Section: Other Grants and Activities

European Initiatives

EuropeAID Project: For Archaeology of Ancient Asian Texts (AAT)

Participants : Marc Csernel, Sergiu Chelcea, Marina Dufresne, Yves Lechevallier, Sattisvar Tandabany, Brigitte Trousse.

This year we started our project called ``AAT'' in the context of the EuropAid (DG1) projects and more precisely of the Asia Information Technology (I.T. Asia). We collaborated mainly with François Patte (UFR Maths-Informatique, UNiv Parie 5 René Descartes) and Pascale Haag (EHESS, Centre d'études de l'Inde et de l'Asie du Sud, Paris).

The objective of the AAT

Ancient texts, whether religious, scientific or philosophic are known to us due to the patient and vigilant work of scribes who, from centuries to centuries, have copied and copied again successive versions of an original text (usually lost for ever).

So there is a chain of copies starting with the original text and continued by an immense tree of hundreds of copies that has grown more or less like a genealogical tree. They are never identical to each other, sometimes extremely different. Parts of the original are missing, fragments are not readable anymore, some have been miscopied, and some others have been voluntarily transformed. This is particularly true for the large Indian subcontinent where at least one third of the manuscript existing through the whole world are supposed to exist, mostly unpreserved , unreferenced , and being at mercy of any accidental event. Even during the 20th century manuscripts were copied by hand by armies of scholars.

Still a question remains unsolved as to how to compare hundreds of different copies of a same original ancient text, and to decide which fragments are original and which ones are not in order to re-build the original document.

Specific software has recently been designed for Latin and Greek scripts which open new avenues to study ancient texts from Roman and Hellenistic periods. It is the aim of the present project to design a most advanced IT tool for ``archaeology of ancient Asian texts''. Such IT Tool will be based strictly on open source.

Contributions to program

This project involves Axis as the applicant of the project and three others partners: University ``La Sapienza'' in Rome (Facoltà di Studi Orientali), the Bhandarkar Institute of Oriental Studies (BORI) in Poona (India) and the Mahendra Sanskrit University de Kathmandu (Népal).

Our three partners will dedicate their force to the collection of manuscripts of a famous Indian grammatical text: The Kâçikâvritti or ``Benares glosses''. This text is the oldest comment (around the 7th century) of the Panini grammar, the world oldest example of generative grammar. It is well known trough hundreds of manuscripts disseminated all around the Indian subcontinent. These manuscripts are dated from the 12th century to the beginning of the 20th century. They are supposed to display the representation of the same text, but because of the time, their completeness is only partially assumed, and they can differ from each other.

Axis is providing the necessary software to reach two different goals which can be completed only one after the other:

• Providing the software tools necessary to help the creation of critical edition of the Sanskrit texts. As a secondary result, a distance between the texts should be established based on the presence/absence of the different words in each manuscript (cf. sections  6.2.2 , 6.2.7 ).

• Using the distance established by the first software sets, trying to establish which are the different cluster set of manuscript (for example via a 2-3 HAC clustering), try to establish more or less a phylogeny of the different manuscripts

One could wonder what is the need for a specific project to compare different Sanskrit texts, as tools such as the famous Unix DIFF exist since a long time. The response is given by some of the Sanskrit writing specificities:

• Sanskrit is written according to a 48 letter alphabet, but, on computer, is written using Latin alphabet using a transliteration such as the Velthuis one.

• Sanskrit is written without blank and the blanks are not very significant

• When two words are written without blank separation, the spelling becomes different, it is the sandhi problem.

Three internships were carried out on this project: S. Tandabany (cf. section  6.2.2 ), M. Dufresne and S. Kebbache (cf. section  6.2.7 ).

ERCIM

B. Trousse presented AxIS researches at the Kickoff meeting of the ERCIM Working group on ``Data and Information Mining'' organised by Christoph Schommer on january 14th at the Campus Kirchberg (University of Luxembourg).

Other Collaborations

• Portugal: B. Trousse collaborated with F. Amilcar Cardoso (amilcar@dei.uc.pt) in the context of the COST Action 282 (2001-2005): ``Knowledge Exploration in Science and Technology''.

  URL: http://www.mpa-garching.mpg.de/~opmolsrv/COST282/ .

• Italy, University of Napoli II (Profs C. Lauro and R. Verde) : Y. Lechevallier, A. El Golli, D. Tanasa and B. Trousse

• Italy, KDD Lab. Instituto ISTI, (Fosca Gianotti): some potential collaborations were identified after the ERCIM kickoff meeting.

• Belgium, Facultés Universitaires Notre-Dame de la Paix à Namur (Profs A. Hardy, M. Noirhomme and J.-P. Rasson) [78] : Y.  Lechevallier.

• Belgium, Université Catholique de Louvain, DICE Laboratory (Prof. Michel Verleysen, Prof. Vincent Wertz, Dr. Amaury Lendasse, Damien François): F. Rossi was invited professor for one month in 2005 [81] , [80] .