It’s been fascinating seeing the responses to Judy Stephenson’s post ‘Spinning little stories: Why cotton in the Industrial Revolution was not what you think’ on the Economic History Society’s Long Run blog (https://ehsthelongrun.net/2016/12/01/spinning-little-stories-why-cotton-in-the-industrial-revolution-was-not-what-you-think/). It’s partly about my East Asian Journal of British History article ‘Fashion, Textiles and the Origins of Industrial Revolution’ (http://www.history.ac.uk/sites/history.ac.uk/files/eajbhvol5.pdf). I’ve followed the ensuing discussion on Twitter and on various blogs (in particular those by Pseudoerasmus [https://pseudoerasmus.com/2016/12/01/allen/] and Vincent Geloso [https://notesonliberty.com/2016/12/07/the-high-wage-economy-hypothesis/]. Like Anton Howes, I’ve been impressed by the quality of what Geloso calls a Twitterminar. So I’m keen to respond.
There are three issues I’d like to take up.
1. First issue: Robert Allen, induced innovation and the High Wage Economy (HWE).
As Pseudoerasmus has noticed, I’m not arguing that Allen’s induced innovation model is inherently misconceived as a model. It’s just that, as far as textiles are concerned, Allen is empirically defective (to put it politely) with regard to wages and capital costs, rendering his analysis / findings redundant.
On capital costs, Allen hugely overestimates the cost of a 24 spindle jenny at 70 shillings. 24 spindles is bigger than the original domestic jennies, which were 8 to 16 spindles. These were crude machines. Any local carpenter familiar with constructing single spindle, non-flyer spinning wheels of the kind used for spinning cotton in Lancashire could have thrown one together pretty cheaply – essentially the same spindle (@ <0.25 shillings) and drive cord multiplied in a frame + bobbins, clamp bar, faller wire and a wheel. A single cotton spinning wheel cost 2 shillings. Simply multiplying the cost of a single wheel by 24 (which is absurd, but serves to make the point) comes to 48 shillings, way less than Allen’s 70 shillings. Aspin, in a book Allen references, pointed out that as late as 1783 (when small, domestic cotton jennies had already probably peaked), the overseers of the poor at Saddleworth, on the Yorkshire border with Lancashire, bought a jenny for 16 shillings. Elsewhere in West Yorkshire, where probate inventories survive for this period, 20 shillings is a typical valuation for the bigger, wool-spinning jennies.
On wages, Allen doesn’t just misrepresent Arthur Young’s evidence on wages, which I show in my EAJBH article. He clearly doesn’t understand the differences in product mix and organization of work between the principal English and French cotton spinning regions – Lancashire and Normandy (I’m not convinced Vincent Geloso will find answers in Alsace – Mulhouse in particular. Pre 1789, Mulhouse was in the Swiss confederation and most of the calico and cotton-linen fabric used for printing there was spun / woven either in India, or in Alpine Switzerland)
The two key differences between Lancashire and Normandy were:-
(a.) Product mix.
The vast majority of cotton-linen fabrics produced in Normandy (checks and striped Siamoises in particular) were loom-patterned. Very few plain cotton-linens or traditional heavy fustians were produced, at least before the 1770s. To provide the coloured pattern, a proportion of the yarn therefore had to be dyed after spinning, but before weaving. This required a three-stage production sequence in the order (i) spinning, (ii) colouring, (iii) weaving, each undertaken by different people in different places in the region. Lancashire, too, hosted large-scale production of loom-patterned cotton-linens (stripes and checks), but there was also a vast manufacture of plain cotton-linens, decorated by dyeing or printing after they had been woven. They included both the traditional Lancashire heavy fustians (dyed in the region) and the lighter fabrics known as Blackburn greys (sent to London for printing). Their three-stage production sequence was ordered differently from loom-patterned fabrics – (i) spinning, (ii) weaving, and only then (iii) colouring.
(b.) Organisation of work.
In Normandy, the women who spun operated as independent small businesswomen, buying cotton wool in small quantities, spinning it, and then selling the yarn, irrespective of their husband’s occupation. In other words, a classic kaufsystem, which made a lot of sense in a context where the yarn would be then passed directly on to dyers, working in relatively centralized premises/locations, before being re-distributed to weavers. In the parts of Lancashire producing plain cotton-linens – traditional fustians and the newer Blackburn greys – a verlag system prevailed, but one that unusually involved putting out materials at two different stages in their production sequence – a prepared linen warp (sourced in Ireland, Scotland, or the Baltic) and raw, uncarded and unspun cotton wool for the weft. This mix of processed materials and unprocessed materials was put out as a package to male weavers of plain cotton-linen fabric, such as James Hargreaves at Oswaldtwistle, at a piece rate for the finished fabric (ie. not the spinning alone). Weavers were expected to arrange the spinning themselves, which in many cases must have been undertaken mainly by their family. No wonder Lancashire evidence of cotton spinning wage rates can be hard to find – for a lot of Lancashire spinners there wasn’t a rate for spinning as such.
A number of consequences for Allen’s HWE flow from these two Lancashire – Normandy differences, at least as far as spinning is concerned:
They suggest that Hargreaves invented the jenny principally to allow cotton spinning to be concentrated in the weaver’s household. This would be entirely consistent with his previous improvement to domestic stock carding, designed to increase output while reducing the need for physical strength. It was all about keeping as much as possible of the spinning work, which included carding, within the family. Why? Partly because, as I argue in EAJBH, there was increasing competition for spinning labour from a rapidly expanding worsted industry in the vicinity (no such competition in Normandy). But also because the later 1750s and early 1760s saw:
(1) major supply problems with the white, all-cotton calicoes imported by the East India Co. from India to be printed near London for export to Europe and colonial North America. One consequence was extra pressure to substitute cotton-linen Blackburn greys, despite their inferior quality.
(2) big price increases for both linen yarn and cotton wool, amplifying the cost of bad / wasteful spinning for weavers and incentivizing the use of trustworthy, easily supervised and trained family labour (a kind of pre-mechanisation Margolin effect).
(3) the introduction of printing on cotton-linens in Lancashire from the 1750s, which probably involved closer, more active monitoring of yarn/fabric quality.
Many of these considerations did not apply in Normandy, given its different product mix and work organization. In Normandy there was certainly less emphasis on measuring yarn quality. Yarn was sold there by weight, not count, perhaps because yarn uniformity was less of an issue for loom-patterned fabrics than for cotton-linens used for printing. Insofar as the considerations outlined above did apply in Normandy, the consequences were different, as there was no particular incentive to adopt an integrated family model of production, sustained by small 8 to 16 spindle domestic jennies. I suspect the French never had many domestic jennies, but simply did a technological frog leap in the 1780s, going straight to bigger, workshop-based jennies (30-40 spindles or more), like the ones with 40 spindles at Oissel in Normandy I mention in EAJBH, which appear to have been housed in a workshop. In that case, when comparing numbers of jennies in England and France, we should be counting spindles, not machines (i.e. thinking like a textile factory production manager).
2.Second issue: hand spinning earnings and work intensity.
I like the way the Schneider-Humphries working paper uses primary sources to demonstrate that most spinners worked, on average, intermittently. Actual earnings were substantially lower than those reported by commentators like Arthur Young. That’s important if what we are interested in is women’s actual earnings compared to men’s – after all, huge numbers of English women were earning by spinning. It enables Schneider-Humphries to offer an effective critique of the way women’s spinning earnings have been presented in recent literature. But I think they over-claim on what their findings tell us about Allen on spinning jennies and HWE.
They lack Lancashire cotton data. Their only manuscript accounts for spinning cotton come from a workhouse at Marlborough in Wiltshire, 200 miles from Lancashire. There it was probably cotton candlewick that was being spun. Candlewick usually had an English cotton yarn count of somewhere about 5 or 6 (or, for technical textile geeks, Ne 5 to Ne 6). This was extremely coarse, similar to the roving fed into the rollers of an Arkwright water frame, but with more twist. Spinning it was considerably less demanding than spinning a Ne 18 weft yarn for a Blackburn grey, let alone a Ne 40 yarn for a coarse muslin.
If what Arthur Young was doing for England was to report what a healthy adult spinner might earn if working 12 or 13 hours a day, six days a week (which is what he requested from his informants), then that’s probably what he reported for France as well. So the international relationship doesn’t change, despite Young’s exaggeration of the hours spinners actually worked (once, that is, we identify the correct Arthur Young data, which, of course, Allen fails to do). It doesn’t change, that is, unless French spinners worked more (or less) intermittently than their English counterparts. Establishing that is not going to be easy, because in Normandy we’re dealing not with piece rates for spinning, but with the profits of independent contractors.
What Schneider-Humphries don’t give sufficient attention to is the variation and volatility of English spinning piece rates, and probably earnings, according to time, place, fibre, and skill. This reflects a general tendency in much of the literature to think about spinning as if were a single activity – unskilled women’s work. I think a more helpful approach is to think about it in the way we think about metal smelting – something that is common to all metals, but with major technical differences and challenges. The analogy can be taken further, because the mechanization of textile spinning was, like the application of coal / coke to metal smelting in England, a long-drawn out process. Each metal posed its own challenges. Applying coal / coke to smelting the whole range of metals took most of the 17th century to complete, from cementation steel in the 1630s to iron ore in the early 1700s, as the late John Harris pointed out.
English spinners were largely paid according to piece rates (ie. pence per lb. of yarn spun to any particular count and / or twist – warps more highly twisted than wefts, therefore slower to spin). Evidence from worsteds and woolens in Yorkshire and East Anglia suggests piece rates in any particular area might vary by c.25% over a three year period. The same was probably true of the availability of work. And rates varied by region and by fibre (spinning linen generally paid less than wool or cotton). There was international and inter-regional trade in yarns, but markets for yarn were at the same time strikingly regionalized. I suspect a key variable here was the difference between generic yarns and specialized yarns, reflecting the precise yarn specifications for different fabrics in what were often fast-moving export markets shaped by constant changes in fashion.
At some periods, as the Schneider-Humphries paper insists, putting-out masters enjoyed considerable monopoly power, especially in East Anglia towards the end of the c.18th, when the Norwich worsted industry stagnated and used increasing quantities of cheap imported Irish yarn. Yet this was not the case everywhere or at all times, even in the later c.18th. There were periodic complaints from many areas about spinning labour being in short supply, during both the 17th and the 18th centuries. These were highly volatile industries, often export-orientated, and the availability of spinning work fluctuated considerably with the state of the market. Commercial correspondence confirms that there were periods when manufacturers could not supply customers due to labour shortages in both spinning and weaving.
So I think that for spinners we can talk about relatively high wages for spinning in parts of the north of England in the 1750s and 1760s, but only relative to spinners elsewhere in England, or at other times. Personally I can’t see much point in describing that as a regional HWE, but others might.
3. Third issue: yarn counts and quality.
For plain cotton-linens and calicoes during the later 18th century and early 19th century , the average fineness of cotton yarns changed rapidly, as Harley has demonstrated. Quality improved, but in ways that make comparisons tricky. John Wyatt discovered in the early 1740s that yarn spun for weft to Ne 12 or less was considered coarse. Above Ne 24 was fine (and unusual). Most hand-spun weft yarn for woven fabrics appears to have been Ne 16 to Ne 20 (Harley’s 1769 data from the Blackburn firm, Cardwell, Birley and Hornby, confirms this). Arkwright water frames could spin up to Ne 60, although generally they didn’t. Ne 60 was good enough for coarse muslins c.1780, but no competition for Bengal hand spinners of muslin, spinning their local short-staple, ultra-fine raw cotton.
The short-staple Old World varieties of cotton would not spin on the Arkwright machine, which would only work effectively with the longer-staple varieties of cotton native to the Americas. Supply of these American longer-stapled cottons came initially from the West Indies and Brazil, and from the USA only after 1815 on a really significant scale (I’ll save my moans about Beckert and History of Capitalism for another occasion).
Michael Edwards’ Growth of the British Cotton Trade is good on the chronology of British muslin production using mule-spun yarn. It was only after about 1810 that top-quality Bengal fine muslins were displaced by English-woven substitutes.