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.

John Wertime on Sumak Bags

by R. John Howe

On September 22, 2007, John Wertime conducted a rug morning program at The Textile Museum on the subject “Sumak: the Very Collectible Bags from Northwest Iran and the Transcaucasus.”

Here is an older photo of John, taken in 1997.  It gives you a more sober view, more appropriate to his scholarly identity and persona,

And provides a useful corrective and counterpoint to the jocular one below that I took one recent Christmas at a party of ruggies.

John and I talked a bit beforehand about whether photography could be permitted.  He said that some of the pieces he would show were not his and so that probably we should not take photos.  So what follows is doubly “virtual.”  You are participating in this rug morning via the internet and, although some of the pieces I will show below are those shown by Wertime this morning,  most are not.  What I can manage for you is to include published pieces of all the geographic areas he treated.

Wertime is an authority in this area.  He lived nine years in Iran as a young person, speaks Farsi and was a member of in early group of rug scholars in Iran what included Parviz Tanavoli and, I think, Jenny Housego.  Wertime was in the Persian studies program at Princeton, and has published several works in the area of his rug morning topic.  He authored an early article on flatwoven structures, co-authored “Caucasian Covers and Carpets” with Richard Wright, and more recently, published “Sumak Bags of Northwest Persia and Transcaucasia,” a volume on which this rug morning was explicitly based.

Wertime began with some general remarks.  

He said that we are disadvantaged in rug and textile studies by a paucity of evidence.  There are few instances in which we have direct knowledge, for example, of where a piece was woven, or of who wove it; what he and Wright called an “anchor piece.”  For this reason, he indicated, we are driven back to more indirect methods of establishing such things as attribution and provenance.  Information from travelers, those who deal in textiles, and various features of the weavings themselves, are the bases for what we, rather insecurely, “know.”  As a result, he said, smiling, nearly everyone is “expert.”

Wertime said that the industrial revolution had an enormous impact on the traditional production of textiles.  He said that no activity occupied more time in traditional society than did the production of textiles.  In traditional society most people, excepting the upper levels of society, often had maybe one change of clothing.  The industrial revolution made the production of textiles much less expensive and as a result, reduced their prestige markedly
He said that the weavings of his subject were made by women in a geographic area approximately reflected in this map.  He would move south to north.

He said that the southernmost region in which the weavings on which his talk would focus, were made is defined by the areas of NW Iran named Kamseh (not the South Persian Kamseh), Garrus, Qazvin, Saveh and Kharaquan.  On the west it bounded roughly by Anatolia and moving north, encompasses the transCaucasus, including Georgia and Armenia.

Wertime noted that the weavings of interest in his session were woven by women who husbanded both the animals from which the fibers (cotton aside as plant-sourced), were taken and the plants used to make the dyes.  They also did the dyeing, the carding, the spinning, and the weaving themselves.  These bags were the result of what we would today describe as a “vertically integrated” production.  There was little “division of labor.”

Next he talked about the character of “sumak.”  He said that “sumak” is a shorthand term for 
a weave in which the warps are wrapped with wefts.  He distinguished it from “tapestry” in which the warp and weft are “inter-laced.”  He held up a side panel of a cargo bag from the Shirvan area made using slit tapestry and asked the audience to note the holes visible in it.  

Then he held up another panel made using “sumak” and asked the audience to notice that there were no holes visible in the resulting material.  He said that tapestry weaving had been mechanized but that “weft-wrapping” has, so far, not been.  Sumak is by its nature tied to and requires the use of a more traditional mode of production and is, in that sense, likely more closely reflective of that tradition.

He said that more sumak is woven in NW Persia and the transCaucasus than anywhere else.  He acknowledged that some is woven in eastern and western Anatolia and in southern Iran but that it was rare in Central Asia.  He asked Kelly Webb and me whether either of us knew of Central Asian uses of sumak.  (The closest thing I could think of quickly is the Turkmen palas, which have extra weft patterning but not, I think, of the wrapped sort.  I have also seen some flatwoven Ersari khorjin, but don’t know what weave was used, perhaps some sort of brocade.  As I wrote today I found a published item of Turkmen sumak, but I think it is a rarer thing.)

Wertime said that, in his view, the fact that more sumak was woven in the NW Persia-transCaucasia area than anywhere else suggested that this structure likely originated there.  He said that he was following a principle analogous to one in biology and botany under which species are usually found to have originated near the area in which they occur most frequently.

He next noted that the groups who wove these sumak bags moved geographically.  Some were nomads and moved following their herds from season to season.  Sometimes whole tribes were moved considerable distances by government fiat.  Third, Wertime said, there seems to have been considerable intermarriage between groups in this area, and often women moved from tribe to tribe, or sometimes some distance, to marry.  He said these various movements of people helped explain both why and how different designs and techniques might appear in different geographic areas.  (The situation he described seems contrasted with what is generally seen to be the predominant pattern of marriage among, say, the Turkmen.  The latter seem to have tended to marry closely within tribal, even kinship groups.)

Wertime noted that flatwoven structures vary far more than do those associated with pile weaving.  And the color palette often associated with flatweave, especially Shahsavan sumak, is considerably wider than that of the most studied textile group, the Turkmen, where there is rejoicing if more than five colors are found in a piece.  This variety makes textiles woven with sumak especially interesting objects of textile collecting and research.

With these introductory remarks Wertime moved to the pieces he had brought as well as to some others had.  He said that he would organize his progression geographically moving south to north.

The “southern region” as defined by Wertime, includes sumak bags woven in the Qazvin, Saveh and Kamseh areas.

Wertime said the southern-most  Shahsavan in NW Persian in the Qazin and Saveh areas are the Inallou and the Baghdadi.  This end panel below is from one of the oldest Baghdadi bags known and was shown in this rug morning program.  It is attributed to the Saveh –Kharaqan area and estimated to have been woven in the 2nd or 3rd quarter of the 19th century.

This example shares the heavy, stiff handle of some other Baghdadi bags, as the result of wrapping two warps rather than one.  They are sometimes wrapped so that one warp is on top of the other.  The line of weft twining countered to form a chevron design at the top of this piece, and the minor “connected buds” border, are also characteristic of Baghadidi weaving.

Wertime said that as wefts are wrapped around warps they “angle” in various degrees.  Sometimes the wrapping in each subsequent row continues so that the angle of the wrapping in a given row parallels that of the row(s) next to it.  But the wrapping can also be made in a subsequent row so that it slants in an opposite or “countering” angle.  When the slant of the wrapping alternates row to row, the sumac produced is described as of the “countered” variety.  This countering affects the texture of the fabric, but is not reflected in its design.  Baghadadi sumak, Wertime said, is characteristically of the countered sort.

In his book Wertime includes only one piece from the Saveh-Kharaqan area, but the next 40 pieces are from the Kamseh area.  So I am going to have to be very selective about the latter.  I’ll show you four.

This is Plate 14 in Wertime’s “Sumak” volume and was not shown in this rug morning.

 It is a side panel of a bedding bag from the Kamseh area.  Estimated to the 1st half of the 19th century.  Notice that it has a border all round.  We’ll talk about that later. 

I find its graphics and colors to be exceptional.  Wertime says this is a masterful use of this field design on a large scale.  Note the “bird-on-a-pole” main border is a version of one that we see elsewhere, often on Yomut Turkmen “envelope” style bags in pile.

The piece below is a second Kamseh area piece that is Plate 26 in Wertime’s “Sumak” volume.  It was not one of those shown in this rug morning.

Wertime describes this as a “masterpiece of design and color”…with its “…five complete and two half Lesghi stars…”  He notes that in some pieces the stars are arranged are aligned vertically with one another.  I think the dark outlining around them is especially effective.

The piece below is a third Kamseh area piece.

As you can see, it is a complete khorjin set, estimated to have been woven during the 3rd quarter of the 19th century.  Wertime says that it displays a distinctive Kamseh palette.  He also says that the border above the closure panel is “infrequent.”

I find it a little crowded and busy, but agree that its colors are glorious.

The saddle bag face below is a fourth Kamseh example (Plate 39 in his book) and was not among those shown in Wertime’s rug morning. 

He says that use of the device in the field compartments of this piece is “widespread and ancient,” and quotes sources suggesting similar usages in 14th and 15th century Anatolian fragments.  Wertime notes this design also occurs in Uzbek weaving.  He estimates that this piece was woven in the 1st quarter of the 19th century.

I am a sucker for compartmented designs and am very attracted to the large scale of the field motif.  I also like the fact that the border does not compete with the field, nor does the larger scale design on the closure panel, the latter, perhaps because it is still somewhat smaller and different from the device used in the field.

The third geographic area Wertime treated in this rug morning was “Garrus” (most of us think “Bidjar” when we hear “Garrus”) and his Garrus example is a repeatedly published and widely praised one.  Wertime included it among his rug morning pieces and it is Plate 42 in his book.  It is a side panel of a bedding bag with a border all round.  It is treated there in a two-facing-page spread so I can’t scan it all successfully, but here is the detail of it that I can manage.

Wertime describes it as very fine and very old, estimating it to the 1st half of the 19th century.  He also notes that the border on the right side is not original and was “most likely) taken from a similar panel on the other side.

At this point Wertime talked a little about the seeming use of design in complete bedding bags of this type.  In general, it seems that if the side panel design does not have borders all around it, the design on the side panel is carried all the way around in the other side panel and both end panels. 

Here, below, is an image of a complete bag of this sort from the Azadi-Andrews volume “Mafrash” that shows the side panel design continuing all round the other four panels.

But, in the case of a side panel with a border all round, usually, the other three panels will be different from it.  The way in which this difference can occur can apparently vary, but here below is another complete bedding bag from the Azadi-Andrews book that shows one side panel with borders all round it and the other three sides in a zigzag design.

If the right side border on the Garrus piece above comes from a matching side panel on its opposite side that would seem to be a rare occurrence, although the right side border is so similar to that on the rest of this panel that that explanation seems plausible.

The next next geographic area that Wertime treated in his rug morning program was that of Hashtrud-Miyaneh.  The piece below is Plate 52 in Wertime’s “Sumak” book but was not one of those shown in this rug morning. 

On the other hand several pieces from this area and with this design were shown.

Wertime indicated that this is another form of sumak in which there is 90 percent displacement of every other warp in the wrapping.  Another, even more unusual feature, he said, was that this piece is made with an “extra-weft wrapping” structure in which “a true knot if formed in the wrapping process.”  Wertime used the example of the knot one uses to tie one’s shoes, but I think what he intends is that a lot of “knots” in the world of textiles are not “true knots” in the sense that they are not “firm on the basis of their own construction.” 

(Both symmetric and especially asymmetric knots in pile weaving do not meet this test fully and require “pinching” by the wefts between knot rows in order to remain firm.  The asymmetric knot is in fact closer to what might be better described as an “in lay” in this respect.  The symmetric knot is “firm” as long as tension it maintained on its cut pile ends.  Otherwise, it too, is dependent on the pinching wefts to retain its firmness.)

I think Wertime is suggesting that this extra-weft version of sumak is one in which the knots are firm on the basis of their own construction.  That is, I think, a fairly unusual occurrence in the weaving world.

Wertime also used this khorjin set to reinforce the point that in general designs seem to move from techniques that are more restrictive to those that are less so.  Sumak, like pile weaving, is one of the less restrictive techniques.  Nearly any design can be executed on it since its minimum requirement is only that one patterning weft must circle one structural warp.  Wertime brought out the example of zili shown below, a species of brocade and a more restrictive technique.

Notice that this zili has this same design.

Wertime said that he felt that this design likely flowed from pieces made in more restrictive techniques, like the zili, to those made with less restrictive ones like the sumak.

A second piece from the Hashtrud-Miyaneh area is the saddle bag half below.

It was in the room for this rug morning and is dated 2nd quarter 19th century.

Wertime said that the provenance of this piece is “problematical” and that its attribution is based in part of the fact that it has warps on two levels.  He says that the central medallion is “rare” and that the birds are drawn somewhat differently than most in this area.  It is Plate 55 in Wertime’s book.  Plate 56 is a very similar piece.

The next geographic area Wertime illustrated was Moghan-Savalan.  He had the piece below in the room (it is Plate 73 in his book) and explained why he had included it rather than the more striking piece (Plate 74) that Wendel Swan owns.

He said that he estimates that this piece is older than Wendel’s piece (he says 1st quarter, 19th century and estimates Wendel’s piece as 3rd quarter) and says that it has a lighter purple or violet in its corners that is derived from madder.  He says that dye is common in Anatolian pieces, but rare in those made by the Shahsavan.  The piece shown lacks the exquisite condition and the saturated colors of the Swan piece, but its fragmentary dark-ground border with stars and cruciform devices still effectively frames the milder field this old weaving.

A second piece from the Moghan-Savalan area is the one immediately below.

This piece was not in the room on this rug morning, but is one of the most beautiful weavings of which I know, so I’m giving it to you from Plate 87 in Wertime’s book.

There he says that it is “a marvel of colour and elegant simplicity.”  He attributes this to use of a somewhat larger scale and a spaciousness created by refraining from making the cruciform devices larger.  He sees it as a “stunning display of the dyer’s art” and especially admires its purple.  Wertime also explicitly admires the simple borders and the “elaborately decorated bridge in slit tapestry.”

It is estimated to have been woven in the 1st quarter of the 19th century.

Wertime showed several smaller bags and the one immediately below, about 10 inches square, may not have been among those shown, but is similar in coloration and design to one that was.

Wertime also attributes this piece to the Moghan-Savalan area, describes it as a “small saddlebag” and estimates that it was woven in the 2nd half of the 19th century.  He notes that the ground weave of this piece is warp-faced.  He noted that these are often seen as children’s items but said that they are made at a level of quality that makes this doubtful.

Wertime next moved to items from the Kuba area.  The piece below is a saddlebag “half” and is estimated to the 1st half of the 19th century.

It was in the room for this rug morning, but is also Plate 104 his his sumak” book.  Wertime said that this is one of only two examples of this type of which he knows.  He said that the “palette and overall feel are unmistakably Kuba.”  In his book, he makes analogies between some design items in this piece and other particular old rugs.

Face to face this piece draws immediate attention as an older and likely sophisticated piece of work.

Wertime’s next example, below, comes from Qarabagh.

It is a salt bag, estimated to have been woven in the last quarter of the 19th century.

Wertime calls it “beautiful” and cites its wonderful graphics.  It was not in the room for his rug morning, but appears as Plate 117 in his book.

Wertime describes the next saddlebag half, below, also from Qarabagh as “one of the greatest sumak bags known.” 

He says that is proportions are “optimal” and praises its “simplicity” and “grandeur.”  It has the sort of spacious drawing that might draw even a determined Turkmen collector to a Caucasian weaving.  We were not fortunate enough to have it in the room. You need to go to his volume and read more of his fulsome praise for this piece.

In Wertime’s book, his next geographic area is Kazkh and he offers the salt bag below (Plate 133) as one such example.

The good, strong colors of this piece are appealing and its graphics are striking.  Here is a little of what Wertime says about it in his book. 

“An attractive feature of this bag is the three-dimensional effect created by the device that fills the hexagonal centres of the repeating motif of the ivory bands.  The middle of this device appears as a negative cruciform that pulls the viewer into another realm.”

Another piece from the Kazakh area was in the room.  It is a side panel of a bedding bag and is Plate 134 in Wertime’s sumak book, estimated to have been woven 3rd quarter 19th century.

Wertime notes the “stiff, thick, heavy handle” of this piece and “its unusual field design.”  He says that it is woven with in a “4/2 extra-weft wrapping with the warps displaced 90 degrees.”  It seems similar in some senses to the first Baghadadi piece we saw but lacks its countering.  It also lacks cotton, something also common Baghadadi weaving.

The field design is very unusual and one of the owners of the piece said that they especially like the “ducks” swimming about in the lower half of the field devices, something that Wertime notes specifically in his description of it.  Wertime says that this treatment of the blue-ground border is rare because it occurs on the sides not just top and bottom.

This is a piece that draws the eye face to face.  I think the use of white especially effective.

At the end of his book, Wertime treats Kurdish pieces separately.  The piece below is Plate 136 and is a complete Kurdish khorjin set.  It is attributed to Northwest Persia and is estimated to have been woven in the last quarter of the 19th century.  We did not have this piece in the room in this rug morning but did have a Kurdish chuval that was quite similar.

Wertime says that the field design is a Turkic one used by weavers in a number of locations.  He says that it is similar to weavings in the Kamseh confederacy excepting that it has “a different palette, its pile bottom (I had forgotten in our recent discussion here that pile bottoms can occur on Kurdish pieces), its use of two ground wefts after each row of wrappings and its frame, particularly the guard stripes.  He makes comparisons with a piece that Housego discusses, her Plate 40, and that Wertime examined directly years ago.  Wertime’s discussion of this piece focuses largely on it attribution.

I want also to show you two pieces from the area of Wertime’s rug morning, but from the Azadi-Andrews book “Mafrash.”  Both of these pieces are end panels from the sort of bedding bags we have been treating.

The first is this powerfully graphic piece below.

It is on page 121 and is described as Shahsavan “North of the River Aras, third quarter.

I find its graphics breathtaking.  We had a similar piece in the room in Wertime’s session.

A second end panel is more sober.

It is on page 123 and is described as “Shahsavan” from “Northwest Iran, mid-XIXth century.”

I like the use of a dark ground and of the color choices and outlining on the hooked lozenges that let them stand out from it.  Azadi-Andrews also draw attention to the simplicity of the “triple horizontal border.”

That is what I have to show you of sumac bags from northwest Persian and the transCaucasus.

This was an excellent rug morning of the more authoritative type.

Our thanks to John Wertime for conducting it.


R. John Howe