La base de données en ligne Trismegistos Places est un outil de recherche indispensable pour la toponymie de l’Égypte gréco-romaine. Elle succède au Fayum Project et appartient à la base Trismegistos, qui est développée au sein du projet Multilinguisme et Multiculturalisme dans l’Égypte gréco-romaine. Dr. Herbert Verreth de la section ‘Histoire ancienne’ de la KU Leuven en Belgique nous en dit plus:
The online database Trismegistos Places of the KU Leuven is a key research tool for the toponymy of Graeco-Roman Egypt. It is the successor of the Fayum Project and belongs to the larger Trismegistos database, which is developed as part of the project Multilingualism and Multiculturalism in Graeco-Roman Egypt. Dr. Herbert Verreth from the research unit ‘Ancient History’ of the same university tells us more about it:
To begin with, could you tell us a few words about your interest in toponymy and about the history of the Trismegistos project.
Already during my studies of Classical Philology and Archaeology I had a keen interest in geography and toponymy and so chose to make my PhD on the historical topography of the northern Sinai in the Graeco-Roman period, the area in Egypt between the Suez canal and the border with Israel along the Mediterranean coast.
I finished my PhD in 1998, but at the same time, from 1993 on, I was involved in the development of a database version of the printed Prosopographia Ptolemaica, which had been compiled at the university of Leuven between 1950 and 1981. The basic concept was simple: we created three separate Filemaker files that were linked in a relational database. The first file contained the metadata for all the ancient documents mentioning persons linked with Ptolemaic Egypt. These metadata included the place of origin, the date, the language and the type of document, its main publications and the museum or institution where the document is now kept. So for every text a card was created; at first we identified the card with its major publication, but after a while we chose for an identification through a unique, sequential number. The second file contained all the persons involved, and the third file contained all the references to these persons in the different ancient texts. One could therefore ‘jump’ from the text card of a certain document to all the persons listed in that document (and vice versa). In a next phase Prof. Willy Clarysse started the Fayum Project, another database which gathered all the toponyms known from the Arsinoites and provided the references to the ancient documents where these toponyms occurred. We used the text file of the Prosopographia Ptolemaica as basis, and added two new files. The first one, which we called ‘Villages’ or ‘Geo’, contained all toponyms from the Arsinoites; the second one, which we called ‘Georef’ contained all the references in the ancient sources to these toponyms. Later on new projects were developed, including more detailed files for literary texts (the Leuven Database of Ancient Books or LDAB), files for the archives to which a document might belong, files for the musea where the documents are now kept, files for the place where a document was found, written or send to, and so on, but all of them linked to the central text file, which kept expanding. When Prof. Mark Depauw entered the research unit, in a collaborative effort between Cologne and Leuven sponsored by a Kovalevskaja Award of the Humoldt Foundation, we more or less left the Ptolemaic Period and the Fayum as main research items and expanded our field of research over the whole of Egypt from the eighth century BC till the eighth century AD, including all papyrological and epigraphical documents from Egypt written in Greek, Latin, Demotic, Coptic, Hieroglyphic or other languages. The whole of all these different files were reorganised in 2006 in a structure that would be easier to consult online. For this we made up the name ‘Trismegistos’, which actually started as a kind of a joke: there is of course the link with the Egyptian Hermes Trismegistos, but Trismegistos in itself means ‘three times great’ or ‘super big’, a name we found fitting for a database that now counts several millions of cards spread over dozens of files, all linked with each other through unique numbers.
I am very glad that I could participate from 1993 till now in one way or another in most of these projects. I could of course not always spend all my time on the geographical items, but I gradually found the time to expand the original Fayum Project to the present-day Trismegistos Places.
What is Trismegistos Places?
Trismegistos Places is the part of the overall Trismegistos database that focuses on geographical issues. Apart from the central text file, it mainly consists out of three files. The first one, Geo, contains (or will contain) all ancient toponyms known in Egypt between the eighth century BC and the eighth century AD, regardless of the language in which they are documented. In this file each toponym receives a unique number. Some of the places in Geo are also plotted on Google maps. The second file, Georef, contains all the references in the ancient sources to the places occurring in Geo. The third file, Geotex, contains for every document in the database the place where it has been found, written or sent to. Quite often an inscription has been found in a place for which the ancient name is not yet known, so we also needed cards in Geo that contained only modern Egyptian toponyms. Because the Leuven Database of Ancient Books includes literary texts from out of Egypt, also ancient and modern toponyms from Ireland till Iran found their way into the database.
Why do you insist on these unique numbers?
It took us a while before we realized that the only practical way for a computer to cope with so much data, is to use unique numbers, without any meaning attached to them whatsoever. This goes counter-intuitive to the widely-spread scholarly habit of grouping certain items by composing specific identification-tags that tell the user immediately to which group a certain item belongs. Computers, however, have huge problems coping with numbers such as ‘3.16 B’, and prefer a continuous numbering. At the same time, however, switching to unique numbers gave us an enormous advantage from a user’s point of view: most of the separate files in Trismegistos need to show for every card e.g. the publication, the date and the language of the document in question. At first we therefore had to copy manually all this information from the main text file in to the other files. This in itself was a lot of work, but the problems really started when a new date was suggested for the text, or when a new edition was published. In that case we had to change the old information manually in all the files where data from this text card were used, and with such a diverse team working on the database it was not easy to keep it all uniform and up to date. With the unique Trismegistos number this problem does not exist anymore: by adding the TM number to a Georef card, all necessary fields of the text card are automatically shown in Georef, so that we always have the most recent information available. And the same goes for the number each toponym receives in Geo: if e.g. a new reading is suggested for a toponym, we make the change once in Geo and automatically the new information is shown in Georef and Geotex. This makes our lives a lot easier!
How does Trismegistos Places work ?
There are two possible ways to search the data. The first one is text based. You read a papyrus or inscription and notice a toponym, about which you want some more information. You look for that text in the Trismegistos database and find a list of all toponyms mentioned in that text, arranged by line number. You choose the toponym you want, and you are directed to a specific Georef card. There you can find several kinds of information, each one presented in a specific field: (1) the most recent publication for the document, followed by the line number where the toponym occurs; (2) the date of the document; (3) the standard name of the toponym preceded by a code representing the nome or country to which it belongs; (4) the toponym spelled in the original language as it occurs in that specific text, including all the square or round brackets or the uncertainty dots; (5) the same toponym but now in the nominative case (at least for Greek and other declined languages) and without all the diacritical signs (a much easier form to sort upon or to look for); (6) a translation of the passage where the toponyms occurs, or a short description of its context; (7) bibliographic references for that specific passage; (8) remarks about old readings or divergent interpretations; if the text provides information about the administrative district to which the place belongs and about the ‘kind’ of toponym it is (e.g. ‘the kome of Philadelpheia in the Arsinoites nomos’), this information is rendered in the fields ‘administrative situation’ (9) and ‘status’ (10). Finally, if the toponym occurs in a text for which the full text is available (e.g. in the Duke Databank of Documentary Papyri or in the Greek Inscriptions of the Packard Humanities Institute), a link is shown to that full text.
From this Georef card one can click further on to the main Geo card, where all the general information about the place is brought together: its standard name in several ancient languages, its variant names, its administrative situation and status, bibliographical references to the general geographical corpora discussing the place and to more specific studies.
The second way to search the data starts from this Geo file. You search for a specific place, you go to the corresponding Geo card and there you find a list, chronologically arranged, of all the references to the documents mentioning this toponym. You click on one of the references and the Georef card appears with the information described above. From there on you can click to the text card for that document.
Now, a few figures… how many place-names are kept in the database?
Here are some figures from the Trismegistos database on 10 September 2012:
- number of texts: 134.943
- number of places (Geo): 14.289
within Egypt: 11.063
outside of Egypt: 3.226
- ancient and modern toponyms where texts have been found or written (Geotex): 1.721
- ancient toponyms with references to the texts where they occur (Georef): 12.140
- number of references to a toponym (Georef): 114.155
in Greek texts: 94.221
in Latin texts: 4.924
in Demotic texts: 13.243
in Coptic texts: 349
in Hieroglyphic texts: 298
in Hieratic texts: 457
in Abnormal Hieratic texts: 263
in Aramaic texts: 280
in Arabic texts: 187
and a few toponyms in Armenian, Hebrew, Minaean, Nabataean, Pehlevi, Phoenician and Syriac texts.
It is clear from these figures that the Greek, Latin and Demotic toponyms are quite well represented, but there still seems to be a lot of work for the other languages?
Yes, the database is far from finished. As so often in the scientific world, it all depends on the funding that can be made available for a certain project. We also started with those languages for which we had the expertise in our own research unit, i.e. Greek, Latin and Demotic. This is also one of the reasons why we were so pleased that Alexander Schütze from the university of München provided us with the list of all toponyms he found in the Aramaic documents from Egypt, an area for which we had no specialists in Leuven.
We are currently developing ways in which the scholarly community can also contribute to the database through an online form to enter new data. This will probably become available late 2012, early 2013.
Those 114.000 references, have they all been entered one by one, manually, just by reading through the ancient texts?
For the references to the Greek and Latin epigraphical and literary texts and for all the Demotic texts, the information has indeed been compiled by going through the indices of the major publications and then by checking the more wide-spread publications in journals, Festschrifte and such. This is enormously time consuming and a full coverage of all toponyms in these texts is not possible without (expensive) years of gathering information.
With regard to the Greek and Latin papyrological sources, however, we could cooperate with the Duke Databank of Documentary Papyri, which kindly put their full text databank at our disposal for the somewhat 50.000 papyri, ostraca and tablets thus far published. Bart Van Beek, our ICT specialist, and Mark Depauw devised a way to automatically identify all words in these texts starting with a capital and linking them with a list of both personal names and place names available from the files People and Geo. This yielded – after a few months of work for the whole team – more than 400.000 references to personal names, and some 70.000 references to toponyms. This automatization saved us a huge amount of time, but on the other hand, the results need to be checked one by one since a computer can e.g. not automatically ascribe a village like Syron Kome to one of the five homonyms attested in Egypt. We also noticed that the computer for several reasons ‘missed’ some 10 to 15 % of the toponyms actually occurring in the papyri being searched, so the remainder needs to be added manually. For the moment we are about halfway checking the automatically imported toponyms, and we greatly appreciate any help in identifying the ‘anomalies’ that no doubt still occur in our database.
The same principle of automatic identification can of course be used for other full text corpora (like the Thesaurus Linguae Graecae for the Greek literary texts), but for texts written in Demotic, Hieroglyphic, Hieratic or Coptic such full text databases are only getting started.
Which are the plans to develop the database? Is there a target for its incrementation?
The Trismegistos database is constantly growing. Numerous partners world-wide help us gathering all the ancient texts that are preserved. It is difficult to put it into an percentage, but probably some 95 % of all published Greek, Latin, Demotic and Aramaic texts from Egypt have already been entered in the database, while also the Coptic and Arabic papyrological sources are quite complete. A lot of work is still to be done for the Hieroglyphic and Hieratic texts. For this last group of texts we are already cooperating with e.g. the Totenbuch-Projekt (TBP) of Bonn and the Deir el-Medina Database (DeM) of Leiden, importing part of their data in Trismegistos and providing each of the texts with a unique TM number. From the last two examples it is clear that we are gradually leaving the time frame of the Graeco-Roman period and expanding it to earlier periods.
Also geographically Trismegistos is expanding. Within a few years Trismegistos will have incorporated through EAGLE (Electronic Archive of Greek and Latin Epigraphy) the texts from several epigraphical databases, so that most Greek and Latin texts from the Mediterranean will have a TM number and can be searched through Trismegistos.
How do you consider the future of Trismegistos Places in particular?
With the geographical expansion of the Trismegistos database it will be necessary to drastically enhance the number of toponyms from outside of Egypt in order to provide every new text with the obligatory Geotex card, which gives information about the place where the text was found or written. But there are for the moment no concrete plans to gather all the toponyms mentioned in these new text. We would be eager to collaborate on ways of automating these tasks, but obviously this will depend mostly on new funding or new partners.
For the moment probably some 80 to 90 % of all the Greek, Latin and Demotic toponyms from Egypt have been entered into the database. We keep on checking the automatically imported toponyms and try to fill in all the remaining gaps for the older publications. At the same time, however, every new publication that comes to our attention, is entered into the database and the toponyms in these texts are immediately listed in Georef. We check the information in Trismegistos with recently published geographical corpora and studies, and try to incorporate as much as possible the new readings that sometimes are proposed for toponyms occurring in older publications.
Another item that will be improved is the online search structure for Trismegistos Places. The Trismegistos files are all made with the program Filemaker, which has great searching facilities. It is no problem, e.g., to search for all toponyms containing the element ‘polis’ in their name in Greek documents of the third century BC, or to search for all toponyms that are called a ‘dmi’ or village in the Upper-Egyptian Hermopolites nome in the Roman period, and to sort the results in whatever way you want. Once a week, the Filemaker database is uploaded into the online html structure, so the online data are usually well up to date. The online searching facilities, however, still need to be improved, although it will be virtually impossible to reach the same options as we have in Filemaker. If someone, therefore, has a specific search he cannot cope with online, it is always worth sending us an email and perhaps we can work it out in the Filemaker version.
Which is the place of the Trismegistos Online Publications (TOP) about toponymy in relation to the database? You are yourself the author of TOP 2, A Survey of Toponyms in Egypt in the Graeco-Roman Period published in 2008 and of TOP 5, Toponyms in Demotic and Abnormal Hieratic Texts from the 8th Century BC till the 5th Century AD, published in 2011. We can also point out TOP 4, written by Amin Benaissa, and entitled Rural Settlements of the Oxyrhynchite Nome. A Papyrological Survey, published in 2009 and updated in 2012.
Although databases are much more up to date than printed volumes and have much more searching and sorting options, scholars often prefer a printed version of these data, which is easier to refer to bibliographically and on the other hand can offer a broader ‘survey’ of the data not always possible in a database. When the end of a certain project is near, we consider the option whether it would be useful to provide an export version of the data in our TOP series. At the same time such a publication forces us to ‘clean up’ the data more thoroughly and to solve every last remaining problem :-).
We deliberately chose for online publications because we consider these exports as useful tools, which are, however, already obsolete the moment they are exported. Creating PDfs online instead of printed books makes it possible to give the information to the user a lot earlier (and cheaper), and it is easier for us to provide an updated edition at regular intervals. Which reminds me that I definitely should start working on an updated edition of The Survey of Toponyms in Egypt in the Graeco-Roman Period :-).
Also other people, such as Amin Benaissa and Klaas Worp, saw the benefit of publishing long lists as a PDF online, while the items in their list can be linked to the online Trismegistos database, so that we both profit from the cooperation.
Vous pouvez également suivre l’actualité de Trismegistos sur Facebook