LogosLink User's Manual · LogosLink version 2.0.0

Parse Argumentation Model Discourse

The first task during argumentation analysis usually involves segmenting the text into locutions. LogosLink Desktop partially automates discourse segmentation, as described below.

Follow these steps to parse the discourse for an argumentation model:

  1. Open LogosLink Desktop.
  2. Create a new argumentation model. Alternatively, load an existing argumentation model by opening the backstage and clicking on Open to select or browse for a file. LogosLink Desktop shows the argumentation model in a new window.
  3. If the argumentation model has no text content, you will need to add some text. Copy and paste some text to analyse into the Content pane.
  4. Make sure the text is in a language that LogosLink recognises. Click the button to have LogosLink Desktop determine the language of the text, or select it manually from the Language dropdown box.
  5. Click the Parse Discourse button in the Home ribbon tab. The Parse Discourse dialog box will appear.
  6. First, select the text to parse in the top text box. You can select the whole text, or just a fragment.
  7. Second, define the text format by inserting tags and free text in the Format text box. This tells LogosLink Desktop how to parse the text and how to extract speaker and timestamp information, if present. You can type directly into the text box, or you can use the button to the right to display a pop-up menu. Check the examples below for details.
  8. Validate that the text format specification is OK by clicking the button to the right. Amend errors as needed.
  9. Finally, select the parse options as you wish. Usually, creating transitions automatically is safe, especially in a monological text or a dialogical one with speaker information. Creating propositions and assertings automatically is a bit more uncertain.
  10. Click OK to parse the text and generate the corresponding argumentation model elements.
Examples

Consider the following text: 

    Alice: I am happy today because it's sunny.
    Bob: It's not really sunny.
    Alice: Well, I'm happy anyway.

Here, each paragraph contains the speaker name followed by a semicolon and then the content text. Consequently, the format that you should use would be {Speaker}:{Content}.

Consider this other text: 

    [01:16] Alice: I am happy today because it's sunny.
    [01:18] Bob: It's not really sunny.
    [01:19] Alice: Well, I'm happy anyway.

Here, in addition to the speaker name and the content, there is a timestamp at the start of each line, enclosed in brackets. The format that you should use to parse this text would be [{StartTime}]{Speaker}:{Content}.

Contents distributed under a Creative Commons Attribution 4.0 International License · About · Terms of Use · Contact Us · last updated on 19/02/2025 10:51