1. Introduction
Historians have agreed that the gender division of labour in rural households was similar across much of northern and western Europe during the early modern period.Footnote 1 Thus, in England:
Men took the main responsibility for field work: ploughing, maintaining hedges and ditches, mowing hay, reaping corn and caring for cattle. Women’s main activities were centred on the house and farm-yard: milking dairy animals, making butter and cheese, brewing beer, raising poultry, caring for pigs, and growing herbs and vegetables … Spinning and lace-making provided work for women and children.Footnote 2
While in Scandinavia:
Men may have devoted about half their working time to agricultural work per se, such as ploughing, harrowing, haymaking, harvesting, threshing and fencing … Female work included indoor duties but also the tending of cattle, sheep and poultry, milking and the work connected with various products such as brewing, slaughtering, salting, smoking meat and the production of textiles from wool and linen.Footnote 3
Similar generalizations are found in farming advice books and descriptions from the early modern period. Thus, in England Gervase Markham wrote about ‘the perfect husbandman … whose office and employments are ever for the most part abroad, or removed from the house, as in the field or yard’, while the ‘English housewife … hath her most general employments within the house’, and provided advice to the housewife on cookery, preparing medicines, processing wool and flax, dairying, brewing and baking.Footnote 4 In Sweden, Olaus Magnus wrote that peasant women busied themselves with ‘spinning, weaving, baking, brewing, cooking, cleanliness and clothing of children, making beds, and caring for lambs, calves and other small animals’ while peasant men carried out ‘heavier tasks such as field work, threshing, taming of horses …, sharpening of weapons, fencing, and supervis[ing] and tak[ing] care of ploughs and fields.’Footnote 5
This article compares new evidence of the gender division of labour in two countries, England and Sweden. It presents a detailed picture of the gender division of labour that not only shows differences between the two countries but also demonstrates flexibility and adaptability in the allocation of work between women and men. As a consequence, we argue that neither broad generalizations nor single causes are adequate explanations for the patterns observed.
Similarities between societies have been simplified to the level of caricature by economists seeking to explain the gender division of labour in the long term. For instance, Alberto Alesina and co-authors suggest that in all pre-industrial societies based on plough agriculture, or ‘plough societies’, ‘men tended to work outside the home in the fields, while women specialized in activities within the home’.Footnote 6 Such generalizations encourage blanket cross-cultural explanations for the gender division of labour. Alesina et al. argue that men’s greater physical strength leads them to dominate plough agriculture.Footnote 7 Sheilagh Ogilvie offered a careful critique of such approaches in her groundbreaking study of the gender division of labour in early modern south-west Germany.Footnote 8 Yet she too favoured a single-cause explanation, citing guild regulations as the main reason for women’s exclusion from many forms of work in that region.Footnote 9
In this article we argue that ‘plough societies’ can encompass a great deal of variation, and that single factors are inadequate to explain the gender division of labour. Studies of small-scale agriculture in the modern world have queried the generalization that agriculture was predominantly a male activity in which women merely assisted and worked within the home. In an article that summarizes the findings of a host of studies in development economics, Carmen Deere notes the complexity of divisions of labour in farming households, pointing to the need to examine not only broad categories such as cultivation, livestock husbandry and food processing but also finer details at the level of particular tasks, such as ploughing, sowing and weeding.Footnote 10 Researchers looking at modern peasant societies emphasize the importance of observing how farm work interlocked with other forms of income-generation, such as craft production and the sale of produce.Footnote 11 Most importantly, Deere observes that the gender division of labour varied ‘cross-culturally and regionally … and also, within given regions in accordance with the prevailing social relations of production and income-generating opportunities’, as well as between households of different levels of wealth, arguing for the ‘importance of material conditions in changing social constructs’ such as the gender division of labour.Footnote 12 It is likely that similar complexities existed in historical societies. Indeed, Ogilvie observes that in early modern Europe, ‘gender specialization in arable agriculture … varied enormously across regions of central and northern Europe, and even among different parishes in the same region’.Footnote 13 From this perspective, it is interesting to notice how Olaus Magnus added, parenthetically, that women too could be involved in the taming of horses: not what we might expect from the generalizations made by historians. Even if it is unclear to what extent his statements reflect Swedish realities, his remark nonetheless suggests that a flexible division of tasks was not improbable.
Until recently there has been inadequate data to compare details in the historical division of labour between European countries and test the validity of cross-cultural generalizations. However, new research that has collected evidence of work tasks for early modern Sweden and England now makes it possible to undertake a direct comparison of two European countries. In this article, the gender division of labour in agriculture, craft production and commerce in Sweden and England is compared to demonstrate the complexity of historical gender divisions of labour. For England the data cover the period 1500 to 1700, while for Sweden it relates to 1550 to 1800. Although these chronologies are slightly different, they represent periods when a similar type of ‘early modern’ economy prevailed in each country, before the introduction of mechanized production. In both countries women and men each made a significant contribution to the early modern economy, and there were indeed many similarities in the gender division of labour at the level of general categories such as ‘agriculture’ and ‘commerce’.Footnote 14 But within those large categories there were important differences in who did particular tasks. As the gender division of labour differed significantly from many of the generalizations offered in the existing literature, we explore the reasons why it was more flexible and varied.
Sections 2 and 3 introduce the broad contours of economy and society in early modern Sweden and England and the methods used to collect evidence of women’s and men’s work. These are followed by three sections (Sections 4–6) looking in turn at gendered work activities in agriculture, crafts and construction, and commerce in the two countries. In Section 7, in conclusion, we argue that once the complexity and the flexibility of the gender division of labour are apparent, it also becomes evident that cross-cultural and mono-causal explanations of these arrangements are inadequate.
2. Sweden and England: Context
Before embarking on the analysis of work tasks, it is helpful to draw some broad comparisons between England and Sweden and their economies in the relevant period. These had similarities but also important differences. Both countries were dominated by mixed agriculture, in which livestock husbandry provided dairy products, meat, traction and the manure needed to keep arable fields fertile. The repertoire of grain crops (wheat, rye, barley and oats) and of farm animals (cattle, sheep and goats, pigs, poultry and horses) was the same. Yet England had a smaller land area and a larger population of 5.2 million in 1700, leading to a population density of 40 people per square kilometre.Footnote 15 In contrast, the population of Sweden in 1750 was 1.78 million, with only 4 people per square kilometre.Footnote 16 The Swedish sex ratio was markedly skewed: in 1750 there were only 88.7 men to 100 women.Footnote 17 In England, sex ratios show approximately equal numbers of men and women in the eighteenth century.Footnote 18 Despite its larger size, Sweden had less arable land than England: it is estimated that Sweden had 1.5 million hectares of arable land in 1800, compared to England’s 3.9 million hectares in 1700.Footnote 19 Sweden’s forests were much more extensive, covering around 70 per cent of the land, while woodland covered only 9–15 per cent of early modern England.Footnote 20 Differences in climate affected agriculture: colder temperatures meant that the growing season was shorter in Sweden.Footnote 21 The Swedish landscape was more open, whereas much of England was characterized, even in 1600, by enclosed fields surrounded by hedges. The balance within the livestock economy differed: cattle were more important in Sweden and sheep in England. It is estimated that there were 0.9 million cattle and 16.8 million sheep in England in 1600, while in Sweden in 1630 there were 1.7 million cattle but only 0.8 million sheep.Footnote 22
Contrasts extended beyond the agrarian economy. In 1751, around 10 per cent of the Swedish population lived in cities and towns, a slight increase compared to 1571 when the estimated share was 7 per cent. The largest city by far was Stockholm with c. 63,000 inhabitants in 1751. Large parts of northern Sweden lacked towns altogether.Footnote 23 The urban proportion of England’s population began at a similar level of 5 per cent in 1540 but grew more rapidly to 17 per cent in 1700, when London’s population reached 575,000.Footnote 24 Throughout the early modern period, the main Swedish export commodities were various iron and forest products.Footnote 25 In 1751, 73 per cent of the exports consisted in iron products and 13 per cent in forest products. The balance had been similar in 1643.Footnote 26 England’s exports were dominated by textile manufactures. In terms of monetary value in 1699–1701 woollen cloth and clothing made up 69 per cent of England’s exports, other manufactures 12 per cent, foodstuffs 11 per cent and primary products such as lead, tin, coal, skins and salt 8 per cent.Footnote 27 This profile was reflected in a high proportion of men with secondary (manufacturing) sector occupations in England: this grew from 22–28 per cent in 1600 to 31–42 per cent in 1700.Footnote 28 Equivalent occupational data are not available for Sweden, but population statistics suggest that in 1751 around 10 per cent of Swedish men were active in the secondary sector.Footnote 29 These figures indicate that England was more commercialized than Sweden, with a higher urban population and a larger manufacturing sector. Thus, although the economies of Sweden and England had some broad similarities, they also had several important differences, making them well suited for a direct comparison of the gender division of labour and its underlying causes.
3. Methods
The data analysed in this article were collected by two large research projects, the Swedish Gender and Work (GaW) project and the English Forms of Labour project.Footnote 30 The methods used by the two projects have many things in common. They both spring from the idea that ‘work’ should be studied as practice: what people do to provide a living for themselves and their families rather than what they are called or how much they are paid. Rather than working with occupational titles or wage data, these projects primarily use court records and court-like records to recover descriptions of sustenance activities that take the form of, or can be easily rendered as, verb-phrases: activities like ‘mow grass’, ‘weave cloth’ and ‘sell herring’, which can then be analysed quantitatively and qualitatively.Footnote 31 In doing so, they are able to capture both paid and unpaid forms of work, giving a fuller picture of the whole economy, and are particularly effective at identifying women’s work, which tends to be less visible than men’s in many other types of source material.
While the two projects share the same set of starting principles, there are a number of important differences between them that need to be addressed to usefully compare their results. There are two key types of difference: those related to the sources used by each project, and those related to different decisions they have made about whether to include, and how to categorize, certain types of activities. The latter set of issues can be more readily resolved by reworking the two datasets to be consistent with one another, as outlined at the end of this section. The differences in the court records used by each project are more difficult to overcome and need to be factored into the analysis that follows. To facilitate this, it is important to consider how each dataset has been produced, and from what records.
The English project is focused on court records in which witnesses, plaintiffs and defendants were asked to describe a series of events connected to a crime or other legally significant incident, such as the making of a will or an accidental death. In such cases those giving testimony not only commented on the incident itself but routinely offered incidental details about what they had been doing at the time or in the lead-up to the event. For example, Millicent Batchelor of Great Bradford, Wiltshire, told a local criminal court in 1673 that ‘she was coming from milking’ when she witnessed an assault which was under investigation.Footnote 32 This would be recorded as an example of ‘milking’. Such testimony was normally taken down by local officials who interviewed witnesses in advance of a court meeting, not given in open court, and the resulting depositions survive in vast numbers for the sixteenth and seventeenth centuries.
As historians well know, legal testimony of this sort cannot simply be taken as an accurate reflection of everyday realities. Those deposing often had ulterior motives for not telling the truth. That said, the method deployed here did not record the crimes or incidents that were under dispute as work activities, focusing instead on the less controversial incidental information provided. While these details may not always have been true either, they needed to be plausible to persuade the court, and can therefore be treated as reliable accounts of plausible everyday behaviours and activities. This is especially important when it comes to investigating the gender division of labour: a witness would be unlikely to describe themselves as performing an activity that was only ever performed by the opposite sex if they wanted to tell a convincing story.Footnote 33
Another potential issue with this material is that it could disproportionately record work tasks that took place in ‘criminogenic’ contexts – a witness might be more likely to see a crime while in an alehouse or at market than when working in their own home, and therefore market work would show up more often than housework. This is more of a concern when comparing the relative sizes of work categories than when considering the gender division of labour within them, which is the focus of this article: where this material captures a type of work, it tends to capture both women and men performing it. It is also important to stress that the court records used for the English project were not all criminal records. There were three main courts whose records were used: the quarter sessions, the church courts and the coroners’ courts. The first of these were criminal courts, and the majority of cases related to theft or assault. Church courts had a rather different profile, though: they had jurisdiction over various moral and administrative issues, and the main types of cases they dealt with related to defamation, matrimonial disputes, testamentary disputes and the collection of tithes. The incidental information reported in such cases originated in a wide range of contexts, including evidence of people tending to the sick and dying when they heard them declare their will, to collecting the harvest and therefore knowing how much tithe corn was due. Few if any of the church court cases were related to criminogenic contexts. Quarter sessions and church courts were the main sources used, but a small sample of coroners’ reports were also included.Footnote 34 These were accounts of the events leading up to an accidental death, and provided examples of tasks as diverse as collecting water from rivers (leading to drowning), to receiving fatal wounds while cutting bread with a knife. Taken together, evidence from these three courts comes from a range of contexts that included outside and inside tasks, group and individual tasks, and public and private tasks.
In total the project collected evidence of 9,650 tasks being performed by specific individuals, averaging 48 observations per year. Of those tasks, 5,692 fell into the categories of agriculture and land, crafts and construction, or commerce, and represent the sample analysed in this article. The precise proportion of these tasks recorded from each type of court is displayed in Table 1. This table also shows that the data cover three broad regions of England – the south west, the east, and the north – which between them include a wide range of rural economies from the lowland arable to the upland pastoral, as well as coastal communities. The majority of tasks took place in small villages or market towns, though roughly 8 per cent came from larger urban areas. No data were collected for the capital, London, but significant provincial cities like Exeter, Norwich and York (between 10,000 and 15,000 inhabitants in 1600) are represented. There is limited data available for the first half of the sixteenth century, but a fairly even spread across the period 1550 to 1700.
Table 1. Work tasks by court, region and time period, England
Note: This table only includes tasks from the three work categories used in this article: agriculture and land, crafts and construction, and commerce. The same analysis for the full dataset of all categories can be found in Appendix A. Percentage totals in this and subsequent tables do not all add up to 100.0 due to rounding.
The dataset offers a unique view of the wide range of everyday tasks, both paid and unpaid, that made up the world of work in early modern England. The resulting picture is not perfect. Certain types of work are under-represented in the English sources. An obvious example is spinning. This appears to be a quirk of the records, whereby this form of work was so ubiquitous and routine that witnesses rarely commented on it explicitly as they did most other tasks. A similar dynamic means that childcare is also under-represented. The most significant element of under-representation in the dataset is of women’s work in general. Of all tasks recorded by the project across all categories, 27.8 per cent were performed by women and 72.2 per cent by men. Given that the method incorporates both paid and unpaid work, it would be reasonable to expect that women did at least as much work as men, and that this number should be close to a 50:50 split. The explanation for the imbalance can be found in the fact that female witnesses were equally likely to describe women’s and men’s work, but men were more likely to describe the work of men, leading to a male predominance in the work tasks recorded.Footnote 35 For the English data we therefore apply a multiplier to the number of female tasks across all categories and subcategories of work, of 2.59, to ensure that the total number of tasks in the dataset is split 50:50 between male and female task performers.Footnote 36 This adjustment relies on the simple assumption that if we were to continue collecting data using the same proportions of each source type, but only of work tasks done by women, until we had an equal number of male and female tasks, those extra tasks would be distributed across the categories in the same proportion as the current sample. By creating a balanced overall sample, we get a more accurate picture of the proportion of work done by women in each category and subcategory of work, that is, of the gender division of labour.Footnote 37
The Swedish court records describe events of legal interest, including both criminal and civil cases as well as some administrative matters (like granting of permits). The broad remit of Sweden’s local courts used for this study is an advantage as it increases the likelihood that any type of work activity will turn up (with the exception of mining, as cases related to mining were handled in special courts that were not consulted here). The records present named witnesses, plaintiffs, defendants and court officials, and frequently provide additional information of tangential relevance to the legal issue but often of great interest to the labour historian. For example, in a theft case from Kålland in 1700, the wife of Anders Persson was described in passing as having spun for payment,Footnote 38 and in a rape case from 1790, Anna Catharina Lenberg of Österby, north of Uppsala, told the local court how she and some other young women had made mortar on a September morning.Footnote 39 ‘Spinning’ and ‘making mortar’ are both verb phrases included in this study.
Unlike the English records, the Swedish court records are clerks’ summaries of proceedings and seldom include separate depositions, making it always difficult to establish exactly whose voice we hear. This in turn complicates the task of establishing people’s ulterior motives for saying what they did in front of the court and, consequently, of assessing their reliability. Swedish early modern court sessions were, however, public events attended by many, and it is unlikely that apparent lies or entirely implausible accounts would have been accepted without protest. When Anna Catharina Lenberg told the court that she was making mortar when a man assaulted her, this could have been her way of depicting herself as an honest and virtuous female worker. She had not been alone on the occasion, though; others would have been able to verify or question the statement about her work. Statements about the subsequent rape were harder to verify as it took place in private, but this was a problem for the court and less so for the labour historian.Footnote 40
When some exclusions are made to make the Swedish data compatible with the English data, such as removing the work category ‘theft’ and others that were not recorded by the English project, there are 14,229 work activities recorded across all remaining categories, averaging 57 observations per year. Of those, 4,573 fall into the three categories analysed in this article: agriculture and land, crafts and construction, and commerce. The proportion of these work activities recorded from each type of court, each part of the realm and each period is displayed in Table 2. The Swedish dataset includes verb phrases from both rural and urban courts, from southern, mid and northern Sweden and from Finland (part of the Swedish realm in the early modern period). Most data come from first instance courts (rural häradsrätter as well as urban kämnärsrätter and rådhusrätter), but some were collected from the appeal court level (Göta Hovrätt 1636–1675). The urban areas include relatively small towns like Örebro, Västerås and Linköping (around 2,000 to 3,000 inhabitants in the mid-eighteenth century), Norrköping (slightly larger) and really small Ekenäs in southern Finland. Unlike the English dataset, observations from the capital are included, but Stockholm was significantly smaller than London throughout the period. The majority of work activities originate from villages, hamlets or small towns. The rural areas include rich grain-producing areas like Västmanland and Södermanland on each side of Lake Mälaren and Kålland on Lake Vänern. They also include areas in Dalarna, Jämtland and Västerbotten where the soil was more meagre and animal husbandry played an important role in the economy.
Table 2. Work tasks by court, region and time period, Sweden
Note: This table only includes tasks from the three work categories used in this article: agriculture and land, crafts and construction, and commerce. The same analysis for the full dataset of all categories can be found in Appendix A.
Certain types of work activities are either over- or under-represented in the Swedish court records. For example, administrative work is over-represented, especially administration by married men, whereas agriculture and forestry, textile work and housework are under-represented; in the case of textile work, comparisons with probate inventories and topographical descriptions make this clear.Footnote 41 It is hard to identify a general cause for under-representation. In the case of agriculture and forestry – types of work one would expect to have dominated everyday life in early modern Sweden – the low accessibility of the legal system in the countryside may be the explanation. As many rural courts convened only three times per year, incidents of potential interest to the historian may have been solved outside the judiciary. As in the case of England, both women and men appear in all categories of work, which reassures us that when the court material does capture a type of work it does so for men and women alike.
The most significant element of under-representation in the data, as in the English case, is of women’s work. Of the total work activities recorded across all categories, 24.3 per cent were by women and 75.7 per cent were by men. The urban records show more women at work than do the rural ones. In view of the skewed gender ratio in early modern Sweden, there is reason to assume that women did in fact carry out a larger share of all work in society than did men. However, as we do not know the exact gender ratio prior to 1750, and since the ratio may have varied from one place to another, we make the conservative assumption that women and men performed the same total amount of work in society, including both paid and unpaid work. Keeping with the approach of the English project, a multiplier of 3.11 is consequently applied to adjust for the under-representation of female work. This has the result of creating an overall 50:50 split between male and female tasks.Footnote 42
Many of the differences between the two datasets outlined so far are inherent to the underlying sources and cannot be easily changed. These differences will be at the forefront of the analysis that follows. Other differences between the two projects and their datasets can be more straightforwardly resolved. The two projects use different sets of overarching categories for sorting work activities into, with the Swedish project using 16 and the English project 9. This can be overcome by recategorizing activities into a common set of categories, and this type of analysis has been attempted previously.Footnote 43 The result was a very similar gender division of labour at the level of major categories such as ‘agriculture’ and ‘care’. Instead of repeating this exercise, in the following analysis we focus on three major categories of work that were already used by both projects, one from each of the primary, secondary and tertiary sectors: agriculture and land, crafts and construction, and commerce. This allows us to move beyond the observation of broad similarities to compare in more detail the gender division of tasks within these key areas of the economy, where more differences become apparent. To ensure as much consistency as possible we have recategorized tasks from both projects into consistent sets of subcategories within each of these categories. We have also removed verb-phrases from the Swedish data that relate to broader responsibilities (for example ‘run a farm’ or ‘work in service’) rather than specific tasks (as ‘mow grass’ or ‘milk cow’) as the English project collected only the latter type of evidence. Activities carried out by people of unknown gender were also removed from the Swedish data, as these were not recorded for England. As far as possible the data analysed below have been made compatible across the two projects. This allows us to narrow our analysis to an exploration of whether any differences identified are the result of our different sources or of ‘on the ground’ patterns of everyday working life, rather than differences in decisions we have made about what to record or about categorization.Footnote 44
4. Agriculture
As the largest sector of the economy, agriculture dominates traditional descriptions of the gender division of labour in non-industrial societies. The work activities described in Table 3 include the core farming activities of animal husbandry and field work (crop production) and associated activities such as farm transport (moving crops and equipment within the farm/local area), hedging and fencing to maintain field boundaries, and digging earth and marl, which in England was used to maintain soil quality. It also includes other land-management and food production activities such as wood husbandry or forestry, gathering food and fuel, and hunting and fishing. Gardening typically involved the small-scale cultivation of fruit and vegetables close to the farmhouse.
Table 3. Agriculture and land, size of sub-categories
Comparison between England and Sweden in the distribution of agricultural work activities shows very similar proportions of tasks related to crop cultivation (field work) but notable differences elsewhere. Some differences are expected. Larger proportions of wood husbandry and hunting and fishing in Sweden are no surprise given the greater extent of non-farmed land, although regulations concerning hunting and fishing boost the recording of these tasks in Sweden. The exclusion of crimes directly related to the case from the English data leads to under-recording, as hunting most often appears in prosecutions for poaching in the quarter sessions.Footnote 45 However, the relatively few tasks related to animal husbandry in the Swedish material are more difficult to explain given the importance of livestock in Swedish agriculture. Care of livestock is clearly not well-represented in Swedish court records.
What is of judicial interest affects what is observable. However, while source biases influence the proportions between categories, they do not necessarily shape the gender division within each category. Table 4 shows the gender division of labour for the subcategories of land introduced in the previous table. The adjusted figures provide the more accurate and comparable figures for women’s involvement. Women’s overall involvement in these agricultural and land management tasks was very similar in the two countries, providing almost exactly one-third of work tasks in Sweden and just over a third in England. There was close similarity in women’s involvement in some of the larger categories: the proportion of women among those doing field work was only slightly lower in Sweden, and the proportions engaged in forestry and hunting and fishing only slightly higher.
Table 4. Agriculture and land, gender division of labour
Note: Here and in subsequent tables ‘adjusted’ means that the female multiplier has been applied (x 2.59 for England; x 3.11 for Sweden), as explained in Section 3. The ‘% by F’ column is based on ‘raw’ data: the number of tasks actually recorded, without adjustment.
However, other categories of work showed substantial differences, demonstrating the importance of looking within the larger categories at the detailed patterns of work. The most notable difference is women’s dominance of animal husbandry in Sweden, where they carried out 70.0 per cent of these tasks; in England women were well-represented at 46.6 per cent, but not so dominant. Women in Sweden also played a larger role in farm transport and hedging/fencing – perhaps in connection with their responsibility for animals.Footnote 46 In England women were much more dominant in the category of gathering food and fuel. Here a key difference was the importance of gleaning in England: it accounted for 48 of the 88 tasks in this category. Gleaning, picking up uncollected stalks of grain from the fields after the harvest, was predominantly a female activity: using the adjusted figures, women carried out 87.5 per cent of these tasks in England. Gleaning was never institutionalized in Sweden, but private arrangements could at times give individuals the right to collect grain from specific fields.Footnote 47
Field work and animal husbandry were the core activities of the farming economy. Tables 5 and 6 examine the gender division of labour within them in more detail. The gendered pattern of field work was remarkably similar in the two countries. Preparing ground (ploughing and associated activities) was male dominated in both countries. Here Alesina and his co-authors were correct; where they were wrong was to assume that this activity dominated agricultural work patterns and consigned women to domestic labour. Weeding crops was dominated by women in both countries. In other key crop-related tasks such as sowing and the hay and corn harvests, women undertook around a third of the labour in both countries. Harvest work was generally carried out by larger groups of workers, and these teams were often described with vague terms such as ‘my people’ or ‘the harvesters’, which reduces the number of examples that can be used for gendered analysis.
Table 5. Agriculture and land, types of field work
Table 6. Agriculture and land, types of animal husbandry
* Unspecified animals.
While arable agriculture emphasizes strong similarities in the gender division of labour, animal husbandry emphasizes the differences. As we have already seen, women did a much higher proportion of this work in Sweden. They made up two-thirds or more of those carrying out these tasks in every subcategory of animal husbandry except care of pigs, where they made up 52 per cent, and care of horses. Caring for horses was the only subcategory where women were not in the majority. The picture for England was much more mixed. Women dominated milking and did the majority of work caring for pigs and poultry. They are well represented in the other subcategories of animal husbandry work, but outnumbered by men. Given the dominance of sheep among English livestock, the contrast in gendered patterns of labour between England and Sweden in this category is particularly notable. It was a male-dominated task in England but a female-dominated one in Sweden. In contrast to the Swedish areas under study, English sheep were often kept in large herds by professional male shepherds or other male servants. In Sweden the typical animal herder was female, the vallpigor (‘herding maids’).
To summarize, the similarity in the overall proportion of agricultural work tasks carried out by women in the two countries hides some important differences. Most notable was women’s dominance of animal husbandry in Sweden. The usual explanations for the gender division of labour offer little explanatory value in this case: caring for livestock was physically strenuous and often took place far from the farmhouse; it was not regulated by guilds.Footnote 48 The scale of farming offers no easy explanation either; while some men in England cared for large herds of sheep, they also cared for small herds. In both England and Sweden cattle were typically kept in small herds, yet women’s work in this area was much more prominent in Sweden.
5. Crafts and construction
Peasant households frequently engage in a broad range of economic activities, not just agriculture. In fact, the non-agricultural activities are often essential for their survival.Footnote 49 The same was true of early modern households in general: the combination of various types of work made households less economically vulnerable. This section looks closer at secondary sector work: the construction of buildings and the production of commodities for own use or sale. These kinds of work grew in importance after the arrival of industrialization and were then often masculinized. In the period we are looking at, secondary sector work tended to be smaller scale and there is reason to believe that it engaged many women, although it has previously proved difficult to measure their contribution directly.Footnote 50
As Table 7 shows, the composition of the two datasets for crafts and construction is similar, with the exception that there is less textile work and more building work recorded for Sweden than for England. We think that textile work is under-reported in both countries, but that the size of the textile industry in England means that it is not surprising that this is a larger category than it is for Sweden.Footnote 51 The difference in the size of the ‘buildings’ category for each country is less straightforward to explain. The sources for both countries seem to capture the same range of activities; for instance, they include not only newly constructed buildings but also a lot of maintenance and repair work. In Sweden some building activities required the completion of an apprenticeship (within a guild), and conflicts about whether or not a person had fulfilled this requirement ended up in court and show up in the sources used here.Footnote 52 While apprenticeships were also required for many forms of building work in England, these restrictions were not enforced in the court records drawn upon for this project, and this may be part of the explanation for why building activities are a smaller proportion of the English dataset.
Table 7. Crafts and construction, size of sub-categories
Another factor has more to do with different building types than source differences. In Sweden, the wooden constructions that both people and animals lived in were relatively simple to take apart and move to a new place (see Figure 1). Court records show people doing this all the time.Footnote 53 In England, more permanent structures meant that processes of construction and deconstruction were more infrequent and thus captured less often. In Basingstoke, Hampshire, in 1585, John Ewens did ‘help tile’ the roof and ‘mend and fill up with mortar certain brakes and holes in the inside of the walls’ of John Potter’s house, where ‘the weather had so driven away the earth from the rods in certain places of those walls’.Footnote 54 While the wet English weather meant that regular repairs were necessary, the use of stone, brick and cob (earth), as well as timber, combined with tiled or thatched roofs, meant that buildings were not easy to dismantle and move (see Figure 2).Footnote 55
Figure 1. An example of typical Swedish buildings from the period. Skruven två mil norr om Stockholm: Tecknat efter naturen av A F Cederholm 1813. Image info: https://www.alvin-portal.org/alvin/imageViewer.jsf?dsId=ATTACHMENT-0001&pid=alvin-record%3A94202&dswid=1790
Figure 2. An example of typical English buildings from the period. Markers Cottage, Town End, Broadclyst. Image info:https://www.geograph.org.uk/more.php?id=5124774
Table 8 shows that in both countries women performed around 42 per cent of all activities in crafts and construction, and that their presence was most striking in textile production and, to a lesser extent, in production of clothing and shoes.
Table 8. Crafts and construction, gender division of labour
What is especially striking, though, is the degree to which women dominated textile production in Sweden, performing 98.5 per cent of tasks. A more detailed breakdown of the component parts of textile production shows the very different organization of this work in the two countries (Tables 9 and 10). In Sweden, women dominated all aspects of textile production, with the weaving of linen the only area with a notable male contribution. In England the sources provide more detail on the component tasks around wool processing and show that women dominated many of these – cleaning, carding and spinning – and were involved in all areas of textile production to some extent, but men dominated the weaving of cloth and the finishing processes. These later processes were more heavily controlled by apprenticeships, which was important to the gender division of labour in English textiles. Yet in Sweden men did not take control of weaving tasks. The striking difference between the two countries here is likely related to the fact that textile production in England was a major domestic and export industry and required the labour of both women and men to sustain it – providing income-generating opportunities for both. In Sweden, where textile production was smaller scale and more domestically oriented, it may have been less necessary, or desirable, for men to direct their labour into textile production rather than into other trades such as timber and mining.Footnote 56
Table 9. Textile production, gender division of labour, Sweden 1550–1800
Table 10. Textile production, gender division of labour, England 1500–1700
* This does not include sheep shearing, which is categorized as animal husbandry.
We can identify other interesting differences where women turned up in other types of craft. In Sweden they were visible in the building sector, where they accounted for almost a quarter of adjusted observations: they made mortar and some – although not many – cases strongly suggest that they took part in actual housebuilding and house repair. Anna, a soldier’s widow, was described as having erected three buildings together with her late husband: the house where they lived, the cowshed and the pigpen.Footnote 57 Like men, women could get involved in conflicts with non-apprenticed craftsmen, or be accused of being non-apprenticed themselves and therefore not permitted to carry out some forms of work.Footnote 58 They also undertook groundwork like putting gravel on bridges.Footnote 59 In England, in contrast, evidence of women working in the building sector, or undertaking groundwork, is very rare. There is one example of a pregnant woman carrying thatch up a ladder to assist a male thatcher who was repairing her roof; one of a married woman working with her husband to build a little dovecote in their field; and one where a man and women worked together to redirect a water course.Footnote 60 Yet these are the only observations of women working in these areas. They suggest that when they did so women in England were usually working on tasks alongside men, rather than undertaking specific tasks in the sector that might be seen as women’s work. There were no doubt more occasions where this co-working occurred, but men did not mention their female co-workers in their testimony, as we know men routinely under-reported women’s work.Footnote 61
That said, there does appear to have been a significant national difference, with women more prominent in construction and repair work in Sweden than in England. One reason may have been the ease with which Swedish timber constructions could be taken apart and moved, reducing the need for specialist knowledge, tools and to some extent physical strength. Yet we should not over-emphasize the importance of building materials, as there were timber buildings in England too and women are known to have taken active part in the construction of large stone and brick buildings in Stockholm. Whatever the reasons, this serves as another example of the flexible nature of the gender division of labour and its adaptability in different national and local contexts. Moreover, there is evidence that women’s involvement in construction work also varied over time in both countries. While women were seldom recorded doing such work in early modern England, there is evidence from the medieval period of women being employed as building labourers.Footnote 62 In Sweden, women were active as producers and carriers of mortar on nineteenth-century building sites and there are images showing them in this role from the late seventeenth century. But after 1900, women left the trade as a consequence of male competition.Footnote 63 This chronological variation reinforces the point that the gender division of labour was malleable and contextual, not fixed and universal.
6. Commerce
The tertiary sector – including the provision of services, commerce, transport, management and administration activities – is generally regarded to form only a small part of non-industrial economies.Footnote 64 Yet both the Swedish and English projects have found work of this kind to be far more prevalent than estimates based on occupational titles would suggest. In part this is because these approaches include unpaid women’s work, which often fell in this sector, but another major factor is that both projects have found plentiful evidence of both women and men engaging in instances of commercial activity, such as the buying and selling of goods, related negotiations and trips to and from the market. This section explores the gender division of labour within the category of commerce. While there has been important research on women’s commercial activity in the period, the data from our projects allow us to quantify their contributions to this sector in new ways.Footnote 65
In Sweden and England these types of observation are overwhelmingly made up of fairly straightforward instances of individuals buying and selling goods, as we can see from Table 11, but there are a small number of other types of commercial activity in both datasets. ‘Going to market’ was recorded more often in the English material, though this may well have more to do with the nature of the sources than such activity being more frequent – witnesses often reported being on their way to or from market when they overheard defamatory exchanges or witnessed criminal activity, so this was often reported as an incidental detail of the sort that is more common in the English sources than the Swedish, which did not include more elaborate witness depositions, but rather summaries of court officials. Both projects record plenty of examples of buying and selling at markets, but in both cases these were recorded as buying and selling tasks, not ‘go to market’ tasks.
Table 11. Commerce, size of sub-categories
Notes: ‘Exchange’ includes ambiguous dealings such as price negotiations or viewing goods to assess/value them, where it is unclear if a sale occurred. ‘Go to market’ is essentially a travel activity of going to or coming from the market. ‘Run shop/stall’ records tasks related to doing so that were not specific sales or purchases.
Once again, the data becomes more revealing when we turn to examine the gender division of labour within commerce and its subcategories, as shown in Table 12. The adjusted figures show that women were well represented in all activities in both countries, and the overall proportion of commercial tasks undertaken by women of 51.5 per cent in Sweden and 48.0 per cent in England demonstrates that the division of labour in this sector was remarkably even. There are some interesting national differences at the subcategory level, not least of which is that women undertook a higher proportion of sales tasks in Sweden than England (56.9 per cent versus 42.9 per cent) and were more dominant in running shops and stalls too (54.2 per cent versus 45.1 per cent). Why women were more prominent in selling activity in Sweden than England is not immediately obvious, but an examination of what women, and men, were selling in each country provides some clues.
Table 12. Commerce, gender division of labour
As can be seen in Tables 13 and 14, in England and Sweden alike the majority of all sales activities for both women and men were connected to the agricultural economy, including agricultural produce, livestock and farm-produced foods. But there are some notable differences within that broader similarity. While women did play an essential role in selling agricultural products in the Swedish case (32.7 per cent), they did so to a lesser extent than in England (44.4 per cent). In both countries men were more likely to sell field crops and animal skins, whereas women made a notable contribution to selling milk and grain. A major difference is that the sale of wool does not appear in the Swedish records, but was a major agricultural product sold in England, by both men and women: it is this involvement in wool sales that explains the higher proportion of women’s sales of agricultural products in that country. As we saw in the crafts and construction section, the relative size of the textile trade in each country affected the gender division of labour.
Table 13. Types of goods sold by women and men compared, Sweden 1550–1800
Table 14. Types of goods sold by women and men compared, England 1500–1700
In the Swedish case, more than half of all goods sold by women (52.8 per cent) were processed farm foods and women undertook three-quarters of all such sales. In England, the respective figures were just 22.8 per cent and 54.0 per cent. These types of sale account for women’s greater prominence in the ‘sell’ category in Sweden overall. The sale of bread made up just over half of the instances of women selling farm foods in Sweden, whereas in England it was less than a quarter. First impressions here suggest that bread selling was a more important part of women’s work repertoires in Sweden than in England, but closer inspection of the cases in which bread selling arises suggests that this may in fact be due to differences in the source base of the two projects. Cases of women selling bread in Sweden appear in the records when they infringed on the rights of the male bakers’ guild, or those exceptions where poorer women who were given a permit to bake and sell bread to support themselves. While bread making and sales were regulated in England, this was done through urban or manor courts, neither of which were used for the data collected here. Instead, references to bread selling in England were incidental rather than regulatory, and this type of evidence was not inherently more likely to capture women’s bread-selling activities than men’s, as were the Swedish regulatory records.
This serves as a reminder that, even with the adjustments to the data we have made for this analysis, the gender division of labour seen in the results can in some cases still be influenced by differences in the records. But that does not mean that such evidence has little value in helping us to understand the reasons behind the gender division of labour more generally: indeed, it is a reminder that commerce, like crafts, could be heavily regulated, and that the more even gender division of work in this category cannot simply be seen as a consequence of commercial activity being in some way ‘freer’ and less subject to patriarchal controls than other areas of work.
A difference between the data for the two countries that appears to tell us more about variations in practices on the ground is that the proportion of women selling livestock is higher in Sweden than in the English case (37.3 per cent compared to 23.0 per cent). Men did still dominate livestock sales in Sweden, which primarily involved the sale of cattle and horses, but it is notable that there was not a particularly distinct division of labour around who sold what, with both women and men involved in the sale of large animals. In a court case in 1763, we are told about a wife who, on several occasions, visited markets and sold and bought oxen, a steer and cows.Footnote 66 In England, not only did women undertake a smaller proportion of livestock sales but the type of animals sold was more clearly gendered: 14 of their 27 livestock sales related to poultry, though the others did include pigs, cattle, sheep and a horse. For men, sheep were by far the largest category, accounting for 65.8 per cent of their 234 livestock sales, followed by cattle (22.2 per cent) and horses (9.4 per cent), in addition to some poultry and pigs. In short, men dominated the sale of larger animals, and animals kept and often sold in large flocks.
These were also the sorts of transactions that involved larger sums of money, which is reflected in the fact that the average value for the livestock sold in ‘sell’ tasks done by men in the English dataset was 696 pence, whereas for women it was just 15 pence. It may be that access to large amounts of capital or credit explains why men dominated trading in this type of animal in England, whereas women generally traded in smaller and lower-value livestock.Footnote 67 In the Swedish data, livestock sales normally involved the sale of a single animal or a pair, rather than large herds, so trading in livestock was less dependent on access to extensive credit or capital; this may help explain why women were more prominent in livestock transactions in Sweden than in England.
According to the Swedish sample, manufactured goods accounted for approximately the same proportion of goods sold by women as they did for goods sold by men (27.4 per cent women and 30.8 per cent men), but here they were partly selling different kinds of goods. Men mainly sold metalware, accounting for 45.7 per cent of 223 instances of selling manufactured goods, primarily ironware of various sorts (iron, bar iron, pig iron). There were, for example, ten men (bergsmän) and one widow who, in 1791, sold pig iron to an iron master.Footnote 68 Only occasionally did women sell iron, and metalware sold by women was mostly different silver items. The most common manufactured goods sold by women were clothes of various kinds, such as hats, gloves, scarves, sweaters, skirts and so on, and shoes and fabrics, most probably not self-made but sold second-hand.Footnote 69
In England, metalware was also the largest category of manufactured goods sold by men, accounting for 24.6 per cent of 161 instances of such sales, but women were also involved in metalware sales, and it was the third largest category of manufactured goods they sold (11 of 71), behind yarn/textiles/lace (27 of 71) and clothing/napery (18 of 71). In 1662, a Wiltshire woman sold three pieces of iron from an old cartwheel to a blacksmith for 20 pence.Footnote 70 So, while it was the case in both countries that men were more likely to sell metalware, and women more likely to sell textile and fabric-related goods, there was not a strict division of who sold what in either country, and the proportionate involvement in selling different types of goods varied between the countries across categories. In the case of commerce, we can see that the gender division of labour was not only fairly equal overall but particularly fluid and flexible in relation to selling specific types of goods.
7. Conclusions
This article sheds new light on the gender division of labour in early modern Europe. It has done so by approaching the issue in several novel ways. One is that it draws on direct empirical evidence of women and men engaged in specific work activities, rather than relying on the generalized observations of prescriptive literature. Another is that it has offered an in-depth comparison of this type of data for two countries for the first time. To make this comparison effective, the datasets for Sweden and England have been carefully crafted to ensure a high degree of similarity between what is being analysed. This includes the application of a multiplier to ensure that the overall proportion of tasks done by men and by women in each country is 50 per cent, which allows for meaningful comparisons of the gender division of labour within the categories and sub-categories that make up the economy as a whole.
There are still limitations to this data. Most obviously, the sizes of some of the categories of work captured by each project are implausibly small. This is true of spinning in both countries, whereas others are recorded more frequently in one country than in the other, such as animal husbandry being less visible in Sweden than in England. There are a variety of reasons why such activities might be under-reported by the sources, and caution needs to be exercised when comparing the relative sizes of work categories both within and across the two countries.Footnote 71 These issues are less significant for the main purpose of this article, which is analysis of the gender division of labour. While some forms of work are more visible than others in the sources, when a type of work is captured, the evidence is more reliable at capturing the proportions of it that were performed by each gender. There are some exceptions, especially where regulations had a strong gendered dimension – as we have seen in the case of women’s bread selling in Sweden – but, overall, we are confident that our analysis of differences in the gender division of labour reflects differences on the ground in each country.
Women’s contribution to all the main areas of economic activity examined here was substantial, and at the level of overarching categories the proportion of work done by women was remarkably similar between the two countries. Women performed 34–36 per cent of agricultural work, 40 per cent of crafts and construction, and 48–52 per cent of commerce. There were also some similarities within these categories: for instance, women’s participation in field work was similar at 29–33 per cent and women in both countries made up 53–54 per cent of those selling manufactured goods. At a superficial level these findings appear to support cross-cultural generalizations that suggest that the gender division of labour was similar across northern and western Europe, or even in all ‘plough societies’.
Yet there were also important differences in the gender division of labour between the two countries. Women dominated livestock husbandry in Sweden, carrying out 70 per cent of tasks compared to only 47 per cent in England. In Sweden women almost monopolized textile production, undertaking 99 per cent of tasks compared to 69 per cent in England, while in England women did more clothing production: 65 per cent of tasks compared to 50 per cent in Sweden. Women played a noticeable role in construction and groundworks in Sweden, accounting for 23 per cent and 17 per cent, respectively; in England women were almost absent from these forms of work. Women were more involved in selling livestock in Sweden, comprising 37 per cent of those carrying out this task compared to 23 in England. These differences may be partly explained by the female bias in Sweden’s sex ratio, with only 88.7 men per 100 women. The more important point, however, is that because both men and women took part in almost all types of work, this gave the gender division of labour flexibility. The differences between the two countries indicate that the gender division of labour was adaptable. Just as in the modern peasant societies surveyed by Deere, the gender division of labour varied with local circumstances.
A close examination of the differences between Sweden and England suggests that neither broad generalizations nor a single cause can explain the gender division of labour. Alesina et al. argue that men’s physical strength allowed them to dominate plough agriculture, resulting in women working in or near the farmhouse. We show that ploughing was a very small proportion of work repertoires even in these two predominantly rural societies, and that women dominated some activities that required physical strength and working well away from home, such as livestock husbandry in Sweden. Ogilvie found that guild regulation structured the overall gender division of labour in early modern south-west Germany, by excluding many women from certain types of work. We found that guilds and apprenticeship regulations affected the gender division of labour in some forms of work such as manufacturing, construction and certain types of commerce. But the all-encompassing nature of guilds in Ogilvie’s study area makes them a special case. Similar regulations did not structure other areas of the economy in Sweden and England, such as agriculture and most of commerce.
In the examination of commerce, we came closest to revealing the complexity and inter-relatedness of the gender division of labour. In England, women’s higher participation in selling farm products was caused by their role in selling wool, which was in turn related to the importance of both sheep farming and the textile industry. Women’s lesser role in selling livestock in England was in part due not only to their lesser involvement in livestock farming compared to Sweden but also to the size of herds being traded, and women’s lesser access to capital and credit, which allowed men to dominate larger transactions. In Sweden the size of transactions was generally smaller and the balance between men and women more equal. Guild regulations led to higher numbers of women being recorded selling bread in Sweden. Any explanation of women’s involvement in commerce needs to encompass all these influences.
Some differences in the gender division of labour between the two countries are inherently difficult to explain. Why did women dominate livestock husbandry and take a fuller part in construction work in Sweden but not in England, for instance? An acknowledgement of the complexity of the reasons behind the gender division of labour also means that we should not always expect to find obvious explanations. Nonetheless, while we cannot explain all the differences observed, we can confidently conclude that the gender division of labour in early modern Sweden and England, and likely the whole of northern and western Europe, was varied, flexible and adaptable. Its causes are likely to have been equally variable, relating to the particular economic, social and cultural circumstances in which women and men engaged in their struggles to make a living.
Supplementary material
The supplementary material for this article (Appendix A) can be found at https://doi.org/10.1017/S0268416025000104.
Competing interests
The authors declare none.
Open access
