KARMIRVANK - A Little Red Church Near Ani

About 3.5 kilometres east-northeast of Ani, at the lower end of a small dry gorge, and some 40 metres from the bank of the Arpa / Akhurian River, is a small and partially-ruined church. The ruins of ancillary buildings suggest it was part of a small monastery, and this confirmed by an inscription on its east facade.

Its architectural style, together with information gained from the inscriptions on its walls, suggests that the church was constructed in the early to mid-11th century, and that the monastery probably continued to function during the eras of Byzantine and Shaddadid rule at Ani though in a diminished state, flourished again during the era of the Zakarids, was abandoned for a period sometime after the Mongol conquest of Ani in 1236, only to be restored to activity in 1271. Its final date of abandonment is unknown.

Karmirvank may be a surviving example of the sort of churches that once existed at Ani in large enough numbers for the city to attain its "1001 churches" epithet, but which exist now as only piles of rubble. They were small and inexpensively built, but intended to be as impressive as their size and limited construction budget allowed so often had domes and followed the designs of larger churches. Their patrons probably were minor local nobility or wealthy merchants.

Karmirvank in Past Accounts

This church is omitted from almost all descriptions of Ani and its original name has been lost. Karapet Basmadjian visited it in 1903 and called it Karmir Vank ("Red Monastery") in his 1911 account of that visit [note 1]. In his 1920s Revue de l'Orient chrétien articles on Armenian inscriptions at Ani [note 2], the wording used by Basmadjian, l'eglise dit 'Karmir Vank', suggests that Karmir Vank was a name the local population called the site rather than a name coined by Basmadjian. Toros Toramanian also called it Karmir Vank in a diary entry from 1909. A photograph of the church exists taken by Toramanian either in 1909 or 1913. Locals in Turkey now call the church Çoban Kilisesi ("Shepherd's Church"), though it is uncertain how long-established this name is (it may be a recent transference from the now destroyed Church of the Shepherd that was located just outside the walls of Ani and close to the track leading to Karmirvank).

J.-M. Thierry in his 1971 article [note 3] on the Karmir Vank that is also known as Kizilkilise mentioned the existence of this other Karmir Vank, though without visiting it. The first detailed study of the monument was the article titled Karmirvank', le Couvent Rouge by Ani Totoyan-Baladian, published in Revue des Études Arméniennes, volume 30, 2005-2007 [note 4], based on a visit made by Baladian to Karmirvank in 2004. Karmirvank is mentioned briefly in the 2011 edition of Araştırma Sonuçları Toplantısı [note 5] and it is also mentioned in the World Monuments Fund's 2014 report Ani in Context [note 6], in each case with a single accompanying photograph. The photographs used for this webpage were taken in 2015. Comparing them against the photographs found in the above three articles reveals that a large amount of damage has occurred to the church in recent years. The monument is currently at risk of complete collapse, and the nature of that recent damage is very suggestive of a deliberate intent to hasten such a collapse [note 7].

Access to the Site

Access to Karmirvank is currently not forbidden in practice, but any visit should be made as discreetly as possible because it is so close to Turkish-Armenian border.

To walk to Karmirvank from the city walls of Ani takes just under an hour. Use the road that runs past the "Ani restaurant" (built on the site of the former army base at Ocakli village) and continues as a dirt track through cultivated fields. The easiest way into the small gorge containing the monastery is by using an old trail that zigzags its way down the slope that overlooks the site to the southwest - this is probably the original access route to the monastery. There is a second rougher trail a little higher up the same gorge on its north side - this probably provided access for those approaching the monastery from the north, from the direction of Horomos monastery. It is also possible to reach Karmirvank from Ani by walking along the top edge of the gorge above the river, starting from the abandoned watchtower oposite the Tigran Honents church. However, this route may attract unwanted attention from local guards at Ani or Russian border guards on the Armenian side of the border at Kharkov. It is not possible to get there by walking along the riverbank from below Ani (an earlier version of this page erroniously that it was).

The Exterior of the Church

The church is small in size (just 8.2 metres by 6.9 metres externally) but has pronounced vertical proportions (the internal height of the dome above the original floor level is about 12 metres). It has a rectangular external plan, sits on a three-stepped foundation, and has two v-shaped niches cut into the east facade that articulate the location of the apse within. The dome has a conical roof, and a drum that is circular inside and outside and pierced by two windows along its east-west axis.
plan of the church
The west facade has the sole entrance to the church, though almost all trace of this entrance has gone. The photograph by Toramanian (see below) shows a rectangular opening under a large tympanum that is framed by a hooded archivolt moulding. The opening is flanked on each side by a pair of engaged columns with globular capitals – the top of the capitals meet the base of the archivolt. The tympanum, a large monolithic block of dark stone, is now lying face-up on the ground. There is carved decoration on it that has become too worn to be fully made out but consists of a large roundel with a wide circular frame of chain interlacing and a central motif that might be a cross.

The church as photographed by Toros Toramanian in 1913

There was originally a window above the entrance, matching the surviving windows in the north, south and east facades. All trace of it has now vanished but the top half of it on the inside survived until at least 2013. There is a small khachkar on the left side of the gable of the west facade. The photograph of Karmirvank in the 2014 WMF report, which was taken in 2013, shows a similar but slightly smaller khachkar at the top right side of the same facade. This khachkar had vanished by 2015. A gap in the facing stones indicates that a large rectangular masonry block was once located vertically under the apex of the gable of the west facade. Visible in the Toramanian photograph, it was a third khachkar, considerably larger than the other two. The surviving khachkar has an arched frame with the arch section containing a scroll of grape vine-like leaves and the sides a fretwork interlace. The now missing slightly smaller khachkar had a very similar design, but with the decoration on the arched frame being a continuous fretwork of interlocking curves. The three khachkars do not appear to be reused gravestones but carved when the church was constructed.

The rest of the exterior is rather plain. Except for the three khachkars and the portal ensemble, decorative elements are limited to just a simple rectangular beaded moulding around the apse window, plain hooded mouldings above all the remaining windows, and ribbing on the roof tiles. Despite that, the building's vertical proportions, the strong reddish-orange colour of the tufa stone used in its construction, and the beauty of the location, give the church an eye-catching appearance.

The south and west facades have sustained serious structural damage – the entire lower half of the south facade is gone, as is most of the west facade with the exception of its north corner. The northeast corner of the church is also badly damaged, the roof has lost most of its stone roofing slabs, and selected facing stones have been carefully removed from various parts of the exterior.

The Interior of the Church

The interior has a "domed hall" design with open western corners (alternatively called an "open partitioned cross"). However, the small size of this church reaches the limits of what is structurally sustainable (there is very little concrete core between the facing stones) or aesthetically pleasing using this design form.
sectional drawing of the church
The western corners are open onto the west bay (hence the term opened partitioned cross), resulting in two rectangular spaces that seem too small to have had a real function and that are visually rather ungainly (they are overly tall and narrow, with off-centre and irregularly-shaped vaults). However, these niche-like spaces do appear to have originally had particular functions that seem to have required their delineation from the rest of the interior space: there are two slots cut into the masonry on the outer edge of each, one near the top, one near the floor, and there is additionally a groove cut into the vault of the southern one at its outer edge. It is suggestive of there once being of a timber screen of some sort that closed off each of these rectangular spaces.

The eastern corners of the church contain small chambers. They are probably small chapels since they have an apse-like eastern end pierced by a narrow window, and have a niche in their north wall. The chambers are accessed from doors through the east sides of the north and south bays.

The apse is semicircular, has a niche in its north side, and has a raised bema that fills the entire semicircle. The outside edge of the bema floor is rough and untidy, indicating missing paving blocks and that the bema formerly extended out a little further (probably as far as a line drawn between the outer edge of the engaged columns, maybe as far as one drawn between the east walls of the north and south bays). Baladian may have misinterpreted the foundation of those missing paving blocks as belonging to the floor level of the central bay of the church because she writes "le béma présente une élévation d'à peine une marche par rapport au niveau de la sale" ("the bema has an elevation of only a step above the room level"). The original floor surface of the church is hidden under rubble, but its level can be estimated by looking at the lintels of the doorways to the corner chambers. Their undersides are 44cm above the level of the bema floor. If the minimum usable height of these doorways is perhaps 1.2 metres, then the bema must have been raised at least 76cm above the floor level of the rest of the church, or the height of three steps.

The dome is supported by a cylindrical drum, with the transition from the square central bay of the church obtained using pendentives.

The interior surfaces are all bare, with smooth-finished and well-jointed stonework. There is no evidence of plastered surfaces. Decoration is minimal and consists almost entirely of architectural elements: at each corner of the bay under the dome is an engaged column topped by a plain cushion capital and there are impost mouldings at the springing line of the vaults. The only other sculptural element is a khachkar set into the wall high above the entrance to the northeast corner chapel. There is not a matching khachkar above the southeast corner chapel entrance. It is a reused khachkar: its left side has been cut down to fit the available space. It also shows traces of having being whitewashed. Its architectonic form - the cross is under a semicircular arched frame and the arch has a scroll of alternating leafs and rests on pilasters – is typical of khachkars dating from the end of the 10th century.

Amongst the rubble on the floor is a fragment of a khachkar and a stone slab with one half of a rectangular slot hole: it is tempting to think they were once part of the same monument, and that it was perhaps set up inside the church.

The Inscriptions

There is a short two line inscription on the south facade. It was recorded and translated by Basmadjian (as his inscription 27):

In 614, I Hairapet, I restored.

Baladian considers that the end of the first line is missing (that the inscription continued onto a facing stone which is now missing), truncating the word hayrapetowt'ea to hayrapet, and thus she renders the inscription slightly differently:

In the year 614 (has been) restored (under the) Ponti[ficate] ...

Baladian's theory is reasonable – the missing facing stone was already missing on Toramanian's 1913 photograph, so was probably missing when Basmadjian recorded the inscription in 1903. The year 614 of the Armenia era corresponds to the year 1165. This date lies within a short period (1161 to 1165) during which Ani was temporarily under Georgian control before reverting back to Shaddadid rule.

Basmadjian recorded a further inscription on the south facade (his inscription 33) that covered three lines. He translated it as:

By the grace of God, I, Zacharia, amirspasalar, I gave to this holy monastery two deniers of karmir-marg.

This inscription no longer survives, but a block of facing stone now lying below the south facade carries two lines of lettering from it (see photograph 19). In 2004 Baladian also noted this stone, as well as another block bearing an additional fragment, and translated the inscription as:

By the grace of God, I Zak'aria, amirspasalar, I gave to this holy monastery two thirds of the revenue of the meadow called Karmir.

Amirspasalar ("commander in chief") Zakare, with his brother Ivane, captured Ani on behalf of the Georgian Queen Tamara in 1199 - so the inscription can be no earlier than that, and Zakare is thought to have died by 1215.

Basmadjian also mentions the existence of two other short and, in his time, fragmentary inscriptions on the west facade (his inscriptions 141 and 142). They no longer survive. He also noted a fallen facing stone bearing a sundial.

The 14-line inscription carved on the east facade

The most important inscription on the church is 14-lines long and carved under the gable of the east facade. It was recorded by Basmadjian in 1903 and, as his inscription 70, published and translated into French. Baladian's translation (also into French) has some improvements, mostly clarifying some confusing words [note 8]. The inscription is dated, with the date section separate and placed inside a circle:

In our era 720.

This corresponds to the year 1271. Below it, the inscription continues as:

By the grace of God, I, Uk'an K'arimadin, and my brother, Papk'an Vaxradin son of Sargis grandson of Hogi, and my wife Dapta Khat'un, we purchased this convent in its entirety with its lands and waters, and we relieved its abandonment, we have enriched it with objects and books and we gave to this church one sixth of the income from a khanap we had bought ...

A khanap was a sort of mercantile building. Baladian considers that it could be a store or workshop or depot, or a grouping of chambers with these functions housed in a single building. The inscription goes on to detail the location of the khanap in relation to adjoining structures (though not stated, it presumably was at Ani), and then lists further donations made to the monastery, including a vineyard at Bagaran, and two-thirds of the income from a location called Khach'ikhor and its mills.

The inscription then asks that:

In return, father Sargis and the other brothers will say Mass every Saturday before the three altars for our father Sargis, and for Sit'i, and for Merch, and four days for Sargis ark'aun.

It concludes by stating:

Those who oppose these donations, they are judged by Christ. Amen. Mxit'arich the engraver.

The donors mentioned in the Karmirvank inscription were important enough persons to be recorded in inscriptions at a number of locations in Armenia. Five other known inscriptions, plus a manuscript colophon, are analysed by Baladian in her article on Karmirvank. Through them, it is revealed that of the individuals named at the end of the Karmirvank inscription, Merch was a third brother, deceased by 1271, and Sargis ark'aun was probably the father of Dapta Khat'un. The term ark'aun appears during the period of the Mongols and is probably a title, designating a Christian having an important function in the Mongolian administration. Baladian examines and then dismisses earlier theories that ark'aun designated a Chalcedonian Armenian, or an Armenian originating from the Byzantine Empire. The identity of the individual named Sit'i remains unknown, but it is a female name.

At Ani, in 1912, excavations in the ruins of a church (M-O site 101) uncovered an undated inscription telling of the construction of a bell tower by Uk'an K'arimadin, Papk'an Vaxradin, and Datpa Khat'un. Because of the inscription Marr named it the Church of K'arimadin. The name K'arimadin also appears in an inscription dated 1269 carved on the narthex of the Holy Apostles church.

The Ancillary Buildings

There is a horizontal double line of shallow holes hacked into the west facade of the church. This may indicate that a crudely-built jamatun was located in front of the church, or just that a simple structure once occupied the corner between the church and the adjoining precinct wall.

There are several large stone slabs below the east side of the church – they probably mark the location of graves. In 2016, holes newly dug here by treasure hunters contained human bones.

Southwest of the church are the ruins of a row of chambers, the outer wall of which runs in a north-south direction and forms a continuous line about 36 metres long. It seems to be the product of several building periods: the northern part has interior and exterior faces of carefully cut and jointed masonry; the southern parts are faced with roughly finished blocks. The chambers at the southern end may have been on two levels.

plan of the monastery

At each end the wall turns 90 degrees to continue in an easterly direction. On the north side this wall probably ended at the northwest corner of the church (a ruinous wall is shown in Toramanian's photograph). The end point on the south side is unclear. The southwest corner is curved on the exterior, and this feature, together with the height and solidity of the wall at that location, gives the visual impression of a defensive wall. However, there are no certain traces of a wall along the eastern side of the monastery, so it would be an unproven supposition to state that the monastery had an actual enclosure wall.

Other Structures near The Monastery

In a cliff face on the heights overlooking the north side of the church is a small rectangular opening into a rock-cut chamber (in photograph 29 it is visible to the left of and slightly above the conical roof of the church). The rock face below the opening has fallen away, making the chamber inaccessible. Access to it appears to have been via an external timber staircase (there is a diagonal line of holes cut into the cliff face). There are more rock-cut chambers higher up the gorge on its south side. Some of these chambers show traces of having had stone-built frontages.

On the opposite side of the Akhurian, in Armenia, there is a stone embankment constructed of fieldstones that carries an old roadway over a section of the river bounded by a cliff (see photograph 32; the roadway it supports is visible in photograph 2). I am uncertain of the age of this structure; it might be medieval, being similar in appearance to the embankment supporting the roadway to the bridge at Ani. The track seems to end about 200 metres to the northeast, close to the river bank. This may mark the location of a ford across the river - though no immediate continuation of the track is visible on the other side, unless Karmirvank itself lies along its route. Its route in the opposite direction can be traced as far as Kharkov. Kharkov is a Soviet-era army base, still in use, but there are the ruins of an old village also on the site.

There is a cemetery site located half way up the slope that overlooks the monastery to the south, a little to the east of the zigzagging track from the plateau to the monastery. No khachkars or inscribed stones are visible in the cemetery, and a number of the graves have been recently dug up by treasure hunters.

At the top of the track that accesses the gorge from the north, at the edge of the plateau, are the overgrown ruins of two clusters of rectangular chambers built of fieldstones. They were perhaps associated with the monastery.


1. Basmadjian, K. J.; Anii drac'i Karmir vank'a, in the Armenian periodical Vostan, issue 1, p233-235, Paris 1911, (n/a to me, but cited in Baladian).
2. Basmadjian, K. J.; Les inscriptions arméniennes d'Ani de Bagnair et de Marmachén, Revue de l'Orient chrétien, volume 22, 1920-1921, pages 337-362 (inscriptions 1 to 15); volume 23, 1922-1923, pages 47-81 & 314-344 (inscriptions 16 to 74); volume 24, 1924, pages 356-371 (inscriptions 75 to 88); volume 25, 1925-1926, pages 156-186 & 372-377 (inscriptions 89 to 167); volume 26, 1927-1928, pages 357-380 (inscriptions 168 to 191); volume 27, 1929-1930, pages 225-287 (inscriptions 192 to 223 and index). There are no photographs of the inscriptions in the articles. The inscriptions that are from Karmirvank are numbers 27, 33, 70, 141, 142.
3. Thierry, Nicole & Jean-Michel; "Karmir Vank (Kizil Kilise)" in A propos des quelques monuments chretiens du vilayet de Kars II, Revue des Études Arméniennes, volume 8 (1971), pages 192-196.
4. Totoyan-Baladian, Ani; Karmirvank, le Couvent Rouge, Revue des Études Arméniennes, volume 30 (2005-2007), pages 301-337.
5. Sağır, Güner; "Çoban Kilise" in Kars İli ve Çevresinde Yer Alan Ortaçağ Ermeni Kiliseleri (Ani Örenyeri Hariç) Yüzey Araştırması, Araştırma Sonuçları Toplantısı, volume 29 (2011) part 2, pages 22-23. This is a report detailing a series of visits made by Turkish investigators to Armenian sites around Kars in 2012. In the article, the stated aim of the visits is to correct "errors" created by "unauthorised research" by "foreign scientists". This seems to be an attempt by Turkey to gain a greater academic ownership of (and thus greater control over) Armenian monuments in Turkey by producing some Turkish-language academic literature on them. Sağır's article on the Taylar church published in Turkish Studies, volume 9/10 Fall 2014, pages 929-940, and his forthcoming paper on Kizilkilise, are part of the same project.
6. Report titled Ani in Context Workshop, September 28 - October 5, 2013, Kars, Turkey, published by World Monuments Fund, 2014. Unsurprisingly, given the destruction World Monuments Fund had already inflicted at Ani, the report displays blatant disregard for modern conservation methodologies and philosophies. For Karmirvank, it recommends "restoration" rather than conservation. The nearby Taylar church may be being set up for an Aghtamar-style show restoration for political reasons because the report recommends its "complete restoration". The Sağır article on Taylar mentioned above may be a production made in connection with this future aim.
7. For example, there is suspicious looking chipping along the edges of many of the facing stones that adjoin stones that have vanished since 2004, as if metal tools had been used to pry off the missing blocks. No such chipping exists along the edges of stones that adjoin older gaps in the facing.
8. Here is Basmadjian's translation: « En 720, par la grâce de Dieu, moi, Ouqan Qarimadin, et mon frère, Papqan Yakhradin, fils de Sarguis. petit-fils de Hogui, et ma bru, Daptakhatoun, nous avons acheté (le terrain de) ce couvent avec toutes ses dépendances, avec les terres et les eaux; nous l'avons construit sur la friche, nous l'avons enrichi d'objets du culte et de livres, et nous avons fait don à cette église de notre propriété, achetée de nos deniers : le denier du marché, qui était la moitié (?), le haut de la propriété du maire, le bas et le coté sud, avec les boutiques, et aussi une vigne à Bagran (avec) le vigneron, le lien d'Aqsout, les quatre deniers de Khatchikhor et des moulins. En récompense, chaque samedi le père Sarguis et mes autres frères diront des messes, devant les trois autels, pour notre père Sarguis; et si ce n'est pas pour notre (père), ils diront quatre jours de messe pour l'archonte Sarguis. Ceux qui s'opposent à ces dons seront jugés par le Christ. Amen. — Mkhitharitcli l'écrivain ».
Here is Baladian's translation: « En notre ère 720. Par la grâce de Dieu, moi, Uk'an K'arimadin, et mon frère, Papk'an Vaxradin, fils de Sargis, petit-fils de Hogi et mon épouse, Daptay Xat'un, nous avons acheté ce couvent dans son intégralité avec sol et eau; et nous (l')avons (re)levé de son abandon, nous (l')avons enrichi d'objets et de livres (liturgiques), et nous avons offert à cette église un sixième des revenus du xanap que nous avions acheté, et qui était (situé) entre (celui) de Sahap en haut et en bas et avec ses échoppes (situées) au sud; et une vigne à Bagran, son vigneron, l'élevage d'hermines, les deux tiers des revenus de Xach’ixor et de ses moulins. En échange, le père Sargis et les autres frères diront la messe chaque samedi, devant les trois autels, pour notre père Sargis et pour Sit'i et pour Merch et quatre jours pour Sargis ark'aun. Ceux qui s'opposent à ces dons qu'ils soient jugés par le Christ. Amen. Mxit'arich le graveur ».


3rd October 2015 - This webpage is published.
26th March 2018 - Survey drawings of the church improved, site plan added.