London Lives Petitions Data

Description

Textual data and documentation relating to approximately 10,000 eighteenth-century petitions addressed to London magistrates, originally digitised as part of London Lives.

Background

London Lives 1690-1800: Crime, Poverty and Social Policy in the Metropolis is an online resource providing ‘a wide range of primary sources about eighteenth-century London, with a particular focus on plebeian Londoners’. It contains more than 240,000 manuscript and printed pages from eight London archives, including records from four of London’s main courts: the Sessions of the Peace for Middlesex, Westminster and City of London, and the Old Bailey.

In the eighteenth century, magistrates were responsible for a wide range of administrative duties in addition to criminal justice, including the administration of poor relief and the settlement laws, economic regulation (including for example, wages, alehouses and fairs), the maintenance of highways, and much local government besides. Sessions Papers are products of this work, bundles of assorted documents which clerks of the peace would file together after each court session.

There are nearly 100,000 pages of material in 1300 session files in London Lives’ four series of session papers. Not only are they the largest single body of manuscript material in London Lives, they are also by far the most diverse. They include many types of document, but probably the most important and numerous are: witness examinations and depositions; lists of those appearing before the court; court orders; and petitions.

The petitions

Petitions are ubiquitous in early modern legal archives across England and Wales and they are rich and important sources that many social historians have used extensively. As physical objects, they range from high-quality professional products to scrawled scraps of paper, reflecting the diversity of petitioners. The language of petitions is deferential, and often highly formulaic; it can also be florid, melodramatic, poignant, distressing, and manipulative. Petitions are not by any means straightforward, authentic ‘voices of the people’, but they can provide insights into plebeian experience, as well as being rich source material for local government and poor relief administration.

The petition was a standard method of communication with early modern representatives of authority. Petitions were instigated by institutions and by individuals, by elites and by paupers, and all sorts of people in between. They were used by convicted criminals to beg for the royal pardon; by officials and contractors to claim expenses for government work; by private individuals to complain about abusive behaviour by neighbours, employers, apprentices, husbands or local officials; by parishes to appeal official decisions about paupers and vagrants; to claim exemptions from local office or taxes; and more. Individuals also wrote less formal letters to officials similarly requesting some kind of relief, aid or mitigation, and this project attempts to gather those documents as well as ‘proper’ petitions, for a fuller picture of the culture and practices of petitioning. (Additionally, there are related documents for many petitions, particularly court orders, which I have not attempted to document at this stage.)

The London Lives Sessions Papers contain approximately 10,000 petitions and petitioning letters addressed to London magistrates (and a few to the judges of the Old Bailey). The sheer scale and heterogeneity of the Sessions Papers, especially the Middlesex Sessions, and the organisation (or lack of it) of the material in London Lives means that they have not been easily accessible as a body of texts for in-depth and systematic investigation of their language and their creators. The London Lives Petitions Project is intended to remedy this.

Methods and caveats

Essential background reading:

The source data

1300 XML files, part of the data for the searchable online resource at www.londonlives.org (v1.1, released April 2012). These files represent four record series held at the London Metropolitan Archives: the Sessions Papers for the City of London (SL), Middlesex (SM), Westminster (WJ) and the Old Bailey (OB) (the latter contains only a tiny handful of petitions).

The material in London Lives was transcribed using a method known as double rekeying, in which two (non-academic) typists transcribe text, the two versions are compared and only where they differ does any further manual checking take place. It is not as accurate as traditional academic standards of transcription and proofreading, but unfortunately that is not a practical option for transcription of such large amounts of text. The accuracy rate for London Lives is approximately 98-99% (at character level), but it can vary considerably between documents.

The rekeyed texts were then marked up for searching, using a combination of automated and manual tagging, for various types of information: the names of people and places, occupations and dates. There is also structural markup for features such as text in margins, deleted, obscured or illegible text, text continuing across page breaks, and so on.

The markup did not, however, include information about document structures and types. General file categorisation such as ‘Sessions Papers’ or ‘Accounts Books’ was based on the original document series’ archival organisation. There is no real structuring information between this type of generic classification and the single page. This is quite problematic in the Sessions Papers because they are both much more varied and much larger than the other documents in the resource (many of which are bound volumes rather than loose papers). Thus, simply finding the petitions was the project’s first challenge.

Data collection

There are far too many pages of material in the Sessions Papers for me to identify petitions manually, and so I employed a number of semi-automated search strategies. The likely result of this is that I have not found every petition in the original data, and there may be a few documents that are not petitions at all. Moreover, there is a degree of bias built in to the method, in favour of certain kinds of documents, which users of the data should be aware of.

The main strategy for identifying likely petitions was based on searching for certain keywords and phrases. The following are common features of petitions:

The preamble: “To The Right Honourable/Worshipful/similar title…”
Less formal letters begin with phrases like ‘Honoured Sir/Sir,’
At the beginning of the document or immediately following the preamble: “The humble petition/prayer/appeal of…”
In the body of the document: “your (humble) petitioner(s)” (also “humbly sheweth”; “May it please you/your honour”)
The sign off: “And your petitioner(s) (as in duty bound) shall ever pray etc”

The list above presents the phrases in standard modern English. But in reality the forms of the words in the data files can vary considerably, making them much harder to find:

spelling variations and the use of abbreviations in the original documents
damage to the original documents which has obscured parts of the text
rekeying errors

This necessitated the use of some complex regular expressions searches. For example, a search query for ‘The humble petition of’ could look like (th|y)e.? humb(le|ell?) peti(t|c)i?on of

Issues and limitations

I initially underestimated how many pages that were not in fact petitions might also contain some of the same phrases (particularly, for example, in orders relating to petitions). I had to cull around 300 items from my original results list and subsequently found further non-petitions (although they are often closely related material - orders or notes relating to decisions, for example). But the few remaining false positives might be less problematic than the documents that these searches might have missed.

The method is biased, obviously, towards documents that contain the selected phrases. It is most effective for finding professionally-produced petitions that adhere closely to the formal conventions of petitioning language and use consistent spelling. Non-professional petitions from individuals of lower social status are more likely to

depart to some extent from those forms
use unpredictable and inconsistent spelling
have less well-formed handwriting (which in turn made them harder to rekey accurately)

I think I have now found the vast majority of petitions and petitioning letters. In both cases, though, future additions are possible.

Keyword tagging

The keyword tagging should be regarded with considerable caution, merely as rough indicators of likely content and/or petitioner types. In most cases a tag simply means the petition refers to that subject but . Eg, the tag ‘apprentice’ may mean the petition was on behalf of an apprentice or simply that it mentions one.

Moreover, some of the tagging is probably unhelpful or simply wrong, and much of it is inconsistent (it should not be assumed that the existence of a tag means I have indentified all petitions that could have that tag). Additionally, I have focused on keywords that are of interest to me; so, for example, crime-related keywords are over-represented.

I would emphasise that the tagging is not reliable enough to use as the basis for statistical analysis.

I felt that some kind of tagging would be better than nothing, but future updates can be expected to improve significantly on this aspect of the data.

Considerations for further analysis

All of the issues described need to be taken into account for anyone trying to use the data. I have created a ‘corpus’ of plain text files but I do not yet know exactly what it will be possible to do with them, given the two compounding factors of transcription errors and spelling variations in the original documents. Corpus linguistic approaches to early modern printed texts - which are considerably more standardised than these non-elite manuscript sources - have to work hard to deal with their orthographic variations. The petitions texts might be more comparable to OCRed printed texts (such as ECCO) than to the rekeyed EEBO-TCP corpus.

To give one brief example: the top 10 variants from a search of the Middlesex Sessions petitions texts for the word “relief”, using the regex rel[ie]+f:

word	count
relief	1100
releife	536
releif	171
reliefe	101
relife	20
releiffe	8
releifed	4
reliefs	4
releiff	3
releift	2

Rekeying errors are frequently hard to distinguish from spelling variations without closer examination of the text. For example, the word ‘agoe’ (‘ago’) has been rekeyed on a number of occasions as ‘aged’. In the ‘relief’ example above, ‘relieft’ could be an unusual spelling of ‘relieved’ - or simply a random rekeying error.

However, at the very least the fact that the petitions exist as a corpus makes it possible to explore just how ‘bad’ the data is, whether it is possible to distinguishing spelling variation from transcription error, and potentially improve the text in the future.

Get the data

The latest release of the data, with documentation, can be downloaded here.

The dataset and all accompanying documentation are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Acknowledgments

The dataset has been created using the transcriptions of the Sessions Papers published at London Lives. I am deeply grateful to Tim Hitchcock and Bob Shoemaker, the London Lives project directors, for willingly agreeing to share the data.

The original documents are held at the London Metropolitan Archives.

The London Lives project (under the name Plebeian Lives) was funded by the Economic and Social Research Council between 2006-2010.