Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and...
Transcript of Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and...
![Page 1: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/1.jpg)
Logic, Interaction andCollective Agency
Lecture 1
ESSLLI’10, Copenhagen
Eric Pacuit Olivier Roy
TiLPS, Tilburg University University of Groningenai.stanford.edu/~epacuit philos.rug.nl/~olivier
August 16, 2010
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 1
![Page 2: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/2.jpg)
Writing a paper together
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 3: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/3.jpg)
Writing a paper together
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 4: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/4.jpg)
Writing a paper together
Problem of Coordination.
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 5: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/5.jpg)
Writing a paper together
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 6: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/6.jpg)
Writing a paper together
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 7: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/7.jpg)
Writing a paper together
Problem of Cooperation.
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
Intuitively, we solve these problem by working together.This is the question of collective agency.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 2
![Page 8: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/8.jpg)
Plan for the week
1. The problem of collective agency:
• Individual and group agency.• Games.• Beliefs (Type Spaces) and rationality.
2. Group attitudes (I): a non-standard introduction.
3. Acting on team preferences, frames and team reasoning.
4. Group attitudes (II): Correlations.
5. Commitments, intentions and cooperative agency?
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 3
![Page 9: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/9.jpg)
Plan for the week
1. The problem of collective agency:
• Individual and group agency.• Games.• Beliefs (Type Spaces) and rationality.
2. Group attitudes (I): a non-standard introduction.
3. Acting on team preferences, frames and team reasoning.
4. Group attitudes (II): Correlations.
5. Commitments, intentions and cooperative agency?
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 3
![Page 10: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/10.jpg)
Individual vs collective agency
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 4
![Page 11: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/11.jpg)
Individual vs collective agency
Different contexts of agency
I Individual decision making and individual action againstnature.
• Ex: Gambling.
I Individual decision making in interaction.
• Ex: Playing chess.
I Collective decision making.
• Ex: Carrying the piano.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 5
![Page 12: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/12.jpg)
Individual vs collective agency
Different contexts of agencyI Individual decision making and individual action against
nature.• Ex: Gambling.
I Individual decision making in interaction.
• Ex: Playing chess.
I Collective decision making.
• Ex: Carrying the piano.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 5
![Page 13: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/13.jpg)
Individual vs collective agency
Different contexts of agency
I Individual decision making and individual action againstnature.
• Ex: Gambling.
I Individual decision making in interaction.
• Ex: Playing chess.
I Collective decision making.
• Ex: Carrying the piano.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 5
![Page 14: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/14.jpg)
Individual vs collective agency
Different contexts of agency
I Individual decision making and individual action againstnature.
• Ex: Gambling.
I Individual decision making in interaction.
• Ex: Playing chess.
I Collective decision making.
• Ex: Carrying the piano.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 5
![Page 15: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/15.jpg)
Individual vs collective agency
Different contexts of agency
I Individual decision making and individual action againstnature.
• Ex: Gambling.
I Individual decision making in interaction.
• Ex: Playing chess.
I Collective decision making.
• Ex: Carrying the piano.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 5
![Page 16: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/16.jpg)
Interaction - Formal models
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 6
![Page 17: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/17.jpg)
Interaction - Formal models
Situations of Interaction
In this course we will mostly study situations of interaction interms of Games in Strategic or Normal form.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 18: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/18.jpg)
Interaction - Formal models
Situations of Interaction
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 19: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/19.jpg)
Interaction - Formal models
Situations of Interaction
DefinitionA game in strategic form G is a tuple 〈A,Si , vi 〉 such that :
I A is a finite set of agents.
I Si is a finite set of actions or strategies for i . A strategyprofile σ ∈ Πi∈ASi is a vector of strategies, one for each agentin I . The strategy si which i plays in the profile σ is noted σi .
I vi : Πi∈ASi −→ R is an utility function that assigns to everystrategy profile σ ∈ Πi∈ASi the utility valuation of that profilefor agent i .
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 20: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/20.jpg)
Interaction - Formal models
Situations of Interaction
DefinitionA game in strategic form G is a tuple 〈A,Si , vi 〉 such that :
I A is a finite set of agents.
I Si is a finite set of actions or strategies for i . A strategyprofile σ ∈ Πi∈ASi is a vector of strategies, one for each agentin I . The strategy si which i plays in the profile σ is noted σi .
I vi : Πi∈ASi −→ R is an utility function that assigns to everystrategy profile σ ∈ Πi∈ASi the utility valuation of that profilefor agent i .
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 21: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/21.jpg)
Interaction - Formal models
Situations of Interaction
DefinitionA game in strategic form G is a tuple 〈A,Si , vi 〉 such that :
I A is a finite set of agents.
I Si is a finite set of actions or strategies for i . A strategyprofile σ ∈ Πi∈ASi is a vector of strategies, one for each agentin I . The strategy si which i plays in the profile σ is noted σi .
I vi : Πi∈ASi −→ R is an utility function that assigns to everystrategy profile σ ∈ Πi∈ASi the utility valuation of that profilefor agent i .
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 22: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/22.jpg)
Interaction - Formal models
Situations of Interaction
DefinitionA game in strategic form G is a tuple 〈A,Si , vi 〉 such that :
I A is a finite set of agents.
I Si is a finite set of actions or strategies for i . A strategyprofile σ ∈ Πi∈ASi is a vector of strategies, one for each agentin I . The strategy si which i plays in the profile σ is noted σi .
I vi : Πi∈ASi −→ R is an utility function that assigns to everystrategy profile σ ∈ Πi∈ASi the utility valuation of that profilefor agent i .
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 7
![Page 23: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/23.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work
Minimal Work
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 24: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/24.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work 1, 1 0, 0
Minimal Work 0, 0 1, 1
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 25: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/25.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work 2, 1 0, 0
Minimal Work 0, 0 1, 2
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 26: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/26.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 27: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/27.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 28: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/28.jpg)
Interaction - Formal models
Games of Interest
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
I Coordination Games.
I Prisoner’s Dilemma.
I In general: games with scope for cooperation.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 8
![Page 29: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/29.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what will the agents do?If they are rational?• Descriptive question.
I Our main focus in this course.I First tenet: As such the question is under-specified.
• One needs to specify the context of interaction (or of thegame). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 30: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/30.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what should the agentsdo? If they are rational?• Normative question.
I Our main focus in this course.I First tenet: As such the question is under-specified.
• One needs to specify the context of interaction (or of thegame). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 31: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/31.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?• Analytical question.
I Our main focus in this course.I First tenet: As such the question is under-specified.
• One needs to specify the context of interaction (or of thegame). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 32: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/32.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?X Analytical question.
I Our main focus in this course.
I First tenet: As such the question is under-specified.
• One needs to specify the context of interaction (or of thegame). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 33: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/33.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?
I Our main focus in this course.
I First tenet: As such the question is under-specified.
• One needs to specify the context of interaction (or of thegame). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 34: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/34.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?
I Our main focus in this course.
I First tenet: As such the question is under-specified.• One needs to specify the context of interaction (or of the
game). This includes:
I (possibly) some additional group- or team-related aspects ofthe game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 35: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/35.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?
I Our main focus in this course.
I First tenet: As such the question is under-specified.• One needs to specify the context of interaction (or of the
game). This includes:I (possibly) some additional group- or team-related aspects of
the game.
I Information of the agents about all relevant aspects of thegame.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 36: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/36.jpg)
Interaction - Formal models
The Main Question(s)
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
I When there is scope for cooperation, what does it mean tosay that they are rational?
I Our main focus in this course.
I First tenet: As such the question is under-specified.• One needs to specify the context of interaction (or of the
game). This includes:I (possibly) some additional group- or team-related aspects of
the game.I Information of the agents about all relevant aspects of the
game.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 9
![Page 37: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/37.jpg)
Information in games
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 10
![Page 38: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/38.jpg)
What does it mean to be (perfectly) rational?
U H M
H 3,3 0,0 H
M 0,0 1,1 H
Ann’s best choice depends on what she expects Bob to do, and thisdepends on what she thinks Bob expects her to do, and so on...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 11
![Page 39: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/39.jpg)
What does it mean to be (perfectly) rational?
U H M
H 3,3 0,0 H
M 0,0 1,1 H
Ann’s best choice depends on what she expects Bob to do, and thisdepends on what she thinks Bob expects her to do, and so on...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 11
![Page 40: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/40.jpg)
What does it mean to be (perfectly) rational?
U H M
H 3,3 0,0 H
M 0,0 1,1 H
Instrumental Rationality: maximize given your current information(Bayesian Decision Theory)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 11
![Page 41: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/41.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 42: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/42.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 43: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/43.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 44: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/44.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 45: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/45.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 46: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/46.jpg)
Information in games situations
I Various states of information disclosure.
• ex ante, ex interim, ex post
I Various “types” of information:
• imperfect information about the play of the game• incomplete information about the structure of the game• strategic information (what will the other players do?)• higher-order information (what are the other players thinking?)
I Varieties of informational attitudes
• hard (“knowledge”)• soft (“beliefs”)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 12
![Page 47: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/47.jpg)
Finding the “rational” choice
Game G
Strategy Space
Player 1’s States Player2’s States
G : available actions, payoffs, structure of the decision problem
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 13
![Page 48: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/48.jpg)
Finding the “rational” choice
Game G
Strategy Space
Player 1’s States Player 2’s States
solution concepts are systematic descriptions of what players do
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 13
![Page 49: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/49.jpg)
Finding the “rational” choice
Game G
Strategy Space
Player 1’s States Player 2’s States
Consider possible information states of the players
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 13
![Page 50: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/50.jpg)
Finding the “rational” choice
Game G
Strategy Space
Player 1’s States Player 2’s States
Restrict to information states satisfying some rationality condition
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 13
![Page 51: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/51.jpg)
Finding the “rational” choice
Game G
Strategy Space
Player 1’s States Player 2’s States
Project onto the strategy space
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 13
![Page 52: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/52.jpg)
Time for some details...
Two general modeling strategies:
1. Harsanyi type spaces: sorted structure with maps betweenplayers’ “states”
J. Harsanyi. Games with incomplete information played by “bayesian” playersI-III. Management Science Theory 14: 159-182, 1967-68.
2. Partition model: single set of states with partitions describingthe players’ (hard) information
R. Aumann. Interactive Epistemology I & II. International Journal of GameTheory (1999).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 14
![Page 53: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/53.jpg)
Time for some details...
Two general modeling strategies:
1. Harsanyi type spaces: sorted structure with maps betweenplayers’ “states”
J. Harsanyi. Games with incomplete information played by “bayesian” playersI-III. Management Science Theory 14: 159-182, 1967-68.
2. Partition model: single set of states with partitions describingthe players’ (hard) information
R. Aumann. Interactive Epistemology I & II. International Journal of GameTheory (1999).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 14
![Page 54: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/54.jpg)
Time for some details...
Two general modeling strategies:
1. Harsanyi type spaces: sorted structure with maps betweenplayers’ “states”
J. Harsanyi. Games with incomplete information played by “bayesian” playersI-III. Management Science Theory 14: 159-182, 1967-68.
2. Partition model: single set of states with partitions describingthe players’ (hard) information
R. Aumann. Interactive Epistemology I & II. International Journal of GameTheory (1999).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 14
![Page 55: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/55.jpg)
Literature
See, for example,
P. Battigalli and G. Bonanno. Recent results on belief, knowledge and theepistemic foundations of game theory. Research in Economics (1999).
B. de Bruin. Explaining Games. Ph.D. Thesis, ILLC (2004).
A. Brandenburger. The Power of Paradox: Some Recent Developments in Inter-active Epistemology. International Journal of Game Theory (2007).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 15
![Page 56: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/56.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 57: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/57.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
A set of information statesand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 58: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/58.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
A set of information statesand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 59: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/59.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
A set of information statesand Bob chooses M
Are these choices rational?Yes.
Ann considers it possibleBob is irrational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 60: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/60.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
A set of information statesand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 61: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/61.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
A common prior and Bobchooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 62: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/62.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 63: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/63.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 64: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/64.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 65: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/65.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
12
18
116
14
18
116
14
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 66: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/66.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
12
18
116
14
18
116
14
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
3·PA(H)+0·PA(M) ≥ 0·PA(H)+1·PA(M)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 67: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/67.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
12
18
116
14
18
116
14
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
3·PA(H)+0·PA(M) ≥ 0·PA(H)+1·PA(M)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 68: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/68.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
12
18
116
14
18
116
14
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
3 · 12 + 0 · PA(M) ≥ 0 · 12 + 1 · PA(M)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 69: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/69.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
12
18
116
14
18
116
14
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
3 · 12 + 0 · 12 ≥ 0 · 12 + 1 · 12
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 70: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/70.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
112
212
18
116
112
812
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
0·PB(H)+1·PB(M) ≥ 3·PB(H)+0·PB(M)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 71: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/71.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
112
212
18
116
112
812
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Ann (Bob) knows thatBob (Ann) is rational
0 · 212 + 1 · 1012 ≥ 3 · 2
12 + 0 · 1012
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 72: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/72.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Bob (Ann) knows thatAnn (Bob) is rational
1·PA(L)+0·PA(R) ≥ 0·PA(L)+2·PA(R)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 73: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/73.jpg)
An Example
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
16
116
18
16
116
12
46
Suppose Ann chooses Hand Bob chooses M
Are these choices rational?Yes.
Bob (Ann) knows thatAnn (Bob) is rational
0 · 16 + 1 · 56 ≥ 3 · 16 + 0 · 56
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 16
![Page 74: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/74.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
116
18
116
12
Zero probability 6= “impossible”
Different “types” of players canmake the same choice
Are Ann and Bob rational? Yes.
Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 75: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/75.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
Different “types” of players canmake the same choice
Are Ann and Bob rational? Yes.
Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 76: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/76.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
2. Different “types” of players canmake the same choice
Are Ann and Bob rational? Yes.
Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 77: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/77.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
2. Different “types” of players canmake the same choice
I Are Ann and Bob rational? Yes.
Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 78: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/78.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
2. Different “types” of players canmake the same choice
I Are Ann and Bob rational? Yes.
Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 79: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/79.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
2. Different “types” of players canmake the same choice
I Are Ann and Bob rational? Yes.
I Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 80: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/80.jpg)
Two Issues
Bob
Ann
U H M
H 3,3 0,0 H
M 0,0 1,1 H
18
18
18
18
0
12
1. Zero probability 6= “impossible”
2. Different “types” of players canmake the same choice
I Are Ann and Bob rational? Yes.
I Do they know that each other isrational? No.(though PrBob(Irrat(Ann)) = 0)
H
M
H M
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 17
![Page 81: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/81.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The other players’ types
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 82: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/82.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The other players’ types
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 83: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/83.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
I Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The other players’ types
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 84: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/84.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
I Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
Player i ’s types
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 85: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/85.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
I Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The set of all probability distributions
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 86: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/86.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
I Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The other players’ types
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 87: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/87.jpg)
Harsanyi Type SpaceBased on the work of John Harsanyi on games with incompleteinformation, game theorists have developed an elegant formalismthat makes precise talk about beliefs, knowledge and rationality:
I A type is everything a player knows privately at the beginningof the game which could affect his beliefs about payoffs andabout all other players’ possible types.
I Each type is assigned a joint probability over the space oftypes and actions
λi : Ti → ∆(T−i × S−i )
The other players’ choices
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 18
![Page 88: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/88.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
One type for Ann (tA) and two typesfor Bob (tB , uB)
A state is a tuple of choices andtypes: (M, tA,M, uB)
Calculate expected utility in the usualway...
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 89: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/89.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I One type for Ann (tA) and two typesfor Bob (tB , uB)
A state is a tuple of choices andtypes: (M, tA,M, uB)
Calculate expected utility in the usualway...
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 90: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/90.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I One type for Ann (tA) and two typesfor Bob (tB , uB)
I A state is a tuple of choices andtypes: (M,M, tA, tB)
Calculate expected utility in the usualway...
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 91: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/91.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I One type for Ann (tA) and two typesfor Bob (tB , uB)
I A state is a tuple of choices andtypes: (M, tA,M, uB)
I Calculate expected utility in theusual way...
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 92: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/92.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
Ann (tA) is rational0 · 0.5 + 1 · 0 ≥ 3 · 0.5 + 0 · 0.2Bob is rational0 · 0.5 + 1 · 0 ≥ 3 · 0.5 + 0 · 0.2Bob thinks Ann is irrationalPB(Irrat(Ann)) = 0.xx
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 93: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/93.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I M is rational for Ann (tA)0 · 0.2 + 1 · 0.8 ≥ 3 · 0.2 + 0 · 0.8Bob is rational0 · 0.5 + 1 · 0 ≥ 3 · 0.5 + 0 · 0.2Bob thinks Ann is irrationalPB(Irrat(Ann)) = 0.xx
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 94: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/94.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I M is rational for Ann (tA)0 · 0.2 + 1 · 0.8 ≥ 3 · 0.2 + 0 · 0.8
I M is rational for Bob (tB)0 · 0 + 1 · 1 ≥ 3 · 0 + 0 · 1Bob thinks Ann is irrationalPB(Irrat(Ann)) = 0.xx
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 95: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/95.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I M is rational for Ann (tA)0 · 0.2 + 1 · 0.8 ≥ 3 · 0.2 + 0 · 0.8
I M is rational for Bob (tB)0 · 0 + 1 · 1 ≥ 3 · 0 + 0 · 1
I Ann thinks Bob may be irrationalPB(Irrat(Ann)) = 0.xx
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 96: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/96.jpg)
Returning to the Example: A Game Model
Bob
Ann
U H M
H 3,3 0,0
M 0,0 1,1
I M is rational for Ann (tA)0 · 0.2 + 1 · 0.8 ≥ 3 · 0.2 + 0 · 0.8
I M is rational for Bob (tB)0 · 0 + 1 · 1 ≥ 3 · 0 + 0 · 1
I Ann thinks Bob may be irrationalPA(Irrat[B]) = 0.3, PA(Rat[B]) = 0.7
tA
U H M
tB 0 0.5
uB 0.2 0.3tB
U H M
tA 0 1
uB
U H M
tA 0.4 0.6
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 19
![Page 97: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/97.jpg)
Bayesian Rationality
Notation:I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) and
t ∈ T1 × · · · × Tn (set of types).I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other
agents’ choices (similarly for types) and si ∈ Si agent i ’schoice.
I Write (si , s−i ) for s.I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Agent i is rational at state (s, t) whenever:
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 98: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/98.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Agent i is rational at state (s, t) whenever:
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 99: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/99.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Agent i is rational at state (s, t) whenever:
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 100: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/100.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Sum over all possible types and choices
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 101: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/101.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
ti ’s 1st-order beliefs
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 102: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/102.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Agent i ’s utility
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 103: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/103.jpg)
Bayesian Rationality
Notation:
I Suppose s ∈ S = S1 × · · · × Sn (pure strategy profiles) andt ∈ T1 × · · · × Tn (set of types).
I Let s−i ∈ S1 × · · · × Si−1 × Si+1 × · · · × Sn be the other agents’choices (similarly for types) and si ∈ Si agent i ’s choice.
I Write (si , s−i ) for s.
I For ti ∈ Ti , let pti = λ(ti ) ∈ ∆(S−i × T−i )
Agent i ’s expected value at state (s, t) is:
EVi (s, t) =∑t′−i
∑σ′−i
pti (s ′−i , t′−i )ui (si , s
′−i )
Agent i is rational at state (s, t) whenever:
si ∈ argmaxs′∈Si (EVi (s[si 7→ s ′i ], t))
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 20
![Page 104: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/104.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 105: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/105.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 106: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/106.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 107: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/107.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 108: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/108.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 109: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/109.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 110: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/110.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 111: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/111.jpg)
Bayesian Rationality
Beyond Game Models: Bayesian GamesThe components of a Bayesian game for a set of agents A:
I S = S1 × · · · × Sn where Si is the set of actions for player i ;
I T = T1 × · · · × Tn where Ti is the set of types for player i ;
I u = (u1, . . . , un) where ui : A× Ti → R is a utility function.
I p : T → [0, 1] is a common prior over types; and
I αi : Ti → Ai is a pure strategy function
I Let si be a mixed strategy and si (ai |ti ) denote the probabilityagent i plays ai given that i ’s type is ti .
I Ex post expected value:EUi (s, t) =
∑a∈A
∏j∈A sj(aj |tj)ui (a, t)
I Ex interim expected value:EUi (s, ti ) =
∑t−i∈T−i
p(t−i |ti )EUi (s, (ti , t−i ))
I Ex ante expected value:EUi (s) =
∑ti∈Ti
p(ti )EUi (s, ti )
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 21
![Page 112: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/112.jpg)
Bayesian Rationality
Rationalizability
Hard Work Minimal Work
Hard Work 3, 3
Minimal Work 1, 1
DefinitionStrategy s ′i of agent i is rationalizable if there exists a state (s, t)in a type structure T such that si = s ′i and s ′i is rational at (s, t).
Observations: “Methodological Individualism”
1. Coordination on Pareto sub-optimal outcome is rationalizable.
2. The cooperative outcome (HW,HW) in the PD is notrationalizable. (Why?)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 22
![Page 113: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/113.jpg)
Bayesian Rationality
Rationalizability
Hard Work Minimal Work
Hard Work 3, 3 0, 0
Minimal Work 0, 0 1, 1
DefinitionStrategy s ′i of agent i is rationalizable if there exists a state (s, t)in a type structure T such that si = s ′i and s ′i is rational at (s, t).
Observations: “Methodological Individualism”
1. Coordination on Pareto sub-optimal outcome is rationalizable.
2. The cooperative outcome (HW,HW) in the PD is notrationalizable. (Why?)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 22
![Page 114: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/114.jpg)
Bayesian Rationality
Rationalizability
Hard Work Minimal Work
Hard Work 3, 3 0, 4
Minimal Work 4, 0 1, 1
DefinitionStrategy s ′i of agent i is rationalizable if there exists a state (s, t)in a type structure T such that si = s ′i and s ′i is rational at (s, t).
Observations: “Methodological Individualism”
1. Coordination on Pareto sub-optimal outcome is rationalizable.
2. The cooperative outcome (HW,HW) in the PD is notrationalizable. (Why?)
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 22
![Page 115: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/115.jpg)
Bayesian Rationality
Teamwork once again
I Coordination on Pareto sub-optimal outcome is rationalizable.
I The cooperative outcome (HW,HW) in the PD is notrationalizable.
Question: can teamwork do better than that?Intuitively, Yes.
“There are these two broad empirical facts about Hi-Logames, people almost always choose A [Hi] and peoplewith common knowledge of each other’s rationality thinkit is obviously rational to choose A [Hi].”
[Bacharach, Beyond Individual Choice, 2006, pg. 42]
See also chapter 2 of:C.F. Camerer. Behavioral Game Theory. Princeton UP, 2003.
But then more machinery is needed...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 23
![Page 116: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/116.jpg)
Bayesian Rationality
Teamwork once again
I Coordination on Pareto sub-optimal outcome is rationalizable.
I The cooperative outcome (HW,HW) in the PD is notrationalizable.
Question: can teamwork do better than that?
Intuitively, Yes.
“There are these two broad empirical facts about Hi-Logames, people almost always choose A [Hi] and peoplewith common knowledge of each other’s rationality thinkit is obviously rational to choose A [Hi].”
[Bacharach, Beyond Individual Choice, 2006, pg. 42]
See also chapter 2 of:C.F. Camerer. Behavioral Game Theory. Princeton UP, 2003.
But then more machinery is needed...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 23
![Page 117: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/117.jpg)
Bayesian Rationality
Teamwork once again
I Coordination on Pareto sub-optimal outcome is rationalizable.
I The cooperative outcome (HW,HW) in the PD is notrationalizable.
Question: can teamwork do better than that?Intuitively, Yes.
“There are these two broad empirical facts about Hi-Logames, people almost always choose A [Hi] and peoplewith common knowledge of each other’s rationality thinkit is obviously rational to choose A [Hi].”
[Bacharach, Beyond Individual Choice, 2006, pg. 42]
See also chapter 2 of:C.F. Camerer. Behavioral Game Theory. Princeton UP, 2003.
But then more machinery is needed...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 23
![Page 118: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/118.jpg)
Bayesian Rationality
Teamwork once again
I Coordination on Pareto sub-optimal outcome is rationalizable.
I The cooperative outcome (HW,HW) in the PD is notrationalizable.
Question: can teamwork do better than that?Intuitively, Yes.
“There are these two broad empirical facts about Hi-Logames, people almost always choose A [Hi] and peoplewith common knowledge of each other’s rationality thinkit is obviously rational to choose A [Hi].”
[Bacharach, Beyond Individual Choice, 2006, pg. 42]
See also chapter 2 of:C.F. Camerer. Behavioral Game Theory. Princeton UP, 2003.
But then more machinery is needed...
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 23
![Page 119: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/119.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not. But interesting phenomena at this level already.
⇒ Coalitional powers (c.f. Pauly 2002).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 120: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/120.jpg)
Bayesian Rationality
What is a team?
Any group?I Surely not. But interesting phenomena at this level already.
⇒ Coalitional powers (c.f. Pauly 2002).
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 121: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/121.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not.
Then a group with:
i A certain (hierarchical) structure?
ii Whose members identify with the group (c.f. Gold 2005)?
• Information about who’s in and who’s out.• Reasoning and acting as group members.
iii Team- or group objectives/aims/preferences?
iv Shared commitments? (Bratman, 1999, Gilbert 1989,Tuomela, 2007)
v Common knowledge (beliefs?) of (i-iv)?
Note: None of these are necessary conditions!
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 122: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/122.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not.
Then a group with:
i A certain (hierarchical) structure?
ii Whose members identify with the group (c.f. Gold 2005)?
iii Team- or group objectives/aims/preferences?
• Shared by the members?
iv Shared commitments? (Bratman, 1999, Gilbert 1989,Tuomela, 2007)
v Common knowledge (beliefs?) of (i-iv)?
Note: None of these are necessary conditions!
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 123: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/123.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not.
Then a group with:
i A certain (hierarchical) structure?
ii Whose members identify with the group (c.f. Gold 2005)?
iii Team- or group objectives/aims/preferences?
iv Shared commitments? (Bratman, 1999, Gilbert 1989,Tuomela, 2007)
• Shared intentions.• Sanctions for lapsing?• Shared praise[blame] for success[failure]?
v Common knowledge (beliefs?) of (i-iv)?
Note: None of these are necessary conditions!
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 124: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/124.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not.
Then a group with:
i A certain (hierarchical) structure?
ii Whose members identify with the group (c.f. Gold 2005)?
iii Team- or group objectives/aims/preferences?
iv Shared commitments? (Bratman, 1999, Gilbert 1989,Tuomela, 2007)
v Common knowledge (beliefs?) of (i-iv)?
Note: None of these are necessary conditions!
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 125: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/125.jpg)
Bayesian Rationality
What is a team?
Any group?
I Surely not.
Then a group with:
i A certain (hierarchical) structure?
ii Whose members identify with the group (c.f. Gold 2005)?
iii Team- or group objectives/aims/preferences?
iv Shared commitments? (Bratman, 1999, Gilbert 1989,Tuomela, 2007)
v Common knowledge (beliefs?) of (i-iv)?
Note: None of these are necessary conditions!
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 24
![Page 126: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/126.jpg)
Bayesian Rationality
Recap
Acting as a team involve:
I Adopting the team’s preferences. (Preference transformation).
I Team-reasoning (Agency Transformation).
Later this week.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 25
![Page 127: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/127.jpg)
Bayesian Rationality
Recap
Acting as a team involve:
I Adopting the team’s preferences. (Preference transformation).
I Team-reasoning (Agency Transformation).
Later this week.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 25
![Page 128: Logic, Interaction and Collective Agencyepacuit/lograt/ESSLLI_slides_L1.pdfLogic, Interaction and Collective Agency Lecture 1 ESSLLI’10, Copenhagen Eric Pacuit Olivier Roy TiLPS,](https://reader033.fdocuments.in/reader033/viewer/2022042007/5e707182d70e421254572712/html5/thumbnails/128.jpg)
Tomorrow
I Building the common perspective: (a non-standardintroduction to) common knowledge, and common modes ofreasoning.
Eric Pacuit and Olivier Roy: Individual and Collective Agency (ESSLLI’10) 26