A fragment held in the Estonian Institute of History is one of the rare exceptions.

The fragment was excavated in 1949 from a Votic cemetery at Jõuga in North-Eastern Estonia. It is likely a part of a mitten cuff, since it was found at the hand of a woman in her grave. The burial has been dated to 1238–1299, making the mitten at least 700 years old and possibly closer to 800. It sat in the collection for most of the twentieth century, mentioned in passing in a 1993 article about textiles from Votic cemeteries, but the knitting technique went unexamined until researcher Anneke Lyffland published a close study in 2005.

The Votics were a Finnic ethnic group who lived along the northern shores of Lake Peipus, west of modern-day St. Petersburg. The fragment gives some sense of what this mitten was: relatively fine gauge (25 sts × 30 rows per 4” / 10 cm), worked in stranded colourwork using red, blue, and natural white. The red was dyed with madder; the blue with indigo, which had to be imported at considerable cost and was blended with undyed fibre to stretch it further. In Estonian folk belief, red was associated with blood and had protective properties; it was thought that evil and disease could not cross a boundary of red. All of this points to the mitten almost certainly being made for a wedding or a funeral, and the knitting throughout is consistently even, with the floats well managed. This was clearly not the work of a beginner.

What Lyffland found, working carefully through the stitches to correct an earlier account that had mis-described the stitch pattern, was a single crossed stitch in the solid red row. Every other identifiable stitch in the fragment is uncrossed. In knitting, a crossed stitch is one where the legs of the loop sit twisted against each other, the result of entering a stitch from the wrong side. It can be done deliberately for texture, but here, in an otherwise uncrossed piece, it is likely a mistake. The knitter slipped up once, in what was probably a row worked by firelight or candlelight, and moved on.

That mistake is the reason we know anything at all about their technique.

The two main methods of hand knitting, what we call Eastern and Western, produce identical fabric when worked correctly. A piece of stockinette made by an Eastern knitter looks exactly like one made by a Western knitter. This means that for the vast majority of surviving historical knitting, including some of the most celebrated pieces that have come down to us, there is simply no way to determine how the knitter held their yarn or wrapped their needle.

The difference between the methods lies in how stitches sit on the needle before they are worked. Eastern knitters, looping the yarn from back to front around the needle, end up with stitches oriented differently from Western knitters. Both compensate automatically as they work, entering each stitch from the appropriate side and producing an uncrossed result. But when a stitch gets worked from the wrong side by mistake, the resulting cross slants in a direction determined by the method: right-slanting for an Eastern knitter, left-slanting for a Western one.

The crossed stitch in this fragment slants to the right. Lyffland also observed that Eastern knitting tends to untwist s-spun yarn, and this effect is visible in close-up photographs of the stitches, which corroborates the conclusion. The knitter was using the Eastern method, but we can’t know whether they held their yarn in the left or right hand. According to Anu Pink, the current authority on traditional Estonian knitting, knitters predominantly knitted “English” style up until influence from Finnish craft teachers led to a switch to continental knitting.

Most of the time, when we make a mistake in our knitting, the question is whether to fix it. It is easy to feel the work is somehow compromised, or that we failed ourselves (or the recipient) by not paying close enough attention. This knitter was clearly paying attention; the rest of the fragment is meticulous. The single wrong stitch probably went unnoticed by them, and everyone around them.

The correct stitches, thousands of them, consistent and even throughout, ironically tell us very little about the hand that made them. The one that went wrong, eight centuries later, tells us quite a lot. Perhaps the mistakes we make now could be just as valuable to someone else in the future.

Further reading

Anneke Lyffland, “A Study of a 13th-Century Votic Knit Fragment” (2005). The primary source for everything in this essay. Lyffland examined the fragment held at the Estonian Institute of History (AI 4008 XXII: 156).

Anu Pink “Knitting style – the grace of noble ladies or the speed of peasant girls?”, Studia Vernacula, vol. 11 (2019). The current authority on traditional Estonian knitting, covering regional styles, techniques, and the historical shift from English to Continental method.

FO: That 70s Toe

Isn’t it strange when swatching a technique turns into a project? When I was writing about Shortrow Toes last month, I started playing around with different methods of creating the wedge toe shape. If you draw a schematic, you will see that it is two trapezoids stacked on top of each other, forming a standard hourglass shape. I immediately saw it could be broken down into 6 equilateral triangles.

Read more

The first time someone explained the symbolism of Aran stitches to me, I believed it without question. Cables for the fisherman’s rope, diamonds for the net mesh, each family’s pattern distinctive enough to identify a man lost at sea: it was a coherent, beautiful story, carrying the hush of something almost ancient, as if the salt and wind of Galway Bay still clung to every stitch. The wool, too, was left unwashed so that the natural lanolin would protect against the Atlantic weather. This tradition, in the telling, reaches back to ancient Celtic craftsmanship, encoded in the fabric for centuries before anyone thought to write it down.

It is a very good story. It is also, as far as anyone has been able to establish, largely a twentieth-century invention. The symbolism attached to Aran stitches appears to have been constructed and promoted from the 1930s onward, beginning with a German textile historian named Heinz Edgar Kiewe, who encountered the sweaters in a Dublin shop in 1936. The tradition of family patterns, the ancient Celtic connections, the mythologised wool: none of it has much documentary support. Alice Starmore, who wrote the most rigorous study of Aran knitting in 1997, concluded that the sweaters in their modern form probably date from after 1946, and that they owe more to the Scottish gansey tradition than to anything uniquely Irish. The story, she found, arrived before the sweater did.

Aran is the most elaborate case but not an isolated one. Fair Isle has a story too: Spanish sailors from the Armada, shipwrecked on the island in 1588, taught the locals their distinctive colourwork technique. The design of the patterns does suggest outside influence (Moorish or Arabic connections have been proposed), and the story has been repeated often enough to feel like established history. The difficulty is that there is no evidence for it before the 1850s, two and a half centuries after the Armada, and the story is disputed even by people on the island itself. Current thinking places the tradition’s origins closer to the Baltic (Estonia, Lithuania, Finland), arriving through trade rather than shipwreck. The Spanish Armada story is a better story: livelier, more dramatic, with a touch of swashbuckling bravado that neatly flatters the national imagination. It is also, almost certainly, a later addition.

Iceland offers a more muted version of the same story. The lopapeysa (the yoked sweater that has come to feel synonymous with Icelandic identity) carries the weight of something much older than it is, as though it has always been there. The yarn it is typically made from, Lopi, now almost synonymous with Iceland in the minds of knitters, was developed only in the early twentieth century from earlier preparations of carded wool. The sweater style itself came together from a magazine pattern a couple of decades later. In this case, there isn’t as much false history surrounding it. Instead, it is a relatively recent tradition that has slowly been woven into a broader narrative of national continuity.

Even the gansey, a fundamentally practical and workmanlike sweater tradition, has accumulated its own embellishments. The pattern on a gansey, it is often said, was specific enough to identify a fisherman drowned and recovered from the sea. Of all the fishing-sweater myths this is probably the most persistent, and also the least supported: historians of the tradition have been looking for a documented case and have not found one. Some have traced it to J.M. Synge’s 1911 play Riders to the Sea, where a body washed up on a beach is identified by a distinctive mistake in his stocking; by the early 1960s that detail had migrated, in print, to the gansey and Aran sweater. Patterns were certainly distinctive in places, and in small fishing communities a familiar garment might sometimes have helped. But as the tradition spread, patterns were copied and shared across regions, and any identifying function they had would have blurred quickly.

Of the four, only the Aran story continues to circulate at full strength. The Armada story is actively disputed by Fair Isle people themselves; the lopapeysa was never quite false history in the same sense, but a recent tradition; the gansey myths are routinely debunked, yet never quite vanish. The Aran symbolism story, sustained by a commercial export industry from the start, was designed to travel, and it keeps arriving fresh to new audiences.

None of this is unusual in the history of folk objects. But why is it that sweaters - arguably more so than any other textile or even garment - so often come adorned with such rich, fanciful, even lurid folklore?

Part of the answer is that a sweater looks as though it should mean something. The surface of a densely patterned Aran or Fair Isle garment is visibly structured, full of decisions: cables, colours, repeating motifs arranged with obvious care. That kind of surface invites interpretation. It looks like a text, and a text implies a meaning waiting to be read. Plain grey jersey does not tend to generate origin myths. A garment that looks like it might be saying something creates a demand to know precisely what it is saying.

Nation-building accounts for another part of it. In almost every case where the mythologising is most intense, there is a community that needs the object to carry more weight than the object can carry on its own. The Aran sweater’s mythology developed in the decades after a hard-won Irish independence, when the country was actively shaping its own national culture as one distinct from its coloniser across the water. The Fana sweater, worn in Norway under German occupation during the Second World War, became a silent symbol of resistance: a garment that could communicate what it was not safe to say out loud. The lopapeysa consolidated around Icelandic identity during the mid-twentieth century, following independence from Denmark. Again and again, the sweater gets conscripted at the moment when a community is deciding who it is.

Commercial motivation has played a role too, particularly in the Aran case. A sweater with an ancient story attached to it sells differently than one without. Kiewe’s mythology arrived at the same moment that Irish cottage industries were trying to build export markets, and it is not difficult to see how a romantic story about ancient Celtic craftsmanship would serve that purpose. This is not straightforwardly cynical: many of the people who repeated the stories believed them, but the commercial convenience was still there. It is, when all is said and done, great marketing.

While I think it is important to correct these myths, I don’t think we should dismiss them. Instead, we should ask what they reveal about what we want from the objects we make. A sweater that carries the weight of a family, a tradition, a place, a history is a different thing to wear than one that does not. The desire for that connection is genuine, even when the story that supplies it is not. The Aran sweater is not less remarkable because the symbolism of its stitches was largely invented in the twentieth century. The knitting is still extraordinary. The craft is still real.

The stories grew up around these garments because the garments were so spectacular that they seemed to deserve them. Pick up a sweater you’re proud of and it is easy to see why: the weight of it, the density of the stitches, the hours of work held in the wool. That was probably always enough on its own.

Further reading

Alice Starmore, Aran Knitting (1997). Examines the sweaters in the National Museum of Ireland directly and traces the tradition’s origins.

Alice Starmore, Book of Fair Isle Knitting (1989). Covers Fair Isle technique, pattern structure, and historical context.

Sheila McGregor, The Complete Book of Traditional Fair Isle Knitting (1981). A comprehensive survey of Fair Isle patterns and motifs, with material on where the designs actually come from.

Gail Ann Lambert, The Taxonomy of Sweater Structures and their Origins (MS thesis, 2002). An academic study classifying sweater construction types and tracing their documented origins across traditions.

Pierce Kehoe, The Knit of a Nation?: How Irishness was marketed to the world by the Aran sweater (1950-1980) (MS thesis, 2019). Examines how Aran sweater marketing from constructed and reshaped ideas of Irishness for global consumers.

Gansey Nation, “Popular Gansey Myths” (ganseys.com). A myth-by-myth analysis of the gansey tradition written by a Wick historian and archivist.

FO: Shetland Hap

Usually when I’m travelling I cast on a sock: small, portable, mindless. But my partner had brought me back a wool-linen blend marketed as sock yarn, and as much as I love linen, it didn’t feel right turning it into socks. So instead I went looking for a shawl pattern to take with me.

Shetland lace has always fascinated (and intimidated) me, and I’d been waiting for an excuse to dip my toe in. Haps seemed like the sensible entry point: the central square or triangle is usually plain garter, with the lace saved for the border. Trouble is, most haps lean on the feather-and-fan motif in multiple colours, and I didn’t want to be haulingn multiple balls of wool around Britain. Then I found the Aestlight shawl: one colour, manageable complexity. Perfect.

Read more

Usually when we learn to knit, the knit stitch comes first, and for many people it feels natural relatively quickly. Then comes the purl: the same motion in reverse, the yarn in front, the needle inserted from right to left. It is often described as the more awkward of the two, and the Continental purl in particular has a reputation for feeling strained at first. There are ways around it: Norwegian purling keeps the yarn at the back and avoids the motion most people find fiddly; combination knitting wraps the yarn clockwise on the purl rather than the usual anti-clockwise, which some knitters find more comfortable. It’s worth trying all of them, and worth trying more than once. I came to love Continental purling, specifically combined Continental, and now feel that my purl rows are often more enjoyable than the knit rows.

But the premise behind all of this, that mastering the purl stitch is necessary to advance as a knitter, is worth examining.

Knitting and purling are two ways of making the same basic stitch. The technique differs - the position of the yarn, the direction the needle enters the loop - but the stitch in the fabric is structurally the same either way. From the right side, a stitch made by knitting shows as a smooth V, which we call a knit stitch; one made by purling shows as a bump, a purl stitch. When you work flat in alternating rows of knit and purl, you get stockinette: smooth on one side, bumpy on the other. When you work only knit stitches flat, you get garter stitch, bumps on both sides, a thicker and more elastic fabric. To many, garter stitch is associated with beginners.

Shetland lace, which is widely regarded as among the finest hand knitting ever produced, is worked entirely in garter stitch. There is not a single purl stitch in a traditional Shetland lace shawl. This does not make it beginner’s work. The patterns are complex and delicate, the finished objects are remarkably fine, and they are built entirely from the knit stitch, worked flat, with decreases and eyelets forming the complex patterns.

Conversely, Fair Isle garments are typically worked in the round, which means stockinette is produced using only the knit stitch: every round goes in the same direction, so the purl side never faces you. These are some of the most iconic knitted garments in the world, and at their most elaborate they involve many subtle colour changes within a single motif. For sleeve types which would usually necessitate working flat above the armholes, such as drop shoulder and set-in sleeves, steeks are added instead. These are extra columns of bridging stitches worked into the fabric at the point where an opening will eventually be cut. The knitter continues in the round for the entire garment without ever purling back; the steeks are cut through at the end to open the fabric. There are even examples of Fair Isle garments knitted flat with the right side always facing the knitter, requiring the yarn to be cut and rejoined on every row.

Working in the round is a very old method of knitting. Surviving early examples of knitted objects, fragments from Egypt, Spain, and the Middle East from the medieval period, tend to be worked circularly. The assumption that flat knitting is the default and circular knitting is a specialist technique has things a bit backwards. Producing stockinette by working flat, knit one row, purl one row, appears to have become more common around the turn of the last century, as fashions changed and the sweater became an item of everyday clothing. The expectation that knitters work flat pieces and seam them together, in the long view of the craft, is a relatively recent one.

The purl stitch itself does appear in some of the earliest knitted fragments found in Egypt, but that knowledge did not carry into European knitting. Richard Rutt, in A History of Hand Knitting, concluded that the purl was a late invention in the European tradition, and the earliest datable European purl stitches are on the stockings that Eleonora de Toledo, wife of Cosimo de Medici, was buried in around 1562. Before that, it seems that European knitting was worked without it.

There is a further complication: Portuguese-style knitting, found not just in Portugal but also in Turkey and parts of South America, treats the purl as its primary stitch. The yarn is tensioned around the neck or on a pin at the chest, with the purl side facing the knitter, making the knit stitch the awkward one. It’s even possible that this style could be the oldest form of knitting, which adds another layer to the question of which stitch is foundational and which came later.

None of this means purling is unnecessary. Ribbing requires alternating knit and purl stitches within the same row, and ribbing is hard to replace. On my knitting machine, which produces only stockinette, the workaround is to ladder down a column of stitches and latch them back up from the other side, turning them from knit to purl one by one. It works, but it is not fast. Textured fabrics, cables, seed stitch, most of the stitches that define the gansey tradition, are built from combinations of the two, and ganseys are some of the most technically complex and historically significant garments in knitting. Even those are often worked in the round with plain stockinette making up the body, so the purling is concentrated in the sections around the chest.

I am not arguing that knitters should avoid the purl stitch or stop working at becoming more comfortable with it. Getting comfortable with purling allows you access to the whole world of knitting, and I’ve come to love it. The point is that a preference for purl-avoidance tends to get treated as beginner behaviour, something to grow out of, when the preference is neither new nor limiting. Some of the most admired knitting in the world was built around this aversion. The current popularity of stockinette sweaters worked in the round is not a modern shortcut; it is a very old preference, and I don’t think it’s going away.

Further reading

A History of Hand Knitting, Richard Rutt: the standard academic reference on the subject, and the source for much of what we know about when the purl stitch arrived in Europe.

The History of Hand Knitting, V&A: a broad overview of the craft’s development, from the earliest surviving examples through to the present day, with pieces from the V&A collection.

British Knitting Traditions, V&A: on the regional traditions of the British Isles, including Fair Isle, Shetland lace, and the gansey.

History of Hand Knitting, KC Guild: a readable overview with good detail on the purl stitch’s timeline and earliest documented examples.

One option is to simply pick up the stitches from the cast-on edge and start knitting, but this leaves a noticeable ridge on the inside of the work. A provisional cast on gets around this by leaving the cast-on stitches live and ready to work in the opposite direction. Depending on the method, those held stitches are kept in place either by waste yarn (which is removed later) or by a spare needle. The most common waste-yarn approach in hand knitting is the crochet chain; the most popular spare-needle method is Judy’s Magic Cast On (JMCO).

I have used both extensively and I am a big fan of JMCO. I have also used a long-tail cast on as provisionally, with waste yarn as the thumb yarn, and main yarn as the index finger yarn. After knitting, you can pick out the waste yarn, though it does not unravel as cleanly as the crochet chain does.

When I took up machine knitting a few months ago, I found that the standard approach to starting most projects there is to cast on with waste yarn, knit a few rows, and then join the main yarn. Waste yarn is king in machine knitting: the only extra step is a short stretch of knitting before the main yarn begins. The result is a provisional cast-on: the first row of main yarn stitches stays open and can be picked up later to work in the opposite direction, make a hem, or bind off. The waste yarn then comes out with a single snip and a pull. The stitch structure is the same in hand knitting, so I decided there was no reason not to try it.

What I like about this is that it requires nothing you do not already have: no crochet hook, no spare needle, no new technique to learn. The cost is a few rows of waste yarn at the start, which is not much for most projects. I did once use this on a 250-stitch cast on and found myself wondering whether the crochet chain might have been faster, but for anything small it is not a consideration.

Note, I have only tested this working flat, not in the round, although I can’t see a reason this would cause a problem.

Tools needed: project needles, waste yarn, main yarn, scissors or snips, a smaller needle for picking up stitches (optional).

Step 1. Cast on the required number of stitches in waste yarn, using any regular method (long-tail, knitted cast on, backwards loop, cable cast on, etc.).

Step 2. Knit at least one row in waste yarn. Two or three rows is better.

Step 3. Knit one more row in waste yarn, keeping your tension looser than usual. This will make it easier to pick up the stitches later.

Step 4. Join the main yarn and knit one row in plain stockinette. This is a set-up row; the following row will be the first row of your pattern.

Step 5. Work your pattern as instructed, until it is time to pick up the stitches from the provisional cast on.

Step 6. With the wrong side facing, pick up the first set of loops in your working yarn. A smaller needle makes this easier, though it is not mandatory.

Step 7. Carefully snip a stitch in the waste yarn on the side opposite to where the tail sits (the final row of waste yarn knitted).

Step 8. Pull on the tail of the waste yarn from that last row. It should come away easily, and the waste yarn will simply fall free.

Step 9. You can now work your pattern in main yarn as instructed.

This simpler provisional cast-on works well when your stitch count is small, since knitting a few rows of waste yarn is not much of an imposition, or when you do not have a crochet hook or spare needle to hand. I have used it when making short row toes, a technique that machine knitting prompted me to revisit in hand knitting. If you would like to find out more about short-row toes, you can read about them here:

I had been keeping an eye on Chase Harpole’s Earthbound for a couple of months and had set my mind on casting on this spring. It is a seamlessly constructed, top-down sweater in sport weight yarn, with wide, even stripes reminiscent of a classic rugby shirt. What drew me to it beyond the look was the sizing: forty adult sizes grouped by shoulder width rather than chest measurement. Shoulder width is, in my experience, the dimension that matters most for a good sweater fit, and seeing it taken seriously from the start was immediately appealing.

I had also been dreaming about a wool I’d seen at my local yarn store. Semilla Melange from BC Garn is a woolen-spun sport weight yarn that comes in a refined selection of muted colours. There were two colourways I couldn’t stop thinking about: a deep chocolate and a dark navy, both shot through with small flecks of orange and red. Fazer, Finland’s most popular chocolate brand, uses a similar navy in its packaging, and I’d always thought the chocolate looked particularly chic when peeking out of the packaging. The pattern and yarn seemed to belong together, and everything was on track for it to be on my needles as the snow was melting.

Then, in mid-March, I ended up getting a knitting machine. I previously wrote about how this was something I never thought I’d get into, but it somehow just happened. After a few simple hats, pairs of socks, and a slightly more advanced lace scarf, I decided it was time to tackle a sweater on the machine. My plan was to ease myself in with something simple: a drop shoulder sweater, worked in pieces, bottom up. No unusual construction, no conversion work, just learning the machine’s habits before adding complexity. This is a reasonable approach for a beginner, so of course I abandoned it almost immediately, and decided I would make Earthbound on the machine and work out the adaptations as I went.

The problem was that it is a seamless, top-down design with set-in sleeves, intended for hand knitting. Working in the round isn’t possible on a single-bed machine, and seamless top-down construction isn’t something I had seen instructions for in machine knitting. Still, I decided it was better to make a sweater I truly wanted than something just for the sake of it. So, away I went, overly confident and perhaps a bit naive.

After six swatches (yes, you read that correctly) I still couldn’t match the pattern’s gauge on my machine. I settled on the fabric I liked best and decided to recalculate the instructions from the schematic rather than keep chasing the original gauge. This also gave me the opportunity to customise the dimensions to my own body rather than choosing the nearest graded size. I took careful measurements and worked through the arithmetic, checking each section against the schematic. The previous time I had trusted a size chart over my own measurements, I ended up with sleeves 10 cm too long. Although I had more confidence in the designer this time, I realised that nobody’s body neatly fits into any one box, and in my case the arms appear to be the significant outlier.

Rewriting the pattern into my spreadsheet (running totals, row-by-row increases, calculated widths and lengths at each stage) was where I encountered something unexpected. The back yoke depth was considerably shorter than the front. I had never seen this in a hand-knitting pattern before and spent some time working out what it meant.

The construction goes by several names, among them European shoulders, English tailoring, and fully fashioned shoulders. The terminology is not consistent, and the sources I found do not agree on what distinguishes one term from another. “European Shoulder Seam” appears to be a relatively recent coinage used as an umbrella for what is variously called English, Italian, or French seams, all of which describe the same basic principle: placing the shoulder seam on the back of the shoulder rather than on top. The front piece carries no shoulder shaping at all and just runs straight across. When the garment is assembled it torques naturally into position over the shoulder. This seems like it shouldn’t work, but the more I thought about it, the more it started to make sense.

This method has its origins in commercial knitwear, with the motivation being efficiency. Concentrating all the shoulder shaping on one piece meant that the other piece could be knitted without any shaping at all, making it extremely efficient. The resulting shoulder seam ends up sitting on the scapular spine, which is said to be where it should be for an anatomical fit. Designers had taken that industrial tradition and brought it into a hand-knitting pattern, and I was then taking it back to a machine. If anyone has a clear source on the distinctions between the different terms for this construction, it would be great to know.

I sat down at the machine the morning after finishing the spreadsheet, planning to get most of the sweater done over the weekend. The top of the back yoke begins with short rows, which on a machine require the fabric to be weighted evenly below the needle bed. Directly after a cast on there is very little fabric to hang weights from. After a couple of false starts I found a solution: cast on with waste yarn first, knit a couple of inches, then work a crochet cast on onto that base. The waste yarn gives enough fabric to attach weights, and once the short rows are done the knitting proceeds cleanly.

The pattern also includes a contrast-colour detail on each sleeve cap, worked in short rows using a small section of intarsia. I had bought a second-hand intarsia carriage for exactly this kind of work, but it was still in the UK. Working the intarsia section by hand while also managing short rows felt like too many variables to track at once, so I left the detail out. It had been one of the things I loved most about the pattern, and dropping it was a bit of a disappointment. The sleeves look fine without it, and I’m glad I made the decision to simplify things slightly.

After those early obstacles, the knitting itself moved quickly. Each sleeve took about an hour on the machine, which makes the case for second-sleeve syndrome quite weak. Seaming and the collar followed without drama. The overall ease was, I think, a product of how the pattern was written: the instructions are clear enough that converting them, while involved, was mostly a matter of working carefully through what was already there. A well-written pattern is one you can adapt.

I have decided it is the best sweater I have ever made. The fit is better than anything previous, partly because I calculated the dimensions to my own body rather than selecting from a chart. The stripes are exactly what I hoped. The Semilla Melange is soft enough to wear against the skin, something I wasn’t certain of until the fabric was blocked and fully relaxed.

The week after finishing it, I wore it on a day out at Seaton Sluice with family. There is something different about wearing a new sweater through a full afternoon outdoors with people who know you well, compared to putting it on for a day home alone. That afternoon is when it started to feel finished.

The European shoulder construction is something I want to try again moving forward. I hadn’t come across it before this project, and now that I understand what it does, I can see myself applying it in future work regardless of whether the original pattern calls for it. The fit across the shoulders is the part I am most pleased with.

I plan to make the next machine-knit sweater in flat pieces, closer to how the machine is typically used. I may come back to Earthbound in hand knitting at some point. For now, the result was better than I had any right to expect from a project I entered by ignoring all of my own advice.

FURTHER READING

Pattern: Earthbound by Chase Harpole

Yarn: Semilla Melange by BC Garn

On European shoulders and fully fashioned construction:

- “Zen Inspirations: The European Shoulder Seam” — East London Knit (2016). Good introduction to the construction and how it came to be named.

- “Full Fashioned Shoulders” by Leslie Gonzalez — TKGA. This covers the history, the mathematics of shoulder slope, and why the seam sits where it does.

Before the Industrial Revolution, almost every household spun. Not as a hobby, and not occasionally: spinning was a constant background activity, fitted around everything else. Women and children picked up a spindle at any spare moment throughout the day, and this yarn would go on to weavers who would produce cloth. Even with the spinning wheel, which had greatly increased productivity over the spindle, it is estimated that eight spinners were still needed to supply the yarn for a single weaver’s loom.

The flying shuttle, invented in 1733, made the bottleneck in spinning worse. By allowing one weaver to do the work of two, it meant looms could run faster and demand more yarn than before. Thirty years later, the spinning jenny arrived, and with it, the ability to spin multiple threads at once. Production began moving out of homes and into mills, amplified by water power and then steam. This is how the familiar shape of industrialisation emerged: factories, mechanised production, large-scale labour systems, and supply chains that reached across the world. Those supply chains were already global before the end of the eighteenth century. The rapid expansion of British textile production depended on cotton, much of it grown by enslaved people in the Americas. Mills processed that cotton into cloth, which was then exported around the world.

Within a generation or two, the domestic spinning that had occupied women and children for centuries had for the most part disappeared. The knowledge stopped being passed down, and weaving followed much the same path. Hand spinning and weaving exist today, but as specialist crafts, not as things most people do or learn as a matter of course. Knitting did not follow that trajectory.

* * *

Part of the explanation is structural. In 1589, William Lee invented the stocking frame, a machine that reproduced the motions of hand knitting using rows of hooked needles. It could produce stockings quickly and consistently, and it eventually spread across parts of England. But instead of giving rise to factories, it gave rise to the framework knitting industry: machines operated in homes and small workshops, with workers producing goods for merchants, often using frames that they rented rather than owned. Mechanisation arrived, but the broader reorganisation of labour that defines industrialisation did not follow.

There was less economic pressure on knitting than on spinning, too. Industrialisation tends to accelerate where systems are slowest, and spinning was the slowest part of textile production by a wide margin. Until that problem was solved, there was limited incentive to mechanise anything else. Woven cloth was also far more central to early modern economies than knitted goods, since it was used for uniforms, sails, trade goods, and household fabrics, all of which were produced and traded on an enormous scale. Knitted goods (stockings, caps, gloves) were vital to daily life but not to the large-scale economic systems that attracted investment and drove mechanisation. The pressure went toward spinning and weaving, and knitting was largely left to develop on its own terms.

It was also well adapted to domestic life in ways that made it difficult to dislodge. Knitting requires very little equipment, it is portable, and it can be set aside without much loss. It had always been done alongside other work: tending animals, managing a household, filling the long winter evenings. This was not a seasonal or marginal practice; across most of Britain, for most of this period, knitting was simply part of how households ran.

* * *

Structural reasons explain why knitting was not industrialised in the same way as spinning and weaving. Why hand knitting survived as a widespread domestic practice when hand spinning did not is a separate question, and more a matter of cultural history than economic history.

Spinning was useful and universal. As it moved into factories, there was no longer any reason to teach it to people who did not need it, and it carried no particular social value that might have kept it alive in homes where it was no longer economically necessary.

Knitting, for most of its history, had been much the same: a form of paid labour or household necessity, unglamorous and widespread. By the Victorian period, something had shifted. It had become a respectable leisure activity for women of the middle and upper classes, done in drawing rooms and at social gatherings. Victorian society had elevated “useful” needlework as a visible expression of feminine virtue and respectability. This drove a publishing economy around it: pattern books, magazines, instructions for increasingly elaborate work. The knowledge was actively circulated to people who came to it for pleasure rather than necessity. At the same time, knitting was being taught in workhouses and charitable schools, to orphans and the poor, as a skill that could generate a modest income. It functioned simultaneously as a polite accomplishment and as a form of poor relief, and both uses kept it in circulation.

Wartime reinforced it further, starting in the 1850s with the Crimean War, where women were encouraged to send “templar hoods” or “helmet liners” to the front, which eventually became known as “balaclavas”, named after the site of an important battle. Both world wars also produced large-scale civilian knitting campaigns, with millions of people producing clothing and equipment for troops. The campaigns were organised, publicised, and socially pressured in ways that sustained knitting through periods when it might otherwise have declined.

None of those reinforcing mechanisms existed for spinning. By the time the factory had made domestic spinning unnecessary, there was nothing to sustain it.

* * *

Crochet’s story is quite different from knitting’s, despite the crafts sharing many similarities.

Its origins are disputed: some trace it to Europe in the early nineteenth century; others argue for older roots. What is clearer is that by the mid-nineteenth century, it was popular in Britain and Ireland both as a hobby and as a way of supplementing income through lace-like work. Queen Victoria crocheted and reportedly gave away items she had made, which gives some sense of how fashionable it had become at the highest levels of society.

Where crochet’s story diverges most clearly from knitting’s is in its relationship to machines. Knitting was eventually industrialised: the flat-bed and circular machines that produce most knitted fabric today are true knitting machines, reproducing what hand knitting does at an enormous scale. Crochet has not been mechanised in the same way. Machines that produce crochet-like fabric typically do so using knitting technology, imitating the look rather than reproducing the construction. Crochet is formed from a single active loop, continuously manipulated with a hook, and that motion has complications for industrial automation.

Crochet’s story, then, is less about surviving the industrial revolution than about bypassing it. Part of the reason it was never industrialised is simply that by the time it became widespread, the textile machinery already in existence had been built around different processes, and there was no obvious incentive to start again from a different direction.

* * *

Most knitted fabric today is made in factories. The circular machines that produce jersey and the flat-bed machines used for shaped knitwear are largely automated, part of the same global supply chain that the textile revolution created. The industrial version of knitting is real and here to stay.

Hand knitting persists alongside it. I previously wrote about how, even with a knitting machine, I would not stop knitting by hand. They scratch a very different itch. That distinction is probably what kept the hand version alive through everything else: there was always a reason to keep doing it, and enough people across enough generations kept passing that reason on.

Further reading

The history of hand knitting, V&A: a broad overview of knitting’s development, from medieval guild workers to the present day.

Knitting for Victory, The History Girls: on the organised wartime knitting campaigns, from the Crimean War through to the Second World War.

Knitting and domestic life, Victorian style, Modern Daily Knitting: on knitting as a respectable drawing-room accomplishment in the nineteenth century,and the publishing economy it produced.

The sleeves turned out to be comically long, about ten centimetres past where they should end. This goes well beyond the territory of folding up your cuffs and into something harder to explain.

My first feeling was embarrassment, which gave way fairly quickly to genuine puzzlement. I rewrite every pattern I knit into a spreadsheet before I start: running totals of width and length at each stage, cross-checked against the schematic, so I can catch problems before they are knitted into the fabric. If there is shaping involved, I track how many increases happen on each row, add them to the stitch count, and update the width. For length, since I know my gauge, I can calculate exactly how many rows each measurement should take. It is the kind of system that is supposed to prevent exactly this, and somehow it did not, and I have not yet worked out why. I don’t feel like investigating the error right now, so my current way of dealing with it is exclaiming “Look at my sleeves!” as soon as I see people. It is not the most rational response, but it does seem to be helping.

* * *

The next sweater I am planning is also a set-in sleeve construction, which puts the same category of error squarely back on the table. This time I am being more thorough than usual. My partner spent Saturday morning helping me measure different parts of my body, tape measure and notepad in hand, patient throughout in a way I was grateful for. I have also been working through Faina Goberstein’s Craftsy course, Sizing Knitwear Patterns, which goes through how to take body measurements and calculate each piece of a sweater. Set-in sleeves are less forgiving than other construction styles, and the extra preparation may feel fussy, but in my case it is clearly necessary.

I am hoping the preparation makes a difference. Set-in sleeves are unforgiving enough that I am not expecting certainty, but at least this time I will have done the groundwork.

* * *

The week was not all fit disappointment and anxiety. A friend turned thirty recently, and I had knit him a pair of cabled socks as a gift. He messaged to say they fit perfectly and he’s already looking forward to the cooler season again. It is a reasonable way to end a week that started with a phantom ten centimetres of ribbed fabric.

I have knit a lot of socks by hand, both cuff-down and toe-up. Over time I have tried most of the common heel constructions: heel flap and gusset, short row, afterthought and instathought (peasant) heels, Fleegle, and a few others. My toes, however, have almost always been wedge-style: four increases or decreases at the sides of the foot, finished with grafting for cuff-down socks or begun with Judy’s Magic Cast On for toe-up.

Since I started making socks on my flatbed machine, I noticed that most machine knitting patterns use a short row toe as well as a short row heel. This makes complete sense on a machine, where short rows are much faster and easier than decreases. Seeing these instructions reminded me that short row toes are equally possible in hand knitting, yet they do not seem especially popular.

In this post, I will explain how to work short row toes for both toe-up and cuff-down socks and suggest modifications for a more rounded shape. I will also look at why they are not more widely used, despite their advantages.

Why Short Row Toes Work

First, consider how a standard cuff-down wedge toe works. You make four decreases every other round until you have roughly one third of the original stitch count remaining, then graft the ends together. This creates a trapezoid-shaped pocket that conforms to the shape of the foot when worn.

If you have ever knit an afterthought or peasant heel, you may have noticed that the shaping process is identical to a wedge toe. In a toe, this shaping reduces the circumference of the sock as the foot narrows. In a heel, the same shaping increases the circumference around the widest part of the foot and introduces the angle that allows the fabric to transition from ankle to foot.

Short row heels can also replace peasant-style heels, producing a heel that is effectively identical in shape to one made with decreases. From this, it follows that the decrease-based shaping at the toe can equally be replaced with short rows.

Toe-Up Construction

For toe-up socks, begin with a provisional cast-on for half of the total stitch count. For a 64-stitch sock, this means casting on 32 stitches provisionally. Any provisional method works. A crochet provisional cast-on is the most common, but Judy’s Magic Cast On can also be used provisionally.

From there, work the toe exactly as you would a short row heel, using your preferred short row method: wrap and turn, German short rows, yarn-over short rows, or shadow wraps. Once the toe is complete, pick up the live stitches from the provisional cast-on and join in the round.

This approach is intuitive if you already knit short row heels. There is no new technique to learn, only a different starting point in the sock. It also appears to have been more common before 2006, before Judy’s Magic Cast On was popularised. For knitters using DPNs or short circular needles, this method avoids switching to magic loop, which is often required for the initial rounds when working toe-up. It also eliminates grafting, which many knitters are happy to avoid.

The main drawback is the provisional cast-on itself. Some knitters find it fiddly, and if you already use Judy’s Magic Cast On comfortably, the short row toe may not feel like a compelling alternative.

Cuff-Down Construction

For cuff-down socks, knit the foot to the point where the toe begins, then place half of the stitches on hold. This can be either the instep stitches or the sole stitches. The remaining stitches are worked back and forth in short rows, exactly as for a short row heel. Once the toe shaping is complete, graft the toe stitches to the held stitches.

One practical advantage is that, if you are using short circular needles, you can complete the entire sock without switching to DPNs or magic loop. You can stay on the same needles from cast-on to graft.

Grafting is not eliminated, only moved. With a cuff-down short row toe, the graft runs across the full stitch count rather than the smaller number used in a wedge toe. This can make it harder to achieve even tension, so it is worth practising before using this method on a finished sock.

Modifications: A Rounder Toe

Some knitters find a standard short row toe, like a standard wedge toe, slightly too pointed for their foot shape. With decrease-based wedge toes, a common adjustment is to work decreases every other round for the first half of the shaping, then every round for the second half. This produces a noticeably rounder result.

A similar adjustment works for short row toes. In the later stages of the shaping, instead of turning one stitch before the previous turning point, turn two stitches before it. This subtly increases the rate of shaping at the tip and produces a rounder toe.

Closing

Short row toes are not a better method, just a different one, and knowing them gives you more options. They are particularly useful if you already knit short row heels, if you work on short circular needles, or if you want to avoid closed cast-ons for toe-up socks. For cuff-down socks, they are a reasonable alternative to a wedge toe, with the trade-off that the graft is longer.

I ended up exploring them more after getting into machine knitting, where short rows are the default for both heels and toes. That experience nudged me to revisit the technique in hand knitting, and from there I started experimenting with stripes. With a short row toe, stripes can curve around the toe in a way that resembles intarsia, without the bother of multiple bobbins. It is something I have been developing and may turn into a future pattern.

Whether any of this makes short row toes worth trying depends on what you already do and what you find fiddly. In sock knitting, it is always good to mix things up, but in the end the method that makes you happy is the right one.

In March I took a long weekend trip to Belgium with my partner. Since it’s such a compact country and the train network is extensive, you can cover a lot of ground without much planning. We made day trips to different cities, one of which was Namur.

Namur is an hour south of Brussels by train, where the River Sambre meets the River Meuse. The regional capital of Wallonia, it has an impressive medieval citadel on a hill, and a pleasant old town. We had no particular agenda, so we just walked.

After a somewhat nerve-racking cable car trip, we continued to stroll in the old town. There are plenty of small boutiques, and the terraces were lively and inviting, even on a chilly March afternoon. We eventually came upon a shop window that grabbed my attention, Atelier-53, and wandered in. It was the kind of place that takes a moment to categorise: high shelves of yarn alongside bolts of fabric, second-hand needles in a jar, a general sense that everyone who came in was there to make something. The yarn selection was small but clearly thought about, the kind of muted colours and unique yarn compositions that make you want to sit down and plan a project.

The owner was happy to chat, and I mentioned that I was drawn immediately to the coned yarn I could see tucked away near the counter. The week before I had acquired a knitting machine, and cones were suddenly relevant to my life in a way they weren’t before. I got talking to the owner, and she told me about the yarn. It was called ‘Mouton ultra locale’ and was unique to the store. The wool was from the Shropshire breed of sheep and originated from a single farm in Havelange, 40 km away, before being washed in Verviers and spun in France. One ply, and a beautiful natural colour, it was finer than I expected. She gave me the details and I was in love.

I spent some time working out whether a cone would feasibly fit in my backpack. Budget airlines have a way of making you do this arithmetic in real time, but I figured it would fit, so I bought it.

* * *

The day before, we had been to the Fashion and Lace Museum in Brussels, and that was where I learned that Brussels lace was a thing, or rather, that it was the thing. A type of bobbin lace, it has a tradition going back to the 15th century, and was considered the finest in the world for a long time. The fabrics were made mostly from local linen, painstakingly fine work produced by hand.

Unfortunately the technique is a dying art, but standing in that shop in Namur with a cone of local wool in my hands, I knew that knitted lace and fine, single-ply yarn would be a perfect match.

* * *

The stitch pattern came from a tutorial by Elena Berenghean, who runs the YouTube channel knitology1x1. She focuses on open work and hand-manipulated stitches for the machine, and her tutorials are very clear, even for a beginner like me. The pattern makes elongated diamonds: cables paired with vertical ladders running the length of the fabric.

On a knitting machine, ladders are straightforward. You push needles out of work and leave them there. The yarn passes over the gap on every row and the column stays empty. A clean vertical line in the fabric, no drama.

I have since tried to translate the same pattern to hand knitting. It is the same fabric, technically, but the process is different enough that it barely feels like it. I made the ladders with dropped stitches, which means you need to plan which columns will drop and keep track of them throughout. On the machine, you notice the empty needles and work around them automatically. By hand, the empty columns are invisible until you drop the stitches, so you have to hold the structure in your head. It requires more concentration than I could maintain for a whole project.

I’ve also ordered a copy of Elena’s book, Lace and Open Fabrics, which covers machine knitting techniques for exactly this kind of work. There are a lot of stitches in there that I can’t wait to try and I hope to incorporate them into more projects in the future.

* * *

The scarf is around 22 cm wide and 170 cm long (8.5“ x 67”). There is more handling of the stitches than you may think in machine knitting, but once I had swatched and calculated the pattern, the knitting on the machine only took about a weekend. If I hand knitted this, I’d probably be writing to mark the 1 year anniversary of the Belgium trip.

The yarn has been a pleasure to work with. Single ply needs a little care with the transfers and cables, as there’s always a risk it will break. Keeping a loose tension helped avoid this, and the final fabric blocked out beautifully. It is not Brussels lace, but it came from that afternoon in the museum, and from a farm just outside Namur.

I did not go to Belgium planning to make a lace scarf, and, if I’m honest, I didn’t know that Brussels had a lace museum, or even that Namur existed. I thought maybe I’d buy some sock yarn if I spotted some, but ended up coming home with much more.

I’ve been hand-knitting for a number of years, and now that I’m comfortable with most techniques, I often find myself with plenty of creative ideas and not enough time to test them. A few weeks ago, a vintage Brother KH585 from the mid-sixties appeared on a local second-hand platform at a price I couldn’t justify refusing. I thought, “This is perfect, finally I can test all of my knitting ideas quickly and easily, without spending hours knitting swatches by hand. It’s the same craft, just quicker”.

While that assumption turned out to be incorrect, I don’t feel misled or disappointed, but rather surprised. I expected to extend a craft I already had, and instead I seem to have started a different one.

The first thing that surprised me was how mechanical it felt. I’m not a very handy person. I don’t have a feel for engines or an instinct for how things fit together physically, despite being a scientist. Opening up a sixty-year-old knitting machine and discovering, from an old manual, that I needed to oil the needle butts was not something I felt prepared for. There’s a whole vocabulary of levers and cams and carriages that has nothing to do with knowing how to knit, and learning it felt less like picking up a new technique and more like becoming, reluctantly, a small-scale mechanic.

I pushed through mostly because I had nothing to lose. The machine was cheap, my frustration was real, and stubbornness is a reasonable substitute for confidence.

The KH585 dates from the mid-sixties, and the core mechanism runs without complaint. Sixty years is a long time for any piece of equipment to remain functional, let alone precise enough to use for serious work. It suggests that the people who designed and built these machines were thinking about longevity in a way that feels uncommon now. A phone bought today will likely be replaced before it breaks. A machine tool from 1965 is still doing what it was built for.

The machine has its own parameters, its own logic, its own constraints. Gauge behaves differently, and swatching becomes non-negotiable. Certain things that are trivial by hand become genuinely convoluted on the machine, and vice versa. Tuck stitches, for example, are remarkably easy on a machine, while doing them by hand can feel like a real slog. On the other hand, knit-purl textures and especially cables, are not automated on most domestic knitting machines. Garter stitch, which is usually the first fabric we learn to create as hand knitters, is not a stitch for beginners on the machine. Something meditative by hand becomes fiddly and prone to dropped stitches.

Cables have been known as “the great equaliser” of knitting machines: no matter how fancy your machine is, you make cables by transferring stitches individually using a transfer tool. It is (relatively) slow, precise work.

I have started to see hand knitting and machine knitting as two separate crafts, that share a common structure. The basic mechanics of both involve a loop pulled through a loop, and they can produce the same kinds of fabric with the same sensibilities, but they ask very different things of the crafter. Rather than one being a faster version of the other, they feel more like siblings. Related, recognisably from the same family, but with their own distinct characters. Modern hand knitters tend to see knitted fabric stitch by stitch, and enjoy great freedom in being able to try garments on intermittently while they’re being made. When you’re machine knitting, you often think in entire rows rather than stitches, and you are inherently limited by the gauge and number of needles on your machine. In my experience over the past few weeks, the actual knitting time on the machine is short, but the planning and finishing can be a much lengthier and less avoidable part than I’m used to.

* * *

I’ve often seen hand-knitters act dismissive of machine knitting in a way that doesn’t really hold up historically. The knitting frame was invented in 1589 by a Nottinghamshire clergyman named William Lee. Elizabeth I refused to grant him a patent, reportedly on the grounds that the machine would deprive hand-knitters of their livelihoods. Eventually, the hand-knitting economy did decline, but in its place sprang entire regional economies based on the knitting frame. By the early nineteenth century, there was a large class of skilled framework knitters whose wages and working conditions depended entirely on how those machines were used.

Those machines produced fully fashioned hosiery: stockings and other garments knitted directly to shape on narrow frames, seamless, made in one piece. It was skilled work, regulated by trade agreements, and the quality of the fabric was the basis of the whole industry. What changed in the early nineteenth century was that manufacturers began using wide frames to produce cut-up goods instead. Wide frames produced a flat fabric that was then cut and sewn into shape. The seams showed. The items wore out faster. Wide frames could also be operated by cheaper, untrained workers.

The Luddites existed in this context. Their name now gets used casually to mean anyone suspicious of emerging technology, which erases what was actually happening: skilled workers in Nottinghamshire, Yorkshire, and Lancashire watching manufacturers use wide-frame machinery to produce inferior goods and bypass apprenticeship rules. The machines weren’t new, and the workers knew how to use them. What they were disputing was what the machines were being used to produce, and who got to decide. They had petitioned Parliament for statutory protection and been refused. When they started smashing frames, they were selective about it: the targets were specific hosiers known to be using wide frames or paying below-agreed wages, not machinery in general.

The government’s response was to make frame-breaking an offence punishable by death. Lord Byron spoke against the bill in the House of Lords in 1812, arguing that the government was prepared to hang starving men over a trade dispute, and asking whether their lives were worth less than the frames they had broken.

Knitting has been tangled up with questions of industry, labour, and technological change for as long as anyone has tried to do it at scale. The handmade and the mechanical have never really been opposites; they have been intertwined for four hundred years.

The urge to legitimise craft nowadays is worth sitting with, because it’s largely shaped by how we think about time. Under the logic most of us have absorbed, leisure needs justification: it was earned, or it produced something, or it demonstrates something about you. Hand-knitting a sweater costs far more time than buying one. What it buys instead is a demonstration of patience, skill, and a certain freedom from efficiency. That’s not a criticism. It’s just worth naming: the insistence on handmade is at least partly a way of showing what your time is worth, or what you can afford to do with it.

The machine sits awkwardly here because it looks like a concession to efficiency, and efficiency, in this framing, is the enemy. But the claim that machine knitting isn’t “really” handmade starts to collapse under any pressure. The five-euro t-shirt we can buy at the click of a button may not feel handmade, but it was sewn by a person operating a sewing machine. The fabric was woven on an industrial loom by someone, somewhere. Fully automated textile production is mostly a fiction; there are human hands in all of it. What we’re actually debating isn’t whether hands were involved (they always are) but which hands, doing what, for how long, and whether that counts as craft or merely labour.

When I bought the machine, I expected to knit faster. Instead, it feels like I’ve gained a second craft, related to the first but distinct from it, with its own history I’m only beginning to investigate, and its own aesthetic possibilities I haven’t fully worked out yet. That’s a lot more than I was shopping for.

* * *

This is the beginning of a longer look at both crafts: machine knitting and hand knitting, their shared history, the communities keeping them going, and what each produces when you push it. More on all of that soon. If you work with either or have been thinking about starting, I’d be glad to hear about it in the comments.

Further reading

Crich Parish: framewreckers, a local account from Derbyshire, near the centre of the dispute.

Kevin Binfield on the Luddites, from the editor of Writings of the Luddites (2004).

Unmaking as Emancipation (CHI 2023), on Luddism, selective refusal, and social justice.

This Substack is reader-supported. To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe now