Insight does not automatically change a social situation. A person may understand why a recurring meeting is draining, why a rule creates silence, or why a social problem cannot be solved with better self-discipline. Yet the step from understanding to action can be unclear. Chapter 37 discusses how a sociological hypothesis can be made testable through a small social experiment.
The word “experiment” is used here in a colloquial and practical sense. You should not conduct research on other people. A social experiment is an open, delimited, and—as far as possible—reversible change to a routine, relationship, environment, or decision-making process. The experiment should be conducted within the existing mandate, with consent when others are affected, and with a set timeframe for follow-up. The goal is not to prove a general theory. The goal is to learn something about a specific situation and adjust the next step.
In the book’s methodology, the experiment follows socioanalysis and a gentle formulation. This means that the course of action is not chosen at random. First, there needs to be a reasonable idea of what is keeping the problem alive. Then, the level is chosen where a change can both address the mechanism and be implemented in a responsible manner. Sometimes the next step lies in an everyday interaction. Sometimes a group or organization needs to change a rule. Sometimes a public or collective approach is required.
The main text refers to the movement from insight to practice as “socioasceticism.” Here, the term denotes a trial-and-error process that is socially embedded. It is a working hypothesis formulated by the author, not a validated treatment. This chapter should therefore not be read as a call for constant self-improvement. A wise course of action may be to wait, seek support, delegate responsibility, or acknowledge that one’s own scope for action is too limited.
An experiment begins with a question, not with a demand
A performance project often begins with the question: How can I improve? A sociological experiment, on the other hand, begins with: What happens if a specific social condition is altered slightly? The difference is important. The first question directs attention toward the individual as the bearer of the shortcoming. The second opens up the possibility that time, space, rules, roles, resources, and relationships may be what needs to change.
A working hypothesis is a provisional explanation that can be revised. If a meeting is dominated by two people, the hypothesis might be that the open-discussion format favors those who speak up quickly. A possible experiment, then, is not for the quieter participants to practice becoming more outgoing. Instead, over the course of two meetings, the group can try a rotation system where everyone first gets a short period of uninterrupted time to speak. If more perspectives emerge, the experiment supports the hypothesis. If nothing changes, the explanation needs to be reevaluated.
An experiment must also have a stopping condition. This could be a fixed time limit, a cost that must not be exceeded, or a side effect that causes the experiment to be halted. Without a stopping condition, a temporary trial can slip into becoming a new requirement. Reversibility does not mean that every consequence can be undone. It means that the change is chosen so that it can be stopped without significant harm and that possible consequences have been thought through in advance.
The mechanism links problems and action
A mechanism is the intermediary through which a condition contributes to a consequence. If a person becomes exhausted from work, “work” is too broad a mechanism. It could be that last-minute schedule changes make recovery unpredictable, that responsibilities lack boundaries, or that a staffing policy distributes the workload unevenly. The clearer the intermediary, the easier it is to choose what to test.
This does not mean that every problem has a single cause. A broad sociological formulation may encompass multiple interrelationships. But a single experiment usually needs to focus on one or, at most, two mechanisms. If time, staffing, leadership, housing, finances, self-image, and family relationships all change at the same time, it becomes impossible to understand what the experiment demonstrated. Delimitation is therefore a method for learning, not a claim that the rest is irrelevant.
The measure must be closely aligned with the desired change. If the goal is participation, it is not enough for the meeting to be shorter. The group needs to observe who had a chance to speak, whether the decision was understandable, and whether anyone was assigned more invisible work. If the goal is recovery, it is not enough for a new routine to be implemented. It must also be possible to determine whether sleep, stress, or access to rest were affected.
Micro, meso, and macro are interconnected levels
The micro level concerns the immediate situation: actions, conversations, bodies, objects, and attention. The meso level concerns relationships, groups, and organizations with recurring roles, routines, and rules. The macro level concerns institutions, social structure, legislation, and historical patterns. These levels are —not three separate worlds. A law is implemented within organizations and is experienced in concrete interactions. A recurring everyday action can simultaneously contribute to maintaining a larger order.
C. Wright Mills distinguished between private concerns and public issues. This distinction helps assign responsibility. If a single person has difficulty understanding a form, individual support may be appropriate. If many people fail because the language is unnecessarily difficult, there is an organizational and institutional issue. A micro-level intervention can still be helpful, but it should not be described as the solution to a mechanism that lies at a higher level.
The level at which a problem is noticed is therefore not always the level at which it is primarily reproduced. Anxiety may be felt physically but is exacerbated by precarious employment conditions. Silence may arise in an individual meeting but recur because the organization rewards speed and punishes objections. A social experiment must therefore identify two points: where the consequence becomes visible and where the rule or resource that the experiment aims to influence is located.
Microexperiments openly alter a situation
A microexperiment can alter the sequence of a conversation, the placement of an object, a time, a means of contact, or the way a role is performed. Small changes can make unspoken norms visible. Harold Garfinkel’s ethnomethodology demonstrated how much effort people put into making everyday life appear self-evident and understandable. His famous norm-breaking, however, is not a template for secretly provoking family, colleagues, or strangers.
In this book, microexperiments should be open and low-risk. If a conversation pattern is being examined, participants can agree to try a new turn-taking order. If a cell phone interrupts a shared moment, those involved can try placing their phones in a visible spot for half an hour. If a role has become rigid, the person can try a slightly different action that fits within the task, without pretending to have a competence or mandate they lack.
The important thing is not to treat others as guinea pigs. Participants should be able to understand what is changing and say no. Nor should a microexperiment be used to prove that a specific person is the problem. The focus is on the interplay between rules, situations, and actions.
Meso-experiments target rules and resources
The meso level often offers the greatest practical leverage. A workplace, organization, educational institution, or family has recurring patterns regarding time, information, responsibility, and conflicts. These may be perceived as personal because they involve an individual, but they are sustained by multiple people and repeated over time. A change in a rule can therefore affect more situations than a piece of advice given to a single person.
A meso-experiment requires a legitimate decision-making process. This might involve a work team testing a new method for distributing background materials, an association trying out a clearer division of responsibilities, or a study group establishing a fixed procedure for handling questions. Someone without a mandate can document a pattern, gather feedback, and submit a proposal. She should not single-handedly introduce a system that others are expected to follow.
Rules also carry values and power. A more efficient routine can make things easier for some and harder for others. Therefore, the follow-up needs to ask who benefited, who ended up with more work, and who had the opportunity to object. A successful trial is not just one that meets a metric, but one that can be defended before those affected.
Macro-level action requires a path between experience and institutions
As an individual, you can rarely experiment with a social structure. However, it is possible to explore a path for collective or institutional influence: formulate a shared question, contact a responsible party, participate in a consultation, organize a small group, or investigate an existing alternative. The experiment then lies in how people coordinate their efforts and bring the issue to the appropriate level.
Paulo Freire’s concept of praxis unites reflection and action in dialogue. The point is not activity for activity’s sake. Those affected need to be able to articulate their situation, test their understanding, and adjust their approach as new experiences emerge. Erik Olin Wright’s “real utopias” similarly direct attention toward actually existing institutions that realize parts of a desirable alternative. An alternative becomes instructive when its conditions and limitations can be examined, not when it is treated as a finished model.
Macro-level action can start on a very small scale. Identifying which committee, union channel, advocacy group, or public statistic pertains to the issue can be the first step. The next step might be to develop a joint position paper or hold a meeting with a clearly defined agenda. It remains important to distinguish between exerting influence and shouldering systemic responsibility alone.
Shared resources require local rules
Some social experiments involve something people use together: a room, a digital forum, an association’s fund, a tool pool, a garden, or a shared knowledge resource. Elinor Ostrom’s research shows that common-pool resources do not always have to be understood as a choice between private ownership and centralized state control. People can create local institutions, but their sustainability depends on factors such as clear boundaries, participation in rule-making, monitoring, and mechanisms for conflict resolution.
This is not a universal checklist. A small housing cooperative, an open digital community, and a natural resource each have different conditions. However, these questions help us see why good intentions alone are not enough. If it is unclear who is allowed to use the resource, who maintains it, and how misuse is handled, costs can easily become invisible or unevenly distributed.
An experiment can therefore consist of testing a single rule regarding a shared resource. This should be done in the appropriate forum and with due consideration for ownership and responsibility. The goal is not to create a commons everywhere, but to investigate whether a more participatory arrangement can solve a specific coordination problem.
Monitor impact, cost, and displacement
An experiment does not simply yield a yes or no answer. It may be feasible but ineffective, effective but too expensive, or appreciated by some but burdensome for others. A sociological follow-up must therefore distinguish between at least four factors: whether the change was implemented, whether the intended outcome occurred, what costs were incurred, and whether the problem was shifted.
Scheduling is particularly important. A sooner appointment may require more preparation at home. A new channel of communication can help experienced users but exclude those who lack technical skills or language proficiency. A private solution can alleviate a temporary concern while simultaneously making organizational shortcomings less visible. Follow-up should not seek out errors in order to halt everything, but rather to prevent benefits from being calculated in a way that obscures the costs borne by others.
Feedback from those affected is a source of insight in its own right. The person who designed the experiment finds it easier to see what she intended to achieve. The person who experienced the change may notice friction, extra work, or ambiguity. When there is a significant power imbalance, a channel is needed where criticism can be offered without anyone being punished.
A failure is not always a personal defeat
An experiment may fail because the hypothesis was wrong, the scale was inappropriate, resources were insufficient, or the timing was poor. The recipient may lack the mandate; other events may have influenced the outcome; or a key group may have been excluded. These distinctions are necessary for the failure to become a source of insight rather than a source of shame.
This does not mean that every failure should be romanticized. A harmful consequence is still harmful. Sometimes the attempt shows that something should be abandoned. Sometimes an apology or restitution is required. Sometimes the lesson is that an issue is too big for an everyday experiment to handle.
The main manuscript’s idea of reflexive action can be summarized as reflection and action correcting one another. Knowledge that never meets reality risks remaining in the planning stage. Action that is never followed up risks repeating the same pattern. The experimental approach is therefore not an identity as a “brave person,” but a way of working that allows for reevaluation.
A good social experiment, then, is not necessarily bold. It is well-defined, focused on a reasonable mechanism, and open to the possibility that the results may change our understanding. Chapter 38 moves on to what often determines whether an experiment is even feasible: a supportive social network.