Saturday, February 05, 2011

Natural Agents and Status-Quo Bias

Carolina Sartorio argues that ordinary morality is committed to a kind of status quo bias, or what she calls 'moral inertia' [pdf]. The basic idea is that we need stronger reasons to justify interfering in a process (e.g. deflecting a trolley) than to justify abstaining from such involvement. The intuitive force of this claim is illustrated in two cases:

(1) The 1/1 trolley: a trolley is going to kill X unless you switch it to the side track, where it would kill Y. All else is equal. Sartorio claims it is impermissible to pull the switch, though it would do no more harm than leaving things be.

(2) The 5/3/1 trolley: a trolley is going to kill 5 people, unless you switch it to one of two other tracks. Side track A would kill 3 people, whereas side track B would kill just 1. Sartiorio claims that it is impermissible to switch to track A, but permissible to either switch it to B or do nothing (even though this leads to more harm than switching to A).

I'd offer a different diagnosis of these cases. In case two, I think the agent has decisive reason to switch to track B. Doing nothing and letting five die is not as justified an option, though it might be excusable if the agent's psychology is such that they would be traumatized by the thought of having "killed" someone (as opposed to merely "letting" five die).

In case one, either option seems permissible to me. Sartorio points out that it would be odd for an agent to pull the switch, since there's no reason to -- we might suspect that they had some special grudge against the person on the side track -- but this at most counts towards evaluating the agent as perverse (or possibly blameworthy, if they acted from malicious motives); it doesn't show the act to be opposed by decisive reasons (or indeed any reasons at all).

To bring this out, suppose that the agent has an ingrained habit of pulling switches, such that it would require more effort on his part to "stand back" than to just let himself go and pull the switch. (I should mention that I wasn't the only person in our seminar to be thinking along these lines.) It seems to me absurd to think that he is morally required to hold himself back in this case -- what would it achieve? Someone's gonna die either way, and you've no reason to prefer it to be the first person. In this case, it seems to me that putting in the effort required to do nothing is what would be perverse.

Sartorio might accommodate this intuition (as she does in a similar case where the switch is triggered by merely breathing) by suggesting that we build the dispositions of the agent into the 'natural circumstances'. Moral inertia then supports the agent's "going with the flow" and triggering the switch. But if we go this far, where do we stop? Isn't the agent's deliberative disposition a part of nature too? So whatever the agent ends up doing, that's the 'natural' thing to happen, given that the agent himself is part of the natural process. But then the moral inertia thesis is trivialized.

A related puzzle: even if (for some mysterious reason) we should not view ourselves as part of the natural process in this way, does moral inertia also give bystanders a reason to try to tackle an interfering agent before they pull the switch? Or should we at least view other agents as part of the natural process, and so refrain from getting involved ourselves?


  1. Every time I think to myself, "surely no moral philosopher would defend that bias", I get proven wrong.

  2. Quick question about your "intuitions".

    In the first case is it permissible to turn the trolley if you are "Y". That is, are there decisive moral reasons against saving yourself at the expense of another?

    In the 5-3-1 case do you believe you are obliged to turn on the 1 even if you are that one? Are you obliged to sacrifice yourself?

  3. It may be considered relevant to the discussion that Sartorio defends a view of causation and the concept "moral responsibility", the logical untenability of which forecloses the drawing of the "permissability/impermissibility" distinction that is required to save her account (of moral inertia) from triviality . It is not logically possible to derive "impermissibility" from a consideration of the agent's motives or reasons for acting (on the assumption that if one is "responsible" for the intention with which one acts, then the act itself is "impermissible") and indeed trivialization eclipses any argument that insists upon taking into account agent intentions, motives, reasons, deliberative dispositions, etc. in order to assess the moral status i.e. the "impermissibility" vs "permissibility" of action which cannot be applied without appeal to a notion of "moral responsibility" -- and it is precisely this notion (of causation as the vehicle of transmission of "moral responsibility") -- to which C.S. must appeal -- that is untenable.

  4. The conclusions in the scenarios rest on two assumptions:

    1) Acts of commission require more justification than acts of omission.

    2) Human lives have an equal value such that saving more people is preferable to saving fewer people, irrespective of the age, status or other characteristics of those people.

    I think you could get a lot of agreement on those points, although it would by no means be universal. However, I think the issue is far more complicated than that. For instance, I think you might find a general sense that an act of commission in which the people who die do so as a side effect of saving others, is viewed as far more morally justifiable than an act of commission where those who die are deliberately killed in order to save others.

    For instance, I expect that throwing a switch to change a train from a track with 5 people on it, to a track with 1 person it, would be far more acceptable to most people than killing 1 person because you know that their organs will save the lives of 5 people requiring transplants.

  5. Tomkow - whether you're permitted (or even required) to favour yourself is a separate issue from that being discussed here, though I personally lean towards an impartial view of reasons.

    Harrell - I find your comment difficult to parse. (More sentence breaks might help!)

    Paul - right, it's a separate question whether harms intended are (in some relevant sense) "worse" than harms merely foreseen. The "loop" version of the trolley case provides a nice counterexample there. But for this post, I'm just addressing Sartorio's arguments concerning moral inertia (or commission/omission, as you put it).

    Two clarifications, then:
    (1) The claim that commission requires more justification than omission isn't meant to be an assumption, but rather is (I take it) the conclusion Sartorio draws from the cases discussed. (Of course, insofar as one finds my alternative diagnoses plausible, that undermines Sartorio's case for this conclusion.)

    (2) Your second assumption is unnecessary. The point of the "all else equal" clause is to put aside such issues. Just imagine that the people are all exactly similar in respect of age, family status, and any other characteristics that might be relevant to how bad it would be if they died.

  6. Put another way, if the agent's volitional disposition is not self-determined, but is a function of "natural circumstances" i.e. of causation, what does it mean to say that an action is "impermissible" (or "permissible" for that matter)? If dispositions to act are not constitutive features of -- are not determined by -- the "natural process" what accounts for their existence? What logical justification could be offered to corroborate the claim that commission requires more justification than omission? As Sartorio schematizes it, "moral inertia" (whether an act of "commission" or "omission") reduces to "going with the flow" -- the thesis therefore cannot explain anything since it is just a tautology.

  7. The question then is whether Sartorio's claim (that "commission" requires more justification than "omission") is grounded in the assumption that acts of "commission" are more acceptable (less justifiable) because they are deemed to be the result of "moral responsibility". Expressed more concisely, is Sartorio's interpretation of acts of commission predicated on the (mistaken) assumption of a causal basis for "moral responsibility"? Indeed, why make a distinction between "commission" and "omission" anyway? The fact that the "moral responsibility" thesis is logically incoherent (as explained by Galen Strawson) implies that the division drawn between acts of "commission" and "omission" is trivial since every act performed is a necessary one.

  8. You seem to be getting the current topic mixed up with skeptical worries about free will / moral responsibility. For the record, I think that agents can be morally responsible without being original causes, and as far as I can see this moderate compatibilist view is, er, compatible with everything Sartorio says here too. It's a different issue.

    'why make a distinction between "commission" and "omission" anyway?'

    As I said in response to Paul, Sartorio can be read as arguing for the moral significance of moral inertia on the grounds that it's the best way to make sense of what seem to be the correct moral verdicts about the two trolley cases discussed. (I then offer an alternative account.) It's an argument from cases, not top-down theorizing.

  9. Richard you say: "...whether you're permitted (or even required) to favour yourself is a separate issue".

    Because you'd rather set it aside does not make it a "seperate issue". Am I to infer from this that you believe that you think that while ceteris paribus it is perfectly permissible to sacrifice A to save B where A and B are perfect strangers to you; you think it may be impermissible to sacrifice a stranger to save yourself?

  10. Tomkow, no, you aren't to infer that - Richard said he favours an impartial view, in which case he would give the same response whether it is him on the trolley or a stranger.

    What's more, even if you could infer that, which you can't, that wouldn't change the fact that it definitely IS a separate issue.

    Also you misspelled separate, which I wouldn't normally remark on except that the phrase which you are quoting has the correct spelling, and also there is a spellchecker here which underlines "seperate" and suggests you replace it with "separate."

    In any case, I have the same confusion Nick expressed. The status-quo bias is a fairly obvious case of where our moral intuition is irrational and wrong - why is anyone talking as if it might reflect an actual moral truth?

    Moral intuition is as flawed and irrational as every other form of intuition, so why pay attention to it?

    I don't buy the argument that "intuition's all we've got to go on in moral philosophy." And even if one somehow proves that true, that doesn't make intuition a reliable indicator of truth - all it does is prove that we don't have any reliable indicator of truth in moral philosophy.

  11. Richard,

    I think your argument is absolutely correct though I don't see where you explain that agents can be "morally responsible" given that we have to face the fact that "a person cannot be ultimately responsible for their initial state of being" i.e., cannot be the ultimate originator of these states. Which brings me to Sartorio who makes the mistake of thinking that disjunctive causation can ground "moral responsibility" ("Causation and Responsibility", Philosophical Compass, 2007). Again, the question is -- is the moral verdict being drawn in cases of acts of "commission" from the assumption of a causal basis for "responsibility"? is the agent who performs an act of "commission" or intervention more "blameworthy? The thrust of your alternative account is established by Sartorio herself: "When do we leave things unchanged and when do we change them? . . . what counts as an intervention and what counts as a failure to intervene?" Clearly, the line cannot be non-arbitrarily drawn between the causally relevant or irrelevant, i.e., between "default" states of the world and "departures" from these states. If we exclude the agents "own antecedent behavior and dispositions" what is left? if the agent does not interfere for a reason, why does he interfere? of what significance is the intervening/failing to intervene distinction if "moral responsibility" is incoherent? is an act of commission more clearly wrong because the "pre-existing process" or "pre-set path" or "default state" of causation is diverted by an act of "moral responsibility"? Doesn't the incoherence of "moral responsibility" as a concept raise the question of whether "moral inertia" -- rather than arising from the conditions of a "specific class of interventions" -- is simply inherent in the deliberative process? I don't have to make these objections; Sartorio herself concedes that her account of inertia can only be conceptualized in relation to the (foundationally incontrovertible, arbitrarily privileged) assumption that causation "obtains between events in the world independently of our subjective or inter-subjective experience" i.e., that there is a clear line between causal continuity and discontinuity. Hence my question -- what logical justification could there be for the claim that "deflecting an old threat isn't causing the outcome, but launching a new threat is" or that reserving my liver "wouldn't be diverting a pre-existent train of events"? After all, my mental states belong to a causal sequence -- they determine the "default" state of the world -- so what determines if I interfere, if not my "antecedent behavior, dispositions, etc."? The thesis founders on the question of antecedent-consequent relationships -- the arbitrariness involved in asserting causal continuity at some points and causal discontinuity at others trivializes it.

  12. A basic status-quo bias would probably have evolved, no? We often don't know what's really going on, or why (not in much detail); etc.

    But the thing about Trolley cases is that we're told a lot of stuff that we wouldn't know if we were in such a situation, and then asked what we would do in such a situation. So they can hardly prove much, and are a bit too confused even to work as intuition-pumps.

    So I agree with you, Richard. From what we're told, nothing else is morally correct, as a matter of simple arithmetic. Intuitions that support Sartorio's view must therefore come from similar situations that aren't the ones described.

  13. I agree with Sartorio in the 1/1 case. One ought to abstain from pulling the switch simply because one would otherwise make oneself a contributor to the death of someone without good reason even at the cost of mustering additional effort. One would be complicit in the death in some sense in pulling the switch. In the 5/3/1 case, one is also a contributor in the causal chain in causing someone's death but it was for the right reasons and IMO justified. I also think it is not only justified and permissible but pace Sartorio, obligatory in that case to switch the track so that 1 person instead of 5 is killed.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)