3.3.4. Establishing a Monolingual Base List

To create a monolingual base list, you must scan for terms all the texts that provide information about one or more of the concepts in the concept system under study. This means that you will highlight terminology units and note their contexts (sentences, paragraphs).

When term extraction is performed in more than one source dealing with the same topic, the lists resulting from the sources scanned can be merged so that more information is readily available for conceptual analysis and so the best textual supports can be selected for entry on the terminology record about a given concept.

In order to note authentic usage, it is recommended that original-language sources in the source and target languages be scanned for terms first, before any translated sources are scanned. You may be able to begin matching (apparently) equivalent terms at this stage, but conceptual analysis and further research are required to confirm your suppositions.

This list shows the bilingual matching of extracted terms.

The term extraction lists often include terms that belong to other subject fields or that designate concepts missing in your original diagram of the concept system. In addition to terminology units, automatic term-extraction software outputs a great deal of "noise" (that is, pseudo-terminological expressions or components that accidentally occur together in discourse but that do not designate concepts). By briefly reviewing the term's contexts, you can eliminate noise, set aside terms that belong to other subject fields, and create a more complete graphical representation of the concept system through the insertion of missing concepts into the concept diagram.

The terminological base list is the list of terms created through term extraction. It contains all of the terms that you will be assigning to "nodes" of the concept diagram in order to group the textual supports by concept. In comparative terminology, the concept system is used to establish a terminological base list for each of the languages and is the main benchmark for matching base lists. A base list is usually subject to further terminological research.

Sometimes, your assignment may be to collect the terminology used in the bilingual documentation of a department or company. In this case, you can carry out bilingual term extraction in order to concurrently identify the terms and their contexts in both the source and target languages. In some terminology services, the translated equivalents are checked for authenticity by comparing them with the terms identified during term extraction in original-language texts. Such verification is not always possible in translation services where terminology files store source language terms and their target language equivalents without their textual supports or source references. Authentication usually involves further research work.

term extraction not only reveals terminology units, but also co-occurrents of the terms. Collocations (two or more terms that typically appear together in a given subject field) show how the term is used in specialized discourse. This information is often recorded in the phraseologisms component of the record.

You can also use the terms that you identified during term extraction to perform a diagnosis of a database. The list of terms can be queried in order to determine the response rate of terms found in the database or file that is to be evaluated.

This is a base list retained following verification of the text on which automatic term extraction was performed.