A bit behind with the taxes

A Manchester team has produced a remarkable 'income return', 700 years later, Jeff Denton explains.

Computer analysis of medieval tax records has detected scribes' errors which went unnoticed for seven centuries. The work was carried out on a University of Manchester database compiled from the many manuscript versions of the 1291-92 assessment of all ecclesiastical income in England and Wales. A grant from the Economic and Social Research Council made it possible to examine the textual relations between the versions, which had not been studied in detail. The result - a new basic text of the assessment - represents a major breakthrough.

Frequent and harsh taxation of all ecclesiastical income in the late 13th century was a stark demonstration of ascendant royal power, one of many examples of the centralising bureaucratic capabilities of Edward I's government. Under the orders of Pope Nicholas IV, eager to support an English crusade, and with the king's assistance a full new assessment (known as a taxatio) was made of all Church income in England and Wales in 1291-92. The Crown took this assessment for its own taxation purposes, and used it for the next two and a half centuries. In addition to the taxation of income from ecclesiastical estates ("temporalities") and tithe-income ("spiritualities"), the Church's own taxation parish by parish was in its turn also taxed by the Crown.

The detailed taxatio of 1291-92 is an essential source for all who work on the pre-Reformation Church. It is unique of its kind for the medieval European Church and has survived in many manuscript copies. One of these, compiled in the royal Exchequer in the 15th century, provided the text which has been used by historians for almost 200 years: it was published to form the Record Commission's Taxatio Papae Nicholai IV of 1802.

It is now clear that this royal Exchequer version was one of the most idiosyncratic and least reliable of texts. The Record Commission editors added in footnotes some variants from a better version in the British Library, but their recourse to this manuscript was hasty and erratic. As a systematic and comprehensive compilation of items of income, diocese by diocese, owner by owner, the census-type assessment is one of the few excellent candidates from the medieval period for electronic treatment.

The database uses Ingres software and is held on the Cray Superserver CS6400 belonging to MIDAS (Manchester Information Datasets and Associated Services), the national datasets service at the University of Manchester.

The database is concerned with the "spiritualities" sections of the assessment: the income of the benefice-holders of some 9,000 churches, and as many again separate items of taxable income, notably the income of vicars and pension- and portion-holders. The information concerning these sections of the assessment has been entered into sets of tables, church by church, using mainly the data provided by the printed edition. Variants found in manuscript versions are being added.

Over the past 12 years work on the database has been carried out under my supervision by a team of researchers, including two volunteer research assistants, Dorothy Ross and Beryl Taylor, working throughout with the expert assistance of Sarah Davnall of the Manchester Computing Centre. Two of the basic tables can be accessed on the Internet. From a church name in its modern or medieval spelling users can find the relevant entries in the assessment, as well as full identification of each ecclesiastical benefice, together with a breakdown of its constituent values, for example the values of any vicarage, prebend, chapel, pension or portion. Nine full versions of the sections of the assessment concerning individual benefices survive: two in the British Library, two in Oxford's Bodleian Library, and others in the Public Record Office, Cambridge University Library, Canterbury Dean and Chapter Muniments, Lincoln Dean and Chapter Muniments and Lichfield Cathedral Library. These are being meticulously collated. The research has produced nothing less than a new basic text.

Entering the different material from each manuscript onto the database revealed the links between the manuscripts: they are grouped into three families which developed soon after the assessment. But the database can be used to show where errors occur. Some variations are important for they provide details or they indicate changes following local inquests soon after 1291-92, but these must be distinguished from scribal mistranscriptions. This is possible because in each manuscript totals of values are provided for each rural deanery, archdeaconry and diocese. Using the Structured Query Language facility of Ingres, discrepancies in manuscript totals and the actual totals of assessed items can be revealed and erroneous items identified. Another process has also proved valuable. Individual items appear sometimes in the form of Pounds .s.d. but most often as marks, a mark equalling two-thirds of a pound or 13s-4d. The rapid revelation of values in both forms can often demonstrate what would otherwise remain obscure. For example, it becomes clear that Pounds 9-6s-8d in one copy of a text is not a variant value for Pounds 30-0s-0d in other copies but a mistranscription of xiv for xlv marks.

The database provides the otherwise unattainable possibility of producing a core text stripped of all errors, to which significant variations from all manuscripts can be added. Although four manuscripts provide the earliest and most reliable texts (British Library Cotton Tiberius C x and Lincoln Dean and Chapter Muniments A/1/11, Cambridge University Library Mm.iii.17 and Lichfield Cathedral Library MS 23), no single manuscript provides what can be regarded as the basic text. The printed version can no longer be regarded as reliable for modern scholarly purposes. A marriage of traditional textual study with systematic electronic analysis has made a new text possible. Modern processes reveal that, while the medieval scribes made many slips, the compilers of the taxation returns were extraordinarily skilful and accurate in calculating values and in producing totals of long and complicated lists of entries. Here is a dramatic vindication of the use of information technology in textual editing.

*Part of the database is on the Internet at http://midas.ac.uk/taxatio.

Jeff Denton is professor of medieval history, University of Manchester (jeff.denton@man.ac.uk)