Modelling digital edition of medieval and early modern accounting documents

Accounting documents seem to be are well suited digital reproduction: they contain a significant number of highly structured information which was meant to be calculated already in its original context . Creating a database for executing statistical and numerical analysis of the data seems to be a promising method to analyse these documents. This assumption is not reflected in the practice of critical scholarly editions of accounting books. The digital humanists creating resources of accounting documents have not succeeded to build a common form of representation apart from the preference of spread-sheet style data storage. This approach neglects important information in the original documents which become in particular clear when studying medieval account books:

The recent scholalry edition of the Luxemburg city accounts 1 shows that accounting is close to everyday language and thus an important source for research on the history of a language. A digital resource concentrating on numbers, short descriptions of bookkeeping entries and classifications excludes that information.
Producing an account in the middle ages was a process which included several steps: collecting the information from informal documents into a fair copy, reading the listed transactions to a supervising committee or to a person using the abacus, deleting debit entries with their fulfilment, entering debits for tax or rent collection and updating the entries during money collection while tax collection or in manorial administration. Medieval accounts thus can be a protocol of acts of accounting and controlling.
The layout of medieval accounts changed in time, i.e. for example in administrative accounting the layout developed from protocols in simple text blocks to sophisticated tabular representations2. The early forms of double entry bookkeeping are based on the position of the entry on the page 3. The development of the visual form of accounting documents is part of the research on the history of accounting.
Account books give an insight in everyday life and its economy which can be researched only when the variety of existing entries is transfered into a taxonomy. Medieval and early modern accounts used variety of taxonomies to structure the data. All this can only be accessed by researcher with a thorough encoding of the “content”.
Traditionally this multiple interest in the account books led to a decision on one preferred approach. On paper the text could be printed economically only either as tables or as full transcriptions.

The current state of digital edition theory promises to solve this problem: an edition as a digital resource can include several layers of interpretation and leave the decision on the presentation to the user. But a brief survey done on relevant digital resources 4 shows these possibilities are rarely exploited in full.

The paper discusses some reasons for this situation, in particular the influence of the dominant schools in digital scholarly editing and the resulting models which focus on textual variance, representation of text and documentary edition. Syd Baumann and Kathryn Tomasek have already suggested changes in the de-facto standard for encoding of historical documents the TEI 5. They propose an element to describe transactions by referring to the textual parts describing the business facts. An alternative standard for the encoding of business facts can be found in XBRL6 which offers a flexible methodology and in the “Global Ledger”-module a taxonomy of basic business facts which can be transferred to many historical documents. While the TEI guidelines are good to encode the accounts on the documentary and linguistic level, XBRL seems to offer a good base to model the economic content. A simple bookkeeping ontology would have to include the following facts: gl-cor:entryDetail(gl-cor:account, gl-cor:signOfAmount(bk:i, bk:d), gl-cor:amount(tei:measure(tei:quantity, tei:unit), gl-cor:debitCreditCode), bk:price(tei:measure[x], tei:measure[y]), bk:transaction(bk:when, bk:where, bk:who, what, gl-cor:entryDetail[bk:debit], gl-cor:entryDetail[bk:credit]) which represents the single entry (gl-cor:entryDetail), their organization in bookkeeping (bk:account, gl-cor:debitCreditCode), the central economical informations (amounts, increase/decrease, prices) and the social act of transaction (bk:transaction).

Departing from the theoretical considerations presented by Manfred Thaller at the DH2012 7 the paper tries to develop a RDF-model which integrates the multiple layers how an account book can be conceptualised in a digital scholarly edition (text as image, as trace, as language and as meaning). The model is based on references between an URI for the physical object, the images of the object, the transcription and the bookkeeping facts (see fig. 1).

Fig. 1: conceptual model for stand-off markup of digital scholarly editions

The model can be aligned with upper level ontologies like the CIDOC-CRM 8 as the basic entities relate to linguistic objects (crm:E33, text as language), to images (crm:E38, text as image), to inscriptions/marks (crm:34/crm:37, text as trace) and to facts which can be described as events (crm:E5, business transactions as the meaning of the accounts), although substantial intermediate steps have to be included (e.g. price relations, measurements and monetary values as subclasses of propositional objects, crm:89). It allows like XBRL the inclusion of taxonomies for commodities or for types of transaction.

The model can be serialized with the help of the feature structures in TEI and converted into explicit RDF with simple XSL which includes the transformation of DOM relations (e.g. ‘contains’) into ontology statements (e.g. <list ana=”bk:d”><item ana=”bk:entry”>For <seg ana=”bk:account”>wood</seg> <measure ana=”bk:amount” quantity=”10” unit=”lb”>x lb</measure></item></list>). First experiences in a project to create a digital edition of a whole series of early modern city accounts show that the model can be used efficiently when the encoding of this kind of structures is supported with a TEI customisation which helps the transcribers to replace repetitive and verbose code with simple XML tags (like <r:lb> for <measure type=”currency” unit=”lb”> or <r:e> for <item ana=”bk:entry”> and when repetitive tasks like the transformation of roman numbers can be included in the XSL transformation from customised code to TEI.

The paper will demonstrate the application of the model to several late medieval account books in a TEI serialization9. The integration in the GAMS repository infrastructures10 allows showing possible functionalities of digital editions based on the model. This includes joint operations on the XML text stored in a Fedora Commons repository and the RDF representation stored in a triple store.

References
1. Die Rechnungsbücher der Stadt Luxemburg, bearb. v. Claudine Moulin u. Michel Pauly, z.Zt. 6 Hefte, Luxemburg 2007 - 2012 (Schriftenreihe des Stadtarchivs Luxemburg ...)

2. Mark Mersiowsky: Die Anfänge territorialer Rechnungslegung im deutschen Nordwesten. Spätmittelalterliche 2. Rechnungen, Verwaltungspraxis, Hof und Territorium (zugl. Diss. phil. Münster 1992) , Sigmaringen 2000 (Residenzenforschung 9); Georg Vogeler: Spätmittelalterliche Steuerbücher deutscher Territorien, Teil 1: Überlieferung und hilfswissenschaftliche Analyse, in: AfD 49 (2003), S. 165-295, Teil 2: Funktionale Analyse und Typologie, in: AfD 50 (2004), S. 57-204.

3. Federigo Melis (1950): Storia della Ragioneria. Contributo alla conoscenza e interpretazione delle fonti più significative della storia economica, Bologna; Franz-Josef Arlinghaus: Bookkeeping, Double-Entry Bookkeeping, in: Medieval Italy. An Encyclopedia, hg. v. Christopher Kleinhenz, New York 2004, S. 147-150; Basil S. Yamey: Scientific Bookkeeping and the Rise of Capitalism, in: EHR N.S. 1 (1949), S. 99-113.

4. Die Cameralia des Stadtarchivs Regensburg, Bd. 3: Ausgaben der Stadt Regensburg 1393 - 1394, ed. by Heidrun Boshof (Fontes Civitates Ratisponensis) www.fcr-online.com/editions/c03/index.htm; Die mittelalterlichen Schuld- und Rechnungsbücher des Deutschen Ordens um 1400. Eine synoptische Edition im Internet, ed. by Christina Link u. Jürgen Sarnowsky, Hamburg 2008, www.schuredo.uni-hamburg.de; Les comptes des consuls de Montferrand (1273–1319)2006 (Éditions en ligne de l'École des Chartes, volume16), éd. R. Anthony Lodge 2006 elec.enc.sorbonne.fr/montferrand/; Comptes de châtellenies, http: www.castellanie.net; Comédie-Française Register Project, http: web.mit.edu/hyperstudio/cfr; Open domesday. The first free online copy of Domesday Book, ed. by Anna Powell-Smith, J.J.N. Palmer, Univ. of Hull http: www.domesdaymap.co.uk; The Alcalá account book project, National University of Ireland, Maynooth 2008, http: archives.forasfeasa.ie; Henry III fine rolls Project, King’s College London et al., 2007-2011, http: www.finerollshenry3.org.uk

5. Encoding Financial Records for Historical Research, paper presented ad the TEI-MM 2012 in College Station, http: idhmc.tamu.edu/teiconference/program/papers#encfin; Syd Bauman, Transactionography Customized Documentation, Encoding Historical Financial Records Open Access Library, accessed October 28, 2013, http: omeka.encodinghfrs.org/items/show/5; hitepaper 2012, NEH Ref: HD-51224-11, omeka.encodinghfrs.org/items/show/4

6. eXtensible Business Reporting Language, www.xbrl.org, Global Ledger Taxonomy: www.xbrl.org/gltaxonomy

7. Manfred Thaller: What is a text within the Digital Humanities, or some of them, at least? digital humanities 2012 (Hamburg) www.dh2012.uni-hamburg.de/conference/programme/abstracts/beyond-embedded-markup

8. Definition of the CIDOC Conceptual Reference Model, ed. by Nick Crofts, Martin Doerr, Tony Gill, Stephen Stead, Matthew Stiff, December 2011. www.cidoc-crm.org

9. Rechnungen des Mittelalters und der frühen Neuzeit, Prototypen, ed. by Georg Vogeler, 2012ff. gams.uni-graz.at/rem

10. Geisteswissenschaftliches Asset Management System [(Humanities Asset Management System], curated by the Zentrum für Informationsmodellierung – Austrian Centre for Digital Humanities gams.uni-graz.at/context:gams/sdef:Context/get?mode=about