In October of 1559 Margaret Grills, a 28-year-old widow, was ‘spinning in a certain entrance of Richard Irish at Berry Pomeroy’ in south Devon. From her vantage point in this doorway, presumably chosen to maximise the autumnal light available to her as she set about her work, Margaret was able to see into Irish’s barn, where she observed his wife Elizabeth and one Joan Myller ‘winding [winnowing] of corn’. These two women discussed a rumour that the local gentleman had fathered a child on one of their neighbours, a claim that the implicated neighbour, when she heard of it, deemed defamatory; she subsequently sued Elizabeth Irish for defamation in the church courts. Margaret Grills was called as a witness to the words being spoken, and in the course of her testimony provided the details cited above.Footnote 1
This kind of evidence represents rich raw material for the work-task approach to the history of work, with deponents providing accounts of everyday work activities being undertaken at the time that a defamation, crime, or other legally significant incident took place. From this case alone, it is possible to record an example of spinning, and two instances of women winnowing, as well as contextual information such as the gender and marital status of the actors involved, and the dates and locations of work. By consulting tens of thousands of such depositions, a detailed picture of everyday working life can be constructed. That, in essence, is what this book does. But uncovering broader patterns of work through the collection and analysis of individual work tasks is not as straightforward as it might sound, and it requires a range of careful decisions to be made about precisely what material to consult, and what activities and information to record – or not record – that all have an impact on the picture that emerges. For readers to fully comprehend the results presented in this book, it is important that they are acquainted with those decisions and the particularities of the approach adopted here.
Whilst the Introduction set out the rationale and underlying principles behind our research methodology, this chapter explains in more detail how these were applied in practice. The first section discusses the types of source material used and the main challenges associated with using them. The chapter then moves on to a detailed account of our methodology for identifying and recording work tasks. This is followed by an overview of the sample of sources used, before a long section discussing the overall results of the study.
1.1 Sources
There is a wide range of sources that verb-oriented methods can and have been applied to, with account books, diaries, petitions, and legal records of various types – from prosecutions for sabbath breaking to disputes over common rights – all containing verb phrases that tell us about the ways individuals made a living.Footnote 2 The focus in this study is on records, like the example above, that contain detailed testimony given by individuals in the process of being examined in legal cases. This type of material has several advantages over other records, especially for the pursuit of the work-task variant of the verb-oriented method used here. These testimonies have a narrative element, which means they tend to be fuller and more detailed in the tasks they describe than many other types of record, and often contain richer and more precise contextual information about who was doing tasks, and where and when, than sources such as wage payments in account books. This contextual information plays a major role in the subsequent analysis. They also provide reports of work tasks that the deponent was either a participant in or a direct witness of, rather than a more abstract account of work recorded at a distance from the activities themselves: this makes it possible to establish with greater certainty who undertook specific tasks.
Whilst diaries might meet some of these criteria, legal testimony has the advantage of being far more socially inclusive and plentiful than diary material. The English legal system in the early modern period was not equally accessible to everyone, but participation in the machinery of the courts was remarkably widespread across the social scale. There were fees involved in bringing a case to the courts. Pursuing a defamation case like the one above could result in plaintiff’s costs of several pounds, or more than one hundred days’ wages for a labourer.Footnote 3 When in 1675, John Cotterell of Salisbury called Mary Bannister ‘whore’, he taunted her that ‘she had not money to begin her suite’ against him for defamation.Footnote 4 Contemporaries clearly knew that justice was more affordable to some than others. Still, Mary Bannister did manage to bring a suit against Cotterell, and the financial restraints on plaintiffs did not apply in the same way to witnesses: their costs were usually paid by the party that produced them. In the criminal courts such as the quarter sessions, men and women of any social status could be asked to give testimony if they were suspected of or witness to a crime. Whilst the testimony of the poor and of women was seen as less trustworthy than that of middling sort men, being present when an incident took place was the most important factor in determining who was called to give testimony in the courts used in this study.Footnote 5 Table 1.1 shows that where it is possible to identify the social profile of male deponents whose testimony is drawn upon in this study – including plaintiffs, defendants, and other witnesses – it is broad based.Footnote 6 As female deponents were rarely accorded social or occupational descriptors, the same analysis is not possible for female witnesses, and although the gender profile of witnessing is not balanced, a point returned to below, women do make up a significant proportion of deponents at 27 per cent; they were not excluded from giving testimony.
a Tradesmen = commercial and transport trades.
Table 1.1Long description
A table with three columns, where the first column lists occupational categories, the second lists the number of individuals in each category, and the third displays the corresponding percentages. Below is the data from the table:
For gentry and professions, the corresponding data are 109 individuals and 5.4%.
For artisans and tradesmen, the corresponding data are 696 individuals and 34.3%.
For yeomen, the corresponding data are 178 individuals and 8.8%.
For husbandmen and agricultural trades, the corresponding data are 572 individuals and 28.2%.
For labourers, the corresponding data are 219 individuals and 10.8%.
For servants and apprentices, the corresponding data are 204 individuals and 10.1%.
For miscellaneous, the corresponding data are 49 individuals and 2.4%.
For total, the corresponding data are 2027 individuals and 100%.
The legal testimony used for this book is drawn from three sources: quarter sessions courts, church courts, and coroners’ reports into accidental deaths. All of these have produced an abundance of surviving vignette-style material with a good geographical and chronological spread. The quarter sessions were county-level courts, held four times a year, which oversaw administrative matters such as alehouse licensing and labour legislation, as well as serving as a criminal court. They dealt with less serious offences, or ‘misdemeanours’, such as thefts and non-lethal acts of violence, although they did at times also try more serious crimes such as murder, which should technically have been referred to higher courts such as the Assizes. The testimony that survives from these courts takes the form of ‘examinations’: statements that victims, suspects, or witnesses gave before local magistrates as the latter investigated a crime, which were later presented as evidence when the sessions met. They are the magistrate’s summary of what the witness said, rather than their verbatim account, but they regularly contain information about everyday activities both related to and incidental to the offence being investigated.Footnote 7
Church courts had an eclectic jurisdiction, dealing with matters as diverse as disciplining clerical offences, arbitrating in disputes over church seating plans, and providing a forum for the settling of local defamation disputes. Officially these were divided into two main strands of business: office cases, brought by the Church against offenders, and instance cases, which were inter-party disputes between members of the community. It was the latter that tended to produce extensive testimonies, as parties called on witnesses to support their claims in disputes over tithe collection, testamentary administration, matrimonial contracts, and defamatory words. These witnesses travelled to the diocesan capital (such as Exeter for the Exeter Diocese) to give testimony to a court official at the cathedral or in a nearby building, though not usually in open court. They responded to a series of questions drawn up by the lawyers – or proctors – of the parties in the case, and their answers were likely taken down in note form before being written up by court officials and signed off by the deponent. Again, they are not verbatim statements, but like quarter sessions examinations those court officials who committed them to paper often included a good deal of quotidian detail that lends itself to the work-task method of analysis.Footnote 8
Between them the quarter sessions and church courts provide the overwhelming majority of the testimony drawn upon, but this material is supplemented with a third body of evidence of a different character: coroners’ reports.Footnote 9 The main point of difference is that these are not depositions given by individuals; rather, they are reports written up by a coroner undertaking an investigation of an accidental death. They do, however, have much in common with the other material, in that they take the form of narrative vignettes of the everyday activities an individual was engaged in at the time of their death, reconstructed by the coroner in consultation with a jury of local men who were tasked with gathering evidence of precisely what had happened from local witnesses. And, of course, accidental deaths could befall anyone, so these reports represent another socially inclusive archive of detailed descriptions of everyday work tasks reported by witnesses.Footnote 10
Finding legal testimonies that describe everyday work tasks in the above sources is not a problem, but that is not the same as saying these sources are not problematic. Historians have long been aware that testimony produced in early modern courts cannot simply be taken at face value; that it was crafted to persuade the judge or jury of a particular version of events rather than representing a disinterested account of ‘what really happened’.Footnote 11 Still, the stories that deponents told have continued to have an enduring appeal to historians as an access point into the lives of ordinary people in particular, which are not easily recovered through other types of source material, and they continue to be used by researchers, albeit rarely without a prefatory acknowledgement that they are not straightforward accounts of the truth. But a recent body of scholarship has called social historians to account for this approach to depositional sources, in which their mediated element is recognised but then largely set aside; something that scholars ‘acknowledge and then ignore’.Footnote 12 Frances Dolan, Heather Falvey, and Hillary Taylor have all shown in different ways that a fuller interrogation of the generic conventions, legal processes, and socio-economic structures that shaped the act of giving testimony casts serious doubt on claims that depositions represent expressions of the ‘voice’ or sentiments of ordinary people.Footnote 13
This provides a timely reminder to historians to be careful and critical when working with depositional material. For the purposes of this book, some of the issues with these sources are more pertinent than others. They are not used here to try to recover the ‘authentic voice’ or ‘interiority’ of a deponent. What they are mined for are accurate accounts of everyday activities. Can we trust that depositions provide this, given that they are in effect stories designed to persuade judges and juries? Part of our approach is to focus on incidental information, rather than the main events of a case; what people were doing at the time of, or in the build-up to, witnessing a crime or an incident. This is less likely to have been wilfully constructed than their accounts of the events at the heart of the case that were being contested by different parties. However, even this information is not always entirely innocent: witnesses – and proctors – may have reached for seemingly incidental details such as ‘sitting in a doorway’ or ‘going to market’ as handy narrative devices that gave shape and credence to an account that was in practice more complex or controversial. Witnesses were less likely to report themselves engaging in idleness or activities that were not respectable or legal. Scholars have picked up on the widespread tendency for witnesses to report seeing illicit activities through keyholes or windows, arguing that this was not simply incidental detail but a conventional motif that put the witness at a physical remove from transgressive behaviour whilst being specified in canon law as an acceptable form of proof.Footnote 14 Moral and legal discourses combined to help shape the way stories of witnessing-through-holes-in-the-wall were constructed. It is not enough, therefore, to simply ask what the details in a deposition are; we also need to think carefully about why they are there.
Accepting that some seemingly incidental details were included because they contributed to good legal storytelling does not mean that this material is incompatible with the aims of this book. Reflecting on the function of such details in testimonies suggests that they have value whether they were strictly true or not. As Dolan puts it, ‘the convention of mentioning the hole through which one witnessed what was happening next door can only function to authorize speech if litigants, defendants, and officers of the court all know that such holes do indeed exist’.Footnote 15 In other words, the details provided need to be plausible if they are to add legitimacy to a witness’ version of events. This plausibility is what makes these accounts valuable, rather than their truthfulness. We aim to recover conventional everyday behaviour: if the stories told needed to sound like conventional everyday working behaviour to persuade judges and juries, then they offer valuable evidence, even if some individual instances may not have been true.
This view of depositions as plausible accounts also provides a way around trying to judge between competing narratives of activity given by witnesses and allows the inclusion of more than just incidental details from cases. Alibis provided by individuals accused of a crime, such as a deponent claiming that a sheep they are accused of stealing was in fact bought at market, are not disinterested statements of activity. However, they are plausible vignettes of everyday activity designed to persuade a court and are therefore useful. Rather than trying to determine the truth in these cases, any account that appears to give a plausible version of an everyday work task is useful for reconstructing patterns of work. Particularly important for this book is that a key element of making an account of everyday work activities plausible would have been the gender division of labour described; it would have had to be plausible that a man or woman would have undertaken the task they were reported as doing. It is likely that even untrustworthy testimony is generally reliable with regard to the gender division of labour it reported.
The plausibility principle can be illustrated more clearly by examining some examples that reveal the dynamics behind the construction of false testimony. In June of 1630 Lewis Symons, a labourer of East Brent in Somerset, was examined by a magistrate in a quarter sessions case. Symons told his examiner that he had been offered money by one Richard Rood to ‘say and swear when he should be required … that a little before the death of Widow Andridge … he together with one John Bacon [another labourer] … were thrashing in the said Widow’s barn’ and at that time heard the said Widow say to her servant that she had left all of her estate barring 40s to the said Rood. Symons refused the offer, and neither he nor Bacon had ever undertaken any threshing for Widow Andridge: it was a ruse on Rood’s part, and that was the cause of a case being brought against him. The case demonstrates the prospect of false accounts of work tasks appearing in the archives. In this specific instance we would not have recorded them as no witness in the case claimed it to be a true account. But it is safe to assume that Rood thought this account – that two labouring men might undertake this work activity, threshing, for a third party, the widow, in a barn – was plausible enough that it would stand up to scrutiny in a testamentary case if he had persuaded the men to swear to it.Footnote 16 If they had taken the bribe, we would have recorded it as a work task.
Defendants were certainly conscious of the need to be plausible. When the butcher John Syms was accused of stealing sheep in Ditcheat, Somerset, in 1638, he admitted under examination that he had initially lied to the local constable when questioned on where he had obtained six lambs in his possession. He had told the constable that he ‘bought the said lambs of one Bennett of Glastonbury’ because ‘he was afraid that if he should have told him that he had them so far off he should have been suspected to have stolen them’. Syms had, or so he claimed, bought the lambs unusually far from home, and thought the truth would therefore sound more suspicious than a lie. He rendered his account so it would be truer to everyday work experiences, even though this technically falsified it.Footnote 17 These contextual details could be challenged in court if they were deemed dubious by other parties in a case. In a defamation case heard in the Winchester diocesan church courts in 1585, the fuller Edward Potter claimed to have overheard the exchange of defamatory words whilst ‘sitting in his shop at work and seeing and looking upon the said parties’ in their neighbouring garden. But other witnesses were prepared to depose that it was ‘not possible’ for Potter to have seen the exchange if he had been working in his shop, given the angles involved and the position of his window. If an account did have a whiff of implausibility about it, it could undermine the testimony.Footnote 18
If the reliability of legal testimony can to a certain extent be overcome by a focus on plausibility rather than truth, there are still other issues with the source material that need to be considered. The most important of these is that different courts, and different types of cases, each have inherent biases towards recording certain forms of work and not others. Most obviously, many quarter sessions examinations arise from contexts in which crimes took place, and therefore tasks undertaken in especially ‘criminogenic’ contexts, such as in alehouses or at night, may be overrepresented relative to their actual occurrence. It is not only criminal records that present such issues. The most common type of case in the church courts was tithe disputes, and as these usually focused on disagreements about the quantity of tithe corn due, this material provides a particularly high concentration of observations of related harvest work tasks. The result is that church court records capture twice the proportion of agricultural work compared to the quarter sessions material. Coroners’ reports are much more likely to include accounts of dangerous activities, such as cutting down trees, driving carts, or fetching water, than they are of low-risk tasks that were less likely to result in accidental deaths. There can also be regional quirks to the records. A particular concern with paternity cases in the Cheshire quarter sessions, for instance, can make the midwifery that was often occurring whilst a mother-in-labour was being examined about the identity of a baby’s father appear more prevalent in that county than elsewhere. Sheep theft cases were very common in the south-west of England; associated activities such as skinning and butchering the stolen sheep are therefore more likely to occur in the records of this region (although this probably reflects the local economy more than a high number of midwives delivering babies did in Cheshire). Another concern is that activities that took place in more public contexts, where they were more likely to be witnessed, may be reported more often than tasks done alone, or in more private spaces. In short, the nature of work that emerges from the depositions is shaped in a number of ways by the court, case, and context from which those depositions emerge. The danger is that the pattern of work uncovered significantly overrepresents certain tasks relative to their actual occurrence in everyday life.
It is not possible to overcome these concerns entirely, and even modern time-use studies are not able to capture work patterns with infallible accuracy, despite using custom-designed survey material. It is possible, however, to take measures that tackle the most egregious forms of overrepresentation and help to ensure that the data collected have a meaningful relationship with actual patterns of work. Crucial is the fact that our results are drawn from a range of courts, and within those courts a range of different case types. An approach that relied entirely on criminal cases would find it difficult to avoid work activities that took place in criminogenic contexts being overrepresented. But, as discussed in Section 1.3, just over half of the work tasks recorded for this study are drawn from the quarter sessions; the rest come from other courts that do not have the same inherent biases. Within the church courts material comes primarily from four types of cases: defamation, tithe, matrimonial, and testamentary. Each have their own inherent biases, but they are very different. For instance, whilst tithe cases are likely to involve examples of harvest work in the fields, testamentary cases are more likely to include care for a dying person in a private chamber. The relationship between different types of courts, cases, and types of work is shown in more detail in Appendix A. The full range of cases drawn upon is shown in Table 1.4, but it is a varied combination that together provides access to work done in a wide variety of contexts. The data captures work undertaken in both public and private spaces. It captures work done in social contexts, but also undertaken alone: an accidental death could happen to an individual working alone, for instance, and a defendant in a theft case might provide an alibi of working by themself. Roughly half of the tasks in the dataset were tasks witnesses reported themselves doing, and the other half were activities reported by someone else. This means not all tasks took place when people were being observed by others. It also mitigates concerns that witnesses may have avoided reporting themselves doing activities that were not seen as respectable: witnesses of the activities of others, who were often opponents in a legal case, would not have been constrained to the same extent. This also applies to tasks that were not technically legal, such as undertaking craft work without an apprenticeship or providing certain forms of healthcare without a license.
Drawing on a range of cases that originate in a range of contexts provides multiple perspectives on the world of work, rather than one view heavily distorted by the characteristics of a particular genre of testimony. That said, it is not as simple as the inherent biases of different types of cases ‘balancing’ each other out, and issues still remain. Other mechanisms have been used where certain activities were being significantly overrepresented in certain types of cases: a series of exclusions and filters that will be explained in more detail in the next section. Certain activities have been excluded from the data because including them would have overwhelmed the material collected. Theft is the most obvious example, but the same applies to the work of clergymen and local officials, and a fuller list of examples is discussed below. There are other types of work that are underrepresented in the material. Childcare, for instance, was often undertaken simultaneously with other tasks and therefore not always mentioned by witnesses as the primary task they were involved in when an incident occurred. Spinning is another. Witnesses, and especially female witnesses, often described themselves as sitting in doorways ‘at their work’. It is highly likely many of these women were spinning, as in the example that opened this section, but the term ‘work’ is not sufficiently specific for inclusion in the dataset. It is much more difficult to devise mechanisms that deal with this underrepresentation of certain tasks than it is to tackle forms of overrepresented activity. As a result, the patterns of work presented and discussed in this book cannot be said to fully capture every aspect of the experience of work in early modern England.Footnote 19
In summary, the material at the heart of this study presents a number of challenges when it comes to recovering the world of work. Nonetheless, the work-task approach still provides the most effective methodology for capturing the most common forms of work activity in this period that is currently available to historians, and its careful application can produce results that provide a close approximation of real patterns of work. The next section provides a detailed account of the guidelines that were used in the collection and recording of our data.
1.2 Methodology
1.2.1 Criteria for Recording Work Activities
In important ways the work-task approach used in this book differs from the pioneering works of Ogilvie and Ågren et al. that first developed the verb-oriented method. For instance, whilst Ågren’s Gender and Work project looks for verb phrases that indicate the use of time with the goal of making a living, we adopt a relatively narrow focus on descriptions of specific individuals engaged in specific tasks.Footnote 20 Examples such as ‘running a farm’, or ‘working in service’, which the Gender and Work project would include, are excluded here. These examples obscure as much as they reveal. An individual described as ‘running a farm’ may rely on others to carry out agricultural tasks – to manage the farm accounts, to take its goods to market, and to sell them. It reveals relatively little about who does what tasks on that farm. Likewise with ‘working in service’: this could cover a range of activities from agricultural labour to housework and selling goods at market that represent very different types of work. By focusing this study on evidence of specific individuals actively engaged in specific tasks, a clearer sense emerges of who is actually doing what at the everyday level, recording individuals caught in the (work) act. We describe this as the work-task approach, and as a variant of the broader verb-oriented method.
To apply this approach, quarter sessions examinations, church court depositions, and coroners’ reports were searched for work tasks that met the criteria of (a) a specific individual (b) professing to have done or observed doing (c) a specified work task:
(a) a specific individual
As a minimum requirement for an individual to be recorded their gender needed to be known, given the importance of examining the gender division of labour. Any other information (name, age, marital status, etc.) was recorded when supplied, but was not essential. It was also necessary to be confident that the individual had engaged in the work task themselves, rather than engaging someone else to do it for them. For instance, a farmer in a tithe dispute might refer to ‘growing oats on his land’, but that would not be sufficient evidence that he himself either sowed or harvested the oats: he could possibly have paid others to do all of these tasks, and therefore statements such as this were not included.
(b) professing to have done or observed doing
As discussed above, the crucial point is that we were not looking for definitive proof that an individual undertook a task, only that there was a plausible account of them doing so. For example, in 1598 one Thomasine Weather, of Uplowman in Devon, was accused of having stolen a sheep, but claimed that she could not have done so because at the time of the crime she had been elsewhere. She deposed that:
upon Thursday was fortnight about five of the clock in the afternoon, she went forth from her mother’s house to fetch a burden of wood, who went for the same to one Henry Burnard’s ground called the Butte Moore about a quarter of a mile off, and then returned home again about six of the clock in the same night, and after that stayed in her mother’s house all night.Footnote 21
True or not – and it seems from other witness statements in the case that it may not have been, as they all had different versions of precisely when she had gone out and come back, and what sort of wood she had gathered – this would nonetheless be recorded as a plausible account of a work activity: fetching wood.
(c) a specified work task
Following Margaret Reid’s third-party criterion, any activity that could be substituted with purchased goods or services qualified as a work task. Such tasks needed to be specific, such as spinning, mowing, or collecting water, rather than a more general description such as ‘service’ or ‘labouring’ that could encompass any range of tasks. Another illustrative example highlights how this applied in practice. When Alice Kingston, of Exeter, Devon, deposed in a church court tithe case in 1634, she said that she:
did live with Mr Street late of St Edmund’s parish three years and half or thereabout ended now about a year since, and in that time the said Mr Streete did keep three milk kine [cows] which commonly all the summer time did pasture upon the grounds of [Mr] Joanes in every the said years … and saith that this deponent hath divers times in the said years milked the said kine in the grounds called the shillowes.Footnote 22
Here Alice Kingston would be recorded as a specified individual, professing to have done a specific work task, milking cows. Nothing else would be recorded. Mr Street may have helped to secure a living by ‘keeping three milk cows’, but there is not enough evidence here to say what tasks, if any, he was involved in as a result. He may simply have owned them, and paid others, such as Alice Kingston, to do the specifics of ‘keeping’ them. Mr Jones may have taken cows in to pasture on his land as a way of making a living but again, the deposition is not sufficiently explicit what specific tasks this involved him in for any task by Mr Jones to be recorded.
1.2.2 Exclusions and ‘Information Status’
Whilst some verb phrases are excluded for not being specific enough, other types of activity have been excluded due to concerns that they would otherwise be drastically overrepresented in the database. The most significant of these exclusions is criminal activity, not because it was not often a form of work, but because it would have been the largest category of work tasks in the dataset given its prominence in the quarter sessions in particular. For similar reasons we also excluded the performance of any form of office-holding, including the official duties of constables and the clergy, because of the regularity with which these appeared in the legal records. For instance, constables searching for stolen goods, or clergymen performing divine service, are frequently referenced in theft and clerical misconduct cases, respectively. Any activities that constituted part of the legal process itself were also excluded due to their intrinsic overrepresentation in the legal sources: this includes activities such as the drawing up of wills in testamentary cases and any duties undertaken by an executor in the process of administering a will; any fetching of witnesses to wills or marriage contracts or any other contract or payment central to the case, or fetching of people to discuss the central features of a case with them, or fetching of suitors in matrimonial cases; the writing, witnessing, discussion of bonds, contracts, and debts when these are central to the case itself.
Other activities were excluded on the grounds of being intrinsic to certain types of commonly occurring cases. For instance, the preparation of tithe corn for collection, and its collection, was excluded as this would have been significantly overrepresented by the large number of tithe cases in the church court records. Similarly, the delivery, exchange, or purchase of gifts in matrimonial cases are excluded. The principle here was to try and avoid the overrepresentation of tasks that were not routine or typical work activities, such as making a will; searching a house for stolen goods; buying wedding gifts; burying a corpse; collecting tithe payments; giving a sermon; and ministers or officials carrying out duties. That is not to say that unusual or rare work tasks were not recorded, or that we do not see these tasks as work. Instead, excluding these relatively unusual activities that appear prominently in the records was important to producing a set of results that captured something closer to the common experience of everyday working life.Footnote 23
Certain types of tasks that were recorded are still likely to be overrepresented. For example, buying and selling frequently appears in theft cases as these activities are given as plausible alternatives to accusations of theft. Yet these were included because, unlike the examples above, the buying and selling of goods undoubtedly were routine and typical work activities, and the material provides valuable contextual detail about everyday forms of commerce. To deal with this type of task, another mechanism was adopted that helps control for the overrepresentation of certain activities without excluding them entirely: each work task recorded was assigned an ‘information status’.Footnote 24 Where the work was mentioned simply in passing and had nothing to do with the central matter of the case, for instance if a deponent was weeding in a field when they overheard defamatory words exchanged by two passers-by, the task was recorded as ‘incidental’ information. Where a work activity was at the centre of a case, and its occurrence itself was either in dispute or closely connected to a criminal activity that was the subject of a case, such tasks were recorded as ‘integral’ information. Common examples include accounts of having bought goods that an individual had been accused of stealing, and individuals conducting legitimate work tasks with stolen goods, such as boiling stolen mutton. Such activities were recorded in the deposition because of their significance in the case, and they are therefore much more likely to be overrepresented in that type of case than entirely ‘incidental’ tasks.
A third status of ‘related’ information was also used for instances that fell somewhere between the first two. A common example of this comes from tithe cases, such as the one discussed above of milking cows, where an activity was not the central matter of the case or under dispute itself but did have some bearing on the case. If an individual gave evidence that they knew how much tithe corn their master should pay because they had helped to reap their master’s corn that summer, it was the value of tithe due that was under dispute, not the activity of reaping the corn itself. Such tasks are not the central matter in tithe cases, but their recording is not completely incidental; they are likely to be heavily represented in that type of case but not to the extent that integral activities are. This classification allows analysis to filter out, where necessary, work tasks that still appear to be overrepresented by certain types of court cases even after the above exclusions have been applied.
It is also important to highlight the approach taken to ‘verbs of command’, those instances when an individual orders or asks another individual to do something, not least because the approach here differs from other projects that have used the verb-oriented method. For instance, in 1598 Edmund Bishop of Culmstock, Devon, deposed that Matthew Foxe ‘willed him’ to deliver a heffer to Tiverton.Footnote 25 In this study the delivery of the heffer would have been recorded as a work task, but ‘will another to deliver heffer’ on the part of Matthew Foxe would not have been. This is because there is not enough detail here about what this ‘willing’ involved above and beyond a speech-act, which does not easily fit within the third-party criterion. This applies where all we have is the verb of command attached to the individual doing the ordering. However, in instances where ordering/asking takes place and there is contextual detail about someone expending time and energy going somewhere to arrange something, this has been included as a work task. In other words, for it to be a work task a verb of command has to be accompanied by another verb that denotes evidence of time-use above and beyond a speech-act. This also helps to clarify that the commander has personally engaged in the task, whereas a command without this could in theory have been communicated by sending another to pass on the instruction. This approach allows the recording of some instances of this type of managerial work, without the requirement to record every, very frequently recorded and often quite abstract, verb of command as a work task.
1.2.3 Information Recorded in the Work-Task Database
When an activity met the above criteria, as much detail as possible was recorded about the case, the individuals involved, and the context in which the activity took place. First, a variety of archival information and references were recorded, including the type of case and the county it took place in. Then information was recorded about the actors and witnesses involved in a task. Where it was provided, and almost all the following categories were unevenly recorded across the sources, their name, gender, marital status, parish of residence, and age were noted. Their relationship to the task was recorded (whether they were the actor who performed the task, a witness of it, or both) and any occupational or status descriptors accorded to them, something that allows us to compare occupational titles with the work tasks an individual actually engaged in.
With regard to the task itself, a transcription was recorded in the database of the relevant passage of the deposition describing the task. This was converted into a standardised task, usually in the verb-noun format such as ‘milk cow’ or ‘plough field’, which was entered and then assigned to an overarching work category, such as agriculture and land, or commerce, and a subcategory within that. As much information as possible was recorded about the date and time at which it took place, as well as the parish and any more specific information about the task’s location. All tasks in the dataset were then categorised as taking place either in the actor’s ‘own household’ or ‘outside the household’, or as ‘unknown’ where this was not clear. ‘Own household’ represented the person performing the task’s house of residence, so a servant working in a master’s household would have been recorded as ‘own household’. ‘Own household’ included outbuildings, yards, and gardens, not just work inside of a house itself. ‘Outside the household’ included work done on, or inside, the premises of someone else, and is not therefore strictly a measure of working ‘outside’, though all of these tasks took individuals away from their own home and grounds. The task was assigned an ‘information status’, as explained above, and we recorded whether the task was done ‘for another’ and if so the nature of that relationship including if it was paid work, or part of service, if this was specified. We also recorded information about rates of pay, or the values of goods in a commercial exchange. Taken together this information allows for analysis of the ways in which both women’s and men’s work was structured by age, seasonality, location, time, employment relations, and so forth, as well as by gender.
1.3 Sample
The material that the above methodology was applied to was drawn from three regions of England: the south-west, the north, and the east. These were chosen to represent a range of landscapes, agricultural economies, and industries. For each region a sample was taken of quarter sessions material, church court depositions, and coroners’ records, focused on a selection of counties from that region. Not all types of records survived for every county studied, so the proportion of tasks recorded from each type of court is not balanced geographically, as Table 1.2 demonstrates. For quarter sessions and church courts, roughly one-year-in-ten was sampled. This meant that one year’s worth of sessions per decade, or one deposition book per decade, for periods where the records survived, was consulted, though in practice this was often adapted depending on the extent of detail in, and survival and chronological coverage of, each collection. For coroners’ records all surviving reports for that county from the period 1500 to 1600 were consulted.Footnote 26 Approximately 40,000 depositions were read in total, which yielded a dataset of 9,650 work tasks.
Notes: These numbers are based on the county the case came to court in. Some tasks took place in a different county, but the numbers of these are very small. Dates are based on earliest and latest tasks recorded from each county, and therefore covered by the data, and not necessarily the earliest and latest depositions consulted. Where a task took place much earlier than the year of the earliest deposition from a county; for example, when a task from 20 years ago was remembered, they have been excluded from the date range to avoid confusion over the years covered by those courts. Populations from Wrigley, Early English Censuses.
a Cambridgeshire quarter sessions from Isle of Ely jurisdiction only.
b Lancashire church court tasks were recorded in the Chester diocesan records.
c Yorkshire quarter sessions from North Riding and West Riding jurisdictions only.
Table 1.2Long description
The table presents a detailed overview of administrative and judicial activities across various regions. It includes nine columns of data for each county, followed by the population in 1600, total tasks performed, people per task assigned, quarter sessions tasks, quarter sessions dates, church court tasks, church court dates, coroners’ reports tasks, and coroners’ reports dates.
1. South-West Region:
For Cornwall, the corresponding data are 100,064, 35, 2859, 7, dated 1610 to 1680, 27, dated 1557 to 1597, 1, and 1516.
For Devon, the corresponding data are 261534, 1449, 180, 870, dated 1594 to 1700, 560, dated 1555 to 1692, 19, which is dated from 1515 to 1598.
For Hampshire, the corresponding data are 105384, 578, 182, 0, n/a, 529, dated 1532 to 1632, 49, which is dated from 1509 to 1599.
For Somerset, the corresponding data are 170910, 169, 101, 1223, dated 1600 to 1699, 395, dated 1532 to 1694, 76, which is dated from 1512 to 1600.
For Wiltshire, the corresponding data are 116475, 750, 155, 346, dated 1602 to 1693, 317, dated 1550 to 1678, 87, which is dated from 1504 to 1598.
For South-West Total, the corresponding data are 754367, 4506, 167, 2446, 1828, and 232.
The quarter sessions dates, church court dates, and Coroners’ reports dates are not mentioned for the total South-West region.
2. East Region
For Cambridgeshire, the corresponding data are 73318, 376, 195, 141, dated 1632 to 1664, 123, dated 1555 to 1598, 112, which is dated from 1523 to 1600.
For Hertfordshire, the corresponding data are 58766, 433, 136, 169, dated 1589 to 1695, 183, dated 1549 to 1612, 81, which is dated from 1501 to 1599.
For Lincolnshire, the corresponding data are 175173, 489, 358, 373, dated 1618 to 1699, 0, not applicable, 114, which is dated from 1509 to 1599.
For Norfolk, the corresponding data are 173113, 1195, 145, 700, dated 1591 to 1687, 328, dated 1555 to 1700, 169, which is dated from 1521 to 1599.
For Northamptonshire, the corresponding data are 93113, 241, 386, 123, dated 1658 to 1699, 0, not applicable, 118, which is dated from 1509 to 1599.
For East Total, the corresponding data are 573483, 2734, 210, 1506, 634, and 594.
The quarter sessions dates, church court dates, and Coroners’ reports dates are not mentioned for the total East region.
3. North Region
For Cheshire, the corresponding data are 74738, 557, 134, 471, dated 1581 to 1692, 86, dated 1551 to 1695, 0, not applicable.
For Durham, the corresponding data are 77355, 151, 512, 0, not applicable, 151, dated 1555 to 1662, 0, not applicable.
For Lancashire, the corresponding data are 183692, 541, 340, 521, dated 1606 to 1697, 20, dated 1558 to 1694, 0, not applicable.
For Yorkshire, the corresponding data are 369781, 1161, 319, 510, dated 1660 to 1700, 391, dated 1553 to 1694, 260, which is dated from 1508 to 1600.
For North Total, the corresponding data are 705566, 2410, 293, 1502, 648, and 260.
The quarter sessions dates, church court dates, and Coroners’ reports dates are not mentioned for the total North region.
The geographical spread of work tasks recorded is also displayed in Figure 1.1, which indicates each parish from which at least one work task was recorded. The activities come from 1,918 different parishes, meaning roughly 20 per cent of all English parishes are represented in the database. The vast majority of these were rural parishes or modest market towns, though a small percentage of tasks, at 8 per cent, occurred in large towns or cities, which roughly reflects the percentage of the population of these counties that would have lived in larger urban areas, as is discussed further in Section 3.2. Specifically, urban courts, such as borough courts, were consciously avoided for sampling precisely because we wanted the sample to be representative of the actual rural and urban division of work in this way, which using specifically urban material would have undermined.
Figure 1.1 Geographical distribution of tasks in the dataset.
Notes: Highlighted parishes are those from which the work-task dataset records at least one task.
Figure 1.1Long description
The map of England is labeled regional sample, and three distinct regions are marked in different shades. The legend in the upper corner includes North, East, and South-west. A scale bar indicates distances of 0, 50, and 100 kilometers. An arrow points north side of the map. The North region, in dark shade, covers large portions of the upper central and northeastern part of England. The East region, shown in a light shade, is concentrated along the eastern coast and parts of central England. The South West region, in medium shade, spans across southwestern counties. The regions are separated by administrative boundaries. Patterns of shading represent the spatial distribution of the areas, with a notable density of shaded areas in the North region.
Table 1.3 shows how the tasks are distributed across court type, region, and period, as well as the proportion of tasks undertaken by women in each part of the sample. Just over half of the sample is drawn from quarter sessions examinations, about one-third from the church courts, and just over one-tenth from coroners’ reports. This largely has to do with ‘hit rates’ – the percentage of depositions from a given collection that contain a work task. These vary considerably across court types, as well as from collection to collection, and even sample year to sample year, but in general they are much higher in the quarter sessions where as many as 70 per cent of examinations contain recordable work tasks. In the church courts the number is more often between 10 and 15 per cent of depositions. In other words, it is much more labour intensive to find work activities in the church courts than in the quarter sessions. Hit rates were also higher for the south-west collections than for the north and east, and the surviving material was also more abundant and detailed, hence why it comprises the largest regional sample. The paucity of activities for the first half of the sixteenth century is a consequence of the fact that only coroners’ reports provide full coverage for this period. Church court material survives in good quantities from the mid-sixteenth century onwards and quarter sessions material from the beginning of the seventeenth century onwards. The coroners’ reports used only cover the sixteenth century.Footnote 27 It is also important to note, as Table 1.2 shows, that the proportion of each type of court material used for each region varied, again due to hit rates and survival. This is important to bear in mind when comparing the findings for each region.
Table 1.3Long description
The table indicates data of tasks by court, region, and time period. The first column is for the types of courts, followed by four columns with subsequent data for the tasks, percentage, and percentage by F.
1. Court Data:
For quarter sessions examinations, the corresponding values are 5454 tasks, 56.5%, 29.4%.
For church court depositions, the corresponding values are 3110 tasks, 32.2%, 27.8%.
For coroners’ reports, the corresponding values are 1086 tasks, 11.3%, 19.8%.
For court total, the corresponding values are 9650 tasks, 100%, 27.8%.
2. Regional Data:
For South-west, the corresponding values are 4506 tasks, 46.7%, 29.1%.
For North, the corresponding values are 2410 tasks, 25.0%, 29.3%.
For East, the corresponding values are 2734 tasks, 28.3%, 24.5%.
For the regional total, the corresponding values are 9650 tasks, 100.0%, 27.8%.
3. Period Data:
For 1500 to 1549, the corresponding values are 457 tasks, 4.7%, 15.5%.
For 1550 to 1599, the corresponding values are 2455 tasks, 25.4%, 23.9%.
For 1600 to 1649, the corresponding values are 3261 tasks, 33.8%, 31.0%.
For 1650 to 1700, the corresponding values are 3477 tasks, 36.0%, 29.3%.
For Period Total, the corresponding values are 9650 tasks, 99.9%, 27.8%.
Table 1.4 shows the types of cases across all the courts that have yielded the most work activities. Theft cases from the quarter sessions produced just over 40 per cent of all tasks in the database, by far the largest proportion from any case type. This undoubtedly has a significant effect on the data and is discussed further in the results section below. Beyond theft cases work tasks were collected from a varied mix of case types and therefore contexts. Table 1.4 also demonstrates the types of cases in which women’s work activities are most likely to be recorded. The relationship between certain categories of work and types of cases is shown in Appendix A.
Table 1.4Long description
The table consists of various case types that happen in the court. The first column represents the case types, which is followed by three columns on tasks, percentage, and percent by F. The data can be arranged as follows:
Theft, 3906 tasks, 40.5%, 26.5% by females.
Tithe, 1117 tasks, 11.6%, 10.8% by females.
Accidental death, 1086 tasks, 11.3%, 19.8% by females.
Miscellaneous, 713 tasks, 7.4%, 24.3% by females.
Defamation, 688 tasks, 7.1%, 41.6% by females.
Unclear, 593 tasks, 6.1%, 26.1% by females.
Testamentary, 478 tasks, 5.0%, 35.1% by females.
Paternity, 288 tasks, 3.0%, 71.9% by females.
Matrimonial, 252 tasks, 2.6%, 43.3% by females.
Physical assault, 234 tasks, 2.4%, 30.8% by females.
Rape sexual assault, 79 tasks, 0.8%, 70.9% by females.
Adultery, 74 tasks, 0.8%, 59.5% by females.
Witchcraft, 68 tasks, 0.7%, 38.2% by females.
Murder, 41 tasks, 0.4%, 31.7% by females.
Church seating, 33 tasks, 0.3%, 18.2% by females.
Total, 9650 tasks, 100%, 27.8% by females.
Overall, these tables show that the data is not evenly spread between cases, courts, regions or over time, and these imbalances need to be kept in mind in the analysis of patterns of work that follows. That said, they also show that the dataset has a good range across the country, the period, and types of cases, allowing multiple avenues of enquiry.
1.4 Results
1.4.1 Categories of Work
To aid analysis, the 9,650 work tasks are sorted into nine main categories and 62 subcategories. Decisions about categorisation can have a significant impact on the overall shape of the results. An activity such as milking, for example, could be placed in either the category of agriculture and land or that of food processing. It involves working closely with animals and might therefore be seen as a form of animal husbandry, a branch of agriculture; but it is also closely related to dairying, which sits within the broader category of food processing. As milking was almost exclusively a female activity, placing milking within the category of agriculture, as we chose to, is more likely to challenge existing perceptions of women’s work, increasing the proportion of women’s work in agriculture, whereas placing it in food processing would have reinforced them by placing milking alongside activities such as dairying and brewing which early modern advice books conventionally allotted to women. Threshing provides a similar example: an almost exclusively male activity, it is normally placed within agriculture. However, as it prepares grain for consumption it can equally be considered a form of food processing, which is how we categorised it.
Given the complexity of categorisation, especially when converting qualitative sources into quantitative data, we were initially reluctant to impose fixed categories onto the tasks in the dataset, favouring instead a flexible approach driven by different principles for different parts of the analysis.Footnote 28 At times in this book the tasks have been reworked into different categories where there was a good case for doing so, and this is made explicit when it occurs. In general, however, it was deemed more practical to develop a set of fixed categories to make the results coherent and intelligible. The resulting categories are intended to be meaningful and useful to historians, reflecting existing sub-fields in the history of work such as agriculture, commerce, and housework. At the same time, we have also sought to consciously challenge some of the assumptions that underlie these categories, especially with regard to gender stereotypes, as the examples of milking and threshing demonstrate. Food processing has been consciously separated from housework, with which it is frequently confused, because goods produced were often intended for sale, despite being seen as forms of women’s housewifery.Footnote 29 That said, we also include activities within the broader housework category that could be commercialised, such as laundry and food and drink provision. It is important therefore for readers to look carefully at our list of subcategories to see what goes where.Footnote 30
Each task recorded was placed in a single category and subcategory for ease of analysis, although this could be a complex decision in many cases. Work activities that involved movement were particularly challenging. For example, an individual ‘carting barley’ could simply be taking it from the fields to a barn, which would logically fall in the category of agriculture. But if they were moving it over a longer distance, say to a market, this might be more appropriately categorised as transport. Contextual information in the case was used to determine the precise character of such tasks, but where the context was unclear a set of guidelines was followed to ensure that the same type of task was consistently recorded in the same subcategory and category. Chapters 5 to 8 interrogate the meanings of categories by examining the constituent tasks within them. Here we look at an overview of the results when arranged into the nine broad categories.
1.4.2 The Gender Division of Labour
Table 1.5 sorts the 9,650 work tasks recorded into the nine main categories and shows the gender division of labour in each category. What is immediately striking is that women undertook between 16.8 per cent and 69.9 per cent of tasks in every category. Whilst the gender division of labour was uneven across these categories, there was no main area of the economy from which women’s work was absent. Arguably the most significant figure in Table 1.5 is that only 27.8 per cent of all the tasks recorded were performed by women. This clearly undermines any claim that these results have a meaningful correlation with actual patterns of work, unless we were to accept that women did considerably less work than men and had more leisure time. Given that the work-task methodology records unpaid as well as paid work, and work performed both inside and outside of the home, important forms of women’s work are not systematically excluded in the way that an exclusive focus on paid work would do. Modern time-use studies of non-industrialised nations that use the third-party criterion of work adopted here produce results in which women do slightly more than half of all work recorded.Footnote 31 It would be reasonable to expect to find a similar pattern in pre-industrial England.
Table 1.5Long description
The table consists of various labour work taken up by men and women. The first column represents the work types, which is followed by three columns on F tasks, M tasks, total tasks, and percent by F. The data can be arranged as follows:
For agriculture and land, the corresponding values are 473, 2162, 2635, 18.
For care work, the corresponding values are 368, 196, 564, 65.2.
For commerce, the corresponding values are 556, 1559, 2115, 26.3.
For crafts and construction, the corresponding values are 206, 736, 942, 21.9.
For food processing, the corresponding values are 131, 476, 607, 21.6.
For housework, the corresponding values are 523, 225, 748, 69.9.
For management, the corresponding values are 172, 473, 645, 26.7.
For transport, the corresponding values are 208, 1029, 1237, 16.8.
For other, the corresponding values are 49, 108, 157, 31.2.
For total, the corresponding values are 2686, 6964, 9650, 27.8.
The main reason women’s work is under-recorded is the gender imbalance in witnesses of work tasks. The tasks in the database were reported by 3,893 different witnesses, with the average witness–both male and female–reporting 2.2 work tasks each.Footnote 32 But 73 per cent of those witnesses were male. As roughly half of tasks, at 53 per cent, were self-reported – that is, undertaken by the witness themselves – having a higher proportion of male than female witnesses produces a gender imbalance in the self-reported activities in the database. But the problem goes deeper than that. Taking only those activities where a witness was reporting someone else doing the work task, very different patterns of reporting for male and female witnesses are evident. When male witnesses reported the work of others, 75 per cent of the activities they reported were done by men, and just 25 per cent done by women. When female witnesses reported the work of others, 50 per cent of activities reported were done by men and 50 per cent were done by women. In other words, female witnesses reported an even amount of men’s and women’s work, but male witnesses reported far more men’s work than women’s work.
One possible explanation could be that everyday working life was to a certain extent segregated by gender, so men were less likely to encounter women at work in the course of their day.Footnote 33 But that does not sit well with the fact that women witnessed plenty of men at work during their everyday activities. Another explanation is that male witnesses had a tendency to overlook women’s work, or at least to underreport it, in their testimony. In January of 1611 Peter Joseph of Crediton, Devon, was spotted wearing a fashionable new falling band collar. His neighbours had a similar item stolen from them and asked Joseph how he had acquired it. He told them he had bought it from Christian Slee, the wife of a Crediton cordwainer. When Slee was herself subsequently examined, she partially confirmed his story: the falling band had been purchased from her, but she had ‘sold the same to the foresaid Peter Joseph’s wife for 12d’, not directly to Peter Joseph.Footnote 34 Wilfully or not, Joseph had obscured the work of his wife in his own testimony, claiming a work task undertaken by a member of his household as his own. In a 1680s testamentary dispute from the Devon village of East Anstey two witnesses were asked to describe the role that John Wood’s wife had played in managing their household. John Avery, a gentleman, told the court that ‘Mrs Wood was very little entrusted with the management of her said husband’s concerns save in the necessary management of the household affairs incumbent on woman to manage’, but he provided no further details. Mary Thorne, wife to a yeoman, concurred that Mrs Woods ‘was not entrusted nor did concern herself with the management of her said husband’s estate or outdoor affairs only with the necessary affairs of housekeeping incumbent of a wife to look after’, but went on to add ‘as the taking care to provide meat and other necessaries for the family and the making of butter and cheese and such like’.Footnote 35 As in many other cases, here the female witness was more forthcoming with a detailed description of women’s work than a male counterpart who described women’s work in a general, unspecific manner.
Thus, there are clearly gendered patterns to the information provided by witnesses. Even if there were an equal number of male and female witnesses in the sample, and it would be possible to apply a multiplier to the data to create this effect, there would still be a much higher number of male than female activities in the dataset due to men’s tendency to overwhelmingly report the work of other men, in a way that women did not do for other women.Footnote 36 Instead, it is possible to apply a different form of multiplier. Making the conservative assumption that at least 50 per cent of all work was undertaken by women, a multiplier can be applied that increases the total number of female work tasks to be the same as men’s. That multiplier, 2.59, can then be applied to the totals for each category. This creates an overall gender division of labour of 50:50. 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. The results of applying this multiplier can be seen in Table 1.6.
Notes: adj. = adjusted (x2.59).
Table 1.6Long description
The table depicts labour by category, and female multiplier is applied to it. The first column indicates the labour work, which is followed by six columns labelled F tasks, M tasks, total tasks, percent by F, F tasks adjusted, and percent by F adjusted.
For agriculture and land, the corresponding values are 473, 2,162, 2,635, 18, 1,225, and 36.2.
For care work, the corresponding values are 368, 196, 564, 65.2, 953, and 82.9.
For commerce, the corresponding values are 556, 1,559, 2,115, 26.3, 1,440, and 48.
For crafts and construction, the corresponding values are 206, 736, 942, 21.9, 534, and 42.
For food processing, the corresponding values are 131, 476, 607, 21.6, 339, and 41.6.
For housework, the corresponding values are 523, 225, 748, 69.9, 1,355, and 85.8.
For management, the corresponding values are 172, 473, 645, 26.7, 445, and 48.5.
For transport, the corresponding values are 208, 1,029, 1,237, 16.8, 539, and 34.4.
For other, the corresponding values are 49, 108, 157, 31.2, 127, and 54.
For total, the corresponding values are 2,686, 6,964, 9,650, 27.8, 6,957, and 50.
This represents a more realistic picture of the actual gender division of labour across the different sectors of the early modern English rural economy. The figures show that women undertook at least a third of the work activities undertaken in each category. They dominated carework and housework, as might be expected, but they were not confined to such work. Women did almost half of the recorded commercial work, food processing, crafts and construction work, and management activities, whilst also playing a significant role in agriculture and transport. Precisely what work men and women undertook within each of these broader categories form the subject of much of the rest of this book, and Appendix B presents the data broken down into our subcategories, including both our raw data and an ‘adjusted’ percentage of activities undertaken by women in each subcategory for which the multiplier has been applied. These subcategories reveal a more pronounced gender division of labour in relation to certain types of activity than the main categories capture. However, only six out of 62 subcategories are gender exclusive (hedging, mill maintenance, woodwork, tobacco processing – all gendered male; dairying; miscellaneous housework – gendered female), and only a further 13 are highly gendered, with more than 90 per cent of activities done by one gender. Over two-thirds of the subcategories have a relatively mixed-gender division of labour.
1.4.3 Work Repertoires
In addition to revealing the gender division of labour in relation to different types of work, the dataset can also be used to analyse the distribution of work activities between different categories: what we call work repertoires.Footnote 37 The relative size of each category is shown in Table 1.7. It shows that agriculture and land, and commerce, were the two largest categories of work tasks, followed by transport.
Table 1.7Long description
The table indicates various labour related tasks and the percentage of total tasks. The first column indicates the various labour work, followed by two columns on tasks and percent of total tasks.
For agriculture and land, the corresponding values are 2,635 and 27.3%.
For care work, the corresponding values are 564 and 5.8%.
For commerce, the corresponding values are 2,115 and 21.9%.
For crafts and construction, the corresponding values are 942 and 9.8%.
For food processing, the corresponding values are 607 and 6.3%.
For housework, the corresponding values are 748 and 7.8%.
For management, the corresponding values are 645 and 6.7%.
For transport, the corresponding values are 1,237 and 12.8%.
For other, the corresponding values are 157 and 1.6%.
For total, the corresponding values are 9,650 and 100%.
An important question is the extent to which these relative category sizes relate to the proportion of time expended on each type of work. Work repertoires cannot claim to be fully representative of worktime budgets, in part because certain forms of work have not been recorded, as discussed above. That said, the work tasks that have been consciously excluded are generally those that were not especially widespread or common, but might appear so from these sources, and would therefore not have accounted for a significant amount of time-use across the economy as a whole, such as the work of the clergy, or the annual collection of tithe corn. It is unlikely their exclusion significantly distorts the average work repertoire recorded here. It is the case that certain tasks may be under-recorded in the material, but this is true to some extent for all time-use surveys, none of which capture a perfect picture of working life.
In many respects, as discussed in the Introduction, the approach adopted in this study closely approximates the random-hour recall approach to modern time-use surveys, which ask participants to recall what they were doing at randomly chosen moments and reconstructs broader time-use patterns from that information. One advantage of this approach is that it is not necessary to know the duration of a specific work task undertaken, information which is typically lacking in our sources, as tasks that took up a greater proportion of people’s time would be more likely to be counted by such checks than tasks that were not so time-consuming. Similarly, those tasks that were more common are captured more often than those that were performed infrequently. It is not possible to control our sample as precisely as a modern time-use study, but we have shown that participants are broadly socially representative, and the results can be adjusted to account for imbalances in gender, as discussed above. Whilst each type of source material used has its own impact on the shape of the results, this can to a certain extent be mitigated by applying the ‘information status’ filters discussed above, and the database as a whole can be seen as three aggregated time-use surveys, each from a different court, that used in conjunction provide a more rounded picture of overall patterns of time-use. The results presented here are undoubtedly less precise than modern time-use surveys, but they approximate them to a high degree, and certainly much more closely than any other historical methodology. The work repertoires presented in this book may not, then, be exact budgets of working time-use, but they should nonetheless be seen as having a strong correlation with actual patterns of worktime distribution.
These work repertoires are especially valuable when comparing the proportions of different categories of tasks done by different groups of actors, such as women and men. However, to render these as meaningful approximations of time-use budgets for men and women, it is important to consider whether certain categories of work are being significantly overrepresented in the sources. Relevant here is our system of assigning an ‘information status’ to each task recorded, with ‘integral’ used to identify activities that were most likely to be overrepresented by certain types of cases. Table 1.8 looks at the distribution of different ‘information status’ tasks across our main categories.
Table 1.8Long description
The table indicates various labour related tasks and the corresponding statistical values. The first column indicates the various labour work, followed by six columns on integral tasks, percentage, related tasks, percentage, incidental tasks, and percentage.
For agriculture and land, the corresponding values are 314, 11.8%, 1,687, 35.7%, and 634, 28%.
For carework, the corresponding values are 27, 1%, 396, 8.4%, and 141, 6.2%.
For commerce, the corresponding values are 1,104, 41.6%, 607, 12.8%, and 404, 17.8%.
For crafts and construction, the corresponding values are 180, 6.8%, 458, 9.7%, and 304, 13.4%.
For food processing, the corresponding values are 313, 11.8%, 177, 3.7%, and 117, 5.2%.
For housework, the corresponding values are 142, 5.3%, 359, 7.6%, and 247, 10.9%.
For management, the corresponding values are 200, 7.5%, 312, 6.6%, and 133, 5.9%.
For transport, the corresponding values are 328, 12.3%, 666, 14.1%, and 243, 10.7%.
For other, the corresponding values are 49, 1.8%, 63, 1.3%, and 45, 2%.
For total, the corresponding values are 2,657, 99.9%, 4,725, 99.9%, and 2,268, 100.1%.
The most notable result here is that commercial activities were far more prevalent in integral material than in either related or incidental testimony. This is straightforward to account for: many of those integral commerce activities are instances of buying and selling that were offered as alibis when an individual was accused of theft. Theft cases were also the most common form of case type by a considerable margin and thus provide many such work tasks. Using the information status filter shows that theft cases are indeed having a distorting effect on the data, leading to the overrepresentation of commercial activities. It is preferable therefore to filter out all integral tasks to create a less distorted picture of the relative sizes of the different categories in the average work repertoire. Whilst related tasks are also to some extent shaped by the case they are recorded in, the pattern of work from related tasks is broadly similar to that from purely incidental observations of work activities, so there is not a strong case for filtering them out. Table 1.9 shows the male and female work repertoires with the integral activities removed, and displaying the main differences between the gendered repertoires.
Notes: Integral excluded.
Table 1.9Long description
The table indicates various labour related tasks and their corresponding statistical values. The first column indicates the various labour work, followed by five columns on F tasks, F repertoire in percent, M tasks, M repertoire in percent, and Percentage difference in repertoires. The data is arranged below:
For agriculture and land, the corresponding values are 411, 20.2%, 1,910, 38.5%, and 18.3.
For care work, the corresponding values are 357, 17.6%, 180, 3.6%, and negative13.9.
For commerce, the corresponding values are 272, 13.4%, 739, 14.9%, and 1.5.
For crafts and construction, the corresponding values are 157, 7.7%, 605, 12.2%, and 4.5.
For food processing, the corresponding values are 80, 3.9%, 214, 4.3%, and 0.4.
For housework, the corresponding values are 464, 22.8%, 142, 2.9%, and negative 20.
For management, the corresponding values are 98, 4.8%, 347, 7%, and 2.2.
For transport, the corresponding values are 156, 7.7%, 753, 15.2%, and 7.5.
For other, the corresponding values are 36, 1.8%, 72, 1.5%, and negative 0.3.
For total, the corresponding values are 2,031, 99.9%, 4,962, 100.1%, and 0.2.
Men’s work is heavily dominated by agriculture, at 38.5 per cent. For both women and men commerce is a significant category. For men transport, and crafts and construction made up a similar proportion of activities. The combined categories of carework and housework account for 40.4 per cent of women’s work: a significant proportion, but notably less than half of all their work tasks. What we might think of as more traditional forms of women’s work did not therefore take up sufficient time to prevent them from working in other areas of the economy, as their repertoire clearly shows. In subsequent chapters we compare the work ‘repertoires’ of different groups of workers beyond comparisons between all men and all women, and consider the influence of age, marital status, region, occupation, and a range of other factors on the experience of working life.
Filtering the data by information status and applying a female multiplier are both valuable adjustments, but they are not applied universally in the analysis throughout this book. For instance, excluding integral tasks does not have a significant impact on the gender division of labour within categories, and integral tasks are less of an issue when looking at patterns within categories rather than comparing across them. For some forms of analysis, the gender division of labour is not itself under investigation, and it has not therefore been necessary to apply the multiplier to female tasks. All subsequent tables and figures explicitly state when they have excluded integral tasks, and when they have applied a female multiplier.
1.4.4 Change over Time
There are a number of major debates about the extent of changes in patterns and intensity of work across the early modern period. The work-task methodology is more effective at providing a detailed picture of working life in the sixteenth and seventeenth centuries as a whole, which can then be compared with studies of earlier and later periods, than it is at capturing change within our period of study. This is because the different source bases are not evenly distributed across the period. Due to survival rates and changes in the details recorded, we have drawn on different proportions of coroners’ reports, church court records, and quarter sessions material for different periods within the two hundred years covered. If the tasks are divided into two roughly equal-sized samples, one covering 1500 to 1630, one 1631 to 1700, the first would draw on 45 per cent church courts, 34 per cent quarter sessions, and 22 per cent coroners’ reports; the latter on 81 per cent quarter sessions, 19 per cent church courts, and no coroners’ reports. Given that each type of court produces a distinctive repertoire of work tasks, as shown in Table 1.10, any differences between the earlier and later halves of the dataset would reflect the composition of the sample as much as any real changes in the experiences of work.
Notes: Integral included.
Table 1.10Long description
The table indicates various labour related tasks and their corresponding statistical values. The first column indicates the various labour work followed by six columns on quarter sessions tasks, Q S repertoire in percent, church courts tasks, C C repertoire in percent, Coroners’ reports tasks, and C R repertoire in percent. The data is arranged below:
For agriculture and land, the corresponding values are 970, 17.8%, 1,192, 38.3%, 473, and 43.5%.
For care work, the corresponding values are 325, 6.0%, 224, 7.2%, 15, and 1.4%.
For commerce, the corresponding values are 1,569, 28.8%, 540, 17.4%, 6, and 0.5%.
For crafts and construction, the corresponding values are 466, 8.5%, 333, 10.7%, 143, and 13.2%.
For food processing, the corresponding values are 457, 8.4%, 106, 3.4%, 44, and 4.1%.
For housework, the corresponding values are 423, 7.8%, 191, 6.1%, 134, and 12.3%.
For management, the corresponding values are 417, 7.6%, 224, 7.2%, 4, and 0.4%.
For transport, the corresponding values are 720, 13.2%, 264, 8.5%, 253, and 23.3%.
For other, the corresponding values are 107, 2%, 36, 1.2%, 14, and 1.3%.
For total, the corresponding values are 5,454, 100.1%, 3,110, 100%, 1,086, and 100%.
It is not possible to apply multipliers to balance the proportions of material used from each court in each of the two periods because the dataset contains no coroners’ data for the seventeenth century, and very little quarter sessions data for the sixteenth. An alternative is to explore change over time using only the incidental tasks in the dataset, as these are less influenced by the type of court material used and therefore less susceptible to changes in the underlying composition of the sources used. This has the disadvantage of removing all of the coroners’ data, which was all marked as related, but does at least allow some analysis of change over time within the quarter sessions and church court material. Table 1.11 uses this method to compare typical work repertoires for two different periods, using 1630 as the mid-point with roughly equal numbers of tasks recorded on either side of this date. What is notable here is how muted changes in work repertoires appear to be, overall and for both men and women. Only the proportion of agricultural work undertaken by men, which increases by 6.9 per cent in the latter sample, represents a change of more than 4 per cent across all other categories. Tentative as this may be, it does suggest that the proportion of different categories of work undertaken within the early modern English rural economy did not change dramatically across our period. This contradicts recent research on occupational structure, a point we return to in the book’s conclusion.Footnote 38
Notes: Incidental tasks only.
Table 1.11Long description
The table indicates various labour related tasks and the corresponding statistical values. The first column indicates the various labour work, followed by periods of 1500 to 1630 repertoire in percent, 1631 to 1700 repertoire in percent, 1500 to 1630 F repertoire in percent, and 1631 to 1700 F repertoire in percent, 1500 to 1630 M repertoire in percent, and 1631 to 1700 M repertoire in percent.
For agriculture and land, the corresponding values are 26.6%, 29.6%, 23.7%, 20.7%, 28.3%, and 35.2%.
For care work, the corresponding values are 6.5%, 5.9%, 9.4%, 11.8%, 4.7%, and 2.2%.
For commerce, the corresponding values are 18.5%, 17.0%, 16.2%, 18.2%, 19.9%, and 16.2%.
For crafts and construction, the corresponding values are 14.3%, 12.4%, 11%, 9.9%, 16.2%, and 14%.
For food processing, the corresponding values are 5.2%, 5.2%, 5%, 2.5%, 5.2%, and 6.9%.
For housework, the corresponding values are 11.3%, 10.4%, 23.5%, 22.4%, 4.1%, and 2.8%.
For management, the corresponding values are 5.5%, 6.3%, 2.9%, 5.2%, 7.1%, and 7.0%.
For transport, the corresponding values are 10.3%, 11.2%, 6.8%, 7.4%, 12.4%, and 13.6%.
For other, the corresponding values are 1.9%, 2.1%, 1.5%, 2.0%, 2.1%, and 2.2%.
For total, the corresponding values are 100.1%, 100.1%, 100%, 100.1%, 100%, and 100.1%.
For total tasks, the corresponding values are 1,220, 1,048, 456, 406, 764, and 642.
1.5 Conclusion
This chapter has set out the methodology and source material of this study, laying the foundation for the following chapters and offering a model that we hope other researchers will follow. Like all primary source materials, the legal testimony used in this project must be handled with care and requires careful reflection. It is never possible to ‘fix’ all of the complexities of historical sources, but we have developed a methodology that allows this complex qualitative material to be converted into meaningful quantitative data about patterns of work in early modern England, shedding light on aspects of the economy that have previously been largely hidden to historical research. The headline results are striking, especially with regard to the extent to which women worked across all the main areas of the economy. But the richness of the data, and especially the contextual information relating to each work task, means that the experience of work in early modern England can be explored at a much finer-grained level than the broad overall results presented here. Proof of the efficacy of the methodology lies in the chapters that follow.
Open access