Home Page Discussion

Salon du Tapis d'Orient

The Salon du Tapis d'Orient is a moderated discussion group in the manner of the 19th century salon devoted to oriental rugs and textiles and all aspects of their appreciation. Please include your full name and e-mail address in your posting.

Manastir Kilims: In Search of a Trail

by Davut Mizrahi and Erhardt Stöbe

Editor's Note:  This essay is excerpted and somewhat modified from a book by the authors, of the same title.


Kilims that were once produced in the European regions of the great Ottoman Empire, now northern Greece, Macedonia, Bulgaria and Serbia, are generally known as Sarköy, Pirot or Thracian kilims.  One other group of attractive, mostly small, flat-weaves made its appearance in the past two centuries: the so-called Manastirs.  Conceived as prayer carpets and mainly with a yellow ground, their harmony of color and form distinguishes them from those that were until recently called Turkish.

In the mid-1980s, the Viennese carpet landscape was enriched by the exhibitions organized several times a year at the Weber Gallery by the carpet merchant, Muammer Kirdök.  This is when I saw my first Manastir, a yellow-based prayer kilim - it was love at first sight.  During the next fifteen years, Kirdök’s exhibitions mainly focused on Sarköys and Manastirs. My interest in both groups and in the geographical and historical facts about them grew and, naturally, I succumbed to the temptation and bought.

What makes the Manastir kilims so special?  They certainly differ from the Sarköy products in color and motif although they both stem from the same geographical region.  There are two subgroups of Sarköy; the west Bulgarian kilims differ from their east Bulgarian cousins in color, density of weave and pattern. To me it is important to note that despite Bulgaria’s five-hundred-year domination by ethnic Turks, a gradual fusion between Bulgarian weaves and Turkish motifs came about.  Slavic ethnic groups adopted the patterns and varied them; Bulgarian weavers in the nineteenth century primarily produced for their Ottoman clients.  The kilims continued to be produced as Bulgarian and Serbian products even after the Turkish domination had ended - well into the twentieth century. Following the civil war in Yugoslavia, Sarköy-type kilims occasionally turned up in Vienna’s flea market, but they were more recent products and poor in quality. The predecessors of this flatweave were nineteenth century west Bulgarian products known in Serbia as Pirot kilims.

Of the many assumptions about the Manastir family tree, those made by S. Berntsson seem the most plausible to me: parts of the Karaman Confederation were deported to Macedonia in the mid-fifteenth century, that is, to the province of Manastir (today around the border between Macedonia and northern Greece), where they were made to settle.  One could ask: In contrast to the Sarköy kilim, why did none of the Manastir techniques infiltrate weaving styles in the Balkans?  Is there a connection with the forced resettling?   Do tribal groups adhere more strongly to their traditions in times of adversity?  By delimiting themselves from their neighbors, do they hold on more to the familiar symbols of their homeland?

It is the simplicity and the ostensibly naïve imagery of the Manastir weaves that makes them especially impressive.  They speak of their wide valleys, far mountains, the heat and the barren landscape - about Asia and not Europe; about a simple and long gone rhythm of life.  The prayer arches are reminiscent of entrances to Lycaonian rock graves. Looking at Manastirs takes one on a journey to far-off lands and long-lost times. They are doorways through time and space.

Heinz Arnold
Vienna, August 2002

The Manastir kilim, a small category of textiles, has primarily thrown up questions and only provided very few hesitant answers.  Though known by this name for the past twenty years, only single items turn up by that name on the market, and very little knowledge about them exists.  A variety of assumptions about their origins have been in circulation: they were either known as Balkan kilims or west Anatolian kilims.  The name “Balkan kilim” made them unattractive to the market and, consequently, also to the collector; it was a devaluation in the hierarchy of provenance because they were not “really Turkish.”  This prejudice needs to be corrected as our comprehension of the connected cultural content strongly deviates from the Turkish interpretation of its own history.

Oral tradition and literature both contain very sparse and contradictory facts. Not so long ago - let’s say about twenty-five years - one called the “Manastir carpet” a monastery carpet because Manastir in fact does mean “monastery” in the native language.  Until recently, Sarköy kilims produced in neighboring regions were traded as Ghiördes kilims.  Such superficial and hasty classifications, however, became insufficient as soon as one began to test and compare.  The questions were always, “What is it?” and “What is it for?”

Finally, it was Sonny Berntsson’s article in HALI 112 that threw light on aspects relating to the geographical provenance and the historical and cultural environment of the ethnic groups who are supposed to have woven the Manastir kilims.   In combination with other opinions that were published earlier, it is currently the most extensive account.   Although some of the pieces of the puzzle did not seem to fit well as they may have either been incorrect, extraneous or inappropriate to the generally accepted information, the above-mentioned article is regarded today as the most comprehensive report.  Amongst all those who have written on the subject, Berntsson was the only one to have found access to members of ethnic groups in the form of an oral history relating to Manastir carpets.

What information can augment these oral traditions?  How does the local history of the Manastir people fit in with the greater Ottoman Empire, the predecessor of today’s Turkey?

Our (Central European) perspective of the expansion of the Ottoman Empire differs from the way the Turkish view themselves.  We see the Balkan States as just a peripheral region of Europe; we must concede that our focus in Europe lies elsewhere.  For the emergent Ottoman Empire, however, the Balkan areas (today’s Albania, Bulgaria, Macedonia, parts of Romania and Serbia) were not peripheral states but its heartland.  In this Empire’s geographical designs, the Bosporus zone (Byzantium as an enclave) was conceived of as the center so that the settlement areas expanded in equal measure to east and west; the Mongol invasions from the east forced the ethnic Turks to move westwards.  In the fourteenth century, when the first regions in the Balkans were arrogated, Ottoman conquests were primarily made with the intention of appropriating land for settlement east of the Bosporus.   Their first capital city Edirne (1366 under Murad I) also lies close to the assumed axis of the Bosporus. After the united Bulgarian and Serb forces were ultimately defeated near Kossova in 1389, nothing stood in the path of the foundation of the Ottoman state; the conquered regions were thus integrated into the growing Empire.

Various Yörük tribes and the Oghuz Turks as well as nomadic tribes from the Saruhan (in Anatolia) - tribes who were the cause of much turbulence - were the first to be made to settle into the Sar region (today Macedonia).   But Ottoman policy did not exclude settlement in the other direction: after 1369, Bulgarians were transplanted to Asia Minor (cf. Nickel, Osteuropäische Baukunst, Du Mont, 1981).  Colonization continued in full force well into the eighteenth century.

Amongst the colonizers were also Karamanogullari tribes from the Toros region, who captured residential and pasture land around Manastir in Macedonia; this supposedly happened between the fifteenth and the sixteenth centuries. As the products were often named after the place of trade, it is assumed that it was this ethnic group that produced the Manastir textiles.

The Turkish settlement policy reached its zenith in the eighteenth century, but also its end; the nineteenth century witnessed the settlement of both Turkish as well as non-Turkish peoples in these Balkan areas in equal percentage.

Historical events in the nineteenth century led to a reversal; the declining Ottoman Empire lost several wars and, therefore land.  A wave of remigration to Anatolia resulted.  The first big setback came with the Crimean War (1853-56), followed by the Ottoman-Russian War from 1877 to 1878, which led to the establishment of the independent Balkan States.  In 1885 came the annexation of Rumelia, i.e., the Tuna province, by the newly founded Bulgaria.  Meanwhile, remigration continued between these important dates as the population of Turkish origin felt unwanted and no longer secure; the situation culminated in 1924 with the agreement to mutually repatriate the mainly Greek and Turkish populations.  There are about one million ethnic Turks living in Bulgaria today. This historical outline is based on Bilal N. Simsir, The Turks of Bulgaria, Rustem & Brother, London, 1988.

The remigrants (muhajir in Turkish) could, of course, not be settled in one area.  They were spread over regions where they could fit in with other tribes living there, whose way of life corresponded to their own, and where they could be held in control.  The tribes were split up: ‘Divide and Rule,’ a legacy of the East Roman Empire that we should know well.  This is also the reason for varying accounts.   Information about the place of origin of Manastir textiles can help reconstruct this ethnic dividing.

These partially conflicting accounts, however, do create altogether a picture: descriptions of old settlements in the Balkans often point at Manastir (Bitolja); neighborly relations to the Sarköy groups are particularly emphasized here.  A clear connection with Bulgaria, where large numbers of ethnic Turks still live, is supported by Istanbul’s bazaar.  Today, Bulgaria is considered to be the richest source of Manastir weaves; there is usually silence about the reason for this connection.  The following could well be the real reason: large segments of the ethnic group obviously did not return to western Anatolia as muhajir, but remained in the Balkans in Turkish settlements.  They apparently took along with them all their household goods, which is why their old textiles could be bought in Bulgaria, and also why their later products show strong Bulgarian influences (cotton warp, pattern elements, etc.).  Turkish dealers usually do not differentiate between the names Manastir and Sarköy, which often causes some misunderstandings.

In the past, one would often come across the terms “Balkan kilim” and “West Anatolian kilim” for products by the same group.  Those that were made in the Balkans (in the former home) shared a technical and aesthetic constant which became subject to change in the new homeland.  As mentioned before, the ethnic groups that were united in certain earlier settlement areas were then divided during resettlement and made to live with tribes already living there.  The new neighborhood, and altered circumstances (access to materials), influenced their store of patterns as well as their techniques.  In all likelihood, every reference to regions that count as settlement areas in western Anatolia has a claim to authenticity - at least in my opinion.

This division of ethnic groups finally led to the dissipation of their specific cultural expressions; it seems that no kilim was woven after the 1920s which was unmistakably a Manastir.

The voices of dealers through whose hands most of the Manastir kilims that have been on the market in the last few decades have passed have disclosed a few new facts: initially, kilims made by the muhajir were chiefly purchased in Izmir, Bursa and Konya, and later exclusively in Bulgaria.  Here, the places mentioned are primarily Deliorman and Silistija.  Without ever mentioning Macedonia or Manastir, the name of the grouping is nonetheless inherent in this information, and everyone knows what is meant by it.  The ethnic groups are always referred to as the Yörüks.

Interestingly, descendants of muhajir families report that while a large number of Turkish emigrants to western Anatolia went back to Bulgaria, some returned to Turkey.  For them, this immigration process was never concluded.

We made an attempt at on the spot investigations into the history of the Manastir.  For personal reasons, Ochrid, in western Macedonia, presented itself as a point of departure for our expedition.  Bitolja, formerly Manastir, is just a seventy kilometer bus ride away from Ochrid.

Our investigations in Bitolja were not very successful.  It would have been futile to hope for any material relics since, in the 1970s and the 1980s, the Istanbul bazaar dealers and their intermediary traders had thoroughly scoured the Balkan countries. A Turkish population still lives there - four mosques are still in use - but our conversations with a Muezzin kindled in him no memory of what interested us.  Nobody knows anything about other Turkish groups that left this country almost a century ago. With the help of just a handful of clues, we tried to reconstruct possible conditions of their lives.  The landscape seemed to suggest that people in this urban area had once made their living from agriculture and sheep farming; there is there plenty of pasture land.  The low mountain range (approximately 2000 meters high) is not hostile and it has vast meadows; sheep-breeding Karamanogullari tribes may have once lived there.  The monasteries, often mentioned in connection with the Manastir carpet weavers, are few and far between.  What is meant by “Manastir” in Macedonia is just a tiny church with very little or no living space.  Monasteries that can be seen in Bitolja today (“manastirs”) are quite new and were probably built on the site of older ones, but they are certainly not monasteries as we know them.  There is very little likelihood that a small ethnic group had once lived in such few and tiny “churches” (see Berntsson on this).  That the word “Manastir” as a name for a weave has anything to do with the word “monastery” is just one of the many tales that are told to add a bit of color to the unknown.  What remains is the assumption that the city formerly known as “Manastir” was once the trading place that gave this kilim its name.

While showing photos of the kilims, we encountered curiosity but not recognition.  Professor Tomovskij, our acquaintance in Ochrid, claimed that the frequently used combination of red and yellow in the kilims was very popular in Macedonia - even the national flag bears the same colors.   The custodian for textiles in Ochrid’s museum, Ms. Marija Kanevce, advised us to continue our search in Bitolja or Prileb since both cities are not far from Ochrid.  The custodian in Macedonia’s Ethnographical Museum in Skopje, Ms. Jasemin Nasim, with whom we continue to exchange ideas, was very interested.  She said that she had heard of a weave by the name of Manastir, and especially about the trade, but she had never seen any pieces.  Later, she wrote us expressing her doubt about the fact that “Manastir” carpets had originated in the Balkans; that is, in Serbia, Macedonia or Bulgaria.   The evidence of their relationship to the Sarköy, even to the Bessarabian kilims (Romania, Moldavia), is unequivocal.  When speaking of a formal relationship to West Anatolian works, however, it is still important to emphasize that they share a common strangeness in color and structure when compared to the Anatolian kilims. On account of their simple and austere forms, the kilims should be seen as akin to other works that were made in the Balkans such as the gravestones of the Bogumiles (a Christian sect that was influential in the Middle Ages in what is today Bosnia).

Were it not for the textiles that turn up every now and again with almost no information, one would be tempted to believe that everything related to the Karamanogullari and their products is mere myth.  A thin veil of facts surrounds the void, that place in memory which has retained nothing about an entire ethnic group.

Let me quote from Dubravka Ugresic on this problem of disappearance and denied ethnic groups: "In the Balkans, the story slips away in the other direction and instead of being about remembering it becomes a story about forgetting."
(Balkan as Metaphor, Dusan I. Bjelic et al, Massachussetts Institute of Technology, 2000)

The colors of the Manastir kilims have been described as exceptional in the few texts about the group.  Their unique palette and some of their technical characteristics have made these textiles distinct from mainstream Anatolian kilims.  It is these qualities that make it possible to differentiate them from their neighboring Sarköy kilims.  But the two share only a few elements from a store of patterns, the origins of which are primarily to be sought in the Sarköy.

Berntsson goes into great depth about color in his article in HALI 112, making it very important source material; from dyers woad, blue, and the shades of green resulting from it are easily recognizable.  While Berntsson only found small traces of it in the knotted items he investigated, yellow plays an important role in the Manastir kilims and is thus a rarity compared to other types.  Only Konya produces carpets with yellow as a main color but again, these are knotted carpets.  The choice did not fall on yellow because it was available in excess compared to red and blue, but because it bore special and exceptional significance.  This significance, however, has not been passed down to us. The sources for yellow vary from region to region.  The many shades of it range from a pale yellow with nuances of green to orange-yellow.

I cannot follow Berntsson when he speaks of the trade with dyestuffs.  According to him, dyers woad was used instead of indigo and cochineal instead of madder lake because both indigo and madder lake were difficult to obtain in the Balkans. Cochineal, however, had to be imported and was expensive, which is probably why it was excluded from trade in the early days.  Why could the reddish blue shade not have been won from the indigenous kermes?   According to available literature, it not only existed in southeastern Europe, but was also being used there.

Our knowledge about dyes is limited to a standard repertoire that was probably determined by the carpet trade itself.  The traders tried to retain the criteria for color and placed great emphasis on time-tested recipes which is partially the reason why “well-dyed” wools resemble each other so much even when their origins vary.  Thus, the prevailing trading system not only linked but also unified the local traditions.  Today, this palette of color that was filtered by the historical process is taken to be “authentic”.  To make it more explicit, what I am trying to say is: it is as though in music the use of one particular key were obligatory to the disqualification of every other.

In Manastir kilims, small quantities of “exotic colors” like purple, violet and pink can be found occasionally.  These colors are not permanent, they have either faded due to exposure to light or to washing.   Although such colors are normally termed as “synthetic”, this classification needs to be examined more closely.  On studying the literature about them, one encounters dyes that are natural, but not permanent (Ziemba et al or Gomersall).  It seems plausible to me that efforts were made in regions outside the main focus of interest to obtain dyes from the neighborhood, for instance, from local flowers, leaves, fruit skins and berries.  The import of small quantities of synthetic dyes would thus be an exception.

The slightly subdued tonality of most Manastirs is probably to be ascribed to intensive use and the resulting frequent washing.  These kilims rarely appeared on the market since they were not made for trade but for personal use (cf. Bausback on this).  Although traces like stains or marks left by pots correspond with their use, only a few pieces indicate their explicit use as prayer carpets.  The other pieces with the motif of a prayer arch often served a different purpose, such as wall hanging or a decorative rug.

The formal language of the Manastir kilims is probably what impresses most.   The first that I acquired (number 40 in the publication) had been returned to the dealer by customers who could not bear the “face” on the carpet.   I, of course, kept it.  I found this play on physiognomy particularly interesting; it was an image that could be interpreted both as a vase in a niche or as a feline/human face.  The Manastirs pictorial language moves on a level of communication where the meaning of the forms remains open to interpretation.   One can perhaps penetrate the pictorial structure of these kilims in order to discover an analogy in other provenances to this type, whose originality defies every form of conformity.

More than any other textile of its kind, the Manastir’s patterns reveal a protective function.  They defy being ascribed to that decorative form which so easily turns a piece of handicraft into an intercultural article of trade; flowers are strewn but amulets are not - they must be used sparingly.   Their store of motifs is replete with apotropaic or shamanistic symbols: hands, finger motifs, amulet forms, eyes, combinations of forms to avert evil, protective zigzags, and ram horns that are reminiscent of a totemic animal.   There is also nothing representative about these pieces; they do not aspire to be public.  I also believe that they were rarely used in mosques; protecting the house of God with symbols mainly drawn from shamanism seems absurd to me.  Unlike most other kilims, the Manastirs strictly remained within their own cultural circle, leaving it only once it had collapsed.    One must probably ascribe these products entirely to feminine aspects of Turkish culture; their language conveys a protective function that the women could allow to attribute to the house.

Berntsson made a major contribution with respect to technique, for instance, the peculiarities of Balkan sheep wool.
The usually brown wool warp is only occasionally mixed with light-colored wool.   Berntsson assumes that most of the textiles with a brown warp originated in the Balkans.  However, brown wool was also used in West Anatolia after remigration.  The brown warp yarn often leads to the assumption that kilims from the Balikesir also bear patterns typical of the region and can be presumably filtered out as Manastir or muhajir products.  Interestingly enough, yastiks usually have a white warp.

Later examples with cotton warp instead of the customary brown wool warp were probably not made in West Anatolia, but by settlers in Bulgaria who preferred to stay on with the ethnic Turks living there.  This assumption is based on a similarity of motifs with Bulgarian Sarköy kilims.   The weave is flat and not a rep, giving them a clothlike texture; this kind of weave structure is often found in the Balkans, for instance, in the Szekely kilims, the Bessarabian kilims, etc.

The ends in most of the older kilims have not been preserved; the warp threads in more recent pieces, particularly in one old piece, are interlaced in two or three rows of knots. One could thus conclude that, apart from a few exceptions, this was the most commonly employed method for ends in the Manastir type of kilim.

All the available quotations are probably compiled here.  Some of them were probably themselves quoted and are not passages of personally heard oral history.

BALPINAR, Belkis and HIRSCH, Udo, Flatweaves from the Vakiflar Museum Istanbul, Uta Hülsey, Wesel, 1982, p. 33.
Balpinar and Hirsch ascribe the Manastir products to immigrants from the Balkans; the wool, the color and the texture of these kilims differ from all other Anatolian weaves.  The example illustrated here is ascribed to Antalya or Fethiye.

BRANDSMA, Arendt and BRANDT, Robin, Flatweaves of Turkey, Crawford Press, Bathurst, Australia, 1995, p. 31.
Brandsma and Brandt also speak of remigrants who adapted themselves to their new environment.  Their weaves changed in the process and they also adopted patterns from their neighbors; they emphasize the proximity to Sarköy - the color schemes of the groups are clearly differentiable.  Kilims with a fine cotton warp count as “Bulgarian”.

BAUSBACK, Peter, Kilim, Klinkhart and Bierman, Munich, 1983, p. 51ff.   Bausback calls Akhisar the place of origin of the Manastir and assumes semi-nomadic life styles; there is no mention of remigration. The knowledge drawn from viewing the pieces leads to the observation that they are usually small in size and that they functioned as decoration or floor covering.

BERNTSSON, Sonny, “Balkans and Back” in: HALI 112, London, 2000, p. 98ff  According to Berntsson, the regions around today’s Bitolja, formerly Manastir, would have been the main settlement areas in the Balkans.  The information he had collected led to the assumption that the most significant settlement areas for the remigrated Turks (“muhajir” is immigrant in Turkish) were Izmir and its surroundings, Canakkale, Kütahya, Bursa, Eskisehir, Denizil, Tokat and Manisa.

ERBEK, Güran, Kilim, Secuk A. S. Istanbul, 1982, cat. 33.  Describing a particular kilim, Erbek says that it was woven by Balkan immigrants who had settled in Mihaliccik in Western Anatolia.

ESKENAZI, John, Kilim anatolici, Milan, 1985, p. 59.  Eskenazi places the weavers around Izmir and Usak.  According to him, the kilims were woven by a tribe that lived in the region around Saloniki and following the remigration to western Turkey, continued to produce the kind of carpets they had woven in their former home.

HULL, Alastair and LUCZYC-WHYOWSKA, Jose, Kilim, Thames and Hudson.   These authors suggest the Manastir kilims are different from other Anatolian weaves and are mostly known in prayer carpets size.  The kilims were woven by the remigrants from the Balkans who had settled in the Mihaliccik region.

KIRDÖK, Muammer, catalogue, Vienna, 1992, p. 8.  The author claims that because of the various influences, the technique and the motifs in the entire group are not homogeneous.  Color and the way pictorial space is dealt with are seen as the most cohesive elements. The producers were tribes that had originally settled in the Balkans and later in the Mihaliccik region in West Anatolia.

KREISSL, Rainer Himmelspforten, Hirmer, Munich/Prague, 1988, pp. 22-23.  Kreissl finds the attempt to locate the Manastir difficult due to its wide range of material, symbols and motifs.  About the kilim he presents in this book, he says that it was not made in the Balkans but in the areas around Afyon, Eskisehir and Sivirihisar.

PETSOPOULOS, Yanni, Kilims, Thames and Hudson, 1979, p. 82f.  Petsopoulos does not speak ofremigration and places the origins in the regions around Izmir, Akhisar, Usak and Alasehir.  Alongside Izmir, he also mentions Istanbul and Lesbos as the places of trade.  Among the three pieces illustrated here, I believe that only the third one is really a Manastir. The others are more likely to be Balikesirs; this is also clear from the white wool warp.

WILLBORG, Peter, Vaväd Magi, Stockholm, 1992, p. 68f.  Willborg shows a kilim from the prayer rug category. Although he ascribes its origin to Fethiye, the piece bears many Sarköy features.  He also speaks of immigration to the Manastir area in the Balkans; remigration to Turkey had already begun in the second half of the nineteenth century.  Nevertheless, precise classification is difficult since several pieces were made in the former settlement areas and then taken to Turkey.  The tradition was continued in their new homeland but has now completely vanished from there.   Willborg points out a technical peculiarity in some kilims: the weft threads encircle two warp threads in the patterned areas.  In areas without a pattern, however, they wrap only one weft thread as usual.  This is why their origin is assumed to be in Europe.

ZIPPER, Kurt/FRITSCHE, Claudia, Orientteppiche 3, Battenberg, Munich, 1989, p 58f.   The authors assume Macedonia is the place of origin of Manastir carpets, with its capital Bitolja/Manastir at its center.  There is no mention of remigration.