TOTh Workshop

Today, innovation in terminology is almost exclusively led by players who are directly involved in professional practice, such as industrial or national bodies producing multilingual terminology which is mostly geared towards translation, interlingual communication and technical writing. Terminology is created for practical purposes.

This leads us to adopt a functional standpoint. We therefore invite all interested parties to a one-day workshop to discuss the prospects of integrating and innovating this specific scenario with what has been done so far.

Context and Notes are no longer considered exclusive features of the term entry, but they are framed within an overall usage dimension according to a series of relevant profiles. The practical use of these two fields is essential to establish quality parameters and standards. How can one decide whether an authentic context is preferable to an artificial one without taking its use into account?

Hence the need to design and innovate these two fields pursuing an enlightened use of terminology. Overcoming generic as well as dogmatic approaches, the discussion can focus on two separate issues:

• the use of the context field
• the use of the notes field

Context Field

To identify and define an optimal context one needs to make concrete assumptions about its real use. The context can be identified and defined by means of reliable and representative corpora as well as in connection with computer-assisted translation tools and machine translation systems (in pre-editing and post-editing). But it can also be provided to support the production or use of documents by non-native speakers of a language.

The terminology lookup tool must therefore almost invariably co-exist and interact with other tools.

In the case of corpora, a properly designed context field would allow the user/translator to reduce the time spent searching very large corpora while translating. The terminologist would thus anticipate some of the corpus analysis work and include the relevant information in the term entry. As a result of this, the translator could use the entries to support his/her translation (by plugging them into computer-assisted translation tools or machine translation systems) and have access to information that has been pre-selected on the basis of his/her needs.

Still within a multilingual terminology framework, one can take advantage of parallel and comparable corpora. With the former, the concordances show aligned source and target texts, so that identifying equivalent terms and contexts is quicker. With the latter, on the other hand, the user formulates equivalence hypotheses, identifies relevant terms and finally searches the corpora to verify collocational profiles and inspects concordances to infer contexts of use. This is made possible by the rational use of corpora, as part of the overall workflow leading to the creation of multilingual term entries.

It is possible to extract contexts in language A and in language B from corpora. But should the field necessarily contain only one example? If multiple examples are to be included, what are the relevant and most useful selection criteria by which they can be chosen? In addition, should the examples be relevant in a translation-oriented perspective, or should they just help to understand the meaning? In either case the question arises as to what relationship should hold between the contextual examples provided for language A and those for language B. The examples chosen for language B are necessarily defined as a result of the type of equivalence that has been established, insofar as they are extracted from parallel or comparable corpora. The context field could also include links to concordances or collocations identified in the corpora, along with fields certifying their trustworthiness (Occurrence).

Another point to bear in mind is that the context contains information that should not be contradicted by the definition; as a result, when using corpora it is preferable to examine and identify the possible contexts first, and then create the definition. This bottom-up approach therefore takes as its starting point the search for contexts as a basis to create definitions that are compatible with the context itself. It is thus also possible to carry out a parallel comparison with already existing definitions provided by domain experts.

As far as machine translation systems are concerned, the use of the Context in pre- and post-editing is of paramount importance. Having access to a Context that has been identified by means of a corpus considerably facilitates the preparation of a text to be fed into a machine translation system (pre-editing) or the revision of raw machine-translated output (post-editing).

Notes Field

Against this background, the need arises to reconsider the official instructions or the ISO norms concerning the use of the Notes field to assess its current effectiveness. Perhaps the level of explanation that can be achieved in the Context might make the content of the Notes field less diverse?

As is well known, the Notes field can contain remarks of various categories: linguistic, stylistic, encyclopaedic, idiomatic, operational, political, etc. But the addition of further information can be managed properly only if one knows the future use of the entry in question. As a result, this field can be filled sensibly only on the basis of an awareness of how the term entry is going to be used.

More generally, the Notes field contains all the information which is necessary for a pre-determined user, but which cannot be provided or identified in one of the other fields. As the other fields (especially Definition and Context) are enriched with more informative and detailed information thanks to the integration with other tools, the content to be inserted in the Notes field becomes more accurate and specific. It is however very likely that mostly non-linguistic information will end up there, with considerable variation resulting from institutional, corporate or professional priorities. However, similarly to the role of the Context field, the Notes field could be useful to select relevant information from the results offered by search engines for a given term, thus providing an effective and reliable summary.

Languages of the Workshop

Italian, French, English