Seeing like a school: categories, maps, streets and winks
How the top down tools of school leadership make institutions easier to manage but harder to understand
We cannot deal with reality in its raw state. Too much happens, too quickly, in too many places at once. This is true everywhere, and schools are no exception. Like any complex institution, a school has to turn the rush of daily life into something that can be named, discussed and acted upon. A school day contains glances, routines, absences, corridor movements, half-understood explanations, forgotten homework, fragile relationships, remembered slights, improvised repairs and small acts of professional judgement. No leader can hold all this in mind. No policy can capture it completely and no spreadsheet can contain all the variables. So schools simplify.
Timetables are created so that time becomes governable. Labels are assigned so that children, teachers and departments can be categorised in ways the institution can handle. Data is collected so that learning, behaviour and attendance can be viewed from a distance. Policies are written so that expectations don’t have to be reinvented every morning. Routines are established so that everyone knows what is supposed to happen next.
This is all eminently practical. A school unable to simplify complexity would be chaotic, exhausting and, very probably, unsafe. Children need order, teachers need shared expectations, and leaders need some way of knowing what’s going on beyond their own line of sight. Trouble begins when these necessary simplifications are treated as truths rather than tools.
Once we have simplified reality, those simplifications begin to acquire force. A label such as “low ability” may begin as a rough description of prior attainment, but it can soon affect the work a child is given, the questions they are asked and the explanations we reach for when progress stalls. A teacher thought of as “weak on behaviour” may be seen through that judgement even when the causes of difficulty sit in the class, the curriculum, the timetable or the school’s wider routines. A department thought to be “underperforming” may find that every piece of evidence is interpreted as confirmation of a settled story. What begins as institutional shorthand becomes an explanation, and eventually a self-fulfilling prophecy.
This is one of the most dangerous features of institutional life. Organisations do more than use categories; they begin to think with them. Mary Douglas, the great anthropologist of classification and institutional life, saw this clearly in How Institutions Think. Institutions, she argued, do more than govern behaviour; they govern attention. They teach us what counts as similar, what counts as different, which distinctions matter and which can be ignored. Over time, these classifications acquire the feel of common sense. They no longer feel like active decisions made by particular people in precise circumstances but like a natural expression of how things are.
Whatever else they are, schools are classification machines. They sort by age, subject, attainment, attendance, behaviour, need, risk, progress and performance. They decide what counts as a lesson, what counts as evidence, what counts as disruption, what counts as support and what counts as improvement. These classifications are unavoidable, but never neutral. Once they are built into registers, seating plans, data systems, interventions, reports and meeting agendas, they shape what we are able to see. They steer our attention, making certain questions feel urgent while pushing others out of view.
A child described as “low ability” is no longer encountered only as someone who did badly on a test or has not yet mastered a particular body of knowledge. The label gathers meaning around it. It may affect the curriculum she is offered, the explanations teachers give for her errors, the pace at which adults expect her to learn and the kind of future imagined for her. We may believe we’re recording factual information, but really we’re creating a shorthand for understanding the world.
Albert Michotte, the Belgian psychologist best known for The Perception of Causality, helps explain why this is so treacherous. In his famous “launching effect” experiments, observers watched simple moving shapes. One object moved towards another, touched it, stopped, and the second object began to move. Under certain conditions, observers did not merely report movement followed by movement. They reported causation: one object had hit the other and made it move. Change the timing, even slightly, and the causal impression weakened.
This is uncomfortably close to what happens in schools. We see low attainment and infer low ability. We see missing homework and infer laziness. We see a noisy class and infer poor teaching. Worse, we see evidence of poor teaching and infer a poor teacher. But in school life, as in Michotte’s experiments, what feels like direct perception of cause is often a story imposed on a pattern. We see effects: scores, behaviours, absences, silences, refusals, messy books and stalled progress. Causes have to be inferred, and those inferences are shaped by the categories we already use.
Douglas’s point is that institutions stabilise thought by embedding categories in routines. This helps because it saves effort. Teachers cannot rethink every distinction from scratch every morning, and leaders cannot run a school without shared names for recurring problems. But stability has a price. As Douglas puts it, “Institutions direct individual memory and channel our perceptions into forms compatible with the relations they authorize.” Once a category becomes institutionally useful, evidence is interpreted in ways that reinforce it. We notice what fits more readily than what troubles the category, and disconfirming evidence is discounted. The labels we use start as tools for coordination and end as artefacts of institutional memory.
Ordinary cognitive biases are recruited into institutional life. The halo effect means that one salient judgement spills over into other judgements. In schools, it often works as a kind of reverse halo. Weak writing is taken as evidence of weak thinking. Poor test performance is read as poor motivation. One difficult class becomes a judgement about a teacher’s general competence. A child’s reputation arrives long before the child. On top of this, the fundamental attribution error deepens the problem. Schools, like individuals, tend to overestimate the role of character and underestimate the role of context. Once a label is attached, behaviour is too easily treated as evidence of disposition. Struggle is read as weakness, resistance as attitude, inconsistency as incompetence, and underperformance as low aspiration. The explanation moves too quickly from observation offers to the inference of causes.
The cost is practical as much as moral. Dispositional explanations allow the school to move quickly from problem to person, but they also hide the mechanisms that might be changed. Rather than asking, “How do we fix this faulty individual?”, the better, though more awkward, question is, “What conditions are making these responses more likely?” This forces greater precision. If the problem is framed as laziness, weakness or low aspiration, the default response tends to be exhortation, surveillance or blame. If the problem lies in missing knowledge, poor task design, weak routines, unstable staffing or badly aligned assessment, the response has to be more exact
Take a child who is unable to punctuate speech accurately. The immediate problem is specific: she does not yet understand how direct speech is marked, how punctuation sits inside inverted commas, or how speech tags attach to what has been said. That problem can be taught. It can be broken down, modelled, practised, checked and revisited. If the child has already been entered into the school’s group mind as “low ability”, the instructional problem is easily swallowed by the category. Teachers stop asking, “What exactly has she failed to secure?” and start behaving as though the label has already answered the question.
Something similar happens with teachers. A colleague struggling with a difficult Year 9 group may need help with entry routines, seating, narration, modelling, practice, curriculum materials or confidence. The problem may sit partly in the class composition, partly in the timetable, partly in the curriculum, partly in the teacher’s habits and partly in the needs of individual students. Once the institution identifies the teachers as being weak on behaviour management, complexity collapses. A calm lesson is treated as an exception and chaotic lessons are treated as further evidence that this is a hopeless case. Although it’s a criminal waste of limited human resources, it’s far easier to jettison the teacher than address the systemic obstacles to effective behaviour management.
Bad ideas survive in schools because they have been built into the architecture. They sit in spreadsheets, meeting agendas, intervention schemes, seating plans and reporting cycles. Once a category becomes part of an institution’s ordinary language, it becomes hard to think without it.
The same danger applies to evidence.
Schools run on evidence. Leaders who ignore data and rely only on instinct will soon drift into fantasy. But school evidence is almost always tells us less than we imagine. A reading age gives us a trace of reading performance under particular conditions. A data drop records how students performed on selected tasks at a certain point in time. A behaviour log tells us which behaviours were noticed, recorded and coded. A book scrutiny gives us a sample of work, shaped by task design, curriculum sequence, student effort, teacher expectations and the distortions caused by knowing that books may be checked. A learning walk offers a brief encounter with teaching under conditions that change the teaching being observed.
Evidence becomes more seductive when it’s stripped of context. Once reality has been converted into numbers, those numbers slide easily through meetings, dashboards and reports, offering an impression of clarity that allows leaders to feel that they’re acting objectively. The danger is that numerical proxies begin to feel more solid than the reality they represent.
James C. Scott, the political scientist and anthropologist of state power, peasant resistance and administrative order, gives us the best account of this temptation in Seeing Like a State. Scott’s interest lies in what happens when governments try to make complex human arrangements easier to understand. “I began to see legibility as a central problem in statecraft,” he writes. Forests, cities, farms, names, occupations and populations are simplified so they can be surveyed, taxed, regulated and improved. The state needs maps, registers, grids, censuses and standard measures because action from the centre depends on simplification. But Scott’s warning is that simplification always comes at a cost. To make reality simple enough to administer, the state has to leave things out. Often, those things are important
Although Scott’s examples are not concerned with education, the pattern is immediately recognisable. A natural forest becomes a managed forest because the state prefers trees of the same species, planted in straight lines, measured in uniform units and harvested on a predictable schedule. A city becomes easier to administer when streets are straightened, addresses standardised and neighbourhoods made visible to planners. The simplification may be useful, but it also destroys forms of local knowledge that made the original arrangement work.
Schools have their own predictable ways to make it simpler for those in charge to pull levers in the hope of making meaningful change. Students’ progress is turned into grades, a term’s teaching into percentages, a department’s work into a spreadsheet, a teacher’s practice into a target, a child’s behaviour into a code entered in a Management Information System, and a curriculum into a sequence of boxes shaded green, amber or red. Although some of this can be useful, the danger comes when we treat proxies as the things they are meant to represent. Scott says, “The more static, standardized, and uniform a population or social space is, the more legible it is.” Legibility improves administrative control by reducing local intelligence.
