Beyond the collection of data:
A corpus of translations and its interactive use in research on cultural dynamics

José Lambert (CETRA, KU Leuven),
Lieven D'hulst (HIVT, Antwerp),
Katrin Van Bragt (KU Leuven)
& Ludo Meyvis (KU Leuven)

Bibliographies were supposed to be (rather simple) passive instruments for consultation only: nothing else than the collection of data. They are rated as such in the evaluation of scholarly work by the relevant comittees within universities and on the national level. This assumption may be confirmed by a first glance at Katrin Van Bragt's Bibliographie des traductions francaises (1810-1840) published by Presses Universitaires de Louvain, 1995. With more than 12.000 items, each with an explicit apparatus of descriptive categories, it is indeed a traditionally looking, heavy and impressive book. The innovation is that the book publication is only the (static) top of the iceberg: completed and put into action by a CD-ROM, which exhibits statistics and graphics about correlations, evolution, tendencies, it is transformed into an instrument for interdisciplinary research and for the formulation of questions and hypotheses. With the aid of the book only, questions asked by the human researcher could never be tested out in the same direct, fast and easy way; many among them would not even be formulated with the same kind of systematicity.

The demonstration will show the use of the CD-ROM in more detail for some research questions to be solved. In this description, we want to first explain why and under what conditions a bibliographical repertoire was transformed into a key towards cultural dynamics. Started in the 80s as a manual approach to a historical corpus, the project was gradually made much easier and more efficient due to the development of new programs in the 90s. It gradually moved from a computer supported to a computer stimulated approach, which in itself is an interesting fact on the relation between computer science and the humanities.

The very possibility of the problem solving orientation was embedded in the research questions from the beginning, since they were focusing on relationships, on the basis of systematic and representative repertoires. The parameters for the selection of the repertoire, for the codification of the items (authors, titles, format, publishing house, series, subtitles, language, historical origin, genre, etc.) needed to be worked out and tested out in terms of hypotheses rather than on the basis of a closed set of questions. They were systematically linked with hypotheses and positions provided by the historical setting itself. Only when reflecting the complexity and the fluctuations of a given period and its cultural production can such a repertoire offer insights into dominant and/or peripheral or superficial trends. The use of systems theories was of great help in this construction, but it had to be adapted to the exact nature of the material and the moment. Under many aspects this analysis seems parallel to corpus linguistics: the difference is that, in this case, the systemic approach applies to a corpus of bibliographical data taken also as a historical corpus and considered to have several (kinds of) systematicities.

Among the questions and hypotheses that can be examined and further refined in a statistical and/or graphic way, let us mention: What are the dominant source languages (at each moment), what kind of genres are fashionable (in terms of book publications), who are the most important authors, translators, publishing houses (and book series), where are they published, in what kind of a format, are there any spectacular shifts during this period, and is true that 1830 is a (cultural) turning point in France? Is it relevant to assume that peripheral genres are published in a specific format, and that more traditional publications tend to be distributed in the French provinces rather than at Paris?

The importance of the questions and hypotheses goes far beyond the research project itself since basic trends in particular disciplines (say theology, political studies) can be tested out on the basis of title patterns and their underlying principles.

Obvious possibilities for a direct transfer of these tools can be envisaged in any kind of cultural production where the observation focuses on the relationships between "production", "reproduction" (or tradition) (both from "within the system", i.e. a synchronic point of view) and "import". The very struggle between these layers in a given society (literature, country, genre, etc.) can easily be observed and analysed with the aid of these bibliographical keys. After a discussion of the internal organization of the Van Bragt bibliography, demonstrations with corpora from different periods and cultures will be used in order to illustrate how a better use of bibliographical corpora may improve our cultural research in a spectacular way.