Episode 4: Blaise Pascal's Pensées
Blaise Pascal's "Pensées" (which translates to "Thoughts" in English) is a collection of fragments on theology and philosophy. Pascal, a French mathematician, physicist, and religious philosopher, began writing "Pensées" as a defense of the Christian religion, but he died before he could complete the work. The fragments he left behind were posthumously assembled and published in 1670.
This Philosophical Minute centers around a passage from Pensées, which delves into the human propensity to neglect the present moment, habitually yearning for the future, or dwelling on the past.
We never care about the present. We anticipate the future as too slow to come, as if to hasten its course; or we recall the past to stop it as too quick: so careless, we wander in times that are not ours, and do not think of the only one that belongs to us; and so vain, we think of those that are nothing anymore, and let slip without reflection the only one that remains.
It is because the present, usually, hurts us. We hide it from our sight, because it afflicts us; and if it is pleasant to us, we regret seeing it slip away.
The present is never our end: the past and the present are our means; the only future is our end. Thus we never live, but we hope to live; and, always preparing to be happy, it is inevitable that we never are.
Workflow for Creating a Network of Keywords
Create a new node: Start by generating a new node named "Blaise Pascal - Pensées". This node will hold the text that you plan to analyze.
Insert the text: Add the selected excerpt into the comment section of the "Blaise Pascal - Pensées" node.
Run the Dimension Elicitor, set the General Context to "Philosophy", and input "Keywords" as the keyword for the analysis of the node comment.
Assess the extracted dimensions: Evaluate the keywords or dimensions identified by Hellixia and eliminate any that are redundant or irrelevant.
Use the Embedding Generator for all remaining nodes. This tool will distill the semantics of the names and comments of each node into a quantifiable form.
Set "Blaise Pascal - Pensées" as the Target Node.
Run the Naive Learning algorithm.
Change the style of all nodes to "Badges". This style will display the comment embedded within each node.
Switch to Validation Mode.
Perform an Arc Force analysis.
While within the Arc Force analysis tool, run the Radial Layout. This will arrange the nodes in a clockwise pattern in relation to their connection strength with the target node.
Show the Arc Comments, which will provide information about the strength of the relationships between nodes.
Workflow for Creating a Semantic Network
Start by making a copy of the node named "Blaise Pascal - Pensées".
Open a new graph and paste the copied "Blaise Pascal - Pensées" node.
Use the following keywords to guide the Dimension Elicitor in its analysis of the node: Arguments, Matters, Milestones, Rules, Themes, Theses, Topics, and the General Context set to "Philosophy".
Inspect the dimensions suggested by Hellixia. Any dimensions that are irrelevant or redundant should be removed from your analysis.
Exclude the "Blaise Pascal - Pensées" node.
Use the Embedding Generator on all remaining nodes.
Run the Maximum Weight Spanning Tree algorithm to create a semantic network based on the text analysis.
Change the style of all nodes to "Badges". This will display the comment within each node.
Run the Dynamic Grid Layout to organize the nodes on your graph. Note that this algorithm's output is not deterministic; it may favor vertical, horizontal, or mixed orientations. Execute this layout multiple times until you find the most suitable arrangement.
Switch to Validation Mode.
As the graph you are building does not represent causal relationships, opt for the Skeleton View. This will remove all arc directions, leaving only the node connections without any specified direction.
Workflow for Node Force Analysis
Switch back to Modeling Mode.
Change all node styles to Discs.
Use the Symmetric Layout to organize your nodes in the graph.
Go to Validation Mode.
Conduct a Node Force analysis to evaluate the strength of associations in your graph.
Workflow for Creating the Hierarchical Semantic Network
Execute Variable Clustering: This operation will categorize analogous variables based on their semantic relationships.
Open the Class Editor.
Run Class Description Generator: Use this function to generate descriptive names for your identified factors. This helps to make the output more understandable and interpretable.
Save these descriptions by using the Export Descriptions function.
Switch back to Modeling Mode.
Run Multiple Clustering.
Run the Taboo algorithm: Use this structural learning algorithm to learn a hierarchical network. Make sure to enable the "Delete Unfixed Arcs" option to remove unnecessary connections and streamline your model.
Use the descriptions you exported earlier as a dictionary to rename the latent variables you've just created. This helps in making your model more understandable and keeps the nodes' names consistent with their semantic meaning.
Switch to Validation Mode.
Apply Node Force.
Last updated
