|
Traitement de données langagières dynamiques par les outils et méthodes du TAL| title | Traitement de données langagières dynamiques par les outils et méthodes du TAL |
|---|
| start_date | 2025/06/30 |
|---|
| schedule | 9h10-12h30 |
|---|
| online | no |
|---|
| location_info | room 3-001 |
|---|
| summary | To produce written text can be a daunting task, posing a challenge not only for high school students or second language learners, but also for most of us—including scientists and PhD students writing in their mother tongue. Authors have to perform several tasks (planning, drafting, editing, revising) often simultaneously and under severe space and time constraints. Next to these constraints, there are other reasons making writing difficult
(i) lack of */meta-knowledge/* and */control/*: without knowing what to do and when, authors may allocate too much time for the wrong task (drafting before analyzing the audience and the topic);
(ii) */memory problems/*. losing sight of the initial goal, author_may_ end up downsliding, shifting focus on the wrong goal;
(iii) /lack of *knowledge* and *understanding*/. Writing requires not only linguistic knowledge but also knowledge of the subject matter, the world, society, etc. This is why experienced writers spend more time thinking than writing. They use their brain before using the pen.
(iv) lack of *strategic /knowledge/*. While many authors recognize the importance of planning or outlining, they often do not know how to do so, or where to put best what information.
(v) /lack of a *clear idea*/: Many writers believe they have a clear idea at the outset, yet their written output often reveals otherwise.
All these factors may to cause stress and lead to poor results.
In my talk, I will propose a potential solution to this complex problem, namely, an /ecosystem/ that breaks the whole process into a set of distinct tasks (or 'windows'), allowing users to focus on one problem at a time. I will also explain the rationale behind these windows, their relationships, and their relevance in facilitating the writing process.
Finally, I will present my reflections on the notion of /topicality/—its definition, its operationalization, and its central role in text production. I propose the construction of a /topic dictionary/, in which each entry (i.e., topic) includes its typical subtopics and the key words that define it. If this intuitionproves valid, the result will be a set of trees (or a topic tree) that can be used either in a top-down fashion (for idea expansion) or in a bottom-up fashion (for integrating ideas or topics into more general ones). To produce a fully developed outline, it will still be necessary to determine the sequence of topics and the relationships between them. |
|---|
| responsibles | NC |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2025/06/23 08:25 UTC |
| |
|