Skip to content

Oxford Uehiro Prize in Practical Ethics: Against Making a Difference

  • by

This essay was the winning entry in the undergraduate category of the 7th Annual Oxford Uehiro Prize in Practical Ethics.

Written by University of Oxford student Imogen Rivers 

I. The Complacency Argument

Some of the most serious wrongs are produced collectively. Can individuals bear moral responsibility for such outcomes? Suggestively, it’s been argued that “all who participate by their actions in processes that produce injustice [e.g. “sweatshop” labour] share responsibility for its remedy”;[1] “citizens… bear partial responsibility for the election outcome. Even if an individual’s vote is not decisive for a given candidate’s victory”;[2] “those who contribute to climate change… (by using… excessive… fossil fuels or by deforestation) should make amends”.[3]

However there’s a prevalent defence: it makes no (significant) difference if I do it. For example, “global warming will still occur even if I do not drive [my “gas-guzzler”] just for fun”;[4] “my polluting doesn’t actually harm anyone, since it doesn’t make a difference to anyone’s health”;[5] “why [should citizens] vote even if… each particular vote does not make a difference to the outcome”?;[6] “British officials… dismiss suggestions that our role on the ground in Saudi Arabia makes any difference [to targeting Yemeni civilians]”.[7] 

This common defence seems to go like this:

P1. P is morally responsible for outcome O only if, for some action φ, P’s φ-ing was a cause of

P2. P’s φ-ing was a cause of O only if P’s φ-ing made a difference to O’s occurrence.

P3. In situation S, P’s φ-ing did not make a difference to O’s occurrence.

C. Therefore, in S, for action φ, P is not morally responsible for O.

Call this “The Complacency Argument”.

P1 seems plausible. To illustrate:

(Faulty button) Athena and Balthazar press their respective buttons, reasonably believing that each button-pressing will deliver Caspar a lethal electric shock. Balthazar’s button is broken but Athena’s button works; Caspar dies immediately.

Intuitively, only Athena is morally responsible for Caspar’s death (although we condemn Balthazar’s attempt) since Balthazar’s button-pressing failed to be a cause of Caspar’s death.[8] Moreover, P3 is a reasonable empirical observation in situations including “fast fashion”, global warming, voting and competitive-market arms-trading. But does not making a difference really make a difference to individual responsibility?

In this essay I defend an account of causation according to which difference-making is not necessary for causation. So P2 is false and the Complacency Argument unsound. I argue that we do in fact have moral responsibility for outcomes to which we don’t make any difference.

II: Against making a difference

Why isn’t Balthazar’s button-pressing a cause of Caspar’s death in faulty button? The difference-making intuition: Balthazar’s button-pressing didn’t make a difference to Caspar’s death, and difference-making is necessary for causation. Here’s one way to analyse this intuition:

Causation (Counterfactual Dependence): c is a cause of e iff (i) if c hadn’t occurred, then e wouldn’t have occurred, and (ii) if c had occurred, then e would have occurred.

But counterfactual dependence, unlike causation, isn’t transitive. For example:

(Early Pre-emption) Athena and Balthazar plan to kill Caspar. Both lethal buttons function. Athena presses first, killing Caspar. Seeing this, Balthazar doesn’t press his. But had Athena’s button-pressing not occurred, Balthazar’s button-pressing would have killed Caspar instead.

Whilst Athena’s button-pressing (c) is intuitively a cause of Caspar’s death (e), e does not counterfactually depend upon c. A straightforward remedy: extend counterfactual dependence to a transitive relation by taking the ancestral.

Causal Dependence (Counterfactual Dependence): e depends causally on c iff (i) if c hadn’t occurred, e wouldn’t have occurred, and (ii) if c had occurred, e would have occurred.

Causation (Chain of Counterfactual Dependence): Let c, d1, d2, …, e be a causal chain, viz., a finite sequence of actual distinct events in stepwise causal dependence. c is a cause of e iff there’s a causal chain leading from c to e.[9]

Now c does count as a cause of e: an intermediate event d (e.g. an electric signal running from Athena’s button to Caspar’s brain) depends causally on c, and e depends causally on d.

What about when there’s no intermediate event?

(Late Pre-emption) Athena and Balthazar both press their buttons; Athena’s signal arrives first, killing Caspar an instant before Balthazar’s can.

Intuitively, (only) Athena’s button-pressing is a cause of Caspar’s death. But the process running from the pre-empted cause is only cut short by the main process running to completion. So there’s no causal chain from Athena’s button-pressing to Caspar’s death: had any of the events constituting the journey of Athena’s signal to Caspar’s brain not occurred, Caspar’s death would have occurred anyway thanks to Balthazar’s signal. The analysis under-extends: there’s causation without a causal chain.

Notice: if Athena’s button-pressing hadn’t occurred, Caspar’s death would have been slightly later. So perhaps we should construe events as fragile, where an event is fragile if, or to the extent that, it couldn’t have occurred differently in time/manner/place. But the analysis now over-extends. Suppose that my sending you a poisoned cake was a cause of your death. Intuitively, if you die slightly later—because, say, the postman paused to tie his shoelaces before delivering it—then the postman’s shoelace-tying is a cause of your death occurring at the particular time that it did. Not so if we construe events as fragile: the postman’s shoelace-tying becomes a cause of your death. A loophole: events are fragile in some cases. Which cases though? Without independent motivation, this looks ad hoc and unprincipled.

Worse, fragility can’t save difference-making. Consider:

(Overdetermination) Athena and Balthazar press their buttons simultaneously. The lethal shock switches on, identical to the shock from Athena or Balthazar individually. Caspar dies in exactly the same time/manner/place as if Athena or Balthazar individually had pressed.

So neither Athena’s nor Balthazar’s button-pressing made any difference to Caspar’s death or to its time/manner/place. But both are intuitively causes. So difference-making is not necessary for causation—P2 is false. What, then, is the correct analysis of causation?


III. From difference-making to effect-producing

Causes don’t always make a difference to their effects; perhaps they always produce their effects. The difference-making analysis of “c is a cause of e” captures the intuition that c’s occurrence was necessary and sufficient[10] for e’s occurrence: e couldn’t have occurred without c, and e couldn’t but have occurred with c. This intuition is too strong; the effect-producing account refines it:

  • Weaken sufficiency: c needn’t be sufficient for e, but merely one of a plurality of actual events collectively sufficient for e;
  • Weaken necessity: c needn’t be necessary for e, but merely necessary for the sufficiency of a plurality of events collectively sufficient for e.

The Minimal Sufficiency (MS) analysis congenially captures this effect-producing intuition:[11]

  • A plurality of events c1, …, cn collectively caused (was minimally sufficient for) e iff:
    1. The instantiation of the plurality (i.e. the occurrence of c1, …, cn) was sufficient for e’s occurrence, and
    2. The instantiation of no proper sub-plurality was sufficient for e’s occurrence.
  • c is a cause of e iff c is one of a plurality minimally sufficient for e.

This effect-producing account succeeds where difference-making fails. Consider early pre-emption. There’s an instantiated plurality of events including Athena’s button-pressing which is minimally sufficient for Caspar’s death. But there’s no instantiated plurality including Balthazar’s button-pressing because Balthazar doesn’t press. So Athena’s button-pressing—not Balthazar’s—is a cause of Caspar’s death.

Consider late pre-emption. There’s a plurality of events (e.g. the instantiation of all events on the signal’s path from Athena’s button to Caspar’s brain) which is minimally sufficient for Caspar’s death—there must be at least one, because the death happens.

One such minimally sufficient plurality contains the following events:

  • Athena’s button-pressing at time t.
  • Athena’s electric signal being at position P at t0+1.
  • Caspar’s living brain being at P at t0+1.

These events are all instantiated, so Athena’s button-pressing is a cause of Caspar’s death. What about Balthazar’s button-pressing? Balthazar’s signal reaches P a split-second after Athena’s (say, at t0+1.1). So any plurality of events minimally sufficient for Caspar’s death which contains Balthazar’s button-pressing includes:

  • Balthazar’s button-pressing at time t.
  • Balthazar’s electric signal being at position P at t0+1.1.
  • Caspar’s living brain being at P at t0+1.1.

But this last event is not instantiated: Caspar is already dead at t0+1.1. So Balthazar’s button-pressing isn’t part of any instantiated plurality minimally sufficient for Caspar’s death.[12]

Here’s a last illustration, namely overdetermination. Now there are two pluralities of events minimally sufficient for Caspar’s death, containing Athena’s and Balthazar’s button-pressings respectively. One plurality contains:

  • Athena’s button-pressing at time t.
  • Athena’s electric signal being at position P at t0+1.
  • Caspar’s living brain being at P at t0+1.

The other plurality contains:

  • Balthazar’s button-pressing at time t.
  • Balthazar’s electric signal being at position P at t0+1.
  • Caspar’s living brain being at P at t0+1.

All of these events are instantiated since the shocks arrive simultaneously. So Athena’s and Balthazar’s button-pressings are each a cause of Caspar’s death, which is intuitively right.

IV: The counterfactual test in law

Here’s a pertinent objection: if causation isn’t difference-making, why is the counterfactual test successful in law? My reply is twofold. First, the effect-producing analysis supplies an error-theory for the counterfactual test.[13] Consider: if N is necessary for A, and something is A, then it’s a sufficient condition for x to be A that nothing other than x be N. Now, being in a minimally sufficient plurality is necessary for being a cause. Assume charitably that passing the counterfactual test is sufficient for causation. In faulty button, Athena’s button-pressing is a cause of Caspar’s death. Therefore, if we remove Athena’s button-pressing from the instantiated plurality minimally sufficient for Caspar’s death then that plurality is no longer sufficient for the death. Since there’s only one such minimally sufficient plurality (no pre-emption/overdetermination), the death can’t occur but-for Athena’s button-pressing. So the counterfactual test returns the correct result in simple causal structures and the MS analysis explains why.

What about complex causal structures? In pre-emption/overdetermination cases, courts explicitly eschew the counterfactual test.[14] Moreover, the Restatement (Second) embraces the “substantial factor” causation-test: something is a substantial factor “if two forces are actively operating… and each of itself is sufficient to bring about harm”.[15] Therefore, tort-feasors who are individually unnecessary for the result—viz., make no difference to it—may be causes in virtue of being individually minimally sufficient for harm.

V: The Motivation Argument

Suppose I joyride my “gas-guzzler”. I drive freely—uncoerced—and reasonably foresee that my emissions contribute to global warming. Moreover my driving is a cause of global warming: it’s in a plurality of events (including industrial contributions) minimally sufficient for global temperature-increase. Consider, therefore, “The Motivation Argument”:

P1’. P bears moral responsibility for outcome O iff, for action φ:

  1. P φ’s freely (e.g. uncoerced);
  2. P’s φ-ing was culpable with respect to O (e.g. P could reasonably foresee O);
  3. P’s φ-ing was a cause of O.[16]

P2’. P’s φ-ing was a cause of O iff P’s φ-ing was in a plurality of events minimally sufficient for O.

P3’. My free, culpable joyriding was in a plurality of events minimally sufficient for global warming.

C. Therefore I bear moral responsibility for global warming.

Analogously, suppose that I freely purchase “sweatshop”-produced clothing. I could reasonably foresee that I’m financing continued exploitation. Moreover, my purchase is in a plurality—e.g. comprising the finite number of purchases required for “sweatshop”-solvency—minimally sufficient for continued exploitation. So—without making a difference to (overdetermined) exploitation—I bear responsibility for its continuation. Likewise, my vote contributes to causing—albeit overdetermined—election outcomes. Finally, officials who sell arms to civilian-targeting states—though pre-empting alternative sellers in a competitive market—freely, culpably and causally contribute to war crimes.

Presumably, degrees of responsibility scale with causal contribution. So individual responsibility may be small or better discharged through institutional change than through abjuring contribution. Nevertheless, contra the Complacency Argument and per the Motivation Argument, we bear responsibility for harms to which we contribute, regardless of whether we made a difference. In sum, I hope to generate practical reasoning: we should remedy, combat and protest harms which we participate in producing, from “fast fashion” to the arms trade, voting to climate change.


[1] Young 2006: 125

[2] Goldman 1999: 217

[3] Caney 2010: 205

[4] Sinnott-Armstrong 2010: 334

[5] Kagan 2011: 109

[6] Sartorio 2004: 331

[7] Merat 2019

[8] Of course, that P’s φ-ing should be a cause of O is not sufficient for P to be morally responsible with respect to O. I’ll come back to this. Note that Sartorio (2004) contests P1, arguing that P can be morally responsible for O without being a cause of it. However her case depends on a controversial account of the causal structure of overdetermined omissions. In charity to the Complacency Argument, I assume P1.

[9] Lewis 1987: 167

[10] From now, the “necessity”/“sufficiency” of events means “necessity/sufficiency, given the laws and circumstances”.

[11] Kaiserman forthcoming—a; Beebee and Kaiserman 2019: 366-367; cf. Wright 2013; Strevens 2007.

[12]This effect-producing intuition underpins Lewis’ quasi-dependences (1987: 206): Balthazar’s button-pressing doesn’t affect the “intrinsic character” of the instantiated causal process which leads to Caspar’s death.

[13] Cf. Kment 2010: 91-96

[14] Given “multiple wrongdoers, the court may treat wrongful conduct as having… causal connection with the loss… though the… ‘but for’ test is not satisfied.” (Lord Nicholls in Kuwait Airways Corp v Iraqi Airways Co (No.6), 2002, 2 A.C. 883 at [74]);

  1. Restatements of Torts 1934, 1965: §431; Moore 2007: 407; Stapleton 2013: 60

[15] Restatement (Second) of Torts 1965: §432(2)

[16] See Kaiserman forthcoming—a, b for discussion of this analysis of moral responsibility.



Caney, S., 2010, “Climate Change and the Duties of the Advantaged” in Critical Review of International Social and Political Philosophy 13(1): 203-228

Goldman, A., 1999, “Why Citizens should Vote: A Causal Responsibility Approach” in Social Philosophy and Policy 16(2): 201-217

Kagan, S., 2011, “Do I Make a Difference?” in Philosophy & Public Affairs 39(2):105-141

Kaiserman, A., forthcoming—a, “Against Accomplice Liability”, forthcoming in Gardner, J., Green, L., and Leiter, B. (eds.), Oxford Studies in Philosophy of Law: Volume 4, Oxford OUP

Kaiserman, A., forthcoming—b, “Responsibility and the ‘Pie Fallacy’” in Philosophical Studies

Kaiserman, A., and Beebee, H., 2019, “Causal Contribution in War” in Journal of Applied Philosophy 37(3): 364-377

Kment, B., 2010, “Causation: Determination and Difference-Making” in Noûs 44: 80-111

Lewis, D., 1987, “Causation” in Philosophical Papers Volume II: 159-213

Merat, A., 2019, The Guardian:

Moore, M., 2007, “Causing, Aiding, and the Superfluity of Accomplice Liability” in University of Pennsylvania Law Review 156: 395-452

Sartorio, C., 2004, “How to be responsible for something without causing it” in Philosophical Perspectives 18(1) Ethics: 315-337

Sinnott-Armstrong, W., “It’s Not My Fault: Global Warming and Individual Moral Obligations” in Gardiner, S. et al. (eds.) Climate Ethics: Essential Readings: 332-346

Stapleton, J., 2013, “Unnecessary Causes” in Law Quarterly Review 129 January 2013: 39-65

Strevens, M., 2007, “Mackie Remixed”, in Campbell, J. K., O’Rourke, M., and Silverstein, H. S. (eds.) Causation and Explanation. MIT Press: 4-93

Wright, R., 2013, “The NESS Account of Natural Causation: A Response to Criticisms” in Stepanians, M. and Kahmen., B. (eds.) Critical Essays on “Causation and Responsibility”. De Gruyter: 13-66

Young, I., 2006, “Responsibility and Global Justice: A Social Connection Model” in Social Philosophy and Policy 23(1): 102-130



Share on