Law, Reason and Value Talk: Alec Walen (Rutgers)

Department: Philosophy

Date and Time: February 8, 2018 | 5:00 PM-6:30 PM

Event Location: Law 3500

Event Details

Title: “Risks and Limited Aggregation”


Please join us Thursday, February 8th , at 5pm for the next meeting of Law, Reason and Value, the new colloquium in jurisprudence co-sponsored by the School of Law and the School of Humanities.  

Our speaker will be Alec Walen, Professor of Law and Philosophy at Rutgers University, who will discuss his paper “Risks and Limited Aggregation.” Please see it attached here, with introduction below.

The talk will run 5-6:30pm in Law 3500. There will also be a faculty dinner following the discussion; if interested please email Thank you!


By Alec Walen

Life is risky. Morality should give agents guidance when they are confronted with choices that involve imposing risks of harm on some to achieve benefits (often uncertain benefits) for others (even if they might be among the others). A peculiar feature of theoretical debate about how that guidance should work is that the participants seem generally to want to shoehorn all cases of risk into one of two models; it’s still one-size-fits-all. I argue here that that is a mistake. It takes both types, and a particular kind of hybrid, to handle all the cases.

The two basic models are the ex ante model and the ex post model. The ex ante model holds that claims not to suffer some harm, H, should be discounted to reflect the probability that each claimant will suffer H, where that probability is to be assessed from the point of view of the agent, A, given what she knows and can reasonably be expected to come to know by the time she should act. The ex post model is more complex. It breaks risk down into two parts: what is the probability that someone will suffer H, and how broadly distributed is that risk across a group of people. The ex ante model rolls the uncertainty regarding whether anyone will suffer harm, and if so who, into one combined form of risk that each claimant faces. The ex post model says that if A is certain that someone will suffer H, then she should credit that person, whoever he turns out to be, with a full, non-discounted claim not to suffer H. In other words, the ex post model looks past risk regarding who will suffer H, and finds risk only to the extent that it is unclear if anyone will suffer H. In that regard, the ex post model borrows from the ex ante model: insofar as A is uncertain that anyone will suffer H, she should discount the claim not to suffer H to that degree.

I will argue here that for cases in which the choice to perform a risky action can be modeled on an act being done to and for the sake of one person, the ex ante model is appropriate. For cases in which the choice involves different groups of people, whose interests are known to conflict, the ex post model is appropriate. But when scaling up to consider social policy choices involving the second kind of case repeated on a regular basis, a kind of hybrid model should be used: ex post at the base, but ex ante on the whole.

To make sense of the last position, and many of the arguments that will occur throughout the paper, one needs to accept a position know as limited aggregation. Limited aggregation is the view that when one person has a claim not to suffer a grave harm, such as death, no number of claims not to suffer a much smaller harm, such as not getting a headache, could morally outweigh it. These weaker claims are held not to be “relevant” to the stronger claim. Limited aggregation is an intermediate position between an extreme anti-aggregationist position, adopted by proponents of Scanlonian contractualism,6 according to which the number of claims never adds up to provide more moral weight than the strongest claim among them (though tie-breaking, when competing claims are otherwise balanced, is allowed), and unlimited aggregation as adopted by classical utilitarians. According to limited aggregation, if claims are close enough in strength, then they can be aggregated to produce a net stronger claim in favor of an agent performing a particular act, but if they are too disparate in strength, then no number of weaker claims can outweigh a stronger claim. To complete the example started above, while no number of claims to avoid a headache would morally outweigh a claim not to be allowed to die, a sufficient number of claims to be saved from quadriplegia would outweigh a claim not to be allowed to die