This is the third part of our four part (I, II, IV) look at the production of textiles, particularly wool and linen, in the pre-modern world. Last time, we processed our raw fibers, removing extraneous material and getting them ready to be turned into thread. Today we’re going to continue with that next step, spinning our fibers into thread before weaving that thread into cloth.
We are going to discuss both of these processes (spinning and weaving) this week mostly in the context of household production. Markets in textiles, as we’ll see next week, were a major part of the economies of the pre-modern world and so the commercial production of cloth is something that appears fairly quickly as agrarian societies grow more complex; almost any society with a significant degree of urbanism is producing at least some cloth for the market. That said, for reasons we’ve already discussed, in most pre-modern societies, the great majority of people (who were farmers) remained substantially disconnected from markets in much of their lives. Consequently, at any given time, while commercial textile production was significant (particularly from a trade-and-commerce lens), generally the majority of clothing being worn was likely to be produced within the household economy (though that doesn’t preclude specialization). So while we will discuss commercially produced textiles next week (along with the more frequently commercially-oriented fulling and dying processes), this week we will focus on spinning and weaving through the lens of household production, which in turn means that we will focus on spinning and weaving through the lens of the women who mostly did it in the household.
Fortunately, the processes for spinning and weaving do not meaningfully change based on the context of production (household or commercial; indeed, in many cases the producers will have been the same people through the ‘putting out’ system we discussed last week), meaning that in many cases a commercial spinning or weaving operation may be understood by ‘scaling up’ the same techniques we’ll discuss here simply to involve more laborers working in parallel (and likewise, production in very large households). That said, spinning and weaving are one area where there is significant technological change through the pre-modern, as we’ll see, which has important ramifications for the process and the availability of cloth. In the interest of keeping this reasonably easy to follow, we are going to use the methods as they existed during the Roman period (distaff-and-spindle spinning, weaving with a warp-weighted loom) as a baseline, discussing other methods and changing technologies as we reach them in the process.
I do want to stress here at the outset that this is not a how-to guide, not the least because I am not qualified to write one, but also because that is not our purpose. Consequently, some of the steps here will be somewhat simplified in presentation and a lot of the complicated processes available to very skilled spinners and weavers (such as weaving complex patterns) are going to be left out for the sake of brevity (or my none-too-convincing attempt at it).
And, as always, if you enjoy the yarn I am spinning for you here, please share it; if you really like it, please consider supporting me on Patreon. If you want updates when each new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for not merely updated but also my occasional ancient history, foreign policy or military history musings.
Spinning a Yarn
The next step for both of our fibers, now that they have been cleaned and sorted, is spinning, when the fibers are turned into thread or yarn (the distinction between the two is purpose; any kind of spun fiber is yarn – though the term is often used particularly of wool – while thread is technically yarn intended for sewing, although even in technical writing, often all yarn is called ‘thread’). In antiquity and for much of the Middle Ages (and into much more recent history in many parts of the world) this was done by hand with a distaff and spindle.
This is an incredibly important process. We’ll come back to time and labor at the end of this post, but spinning was by far the most time intensive part of making textiles (in the estimates I’ve seen, spinning tends to take up around 85% of the labor time of textile production, see below for figures). And textile production was a major activity (indeed, the major activity) for probably around 40% of the population in most pre-modern societies – not merely almost all adult women, but girls too would start spinning and weaving (spinning in particular, again, we’ll get there) at very early ages. Consequently, spinning thread may have been the single most frequently performed work-task in the ancient world (the various farming tasks being more varied and more seasonal, while spinning was being done continuously all year round). We tend to think of the pre-modern world as a world of farmers (and it was) but we ought just as well to think of it as a world of spinnners.
(As an aside: weaving, rather than spinning, tends to dominate the modern imagination when it comes to women’s work in the pre-modern world, but this is a product (as we’ll see) of the rather late-arriving spinning wheel (and in particular the even later arriving mechanical spinning wheel), which altered the time-and-labor economics of textile production. From the Neolithic down to c. 1500 AD, by far most of the labor-time spent producing textiles – probably even if shepherding and flax growing are included – went into spinning fibers into yarn.)
This spinning was done with a distaff and spindle (there was an even older pre-distaff spinning process used for flax by the Egyptians very early on; on this note Barber (1991), 44-51). The distaff was a rod, typically made in wood, sometimes with a hooked end (but as frequently without) which holds the fibers to be spun (those fibers are typically in the form of a long, almost rope-like mass called a roving). The roving is wrapped around the top of the distaff and then typically tied in place with cord or ribbon so that the fibers may be drawn out slowly (and so it doesn’t fall off). While some distaffs were hooked or otherwise shaped to hold the fibers, both wool and linen may be spun using a distaff that is just a straight rod (indeed, a simple stick of the right length and thickness will do).
Spindle design is a bit more complex. A spindle has three main parts: the shaft, around which is set the whorl, and sometimes a hook which allows the spindle to be suspended and holds the yarn (some spindles do not have hooks; drop-spindles, which often require hooks, are clearly in use in Greece by the end of the Archaic and possibly much earlier). Finally, something that is on the spindle but not part of it is the cop, which is the tight mass of spun yarn wound around the shaft of the spindle (that is, completed yarn being effectively ‘stored’ on the spindle).
The way a spindle works is that it spins (imagine that!). The spinning is set in motion by the hand of the spinner. The purpose of the whorl, which is generally a large, often flat, round and fairly massive (compared to the shaft) object mounted on the shaft, is to preserve the momentum of the spin, allowing the spindle to be kept spinning with less effort (particularly less frequent turns of the hand). The spinner draws out (‘drafts’) the fibers (this is the hand motion you see spinners doing; it allows them to carefully control how much fiber is being drawn into the spinning action of the spindle), while the spinning action of the spindle twists the fibers, causing the microscopic barbs (‘scaling’) on the wool to hook together as the fibers twist around each other, giving the yarn its coherence and strength. The spin then finally pulls that spun yarn down on to the shaft of the spindle and into the cop, where it can be later removed and used as thread or to weave fabric.
This basic process is largely the same regardless of if it is wool or linen. The direction of the spin determines if the resulting yarn is ‘s’-twisted or ‘z’-twisted and a skilled spinner can also control the rate of spinning and thus the twists-per-inch (lower twists makes for softer but weaker yarn, higher numbers of twists makes for stronger but harsher yarns). Skilled spinning requires careful control thus both of the rate at which fibers are drawn and the rate of rotation of the spindle.
Unsurprisingly, spindle and distaff designs vary a fair bit, place to place, purpose to purpose and culture to culture. For distaffs, the main variation is the length of the rod; distaffs from the ancient world only rarely survive archaeologically but fortunately spinning scenes are some of the most common artistic motifs depicting women (no surprise, it is the thing they do the most of) so we still see a lot of them. The most common distaff length seems to be around three feet, an ideal length for the staff to be couched under the arm when in use for spinning (as you can see in some of the pictures). That said, we also see artwork of much longer distaffs which were long enough to be rested on the ground while the spinner was either seated or in some cases standing, and much shorter distaffs designed to be held in the hand rather than couched under the arm.
Spindle design is more varied still. Archaeologically, spindle whorls – the circular weight, which could be a part of the spindle itself but was often a separate piece attached to the spindle-stick – are recovered in significant numbers because they tend to be made of durable materials like pottery or stone (while the spindle itself is generally wood). Spindles might mount the whorl at either the top or bottom of the shaft; high-whorl spindles seem to have been more common in the Near East, while low-whorl spindles were more common in the Mediterranean and Europe, and there are even some ‘middle whorl’ spindles from Asia Minor. The base of the spindle-stick generally thins almost to a point – this was to gain mechanical advantage when starting the spinning motion by flicking the spindle between the thumb and forefinger – the narrower the spindle at that point, the more rotations are induced by a single flick.
The spindle also needs to be supported in its rotation, which has an impact on design. ‘Drop-spindles’ – where the weight of the spindle is supported by the thread itself (they require some form of catch at the top of the spindle-shaft to avoid falling off; this can be a small metal hook or just a notch in the top of the spindle which can catch the yarn) seem to have been the most common in Europe, but by no means the only spindles in Europe. Alternately, a spindle may be rested on something – the thigh of the spinner or even the ground; this is particularly important for very fine, light or slippery fibers (some flax or wools, but also cotton) where the weight of a drop-spindle would draw the fibers too rapidly. The advantage of supported spindles like this is they allow both hands to manipulate the fibers, but there were also hand-held spindles which were turned in the hand (and thus may not need a whorl, though they they sometimes have them anyway); these seem to have been used exclusively for short-staple wool as both hands are required for drawing out long staple wool (Barber (1992), 43-4).
The oldest depictions of spindles I know date to the Egyptian Middle Kingdom (2048-1782 BC, Barber (1992)), 45), but archaeology tells us the technology is even older, with spindle whorls appearing in the archaeological record in pre-historic Neolithic sites (Barber (1992), 55). The distaff seems to come later; clear evidence for the distaff shows up in the visual record for late Archaic and Classical Greece (and in the Danube basin in the 7th century), but doesn’t arrive in the Egyptian record until the Roman period (though there may – interpretations are difficult – be evidence for the use of distaffs in 3rd-millenium Mesopotamia, far earlier).
The great advantage the distaff provides is portability (note this video for a demonstration). With the wool or bast roving wound up on the top of the distaff, a woman spinning with one can stop the spindle, gather up the drawn fibers, tuck the spindle into the roving and get moving fairly quickly. This was a tremendous advantage for a woman looking to spin thread in whatever small gaps of time she had in the midst of other household work (e.g. food preparation, child care, or overseeing other younger spinners in the household). Consequently, a woman could take her spinning with her almost anywhere she needed to go, setting up, getting working and then packing up and moving again as needed.
The Spinning Wheel
And for more than 4,000 years, from the beginning of the Bronze Age (at least!) to perhaps 1000 AD, that was simply how yarn was produced. The major change comes with the development of the spinning wheel. The exact dates for the development of the spinning wheel are unclear; it may have been invented in India or the Islamic world. In any event it is known in the Near East by 1030 AD and we have clear depictions of it (see below) by 1237. The innovation travels fast, with the first references in China by 1090 and the first depictions in 1270 and the first clear illustrations in Europe in 1280, though the spindle-and-distaff were not fully displaced (they retained advantages in portability and cost, as well as being able to handle a wider range of fibers and yarn-types) until the industrial revolution in some parts of the world.
The basic design of the spinning wheel was heavily iterated on, but at its core the wheel has a wheel (rotated by hand, or foot or later peddle) which in turn rotates a spindle (there is still a spindle) which is held horizontally rather than vertically. By holding the fiber at an angle to the spindle, its turning induces the necessary twist in the fibers to produce yarn while, as before, the spinning of the spindle draws the freshly spun yarn on to the spindle’s cop. The spinner controls both the rotation of the wheel (thus the speed of the twist) and with the other hand the rate at which fibers are fed into the twisting. A major development was the addition of the treadle (developed in Germany c. 1533), a foot-pedal which spun the wheel and allowed for very steady, constant rotation while freeing up the spinners hands to focus on drawing the fibers. You can see a good demonstration of this here.
The spinning wheel, at least in these early forms, had serious limitations. Yarn spun on early spinning wheels tended to be looser and weaker and so was unsuitable for warp threads when weaving (see below) which meant that the spinning wheel persisted alongside hand spinning for some time. The treadle-driven wheel (known as the ‘Saxony wheel’ in England) enabled the spinning of much stronger yarns (suitable for the warp) and it was this development that set in motion the replacement of hand spinning with spinning wheels though as noted hand spinning retained enough advantages to still continue, especially in poorer areas, until the industrial period.
The spinning wheel dramatically altered the labor economics of yarn production. Even the hand-turned spinning wheel was probably around three times faster than hand spinning (J.S. Lee, op. cit., 47); the treadle wheel was at least twice as productive as that and some estimates for mature spinning wheel technology over hand spinning suggest as high as a ten times increase in spinning speed. Needless to say, a reduction in labor time potentially close to an order of magnitude in the most labor-intensive part (again, c. 80% of the labor time!) of textile production had enormous economic impacts once the spinning wheel was widely dispersed. J.S. Lee (op cit., 9) estimates, for instance that English cloth production tripled (measured by weight) between 1315 and 1545 and cloth produced per capita increased five-fold (the English population declined during the period due to the Black Death). Income and the status of spinners consequently declined (wages may have dropped by as much as half) and the sudden relative abundance of cheap yarn put pressure on the other stages of textile production – both to feed wool into the spinning wheels and then more weavers to make cloth out of it.
We’ll talk more about the status of the workers doing all of this spinning at the end of the post, but for now this brings us to the next stage in textile production: weaving.
Some of the yarn we have made is going to be used in sewing (though as we will see, really only a tiny fraction of it), but most of our yarn is going to need to be woven into fabric. Of course yarn, particularly woolen yarn, can also be knitted into a knitted fabric, but the great majority of textiles are woven and not knitted so we are going to focus on weaving to produce fabric.
First, we need to clarify some key terms because otherwise describing various weaving methods won’t make much sense. Woven fabric is made up of two components: the warp threads (which held still in tension on the loom) and the weft thread (which runs through the warp and loops back as it reaches the warp’s end on each side); note the numbers here: there are many warp threads, but typically only one weft thread. Fundamentally, the act of weaving is about passing the weft thread through the warp, alternating the warp threads it passes in front of and behind of (following a pattern; we’ll discuss some in a moment), looping back around after each completed pass to create a tightly woven fabric. The frame used to hold the warp during this process is called a loom.
To facilitate this process, a loom has to do a few things. First, it has to have a system for holding the warp threads taught under tension; this must be done along a bar at both ends rather than being tied off at a single point in order to weave fabric wider than just a few inches. Second, it needs a way to separate the upper and lower threads of the warp (so that the weft can be passed through between them, passing over some and under others), creating what is called the ‘shed’ (not, by the by, in the sense of a house – that meaning comes later – but actually from German scheiden, ‘to divide,’ as the shed divides the warp threads), and it needs to be possible to change which warp threads are on top and on the bottom of the shed. Without the ability to create the shed, it is necessary to ‘darn’ in the weft, manually passing it over some warp threads and under others, which would be slow and tedious to say the least. Next, to assist the passing of the weft thread it is generally connected to some convenient holder called a shuttle. Finally there needs to be some way, once each loop of the weft has been run through the warp, to compress the weft in order to form a tight, strong fabric. Different looms are essentially defined by their different solutions to these problems.
Probably the earliest looms were in the form of the backstrap loom (where one warp beam is tied off to a fixed object while the other end of the warp is secured around the waist of the weaver; later more sophisticated variants use two bars, but secure the second bar around the back), but the evidence is tricky; we see such looms in the pre-contact Americas (for instance in Inca sites in the Andes) and in China during the Han dynasty, but this evidence is chronologically after our evidence for ground looms appearing in Egypt. Since we’re focused on textile production in the broader Mediterranean, I am going to otherwise leave the backstrap loom out (but see Barber (1992), 80-82 on it if you like).
We have evidence for the ground-loom from the late Neolithic (early 4th millennium BC) in Egypt (Barbar (1992), 83). In a ground loom, the warp lays flat parallel to the ground, attached at both the top and bottom to warp-bars which are in turn secured to the ground via stakes, thus providing the necessary tension in the warp. By using two warp bars like this, the fabric may be arbitrarily wide or long, though making it wider than a person’s reach poses difficulties (the Egyptians seem to have generally had two workers on the loom for wide fabrics).
The solution that appears here to quickly creating the shed and counter-shed are the heddles and the heddle bar, along with a shed bar. The shed bar is a rod that passes through the warp such that every second thread (or whatever pattern is desired) passes over the rod while the other threads pass under it; when that rod is pulled forward, the threads are separated to make that passageway (the shed). The countershed is then created through the heddle bar, which is attached to the return row of the warp by heddles (these days, little metal hooks, but in ancient weaving, heddles were short strings looped around behind the return-row warp threads and then tied to the heddle bar). Let the heddle bar down while pulling the shed bar up (or in a ground-loom just having a tall enough shed bar that it creates a ‘natural shed’ by simply being there) and you get the shed; pull the heddle bar up while leaving the shed bar down and you get the counter shed.
Thus the weaver can use the two bars to make the shed, then passes the weft through the shed on the shuttle (called shedding), then shift the bars to make the counter-shed, pass the weft back the other way on the shuttle (called picking). Finally, another rod (lots of rods here) called the beater – generally a very flat rod which can span the entire way over the warp and still stick out on both ends – is used to compress the row of weft thread by being pushed or pulled hard to compress the weft into the completed sections of the weave (a process called ‘beating’ or battening). By repeating that basic sequence, our carefully spun thread can be slowly transformed into fabric.
Naturally, different looms were common for different purposes and also at different times and places. Because weaving, like spinning, was a skill learned locally (passed down from mother to daughter, generally), there tends to be a lot of regional variation in loom design and even very old loom types often persist in certain areas even while being replaced in others (the same is true, by the way, of spinning implements).
Dating to the late Neolithic, the Warp-weighted Loom was the standard loom of the ancient Mediterranean and while it was supplanted in most of Europe during the Middle Ages, its use persisted in Scandinavia through the Middle Ages and into the modern period. This was the principal loom for the Greeks and the Romans, as well as for much of Europe during the early Middle Ages (especially in Northern Europe). These looms, being made of wood, are only very rarely preserved archaeologically, but their loom-weights, being made typically of fired clay or stones, are preserved in great number, often allowing archaeologists to identify where in a structure or settlement weaving was taking place by the presence of the loom-weights.
In the warp-weighted loom, the warp is attached at the top to a warp bar which is supported by a frame (which can either be free-standing or designed to lean against a wall); the attachment of the warp is managed through the use of a heading band (a bit of already woven cloth as a ‘starter,’ the weft of which became the warp of the finished weave; I think I’ll do a short addendum on tablet weaving to discuss one way these heading bands were made.). At the bottom, groups of warp threads are tied to loom-weights (often stone or ceramic weights with loops for tying off the warp threads), which provide the necessary tension, but also give the weaver more opportunity to manipulate the warp threads if necessary. Then towards the bottom of the frame, the shed-bar can be mounted on the frame itself (as with the ground-loom, the shed bar passes under every other warp thread and over the rest), while the heddle bar (which otherwise functions exactly like that in the ground-loom) is mounted around the center of the frame, connected to half of the warp threads with string. Weaving started from the top and moved down the frame.
The great advantage of the warp-weighted loom as compared to the ground loom is that it allows the weaver to weave sitting or standing, which is a lot more comfortable than having to weave from a squatting position as shown in Egyptian artwork on the ground-loom. Such a loom could also be more easily set up or moved to a new location if needed. The limitation was that the size of the frame fundamentally limited the size of the textile which could be woven and most frames tended to stick to a size small enough to be used by a single person, though two-story frames or extra-wide frames worked by multiple weavers for very large fabrics are known in art (Gleba, op. cit., 123-4).
Horizontal framed looms, developed in Asia, began to spread into the Mediterranean in the 10th century and seem to have become reasonably common in Europe by the thirteenth. In this kind of loom, the warp is held horizontally, often using rollers on each side to allow the the weaving of a fabric of any length (there a vertical loom like a warp-weighted loom was restricted in fabric length to the height of the loom). The warp beams are then attached to a rectangular frame (see image) and the heddle bar (or bars) are suspended from the top of that frame. Often this allows the heddles to be controlled not by hand (as with the warp-weighted loom) but by a foot-controlled treadle, allowing shedding to be done quite rapidly (though passing the shuttle and battening are still done by hand; that won’t be resolved until the invention of the flying shuttle loom in 1733, outside of the scope of this essay). It is not hard to guess that the adoption of better looms (J.S. Lee, op. cit., 51 supposes that the horitzontal frame loom could be worked perhaps three times faster than the old vertical warp-weighted looms) and the adoption of the spinning wheel motivated each other, as faster production on one end created the demand for faster production on the other.
In any event, the end result of our process (we have skipped a few finishing steps, but this is not a how-to-guide) is a lattice-work of fine fibers of either wool or linen (you can mix the two, creating linsey-woolsey, but my impression is that ancient weavers generally didn’t do so), which is to say, fabric, ready to be dyed or bleached (if the thread wasn’t dyed already) and cut or sewn into garments, blankets or any number of other uses.
Now I’ve alluded to the labor demands of all of these processes before, but it is time to put some sense of numbers and scale behind that. Time-and-labor studies with actual historical spinning and weaving equipment are few, but there are a handful, so we can to some degree model the household economics of a family looking to produce its own textiles.
First we need to think about minimum textile requirements. At the absolute bottom end, Cato the Elder (De Ag. 59) who was well known to be parsimonious and cruel towards his enslaved workers, recommends each worker get a long tunic and a cloak (the sagum – sometimes oddly translated as ‘blanket’ I do not know why) every other year; obviously a fairly miserable minimum but it adds up to about 21,650cm2 of fabric per year. Probably even most enslaved workers were somewhat better clothed than this, given Cato’s reputation, and free farmers must have aimed a fair bit higher for sure. Roman soldiers – initially drawn from the modest-but-not-poor freeholding Roman farmer class (the assidui) seem to have had perhaps two complete sets of garments per year (following J. Liu, “Clothing Supply for the Military. A Look at the Inscriptional Evidence” in Wearing the Cloak, ed. Marie-Louise Nosch (2012), 19-28)), which might be more reasonable; presumably poor farmers would have somewhat less than this, but someone more than Cato’s poor enslaved workers.
So let’s assume a standard, somewhat extended household, of perhaps six individuals; a married couple, one of their elderly mothers, an adult son and two children (a decently plausible small farming household). A complete set of Roman clothing (I’m using the Romans because I’m more familiar with their dress), excluding formal wear (read: the toga, though I am also not counting the woman’s palla either) for this family of six might require something like 220,000cm2 (26.3 square yards) of fabric at a minimum pear year – a single complete change of clothing. Comfort might look two or three times this much. How much labor is that?
We can look at a few different estimates (skip one paragraph ahead if you hate lots of numbers). Aldrete et al., (Reconstructing Ancient Linen Body Armor (2013) do a complete labor study of the time it took to make a linothorax, a Greek style of linen armor, including fiber preparation, spinning, weaving and sewing. For the roughly 65,000cm2 amount of (admittedly quite rough) linen required (which in turn required 12,600m of thread), they figure it took 25 hours to break, scutch and hackle the flax, 575 hours to spin it into thread, 75 hours to weave the thread into fabric (including loom setup time), and 8 hours to measure and cut the fabric (alas, the linothorax is laminated, not sewn, so they have no data for the sewing portion). Eve Fischer has done a similar calculation (but with back-of-the-envelope estimates rather than a detailed study) estimating that 41,804cm2 (5×1 yards) of fabric would require c. 8,230m of thread which would in turn demand something like 7 hours of sewing, 72 for weaving, 500 or so for spinning. J.S. Lee (op. cit., 51) figures a 14th century weaver (with those fancier looms and spinning wheels) could weave around 2 yards of fabric per day from roughly 6lbs of spun yarn while a given spinner might spin about 1lbs of yarn per day; assuming a 12-hour work-day that comes out to about 6 hours per yard weaving (a little more than twice as fast as Fischer of Aldrete’s vertical loom weavers) and 36 hours per yard spinning (three times faster than the hand-spinners).
So we have our range if we harmonize the estimates: something like a yard-square (8.361.27cm2) might require labor time roughly on this order:
|Fiber Preparation||Spinning||Weaving||Sewing||Total||% spent spinning|
|Aldrete et al||3.25 hours||74.7 hours||9.75 hours||2+ hours?||89.7 hours||83.2%|
|Fischer||–||100 hours||14.4 hours||1.4 hours||115.8 hours||86.35%|
|J.S. Lee||–||36 hours||6 hours||–||42 hours||85.7%|
Now we can very roughly calculate the labor time our family needs (keeping in mind this is rough, with the largest problem being that we have an estimate for flax preparation but not for wool preparation; if you are a traditional wool worker, I would love detailed, rigorous time-labor data for all stages of wool preparation as doing a rigorous assessment of the textile production needs of a Roman smallholder is a long-term research goal of mine). Using the average of Aldrete and Fischer’s figures (erring a little high to account for Fischer’s lack of preparation time) we might figure something like 2,683 hours to produce our 220,000cm2 minimum requirements. Our upper ‘comfort’ level might be three times this or 8,049 hours.
Put into working terms, the basic clothing of our six person farming family requires 7.35 labor hours per day, every day of the year. Our ‘comfort’ level requires 22.05 hours (obviously not done by one person). These figures come way down once we get the spinning wheel and horizontal loom, but what seems fairly readily apparently is that women did not necessarily work less so much as produce more, selling the excess via the ‘putting out’ system we mentioned last time and using that to support their families.
That data has some implications.
A lone woman could, if she spun in almost every spare minute of her day, on her own keep a small family clothed in minimum comfort (and we know they did that). Adding a second spinner – even if they were less efficient (like a young girl just learning the craft or an older woman who has lost some dexterity in her hands) could push the household further into the ‘comfort’ margin, and we have to imagine that most of that added textile production would be consumed by the family (because people like having nice clothes!).
At the same time, that rate of production is high enough that a household which found itself bereft of (male) farmers (for instance due to a draft or military mortality) might well be able to patch the temporary hole in the family finances by dropping its textile consumption down to that minimum and selling or trading away the excess, for which there seems to have always been demand (from families with the wealth to demand more textiles than the women of the household would produce; we know that it was rare in practice for aristocratic Roman women to do much spinning for instance – exceptions were notable, see below). Consequently, the line between women spinning for their own household and women spinning for the market often must have been merely a function of the financial situation of the family and the balance of clothing requirements to spinners in the household unit (much the same way agricultural surplus functioned).
Moreover, spinning absolutely dominates production time (again, around 85% of all of the labor-time, a ratio that the spinning wheel and the horizontal loom together don’t really change). This is actually quite handy, in a way, as we’ll see, because spinning (at least with a distaff) could be a mobile activity; a spinner could carry their spindle and distaff with them and set up almost anywhere, making use of small scraps of time here or there.
On the flip side, the labor demands here are high enough prior to the advent of better spinning and weaving technology in the Late Middle Ages (read: the spinning wheel, which is the truly revolutionary labor-saving device here) that most women would be spinning functionally all of the time, a constant background activity begun and carried out whenever they weren’t required to be actively moving around in order to fulfill a very real subsistence need for clothing in climates that humans are not particularly well adapted to naturally. The work of the spinner was every bit as important for maintaining the household as the work of the farmer and frankly students of history ought to see the two jobs as necessary and equal mirrors of each other.
At the same time, just as all farmers were not free, so all spinners were not free. It is abundantly clear that among the many tasks assigned to enslaved women within ancient households. Xenophon lists training the enslaved women of the household in wool-working as one of the duties of a good wife (Xen. Oik. 7.41). Columella (De Re Rust. 12.3.6.-7) advises that the enslaved wife (the vilica) of the enslaved manager (the vilicus) ought to be working wool (which is to say, spinning) at any time when she could not be doing agricultural work and moreover assumes that she would also be supervising other enslaved women on the estate doing the same. Columella also emphasizes that the vilica ought to be continually rotating between the spinners, weavers, cooks, cowsheds, pens and sickrooms, making use of the mobility that the distaff offered while her enslaved husband was out in the fields supervising the agricultural labor (of course, as with the bit of Xenophon above, the same sort of behavior would have been expected of the free wife as mistress of her own household).
Large and wealthy households might have many enslaved women producing not only the textiles required of the workers, but also for the family itself. Columella (loc. cit.) is quite clear that the aim of his vilica and her assistants in spinning was to make the villa self-sufficient by providing domestically produced clothing for the male workforce. Meanwhile, Homer in describing Alcinous’ household (presumably meant to represent an opulent late-Dark-Age/early-Archaic Greek household) notes that Alcinous had fifty serving women who both ground grain in hand-mills and produced textiles (Od. 7.103-5). Notably, earlier in the poem we hear that Helen (of Troy, though now back with her husband Menelaus) works wool and spins herself (Od. 4.130f), as of course does Odysseus’ own wife Penelope.
Consequently spinning and weaving were tasks that might be shared between both relatively elite women and far poorer and even enslaved women, though we should be sure not to take this too far. Doubtless it was a rather more pleasant experience to be the wealthy woman supervising enslaved or hired hands working wool in a large household than it was to be one of those enslaved women, or the wife of a very poor farmer desperately spinning to keep the farm afloat and the family fed. The poor woman spinner – who spins because she lacks a male wage-earner to support her – is a fixture of late medieval and early modern European society and (as J.S. Lee’s wage data makes clear; spinners were not paid well) must have also had quite a rough time of things.
Spinners, Spinsters and Distaff Lines
It is difficult to overstate the importance of household textile production in the shaping of pre-modern gender roles. It infiltrates our language even today; a matrilineal line in a family is sometimes called a ‘distaff line,’ the female half of a male-female gendered pair is sometimes the ‘distaff counterpart’ for the same reason. Women who do not marry are sometimes still called ‘spinsters’ on the assumption that an unmarried woman would have to support herself by spinning and selling yarn (I’m not endorsing these usages, merely noting they exist).
E.W. Barber (Women’s Work, 29-41) suggests that this division of labor, which holds across a wide variety of societies (though commercial textile production was often done by men in pre-modern societies, something we’ll discuss next week) was a product of the demands of the one necessarily gendered task in pre-modern societies: child-rearing. Barber notes that tasks compatible with the demands of keeping track of small children are those which do not require total attention (at least when full proficiency is reached; spinning is not exactly an easy task, but a skilled spinner can very easily spin while watching someone else and talking to a third person), can easily be interrupted, is not dangerous, can be easily moved, but do not require travel far from home; as Barber is quick to note, producing textiles (and spinning in particular) fill all of these requirements perfectly and that “the only other occupation that fits the criteria even half so well is that of preparing the daily food” which of course was also a female-gendered activity in most ancient societies. Barber thus essentially argues that it was the close coincidence of the demands of textile-production and child-rearing which led to the dominant paradigm where this work was ‘women’s work’ as per her title.
(There is some irony that while the men of patriarchal societies of antiquity – which is to say effectively all of the societies of antiquity – tended to see the gendered division of labor as a consequence of male superiority, it is in fact male incapability, particularly the male inability to nurse an infant, which structured the gendered division of labor in pre-modern societies, until the steady march of technology rendered the division itself obsolete. Also, and Barber points this out, citing Judith Brown, we should see this is a question about ability rather than reliance, just as some men did spin, weave and sew (again, often in a commercial capacity), so too did some women farm, gather or hunt. It is only the very rare and quite stupid person who will starve or freeze merely to adhere to gender roles and even then gender roles were often much more plastic in practice than stereotypes make them seem.)
Spinning became a central motif in many societies for ideal womanhood. Of course one foot of the fundament of Greek literature stands on the Odyssey, where Penelope’s defining act of arete is the clever weaving and unweaving of a burial shroud to deceive the suitors, but examples do not stop there. Lucretia, one of the key figures in the Roman legends concerning the foundation of the Republic, is marked out as outstanding among women because, when a group of aristocrats sneak home to try to settle a bet over who has the best wife, she is patiently spinning late into the night (with the enslaved women of her house working around her; often they get translated as ‘maids’ in a bit of bowdlerization. Any time you see ‘maids’ in the translation of a Greek or Roman text referring to household workers, it is usually quite safe to assume they are enslaved women) while the other women are out drinking (Liv. 1.57). This display of virtue causes the prince Sextus Tarquinius to form designs on Lucretia (which, being virtuous, she refuses), setting in motion the chain of crime and vengeance which will overthrow Rome’s monarchy. The purpose of Lucretia’s wool-working in the story is to establish her supreme virtue as the perfect aristocratic wife. This didn’t change in the Middle Ages either; for instance by the 12th century, popular memory of Bertha of Swabia (907-966; queen of Italy 922-926, 937-948; this memory may have been in part invented by monastic communities which claimed her as a sponsor) represented her as an idealized figure, depicted her with distaff in hand and the phrase “when Queen Bertha was spinning” became a trope to evoke an idealized past. Betha’s is hardly the only legend that has to do with women’s role in textile production.
Moving from legends, textile production was an important point of status for many pre-modern women. It was a thing which might well be bragged about (e.g. Plut. Mor. 241d, where an Ionian woman does just that to a less-than-receptive Spartan woman). Augustus (the first emperor of Rome) and his wife Livia recognized the value of the imagery, letting it be known that Livia (herself fantastically wealthy even without being the wife of the sole ruler of Rome) spun his clothing herself (Seut. Aug. 73); we may safely judge this to be a bit of theater, but the usefulness of the theater (much like the manufactured memory of Bertha) tells us something.
Indeed, even in death a great many pre-modern women did not leave behind the distaffs and spindles. Perhaps the single most common compliment given to Roman women on funerary inscriptions was the simple but ubiquitous ‘lanam fecit’ (“she made wool;” for an example, note). Spinning and weaving tools are extremely common grave goods (almost always, but not always for female burials) in numerous pre-modern cultures. Sanna Lipkin (“Textile Making in Central Tyrrhenian Italy – Questions Related to Age, Rank and Status” in Making Textiles in Pre-Roman and Roman Times, ed. M. Gleba and J. Pásztókai-Szeöke (2013), 19-29) notes numerous examples in pre-Roman Italy (interesting age seems to have been a factor, with younger individuals more likely to have weaver’s sets, while older individuals were more likely to be buried with a distaff; Lipkin suggests quite plausibly that burial with a distaff and spindle was the clear marker of the mater familias – the female head of house).
For myself, I find that students can fairly readily understand the centrality of farming in everyday life in the pre-modern world, but are slower to grasp spinning and weaving (often tacitly assuming that women were effectively idle, or generically ‘homemaking’ in ways that precluded production). And students cannot be faulted for this – they generally aren’t confronted with this reality in classes or in popular culture. For instance, in HBO’s Rome, we see men work, build and farm (as well as fight), but we do not see women spinning or weaving at any point, even in very modest households where this would have been a continuous activity (the opportunity to have Livia, herself a character in the show, be shown spinning Octavian’s clothes was also sadly missed). Even more than farming or blacksmithing, this is an economic and household activity that is rendered invisible in the popular imagination of the past, even as (as you can see from the artwork in this post) it was a dominant visual motif for representing the work of women for centuries.
Needless to say, there is much more to discuss when it comes to these processes (I have, for instance, left out some of the finer points of weaving, in particular the mechanisms of producing more complex weaving patterns and the loom set-up and finishing techniques; seriously, weaving is complicated but I leave it to modern practitioners to explain those details), but with any luck I have at least convinced you that pre-modern household textile production is not a thing to be neglected in the study of the past.
Next week, we move more fully into the commercial space and look at the fabric trade and commercial production, along with the various color treatments that might be applied to either fabric or yarn.