1 / 24
Scroll
Rotate your phone for a better view
Navigate Esc Close
Click to view sentences

NLP-POL · Political Discourse Analysis

François Legault Resigns

NLP analysis of Quebec's Prime Minister resignation speech

January 14, 2026

Antoine Lemor | NLP-POL PROJECT ↗
Swipe for the interactive analysis

What is NLP-POL?

An open source project applying NLP to political discourse analysis

Goal

Demonstrate the possibilities of NLP methods applied to political discourse study

Tools

Transcribe-tool ↗ + LLM_Tool ↗

Frequency

Regular analyses published on my blog ↗

Methodology

Methodology

Sentence-by-sentence annotation with LLM

122
sentences
10
dimensions
154
actors
105
justifications
View analysis

Interactive Dashboard

Click on an indicator to navigate to the corresponding analysis

Key finding
A political legacy speech more than a resignation, centered on party politics: pride in achievements, defense of CAQ record, gratitude to teams
Key finding
72%
of the speech dedicated to political legacy
It's not nothing, building a party and then governing for 6 and a half years.
François Legault explains less than he defends. Party before public policy.
Discover the analysis

Speech at a glance

The speech at a glance

View chart
Sentence-by-Sentence Emotional Evolution
Positive
Neutral
Negative
Hover Hover over the chart to see each sentence with its tone, speech act and theme

Analysis #1

Which policy themes dominate?

View results
Policy themes
24
mentions of party politics
I want to thank those who embarked on this great adventure of creating a new party with me.
Party before public policy.
Explore the data

Explore policy themes

Click a sector to discover associated sentences

Try it Click a sunbeam to reveal all sentences
Policy themes
122 sentences
Click a ray
Select a theme
All sentences
Pick a sunbeam to explore the classified sentences.

Analysis #2

What is the balance of justifications?

View results
Justification balance
86%
of justifications defending the record
We worked hard, we did many beautiful things together.
François Legault explains less than he defends. Party before public policy.
Explore the data

Justification Strategies

How Legault justifies his decision and record

Does he focus on resignation reasons or his record?
0
0
Resignation Record

Analysis #3

How are actors portrayed?

View results
Actor sentiment
+0.94
actor sentiment index
Thank you to so many people. I want to thank Brigitte, Stéphane, Claude, Lynn, Nicole, Pascal, Charles, Mario...
Little sadness or regret expressed. A speech of thanks rather than political analysis or contrition.
Explore the data

Sentiment towards actors

Click on an actor to discover the sentences where they are mentioned

Sentiment index
0.00
Loading...
Positive
Neutral
Negative

Analysis #4

What place for Quebec identity?

View results
Quebec identity
16%
of the speech on identity
We must also stop the decline of French. Particularly in Montreal, the future of our nation is at stake.
Identity is present but secondary to party politics.

Identity Themes

Click on a theme or stance to discover the associated sentences

Identity proportion
0%
Loading...
Themes
Stances
Quotes
Analysis #5

What rhetorical strategy?

Dominant speech act
51
justifications (26.4%)
"I tried to reduce the wealth gap between Quebec and the rest of Canada."
Justifying (51), claiming (35), praising (22). Defending his record is the speech's primary function.
Explore the data

Rhetorical Strategy

Click on a speech act or political legacy framing to discover the associated sentences

Dominant act
Loading...
Speech Acts
Political Legacy Framing
Quotes
Conclusion

A political legacy speech

Quantitative analysis shows a political legacy speech rather than a resignation. François Legault defends his heritage and political partners, and above all protects the image and political legacy of the CAQ. A departure explanation tied to the party whose theme is dominant, little self-criticism or analysis of failures.

+0.33
Tone
24
Party politics
72%
Pol. Legacy
+0.94
Actors
16%
Nationalism

Resources

💻
nlp-pol
Code & data ↗
🤖
LLM_Tool
Annotation ↗
📝
Blog
More analyses ↗

Antoine Lemor · Everything is open source (and open to criticism). That's science.