the niche contrasting with the pristine complete opening of the central area, in addition to the symmetry in the distribution of the cupolas, the connections between the rooms, the murals (see below), all combine to demonstrate that this structure was created as the five-fold sanctuary of an exceptionally important Coptic church, with the central area being the chief sanctuary14 (fig. 6 and 7).

• 15 Insofar as our observations were almost made in the dark, it is likely that some traces of paint e (...)
• 16 It should be noticed that the cupolas of similar Egyptian buildings like St-Macarius, are not pain (...)

20In the eastern sanctuaries, traces of paint15 are still visible on certain preserved plastered areas. In the main sanctuary, a horizontal red stroke can be observed high up on the cupola.16 In the southeastern sanctuary, superimpositions that imply a diachrony can be seen on either side of the doorway that leads to the next sanctuary, and in the lower area of the wall, a thick earth mortar covered coating displays traces of yellow paint. When standing inside the sanctuary—to the right side of the door that opens northwards—one can observe beneath areas of loss, a fine, and therefore earlier coating, covered with a background of red paint that appears to form a band. To the right of the doorway that opens westwards, in the lower area of the wall, once again, strokes of vertical and horizontal yellow paint were applied on a clear sublayer. These observations bear witness to, at the very least, two phases of ornamental treatment of the walls with flat tints of colour. The strokes, simple as they may look, could in fact betray another project, perhaps figurative. In the current state of preservation, the strokes appear as separation lines; parts of bandeaux which separate the registers in Coptic or Ethiopian paintings. In addition, it is intriguing that it is in the lower sections of the walls that these flat tints of colour can be observed. This may not be without meaning—for one might infer that the coloured register would have been more appropriate to the lower parts of the walls, with figurative decoration only applied at body height and above.

• 17 Another ʿAddi Abun exists in Tigray, north of Adwa, also referred to in traditions related to the (...)
• 18 See the famous sites mentioned by P. Miquel describing the surroundings of Lālibalā and which, in (...)

21The toponym of ʿAddi Abun came to the attention of Antonio Mordini and his followers in connection with the site of Nāzrēt. The place would have received this name because it was supposedly the grave of several metropolitans of Ethiopia.17 This toponym is still in use at Māryām Nāzrēt. A brief inquiry conducted while visiting during October 2014 points to other toponyms related with sites or places of scenic interest that echo the name of ʿAddi Abun. Among those, the following are particularly remarkable: Māy Ṣaʿadā, which reportedly received that name because Whites were living there, and ʿAddi Gebṣi because Egyptians were dwelling there. We were also informed of a rock or cave-church named Māryām Westē, higher to the east in the mountain, described as being divided into three sections. None of these three sites were visited by us, but each time we asked people about them they stated that the toponyms were associated with stone-cut remains. This line of inquiry bears fleshing out, in order to determine whether we are dealing with actual archaeological sites or a rather classic case of historical interpretations of natural scenery (e.g. as with basalt organs once seen as cut stones and archaeological remains).18 Nonetheless, this same expedition allowed us to visit a fourth site, that of Golegotā.

• 19 The dating of the building is unsure but it could be traced back to the eleventh century (Grossman (...)
• 20 The fact that it was painted in 1232-1233 provides one with a terminus ante quem.

30The church of the Martyrs (Dayr eš-Šohada near Akhmim) could serve as a useful model for visualising the type of church Golegota may have been.19 Irrespective of scale, the church of the monastery of St Antony-the-Great near the Red Sea (fig. 9), after rearrangement in the twelfth century,20 also serves as a classic reference illustrating the structure of the churches of the Middle Ages. Mutatis mutandis, it presents a pattern fitting the church of Golegotā.

• 21 A partial translation of this text, without edition, was published by Sergew Hable Sellassie (1972 (...)

33A now well-known document, hitherto relatively neglected due to concerns of authenticity,21 confirms the role of an Egyptian in the construction of a church at Māryām Nāzrēt. This document is a note from a Coptic metropolitan, Mikāʾēl, dated to the twelfth century. It is found in the Gospels of the church of Mikāʾēl Ambā, a monastery lying at about 100 km to the north. These Gospels are a rather late manuscript assigned by its typically G^welḥ script to between the mid-seventeenth and the eighteenth century. This dating is confirmed by the layout of the frontispieces of each of the four Gospels in which a few introductory lines are distributed between the two columns preceding the main text, a practice mainly observed in the manuscripts of that time, even if the paragraph signs in the margins witness the survival of a practice extinct elsewhere (Uhlig 1988: 450, 564, 828). The document related to Mikāʾēl was entered later still, and by another hand. This sequence raised many doubts, leading many to believe the document a forgery (Nosnitsin 2007: 953). In addition, doubts were raised by the facts stated in the text, in which the metropolitan claims to have arrived in Ethiopia at the time of the little known King Anbasā Wedem, constructing many churches—among them Mikāʾēl Ambā—, ordaining 27,000 priests and baptising 50,000 persons.

• 28 C. Conti Rossini (1909-1910, II: 317), who edited that land grant, judged that the document could (...)
• 29 These royal lists cannot be trusted for, as É. Godet (2003: XXXII) put it so well, “les querelles (...)

38Mikāʾēl mentions an Ethiopian king, Anbasā Wedem, but we are not certain he was still in place at the time of the consecration of the church of Mikāʾēl Ambā in 1149/1150. The text only states that he was the sovereign of Ethiopia when Patriarch Macarius ordained Mikāʾēl metropolitan of Ethiopia between 1102 and 1129. It must be emphasized here that, while Mikāʾēl’s commission applied to all Ethiopia, his see is not mentioned. Among the land charters compiled in the church of Aksum and known as the Liber Aksumae, there is a donation made by King Anbasā Wedem for this church (Conti Rossini 1909-1910, II: 5-6).28 However, this document does not include any date in either of its two versions, or any list which would allow the reign of Anbasā Wedem to be dated. Antonella Brita has recently mentioned a manuscript found in a church of Takla Hāymānot in Tegrāy, dating back to the nineteenth or twentieth century and dedicated to the Life of the Justs of Ḥawzēn, a group of missionary monks associated with the second Christianisation of Ethiopia starting from the sixth century. As it happens, this text places the arrival of these missionary monks in the twentieth year of Anbasā Wedem’s reign, during the year of grace 910 (i.e. 1194), which in turn places the beginning of the reign of Anbasā Wedem in 1174 (Brita 2010: 4-5). Could this be the same king, despite Mikāʾēl’s claim to have anointed seven sovereigns during his ministry? Such a claim suggests great monarchial instability. Or did two kings bear the same name? Or was the one and only Anbasā Wedem removed from the throne at one stage and reinstated later? These questions must be left aside for the moment, while keeping in mind that even though King Anbasā Wedem left very few solid traces from which to deduce his reign, he did leave his mark on the kingdom. According to the lists of kings that were tampered with after the sixteenth century, he reigned just before Delnaʿād, regarded as the last Aksumite king before the fall of the dynasty at the end of the tenth century, an event associated with the intervention of Queen Gudit.29

• 30 Gregory of Nazianzus’ will was written in Greek in 381. The bishop bases himself on Roman law in o (...)

39The text written by Mikāʾēl may be related to a testament (Fritsch 2016: 59-60), even though that word is not employed in the text and certain legal dispositions are missing from it, as compared to the better known will of another bishop, Gregory of Nazianzus.30 Nevertheless, it seems that Bishop Mikāʾēl did in these few lines mean to state his last will, and in particular to release his slaves. In fact, a law promulgated by Constantine to the intention of Bishop Hosius about the manumission of slaves by the members of the clergy appears in the Theodosian code, compiled in 438. This law provides for the case of manumission in the testament of an ecclesiastic and states that any slaves must be freed at the time when the will is published shortly after the testator has died, without any need for witnesses or a law enforcing officer:

• 32 For easy reference, see Hanna (1996, chap. 32, § 9: 120).

40The manumission of slaves in ecclesia therefore included the possibility for a bishop to free his slaves without any procedure other than declaring so in his written will (Esders 2012: 58-59). How were these legal principles transmitted to the Coptic Church to which Bishop Mikāʾēl was accountable? Since the eleventh century, the Coptic Church has developed a particular jurisprudence under the form of patriarchal constitutions and, since the twelfth and thirteenth centuries, collections of nomocanons (Aoun 2002: 190-195). The constitutions of the patriarchs of Alexandria, which Mikāʾēl may have known, were elaborated by Cyril ii (1078-1092) and Gabriel ii ibn Turayk (1131-1145). Neither mentions a bishop’s right to bequeath. In contrast, all do mention the matter of slaves. In the eleventh century, Cyril ii forbade any lay people and presbyters to sell slaves to Muslims under the threat of excommunication if they did (Burmester 1936). Gabriel ii established that no bishop or monk could own a maid or a female-slave and that, if he happened to have inherited or owned one before becoming a monk or a bishop, he must either sell her or set her free (Burmester 1935). However, the way in which manumission was supposed to occur is not described. The least we can say regarding his female-slaves is that Bishop Mikāʾēl did not abide by the constitution set out by the patriarch who ruled the Church during his lifetime. For the rest, these texts provide us with no indication as to the source of the jurisprudence referred to by Mikāʾēl in releasing his slaves although it seems he may be referencing Roman law as included in the collections of nomocanons. Legislation concerning slaves was eventually included in Chapter 32 of the nomocanon compiled in Arabic around 1240 by Al-Ṣafī abū l-Faḍāʾil ibn al ʿAssāl, who collected laws both from apostolic writings and Byzantine codes.32 This collection of canons was then adopted as law by the Coptic Church despite its Chalcedonian character. Translated into Ge’ez almost four centuries after the facts which concern us here, this book brought to Ethiopia the same dispositions on manumission, thereby testifying to the continuity of tradition in this matter (Paulos Tzadua & Strauss 1968: 176).

• 33 One example of this is found in eighteenth-century historical documents (Getatchew Haile 1988: 11, (...)

41Considering the document at hand as a kind of testament, Bishop Mikāʾēl’s list of achievements definitely makes sense. He recalls the number of kings he anointed, the number of churches he consecrated, how many presbyters he ordained, how many monks and nuns he initiated and how many people he baptised. His intention is not to overstate the extent of his action but to give an account of it on the eve of his passing. He does not claim to have visited one thousand and nine churches but to have consecrated the altars for them (in fact an altar-tablet, the tābot) brought to him (Fritsch 2012: 447-453).33 Similarly, priests were not ordained one by one but in groups, as witnessed by a sixteenth-century Portuguese who attended the ordination of more than two thousand presbyters streaming from all over the kingdom towards the Egyptian metropolitan, in one sole ceremony. It seems that five or six thousand priests could usually be ordained within a few days (Beckingham & Huntingford 1961: 350). Read in this contextual light, the numbers presented by Metropolitan Mikāʾēl appear far more reasonable than at first glance.

• 34 For the mausoleums of the cemetery of Aswān, especially that of the Saba’ Banat (Creswell 1952, vo (...)

44Niches and squinches are all constructed with four-centered arches. Around them, characteristically, twin rows of voussoirs made of cut stones imitating bricks radiate outwards, with angles protruding to shape the eight niches placed at mid-height in each of the four side-sanctuaries they adorn, resting on horizontal timber boards. The wall surfaces were covered with coats of plaster supporting murals. Devised under the Abbasids (from 750 onwards), this type of wide arch, with a pointed low apex of Persian origin, the four-centered arch necessary for niches and squinches, was commonly exploited at the time of the Fatimid dynasty (909-1171) to enhance such famous buildings as Al-Azhar mosque, the Mosque of al-Ḥākim (r. 996–1013) (Creswell 1952, 1969; Blair & Bloom 1994; Blair 2010). The overall aspect calls to mind such mausoleums in the cemetery of Aswān.34

• 35 The first mention of the sanctuary of St Mark apparently goes back to 1133 (E.M. 849) in the story (...)

45In ecclesiastical terms, the mode of construction of Māryām Nāzrēt is strikingly similar to the sanctuary of St Mark in both timeframe and construction style. St Mark’s sanctuary was built north of the sanctuary of Benjamin in the main church of the monastery of St Macarius-the-Great (Wādī al-Naṭrūn, Egypt) between the years 1010-1050, when the relic of the evangelist’s head was entrusted to the monastery. Typically, it displays similar pointed arches and squinches upon which the cupola rests. In total, four sanctuaries were eventually constructed side by side, all linked by the choir (Leroy 1982: 11-12; 28ff.; plates 28-70)35 (fig. 10).

• 36 It is the head-secretary of the diwān al-inšāʾ, Ibn ʿAbd al-Zāhir, who mentioned the letter of Dāw (...)

49Stuart Munro-Hay (2004: 24-25) proposed to identify Abalfarag (Abū al-Faraǧ) with Abū al-Faraǧ Ibn al-ʿIbrī, also known as Bar Hebraeus, the famous bishop of the Syriac Church of Antioch, author of many works in both Syriac and Arabic. However, this identification implies that the translation from Coptic to Arabic was done outside Ethiopia because Bar Hebraeus is not known to have travelled to Ethiopia. The colophon is, however, very clear about this point: it was in Ethiopia that the translation was done. Munro-Hay’s identification is therefore far-fetched. Regarding the second translator—Abalʾez (Abū al-ʿIzz)—, Stuart Munro-Hay adopts a hypothesis put forward by Carlo Conti Rossini (1923: 506) who proposed viewing Abū al-ʿIzz as the father of a certain Dāwud ibn ʿIzz. The latter is known to us through a letter sent to Sultan Qalāwūn (1279-1290), in which he introduces himself as the “vizir” of King Yāgba Ṣeyon (1285-1294) and requests the sultan to appoint a new Coptic metropolitan in Ethiopia.36 This identification looks more likely than the former one, even though it would require a certain longevity on the part of the father and/or son in order to be able to be active in both 1225 and as well in 1285. The merit of this hypothesis is the light it sheds on a community of Egyptian Christians who were fluent in both Coptic and Arabic and who entered the service of Ethiopian kings during the thirteenth century. One of the centres of that community seems unquestionably to be the church of Māryām Nāzrēt and surroundings and it was probably derived from the entourage which accompanied every new Egyptian metropolitan into Ethiopia and then settled in the country. If Dāwud is indeed Abū al-ʿIzz’s son, it must be because the latter took wife in Ethiopia and had children, perpetuating a kind of Copto-Ethiopian élite in the service of the kings.

• 37 BNF Éthiopien 160, fol. 90r, Zotenberg (1877: 263), Guidi (1899: 2).

50To provide even more evidence, an early sixteenth-century manuscript37 lists a series of metropolitans, adding that six of them were buried in Nāzrēt. Only two metropolitans are mentioned by name: Sawirus and Mikāʾēl. The former was in charge during the middle of the eleventh century, which may imply that the Egyptian community already occupied the region before the construction of the church of Nāzrēt by Mikāʾēl. One may also wonder whether the names recalled are the correct ones as the list was compiled long after the facts. The presence among those names of Mikāʾēl, the founder of the church of Nāzrēt, nonetheless indicates a certain probability, although it is quite possible that several Mikāʾēls served as metropolitans of Ethiopia.