Agentism and the limits of control
A theory from the 1960’s can help us understand contemporary changes in work culture and the perception of our social usefulness.
This article is part of a two-episode series on agentism. Read the second part here.
The phrase “agentic state” is one of the theoretical outcomes of the famous “Milgram experiments” on obedience and authority. In the summer of 1961, Yale psychologist Stanley Milgram investigated how far everyday men and women would apply a harmful procedure to a peer sitting on the other side of a wooden panel. The harmful procedure in question, which involved electric shocks of gradually increasing intensity, was essentially a form of torture. The experiment concluded, surprisingly, that a vast majority of the volunteers didn’t stop applying the shocks even when it became clear that they were inflicting unbearable pain, if they stopped at all. This temporary suspension of reason, and to a certain extent, moral concepts, that seemed to occur when a person is following orders or a well-defined procedure, was dubbed agentism. The idea was that in such conditions, humans would stop behaving as self-determining entitites, and become agents of a system, entirely devoid of critical reason.
Although the extreme conditions of the Milgram experiment were necessary to prove the reality of agentism as a social phenomenon, its pattern can be easily detected in countless real-life situations. Agentism is not only an abstract term in psychological parlance, it is a concept that manifests itself in our daily lives. Sentences like “I’m sorry, Sir, I can’t do that”, “that’s company policy”, or even “it’s above my pay grade” are all clues of agentic behaviour. The people who pronounce them are acting under orders: in the first case, the orders prevent the agent from doing something, in the second, they are being used as justification for one’s actions, and in the last case, orders justify refusal to take a decision. This last case is perhaps the most emblematic of the three: although orders are not present, it is specifically this absence of any prescribed action that is preventing the agent to act. In all these cases, it is tempting, although wrong, to impute this behaviour on the agents themselves. All three agents are acting within a well-defined framework of guidelines and rules that have been written for them, that they only chose to follow as part of their occupational practice. And there, we are touching on the heart of the issue: all three of these examples occur in professional situations, where a free element, the client or colleague, is confronted with an agent, the clerk or customer representative.
Better safe than sorry
In all three cases, organisational agentism is a response to perceived issues of safety: what would happen if the employee performs an action that could carry risks for themselves, the client or the company? How can we mitigate unpredictable consequences of employees’ on-the-spot reactions to new challenges? As organisational entities grow in size, rules and regulatory frameworks are a convenient way of ensuring that the same standards are applied in every branch or office, that customers are treated exactly the same in Paris and London, New York and Tokyo.
The objective of an agentic system is therefore to reduce unpredictability, both for the agents themselves, the upper layers of management who must be able to rely on constant standards, and the multiple points of contact between the agentic system and the outside world. This, in general terms, is not a bad thing. Mitigating uncertainty is crucial for a large system, which cannot be expected to react as swiftly to unpredictable events as the individual elements that constitute it. In a professional setting, agentism can provide a certain flavour of professionalism to a process where all actions are executed flawlessly and by the book. And this is true even outside of the workplace ethos: standards and frameworks do make things go smoother — and faster, up to a certain point. After all, it was the standardisation of mechanical parts and the strict division of labour that enabled great technological leaps such as the first and second industrial revolutions.
Now that it is certain that agentism and standardisation are, in essence, good things, we must examine the dosage at which we wish to apply these factors of cohesion and reliance in our fast-changing societies. If little doses of poison can act as a medicine, we should remember that medicine taken in excess can be just as lethal as the worst of poisons : too much cohesion becomes rigidity, too much reproducibility leads to conformity and uniformity.
The imitation game
A rule provides guidelines, it is designed to channel actions and thought processes through a well-thought out, predetermined workflow. If it is well designed, a rule will rarely, or even never be wrong. Let’s take the example of a store whose manager stands by only one rule, “be polite to the client, make sure they enjoy the best service during their visit”. Everyone, manager, clerk and customer would be better off following that rule, since it maximises benefit for all: the customer feels happier and spends more, the manager makes more, and the clerk gets paid more and gets to spend a nicer day at work. However, if one day the manager notices that the clerk is not welcoming every client when they enter the store, they might decide to enforce more precise rules, such as “Smile at the customer and ask them if there is anything you can help them with today, smile again when they leave and wish them goodbye”. Although theoretically, this sounds like a more refined version of the previous incentive, it might prove disappointingly ineffective. The clerk might feel pressured and will certainly not think so highly about their job anymore, the client experience will feel artificial and sales might decrease. But if we focus on the agent (the store clerk, in this case), they suffer a more subtle consequence: their understanding of the underlying issue has changed. “Be polite” became “smile and say hello”, and through that process of authoritative explanation, the manager has denied the clerk the understanding of politeness and good manners. Put less antagonistically, he has made it superfluous for the clerk to know or understand what these notions mean. The focus of the rule is not anymore on the relationship that links the three parties involved, but on the rule that was supposed to improve it: rule-making easily suffers from a loss of focus that drives it away from the matter at hand to the incidental details of its own application. This loss of focus makes it harder for agents to actually understand what the rules are about or why they are there in the first place.
But this phenomenon leads to another, more worrying fallacy: the concept whose meaning is taken away by a rule tends to get associated with the rule itself. In our example, the concept of politeness is stripped of its personal definition and replaced by a reproducible behaviour that emulates it. If this behaviour is successful, the agent might adopt it in everyday situations, even their own personal life. Bystanders and observers who witness the use of this guideline might then propagate it in their own personal circles: they need only to be fooled by the appearance of professionalism and efficiency that is given off by the agent. We might call this process over-generalisation: a quid pro quo that allows a regulated process to escape the small circumscribed universe it was designed for and spread into the real world. Through this chain of events, overregulation does not simply streamline isolated processes or concepts: it also redefines them, causing profound semantic and cultural mutations within the societies that host them.
Trial without error
An agent’s most desirable characteristic is obedience, and what is obedience, essentially, but a reduction of autonomy? Autonomy is the precursor of self-initiative, and self-initiative in the traditional sense is not encouraged in an agentic system. Instead, such a system promotes an incremental version of free will, where the agent is carefully experimenting with small variants of a rule. When they start understanding how this rule works, they can start experimenting more, eventually they are able to direct junior members in their dealings with this rule, which signals they are ready to move up and deal with the next level of the organisational framework. This tunnel-vision approach prompts its agents to develop their understanding of the system, not so much of the underlying dangers that justify its existence. This, again, is not a negative thing. If all elements of a system start behaving solely according to their own free will, chaos will ensue. In any organised process, some manner of order is necessary.
However, that system needs to strike a certain balance between initiative and predictability, curiosity and compliance. We mentioned that highly agentic systems could lose understanding of critical notions; similarly, a fall into agentism can lead to a loss of critical sense, curiosity and initiative. The link with curiosity is self-explanatory: if you only ever encounter the same situations in which you experience no doubt and know exactly what to do, there is no need to go beyond your obligations. At the same time, repetition has created a habit, and habits are comfortable; they dull our sense of enterprise because they never solicit it. Following a workflow does not harm an agent’s curiosity per se, it just removes all incentives for it, while at the same time discouraging the initiative that would be necessary to trigger it.
But a system’s degree of agentism is also linked to its agents’ capacity for critical sense. Since well-observed rules lead to safer situations, there is a strong incentive to extol their virtues and repress any in-depth examination. We’ve already shown that agentism creates habits, and there is no worse habit than the habit of safety. Criticising the rule means criticising the safety it provides. Rather than rethinking well-established processes, we’d rather live with the flawed ones that are already in place to protect us.
Agentic rules are not only reductions of autonomy, they also act as diluters of responsibility. For a manager, a rule is a way of controlling their employees’ responsibility: if they don’t follow a given procedure properly, they are unquestionably responsible for their mistake and can be punished accordingly. For the employee on the other hand, a rule is a way to deflect responsibility upwards, for if a mistake results from the application of a rule, then the rule is faulty and the mistake can be blamed on their superior. Both sides profit from the creation of this durable framework, the overall result, it seems, is one of increased stability. The issue however, is that a multiplication of rules threatens to weaken the very sense of responsibility in all parties involved. The fact remains that mistakes do and will happen, often quite randomly or caused by a number of factors that are out of a single person’s control. In order to be able to react to failure in a constructive way, agents must safeguard their ability to quickly assess a situation and take ownership of its consequences, acting as its de facto manager and taking responsibility for it. We might refer to this kind of responsibility as a manifest one, whereas the careful definition of roles embodies the notion of an orchestrated responsibility. However, in a context of regulatory overload, agents will instinctively look for rules or guidelines that would tell them who should feel responsible for a given situation and how they should react. In other words, they will tend to look for a culprit, someone who could be made responsible for a given unpleasant situation. This mental process is the radical opposite of the manifest kind of leadership mentioned above, the main point here again being that a desire to mitigate all possible eventualities effectively destroys the agents’ capacity to do so when and where it is needed most.
The highly agentic state is a weakened state of existence, that weakens agents’ reasoning, curiosity and sense of responsibility, three abilities that are key to universal personal development and social progress. By attempting to ensure high-levels of professionalism across all levels of society, it is instead deeply redefining the meaning of professionalism and personal ethics in ways that are questionable at best. When this state gets overused and affects a large system like a national or continental economy, the consequences can be truly devastating. More relevantly, agentic systems are a tool of peace and can be afforded only in peaceful times, because they can only sustain themselves in situations that do not require a high reactiveness. However, when tensions arise, be they internal or external, the agentic structure is put to the test. More and more unforeseen circumstances arise, and soon unresolved situations start to accumulate. By that time, agents will have lost a great deal of their personal initiative, which, confronted with rampant inequality, could degenerate into widespread feelings of frustration and helplessness.
It is therefore important to realise the dangerous implications that our everyday behaviours can have. Technically, putting the brakes on agentism is not a hard task: all it takes is a bit of awareness, and an ability to view our surroundings with a critical eye. Agentism has been thriving because it makes us feel better, it provides us comfort and gives us the impression that we are living in a safe, problem-free environment. We must learn to look beyond that and, occasionally, allow ourselves to step out of our comfort zone. The weapons are old, and well-tried: irony, humour, and repartee. In the ensuing battle of wits, we might rediscover a feeling that we have wilfully been trying to forget, the satisfying tingle of insubordination.
This article is part of a two-episode series on agentism. Read the second part here.