Some related asmalyks

Here are three images of similar Yomud asmalyks to the one in the current mini-salon. Would be interesting to see what aesthetic differences people see in the four pieces.


The Pinner asmalyk is lot 41 in RB's Pinner auction, dated 1st half 19th century.


The Rickmers asmalyk is from the Rickmers collection, dated mid 19th century.


The Thompson asmalyk is from Mackie and Thompson's book 'Turkmen', page 164, dated third quarter 19th cent.

I have added this image of the asmalyk that was the subject of the mini-salon, for the convenience of the readers. Steve Price

Hope this will yield an interesting discussion.

Tim

Thanks for the three images. I think that they show a relative age difference quite nicely. If you look at the detail of the running dog borders, there is an obvious degeneration of the hooks (dog heads?) from the top to the bottom. The same can be said for the detail of the spikey hooks coming off the lattice. I would agree with the REALTIVE ages, not necessarily the 25 year units.

Hi People

I agree with Marvin's comments about the design details and how they are generally related to age. On the other hand, the Pinner piece has the most flattened guls, the largest number of borders, and symmetric placement of the lattice. These are all things I associate with younger pieces. So, the relative age attributions are not so unambiguous to me, and I'm inclined to suspect that the Pinner piece is the youngest, the other two more or less the same age (Thompson's the oldest, if forced to make a choice) but somewhat older than Pinner's.

It would be good to have all three available for hands-on examination.

From an aesthetic standpoint - again, with nothing to go on except the images on my screen - I find it a tough call between the Rickmers and the Thompson pieces.

Regards

Steve Price

I agree with Steve's aesthetic take on these three. I also agree with his description of the guls and the layout; all of which beg the question,"who can tell the age of these or any piece?" - even in relative terms

When is a border a border?

I think one could view the Pinner piece as having 5 or 3 borders. I personally feel the two inner side borders are really just part of the main border. They have the same background color as the main border, and are separated only by a narrow black line. From this perspective, the Thompson and Rickmers pieces display a simplified main border.

A couple more differences may be noteworthy (don't know whether they are important though): (i) The angles of the lattice of the Pinner asmalyk are less steep than the angles of the other two. (ii) The weaver of the Pinner piece used half ashiks near the border, while the weavers of the other two pieces used new motives to fill in the half space. (iii) The gable of the Pinner piece is steeper than the gables of the other two. RB state that this is a sign of age.

Steve: I think it would be nice to have the asmalyk that started the mini-salon in this thread for comparison as well.

Tim

Hi Tim

I just added it to your opening post in this thread, so it can be seen with the others without a lot of scrolling.

Steve Price

Hi

I am not a Turkmen fan, but just out of interest;

Steve, you wrote:

"On the other hand, the Pinner piece has the most flattened guls, the largest number of borders, and symmetric placement of the lattice. These are all things I associate with younger pieces. "

Why do you think later piece guls became flattened?

Regards
Richard Tomlinson

Hi Richard

The conventional explanation for the proliferation of borders (and other details) and flattening of motifs, which I suspect is correct, goes something like this.

By 1850-1875, Turkmen weavers knew that there was a western market for their work. The merchants told them that western buyers valued very fine detail, and that their goods would sell more easily (hence, for more money) by adding detail and being more finely woven (which goes with adding detail, to some extent).

The weavers increased knot density by putting in more rows of knots in the vertical direction, making the motifs smaller in that direction (using the same number of rows for a given motif). This resulted in the flattening.

Regards

Steve Price

Hi Steve

Thanks for your answer.

So, can we then assume that a high knot density is not necessarily a marker of age, or even skill in weaving?

And one more question; do you think that 'flattening' or 'distortion' of designs (in general, not only Turkmen pieces) is a sign of later pieces?

I have very limited experience, but I have noticed that many early Shahsavan pieces tend to have 'fuller' designs.

Perhaps other readers might wish to comment.

Regards
Richard Tomlinson

Hi Richard

I don't know how relevant knot density is to age for things other than Turkmen pile woven items. Even within the Turkmen, typical knot density varies from one group to another.

Within any Turkmen group, the ratio of vertical to horizontal knot counts is probably a better indicator of age than knot density per se, anyway. Younger pieces generally have higher ratios of vertical to horizontal (because vertical counts rose with time).

Vincent Keers has expressed some thoughts about design compression and age from time to time. Perhaps he can jump in here.

Regards,

Steve Price

Here is the technical analysis for the first three asmalyks:

Pinner asmalyk
Warp: camel hair, Z2S
Weft: camel hair, Z2, 2 shots between 2 rows of knots, alternately waved
Pile: wool, Z2
Colors (7): red-brown, brown-red, orange, blue, green, brown, ivory
Knots: symmetric
Knot count per dm: hor: 32, ver: 50, 1600 per dm2
Knot ratio: hor : ver = 1 : 1.56

Rickmers asmalyk
Warp: wool, Z2S
Weft: wool and cotton, Z2, 2 shots waved
Pile: wool, Z2
Colors (7): brown-red, dark red, orange, blue, green, brown, ivory
Knots: symmetric
Knot count per dm: hor: 40, ver: 60, 2400 per dm2
Knot ratio: hor : ver = 1 : 1.5

Thompson asmalyk
Warp: wool, Z2S, warps uneven, some alternate warps slightly depressed
Weft: cotton, Z2
Pile: wool, Z2
Colors (7): red, dark red, yellow, medium blue, dark blue, brown, ivory
Knots: symmetrical
Knot count per dm: hor: 65, ver: 117, 7600 per dm2 (10:18 per inch, 180 per square inch)
Knot ratio: hor : ver = 1 : 1.8

According to the knot ratio indicator, the Thomspon asmalyk would be the youngest. It also appears extraordinarily finely woven (hope I got the conversion into dm correct). Maybe also a sign of a later piece?

Robert Pinner writes about the Rickmers asmalyk that the steep diagonal lines are obtained by using "offset knotting." Possibly the same technique has been used for the Thompson piece. So, the guls of the Pinner and Steve’s pieces are not flattened due to a higher knot count, but the guls of the Rickmers and Thompson piece are unusually steep.

One more question: Does the presence of camel hair and cotton help in determining the age of Turkmen/Yomud pieces?

Dear Steve, and all,

Forgive me for being late.

Think the explanation you gave, is the only one that comes to mind.

Had an other quick look in Hoffmeister's Turkoman Carpets etc.
Pieces 19'th century and some 18'th century.
Must say, it doesn't make sense.
Piece 17: About 1800 / 36x60 warp direction = 1:1.6666666666 (Almost perfect)
Piece 18: About 1800 / 35x88 warp direction = 1:2.514285714
Piece 19: Nineteenth century / 32 x36 warp direction = 1:1.125
Piece 23: About 1800 / 48x94 warp direction = 1:1.958333333
Piece 24: Nineteenth century / 46x62 warp direction = 1:1.3478 etc.
Piece 25: Eighteenth century / 38x90 warp direction = 1:2.3684 etc.
Piece 26: Nineteenth century / 48x70 warp direction = 1:1.4583 etc. / "Round Güls" Hoffmeister says. I see only the top row is elongated.
Piece 57: About 1800 / 45x85 warp direction = 1:1.88888888 and extra are the depressed warps.
Piece 58: About 1800 Salor has depressed warps as well.
It's open spaced, round design as expected. The extra is the depressed warps.
The more warp depression the more knots per dm. horizontal. This leads to 1:1
And 1:1 is square so it contains a circle.

Something else that bothers me is:
If the warps are all at the same level, I expect two sinuous wefts.
But maybe only one weft is applied. That should give some extra knots vertically. But Hoffmeister didn't look at the wefts!

Two sinuous wefts need more warp space than one straight weft combined with a sinuous weft.

So maybe we should have a better look at the construction before looking at the product. Round or flat?

About design compression?
If the design gets elongated bit by bit in the vertical direction. It more likely to be pre 1850.
If the design gets compressed or doesn't change at all, it's standard production.

Oops, I did it again.

Best regards,
Vincent

Hi Vincent -

I may be the only person in the world who had not noticed it but your observation about the effect of warp depression on drawing is excellent.

Salor pieces often have some of the highest knot counts of any Turkmen rugs but one is struck repeatedly standing in front of them for how often the "circles" in them (for example the major guls on a Salor main carpet) are close to perfect.

Your observation suggests that it is not just knot count per given square area that needs to be treated with caution, but that knot ratios (nowadays thought of as a better basis for comparison) can really only be compared equitably between pieces with similar degrees of warp depression.

It's good to have someone around who "goes to the data."

Regards,

R. John Howe

Hi John, Steve and all,

Think it bothers me that: " The weavers increased knot density by putting in more rows of knots in the vertical direction, making the motifs smaller in that direction (using the same number of rows for a given motif). This resulted in the flattening. " as the thing that comes to mind first, doesn't seem to work.
I made this drawing.

All nodes are 2 mm square.
So knots are 2 x 2 mm if the warps are depressed and 4 x 2 mm. if the warps are level.
Both circles are the same. The left one is perfect. The right one looks flat but it is not.
This suggests that changing from depressed warps to level warps does the trick.
Skipping 50% warps means less work. Ex- and importers love that.
As an aside, most borders at the left and right side are wider compared to the borders at
the top and bottom side in the Turkmen world.

The idea is that by changing the knots into 4 x 1 mm. and keeping the same amount
of lines per distance, it creates more space for extra design.
So all knots are changed. So the negative space changes as well (It gets smaller).
Now we should see more borders if the total area stays the same.
More borders at the top and bottom and the left and right side.
In order to get in more borders at the left and right side, it doesn't seem to work if
only the vertical proportions are changed.

Think someone should get out and count warps per cm in all the pieces available.
Because what happens vertically, should happen horizontally as well?

Best regards,
Vincent

Hi Vincent

Your analysis is very interesting, and I like it a lot. But there are some problems in it.
1. Turkmen weaving (with some exceptions) is done on level warps. Depressed warps, when they occur, are usually on late pieces, not on early ones. This is exactly opposite to your prediction.
2. The motifs in later pieces don't just look flattened, they really are flattened.
3. I don't have the data or reference handy (perhaps someone out there knows it and can post it), but it is decently documented that the ratio of vertical to horizontal knots in Turkmen pieces (at least, within some groups - my memory isn't reliable enough to be sure whether it's for all of them) gets larger as time progresses. That is, vertical knot counts go up in later pieces. If horizontal counts go up, they don't go up as much. Like most indicators of age, the correlation isn't perfect, but my recollection is that the trend is pretty clear if enough specimens are used.

This is a nice example of a beautiful hypothesis coming face to face with some ugly facts.

Regards,

Steve Price

Hi Steve,

I never predict anything.
It's Hoffmeisters data that I rely on. So I've been reading a book! Please don't tell me it's the wrong book or I will skip reading rugbooks for the rest of my life.

Isn't Salor, Turkmen or can't the pieces Hoffmeister shows be 18'th century?
Ratio meaning: 30 is to 60 as 40 is to 80 and 50 is to 75 as 60 is to 90. So more warps if the vertical knot count goes up. If the warps stay the same we can't speak of ratio.
But maybe Turkmen ratio is: 30 is to 45 as 30 is to 90.

Vertical knots go up.
So in older pieces we see Guls, 4 inches vertically = 360 knots
In younger pieces we see Guls, 3 inches vertically = 360 knots.
Again the data do not show this. Sometimes it's the other way around. Vertical is higher and horizontal gets down. (Piece 25: Eighteenth century / 38x90 warp direction = 1:2.3684 etc.)

But if you say the warps have all been counted , I'll let it rest.
Hope someone checks in, that did this.

Best regards,
Vincent the Depressed

A hump in the road

Hi

Tim Adam wrote;

"Does the presence of camel hair and cotton help in determining the age of Turkmen/Yomud pieces?"

He also presented information that suggests the Pinner piece has camel hair warps.

There was very recently a Khirgiz piece sold on ebay that the seller believed had a camel hair foundation.

I recently asked a very similar question to Tim's - are camel hair warps in Shahsavan pieces a sign of age. I was told that camel hair is not strong enough for warps, and that it is simply brown wool.

Now I am completely lost.

Thoughts?

Richard Tomlinson

Hi Richard,

Be happy, you aren't the only one that's lost.

Camels don't celebrate the spring or autumn shearing party.
Camel hair is gathered by combing or just picking it up from the ground.
Camel hair doesn't have the lengthy fibre's that sheep and goats have.
So camel hair can be spun into a thread but the short fibres make it unstable and breakable.
Now, I've never studied a camel. But I think it has a few parts where the hairs are longer.
So, in order to make a staple-fibre from camel hair, a flock of camels is needed.

This doesn't seem to be the problem with the the pile.
But I think most rugs that look camel coloured, are made from wool.
The Mehraban-, the Karakul sheep etc. all show camel colours.
And I can make a rug from the hair my Afghan dog looses every day.
A beautiful camel/reddish/blond mix.

And sometimes, only old isn't enough for a rug. It has to be extra, extra special.

Best regards,
Vincent

Originally posted by Vincent Keers
I never predict anything. It's Hoffmeisters data that I rely on. So I've been reading a book! Please don't tell me it's the wrong book or I will skip reading rugbooks for the rest of my life.

Hi Vincent,

Hoffmeister's book is very good, and I'm sure his information is accurate. But the sample size (the number of pieces) isn't very large.

Ratio meaning: 30 is to 60 as 40 is to 80 and 50 is to 75 as 60 is to 90. So more warps if the vertical knot count goes up. If the warps stay the same we can't speak of ratio.
Of course there's a ratio change when the vertical knot count changes and the horizontal count doesn't. The ratio of vertical to horizontal is simply one count divided by the other: 25 vertical per inch, 12.5 horizontal per inch: ratio = 2.0. Change either one of the numbers and the ratio changes.

I was out last night and didn't have time to root around for the reference to the data on knot count ratios and age, but I did notice that ithe relationship is mentioned in the Attribution Guides on Turkotek. I believe John Howe wrote those, and if he remembers the source, perhaps he will post it before I get a chance to look for it tonight.

Regards

Steve Price

Dear folks -

Here from the Tekke attribution guide is one of the passages Steve is pointing to:

"Weave structure. Older pieces tend to have a lower knot count than more recent pieces. In addition, earlier pieces may tend to have a vertical to horizontal knot ratio that is more equal as compared with more recent pieces, which tend to have a vertical to horizontal knot ratio that sometimes approaches 2:1. "Thus Tekke main carpets which have the roundest guls are considered to be among the earliest; following from this, it is often argued that a squarish format is also indicative of an early age." Jourdan, at 17."

As an aside, I acknowledge that I did some of the work on these attribution guides, especially on the Salor one, but someone else has now done a lot more. Moreover, this person has drawn heavily on Loges, a good source but hardly the only one. I have heard Robert Pinner question Loges' analyses repeatedly.

But I think Steve's point is sound. The vertical knot count does tend to increase on pieces seen to be younger. Only the Salors seem to use deeply depressed warps. Any weaver using asymmetric knots will likely have a slight degree of depression. It's in the character of the knot to be weaker on one side. But Tekke weaving, especially (and that is one place we see the phenomennon Vincent is analyzing) is famous for having warps as flat as those on Shirvans.

So I think Vincent's imaginative analysis may not lead to the correct conclusions in general.

Regards,

R. John Howe

Hi Again, Vincent

You commented about the effect of changing vertical knot density on the number of borders, pointing out that In order to get in more borders at the left and right side, it doesn't seem to work if only the vertical proportions are changed.

You are absolutely correct about this. But the number of vertical borders do change. This is accomplished by reducing the width of the field to accommodate the added borders.

You can see this pretty clearly in the illustrations on a page in the Oriental Rug Review website. http://www.rugreview.com/1juvals.htm
It includes an old and a young Tekke juval, and an old and a young Yomud juval. These are quite typical of pieces generally attributed to their respective ages.

Regards

Steve Price

Hi Steve,

Thanks for the link.

Happy that figure 3 / 12:20 and 4a / 8:20 seem to act as I, being Dutch, expect it. Salor 4a made more profit.

Ratio is: Vertical : Horizontal = x ?
Or:
Ratio is: Vertical : Horizontal piece A = Vertical : Horizontal piece B. This is what my dictionary told me? Last night.

"This is accomplished by reducing the width of the field to accommodate the added borders."
All right, so now the Güls are round again Or güls are skipped? Or less negative space?

Best regards,
Vincent

Hi Vincent

The piece with 20 knots per inch vertically and 12 knots per inch horizontally has a vertical/horizontal knot density ratio of 20/12 = 1.67.

The piece with 20 knots per inch vertically and 8 knots per inch horizontally has a vertical/horizontal knot ratio of 20/8 = 2.5.

This is usually what people are talking about when they cite vertical/horizontal knot ratios. You could, of course, take the ratio in one piece to the ratio in the other (what your dictionary told you last night). Mathematically, it's a legitimate operation. But it isn't customary.

"This is accomplished by reducing the width of the field to accommodate the added borders." All right, so now the Güls are round again Or güls are skipped? Or less negative space? There is less negative space. Sometimes the number of guls is increased - younger pieces tend to have more guls than older ones do (lots of exceptions, of course). People often describe the fields of younger Turkmen pieces as "crowded" or "busy".

Regards

Steve Price