But as we've gone along (although we've admittedly not looked at many instances yet) it has seemed that our ability to apply weave pattern has not worked very well in helping us make close distinctions.

This is an effort to try one more of these. Neff and Maggs explicitly claim that weave pattern helps to distinguish old Sarouk rugs (defined as those made in the mid-19th century) from those made more recently (i.e., in the 20th century). Here is the information on the two examples they provide.

Neff and Maggs seem to acknowledge that wefts that are usually visible the entire width of the rug and that vary in width are present in both new and old Sarouks. And while they mention fully depressed warps in their description of new Sarouk but not the old one, I can see single nodes of a given color in some places on the old piece, indicating to me that the warps are pretty completely depressed in it too.

Blue wefts might identify and newer piece but Neff and Maggs admit that natural colored wefts occur in these as well. So what is the difference? We seem to be left primarily with the shape and size of the individual knot nodes.

I suspect we are disadvantaged in trying to make this comparison on the basis of these two photographic images because I would guess that the image of the new back has greater magnification (or was simply taken from a closer distance) than does that of the old Sarouk.

But let's try anyway. It would appear that the knot nodes on the older piece are quite tall in relation to their width, perhaps twice as tall (something Neff and Maggs say is also true of Kazaks). The shape of the knot nodes on the newer Sarouk seems more square. And although the magnification of the two backs is different, it may also be that the size of the knot nodes on the newer piece is larger than that of those of the older Sarouk. If these two things are true, they would certainly affect how the backs of these two kinds of Sarouks would "look" at the level of "synthesis."

So what do we think? Could we reliably distinguish newer Sarouks from and older ones on the basis that Neff and Maggs propose?

Regards, Yon

You may be resisting or not taking in fully the center of what Neff and Maggs are saying.

They ARE arguing that the PICTURE (weave pattern) is different from VERBAL DESCRIPTIONS of why a given weave pattern looks different. It seems to me that this is both inescapably and usefully true.

I think that they also feel that if you get good at recognizing weave pattern, that might be the best level at which to operate in making attribution decisions.

Now we can disagree with this second assertion but the fact that I can explain WHY a given weave pattern looks different from another, and even the fact that I do this by reference to closer-grained indicators, does nothing to suggest that there is anything artificial about their distinction.

So far, it doesn't operate for me in the way they claim, (although I haven't yet had much practice) but for me, their "synthesis" perspective is a real and noteworthy difference, not a forced or artificial one.

Regards,

R. John Howe

The first rug can be an old Sarouck, but without more detailed info, I wouldn't bet my life on it. The seccond rug seems to be Jozan: Sarouck region but with a symmetrical knot.

The first rug is old, because it's elongated in design and creates the perfect perspective for me. The seccond rug is compressed in design, so the perspective doesn't work for me in a right way. But I will not write a book about this subject, because I can find lots of loose ends in my "Discovery".

Best regards,
Vincent

Willborg includes a Jozan in his catalog and so I compared the backs and his is very similar to this "new Sarouk."

I'm curious, what makes you suspect that the "new Sarouk" from Neff and Maggs is a Jozan? What are your indicators?

In his description, Willborg only says "...generally Jozans are a bit more geometric and the have a main border with an undulating vine containing small white "pearls" that is typical."

Neff and Maggs do not offer a Jozan.

Regards,

R. John Howe

Weave pattern identification and traditional technical analysis are not mutually exclusive but they are, as John has said, different.

Vincent has quite correctly observed that the "new Saruk" may be a Josan. Those little dots around the perimeter of the blue field tell me that it probably is. In this instance, however, the lack of a technical analysis prevents us from deciding. The back of a Saruk may have a weave pattern virtually identical to that of a Josan. And so might a new Indian copy of a Saruk. We can only distinguish by having the pieces in hand.

Some identifying features, such as the silk warps that Marvin referred to, are impossible to detect by weave pattern.

On the other hand, a technical analysis of knot type, knot count, warp depression and spin and ply of the wool doesn't really convey much of an image. I may later post an example of this from one prominent rug book.

Sometimes identification is made by end or side finishes, handle or even color. Maggs and Neff do not deprecate technical analyses; they primarily say that they don't readily convey important information.

John - I prepared this before seeing your post about Josans. Not only might the new Saruk be a Josan, I think the old Saruk might be as well. I've seen this simple two color field used on many Josans, but not on many Saruks. Maggs and Neff may not offer the term Josan, but I think they offer images of two of them.

Wendel

John, I said it seems to be. This I did on the basis of design, color and the geometrical look the seccond rug has. Can't be sure, because I do not have any technical data. Could be a Jozan design made in Sarouck.

Same goes for American Sarouck and Lilian.
Same goes for Chinees Tabriz and Persian.
Same goes for Hereke and China Hereke.
Same goes for China and Brittish India and Vietnam.
Etc.Etc. You just have to see a couple of rugs, make mistakes, get lost, start all over again, make more mistakes etc.

But, most important, I've never trusted any book since the time I was told to read.

Regards,
Vincent

Rug #1. Warp: ivory wool, 2 strands, Z-spun, S-plied. Weft: dark wool, same as warp. Pile: wool, 2 strands, Z-spun, S-plied. Knot: Turkish, with left warps depressed 90 degrees; h. 10, v. 10, 100/square inch.

Rug #2. Warp: wool, 3 strands, Z-spun, S-plied. Weft: undyed wool, 2 strands, Z-spun, S-plied. Pile: same as weft. Knot: Turkish, h. 7, v. 10, 70/square inch.

Rug #3. Warp: wool, 2 strands, Z-spun, S-plied. Weft: mostly wool, same as warp, but some cotton strands, 2 shoots. Knot: Turkish, h. 8, v. 14, 112/square inch. Prominent use of diagonal weaving.

Wendel

Serious?
The prominent author likes Turkish rugs because he's talking about Turkish knots.
He's isn't from the European continent.
Thinks that all 2 ply wool used in knots is twisted. This is, even in new rugs, hard to see.
One thing we can tell for sure: The rugs aren't knitted.

Best regards,
Vincent

Am I even luke warm?

Best regards,
Ken

The good news is that you were the only one brave enough to venture a guess. The bad news is that you weren't even close. These are technical descriptions used by Murray Eiland in his 1981 edition.

Rug #1 is a fine Afshari rug, late 19th Century. Plate 7.

Rug #2 is a 19th/20th Century Star Kazak. Figure 232.

Rug #3 is a C-gul Yomud main carpet, 19th Century. Figure 182. Yes, "diagonal weaving" does mean offset knots.

Sometimes written descriptions can be helpful, even determinative. For the most part, technical analyses tend to read almost alike to me. While I don't read them as much now as I used to, I still feel cheated if a book doesn't have them.

General rug books, especially those intended for a wide audience and relative beginners, ought to include images of the backs.

Wendel

Diagonal weaving, looks like lazy lines to me.
This allows the rug to ly flat. This is still done in low standard, Turkish Kazak production. If two kots share three warps, it's more like dioganal knotting, not diagonal weaving.

Best regards,
Vincent

Yarn spin: Z
Knots: symmetrical, 2 singles, height 13mm, H 9-10 pi V13pi H 130 psi
Warp: 5 ply ivory cotton (Z5S) - no warp depresssion
Weft: white or blue cotton, 4 or 5 ply (Z4S or Z5S), one pick, uneven packing.
Selvage: 5 units (1,1,1,1,1) overcast with dark blue singles
Ends: warp fringes (may be cut) - 1,5cm balanced plain weave - wefts are a mixture of white and blue cotton at one end and pairs of cotton yarns at the other end.

Best,

Daniel

One of the things your little exercise demonstrates pretty nicely is that most of us make attributions first on the basis of design and palette, and are probably correct when using those criteria alone more than 90% of the time. The more technical approaches are only needed for a small percentage of rugs (which may, nevertheless, be among the most interesting ones), and the technical information in the absence of a look at the design and palette get us almost nowhere beyond being able to eliminate some possible attributions.

Steve Price

The back of the rug I posted is the back of a circa 1930 - "Tafresh" rug (Hamadan area - Southwest of Saveh) exactly similar to plate 19 in Willborg's catalog.

Best,

Daniel

I think what Steve says above is an accurate description of what we usually do when we make attributions (although some might quibble with 90%).

Another interesting thing about it, though, is that it violates another rule about attribution that we frequently cite to one another, that being that designs travel too readily to be accurate attribution indicators, while, on the other hand, the potential costs to the weaver of making structural changes are so high (if you change the number of warps, or begin to depress them or move to a finer weft, etc., literally everything may "thown off") that weavers are reluctant to make such changes and that this makes structure the only really appropriate basis for attribution.

What we actually usually do when we make attribution judgments is potentially both simpler and more complicated than we often say.

This is one of the ways we confuse each other (and surely novices) in our descriptions of what we are about. Our "espoused theories" are often noticably different from our "practice."

This is an instance of a phenomenon I alluded to elsewhere in this salon: that expert practitioners often do not describe accurately, even when they try to do so self-consciously, the rules they actually follow in their area of expertise.

Regards,

R. John Howe

I think you're close, but a little off the mark in saying, ...it violates another rule about attribution that we frequently cite to one another, that being that designs travel too readily to be accurate attribution indicators, while, on the other hand, .... (structure is) the only really appropriate basis for attribution.

I think the essential point, almost always ignored or forgotten, is that every attribution, whether geographic, tribal or date, has embedded within it an implicit probability statement.

Design and palette alone usually suffice for a high probability of being correct. Nobody would ever call a typical Peking carpet anything except Chinese, and you don't have to examine it or see the back. A grainy black and white photo would do just fine, and the probability of being incorrect on this basis is very small.

Not every piece is so easy and obvious, and the finer the desired attribution (i.e., it's Turkmen, but is it Yomud or Chodor?; is it mid-19th century or much later?) the more information we need to achieve a comfortable level of certainty. No matter what we do, unless we saw it come off the loom, we are making a guess that could be incorrect, although the probability of that being the case might be small.

People really hate being reminded that their fondly held beliefs have an uncertainty attached to them, but that, in the words of Winnie-the-Pooh, is How Things Are.

Regards,

Steve Price

I appreciate your diplomatic comments on my “bravery” in making a guess. I wanted to take a wild stab at it since you were kind enough to set it up. (You can see why I did not take up that promising career as a blindfolded knife thrower in a traveling circus.) But it was instructive to see how important context is when one is looking at technical analyses.

To all:

It is worth noting that nearly all the carpets we have discussed in this thread are commercial pieces made for sale or export. The people buying these would have been conditioned to be familiar with a number of set designs, so associating a rug with a known pattern and region would have been very important for both customer and producer. When you invite your friends over to gloat over your new expensive Sarouk, you don’t want to be told it is something else. Likewise,for the manufacturer, it is not good business to have your own production mistaken for that of another region—unless you are trying to produce convincing fakes. A conservative buying public is unlikely to go for radical variants or hybrids without considerable brainwashing. So it is not surprising that certain patterns, weaves, and materials may be more trustworthy for identifying Sarouks, Serapis, Bijars, etc. than for “tribal” utilitarian and ceremonial weavings made for personal or local use. Since in the tribal items commercial value is not linked to a repertoire of specific patterns, attribution by design and pattern is likely to become that much more difficult. And the pieces correspondingly more interesting, of course.

Regards to all,

Ken