In 1953 or 1954, Edna Martin of Friends of Handicraft — Handarbetets Vänner — in Stockholm designed a large pile rug for Dag Hammarskjöld’s apartment in New York City titled Läggspel. I don’t know when Martin actually designed the rug, or how long it took Handarbetets Vänner to weave it, but I think it is likely to have been ordered in 1953 after Hammarskjöld was appointed to be Secretary-General of the United Nations organization. It was apparently delivered to him early in the summer of 1954, in time for the New York Times Magazine to show a photograph of his apartment living room, with the rug in it, for an August article (discussed in Part 2 of this post).
As noted in part 1 of this blog post above, the title “Läggspel” is an infrequently-used Swedish word which means means “puzzle,” or a the kind of game where pieces have to be assembled to make a whole. Apparently it can also mean the tricky sort of process of trying to work out a time schedule, or a logistical problem. Even if intended to describe the design of the rug, one can see how the idea could also apply to the kind of constant diplomatic balancing act required by Hammarskjöld in his role as Sectretary-General as well.
This post is a brief design analysis of this large pile rug and provides at a little more information about Martin’s further exploration of this design. The rug is certainly visually complex, but it does not feel demanding or busy: its interlocking puzzle-like architectural elements are offset by its quiet, easter-egg-like coloration and its overall soft appearance.
Looking at the rug more carefully, we can see that it is in fact a carefully staggered arrangement of two repeated pattern blocks. Each of these “blocks” consists of some 20-40 smaller pieces, thus the fact that there is, in fact, a repeat in the pattern is not at all clear. These two larger blocks of pattern repeat are identified by the images below:
The first element consists of a very abstract church-like shape, set among a field of pale trapezoidal shapes and punctuated by small yellow square, rectangular and diamond shapes outlined in a creamy white.
The second element is smaller and consists of more pale trapezoidal shapes and a box-like shape with two pale blue “handles.” This has an odd blue shape with a pointy hat and a yellow window-like shape inside that.
The following diagram shows how these A and B elements fit together to make the overall rug pattern. Each of the pattern blocks appears four times, but alternates from side to side, and because the first block is larger than the second, the pattern is staggered or offset.
Up the middle, almost as a sort of seam, is a further repetition of the church-like shape, but the two different blocks “share” this shape. And at the two ends of the rugs, the pattern tails off into simple color blocks, and not a continued pattern repetition. To increase the jazzy quality of the rug, the puzzle elements, even if repeated, vary in color and also slightly in their shape or size.
The border with its crisp black and white edge seems almost to borrow from the kind of distance measures on maps — is this rug a fantasy landscape or townscape?
Unlike many Swedish flat-weave rugs (“rölakan”) which use mixed skeins of yarns to create great depth of color, this pile rug uses a kind of mosaic of individually colored pieces, with at least 15 or 16 different shades of pale color apparent, predominantly in the blue, grey-green, lavender, eggplant and pink families. The yellow seems to be one consistent clear color.
The pattern of this rug seems to have been one which fascinated Martin for a number of years. She may have even been playing with it before Hammarskjöld requested a carpet for his apartment. At some point, date unknown, a large wall hanging incorporating just one of each of these elements was woven in similar colors but without the flashes of yellow which so enliven the carpet. This was also woven in a wool and linen flat weave, giving a very different appearance than the Dag Hammarskjöld pile rug.
My final discovery about the design of Läggspel, the carpet in Dag Hammarskjöld’s apartment, was accidental— and is often the case with research, it leaves open as many questions as it solves. I was in Stockholm earlier this year, and had a chance to visit some of the archival material from the Friends of Handicraft. I was not looking for sketches for this rug, but when I found a pile of Edna Martin’s sketches, I could see that these were unmistakably related to the design of the Hammarskjöld carpet. But they were not, apparently, for the Läggspel rug itself. There were several sketches and even wool samples for variations on a rug called Mosaic (“Mosaik,”) and these were dated 1954. I assumed at the time that these were ideas which Martin was playing with and that when Hammarskjöld wanted to purchase a rug, that she was able to generate the design based on these earlier efforts. But with their 1954 date, it looks like these designs were drawn after his carpet was already being woven, or had been delivered.
It was not uncommon at all for designers to offer color variations on a particular design. Even changes in technique: from tapestry to flat-weave or flat-weave to pile, although infrequent, are not unheard of.
So it looks like the Mosaik sketches and documents were variations on this pile rug already produced. But these are not just sketches for a single alternative. They represent instead five different alternative designs. They are different in several significant ways: pattern orientation, scale and color. Three at least are designs for flat-weave rölakan rugs, all variations on the Mosaik pattern. Let’s look at these.
One working drawing in pencil on graph paper has colors noted by numbers in several segments of the design. There is no color code, so it is hard to know what colors were originally intended. But the pattern itself is somewhat similar to the wall-hanging shown above, in that is uses just a piece of the larger pattern we know from the Läggspel carpet. It is labeled Mosaik and signed by Martin but it is a design for a large flat-weave carpet designed to be 324 x 236 cm, or about 10’-8” x 7’ 8”—about half the size of the Hammarskjöld rug,
Details of this design differ slightly from elements A and B discussed above which make up the Läggspel rug design. Because the details here seem less resolved, it seems like this sketch may predate the Läggspel design, but is undated, so it is hard to know. It is the only one of the sketches in this collection to use the alternating black and white border used on Läggspel.
Two other watercolor sketches for similar carpets use a blue, purple and pink
color palette very much like that of the wall hanging shown above, but where that drew just on a few elements, these show the pattern reproduced to a much larger extent. Both dated 1954, they seem to take the design of Läggspell and turn it ninety degrees, running the pattern sideways with the length of the rug. The diamonds of various sizes which punctuate the pattern are black like those in the wall-hanging, rather than the yellow in Hammarskjöld’s carpet. In this same large
Curiously, though, these two sketches differ in the direction of the pattern…that is, what is down and what is up. Judging from these images, Edna Martin did not seem to have a preference for the direction of the pattern. The first shows the “church-like” elements of the composition inverted; the second reverses that to the more familiar direction. But in fact, in the advertisement for Handarbetets Vänner which I had found earlier, which had given me the name of the rug (this is reproduced in Part 1 of this post) also uses the “upside down” position. Based on the photographs of his apartment, Hammarskjöld chose to place his own carpet with the “church-like” elements of the pattern pointing up.
It is also interesting to note that in both of these sketches, Martin seems to be exploring a different border than the black and white one she used for Läggspel. This one looks like a dark blue border, into which some of the small black diamonds project at irregular intervals.
Another unsigned and undated fragmentary Mosaik sketch in oil crayon investigates a very different color palette. Here the outlines of the shapes are black rather than white and the colors proposed are dark greens and dark blues, An additional card of proposed wool color samples for a Mosaik rug seems related to this sketch, but in a startlingly different colorway. It displays a range of reds, oranges, and purples, with smaller amounts of black grey and blue. These two sketches are so different in character from those shown above, and given the difference between their mediums, water color and oil crayon, it seems likely to me that these suggest another application for this design– as rya rugs, with deeper colors and more blurred pattern. And yet, the color card is labeled as a röllakan, that is, flat-weave and not rya rug. So perhaps the blue-green-black was also intended as a flat-weave.
The brief analysis above about the design of the colorful and handsome Hammarskjöld pile rug called Läggspel tells us more about it as a design. But the discoveries of a similarly patterned wall hanging and sketches raise other questions. We have the interesting linen wall-hanging, but were there any flat-weave or rya rugs made to any of these versions of the Mosaik design? Were there early versions of Mosaik which predated Läggspel, or did the design for that pile rug begin the various iterations of pattern and type of rug indicated by this collection of sketches? Is there any more information that would let us put these versions in a chronological order? I certainly wish I could go back and ask Edna Martin herself about her own creative explorations of these different versions of this wonderful pattern!
FJ Hakimian Antique Carpets, New York
Nordiska Museet archives, Stockholm
Please reference as follows:
Whidden, Anne, “Furnishing a New York Apartment: Edna Martin, Part 3 – theswedishrugblog (9/23/20); theswedishrugblog.wordpress.com; accessed (day/month/year)