Posted by R. John Howe on 10-08-2005 09:06 AM:

Central Asian Attribution Puzzle

Dear folks -

I recently bought a Central Asian piece that is of the sort I collect frequently: bad condition, but interesting in character.

I do not yet have it in hand, but expect to in the next week or so. Meanwhile we can speculate a bit about it, perhaps even offer an informed opinion or two. Here are some images.

First one that shows it overall.

It's about 3.5 ft by 5 ft if I recall correctly.

Here are a couple of images of the mina khani field.

And this is a shot that lets us see pretty well what its borders are like.

Finally, here is a fairly close image of its back.

Now the temptation may be to look at just a couple of the indicators here and to move to ready conclusion. But a correct attribution of this piece may require a little more patience and tentativeness than that which might be tempting initially.

I will gradually reveal what I have been told about this piece, but want to begin by asking for your own attribution suggestions and the reasoning (please) on which they are based.


R. John Howe

Posted by Filiberto Boncompagni on 10-08-2005 10:42 AM:

Hi John,

Okay. It looks like a nice and worn Ersari Beshir, but from what you write you obviously don't think that it is an Ersari.

Jourdan wrote that this kind of design was used also in the Amu-Darya region. Do you presume your rug was woven by Uzbeks, perhaps?



Posted by Steve Price on 10-08-2005 11:20 AM:

Hi John

Looks pretty Ersari to me, too.

But wait! You imply that the obvious is misleading. So it must not actually be Ersari.

But you probably said what you said just to throw us off track. Yes, that's it. Now I'm onto your little game.

Ersari it is. Beshir type.


Steve Price

Posted by Chuck Wagner on 10-08-2005 11:42 AM:

Out on a limb

Hi John,

The picture of the back is sort of fuzzy, but it sure looks like Baluchi work to me. It'll be interesting to read the full story.


Chuck Wagner

Posted by James Blanchard on 10-08-2005 12:37 PM:

Out on a limb...2

Alert: Novice opinion coming up....

The back does look "Baluch" to me too.

That "S" border also strikes a "Baluchi chord" to me also.

Ditto for the white warps.


Posted by David R.E. Hunt on 10-08-2005 01:07 PM:

Compare and Contrast...

Hi John

Find below an image of a classic Beshire chuval in the Mina Khani pattern,ancestral to Herat if I'm not mistaken, and a detail from the corner of a Beshir from one of Joe Fell's "Rug Mornings"

Notice the similarities, the kindered field, the "S" border and floral forms of the elem, as mirrored by the end treatments of the rug.

In the rug we have the "dice" border and "judor"or almond borders(which is also seen in various incarnations in products of other tribes) which say Turkmen. The camel ground for the judor border seems unusual, and might indicate that there has been loss to the sides. Is this weft material we see to the left of the third detail image? Could this be cotton?

With it's profusion of small flower petals, the field more resembles the drawing of the lattice field of Baluchi renditions,

and the following from a discussion here on Turkotek entitled
The Internal Elem Configuration and its Markings.

And our subject rug again.

Next is from the Lattice thread here on Turkotek.

While on the lattice subject, let's not forget the following images from 1420 and 1429 miniatures in von Bode's "Antique Rugs From The Near East" (pg.85) which is accompanied by the following statements

"then regular repetition of certain definite ornamental forms by the most diverse book illuminators permits us satisfactory conclusions regarding their actual decorative stock-in-trade. According to these reproductions the field appears to have been filled with either with a simple scale-pattern or latticework, or else with a continuous and usually rather loose plaited design, in which stars, rosettes, hexagons or other incidental motives were interspersed".


Posted by Marvin Amstey on 10-08-2005 04:51 PM:

If those are cotton warps, its not Ersari; probably some Persian knock-off, but with outstanding colors (presuming that my monitor is accurate). Actually the back looks almost Senneh like.

So, John, what was your initial opinion, before some one told you something that you plan on revealing?

Posted by R. John Howe on 10-08-2005 09:44 PM:

Dear folks -

This piece arrived today and so I can supplement (and correct) some information I've given on it.

It is about 44 inches (approximately 111 cm) wide and 75 inches (about 189 cm) long.

The knot is asymmetric open left. There is pretty marked depression of alternative warps (there is definite "ridging"). The piece is quite finely knotted. Each knot node (looking at the back) is roughly square, so a given knot is about twice as wide as it is tall). So there are going to be about twice as many knots per running inch vertically as there are horizontally. I just counted in one area. There are 8 knots in a horizontal inch and 15 in a vertical one. So it's about 120 knots per square inch (1860 knots per square decimeter).

The warps are not chaulky white with this piece in my hands. They are at least what would be called a medium to dark ivory. They have some of the brushy quality of goat hair warps (I've been feeling them side by side with those on a Tekke torba I have that definitely has fine white goat hair warps). The warps on this piece feel softer than the goat hair, but are also softer than some wool warps on other pieces here. I am not sure (Eiland would demand a microscope) but I wonder if these warps are not camel hair (I also have a piece of camel hair to feel and compare).

The marked whitish fuzzy looking quality of the back in the image above is not general (it may come from some damage in some areas that was not apparent to me as I was considering this piece) and I will provide some direct scans of other areas of the back in the morning.

With this piece in hand, the wefts seem brown in areas of darker pile, but gray in areas of white pile. We'll have some close ups of bare areas tomorrow so you can help me decide that. I'll also ask whether you see any sign of cotton in the wefts.

The white cotton selvedges, most of you will know, have nothing to do with what this rug is and were added at some time for reasons unknown.

The colors of the piece are pretty close, in my view, to what I saw on my monitor. I have not seen it in full sun, but so far see two reds, two blues (both darker), a dark brown, white, yellow-tan and a darker blue-green. That's eight colors if I am right.

The handle, feeling the surface of the back, is firm, even a little hard. Still the piece is thin and flexible.


R. John Howe

Posted by Steve Price on 10-09-2005 06:20 AM:

Hi John

The characteristics you just described probably eliminate Ersari as an attribution. My best guess now would be Belouch, who use lots of Turkmenoid designs and colors.


Steve Price

Posted by R. John Howe on 10-09-2005 07:07 AM:

Dear folks -

We can't really say "by dawn's early light," since it's dark here yet, but here are a few more images of, and thoughts about this, piece.

First, here is an effort to give you a somewhat closer look at the designs on the ends of this rug.

Secondly, notice that the yellow ground minor border which entirely bounds the mina khani field is also placed as a horizontal band above these endings.

Could this piece be an engsi (Reuben offers one with a mina khani field in one of his catalogs)? And if not why not?

The two images below are direct scans to try to give you a basis for helping me decide what I think is the case: that there is no cotton in the structure of this piece.

The first darkish image is from the back and it's likely hard to see wefts adequately.

The white flecks visible in some places are I think the ivory warps peaking through.

This second direct scan of a bare area from the front and wefts may be seen better in it.

I don't see flecks of white in them.


R. John Howe

Posted by Steve Price on 10-09-2005 07:43 AM:

Hi John

You ask, "Could this be an ensi?" I guess it could be, but I see no compelling reason to conclude that it is one. The knotting - asymmetric left - nearly eliminates most Turkmen origins except Salor or "Eagle group Yomud", and it is pretty clearly neither of those.

Are you sure there aren't some missing borders along the sides? The white selvage is unlikely to be original, and the horizontal minor border that separates the skirt from the other borders looks incomplete.


Steve Price

Posted by R. John Howe on 10-09-2005 08:31 AM:

Hi Steve -

The asymmetic open left knot does not in and of itself eliminate the possibility that this piece could have been woven by Ersaris. There are Ersari pieces with both level and depressed warps that have asymmetric open left knots. Loges in his chart at the end shows four, two each with the different warp treatments.

Now about the engis question: you're seeing what I'm seeing but describing it better. The initial things that might raise an engsi suspicion is the general size (Ersari engsis can be quite large) and the elem-like endings, although some purists argue that an engsi can have only one elem and this piece has this elem-like treatment at both ends. (Rueben's engsi with a mina khani field has only one elem and that at its bottom.)

But the piece also seems pretty narrow in relation to length for an engis. Here your suggestion that it might well have been wider since the yellow ground border systems above the row of elem design might indicate that this border system also extended up both sides. If so, (and I think that likely) then this piece was originally quite a bit wider and on the length to width dimension (only) the possibility of its being an engsi is increased. But this same possibility also reduces the chances of its being one since the separated "elem-like" treatment of the ends disappears if this border extended also around the sides. The piece becomes simply a rug.

So where are we?

What pieces show an asymmetric open left knot?

Salor (no one would suspect that this piece is Salor)
Some of the "eagle group" pieces

I think we can eliminate "eagle group" on the basis of knot count. Eagle groups I and III are those with an asymmetric knot open left. O'Bannon in his summary of eagle group characteristics says that group I pieces vary from 2537-4240 per sq. dm. and that group III varies from 2475-2898 per sq. dm. both of these ranges are well above the rather modest 1860 per sq. dm. we calculated for this piece.

The yellow ground meander border is often thought to be nearly signature Ersari.

But the knot count also makes Ersari questionable in the other direction. Ersari pieces do vary a lot in their knot count (in part because we have not been able, adequately, to define this group so far) but in his convenient table Loges reports Ersari knot counts mostly in the 700-1200 per sq. dm. range. Only occasionally do Ersari pieces in his table rise to the 1500-1700 per sq. dm level and only one piece, that Loges includes, reaches 1800 dm. My sense is that this piece is likely too fine to be an Ersari.

So we seem to be pressing our selves toward Arabatchi, but there are problems here too. Arabatchis are reputed usually to have silk and/or cotton in their wefts and we seem not to have any here.

So what do we come to? Balouch, I suppose, might be defended, but I'd want to hear a better argument for it than what I've heard so far. Or are we forced to retreat to "Middle Amu Dyra" (that new term of uncertainty) or even to "Central Asian" indicating that we really don't know?


R. John Howe

Posted by Walter Davison on 10-09-2005 11:52 AM:

Hi Mr. Howe,

In your analysis, you are relying heavily on the structural aspects of this rug. But what about all those pictures above of the Beshire design that this rug exhibits? Why is it that you are ignoring - maybe the word ignoring is not correct - or at least not taking into account more the design features of this rug? It does not appear to me that you have addressed this point. Aside from that, are not the Ersari made up of as yet not completely well-defined subgroups that perhaps made more densely knotted rugs that would compare to this piece? And if that is so, why would an Ersari attribution be ruled out on that basis?


Posted by R. John Howe on 10-09-2005 01:17 PM:

Hi Mr. Davison -

Steve Price chides me from time to time about my tendency to treat structural factors more importantly than similarities in design.

I do do that (its far more costly for a weaver to change structural things than it is for her to weave a new design, so structural characteristics tend to be more stable and to change more gradually) but I also acknowledge that design indicators can be important sometimes.

One of the problems with some of the general tribal designations that we currently use (like Yomut and Ersari especially) is that they are likely far too broad and we're not at all sure of what they are composed. People have, for example, begun more frequently, and a little self-consciously, to say "Yomut groups" rather than "Yomut."

And although it seems likely that "Beshiri" is a more urbanized kind of "Ersari" (there is a town with the name Beshir) the distinction has not yet been sorted out. Robert Pinner and Elena Tzareva were actively working for a time on a book on "Beshiri" weaving, but Robert died, and I am unclear how that work stands.

More recently some experienced collectors and students have begun to say such things as "not Ersari" without saying much further about what it might be instead. "Lakai" too has been criticized as being far too widely applied. And recently on Turkotek, Andy Hale suggested that some saddle covers that some of us have blithely been calling "Beshiri" are probably instead some sort of Afghan production. And I referred in my posts so far to another recent tendency to say that something is probably "Middle Amu Dyra" without going further.

This is because nearly all the Turkmen tribes went through there at some time or other, as did lots of other non-Turkmen Central Asian folks. This "Middle Amu Dyra" usage seems to substitute a likely geographical attribution for the more frequent tendency since the 70s to make attributions in terms of tribes. Apparently, the Amu Dyra polygot is still too rich to sort out, although the indicators for including something in it are not entirely clear to me. (I have jokingly suggested that is it suspeciously close to such usages as "Shiraz" for south Persian rugs or even to "Bokhara" (horrors!!!) for Turkmen ones. I am not at all sure that "Middle Amu Dyra" is a good move. It changes the base indicator without seeming to offer much evident advantage.)

So part of the reason I do not respond to many images that might seem to have designs on them similar to some on my piece is that I'm trying to go as far as I can go with the structural factors first, since I think they are more stable.

Another reason is that my own tendencies to see design similarities are quite closely restricted, and lots of folks here see useful comparisons far more broadly than I do.

I think we have to be careful rather than adventurous with design comparisons. At some point nearly everything can be said to be similar, on some basis, with nearly anything else and when that occurs one of the chief functions of comparison (to distinguish) evaporates.

That is not to argue that my own view is correct, only that it is different.

On the other hand I have not, in this analysis, entirely ignored design factors.

I have referred explicitly to the yellow ground minor meander border, acknowledging that it is often seen to be a nearly signature Ersari usage. It does not in this instance seem to be supported by the knot count indicator, but some might argue persuasively that it may strongly suggest that this piece was woven in the Middle Amu Dyra area where the Ersaris predominate. If that is accepted, the question becomes (always probablistically) what group that we know was in the Amu Dyra is the best candidate to have woven it and why do we think so?

I also treated the "elem ends" seriously as possible indicators of format (an engsi).

Everything is an estimate based on one or more indicators. I'm just seeking the best confluence of indicators I can find and a free form comparison of design features does not seem particularly hopeful to me. But I invite anyone else who wants to make such an argument.

Does that help explain?


R. John Howe

Posted by James Blanchard on 10-09-2005 01:37 PM:

Hi all,

I hope you will forgive some more novice musings...

I recall a discussion about an "Ersari" rug that I showed on Virtual Show and Tell in January, 2005 (some might remember the discussion about stencilled numbers on the back, and Vincent Keers' sleuthing through old log books in Amsterdam). Based on Vincent's investigations, the rug was established to be from 1936 or earlier.

Anyway, here are a couple of pictures of the rug again...

The knotting on this rug is also asymmetric, open left. The warps are ivory, with some depression. The wefts are brown. It is not as finely knotted as John's, averaging 8.5h x 11v (approx. 90-95 kpsi).

Why might this be relevant?

Vincent Keers made the following comment on this rug...

"The white warps make it less Afghan, more workshop and maybe Iranian (Turkman) production on demand."

Perhaps Vincent could elaborate if he is looking in on this discussion, but if there was Iranian (Turkman) production of "Ersari design" rugs with ivory warps and asymmetric-left knotting....


Posted by Walter Davison on 10-09-2005 03:23 PM:

Hi Mr. Howe,

Yes, thank you. That helps a lot, and that makes sense.

But, ... you write that "'s far more costly for a weaver to change structural things than it is for her to weave a new design, so structural characteristics tend to be more stable and to change more gradually.."

OK, now that seems correct to me, the "far more costly" part. But do we know, and I certainly don't, whether it was commonplace for rugs of dramatically different designs, 50, 60, or a 100 years back, to be produced by a village woman, perhaps in a tent, located at the end of a dirt track if any track at all, to be made by a weaver who might and did ever switch designs? Can we know this?

I confess zero experience in this area, but it looks to me that this might be a tough question to answer with any great confidence. Could you address this question for me? Has such design switching been documented?



Posted by Steve Price on 10-09-2005 03:55 PM:

Hi Walter

As many of you know already, I'm pretty skeptical about the supremacy of technique as a criterion for attribution. There are a number of reasons for this, let me mention just a couple here.

1. Technical criteria arose when it was noted that groups or rugs with similar designs, layouts and palettes had common structural characteristics as well. That is, the epistemology of the structural criteria is such that their basis is design, layout and color criteria. There are exceptions aplenty, but in general, this is accurate.

2. One route of design and layout diffusion was intermarriage. A woman who married into a "foreign" culture was likely to weave with the designs, palettes and layouts typical of her new culture. She was unlikely to change the way she worked, though. If she learned asymmetric knotting as a child, I doubt that she began knotting symmetrically after marriage. Thus, a Chodor woman marrying into a Yomud community would probably wind up weaving Yomud-like things, but with asymmetric knots open to the right rather than with the symmetric knots more usual in Yomud work. In such cases, the structural criteria would mislead attributions.

In fact, we recognize technical variations within a community. For example, it is pretty much agreed that a relatively small percentage of weaving done with asymmetric knots can be attributed to the Yomud primarily on the basis of design, layout and color.


Steve Price

Posted by R. John Howe on 10-09-2005 06:45 PM:

Hi Steve -

You wrote in part just above:

"As many of you know already, I'm pretty skeptical about the supremacy of technique as a criterion for attribution."

My thought:

I don't know anyone who takes that position. It seems like a kind of debating class technique that mistates a position the better to oppose it.

My own sense is that the more frequent claim is that structural indicators are often better indicators for making attribution estimates, but few, if any, I think argue that they should be taken alone or that design can never be a deciding indicator.

If design is to be seen as the usually reliable basis for making attributions should we begin again to suggest that What we now call Saryk or Tekke or Ersari or even Balouch pieces with turreted similar looking "Salor" guls should be reconsidered, after all, as likely instances of Salor weaving?

You say that the source of technical analysis began when folks saw that weavings with similar designs had similar structures. That surely happened, but it is also true that it began when people also noticed that weavings with quite similar designs sometimes had different structures.

Would you recommend in the current instance that we examine minutely the similarities and variations in the various mina khani designs posted here or that we could collect? And, if so, what might we conclude from that?

My own position has always been that I tend to weight structural factors more heavily than designs when attempting to estimate a attribution, but it's not a one-variable "structure is supreme" argument.

I think it happens that design indicators can be sometimes be nearly conclusive. But even then design usually does not operate alone.

The first Turkmen chuval I bought, when I began to collect (somehat more) seriously has guls that are identical to those on some Tekke chuvals and has an asymmetric knot open right. There is a very similar piece here in the DC area that is pretty clearly Tekke (Plate 30 in Mackie and Thompson, 1980).

So there was a chance that mine might be Tekke too until George O'Bannon pointed out that there are no known Tekke pieces with the border that mine has, but there are a goodly number of Yomut group chuvals that do and some Yomut pieces have asymmetric right knots as well. So the design was conclusive as long as the knot was not disqualifying.

Now this piece may still be part of some as yet undefined Yomut group, but I have never, since learning that it's border is not in the Tekke vocabulary, suspected that this piece is Tekke.


R. John Howe

Posted by Tim_Adam on 10-09-2005 07:02 PM:

Hi Walt,

I think the fact that tribes copied designs from other tribes, e.g. Balouch from Turkmen, or copied designs from city rugs pretty clearly indicates that weavers experimented with new designs.


I am not sure I understand your points against structural analysis. Are you saying that structural features don't provide much new information beyond design, layout and color criteria?

On your second point, it is my understanding that inter-tribal marriages were extremely rare among Turkmen. And how an inter-tribal marriage would affect a woman's weaving style is highly speculative. Or are there any historical sources on this?



Posted by Steve Price on 10-09-2005 07:56 PM:

Hi Tim

Most of the time, structural information adds very little beyond what can be said on the basis of design, layout and palette. If that weren't so, it would be almost impossible to make an attribution on the basis of an image on a monitor, and it is usually possible to do that with a very high degree of certainty.

I don't know what the frequency of intermarriage was between the Turkmen tribal groups, nor do I have any hard information about how it affected their weaving when it happened. I doubt that anyone else has hard information, either. It is generally assumed, though, that the technique a woman was taught as a child went with her (in John's words, earlier in this thread, "'s far more costly for a weaver to change structural things than it is for her to weave a new design ..."). The conventional wisdom seems to me to be likely to be correct in this case.

Hi John

My wording was poor in saying that some consider technical criteria to be supreme. But I think it is accurate to say that most collectors and dealers give the nod to technical factors when faced with a dilemma.

The Belouch pieces with turreted guls would never be mistaken for Salor work despite the superficial similarity of the guls, even in a photograph where technical information was absent.


Steve Price

Posted by R. John Howe on 10-10-2005 06:39 AM:

Hi Steve -

You wrote in part:

"...The Belouch pieces with turreted guls would never be mistaken for Salor work despite the superficial similarity of the guls, even in a photograph where technical information was absent."

My thought:

You alertly pick the weakest point in my comment, but I think you know from some pieces you own yourself that Belouch weaving can be formidable.

You have in fact suggested that this piece of mine might well be Belouch, despite its also looking much like Turkmen weaving to some of us.

Now it is true that the mina khani design likely has Persian roots and that both the Turkmen and the Belouch are weaving it "second hand" so to speak. So maybe it is not a good test of whether a Belouch rendition of some Turkmen devices might deceive looked at at the design level alone.

But I happen to know of one piece (a torba with two large turreted "Salor" guls on it) in a major collection that has (admittedly after some alternative considerations) been attributed to the Belouch. The chief alternative considered was Afghan Ersari.

So it's not always easy once one starts down the slippery slope of suggesting that designs are usually adequate for attribution decisions.


R. John Howe

Posted by R. John Howe on 10-10-2005 07:24 AM:

Dear folks -

Several of us have offered suggestions about who might have woven this piece and we have in the process explored a number of indicators that might be used to determine this.

Perhaps now I should indicate what the person from who I obtained this piece thought it likely is. We've had several exchanges and I'll give you his actual language as closely as I can.

First, I should say that this dealer is a long-time student of Balouch weaving and hasn't suggested at any point in our conversation that he thinks that this piece might have been woven by Balouch weavers. That for those with that suspicion.

But here is what he thinks this piece likely is. First his description of it:

Materials: wool with camel wefts (ed. I can't detect this)

Structure / Technique: pile, asymmetric knot open left

Condition: Poor

Comments On Condition: low pile, holes, skinned areas, fold wear, and hardened skid inhibitor on the back.

Full Description: wrong dimensions for an engsi, perhaps a funerary rug? Good age and great color.

Next he draws on these indicators and some knowledge of his own to make an attribution argument:

"Here is a sketch of my reasoning.

"Early Russian maps and two German travelogues mention a group of Arabatchie situated near the Ali Eli in the south middle Amu Darya region.

"This group was connected with the main Aral group until the Khan of Khiva again became powerful.

"Part of his policy was to divide and remix the Turkoman tribes in his realm to keep them from forming alliances that might jeopardize his power. Isolating the southern Arabatchies was part of the program.

"I have encountered a number of weavings from this Amu region that have a thin, long, knot almost identical to the Salor knot, though not depressed. These weavings are open left.

"Rarely is cotton seen in them but then cotton wasn't grown down there...only in the north.

"Since only three tribes use this type of knot (ed. he means this long, thin shaped open left knot); the Salor, early Imrelli ("Eagle"), and the Arabatchie, I strongly suspect that pieces similiar to the one you are interested in are from the latter tribe.

In a subsequent message he added:

"I don't think that the south/middle Amu Darya group has cotton. I don't think I've ever seen it in their wefts and only once or twice, in exceedingly small amounts, I've seen it as pile.

"The thing is, that they were there and the knot you see in these weavings is the same sort of long, thin knot that we see in their northern cousins."

So there we have this dealer's thoughts. He works with his knowledge of the history of the Amu Dyra area, including who was where when. What moves the local politicos made that could have affected tribal movements. He has an explanation for the seeming absence of cotton (usually an Arabatchi indicator) and he hangs his argument heavily on both the character of the knot and its width and shape.

Do you find this argument convincing? It seems a little tenuous to me, but it's interesting.

Further frank comments are invited.


R. John Howe

Posted by Steve Price on 10-10-2005 07:24 AM:

Hi John

Your concluding remark, which I think is the "take-home" message in your post, is "So it's not always easy once one starts down the slippery slope of suggesting that designs are usually adequate for attribution decisions."

You're right, it isn't always easy to make attributions with confidence on the basis of designs, layout and palette (I don't think I ever suggested that designs were sufficient; I surely don't believe that they are). But this doesn't negate the fact that designs, layout and palette are usually adequate. And by "usually", I mean, more than 99% of the time. That is, I think less than 1% of the antique rugs and related textiles from western and central Asia present attribution difficulties that require more than a photo of the front of the piece. Indeed, the most common practice is to make attributions by comparing one or more photos of the piece with photos in published sources.

As a related side issue, there has been a recent exchange of unpleasantries on line between two self-proclaimed experts. The dispute was whether one of the rugs offered at a recent major auction was made in Talish (as catalogued) or somewhere else in the Caucasus (I don't remember whether the proposed alternative was Genje or Karabagh, and it doesn't really matter for this discussion). My recollection is that neither "expert" had handled the rug or even seen it in the wool (although the auction house person who wrote the catalogue entry had, obviously), and the debate revolved entirely around design. Neither party even raised the question of supplementary wefts at the edges, a feature nearly universal to Talish textiles and nearly unique to them. In that instance, a techical detail could have resolved the question to my satisfaction, although it got pretty well narrowed down by the color photo.


Steve Price

Posted by David R.E. Hunt on 10-10-2005 04:21 PM:

Arabachi? = Ersari Type?

Hi John ,Steve, All

It seems, as I had expected, that this rug does not clearly fall into any specific category. Looking back I noticed I had failed to make my attribution, but as the content of my post suggests, I had assumed it to be of the Ersari Type.

I had arrived at this conclusion based upon the color, especially of the yellow ground "judor"border (granted I had suspected that I might have been camel hair), and the design in general, which has much affinity with the Ersari chuval depicted. True, the petal arrangement may have been of the Belouch ( bearing the scope of my experience in mind ) but that double judor border says Ersari to me. For what it's worth, I think we would do well to consider attribution as a process proceeding from the general, as in color, design and materals, to specifics of structure.

If find this study of the history (and topography) of the Amu Darya fascinating for it's implications regarding Turkmen weaving.

Oops, gotta run, will continue this line of inquiry in a while.


Posted by David R.E. Hunt on 10-10-2005 05:41 PM:

Amu Darya as Pipline of Commerce

Hi John. All

Find below a map of the Turkmen regions from Khalter's "Art and Crafts of Turkestan",

making note of it's origins in the Pamir mountains, it's passage through the desert and adjacent to Bokhara in the middle region, and it's terminus as a delta feeding the Aral Sea. The following is a brief synopsis of the Amu Darya from Encarta.

Amu Darya (ancient Oxus; Russian Amudar’ya; Turkmen Amuderya; Uzbek Amudaryo; Tajik Dar”yoi Amu), largest river of Central Asia. The Amu Darya is formed by the junction of the Pandj and Vakhsh rivers in the Pamirs mountain region, on Tajikistan's southwestern border. The river measures 2,540 km (1,580 mi) in length. It follows a northwest course between Tajikistan and Afghanistan, continues northwest between Turkmenistan and Uzbekistan, and then flows north through Uzbekistan into the Large Aral Sea (Russian Bol’shoye Aral’skoye More), which separated from the northern Small Aral Sea (Russian Maloye Aral’skoye More) in the late 1980s. The Amu Darya’s main tributaries are the Panj and Vakhsh rivers, which both rise in the Pamirs. The Panj forms part of the boundary between Tajikistan and Afghanistan, and the Vakhsh flows through southwest Tajikistan to join the Amu Darya at the Afghan border. Since the 1950s the Amu Darya has been heavily tapped for irrigation, which has greatly reduced its water level and the amount of water reaching the Aral Sea. During the 1980s several years passed in which little or no water reached the Aral Sea from the river. Inflows from the Syr Darya River, which empties into the Small Aral Sea from the east, have also drastically diminished in recent decades. As a result, the volume of the Aral Sea dropped by about 80 percent between 1960 and 1996. The largest single cause of the decline in the Amu Darya’s water level is the Garagum Canal, the longest canal in the former Soviet Union and one of the longest in the world. Near the town of Oba the canal diverts water from the river at the rate of about 12 cu km (about 5 cu mi) per year—about one-ninth of all the water diverted in the Aral Sea basin. Reduced water flow has restricted water transportation on the Amu Darya, which was once navigable for light draft vessels over nearly half its length. The lower reaches of the river once contained a large delta that supported extensive vegetation, but most of the delta has dried up due to reduced water flow. Over the centuries the river has shifted its course several times. In the 3rd and 4th millennia bc the Amu Darya flowed westward from the Khorezm Oasis into Lake Sarykamysh, and from there to the Caspian Sea. From the 17th century until the 1980s the Amu Darya emptied exclusively into the Aral Sea, except during periods of intense flooding, when overflows went into Lake Sarykamysh.

For now let's concentrate upon the rivers signifigance as a major conduit for the transportation of goods and products in the Turkmen regions, being the largest river of Central Asia, and some possible implication for the carpet industry. I would assume that the river could/would have been used to move raw carpet materials, such as wool and dyestuffs, as well as carpets, along the middle and lower reaches of the Amu Darya, which are nagivable by larger commercial vessils, and possibly the upper extremity in which rafts might be plied ( as they are on some other mountain rivers in this area). All conveniently leading to Bokhara.

In a contemporary thread here on Turkotek Significance of color variations we are discussing the origins of yellow dyes( among other things), and in the process came across a discussion of the yellow dyes in Turkmen carpets in the Whiting essay in the back of the Textile Museum's "Turkmen"

"Seems the primary suspects in Turkmen weaving (at the time of printing) are isparuk (Delphinium sulphureum), which grows wild in Afghanistsan, and weld (Reseda Luteola L.) which is an "essentially cultivated plant. Tekke rugs, both older and newer, which "retained a good green" demonstrated properties of weld dyes, and those exhibiting this green "fading out to blue", of isparuk. Conclusion? First, isparuk is prone to fading, and weld more fast. Second, that isparuk was rarely "less than an important ingrediant except when replaced by weld.". And third, the weld was obtained from setteled people with which the Turkmen were on good terms. Now under what conditions and hence locations does one grow weld?"

What does all of this mean? In short, that dyes, wool and rugs could have been shipped along the middle and lower Amu Darya with ease, possibly facilitating the transmission of design patterns (between Yomud and Ersari) and color selections (especially that of this cultivated Weld) in both directions, up and down river. From the upper extremity of the Amu Darya, transportation would be one way, from higher to lower elevations,say from Mazar-e Sharif or Andkhoi (and hence these indigenous Ersari patterns and color schemes) to Bukhara and points north.

Now for this transmission of design from Ersari to Balouch, I submit the following, from the Middle Amu Darya weaving thread here on Turkotek.


Posted by Tim_Adam on 10-10-2005 06:39 PM:

Hi John,

I am a little lost now. From what I could gather it is only the relatively high knot count and the dealer's opinion that speak in favor of an Arabatchi attribution. Or am I missing something?

I find it rather unconvincing to rule out an Ersari attribution based on a knot count that it is higher than the range provided in Loges' book, especially since there are other indicators that speak pretty strongly in favor of Ersari. After all, Loges' range is based on a relatively small sample of rugs. And the color scheme of your rug is completely different from any piece I have seen that has been attributed to the Arabatchi.


Posted by R. John Howe on 10-10-2005 10:05 PM:

Hi Tim -

The attribution of this piece can definitely still be debated.

I think most people would be tempted to say Ersari on the basis of the yellow ground minor border alone, since that is seen as signature Ersari.

It is true that Loges offers only three Arabatchie pieces, but the knot count in mine centers nicely in their range and you are welcome to find and cite Arabatchie's that do not. And most Ersari pieces are quite a bit coarser. Most are in the 50-60 KPSI range and the center of the admittedly quite wide range of Ersari pieces is quite a bit lower than that indicated for Arabatchis. So I think the knot count difference is not something to be ignored quickly.

Lots of Arabatchie's are pretty ugly, but some (look at the famous Ballard chuval in Mackie and Thompson, Plate 54 and some of the examples Jourdan offers, Plates 202 ff.) can exhibit better, clearer reds. So a better color palette isn't necessarily disqualifying.

The lack of cotton in the structure is a departure but this dealer offers a plausible (who can say if it is true?) explanation for holding on to his Arabatchie attribution despite its absence.

I also think this dealer's noticing the shape, length and thinness of the knot nodes and what other knot nodes resemble them pretty careful and sophisticated.

Now, if you insist on seeing this piece as Ersari rather than one made by lower middle Amu Darya Arabatchies, look at Jourdan's indication on page 228 and Elena Tzareva's parenthetical comments on all of her Arabatchi pieces. As Jourdan notes, Tzareva explicitly indicates that she sees Arabatchis as a part of the Ersari group. That might make this dealer's Arabatchie assertion less difficult to accept.

You can have your attribution cake and eat it too.

Last, let me play Steve's design-guided attribution game a bit, but in a way he probably wouldn't accept.

I own a few Ersari pieces, including one chuval fragment with a mina khani design and lots of silk pile decoration. One thing that struck me about this piece I have just bought is that it didn't quite "look" Ersari to me. Spooky, huh? And maybe wrong.

Examining my feelings, I noticed that there are odd things about the drawing of the mina khani field in this piece. The drawing "decays" a bit at the sides and there is an odd drawing of some of the central mina khani elements. And although David Hunt supplies some instances of rather stiff drawing in Ersari elems, the overall effect of the drawing in this piece had a kind of stiffness about it that I experience as stately rather than just static, and distinctive from what I see in Ersari drawing (I suspect this latter is part of what people who said "maybe Balouch" are seeing too). Now do any of these "doesn't look quite Ersari to me" feelings indicate that it is Arabatchi instead? No.

Maybe I should just say that I suspect that it was probably made in the Amu Darya area (minor border and asymmetric open left knot), but that it doesn't quite look Ersari to me.


R. John Howe

Posted by James Blanchard on 10-10-2005 11:11 PM:

Dear John,

This Discussion is living up to its name... quite a puzzle. The explanation from the dealer seems to point to a particularly idiosyncratic attribution construct that will continue to frustrate those of us who like to master information, and will delight those of us that still like a bit of the unusual!

Regardless, I very much like the rug.

But I do have a few residual questions that are based not on expertise in this area, but rather on some personal observations about the deductive process for attribution.

Please forgive me for dragging us all back to my "stencilled Ersari", but I think it is relevant since it does seem to share some features with yours. It has design features which could really only be called "Ersari". In my hands, the colours and wool seem to be of very high quality. It has ivory wool warps and brown wefts (wool, I am persuaded). The knotting is asymmetric-left and finer than is usual for Ersari, though certainly not "out of range" like yours. I really don't know what is meant by the characteristic "long thin knots" mentioned by the dealer, so perhaps you could elaborate for my benefit. I do see similarities between the knotting on your rug and mine (especially looking at the yellow knots in yours). I don't know the exact age of mine, but although its condition is "nearly mint" it is not "new", having been sold once in 1936. You haven't mentioned an age estimate for yours. One last similarity -- I purchased it from a dealer who has quite a bit of experience in the region, and held the rug in high regard (mostly for the high quality wool and colour). He even nailed the "stencil" date right on the head. He attributed my rug to the Ersari, though he felt that the structure and particularly the size and shape were out of the ordinary.

In discussing my rug, only one person offered a "non-Ersari" attribution. Vincent Keers suggested "Iranian (Turkman) production on demand". I am certainly in no position to refute or affirm that assessment. I have since simply considered it an "Ersari" rug based on design, if not attribution.

I have never entertained a "Middle Amu Darya Arabatchi" attribution for my rug. Should I now? If not, why not?



Posted by David R.E. Hunt on 10-10-2005 11:36 PM:

Arabachi vs Ersari Part II

Hi John, Tim All

I just noticed that we have not one image posted of an Arabachi weaving, so witness the following Engsi, Plate 205 from Jourdan

and the next, from Turkotek's The Turkmen Engsi : Doorway To Paradise salon.

Followed by the rug in question.

Jourdan makes some assertions regarding Arabachi weaving including;

1) very distinctive weavings

2) asymmetric left knotting and wool/cotton plied wefts as sine qua non

3) ribbed backs and scattered flecks of white cotton

4) "camel train" in elem typical for Arabachi engsi

and to judge from what we have seen so far, it fails on most counts.

Or does it? We do have this tenuous yet plausible explanation for the absence of cotton (besides, I wouldn't be suprised at the occasional Arabachi sans cotton in the wefts, experience suggests weavers simply use whatever is available), and a close proximity to a tribe notorious for their use of this "Judor" border, the Ali Eli ( Jourdan also cites the kochak border found in Arabachi plate #204 as an Ersari usage). Number of colors seems about right, and their does seem a kindered use of color in regard to the engsi above. Yet the open left vs. open right knot structure,while a rather striking departure from the "typical" (whatever that is ) Ersari, is hardly unknown. Maybe this is much of the problem, our inability to further qualify and quantify our definition of Ersari. If not for the existence of considerable numbers of high Kpsi open right Ersari, I would be inclined to find this high Kpsi characteristic in of itself more persuasive. And of course the Arabachi are, or in the least some authors contend, Ersari.

For myself it really fails the engsi qualification, as both the dimensions and design don't to conform to the standard (in the least to that degree with which I am familiar ). Now these dimensions,which I think worthy of further comment, say as much about this piece as does the use of the mina khani field, and both point me toward modern as opposed to traditional production. The size strikes as more the 4x6 area rug, and I strongly suspect that, as Steve stated earlier, the outer "judor" guard border is missing. The mina khani field seems a standard of modern Turkmen production, as a visit to the local flea market will attest, so the size, combined with the numerous borders and what strikes as a more modern field design says to me early commercial production. Early as in 19th century

It does seem to share several characteristics with the Arabachi, and I wouldn't argue with an Arabachi? designation (rememberimg that I am proceeding from photographs )but just as the design somehow doesn't "look " Ersari, if the design was a little more traditional I would be of greater inclination to designate as Arabachi. Or perhaps a hybrid?


Posted by Steve Price on 10-11-2005 06:04 AM:

Hi People

It might be worth reminding ourselves that attributions are rarely statements of fact, but statements of probabilities. Sometimes the probability of a particular attribution being correct is high enough to leave us comfortable ignoring all alternatives.

Clearly, this rug is not in that category. It has lots of Ersari characteristics, but some that are atypical for Ersari. The dealer who sold it is able to present arguments for Arabachi origin that, while not compelling, are at least plausible. In the absence of anything more concrete than the fact that someone regarded as an expert on Belouch weavings didn't mention Belouch among his possible attributions, I see no reason to eliminate Belouch as an origin. The Belouch wove so many Turkmenoid rugs that, until fairly recently, many collectors thought the Belouch weavings were all derivative.


Steve Price

Posted by Rob van Wieringen on 10-11-2005 06:39 AM:

Hi All,

For me this John Howe rug is just a normal Ersari ( Beshir, Amu Darya ) rug.
The fineness of the knotting and the openness to the left does not contradict at all with a Beshir attribution, while the design and especially the colors are so Beshir as Beshir can be.

The speculation that it could be " Arabatchi " or " Baluch" is in my opinion hilarious speculative and seems in essence based on a mythical " long, thin knot ".

Would love to see one,

Best regards,


Posted by R. John Howe on 10-11-2005 08:00 AM:

Hi James -

You ask first should you reconsider the indications you have been given and your own impression that your piece below is Ersari.

I think not. All of the design features in your piece are part of the Ersari vocabulary, the knot count is lower and there are some Ersaris with an asymmetric open left knot.

You also ask for a better description of what is being pointed to when this dealer says that the knot is "long and thin" and resembles closely the shape of the knots not only in northern Arabatchie weavings, but also some "eagle" group pieces and the Salors (although he makes clear he is not considering a Salor attribution at all).

And such a description may be useful because Rob van Wieringen sees this dealer's rationale for his attribution "hilariously speculative" and depending too much on a "mythical ' long, thin knot ' ".

Now I think this attribution is a shade aggressive and it is visibly speculative, but I think we should refrain from doubling over in laughter about it.

Knot node size and shape are an established aspect of technical rug analysis. The characteristics of knot nodes are a part of "weave pattern" proposed and discussed seriously by Neff and Maggs, in 1977 in their "Dictionary of Oriental Rugs."

Although used widely by dealers and people in the trade (they will frequently quickly look at the back of a piece and say, often with surprising precision, where it was likely woven) it has not caught on much yet with rug collectors and scholars (you do not often see descriptions of knot node shape in even quite detailed technical analyses; it does not, for example, earn a place on Marla Mallett's guides for analysing pile rugs) but one does sometimes hear reference to it by experienced people.

But some authors have taken it seriously enough to begin to supply images of the backs of their rugs (Willborg's and Runge's catalogs on Hammadan are the readiest examples but Neff and Maggs presented a number of examples in 1977).

My own view is that there is likely something to knot node shape and size as an attribution indicator, but that we just haven't been looking at it long enough for the differences to be recorded, systematized and related to other attribution indicators that we use. But there is nothing "mythical" about the shape and size of a rug's knot nodes. They are very much "there" if you want to examine them.

But to try to answer James' question about what is a "wide, thin" knot. And why might its presence supply less than laughable support for this dealer's attrubtion estimate?

When a weaver ties a knot around two warps, she is doing so with a strand of wool that (not to put too fine a point on it) has a given width. That this width can vary seems beyond debate. After a row of knots has been completed a weaver beats them down with a comb tool. Then one or more wefts are added with additional beating after the insertion of each one. The heaviness and intensity and length of time that this beating goes on compresses the (now tied) knots in varying degrees. That this compression will also be a variable also seems to me beyond debate.

The result of such variations is that when one looks (closely) at the back of a pile rug one will be able to discern (even if magnification is needed) what the shape and size of the two knot nodes that together make up each knot are.

When we looked at my my yellow-ground Caucasian wreck in another thread here I joked (but also not quite) that I might hang a Kazak attribution primarily on the fact that knot nodes looked to me more than twice as high as they are wide. I was drawing on the indication by Neff and Maggs that one of the main features of Kazak weave pattern is that Kazak knot nodes are visibly much taller than they are wide.

What this dealer has noticed is something like the reverse of that.
He says the knots are wide and thin. That is, the wool strands used to make them were noticeably narrow and/or have been tightly packed down by beating. This results in a knot node that is quite short vertically. He also notes that the knots are wide. This is a function of the size and the closeness of the warps. The resulting shape he is talking about is one in which each knot node is wider than it is tall and this sense of width is accentuated when one looks at the two knot nodes that comprise each knot.

Now Rob's laughter about this suggestion has been functional for me because he made me look more closely at what Neff and Maggs say about their two Ersari examples (they offer an "Afghan" example and a classic "Beshiri" prayer format). Here is their description of both of these Ersari knot nodes, which they explicitly say are the same. "On the vertical line, the nodes of the asymmetrical knots are longer than they are wide."

Reading this description of Ersari knots side by side with this dealer's description of my just purchased piece I want to suggest, however modestly, that the knot nodes on mine are different precisely in that the individual knot nodes are wider than they are tall and the tallness itself is quite short. This dealer's description of this lack of height as "thin" seems appropriate.

Now this is just one factor and this dealer's attribution is, I think, still very much in debate, but the knot node size and shape on my piece, ARE different from what Neff and Maggs say is more typical for Ersari pieces.

I have two last questions for Rob.

Please give me your indicators for a "normal Ersari (Beshiri, Amu Darya)" rug?

Do you not distinguish Ersari from Beshiri, and if not why do we need two terms?


R. John Howe

Posted by Rob van Wieringen on 10-11-2005 11:46 AM:

Hello John,

I am sorry if I upset you by my remark; it wasn't my intention.

Of course knot nodes are there and there variety of forms is something not to neglect when studying rugs etc.
My point of view, however, is that the colors and the design in this rug are allover Beshir ( as part of the Ersari-group, settled in the Amu Darya valley ) and rules out any other possible attribution. For me especially the colors, as seen in the close-up, are a very convincing indicator.
The focus on a supposed abberation in knot node form ( and at the same time neglecting strong indicators as colors and design ), seems to me a very thin line to attribute otherwise.

It is like a dog with six legs : I would call it a dog with six legs, but I get the impression the rationale of the dealer would make him to call the dog a "possible spider".

In my opinion your rug is a 19th. cent. Beshir with long, thin knot nodes.



Posted by R. John Howe on 10-11-2005 06:26 PM:

Hi Rob -

We don't get "upset" per se much here on Turkotek. We do engage in vigorous debate, if we feel that is called for. But the future of civilization does not rest on the outcome of our discussions here so no harm is usually done.

You are not alone in insisting that this piece is Ersari. I have Americans traveling in Europe writing me on the side to say that.

But those claiming most vigorously that the color palette and designs indicate pretty conclusively that this piece is Ersari (and it might well be) are oddly elliptical.

Their conclusion is firm, but the evidence on which it is based is described in a conclusionary way without showing itself for examination.

What is a "typical Ersari palette?"

What are the "typical Ersari designs?" (I'll give you the yellow ground minor border.)

Some recognitions are beyond language, but we should at least try.


R. John Howe

Posted by James Blanchard on 10-11-2005 11:08 PM:

A picture's worth...

Dear John,

Would it be possible for you to post a clearer (brighter) picture of the knot nodes of the "long, thin knots"? Since I am sure that many Turkotekkers have a variety of Ersaris in their own collections it might help to assess how extraordinary the knotting technique on your rug is.



Posted by David R.E. Hunt on 10-12-2005 02:33 AM:

City Dye, Country Dye

Hi Folks

Is it just me, or does the yellow in John's rug seem lacking in intensity for an Amu Darya or Beshiri weaving? I had suspected it might be camel hair, to tell the truth. I don't know, the colors overall seems a little murky for above said.

It seems that misperception often proceeds from the confusion of weaving habits of the past with those more recent. As an example, I had noticed a passage ( from the T.M.'s "Turkmen" I believe), which had asserted that the mina khani pattern is frequently seen in Amu Darya weaving, which is of course true, but does not negate the fact that in more recent times this field pattern has figured prominently in the repetoir of Afghan Ersari (hence upper Amu Darya) weaving. Perhaps a characteristic by which one could distinguish between the two is the quality of and use of color? I see this confusion regarding the habits of older weaving culture and the more modern in our discussions constantly.


Posted by R. John Howe on 10-12-2005 11:21 AM:

Dear folks -

I have received, on the side, two email messages from Richard Isaacson.

Richard, some of you will know, is a serious collector and student of Turkmen and other Central Asian materials, and has, explicitly now, undertaken some scholarly study of them.

He has given me permission to quote what he said.


"You might be interested to hear that ALL Mina Khani pieces of this design group that I have ever seen have knots which are asymmetric, open left. This applies to 18th and 19th century main carpets as well as bags.

"The colors couldn't possibly be Arabatchi or Salor. This is an as yet unidentified Middle Amu Darya group, usually lumped in with the Ersaris in most books.

"Elena's forthcoming book may attempt to sort this out."

Then in a subsequent email he adds:

"As far as the colors go, I am guessing from pictures. However, Arabatchis have distinctive orange-reds and other bizarre shades. Your piece looks like it has more conventional color tones."

My thanks to Richard for these informed indications.

The first seems an instance of what Steve Price pointed to when he said that one way in which structural analysis began was that folks looked at particular designs and noticed that they often coincided with particular structural features.

And Richard is more specific about why he estimates that the color palette on my piece is unlikely to be one chosen by Arabatchi weavers.

He also confirmed in this second message that the Elena Tzareva book on Beshiri weaving to which he referred is the one on which she began work with Robert Pinner. It's been in the works for a long time. (She was at Pinner's actively working on it with him when my wife and I visited Pinner in (I think) 1999.)


R. John Howe

Posted by Chuck Wagner on 10-12-2005 01:46 PM:

Great Minds Buy Alike

Hi John, et all,

You are not alone, John. Recently (very recently, but before this thread was started), I acquired a mixed heritage piece with a Turkoman design having the following characteristics (some in common with Johns piece. Note: no cotton anywhere):

Size: 63 in. X 52 in.

Warp: Z4S, Almost all handspun ivory wool, occasionally a warp is comprised of mixed handspun gray & brown wool.

Weft: Z, Three thin shots Z-spun undyed handspun brown wool. The yarns are not plied. There is little no warp depression.

Pile: Asymmetrical left, Z2S, Handspun wool, various colors as in images. Knots: 7 to 8 Vertical, 7 Horizontal. Where intact, the pile is 1/2 inch long

Selvage: Animal hair wrapped around 2 cords, each made of animal hair and wrapped in the same undyed brown wool that was used for the weft.

The seller postulated that this is a product of Chodor/Baluchi intermarriage. I don't buy the Chodor connection; Chodor work is typically open to the right. This piece is open left.

Here are the images; The whole thing:

Front & back:

Closeup of back 1; the thick weft layering allows for a largely symmetrical knot layout:

Closeup of back 2; Knots were not always pounded down very well, leaving warps exposed in some areas:

Closeup showing end of rug with warp & weft detail:


Chuck Wagner

Posted by James Blanchard on 10-13-2005 03:57 AM:

Tying up loose ends...

Dear John,

Thanks for sharing the opinions of Richard Isaacson. I have a few residual questions, that perhaps you or Mr. Isaacson could address.

1. Did this "unidentified Middle Amu Darya group" weave other designs, and if so, would they be generally consistent with Ersari weavings in terms of design and palette?

2. Are there other notable structural features associated with this group (such as ivory wool warps, tighter knotting, "long thin knots")?

3. What are your conclusions regarding the opinions of the dealer (long thin knots = specific Amu Darya Arabatchi), who you indicated also has quite a bit of experience in weavings from that region?

I ask these questions mainly because I am still very much on the steep part of the learning curve and I tend to like to extract at least some "take home messages" from these types of discussions.

On the topic of knot nodes, here are three different rugs that seem to differ somewhat, though the knot counts are similar (8h by 11-13v per sq inch).... The first Ersari and the Baluch do seem to have somewhat "thinner" knots than the middle (ensi).



1. Ersari (asymmetric left, 8h by 11-12v)

2. Ersari (ensi) (asymmetric right, 8h x 11v)

3. Baluch (asymmetric left, 8h x 13v)

Posted by R. John Howe on 10-13-2005 06:08 AM:

Dear folks -

There are at three posts above to which I should respond.

Mr. Davison -

One of the things I do without exception with a new piece is to put it for a time into a freezer to rid it of any possible living moths and moth larva. The piece is now in that freezer which is about an hour and a half from where I am typing. And while I have a digital camera with great close-up capabilities, these can only be realized with a tripod and timed, self-operated shots, something I have not yet equipped myself to do. Close-up shots hand held are in my experience very chancy. But James Blanchard has given you some examples to look at and if you can find someone who has a copy of the Neff and Maggs book they provide a number of close-up pictures of back in which knot node size and shape can be examined.

Mr. Blanchard -

Richard Isaacson is not readily available for questions at the moment, but the logic of the notion of a "as yet undefined group" does suggest that there are likely other members. The problem with an undefined group is that things are still generally murky about it and group membership and characteristics might well be among them. I think Richard is acknowledging that there may be things about this piece that might not let us group it with complete comfort with many of the pieces we currently call "Ersari." I think he is also acknowledging that the group we call "Ersari" is a very large construct and we think it has (Yomut too) lots of parts, we just don't know yet what they are.

This is what leads to such usages as "not Ersari" or "middle Amu Darya." What this dealer of mine has done is to put a possible tribal name on a piece that doesn't seem Ersari to him but that he admits seems likely to have been woven in the "lower middle Amu Darya."

Richard is disagreeing and cites palette (as others have here) as one basis that. He has done a better job of describing than others so far why he thinks the colors exhibited by my piece on his monitor are likely "not Arabatchie" but more within the range of pieces we usually call Ersari.

He says that this piece lacks some "distinctive orange-reds and other bizarre shades" that appear in Arabatchi weavings. The darkish red ground of this piece IS more on the cherry side than on the "brick" side of red and there are no greens or "off colors" (e.g. purplish browns) that Arabatchis often exhibit. (I would counter that the Arabatchis are not unknown to have produced pieces with clear reds and without off-colors. Plate 202 on page 229 in Jourdan is one example.)

You ask about other features of the "group" to which Richard refers and he does not define it really at all. And he has not mentioned this dealer's claim about knot node shape and size, but that may be one reason why he says it may be an as yet undefined Ersari subgroup.

You usefully provide three examples of rug backs and say that you think you can tell something about thinness from looking at these images. Technically, the relative height and width of knot nodes would need to be measured in order to be compared usefully and even my dealer did not offer such measures. He's just looked a lots of rugs and knots for a long time and sees the height of the knots in this piece as noticeably shorter ("thinner") than Ersari knot nodes he has seen, and similar and shape to the knot nodes on two other Turkmen tribal groupings he names. What we can do, looking at these images is to describe the shape of the knot nodes. But we cannot reliably comment on their relative size.

I would describe the shape of the individual knot nodes in the top one as slightly wider than they are tall. The second piece down seems to me to have knot nodes that are somewhat wider in proportion to their length than the top example. The third piece seems to me to have knot nodes that are the closest to square although they, too, may still be slightly wider than tall.

Now about my opinion of this dealer's attribution of this piece, I think I have already given that. He is known to be aggressive in his attributions and is often quite alone. I think it very alert of him to have noticed that what he calls the "thinness" of the open left knots and to have mobilized it (coupled with some history) in an argument for distinguishing this piece from Ersari weavings with open left knots. He is tentative rather than dogmatic about his attribution and has clearly noticed the color pallette of this piece because he describes its colors as "great."

I think it an imaginative suggestion and argment (nearly all the Turkmen tribes WERE in the Amu Darya at one time or other) but I have no way of telling whether he is right. He could be dead wrong (he still actively supports Jon Thompson's "Imreli" attribution that Thompson himself has withdrawn) but I like very much his invocation of knot node. I think it is a variable that we haven't looked at much yet in technical analysis but that might sometimes help us sort some things out.

Chuck -

I would guess that your engsi with a hatchli design is most likely Balouch, but might also be (because of its open left knot) a younger Yomut piece woven by the groups who wove the "eagle" pieces (none of the "eagle " pieces, I think, are estimated to have great age, i.e. before 1850). There are no reds in this palette and I have seen both Baluch and Yomut pieces with this sort of brown. There seems also some hard camel hair used (selvege?) and if so that would be a strong Balouch indicator.

Just to demonstrate, though, that the world is not tidy, compare the border that surrounds the hatchli panels on your piece with the one that occurs in the same place on an Arabatchi (Plate 205) in Jourdan's Turkmen volume, page 232. I can't really see yours but they look possibly very similar. I'm not suggesting remotely that your piece might be Arabatchi, only that the world of attribution estimating is rather messy in some respects.


R. John Howe

Posted by Rob van Wieringen on 10-13-2005 07:27 AM:

Hi John,

With all respect John I think you make things more complicated then they are.
You pick out single elements in this Chuck's rug, such as a certain present brown color, open left knot and a single design element to refer to Yomut, Eagle and Arabatchi ( and what about the lower tree panel : Tekke? ).
I think this is unnecessary confusing. Design is fluid and much more affected by the individual interpretation of the weaver.
The main indicator, for me at least, is the total color palette of the first close-up : it reveals clearly this is a ( Ensi inspired ) Baluch, and nothing else.

Now you will ask, as you did before : what is a typical Baluch palette?
To be honest I am not able to put that in words. I can only say, well this is, as with the Beshir.
I think it is wiser to let the colors speak, then trying to catch them in slippery words.



Posted by R. John Howe on 10-13-2005 07:45 AM:

Hi Rob -

You may be right.

I asked myself (although the hard camel hair edges don't fit at all; but I can't tell if they are that in Chuck's photos) what kind of piece could have an open left asymmetric knot and this brown without red palette?

I mentioned Yomut (although I don't think it likely) because there are Yomut pieces that meet those two tests and it's important to eliminate that possibility, at least. (And the hard camel hair selveges do that, if that's what they are.)

And the Arabatchi reference was not (as I said) to suggest anything about what this piece might actually be. It was only to signal that a rather infrequent border seems to be present and in the same location on both Chuck's piece and on an Arabatchi engsi.

I take it as a sign that design indicators are very messy indeed.


R. John Howe

Posted by Steve Price on 10-13-2005 09:01 AM:

Originally posted by R. John Howe
I take it as a sign that design indicators are very messy indeed.
Hi John

Right, design indicators are messy and unreliable in the absence of other criteria for attribution. If layout and palette are included with design, the combination is almost 100% reliable. That is, they will lead to a reasonably specific attribution in terms of geography (or tribe) and date of origin of almost every piece to which they are applied, and nearly all collectors and "experts" will agree that those attributions are correct. Whether the attributions really are correct, of course, is a set of questions with ever-evolving answers.


Steve Price

Posted by David R.E. Hunt on 10-13-2005 03:53 PM:

Availability = Utility

Hi john, All

Just a couple of clairifications regarding the location of the tribes
which tenatively, as suggested by the dealer, produced your rug.

You had stated that

"Early Russian maps and two German travelogues mention a group of Arabatchie situated near the Ali Eli in the south middle Amu Darya region".

The source of the Amu Dary lies in the Pamir mountains, and the river itself passes through Northern Afghanistan on it's way to Turkmenistan, hence a roughly south- north, upper- lower relationship. So far so good. Then next we read

"What this dealer of mine has done is to put a possible tribal name on a piece that doesn't seem Ersari to him but that he admits seems likely to have been woven in the "lower middle Amu Darya."

Not that it is of such importance, but this characterization of a southern portion of the Amu Darya as "lower middle Amy Darya" does seem confused and should read " upper middle Amu Darya", emphasising the proximity to the Ali Eli in northern Afghanistan and near Mazar-i- sharif.

Next, Richard Iscaason had stated that

"You might be interested to hear that ALL Mina Khani pieces of this design group that I have ever seen have knots which are asymmetric, open left. This applies to 18th and 19th century main carpets as well as bags"

yet I am not sure how to interpret the statement "ALL Mina Khani pieces of this design group". Is this to mean pieces with the mina khani field design, or those sporting the borders and end treatments as in your carpet (i.e.,not the mina khani field). Indeed, as quick look through the Turkmen library reveals that most all mina khani field weavings do seem to have an open left knot, yet these seem to be for the most part of middle Amu Dary and mid 19 cent. or earlier rugs. As I had stated earlier

"It seems that misperception often proceeds from the confusion of weaving habits of the past with those more recent. As an example, I had noticed a passage ( from the T.M.'s "Turkmen" I believe), which had asserted that the mina khani pattern is frequently seen in Amu Darya weaving, which is of course true, but does not negate the fact that in more recent times this field pattern has figured prominently in the repetoir of Afghan Ersari (hence upper Amu Darya) weaving. Perhaps a characteristic by which one could distinguish between the two is the quality of and use of color? I see this confusion regarding the habits of older weaving culture and the more modern in our discussions constantly".

It seems that an Arabahi/Ersari hybrid is still tenable pending further clairification of how the this mina khani pattern equates to structure and tribal affiliation.

Also, I think it likely that geographic location and access to dyestuffs could account for differences in palettes of various weavers and weavings. Arabachi who are removed from their point of origin to another locality might well be expected to use different dyestuffs than their relations in their ancestral lands.


Posted by R. John Howe on 10-13-2005 05:29 PM:

David -

I agree. Since the Amu Darya flows roughly north west, ending in the area of the Aral Sea (I say "area of" because it appears nowadays sometimes not to get there, as a stream, at least) characterizations like "lower" and "upper" can be confusing, since lower turns out to be further north and upper is southerly.

And someone like me, who hasn't been there, and who therefore has a deprived sense of the area's geography can be careless and simply wrong in his descriptions.

I do think you are correct in believing that this dealer's argument implies that southern Arabatchi dyes different from those on the northern Arabatchi pieces (he doesn't actually say that) might be expected, but he thinks they brought their knot and knot node width tendencies with them.

Don't know if he's right. Lots of folks here are dissenting, but his argument is offered with some bases.


R. John Howe

Posted by R. John Howe on 10-13-2005 07:05 PM:

Dave -

I only responded to one of your questions.

I don't know either how to decode Richard's sentence about "ALL Mina Khani pieces of this design group that I have ever seen" either unless it just a redundancy.

The pieces with a "mina khani" field design in this instance would seem to be the "design" group.

If Richard writes me about this I'll pass along what he says.


R. John Howe

Posted by Chuck Wagner on 10-13-2005 11:22 PM:

Hi John, and everyone else,

First, I should note that the colors shown in my images are not to my satisfaction. All but the first image are direct scans, and even with some attempts at tuning the images to what my eye sees, I can't get the reds from the scans to display correctly, that is, with greater saturation. I mention this because, if I'm reading the comments correctly, the perception is that there is no red (only shades of brown) in my piece. The first image, the whole rug, is more representative. My digital camera died, so I can't do much more than scan for the time being.

It is true that the reds are rather pale in comparison with most Turkoman pieces, but I'll point out that there are many examples of later Yomud pieces with light reds that are almost light pink in good light.

There are some things that don't fit well with the Baluchi pieces I've handled. As I noted earlier, the weft strands are not plied into a single yarn; I've never seen that in a Baluchi piece. Then, there is the warp yarn: four strands of ivory wool plied together; also a little odd for a Baluchi piece. The weave is pretty coarse for Baluchi work at roughly 50 kpsi. Last, the animal hair cords that form the foundation for the selvages: It's common for Baluchi pieces to have selvages wrapped with hair, but the cords are usually additional warp yarns and are made of wool.

It's useful to note that it is quite common for Ersari, Chodor, and "Bokhara" (workshop Beshiri) pieces to have animal hair in both the selvage and warp fibers.

For my piece, I'll accept Baluchi as the most probable source, but the jury is still out. As for yours, I'd like to see those knots up close after it thaws out...


Chuck Wagner

Posted by Patrick Weiler on 10-23-2005 05:08 PM:

spittin' image


I have here a photgraph of your rug before it was mercilessly subjected to use as a car-wash drying rag:

You can see that the outer "badam" border continues all the way around.
Otherwise you can see the same elem features, major border, minor border and field.
It is from the 1994 issue 72 of Hali, in an ad from Hazara Gallery. It is described as: Ersari-Beshir carpet, Amu-Darya region, mid 19th century, 1.8x2.51 m (5'11" x 8'3".

Patrick Weiler

Posted by R. John Howe on 10-23-2005 07:18 PM:

Hi Pat -

Been away for the weekend (and no rugs found).

Thanks for this image.

It indeed seems pretty close.

The person I bought mine from who not be surprised or impressed by this attribution. He is being quite tentative but would say that other folks are just not noticing some useful indicators.


R. John Howe