7. Jim Joyce: The Proper Role of Ratifiability Considerations in Decision Making
TLDRThe speaker reflects on the influence of Bill's philosophy on decision theory, particularly causal decision theory and the principle of causal ratifiability. They recount a vivid memory of Bill's engaging lecture style and delve into the complexities of decision-making under uncertainty, using thought experiments like the 'Murder Legion' case. The talk challenges the traditional views on ratifiability, arguing that unconditional causal expected utility should guide decisions, even when acts are unratified, and emphasizes the importance of incorporating all causally relevant information.
Takeaways
- 😀 The speaker first encountered Bill's ideas in the mid-1980s during a symposium on decision theory, which had a profound impact on their professional life.
- 🧠 The concept of 'causal decision theory' was introduced, which emphasizes choosing actions that promote desirable outcomes based on their causal relationships.
- 🔍 The speaker discussed a principle of causal decision theory called 'causal ratifiability,' suggesting that if an action is regretted in anticipation of its performance, it's a reason not to choose that action.
- 🤔 The speaker has come to have second thoughts about causal ratifiability after 25 years, indicating a shift in their perspective on this principle.
- 🎯 A detailed explanation of causal decision theory was provided, including how to calculate expected utility by considering partitions of states of the world.
- 🤷♂️ The speaker highlighted a crucial aspect of causal decision theory: the irrelevance of information about what act one is likely to perform, except as it provides evidence about the act's causal outcomes.
- 📉 The transcript explored the 'Murder Legion' scenario, illustrating the complexities and potential issues with causal decision theory and the concept of ratifiability.
- 🤔 The speaker questioned the assumption that causal decision theory recommends a specific action in the 'Murder Legion' scenario, suggesting that the theory's application is more nuanced.
- 🔄 The concept of dynamic deliberation was introduced, where beliefs and probabilities are updated iteratively until an equilibrium is reached, reflecting a comprehensive consideration of causally relevant information.
- 🚫 The speaker argued against the idea that unratified acts are impermissible, stating that the mere fact an act is unratified does not make it an incorrect choice, even when ratified alternatives are available.
- 🔄 The importance of tempering the implications of ratifiability with beliefs about the potential and warranted nature of regrets was emphasized, suggesting that unconditional causal expected utility already incorporates the magnitude of regrets.
Q & A
What is the speaker's first memory of hearing Bill speak, and how did it impact their professional life?
-The speaker can't remember exactly when they first met Bill but vividly recalls hearing him talk at a symposium during their graduate school days, possibly in 1984 or 1986. This talk had a significant impact on the speaker's professional life, as it excited them about the subject and motivated them to learn more, setting a standard for how they view philosophy should be conducted.
What is causal decision theory, and how does it relate to the speaker's interest in decision theory?
-Causal decision theory is a framework for understanding decision-making that emphasizes choosing actions that causally promote desirable outcomes. The speaker became interested in decision theory during their graduate studies and attended a talk by Bill to explore the field further, where they were introduced to the concept of causal decision theory.
What principle did Bill advocate for in causal decision theory that the speaker found convincing?
-Bill advocated for a principle in causal decision theory called 'causal ratifiability,' which suggests that if you recognize you'll regret performing an act in light of the fact that you're going to perform it, then that's a reason not to choose the act. The speaker found this principle convincing upon first hearing it.
What is the 'murder legion' thought experiment, and why is it relevant to the discussion?
-The 'murder legion' is a thought experiment involving a scenario where a person has the opportunity to kill a king named Alfred using a gun pointed at his head. The person has a condition that either gives them the nerve to shoot but makes their hands shake, or they lack the nerve but have steady hands. This experiment is relevant because it challenges the causal decision theory and the concept of causal ratifiability, leading the speaker to rethink the principle after 25 years.
What is the significance of the 'Nukem problem' in the context of causal decision theory?
-The 'Nukem problem' is a hypothetical scenario often used to illustrate issues in causal decision theory, where one action is dominant regardless of the outcome. It's significant because it highlights the limitations of causal decision theory when it comes to cases where the theory's recommendations don't align with intuitive decision-making.
How does the speaker's view on causal ratifiability differ from Egan's interpretation?
-The speaker argues that causal decision theory does not categorically recommend shooting in the 'murder legion' scenario, as Egan suggests. Instead, the speaker believes that the theory allows for the consideration of all causally relevant information, leading to a more nuanced understanding of the decision at hand.
What is the 'freedom principle' mentioned by the speaker, and how does it relate to decision-making?
-The 'freedom principle' is the idea that if a person takes themselves to be free to perform an action, then the information that an action has a higher expected utility than its alternatives should increase their confidence in performing that action. It relates to decision-making by emphasizing the importance of incorporating all available information into the decision process.
How does the speaker address the concern that unratified acts might lead to regret?
-The speaker argues that the mere fact that an act is unratified does not make it impermissible for choice. They suggest that the relevant considerations of regret are already taken into account in the unconditional expected utilities, and that one should discount the magnitude of regrets by the probability that these regrets are warranted.
What is the 'weak desire reflection principle' and its relevance to the discussion on decision-making?
-The 'weak desire reflection principle' posits that a person's expected utility for an action at a given time should be based on their expectations of their future utility for that action when they have more evidence. It's relevant to the discussion as it provides a framework for understanding how new information should influence decision-making and the potential for regret.
What is the speaker's conclusion on how to approach decision-making in complex scenarios like the 'murder legion' case?
-The speaker concludes that one should take all available information about what their acts are likely to cause into account and then maximize unconditional causal expected utility. They argue that in cases like the 'murder legion,' where every choice is fine, the theory supports the idea that one can choose any action without making a mistake.
Outlines
🎓 Reflecting on a Formative Academic Experience
The speaker recalls a pivotal moment in their graduate school days when they first heard Bill, a philosopher, present a compelling argument on causal decision theory at a symposium. This experience was transformative, igniting a passion for the subject and setting a standard for how philosophy should be conducted. The talk by Bill revolved around the concept of causal decision theory, which emphasizes choosing actions that promote desirable outcomes and disregarding the evidential value of actions. Bill defended a principle of causal decision theory called 'gratifi ability,' which suggests that if an action is anticipated to be regretted, it is a reason not to perform it. The speaker has since used this principle in their work but has developed some reservations after 25 years of reflection.
🧐 Delving into Causal Decision Theory and Its Implications
The speaker delves deeper into causal decision theory, contrasting it with traditional decision-making approaches. They discuss the importance of maximizing expected utility while considering the causal impact of actions, and the theory's evolution through contributions from various philosophers. A key aspect is the irrelevance of information about the likelihood of performing an action, unless it provides evidence about the action's causal outcomes. The speaker begins to express concerns about the 'gratifi ability' concept through examples, including a scenario involving a character at risk of encountering the Grim Reaper, highlighting the complexities and potential issues with causal decision theory.
🤔 Challenging the Gratifi Ability Principle with Counterexamples
The speaker presents a counterexample to the gratifi ability principle using a scenario involving a character with a brain lesion that affects their ability to shoot accurately. This example is meant to challenge the causal decision theory's recommendations and the speaker's own convictions about the theory. They discuss how the expected utility of shooting exceeds that of not shooting, based on certain probabilities and utilities, but also express reservations about the intuitive response to the scenario, suggesting that shooting may not be the correct action after all.
🔍 Analyzing the Dynamics of Rational Deliberation
The speaker explores the dynamics of rational deliberation, discussing how beliefs and probabilities should be updated in light of new information and how this process can lead to an equilibrium state. They argue that causal decision theory does not necessarily recommend shooting in the counterexample scenario, contrary to what some might assume. The speaker emphasizes the importance of incorporating all causally relevant information into decision-making and critiques the idea of relying solely on initial probabilities, suggesting a more dynamic approach to evaluating actions.
💡 The Importance of Fully Informed Beliefs in Decision Making
The speaker underscores the significance of having fully informed beliefs when making decisions, arguing that causal decision theory instructs individuals to act based on comprehensive information about the likely causes of their actions. They introduce the concept of 'freedom of action' and propose that one should only act when their beliefs have taken into account all readily available and relevant evidence. The speaker also discusses the implications of this principle for the 'murder leion' example, suggesting that neither shooting nor not shooting is the unequivocally correct choice.
🤖 The Role of Regret in Decision Theory
The speaker examines the role of regret in decision-making, discussing the concept of 'ratified' acts—those that remain the best choice even after being chosen. They argue against the idea that acts should be dismissed solely based on the potential for regret, especially when other ratified alternatives are available. The speaker challenges the notion that unconditional causal expected utility should always be maximized, suggesting that regret should be considered in a more nuanced way.
🎯 Navigating the Pathology of Unratified Decisions
The speaker addresses the perceived pathology of unratified decisions, where choosing an act does not seem to be the best choice after the decision is made. They discuss the 'murder leion' case and the 'death in Damascus' problem, suggesting that causal decision theory allows for the possibility of choosing unratified acts. The speaker also critiques the idea of using regret as a decisive reason not to choose a particular act, arguing for a more balanced approach to decision-making.
🤝 Integrating Regret and Probability in Decision-Making
The speaker integrates the concepts of regret and probability into the decision-making process, arguing that unconditional causal expected utility already takes into account the potential for regret. They provide a mathematical representation of this idea, suggesting that the equilibrium state in decision-making balances the magnitude of regret with the probability of making certain decisions. The speaker concludes that while one may regret their choice, this regret should be discounted by the likelihood that it is warranted.
📉 The Impact of Dynamic Deliberation on Decision Theory
The speaker discusses the impact of dynamic deliberation on decision theory, suggesting that engaging in a process of iterative reasoning can lead to an equilibrium where every choice is equally valid. They explore the implications of this for the 'murder leion' scenario, arguing that causal decision theory supports the idea that any action can be the right one under certain conditions. The speaker also addresses the philosophical and practical challenges of this approach, including the potential for regret and the need for a rule to choose an action when all options are equally valid.
🤔 The Philosophical Implications of Causal Decision Theory
The speaker engages in a philosophical discussion about the implications of causal decision theory, particularly in relation to the concepts of regret and unratified acts. They challenge the audience to consider whether the potential for regret should be a decisive factor in decision-making and argue for a more comprehensive approach that considers all available information. The speaker also addresses the complexities of applying causal decision theory to real-world scenarios, including the need for a deeper understanding of the causal relationships between actions and outcomes.
🎲 Exploring Game Theory and Decision Making Under Uncertainty
The speaker explores the application of causal decision theory to game theory scenarios, discussing the challenges of making decisions when there is uncertainty about other players' actions. They examine cases where common knowledge assumptions may not hold, leading to situations where traditional decision-making strategies might not yield equilibrium solutions. The speaker suggests that causal decision theory can provide a more robust framework for making choices in such complex environments, even when faced with the potential for regret or unratified actions.
📚 Concluding Thoughts on Decision Theory and Rationality
In conclusion, the speaker reflects on the broader implications of causal decision theory for rational decision-making. They emphasize the importance of taking all available information into account and maximizing unconditional causal expected utility, regardless of the potential for regret or the presence of unratified acts. The speaker also acknowledges the complexities of applying these principles to real-world situations and the need for further exploration of how dynamic deliberation and game theory can inform our understanding of rational choice.
Mindmap
Keywords
💡Causal Decision Theory
💡Ratification
💡Expected Utility
💡Unratified Actions
💡Regret
💡Equilibrium State
💡Dynamic Deliberation
💡Game Theory
💡Common Knowledge
💡Nash Equilibrium
Highlights
The speaker shares a personal anecdote about the impact of Bill's philosophy on their professional life.
Introduction to causal decision theory and its significance in choosing actions that promote desirable outcomes.
Discussion of the causal decision theory's principle of regret, suggesting not to choose actions you anticipate regretting.
The speaker's initial conviction and later second thoughts on causal decision theory after 25 years.
Explanation of the partition method in calculating expected utility within causal decision theory.
The importance of ignoring evidential implications of actions that do not causally promote a state.
The dilemma of decision-making when actions indicate outcomes they do not cause, as illustrated by the 'Death in Damascus' scenario.
Critique of causal decision theory through Andy Egan's counterexamples, challenging the theory's recommendations.
Analysis of the 'Murder Legion' case, questioning the rationality of actions based on causal expected utility.
The concept of ratified acts and their role in decision-making, where an act is considered ratified if it still seems best upon decision.
The speaker argues against the view that unratified acts are impermissible for choice, even when ratified alternatives are available.
The role of regret in decision-making, and how it should be factored into the assessment of actions.
Critique of the no-regrets principle, suggesting that the mere fact an act is unratified does not make it impermissible.
The importance of tempering the implications of gratifiability with beliefs about the potential and warranted nature of regrets.
The conclusion that in cases like 'Murder Legion', any action is permissible if it is based on a balanced consideration of regrets and probabilities.
The speaker emphasizes the need to take all available information into account and maximize unconditional causal expected utility in decision-making.
Discussion on the implications of the theory for game theory, particularly in scenarios lacking common knowledge or equilibrium.
Transcripts
Browse More Related Video
Robert Spekkens: Leibniz’s principle of the identity of indiscernible as...
8. Alan Hájek: The Extremely Improbable
Laura Felline: Scientific explanation in information-theoretic reconstructions of quantum theory
Jenann Ismael: The link between time, totality, and determinism
9. Brian Skyrms: The Evolution of Signalling Systems
James Fraser: Rethinking Perturbation Theory (in the 1950s)
5.0 / 5 (0 votes)
Thanks for rating: