I’m still working on this page. Please leave comments, corrections, and suggestions in comment field below…
1. Terms used in instructions (below), and
2. Structural terms – used in describing braided structures (“weaves”).
1 – Terms used in instructions:
Any movement of loops by the fingers.
A very specific kind of move—a move in which a loop is interlaced with or through another loop or loops. A square braid has two loop transfers that repeat over and over again. One is on the left and the other on the right. In each of these loop transfers, a loop is pulled through all the one or more loops between its starting point, and its new finger on the other hand. There are more moves than this in each cycle, but only 2 transfers.
A move that rearranges the loop or loops on the fingers, without interlacing them through any other loops. Ideally, this move is performed by the finger that the loop is shifting onto, not by a “helping” finger of the other hand. It’s usually done in order to put the loops back into their correct starting-position set-up after a loop transfer occurs.
In the 15th C. manuscripts, shifting loops was called “walking the bows,” which is how it feels, a very reflexive and easy motion that looks complex—like a centipede walking. It was also called to “hi” or “low” the bows, depending on which direction the loops were being shifted.
A move that connects the two halves of one braid—traditionally done in team braiding. Typical example: After each of the individual braiders has finished the moves of one braiding cycle, the two (or more) partners then exchange their closest loops to connect the individual braids into one larger braid.
Noémi Speiser also uses the term “loop exchange” for one particular type of loop transfer: in which a loop on a finger of one hand is exchanged with a loop on a finger of the other hand, one of the two loops passing through the other in this process. The Spiral braids (video links at the bottom of my Tutorials page) are made with this type of loop transfer, as are several other known loop braids.
One repeat of the braiding movements, after which the same exact movements are repeated.
Like a pigtail braid, square braids have two loop transfers in each cycle of braiding, one on the left and one on the right. Spanish braids have 4 loop transfers in each braiding cycle (as do my double braids made by a single braider).
The 8-loop spiral braid—or ‘lace bend round,’ also has 4 loop transfers in each braiding cycle, before the same sequence of moves repeat. (These transfers are very different from square or spanish braid loop transfers—they might also be called “loop exchange moves”.)
‘Braiding cycle’ only refers one repetition of the hand movements, not to a repeat of the pattern on the braid, or to all the loops returning to their original finger-positions.
That usually takes several braiding cycles, depending on the number of loops and the color pattern of the braid.
This is the term for the length of braid, or of braiding, that it takes before all the colors return to their original starting positions on the fingers. This will equal one full expression of the color pattern on the braid.
For a five-loop square braid, it takes 5 braiding cycles (= 10 loop transfers) before all the loops return to the fingers where they started. So the pattern repeat for this braid is often five braiding cycles, seven cycles for a 7-loop braid, etc. However, if the braid has a very limited number of colors, the pattern repeat might be shorter: the few colors might return to their original starting arrangement on the fingers before each particular loop comes back to the finger where it started.
Bicolor loop patterns add another factor. Depending on how often the loops are turned, it might take twice as many cycles before the the loops come back to their starting fingers with all the colors in their original positions on the fingers. For example, in the flat version of a square braid, the first time the loops return to their starting fingers, each loop has only been turned once, so the colors of any bicolor loops are now in reverse on the fingers (and in the braid). A full pattern repeat for a 5-loop flat braid with bicolor loops is usually 10 braiding cycles (= 20 loop transfers).
Or a pattern repeat might be as short as only one cycle (as with an “Edge” -type of bicolor pattern in a square or double braid). In an “edge” pattern braid, the colors return to the same set-up on the fingers by the end of each braiding cycle.
Turned loop transfer (same thing as Reversed or Crossed loop transfer):
Turning or not turning a loop over while transferring it can make a huge difference to how a braid turns out. Doing one or the other, or mixing the two types of transfers, is how you make a braid square, flat, or divide into two braids. ‘Turning’ the loop means taking it off its previous finger in such a way that it gets rotated by 180 degrees (a half-turn) as it moves onto its new finger. This makes the upper and lower shanks of the loop switch positions.
Reversed (vs unreversed) were the 15th C terms for this; Turn was used in the 17th C manuscripts; Crossed (vs open) are Noémi Speiser and Masako Kinoshita’s terms. To me, ‘turned’ (vs ‘not turned’ or ‘straight’) seems clearest, and when I teach, students seem to understand it more easily. In writing and in my tutorials I usually cite all three terms because people have different backgrounds in loop braiding.
Two types of linking for color effects are possible with loops. Both these linking techniques were used in fingerloop braids described in 15th and 17th C. English manuscripts, as well as in old Japanese “kumihimo” braids and Andean/ Peruvian flat braids that are now known to have been made with hand-held loop braiding.
(1) Linking the two shanks of one loop (“turn twice”).
With bicolor loops, an optional way to turn a loop is to turn it twice (ie turning it a full 360 instead of 180 degrees). This has a negligible effect on the shape of the braid compared to turning the loop once, but creates a very different color effect, since the colors of the upper and lower shanks of the bicolor loop will not switch positions. When a loop is turned twice, its upper and lower shanks become linked around each other, instead of crossing each other. At that point in the braid, each shank links around the other shank and goes back to braiding on the same layer of the braid where it had been, rather than the two shanks crossing each other to each emerge on opposite sides of the braid (as when a loop is turned only once). This can create lengthwise columns of the two bicolors, for example in a flat braid (where lengthwise columns of color would not be possible otherwise). I teach this in a few posts: 1 (see third video), 2 (see third and 6th videos).
(2) Linking two loops (“exchange twice”/ “transfer twice”).
A different kind of linking can be done as part of the loop transfer or loop-exchange to create interesting color effects (as in the braid in my sidebar →). This type of linked color-effect doesn’t require using bicolor loops. In this case, two loops of 2 different colors are linked together, instead of crossing each other as they would normally do. That is, each shank ends up linked to the equivalent shank of the other loop. Linking the two loops results in each color turning away from its ‘normal’ course, back in the direction it just came from (each loop now heading where the other loop would normally have gone). It’s a way to prevent selected colors from migrating all the way across the braid, so as to keep them within certain portions of the braid. It is done by exchanging or transferring the loops two times in immediate succession as part of one loop exchange or loop-transfer move. Linking for color-effects can be done between any two adjacent loops. I show video examples of this in three posts:
1 Linking loops at the loop exchange of a flat double braid (see the second 8-loop video and the second 10-loop video)
2 Linking loops during loop transfers of a 7-loop flat (square-type) braid (videos and text)
3 Linking loops during loop transfers of a 13-loop flat (square-type) braid (video and text)
“Loop braiding” vs. “Finger loop braiding”
I usually use the more general term “loop braiding”, unless I’m contrasting finger-held loops with hand-held loops (say, comparing a European braiding method to a Japanese hand-held loop braiding method), in which case I specify “finger loop braiding”.*
2 – Structural terms (‘weaves,’ etc)
Braiding interlaces threads on a diagonal, often called “oblique” slant, as compared to weaving, in which threads interlace on horizontal/ vertical axes. Plaiting means the same thing as braiding—it seems to be a regional or slightly old-fashioned synonym.
Braiding done as a textile technique, and without tools or a braiding stand, has often (confusingly) been referred to as “finger weaving.” This is probably because people tend to associate the word “braid” with hairstyling rather than textiles. When they see yarn or thread being intricately braided, they might correctly notice that the technique is a type of over-under interlacing like weaving, though done by fingers alone, rather than on a loom. The fact that the threads interlace on the diagonal is unlikely to be noticed or considered especially important. So far, every time I have read or heard the term “finger weaving” it has been used for some form of braiding, rather than weaving.
Loop Braiding* is a type of braiding in which the threads are connected into pairs (loops). These are held and manipulated at the ends of the loops, rather than close to the braid itself. Loops are held by the fingers, or less commonly over the hands. Tightening is usually accomplished by spreading the left and right loops apart in a broad stretching motion, or–in certain circumstances–by ‘beating the fell’ as in a weaving. The top of the braid must be attached to a firm fixed point. (The braider’s big toe has been a frequent choice worldwide, but I don’t find this very comfortable, myself!)
Braids — including loop braids — can have some of the same types of interlacing as weavings. There is a difference, of course, since in braids that interlacing is on a diagonal, not straight across as in a weaving. But weaving terminology can sometimes be useful for braids as well.
To be very exact, for braids you would use the term “oblique interlacing” instead of “weave.” “Plain oblique interlacing” would be the more technically correct term for “Plain-weave braid”.
Braiders are sometimes very adamant that weaving terminology should not be used for braiding. But even though I don’t call braiding “finger weaving,” I am not averse to using some weaving terminology in the context of braids. Braiding and weaving have much in common structurally, and this can be helpful in describing and understanding braided structures.
Fell: The fell is the lowest edge of an in-progress braid or weaving–the “growing” edge. Weavings usually have a straight, horizontal fell. Braids usually have variously-shaped diagonal fells. Any braid in which the outer strands are continually brought/braided toward the center of the braid will have a fell shaped like a V. Here’s an example (in-progress kumihimo braid being made on a takadai).
Any braid in which the innermost strands are continually braided toward the outer edges of the braid will have a fell shaped like an upside-down V–generally called an A-shaped fell. The actual braiding structure or “weave” that is being created may be identical in either case. The structure is not necessarily dependent on the order or direction of the braiding moves (which is what determines the shape of the fell). Some braid structures can be braided with either an A, V, or multiple-W or M-shaped fell—depending on the order and direction in which the loops are moved, but not making any difference to the actual “architecture”/ structure/ thread passages of the finished piece.
The two loop braiding methods called “A-fell” and “V-fell” were named by Noémi Speiser for the fell-shapes they produce, but either of these fell-shapes can also be produced by many other methods, including other types of loop braiding, free-end braiding and kumihimo. For example the Slentre loop braiding method is a different method from the “V-fell method”, but both methods make braids that grow with V-shaped fells. This slightly confusing terminology is (I think) why Masako Kinoshita refers to those three methods as Method 1, 2, and 3 rather than A-fell, V-fell, and Slentre.
Warp and weft are the two sets of threads in a weaving. Warp threads are the ones laid lengthwise onto the loom before weaving. Weft threads are woven back-and-forth–through, and at a right angle to the warp threads. These terms are occasionally useful for braids, too, even though usually all threads take turns being ‘warp’ and ‘weft’, they aren’t completely separate sets of elements as in a weaving. (Some braids, like certain tubular braids, do have two separate sets of elements.)
In a braid, the thread or loop that is being moved/transferred is similar to weft while the more passive threads/loops it is being moved through are similar to warp elements. This is not a completely arbitrary analogy. In some braids—for example my plain weave braids here, and somewhat similar single-course twined braids (described below)—a “weft” strand or loop is carried a long distance between a shed of upper and lower “warp” strands. This is completely analagous to weaving, and I don’t see a need to invent new and opaque terms to be used in braiding when well-known and equivalent terms already exist.
Plain weave or tabby are weaving terms for the simplest type of interlacing:
Each strand goes over one strand, then under the next, then over the next, etc. Over 1, under 1. On the next row, it is UNDER 1, over 1—opposite of the first row. (I remember doing this with strips of colored paper in kindergarten.)
The most common hairstyle braid is plain weave with only 3 parts. Over 1, under 1 again and again.
There are probably many types of plain weave, but the following three are the only types I’m aware of in loop braids:
Balanced, warp-faced, and repp (or repp-like, anyway)
In balanced plain weave, the warp and weft threads are equally spaced/ equally visible in the finished piece.
In warp-faced plain weave, only the warp threads show, while the “traveling” threads—the back-and-forth weft threads—are completely hidden. This is because the warp threads are much more closely spaced than the weft threads. The result is that each weft thread tunnels in a straight path between the “over” warp threads and the “under” warp threads, and cannot be seen on the surface of the piece (though they are usually visible at the two edges). So, in warp-faced pieces, it’s the closely-set warp threads that form the visible color-patterning. Tablet weaving and backstrap weaving is usually warp-faced and so are many braid types.
Weft-faced woven fabrics are very common, but I don’t know of any examples in braids. I doubt it would be possible with loop braiding, because of the tension between the top of the braid and the ends of the loops while braiding.
‘Balanced’ and ‘warp-faced’ are actually two points along a continuum…A braid or weaving might be slightly to strongly warp-faced, depending on how closely the warp threads are set, or packed together in braiding.
In a “repp*” weave, the weft – the horizontal, “traveling” thread – is much thicker than the warp, so it forms a horizontal bulge with each row. Repp is usually warp-faced. The thicker weft thread is completely hidden by the close-set, much thinner warp threads that it tunnels through, and only the bulge it creates is visible. In a braid, these bulging rows are slanted, not straight across. See photo here—then click on the photo to see the “repp-like” slanted, bulging rows in the larger braids. (I’m referring to the textural rows, not to the much more steeply-slanted color-patterns that cross them.)
In these repp braids, every ‘weft’ element tunneling between the the upper and lower warp threads comprises two threads—both shanks of a whole loop, making the ‘weft’ elements twice as thick as the ‘warp’ elements they tunnel through. The alternating upper and lower warp elements are single threads (the two shanks of each loop move apart in opposite directions for the weft to go through).
This produces a thick, flat, ridged, single-layer braid, looking similar to single-course twined braids (described below).
Repp-type braids are flat overall, but they are a very different kind of “flat” from the flat twill braids I teach in my Start Here, Continue Here, and flat double braids tutorials, or the flat plain-weave braids in my header photo above (the widest 4 braids in the header photo). In those braids, the two shanks of a traveling (weft) loop do not travel together as a single weft. One shank of the weft loop becomes part of the upper layer of the braid, and the other shank becomes part of the lower layer. The two layers are only connected along one edge of the braid, the way the spine of a book connects the front and back cover. That edge will become the mid-line of the braid, once the braid is spread open to form a single-layer braid. After the (originally) two-layer braid has been spread open into a flat, single-layer braid it is twice as wide, and half as thick as an equivalent repp braid made from the same yarn.
Twill —Square, Spanish, Double braids (of more than 6 loops), the 14- and 18-loop Letterbraids, etc
These are two-layer braids of (usually) twill interlacing, rather than plain weave. The basic square or “rounde” 5-loop fingerloop braid has a twill structure. It is strongly warp-faced. On the upper and lower surfaces (the sides facing to and away from the braider) only the “warp” loops are visible (the passive loops that the travelling loops are taken through by the braider). On the two side surfaces of the braid, each loop forms a V as its two shanks (upper and lower) meet and dive into the braid, changing from passive “around” (warp) loops to active “through” (weft) loops.
Those V or A-shaped patterns created by the two parts of a loop meeting on the sides of a square braid is not what is meant by a V or A-shaped fell, by the way. In most loop braids, the fell (lowest part of an in-progress braid) isn’t very obvious or visible—it’s obscured by the loops that extend down toward the fingers.
2/2 twill: A 5-loop square braid has (for the most part* ) 2/2 twill: each loop goes first through two loops, then around two loops in its trip across the braid. In each successive row, the pairs of passive loops (that the active loop goes through) are shifted over/offset by one loop—which is what makes it a twill.
A 5 loop square (or flat) braid has 2/2 twill because each loop is always pulled through two other loops—the ones on the middle and ring fingers.
Over the course of braiding, on each layer of the braid, a single thread (say, the upper shank of the traveling loop) goes under two threads to the middle of the braid, then over 2 to the other side of the braid, where it changes direction and dives into the side of the braid again, goes under 2 and then over 2 threads back to the side where it started, now lower down in the growing braid. (This describes V-fell braiding, in which outer loops are pulled to the center of the braid. I am ignoring any turning of the loop while transferring it, so this most closely describes making the divided form of a square braid.)
It doesn’t do all this in one move. In the first move, the thread comes (as one part of a loop) from the left halfway across the braid (through two loops). In the next move, this traveling loop becomes a passive “warp” loop as a loop from the right is brought to the center, through 2 loops, including the loop we are tracking. (This creates our loop’s first “around”, as it inches one short step rightward.) Now another loop from the left follows our loop halfway across. Finally a second loop from the right comes halfway across, through our loop and the one following it. (This is our loop’s second “around” move.)
At this point our loop has made it all the way across the braid and is ready to be the next loop taken through the braid from the right side.
From where it started on the left, it will take 5 full braiding cycles (10 loop transfers), before our loop will arrive back where it started
During this time, the upper shank of each loop is doing its own “unders” and “overs,” while the lower shank is doing mirror-image “unders” and “overs” on the other side of the braid. That’s what makes it a two-layer braid. The two parts of each loop move in parallel, but they don’t stay together in the braid, even though the braider moves them as one. The loops are divided into an upper and a lower braided layer by the consistent, repeated pulling of loops through loops.
A 7-loop square braid has 3/3 twill. Each loop is always pulled through 3 loops, after which 3 loops are pulled through it (over the course of three successive braiding cycles) before it reaches the other side of the braid and is ready to be an active ‘through’ loop again. After seven braiding cycles, the loop will be back where it started.
Twining is when a pair, or many pairs of threads (each pair can obviously be a loop) are twisted a half-twist before another thread or loop is passed between them, or before being passed around another thread or group of threads–as is often the case in woven-type twining done on horizontal/vertical axes. Braided twining is sometimes called oblique twining, to differentiate it from horizontal/vertical twining.
Single Course Oblique Twining (often abbreviated as SCOT, especially by ply-split braiders) is the most common type of twining in European loop braiding, and was also used in other loop braiding traditions (kute-uchi and ancient Andean). For example, the many Medieval braids called “bends” and “chevrons” at the last portion of Cindy Myers’ Fingerloop Braids by Type page. The directions for any of these braids will include the following step: “Twist all loops (on a hand) 180°.” After that step a loop is then drawn through all the twisted loops. All the loops on one hand are usually twisted in the same direction.* However, single-course twining can also be done with counter-twining (see below), in which pairs of adjacent loops are twisted in opposite directions.
The traveling loop–the “weft”–is always at least one whole loop (2 strands) or more. Unlike square braids, these strands do stay together in the braid’s structure while they are “weft”–they do not separate into an upper layer strand and a lower layer strand. However, the upper and lower “warp” elements that the “weft” loop tunnels through are single strands, so the overall texture is similar to a repp-type plain-weave braid (described above, under plain weave)*.
Plain Oblique Twining (POT) is another, more intricate type of twining. In POT a loop is twisted before going through one loop, and then passes untwisted around the next loop (which itself gets a twist) and so on across the braid. Here is a photo of very open, meshlike POT on L-MBRIC. Although it looks very different from the meshlike example, my Rainbow Girl braid was also made with plain oblique twining (4th braid from left in my header photo, top of page). When the braid is tightened firmly, the twined structure is no longer visible. The surface color-patterns usually look similar to plain weave patterns. However, the fabric is twice as thick as a plain weave would be.
Oblique Countered (or Counter-) Twining, common in ancient Andean flat braids (as well as in some Japanese braids, as Rodrick Owen has just informed me!) is a type of twining in which two adjacent loops are turned in opposite directions, before another loop or loops is passed through them as the traveling or hidden course. In Andean braiding, it was often another pair of loops that was passed through the countertwined pair.
If these two pairs of loops are of different colors, the braider can choose which color will show in that particular stitch / “pixel” of the braid by selecting which of the two pairs goes through the other. The color of the ‘around’ pair will be the color of that stitch. This allows the braider to plan and manipulate the color pattern of the braid.
Each pair of counter-twining loops typically stays together throughout the braid, the right loop of the pair always turning in a clockwise direction and the left loop counter-clockwise. [Yes, this can result in under- and over-twisting of the yarn in longer pieces!]
There are many examples of counter-twined braids in my Halloween post last year — Rodrick Owen and the Braids of the Mummies.
Unorthodox braid structures
These are not simple twills. As far as I know there are no corollary woven structures to compare them to. These are structures that are possible and often even very easy to make with loops held on fingers, but not easy at all to make by any other braiding method, including hand-held loop braiding.
This is because a loop held on one finger can be treated as a single element (made up of two strands), while a loop held on another finger can be treated as a shed of an upper and a lower strand. This “shed” is not equivalent to one of the alternating sheds that heddles create on a loom. Rather, that upper and lower separation marks a division between an upper braided layer and a lower braided layer.
Imagine that the loops on the second and third fingers are consistently “woven around” as if those two loops were single, 2-strand units (creating a single-layer structure of doubled braiding elements), while the loop on the fourth finger is consistently treated as a shed of one upper and one lower shank that the traveling loop is brought between (as in a double-layer braid in which the upper and lower strands of a loop interlace separately with other separate, individual strands). With each successive row, the loops on the fingers shift over by one loop—this might seem similar to a twill, but the “upper” and “lower” separately interlaced strands of the fourth finger’s loop adds a whole different dimension to the structure. See my info page on unorthodox braids, and see Noémi Speiser’s Old English Pattern Books for Loop Braiding for a more detailed analysis (you may have to learn her “track-plan” diagramatic system to follow her analysis).
* Masako Kinoshita and Noémi Speiser prefer the term loop-manipulation braiding, or finger-held loop-manipulation braiding when they want to differentiate it from hand-held loop braiding. (Masako Kinoshita then shortens this to f-h l-m braiding).
To me, the meaning of the word ‘braiding’ already includes the sense of “manipulation”, so I prefer “loop braiding” for simplicity’s sake. “Finger-held” also feels a bit cumbersome to me, but to Speiser and Kinoshita the compound word “fingerloop” or “finger loop” has a childish/ baby-talk connotation that they want to avoid. (I totally appreciate this, since loop braiding has often been overlooked as a childish string trick rather than considered a real textile technique.)
I’ve occasionally seen others (never Speiser or Kinoshita) drop the word “braiding” altogether, leaving “loop manipulation” alone. I have to admit that this really bugs me–it omits the most important word! ‘Loop manipulation’ alone could apply to knitting, crocheting, many Chinese decorative knots, a popular type of potholder-weaving, or even linking rubber bands together.
*update on definition of repp (also spelled rep): I just found out from Richard Sutherland, who researched this while writing his book Takadai Rep Braids, that the term repp is not very restrictive—it just means a ridged type of texture in a woven or braided piece. It doesn’t have a consistent structural definition, so can be used for plain weave or twined fabrics… (thanks, Richard!)
*Re twisting continually in the same direction for single-course oblique twining: If the direction of twist were to alternate between each row (effectively reversing the twist of the previous row), the result would actually be plain weave rather than twining. I have occasionally used this as a plain-weave braiding strategy, both with hand-held and finger-held loops [see my Warp-faced plain weave braids]. I haven’t heard of this “alternating turning” being used before in loop braiding to create plain weave, but I strongly suspect it was the method for certain Andean warp-faced, repp-like plain-weave braids. If so, they too may well have been made by hand-held loop braiding, which is how twined Andean flat braids are assumed to have been made. (At first Rodrick Owen disagreed with me on this but he may be coming round!)
© 2012–2015 Ingrid Crickmore
See full copyright restrictions and permissions at the bottom of the sidebar (if you are on a small screen, the ‘sidebar’ may appear somewhere other than the side of the screen). Content of this website may not be posted or “reposted” online, sold, or used in fee-based workshops without my permission. It may be shared off-line with certain restrictions – see full copyright info.
Page 1: About Loop Braiding
Page 2: A-fell, V-fell, Slentre, and hand-held loop braiding
Page 3: Too-Many-Loop Braids
Page 4: Unorthodox Braids
Page 5: Old English Pattern Books for Loop Braiding
Page 6: Alphabet braids of the 17th Century
Page 7: About Me
Page 8: Terminology (this page)