Opening Speech of the Conference on Early Dictionary Databases (Toronto, October 1993) [1993, rptd. 1996, 2008]

How to Cite

Quemada, B. (1996). Opening Speech of the Conference on Early Dictionary Databases (Toronto, October 1993) [1993, rptd. 1996, 2008]. Digital Studies/le Champ Numérique, (4). DOI:


Download HTML





Dear Colleagues, Dear Friends,

I was very touched by the honour of being asked to give the opening speech on the first day of the conference. I won't say that this second conference is welcome -- that goes without saying. I will say it was necessary and had to be held here.

Here, because favourable circumstances have forged, in this privileged place for exchanges between anglophones and francophones, a particularly effective alliance of philological, lexicographic and computing competences.

Necessary, because it is indispensable at some point to undertake a review of our projects. The opportunity to exchange experiences, compare methods and results, and learn about other practices, is invaluable, for it also makes possible new forms of trans-national and cross-disciplinary cooperation.

I should like, if I may, to call on my own experience, which in part is bound up with what we can today call the history of New Learning, in order to make a few recommendations.

The first, the fruit of long observation, concerns the decisive role of exchanges between specialists from related disciplines, whether they have worked on the same, or different, materials. The links established, often formerly through computing centres, between researchers from different disciplines have made apparent more common points than was hitherto imagined. It is this innovative synergy that has allowed the most decisive steps to be taken, and in particular has inspired new approaches and made possible the devising of new solutions to problems that otherwise would have continued to be formulated in traditional terms and would have remained mired in the twists and turns of repetition. I greatly appreciated the contributions of the first Toronto conference, which gave indication of initiatives engaged in this direction. I am pleased to see that the programme of this conference will develop along the same lines.

No-one here doubts that it is a good choice to work on early dictionaries. It is most certainly amongst the most useful of those that can be made today by philologists, historians or lexicographers, given the diverse riches they contain and the extent of the applications these riches allow. Your work, more than any other, has already shown the importance of these documents both for linguistics and for cultural history.

It has also shown that it is not just a matter of adapting these texts for electronic processing. Early dictionaries are a source of usable data only if critical studies have first established their identity and worth. On this condition alone can they provide valid, reliable data. This may be a considerable constraint, but it is the price one must pay in order to have their multiple riches at one's disposal. In France, in the first half of this century, Ferdinand Brunot, Mario Roques, Lucien Febvre had glimpsed the potential, but were unable to realize it. We must use our present means to help us activate the philological study of dictionaries and answer the questions such study raises. Would each programme not profit, in the short term, from the inclusion of a scholarly re-publication? Should this latter be paleographic or a post-analytical restructuring? On what medium should it be published? etc. The answers will doubtless be different for each project.

Although in its infancy, our knowledge of early dictionaries has largely benefitted from contemporary metalexicography. This has certain consequences, such as to call into question cumulative studies that would result in the association of data taken from works whose objectives and typological bases are distinct, even incompatible, unless the corresponding precautions have been taken into account. Do we however have at our disposal all the corrective parameters?

Early dictionaries belong to the category of texts that are the most difficult to analyze "automatically". The reasons for this are the mass and variety of the data to be processed, compounded by their complex, heterogeneous and often muddled organization. If this explains in part the relative lack of progress compared to text processing as a whole, it is just a question of specific complications that will gradually be resolved. What are the most important levels of analysis? What types of product correspond to the most basic requirements? Can immediate objectives be defined to meet the widest needs? etc. These are some of the questions open for debate.

Despite the possibilities offered by advanced technology, experience teaches one to beware of quasi-ideal, global solutions that more often than not turn out in the end to be quite disappointing. Constant vigilance, allied to a realistic and reasoned optimism, has always proved more profitable.

The correlation of investment and return has obviously to be taken into account. Yet economies of time, like those of money, should not be made at the expense of quality. Scholarly requirements must stand up to temporary computer limitations.

While it is important to optimize our programs and products according to the current functionality of computerization (hardware, software and practical knowledge), one should also keep in mind that the most sophisticated possibilities are often more suited to the final stages than to the initial ones. The important thing is to devise open formulas that allow for cumulative developments (for example, harmonizing the static and the dynamic versions of a documentary database, or functionally linking the advantages of image and text, etc.).

The intellectual input required by new approaches is often as costly as material investment, which is not a reason for giving up or postponing, but rather an incentive to seek the apportioning of abilities, the pooling of resources and knowledge, and especially the planning of the sharing of expected results. Cooperation, encouraging the normalization of data and methods, is not a personal hobby-horse that I persist in putting into practice and preaching from conference platforms -- it is an objective necessity, as time has shown more and more clearly. When related or similar ventures propose effective norms that have already been tested internationally, we must quickly examine their adaptability to our own needs and their general applicability. There is strength here through unity.

We can also take advantage of automatic language-processing research (in the language industries) that applies particular norms and techniques to modern dictionaries through the development of dictionary databases, computerization, retroconversion, dictionary recycling, etc. These methods and tools may be of use to us, even for the analysis of early text corpora. The same is also true for different forms of association and collaboration that many of our projects should establish, without delay, with certain large-scale textual computerization and early corpus digitization programmes, such as those of the National Library of France (scholars' work-stations integrating hypertext and multimedia, and in the foreseeable future permitting the reading-consultation of early dictionaries).

It is time to give the floor to the other speakers. I believe I have remained true to my role in limiting myself to the obvious, conscious thereby of the French dictum that "that which goes without saying is the better for being said". It's done. You will assert once again the validity of the issues of the young confraternity of scholars you represent so admirably. Your written and spoken words leave no-one in doubt that it is ambitious, demanding, efficient, productive and welcoming. It is up to you to complete the paradigm -- I have said my last word.



Bernard Quemada (Trésor de la langue Française, CNRS, Paris)





Creative Commons Attribution 4.0


File Checksums (MD5)

  • HTML: e55303fa5190079083dc171cdc1fac2e