[article]

Mechanical Ethics

1 Oct 2025

Vincent’s S-CALM model describes the situational and cognitive factors that undermine ethical behaviour. Mechanistic thinking helps explain how those factors might operate, and thus, where we might intervene on them.

I teach ethical leadership at Sandhurst. We use Dennis Vincent’s S-CALM model, which, like Kem’s more famous Ethical Triangle Model, is probably one of the better attempts at ethical decision-making frameworks I’ve come across. Vincent in particular does something most models don’t—something even Kem doesn’t do—he explicitly engages with behavioural science literature that tries to explain poor decision-making, and uses those to think about ethics.

The S-CALM acronym stands for:

• Situational influencers: situational factors that lead to unethical acts, specifically hostile environments, normalised violence, weak leadership, lack of resources, enhanced emotional states;
• Common behaviours: the cognitive biases that lead to unethical acts (conformity, deindividuation, obedience, cognitive dissonance, bystander effect, groupthink, dehumanisation, moral disengagement, social identity, and a few others);
• Accountability: which speaks to understanding different forms of responsibility;
• Leadership: which speaks to the role of leaders in preventing ethical drift; and
• Moral compass: which speaks to the values and standards that should guide behaviour.

Altogether, fairly comprehensive. Vincent draws from the classic social psychology experiments—Milgram’s shock studies, Zimbardo’s prison experiment, Asch’s conformity tests. I’m not super enthusiastic about that, myself, but he also draws from actual UK military disasters like Deepcut, Abu Ghraib, and Sergeant Blackman’s killing of a wounded Taliban fighter in Afghanistan. He doesn’t just give you abstract principles; he gives you case studies with names and faces and real consequences.

But here’s the thing. The S-CALM model is fundamentally diagnostic. It tells you what goes wrong—which situational factors corrupt judgement, which cognitive biases lead people astray. It’s excellent at helping you recognise when you’re in dangerous territory. What it doesn’t quite do is tell you how to fix it. It tells you what the problems are, but not the mechanisms by which those problems arise, and thus, not quite how to intervene.

For that, you need to think mechanistically. Lucky I just wrote an enormous paper that does just that. The ETHIC stack is my illustration of that that looks like. Not the only illustration, I should add, nor even probably the most correct. But it does show you how we need to push past identifying ingredients of ethical failure, and into identifying how to put them together in less unethical configurations.

¶The mechanistic turn in ethics

I’ve written before about the three problems of ethical education. The first is that the catastrophic examples we use—Abu Ghraib, the Stanford Prison Experiment—are terrible illustrations because they’re both rare and extraordinarily complex. The second is that “thinking it through” is almost never the answer, because our faculties of reason are particularly vulnerable to rationalising our existing intuitions rather than challenging them. The third is that most models tell us what to do, not how to do it.

Vincent’s model addresses the first two problems quite well. He doesn’t pretend ethical disasters are simple, and he explicitly recognises that deliberative processes are often slaves to intuition rather than masters of it. But the third problem—the how problem—remains.

The difference between telling people what they should do and how they can do it is the difference between prescription and mechanism. A mechanism requires three things:

1. Entities: the (relatively) stable ingredients or parts of the system
2. Activities: the things those parts do
3. Organisation: the way those things are linked together

Think of a radio. The antenna (entity) captures radiowaves (activity). The tuner (entity) filters the wave to a specific frequency (activity). The heterodyning circuit (entity) drops that frequency into the audible range (activity). The amplifier (entity) boosts the signal (activity). The speaker (entity) converts it to sound (activity). The organisation is clear: radiowaves become intelligible sound.

Why this incessant detail? Two reasons. First, without it, you’re not particularly closer to making radiowaves into something you can listen to. You know what you should have—audible sound—but not how to get it. Second, by understanding the relation between activities and entities, you can see where to “wiggle” something to change the outcome. Swap a short antenna for a longer one, and you’ll capture more waves.

The ETHIC stack is my attempt to sketch one possible account of the entities, activities, and organisations that make up ethical behaviour. It’s a five-level mechanism—or rather, a sketch of what five-level mechanism might look like:

• E: Early emotional circuitry—the ‘pattern detector’
• T: Thought-level schemas—the ‘lazy controller’
• H: Habitat (the situation)—an ‘affordance auction’¹
• I: In-group dynamics—the ‘prestige engine’
• C: Cultural scaffolding—the ‘top-down bias’

Now, these levels aren’t the only levels. In the publication I wrote this up for, I make clear that you could easily make a case for collapsing or extending them. The point wasn’t to be exhaustive, but illustrative.

I illustrate a single mechanism at each level—a lever, if you like, to be able to change behaviour, drawn from theories from cognitive neuroscience and social psychology that seem plausible. Should you dislike my layers, or my candidate theories, then by all means choose your own. The point isn’t that these are the mechanisms, but that thinking in terms of mechanisms—entities, activities, organisations—gives you something prescriptive models don’t: potential intervention points.

With that in mind, Vincent’s S-CALM model and mechanistic thinking (illustrated here by ETHIC) aren’t competing approaches. They’re complementary. Vincent identifies what goes wrong; mechanistic thinking helps explain how it might go wrong, and thus, where you might intervene. Let me show you what I mean.

¶Situational influencers through a mechanistic lens

Vincent identifies five key situational influencers that increase pressure on soldiers to act unethically:

1. Hostile environment: when there’s a lot of threat
2. Normalised violence: when our thresholds for acceptable violence are lowered
3. Weak leadership/lack of supervision: when there’s an absence of moral reference points
4. Lack of resources: when we’re lacking time, personnel, equipment, sleep, etc
5. Enhanced emotional state: when we’re emotionally compromised because of anger, rage, frustration, disgust, and so on.

I would argue, perhaps, that an enhanced emotional state is mostly just a special case of resource-scarcity, but in the main, I think Vincent has provided a really legible account of the situational factors at play in many of the real military ethical disasters of the past half century or so.

The difficulty is drawing a line from what factors were present to why they influence bad ethical behaviour. Poor sleep doesn’t always make us unethical. Many soldiers maintain their moral compass even under highly kinetic circumstances. What’s the difference between times where these things corrupt us, and the times where they don’t?

Applying a mechanistic framework is really the only answer. Let me use my ETHIC stack to illustrate how.

¶Hostile environments and normalised violence: a situational mechanism

Since the “[H]abitat”-level of my stack covers situational factors, it seems like a good place to start. In my paper, I talk about the “affordance auction” described in cognitive neuroscience.

The notion of affordances comes from the work of James Gibson. The basic idea is that the environments we’re in don’t really make us act so much as they invite us to act. A chair doesn’t make us sit on it, but it affords an opportunity to sit, so we are invited to sit on it. We aren’t however, invited to start eating it, or using it to play table tennis, or any of the other possible actions we could engage in at any given time.

That’s the essential point. We can do lots of things. Lots of behaviours. But environments invite some more than others through the affordances for action they provide.

The ‘affordance auction’ idea takes this a step further. Environments often invite more than one action at once—there are usually heaps of affordances for action present. The way the brain seems to decide which to actually carry out happen is to auction them against one another based on their salience (how obvious they are) and their utility (how useful they are versus how costly they might be).

So you get on the subway and there’s a few affordances. The chair affords the opportunity to sit. The pole in the middle of the carriage affords you the opportunity to grab it. If the carriage starts lurching about, then you need to make a decision. Maybe sitting is more useful because you can sit through the turbulence, but it’s going to be harder to get there without falling over. The pole is right there, so it’s pretty salient, and it might be less useful because your arms will get tired, but there’s less chance of you falling over—less cost. So, you grab the pole rather than stumble over to the seat. An auction.

Hostile environments, in which there are many threats, afford many more defensive and aggressive responses than calmer environments. It’s always possible to treat loud bangs as gunfire, but usually there’s not a lot of utility in doing so. Dropping to the floor during a fireworks show seems like it’s not going to do much but accumulate dirt and strange looks.

When every person can be an enemy, though, the utility is much higher. Not least because the cost of inaction—getting killed—is not really sweet. So, aggressive or defensive action wins the auction—they’re high utility.

Normalised violence is a bit more insidious. It doesn’t change the utility of actions, so much as their salience—what actions are worth considering. You wouldn’t typically have “shoot everyone dead” in your behavioural repertoire, regardless of what affordances existed in your environment. When violence is more normal, though, more violent actions come into the window of possibility. They become more salient.

In combination, aggression wins more auctions in hostile environments with normalised violence. To do anything else would require action-possibilities that are of equal utility and salience under the same conditions. Something you can only do by training those patterns of action until they become as salient and useful in the moment as aggression is.

¶Lack of resources: a cognitive control mechanism

Lack of sleep, time pressure, inadequate people or resources to do things. Even an enhanced or compromised emotional state. All of these things give you less to do more.

In cognitive neuroscience, it’s well-known that deliberate, systematic thought is expensive. It much prefers to lazily rely on intuitive and gut-responses than this more costly, if more thoughtful, process.

The basic idea of my “[T]hought-patterns”-level of my stack is the obvious extension. If we’re lacking resouces, then we have less resources to encourage the brain to engage in more thoughtful decision-making processes.

This is rather a problem, because most decision-making models ask you explicitly to “think more” under stress, for exactly this reason. If a lack of resources means you’re going to make more gut-decisions, then it’s exactly then that you should stop and think.

Sadly, the brain is under no obligation to conform to such a wish. Even if you recognise that you should be stopping and thinking, your processes of thinking are compromised by the same lack of resources that are prompting you to worry about it. You might think that you’re thinking harder, but your lazy brain is taking whatever shortcuts it can.

Vincent notes that “the human body requires six and a half hours sleep a night to regenerate and that prolonged periods of less than this reduces cognitive ability. Leaders who lack sleep make poor cognitive decisions.” Now I’m not sure how sold I am on 6.5 as a useful number. Matt Walker’s Why We Sleep might have been a good book, but his 6.5 hours, alongside much of the rest of his ‘facts’ appear to have very little factual basis. But the general point remains. This isn’t just about being tired. On something like the mechanism I’ve sketched, deliberative moral reasoning is one of the first things to go when you’re resource-depleted, because it’s so expensive to run.

It’s a simple fix on the face of it. Just give people more resources. Sadly, not really likely. If we could resource people properly, the whole thing would be a non-issue. The alternative, then, is very similar to our last section—train intuitions better. If deliberative thinking sucks under resource-scarcity, then just improve the intuitive system it falls back on.

¶Weak leadership: social and cultural mechanisms

At the “[I]n-group-dynamics”-level of my stack, I describe social mechanics that drive behaviour. Specifically, I walk through:

• Social norms: the behaviours groups promote
• Group representatives: the people who embody and illustrate those norms
• Social capital: the status/belonging/resources you gain from conforming to the norms

The three things go together, inextricably. The most critical leaders in any organisation of people are the identity leaders. The leaders people look to because they most represent what the group is about. They’re the exemplar of what that group is all about. These people, by design or by accident, set and illustrate the behavioural norms that constrain group behaviour. Everyone wants to do what these people are doing. Then, by doing these things, you start seeming like a better member of the group, and so you get rewarded for it—you belong better, gain status as a consequence, and probably are given more resources or responsibility to boot.

Now, if these leaders are shit, or are shitty people, you’re going to have a problem because the entire structure deranges around their behaviour. The norms people are trying to model, and the behaviours that get rewarded begin to drift towards whatever these identity leaders are doing.

The point, of course, being that not all leaders are equally valuable levers in a given team. The identity leaders are the ones you want to target.

To target them, I’d look to the “[C]ultural-scaffolding”-level of my stack. I point out that every organisation of people have, or inherit, a set of values. Aspirational states that the group should want to achieve. The British Army, for example, has its Values and Standards—Courage, Discipline, Respect for Others, Integrity, Loyalty, Selfless Commitment.

These values—whether explicit, as the British Army has them, or implicit, like many social groups do—act as a ‘top-down bias’ on every other mechanism that drive human behaviour. Every other level of my stack. The intuitions, the thought patterns, the affordances, the in-group dynamics.

But the strength of this bias is directly related to the visibility of the values in the behaviour of the people, and the formal structures the group builds to support those people. If your identity leaders don’t embody the values, then people aren’t going to be very biased by them, and if off-target behaviour isn’t punished but rewarded, then the values aren’t going to be reinforced.

Altogether then, get your identity leaders, and the structure they need to influence people, aligned to your values, or the values simply won’t matter.

¶Common behaviours through a mechanistic lens

Vincent goes on, after detailing his five situational factors, to identify ten common behaviours that lead to unethical actions in stressful situations. He splits them into five individual behaviours (Social Comparison/Conformity, Deindividuation, Obedience, Cognitive Dissonance, Bystander Effect) and five group behaviours (Groupthink, Moral Disengagement, Dehumanisation, and others).

If I’m being honest with you, I’m not at all excited about cognitive biases. I’ve written often about how useless they are in understanding behaviour, and what we should do instead, or what other people think we should do instead.

The essential point being that with any arbitrary list of biases, drawn from the 200+ that exist in the literature, you’re never going to really be able to prevent poor decision-making. You’ll only ever be able to apply them after the fact. There’s nothing inherent to any given bias that tells you why it’ll fuck up your day in any given situation over some other bias.

This is particularly true of social biases. Social biases are good long before they’re bad. You want people to conform and deindividuate and engage in groupthing—you’d never get anything done otherwise. The question is what makes these things become a problem, not whether they exist.

Mechanisms are the things that help us here. Since I describe these more in my various articles on this subject, including an entire series on the subject, I won’t detail them here. But much in the same way as situational factors can become risks by specific mechanisms, so too can useful biases become problematic ones.

Instead, let’s look at a case study, to see it all play out.

¶Sergeant Blackman: the situational factors stack up

In 2011, Royal Marine Sergeant Alexander Blackman shot a wounded Taliban fighter in Helmand Province. Vincent uses this as his extended case study, and maps the Telemeter Report findings onto his situational influencers:

• Hostile environment: the group were under constant threat from ongoing kinetic action; (constant threat)
• Normalised violence: they were constantly in battle, shooting or being shot at;
• Weak leadership; they were geographically dislocated from a broader command structure, trying to adapt distant directives to the situation on the ground;
• Lack of resources; they were exhausted, and had taken enough casualties to be significantly understaffed; and
• Enhanced emotional state: they were understandably pretty frustrated, angry, and dismayed by all of these things. The casualties alone were incredible.

All five of my mechanistic layers helps us understand why they executed the wounded Taliban fighter, against the Geneva Convention:

• I don’t describe it here, but the earliest layer in my stack is the emotional pattern detector. Our brain seems to detect patterns as a fundamental aspect of it’s function. Under these circumstances—constant battle against a persistant and deadly enemy—the strongest pattern would have obviously been “justified target” rather than “protected person under distant pressure to abide by conventions and protocol”;
• The resource problem (fatigue, exhaustion, emotions running hot) would have flattened the expensive deliberative thought processes that might have encouraged a different outcome;
• The situation itself would have afforded quick violence—the wounded fighter has just finished shooting at them, weapons were in hands, and the costs (getting caught) seemed very low indeed;²
• The norms of the squad were shaped by the events, with no recourse to leaders who were uncontaminated by the constant threat and danger. The only identity leaders available had gone through extraordinary violence, and produced new norms to deal with the pressures they faced. Status was more about solidarity against the enemy than anything else.
• The top-down bias of culture was still present. Blackman famously points out “I just broke the Geneva Convention” after the execution. But their effect wasn’t strong enough in the face of these other mechanisms at play to change the outcome.

A tragedy, but an overdetermined one. I play parts of the tape in every lecture on the subject I present, and it’s very clear that any other outcome would have been unlikely indeed.

The intervention would have had to come far sooner than anything that could have happened in the moment. Reinforcing convention such that it outweighed the more local operational demands, and less distance between identity leaders who weren’t contaminated by events to reinforce more tasteful (and legal) behavioural norms.

¶Outro

Vincent’s S-CALM model is a thoughtful attempt to characterise ethical behaviour. His situational influencers are concrete and recognisable. His common behaviours are grounded in real psychology and real disasters. His examples illustrate things beautifully, and his solutions—more accountability, an emphasis on leadership, and the ongoing development of a sense of right and wrong—are all excellent ones.

But describing these things doesn’t do enough to help us understand how to go about them, nor when. For that we need mechanisms, illustrated by something like the ETHIC stack. Not as a replacement, but an explanation. A guide to when a hostile environment becomes an unethical one, rather than a warning. It helps us know where to concentrate our efforts:

• Want to improve the way the brain detects patterns? Run moral case deliberation scenarios to train them in;
• Want people to use their more costly deliberative processes? Resource people adequately;
• Want to change situational affordances? Alter the environment to make the affordances you want to win out, and train them before people get there to make them more salient;
• Want to harness social dynamics? Find out which leaders are identity leaders, and help them understand what behaviours they should be illustrating;
• Want to make cultural values visible? Teach your identity leaders what they look like in practice, and build structures to help them reinforce those things in others.

That’s the promise of mechanistic thinking in ethics. Not a definitive account of ethical behaviour, but an actionable one. Ones we can test, refine, and gradually improve.

Which is, after all, the whole point of teaching this stuff in the first place.

Anthologies: Betterment, Somatic Architecture, Collective Architecture, Moral Terrain, On Ethics, On Leadership

View on main site »