Bergen, Sep. 28, 1998.
Participants: Bill Black, Gerrit Bloothooft (chair), Paul Dekker, Wim
van Dommelen, Anders Erikson, Klaus Fellbaum, Knut Hofland, Paul McKevitt,
Julia Lavid, Tony McEnery, Joakim Nivre, Torbjørn Nordgård,
Guy De Pauw, Koenraad de Smedt, Michael McTear, Andy Way.
The workshop was organized at the conference The Future of the Humanities in the Digital Age and was planned in conjunction with the TNP on Speech Communication Sciences. The aims were:
After a welcome by the workshop proposers (Gerrit Bloothooft and Koenraad de Smedt), the participants present themselves.
The TNP on Speech Communication Sciences has published two books, the first one being Analysis, the second one Proposals. They are preparing a third volume, Recommendations. Gerrit Bloothooft comments on the experiences in writing these books and the prospect for a publication in the area of computational linguistics. ACO*HUM envisages such a publication in the form of a book chapter.
The TNP on Speech Communication Sciences has good coverage for speech, ACO*HUM has less good coverage for computational linguistics. A number of members in the TNP on Speech Communication Sciences are also active in computational linguistics and therefore, both networks together should reach good coverage. Geographical coverage of activities must be considered, if possible also including Central and Eastern Europe, and maybe wordwide activities. The inclusion of other typological language groups is relevant. Coverage of countries is not equal to coverage of languages.
Based on experiences from Speech Communication Sciences, the writing of a book may take about nine months in all. People have to be made responsible for country overviews. A cooperation with the ACL and EACL could be considered; they have their old surveys of courses and programs but these are not very thorough.
Andy Way gives an overview on Erasmus activities including his report on Curriculum development in the area of computational linguistics in a European framework (postscript file). The report is not so much an effort towards standardization, but a comparison necessary to establish if a student has the necessary prerequisites in mobility schemes. Also, sometimes it can be beneficial for exchange students to take a course in a different country similar to one they have already taken, but it may be difficult to get credit for it. ECTS should be implemented wherever possible.
The proposal of a European masters degree in language and speech is discussed. Advantages of such a masters degree would be:
The question of accreditation was put forward. Degrees are normally awarded by institutions, but a masterd degree is not usually awarded in many European countries. Degrees are nationally recognized and often approved by professional bodies which promote their content and legal status. However, there are no true professional bodies in natural language and speech processing, although there are learned societies such as the EACL and ESCA. At NTNU in Trondheim, a 2-year masters degree in linguistics is offered with a content corresponding to the national degrees, but all courses taught in English. At Aalborg University, a 1.5 year masters degree is offered in Intelligent Multimedia, also taught in English. The international or European character of a degree may have several meanings: either it indicates it is for foreign students and usually taught in English, or it implies a cooperation leading to corresponding degrees in different countries.
The shape of study programs was discussed. The current proposal by Gerrit Bloothooft operates with a proposal for a one-year intensive program leading to a masters certificate. This year includes a compulsory visit abroad of at least 3 months. It might, however, be easier to let students study abroad for a whole year. An alternative would therefore be to let students spend a year abroad studying speech processing to supplement a previous background of about 3 years of natural language processing, or vice versa. The final year is likely to be attractive for students wishing to study abroad but the prefinal year may be preferable for several reasons. Math requirements might be a difficult barrier for computational linguistics students wanting to take speech components.
It is agreed that a continued cooperation between the ACO*HUM TNP and the TNP on Speech Communication Sciences, and between the two communities in general, is desirable for the following purposes:
It is agreed that there will be joint presence at a meeting in January 1999, to be planned by the TNP on Speech Communication Science (which is also planning a workshop in London in April 1999).
Report written by Koenraad de Smedt, October 26, 1998.