Sand's 'A Defence of the Control Principle' (Philosophia, 2021) confines the control principle to blameworthiness — CP_B: people are blameworthy only for things within their control — and defends it against moral luck views on the ground of simplicity: a simpler theory of blameworthiness is more likely to be true, and not being falsely judged blameworthy is desirable, whereas luck-inclusive rivals (of the shape MLV1–MLV3) must positively enumerate the beyond-control factors that affect blameworthiness and grow correspondingly complex. The paper itself concedes the account is schematic: what exactly is within human control is 'still debatable' (control over habits is 'hardly determinable from an external point of view'), and excusing and epistemic conditions are distinguished from blameworthiness simpliciter but never integrated into a theory. The problem is the completion this schematic defence awaits: a full control-based theory of blameworthiness — a control condition with determinate extension in the hard cases, integration with the epistemic condition and excuse-handling, and a defensible theory-choice comparison with luck-inclusive rivals — or a demonstration that no such completion can retain the simplicity advantage.

Background

The problem of moral luck originates in classic papers by Bernard Williams (1981) and Thomas Nagel (1979). Moral luck is luck that makes a difference to how praiseworthy or blameworthy a person is: two agents may be alike in everything under their control yet differ in blameworthiness because of factors beyond it. The literature distinguishes resultant luck (in how one's actions turn out — the drunk driver who kills a child versus an equally drunk driver who, by chance, does not), circumstantial luck (in the situations one faces — two judges who would both accept a bribe, of whom only one is ever offered one), and constitutive luck (in one's inborn temperament and capacities). The control principle (CP) says that we are morally assessable only to the extent that what we are assessed for depends on factors under our control; if CP holds, moral luck is impossible. Yet particular judgments — that the unlucky driver is more blameworthy — appear to manifest such luck, and the tension between the two families of judgments is the problem of moral luck (surveyed in the Stanford Encyclopedia entry by Nelkin, revised 2025). Blameworthiness is the verdictive status of deserving blame (a judgment about an agent), as distinct from the practice of blame (resenting, reproaching, punishing). An excuse is a factor that makes it inappropriate to blame someone in the ordinary proportionate way even though they are blameworthy simpliciter — as when someone fails to rescue a stranger but only at the cost of sacrificing their own child. The epistemic condition on blameworthiness is the requirement that an agent not be ignorant of the morally relevant facts (or not be culpably ignorant) for blameworthiness to arise. Sand's paper (Philosophia 49(2), 2021) makes three moves. First, it restricts CP to blameworthiness, because responsibility and moral assessment plainly extend beyond control (obligations can be incurred by chance; and our emotional responses to evil — agent-regret and remorse, for instance toward evils committed by one's own family members — are expected regardless of control), giving CP_B: 'People are blameworthy only for things within their control.' Second, it defends CP_B not by intuition but by simplicity as a guiding norm of theory choice — where a simpler (more elegant, more compactly formulable) theory is held to be more likely true — together with the value of not being falsely judged blameworthy. Third, it argues that luck-inclusive theories cannot merely negate CP_B (that would make people blameworthy for anything at all, including others' wrongs); they must positively enumerate which beyond-control factors matter. Iterating Robert Hartman's Moral Luck View (MLV) illustrates the cost: MLV1 (blameworthy for reasonably foreseeable consequences that actually occur) misjudges the Stauffenberg case — Claus von Stauffenberg's failed 1944 attempt to assassinate Hitler, after which the foreseeably continuing war and its further deaths flowed from Hitler's and other agents' acts rather than his own; MLV2 adds what one actually does; MLV3 adds the traits one actually possesses — each iteration longer and more complex than CP_B. But the paper concedes its limits: 'it is easy to imagine tricky examples to which CP_B cannot provide a definite judgment either. What exactly is within human control is still debatable and whether someone is, for instance, in control of a certain habit is hardly determinable from an external point of view' (citing McKinnon 2013 on fixing the luck/control boundary); and while excuses and epistemic conditions are distinguished from blameworthiness simpliciter, no integrated theory is given. The simplicity argument is also explicitly comparative, setting other arguments 'for the sake of argument' aside. So the defence awaits the theory it promises: a complete control-based account of blameworthiness, and a settling of whether it wins the theory-choice comparison.

Problem Statement

Complete the control-based theory of blameworthiness that Sand's defence of CP_B ('People are blameworthy only for things within their control') only schematizes, and that remains wanting after moral luck's challenge. This is a derived problem within the moral luck debate, not the general problem of moral luck itself. A complete control-based theory of blameworthiness has three jointly necessary components. (1) A control condition with a specified extension: a substantive account of what it is for something to be within an agent's control that yields determinate blameworthiness verdicts for the hard cases the source concedes it cannot decide — at minimum habit-type cases, where control over an entrenched disposition is 'hardly determinable from an external point of view' — as well as for the canonical moral luck case pairs: the equally drunk drivers with and without a victim; Judge Actual versus Judge Counterfactual, who would both accept a bribe but only one is offered one; and the Stauffenberg case of foreseeable consequences flowing from other agents' acts. (2) Integration of the subsidiary conditions: how the control condition composes with a badness condition — one can be blameworthy only for things that are in some way or another bad, and not all things within a person's control have such a negative quality — and with the epistemic condition (when and which ignorance exempts), without reintroducing the luck-sensitivity the control theory exists to exclude; and how excusing conditions are distinguished from blameworthiness simpliciter and operate downstream of it, bearing on whether and in what proportion blame is appropriate rather than on whether the agent is blameworthy simpliciter — stated so that the theory fixes, for a described case, whether the agent is blameworthy simpliciter. (3) The theory-choice comparison on which the source's defence rests: an explicit account of whether the completed theory retains its advantage in simplicity — understood with the source as the slenderness of a formulation, whose operative measure the account must itself state, defend, and apply uniformly to both theories — over its best luck-inclusive rivals (views of the MLV1–MLV3 shape), the comparison being made, per the source's own note that the demand to incorporate the subsidiary conditions 'ha[s] to be addressed to both the luck-free account and the luck-sensitive account', at matched levels of elaboration between the two theories. The common objective is a single, complete control-based theory of blameworthiness able to answer, for arbitrary described cases, who is blameworthy and why — thereby either vindicating the simplicity defence of CP_B or showing that it fails upon completion. The question is whether a completed control-based theory can decide the canonical moral luck case pairs and the habit-type cases of component (1) while retaining, in the matched-elaboration comparison against luck-inclusive rivals of the MLV1–MLV3 shape, the simplicity advantage the source claims for CP_B, or whether no completed control-based theory of this kind can do so.

References

  1. Nagel, Thomas (1979). Mortal Questions. Cambridge: Cambridge University Press (Cambridge Core online edition). DOI: 10.1017/cbo9781107341050. https://doi.org/10.1017/cbo9781107341050
  2. Williams, Bernard (1981). Moral Luck: Philosophical Papers 1973–1980. Cambridge: Cambridge University Press. DOI: 10.1017/cbo9781139165860. https://doi.org/10.1017/cbo9781139165860
  3. Nelkin, Dana K. Moral Luck. The Stanford Encyclopedia of Philosophy (substantive revision of 20 January 2025). https://plato.stanford.edu/entries/moral-luck/
  4. McKinnon, Rachel (2013). Getting Luck Properly Under Control. Metaphilosophy 44(4): 496–511. DOI: 10.1111/meta.12044. https://doi.org/10.1111/meta.12044
  5. Hartman, Robert J. (2017). In Defense of Moral Luck: Why Luck Often Affects Praiseworthiness and Blameworthiness. New York: Routledge. DOI: 10.4324/9781315232041. https://doi.org/10.4324/9781315232041
  6. Sand, Martin (2021). A Defence of the Control Principle. Philosophia 49(2): 765–775. DOI: 10.1007/s11406-020-00242-1. https://doi.org/10.1007/s11406-020-00242-1
  7. Sand, Martin, and Klenk, Michael (2023). Moral Luck and Unfair Blame. The Journal of Value Inquiry 57(4): 701–717. DOI: 10.1007/s10790-021-09856-4. https://doi.org/10.1007/s10790-021-09856-4
  8. Hartman, Robert J. (2024). Moral Luck and the Imperfect Duty to Spare Blame. Erkenntnis 91(2): 475–491 (published online 13 December 2024; print issue February 2026). DOI: 10.1007/s10670-024-00907-3. https://doi.org/10.1007/s10670-024-00907-3
  9. Heinrichs, Jan-Hendrik, and Serap Ergin Aslan (2025). Agent regret and the moral responsibility for the misuse of research results. Research Ethics 21(2): 245–266. DOI: 10.1177/17470161241272760. https://doi.org/10.1177/17470161241272760