Working with databases: The case of the Epigraphik-Datenbank Clauss/Slaby (EDCS)

Archives and edited sources have traditionally represented crucial resources for historians. In the digital age, their function has been partly taken over by online databases. Databases allow the storage, categorisation and management of data. If they are online, they also make the data accessible to a large number of researchers. But working with databases can be problematic. As with archives, the historian has to critically assess how the content of the database was created and how the technical environment influences the way in which we access the data. Since the Epigraphik-Datenbank Clauss/Slaby (EDCS) is the most important source collection for my thesis, I would like to give a short overview of the EDCS and, after a brief introduction, reflect on its critical approach.

The EDCS database was created by Manfred Clauss and Wolfgang A. Slaby at the Catholic University of Eichstätt-Ingolstadt during the 1980s to collect all known Latin inscriptions. The database is available in German, English, French, Italian and Spanish (although the documentation is mostly in German). The database offers various search options. To search for individual inscriptions, users can either use the field “EDCS-ID” (the unique identification number of an inscription in the database) or the field “publication”. It is very important that the text entered in this field follows EDCS guidelines. Most of the time, the search term consists of the abbreviation of the publication followed by the volume and the number of the inscription in the volume (five digits, completed with zeros if necessary). For the Corpus Inscriptionum Latinarum, the most important collection of ancient Latin inscriptions, the search term has to take the form “CIL 03, 00451”. “CIL 3, 00451” or “CIL 03, 451” will not work. Another important collection is the Année épigraphique, abbreviated as AE 1973, 00375. Exceptions that are relevant for my research include the Carte Archéologique de la Gaule. To find an inscription published there, the volume and the page of the inscription is needed, e.g. “CAG-69-02, p 495”. The correct structure of the publication entry can be found in the list of abbreviations. It is also possible to search for specific words within the inscription text, which can be linked by “and”, “or” and “and not”. When searching within the inscription text, it is also possible to find non-Latin inscriptions, marked with e.g. “GR” for Greek or “HEBR” for Hebrew texts, but the actual inscription text is not shown and instead consists only of the remark “GR” or “HEBR”. Other search options include the place and province where the inscription was found, date, material and type of the inscription as well as the personal status of the people mentioned in the inscription. Furthermore, it is also possible to visualise the geographical distribution of the inscriptions using OpenStreetMap.

At the moment, Manfred Clauss, Anne Kolb (University of Zurich), Wolfgang Slaby and Barbara Woitas (Catholic University of Eichstätt-Ingolstadt) are responsible for the database. According to them, the inclusion of Latin inscriptions is “largely complete”.1 The Heidelberg Epigraphic Database (EDH), another important database of inscriptions, contains a total of 80,905 inscriptions, while the EDCS consists of 521,553 inscriptions, thus making it currently the most extensive digital collection of Latin inscriptions. Another important aspect is the question of funding and the sustainability and long-term availability of the databases. Funding for the EDH will expire in 2020 and its future is unclear. However, a long-term follow-up project under the name is planned, in which the databases of various digital epigraphic collections will be merged. In contrast, the funding and thus the availability of EDCS seems to be secure in the long term, thanks to contributions from third parties and the universities of Zurich and Eichstätt-Ingolstadt.

The following figure gives an impression of the EDCS and shows several search options to limit the search.

screenshot EDCS search options (, edited by the author).

A result looks like this (the red frame shows the transcribed inscription):

screenshot EDCS search result, edited by the author

In this case, the inscription is an epitaph found in Lyon of a Treverian wine and ceramics merchant (CIL 13, 02033). The different symbols and numbers in the text (e.g. [, ], (, ), /, ?) are not part of the inscription itself but are text-critical symbols mostly based on the Leiden system. “[3]” stands for a gap within a line, “[6]” for a gap in the scope of a line. Text between square brackets is missing on the inscription in its current state of conservation, and text between round brackets is the completion of an abbreviation. [D(is) M(anibus)] means that the text only says D M, short for Dis Manibus, but since the words are in square brackets, the part of the inscription with these words is lost. A detailed explanation of the search options in the EDCS and the resolution of the characters can be found here.

However, the EDCS is not flawless, and during my PhD, in which I am trying to reconstruct trade and transport networks during the Roman Empire in Gaul and Germania based on inscriptions, some disadvantages, which are important to know about when using the database, have become clear.2 The disadvantage that is most striking when compared with the EDH is the number of dated inscriptions. Of the 521,553 inscriptions in the EDCS, 182,647 are dated, which is roughly 35%. In the EDH, on the other hand, almost three-quarters of the 80,905 inscriptions are dated (58,905).

Another problem becomes clear when looking for inscriptions from Gaul and Germania: a search shows that there are 109,421 inscriptions from this area in the EDCS.3 Nevertheless, if you search for inscriptions in the individual provinces and then add up the number of inscriptions, the result is the following:

Province Number of inscriptions EDCS
Aquitania 16.891
Gallia Belgica 23.259
Gallia Lugudunensis 15.004
Gallia Narbonensis 22.394
Germania inferior 11.830
Germania superior 23.967
Total 113.345


So using this method generates 4,000 additional inscriptions. The reason for this is the way in which the database assigns locations to provinces. Some locations, e.g. Tongeren, Dijon, Langres or Châtillon-sur-Seine, are assigned to several provinces (labelled as “Belgica | Germania inferior” or “Belgica | Germania superior”), and as result, they appear more than once in the table above. After removing these multiple entries, the result is as follows.

Province Number of inscriptions EDCS
Aquitania 16.891
Gallia Belgica 19.335
Gallia Lugudunensis 15.004
Gallia Narbonensis 22.394
Germania inferior 11.347
Germania superior 20.526
Belgica | Germania inferior 483
Belgica | Germania superior 3.441
Total 109.421


In some cases, a place may even be geographically assigned to the wrong province. The EDCS classifies Colijnsplaat as part of Gallia Belgica, although the course of the border between Gallia Belgica and Germania inferior is not clear.4 Other geographical problems with the database become particularly visible in the representation of the geographical distribution of the inscriptions. The Porcupine Bank is located in the North Atlantic, approximately 200km west of the Irish coast, but the inscription found at that place (RIB-02-08, 02503,379) is assigned to the province of Britannia. The same applies to all inscriptions north of the Antonine Wall and from Ireland, whose find spots, according to EDCS, were part of the province of Britannia although this region was not a part of the Roman Empire.

The next problem concerns the text of the inscriptions provided by the EDCS. Although there is a list of references to various editions of the text or publications dealing with the text given, it is not clear which reading the database follows. Possible readings deviating from the EDCS are not mentioned, and critical or unclear points in the text are not marked. At the same time, the database follows some readings and additions, which are uncertain or questionable and sometimes can be proven wrong by the drawing of the inscription in the CIL or by comparison with the linked image in the EDCS. There is no documentation of changes made to entries in the EDCS. In the EDCS, CIL 13, 06496 is seen as “] / an[3]r[3]tiorii S[e]i/opensis ob(i)it morte / sua Cappadoci(a) an/norum XXXV et Nerto [”.But on the linked image of the inscription in the database “sua×Cappadocia×n” is clearly readable.

CIL 13, 06496 (EDCS-ID: EDCS-11000520,$TR_CIL_13_06496_12.jpg and$TR_CIL_13_06496_19.jpg, edited by the author).


The EDH follows the second reading. A possible reconstruction of the text is “is sua Cappadocia n[egotiator …]. This addition is only speculative, but fits the preserved text better than the EDCS version. The EDCS reads CIL 12, 04701 as “na] uclarius (?) [” and follows the CIL, although there is also a separator indicated (“V×CLARIUS”). The interpretation of the CIL can’t be ruled out but the EDCS doesn’t mention this uncertainty.

CIL 13, 00264 (EDCS-ID: EDCS-10400417,$CIL_13_00264.jpg).

Something that is missing is a description of the decor of the inscription and the option to search for inscriptions based on elements of the decor, in the same way as it is possible to search for material. The social context of the people mentioned by the text is not always clear but the decor can help. The epitaph of Andossus found in Saint-Bertrand-de-Comminges (CIL 13, 00264: | (Obito) Andosso Primuli f (ilio) / Sabina Frontonis f (ilia) / coniugi ex testamento) provides no information regarding the profession of the deceased. But the relief decorating the tombstone shows Andossus surrounded by dolphins and transporting of large wine barrels in two wagons pulled by mules, which suggests that he was active in wine transport and perhaps in the wine trade. Another well-known example is the Igel column. Its inscription gives no information about the owners except for their names, but the large reliefs on the column reveal their profession as textile merchants.


Despite the shortcomings mentioned, the EDCS is nevertheless a very important and valuable research tool in ancient history. Most of the disadvantages mentioned can be avoided by adopting a critical approach to the EDCS, especially since they do not occur very often compared to the size of the database.

Edited by Thomas Durlacher and Sarah Cooper



Wim Broeakaert, Navicularii et negotiantes. A prosopographical study of Roman merchants and shippers (Rahden/Westf. 2013).

Andreas Kakoschke, Ortsfremde in den römischen Provinzen Germania inferior und Germania superior. Eine Untersuchung zur Mobilität in den germanischen Provinzen anhand der Inschriften des 1. bis 3. Jahrhunderts n. Chr. (Möhnesee 2002).

Wolfgang Spickermann (ed.), Religion in den germanischen Provinzen Roms (Tübingen 2001).

  1. (23.02.2020). According to the database coordinators, the database consists of 99.5% of all published Latin inscriptions.
  2. There are also some other disadvantages and problems with the EDCS that I won’t go into in any detail because they were not relevant for my project, e.g. the EDCS contains multiple entries of the same inscription.
  3. February 2020.
  4. i. a. Spickermann, W. (ed.), Religion in den germanischen Provinzen Roms (Tübingen 2001), p. 8-13. For example, Broekaert, W., Navicularii et negotiantes. A prosopographical study of Roman merchants and shippers (Rahden/Westf. 2013), no. 14, 35, 37, 48, 50, 65, 74, 91, 94, 154, 157, 163, 168, 191, 203, 353, 1229, 1237, 1246, 1248, 1257, 1267, 1285, 1291, 1304, 1317, and Kakoschke, A., Ortsfremde in den römischen Provinzen Germania inferior und Germania superior. Eine Untersuchung zur Mobilität in den germanischen Provinzen anhand der Inschriften des 1. bis 3. Jahrhunderts n. Chr. (Möhnesee 2002), p. 6 (with further literature on the topic of the border between Gallia Belgica and Germania inferior) treat Colijnsplaat as part of Germania inferior.

Leave a Reply