When Bombing Becomes “Servicing the Target”

In his 1999 paper Albert Bandura gives an example that no explanation can dissolve, because it explains nothing and only names. Bombing raids, he writes, are described as servicing the target. And the civilians killed by the bombs are turned, in language, into collateral damage.

It is the same operation, twice over. An event with someone acting and someone killed is translated into an event with neither. Something is serviced; something is incurred. Whoever speaks the sentence has not given up his moral standards — he has merely put them out of reach of this act.

Bandura draws no excuse from this. His paper makes the limit plain at the very start:

That is precisely Bandura’s thesis. The most comfortable explanation for cruelty is a cruel person; it keeps evil at a distance and makes character the cause. Bandura asks the less comfortable question: how can someone with intact moral standards act without applying them to himself in that moment?

His answer starts from how those standards work. In Bandura’s theory, morality operates not as a tablet of law but as ongoing self-regulation: whoever intends something measures the act against his own standards and sanctions himself — through shame, self-respect, self-reproach. Where these self-sanctions take hold, no supervision is needed. Where they fail to appear, cruelty becomes possible without anyone having to change his convictions. Bandura calls the process the selective deactivation of self-sanctions. The standard remains in place. What is severed is only the link between it and the particular deed.

Eight ways round your own standard

Bandura lists eight such cut points, and they sit at quite different places in the chain between person, act and consequence.

Three of them attach to the act itself. Moral justification reinterprets the deed: as protection, as duty, as necessary sacrifice for a higher good. Sanitising language — Bandura’s euphemistic labeling — does the same thing without argument, by word choice alone; servicing the target is its model case. Advantageous comparison sets the deed beside a worse one and thereby makes it look measured: what we are doing is nothing against what the others are doing.

Two attach to authorship. Displacement of responsibility passes it upward, to orders, regulations, superiors. Diffusion of responsibility spreads it sideways. Here Bandura is very concrete: a sense of responsibility can be dispersed by division of labour, because most undertakings require the services of many people, each of whom performs a sub-task that looks harmless taken on its own. One person draws up lists, a second transports, a third approves, a fourth carries out. Each sees a small, technically named step. The outcome has no single author and yet many contributors.

The last three attach to the consequence and the victim. Distortion of consequences keeps the harm small, denies it, or makes sure nobody sees it. Dehumanisation strips the victim of the status that compels consideration. And attribution of blame reverses the relation: the injured party provoked it, deserved it, brought it on himself.

What makes these eight mechanisms interesting as a group is how inconspicuous they are. None of them requires anyone to set his morality aside. On the contrary: most of them work only because the standard is still there and still wants to be satisfied.

What the experiments show — and what they do not

Bandura examined two of these mechanisms in the laboratory. In 1975, with Bill Underwood and Michael Fromson, he published a study whose title already contains the programme: the disinhibition of aggression through diffusion of responsibility and through dehumanisation of victims. The list later became a measurable trait — in 1996 Bandura, Barbaranelli, Caprara and Pastorelli presented an instrument that records moral disengagement as an individually varying disposition and tests how it relates to conduct.

The most prominent witness in the matter, however, is somebody else’s: Stanley Milgram’s obedience experiments. They display a laboratory form of displaced responsibility, and they are regularly cited as proof that almost anyone will do almost anything when an authority orders it. Bandura himself rejects that reading. What is rarely noted about Milgram, he writes, is that most people refuse cruel conduct once the situation is personalised. Obedience in those experiments was no constant: it varied with the legitimacy of the authority, with its physical proximity, and with the visibility of the victim.

Stephen Reicher and Alexander Haslam sharpened the point in 2012. What Milgram’s and Zimbardo’s studies really showed, they argued, was not blind automatism but a situation in which scientific authority, gradual escalation and shared responsibility made carrying on defensible. Participants doubted, objected, and stopped at different rates depending on the arrangement. Whoever reads obedience as a constant of nature overlooks exactly the adjusting screws by which something can be changed.

How far the description carries Bandura himself demonstrated in 2016. His book leaves the battlefield and walks through whole industries: the arms trade, the tobacco business, the handling of environmental destruction. The mechanisms there are the same, they merely take other names — justification becomes securing the site, sanitising language becomes technical vocabulary, distortion of consequences becomes an expert report. The finding is more uncomfortable than any story of war, because it no longer needs a state of exception: the same anatomy is at work in institutions that are entirely lawful and whose participants rightly consider themselves decent.

That also says what moral disengagement is not: an offender profile. The same mechanisms appear in war, in bullying, in companies, in politics and in online groups, and almost anyone can use them. It is precisely this generality that makes the concept practical. Prevention then looks in the first instance not for bad people but for bad translations between act and consequence.

The antidote is connection

Eight cut points yield eight points of attack, and they all come to the same thing: restoring the severed link.

Responsibility must be named rather than diffuse — somebody signs, and signs with a name. Consequences must be visible and named in ordinary language; “servicing the target” is the phrasing that has to disappear before anything else can change. Those affected need a face, a voice and agency, because dehumanisation fails where someone can answer back. Dissent must not count as betrayal of the group. And tasks must be cut so that nobody sees only a morally neutral-sounding fraction — division of labour is the most powerful of the eight tools, because it works entirely without ill intent.

What remains

What is secure is the description, not the prediction. Bandura’s eight mechanisms are a well-documented anatomy of what people do before and while they cause harm; they do not predict who will do it. The classic piece of evidence remains contested: what Milgram’s experiments prove about obedience is still being negotiated, and Bandura himself stands with those who take the finding to be conditional rather than universal.

In everyday life the thing is easiest to spot in the grammar. Where actions appear without anyone acting, where consequences are incurred rather than inflicted, where a responsibility dissolves into a description of a procedure, a cut point is in operation. The test is cheap: state the same event with a subject, a verb and an object, all three named. What survives that sentence was a decision. What does not survive it was a translation. Evil is not banal because it is small. It can be banally organised — into forms, euphemisms and jurisdictions that let nobody speak the whole sentence.