Is effective altruism just for consequentialists? | Andreas Mogensen (2022)

Andreas Mogensen's Background & Current Work

• Senior Research Fellow at Oxford University's Global Priorities Institute with a PhD from All Souls College.
• Early adherent to the Giving What We Can pledge (joined in 2010).
• Current research focuses on the "Hinge of History Hypothesis," specifically critiquing arguments that optimal timing requires waiting for a future moment rather than acting now.
• Awaiting publication of a critical reply to Will McCaskill's paper Against the Hinge of History Hypothesis.

Deontology vs. Consequentialism

• Core Distinction: Consequentialism defines right action solely by the maximization of the good (e.g., total welfare), whereas deontology posits multiple sources of moral reasons, including rights, promises, and partiality, which can constrain the pursuit of the good.
• Rejection of Singular Criteria: Mogensen favors W.D. Ross's pluralistic deontology, which rejects the idea that a single property defines all right actions.
• Kantian Critique: He finds Kant's "universal law" formulation unclear and potentially absurd, noting it implies impermissible restrictions on charitable giving if the average person cannot also give more than the average.
• Practical Overlap: Despite theoretical differences, Mogensen argues that non-consequentialists can still hold strong duties of beneficence and support effective altruism, provided actions do not violate specific rights or justice principles.

Effective Altruism and Deontological Constraints

• Pledge Support: Mogensen endorses the Giving What We Can pledge, believing strong obligations to aid those in extreme poverty exist even without full consequentialism.
• Side Constraints: He agrees with Effective Altruism (EA) that actions like stealing or killing to save others are impermissible, noting that real-world cases rarely align with the "transplant case" thought experiments where consequentialism might demand such acts.
• Aggregation Disagreements: Deontologists may reject utilitarian cost-effectiveness calculations that aggregate many trivial benefits against one severe harm (e.g., saving one life vs. giving lollipops to billions).
• Justice vs. Beneficence: Deontologists may prioritize rectifying specific injustices where they are complicit over impartial beneficence to strangers, a tension Mogensen notes is less pronounced in global poverty interventions where global justice systems are implicated.

Long-Termism and Population Ethics

• Deontic vs. Axiological Long-Termism: Mogensen distinguishes between the view that the future is better (axiological) and the view that we are obligated to secure it (deontic). He is more skeptical of deontic long-termism.
• Creation of Lives: He is skeptical of the claim that we have moral obligations to bring additional happy lives into existence or to choose between potential people based on which life would be "better."
• Non-Identity Problem: Mogensen aligns with David Boonin's view that while it is better to bring into existence a person with a longer, happier life, it is not morally obligatory to do so compared to bringing a person with a shorter life into existence.
• Discounting Future Welfare: He proposes that partiality to nearer descendants could theoretically justify a positive, though not necessarily extreme, discount rate for future welfare, challenging the standard economic practice of exponential discounting.

Large-Number Skepticism

• The World Cup Thought Experiment: Based on T.M. Scanlon, this scenario asks whether one should save an electrocuted individual or allow a broadcast to continue to save the pleasure of 500 million+ viewers. Mogensen argues many intuitively believe no number of viewers justifies the death.
• Critique of Skepticism Arguments: In his paper Against Large-Number Skepticism, Mogensen refutes arguments that humans cannot comprehend large numbers:
  • Evolutionary Analogy: He argues resistance to evolutionary theory is not primarily due to an inability to grasp time scales, but rather specific disputes over the Earth's age.
  • Time vs. Numbers: He contends that the "Hume" astronomy anecdote (confusing 5 million vs. 5 billion years) reflects time discounting rather than an inability to grasp number magnitudes.
  • Scope Insensitivity: He critiques the use of contingent valuation studies (e.g., saving birds) as evidence, noting these rely on "separate evaluation" rather than the "joint evaluation" used in moral thought experiments.
• Conclusion: He concludes that the case for large-number skepticism is currently inconclusive and that existing empirical literature does not warrant dismissing intuitions about large numbers.

Evolutionary Debunking Arguments (EDA)

• The Challenge: EDAs claim that since moral beliefs were shaped by evolutionary selection (indifferent to truth), we lack justification for believing them.
• Mogensen's Rebuttal:
  • Epistemic Principles: He argues that applying EDA principles (like Occam's Razor or sensitivity) often leads to generalized skepticism about the external world (e.g., "brain in a vat"), which most philosophers reject.
  • Proximate vs. Ultimate Explanation: He distinguishes between why a belief was selected for (evolutionary history) and why it is held (current causal explanation). Moral facts could explain why we hold a belief now without having been the cause of its evolutionary selection.

Personal & Community Reflections

• Philosophical Evolution: Mogensen has shifted from anti-realism (moral error theory) to non-naturalist realism, viewing objective moral facts as fundamental but not empirically reducible.
• Virtue Ethics: He speculates that Effective Altruism might be best understood as a virtue ethics framework focused on cultivating dispositions like intellectual integrity and "scout mindset" rather than a strict consequentialist doctrine.
• New Research Directions: He notes ongoing work by Jacob Barrett and Lauren Frixell challenging the "neglectedness heuristic," suggesting it may mislead prioritization in political contexts where "tipping points" exist.
• Parfit's Influence: While Mogensen did not know Derek Parfit personally, he cites Parfit's work as a profound intellectual influence, particularly regarding exam design (favoring depth over "trick" questions).