Perception is not the real

We live with the impression of gaining direct access to the world around us. We see a colour, hear a voice, recognise a face, assess a situation and quickly draw conclusions from what our senses transmit. This experience seems immediate. It nevertheless rests on a succession of selections.

Our vision covers only part of the electromagnetic spectrum. Our hearing remains confined to a determinate range of frequencies. Other species detect signals to which we are insensitive. Scientific instruments have made it possible to extend our field of observation considerably, from the infinitely small to the most distant astronomical objects, but each instrument likewise has a resolution, a domain of use and a particular way of translating what it measures.

To this physical limit is added the work of attention. The brain continually receives more information than it can consciously process. It selects, ranks, compares and relates what it perceives to what it already knows. Experience progressively transforms this selection. A mechanic sometimes hears in an engine a variation that remains unremarkable to its driver. A photographer immediately spots a light or a perspective that others do not notice. An experienced physician may detect in a set of symptoms an association that means nothing yet to the layperson. Learning enriches knowledge, but it also sharpens the eye.

Perceiving thus implies an encounter between what surrounds us and what we are able to distinguish within it.

Memory intervenes, as do affective state, habits, expectations and culture. A situation experienced as threatening draws attention toward certain cues. Familiar information more readily finds its place within an already constructed representation. Our values likewise orient what strikes us as serious, acceptable, unjust or worthy of interest. Two people present in the same place may receive comparable information and yet give their attention to very different aspects.

This diversity explains part of our disagreements, without sufficing to justify them all. Some divergences stem from an error, from false information or from fragile reasoning. Others arise because several people are observing different portions of the same situation. Distinguishing them requires returning to the available facts, to the conditions of observation, to the perimeter adopted and to the way the conclusions were built.

Position modifies what becomes visible

This idea appears with particular clarity in organisations. Someone who works daily on the ground knows constraints that management almost never meets directly. They see the useless movements, the material problems, the absurd deadlines, the procedures that work badly and the informal adjustments that nonetheless allow the work to continue. A middle manager has another view, made of coordination, resources, deadlines and trade-offs. At a higher level appear finances, contractual commitments, investments, legal risks and strategic choices.

These views do not exclude one another. They correspond to different positions within the same whole.

The problem arises when one of them is taken for a sufficient description of the whole. Indicators can give the picture of a satisfactory operation while those who carry out the tasks observe a daily deterioration. The reverse also occurs. A difficulty obvious from one particular post may be bound up with a wider constraint unknown to the person who suffers it.

Information is transformed as it circulates. A lived situation becomes a report. Several reports are grouped under a category, and that category sometimes joins a dashboard. Passing from one level to the next makes comparison possible, at the price of a loss of context. A singular experience, rich in detail, ends up fitting into a box or a percentage.

This reduction is unavoidable as soon as an organisation seeks to process a large volume of information. It becomes dangerous when the summary progressively replaces what it was supposed to represent.

Hierarchical relations complicate this circulation further. A pertinent observation may be blocked because it disturbs a local objective, brings an error to light or threatens the image of a department. Some employees also end up giving up on reporting problems when their remarks disappear without response or bring them more trouble than recognition.

Creating a reporting channel offers no guarantee in itself. Its value depends on the possibility of bypassing a level that would obstruct it, of following what becomes of a report, and of protecting whoever transmits embarrassing information. Depending on the context, several routes can coexist, for example the classical hierarchy, a collective representation, a confidential channel or an independent body. The interest of such plurality lies less in the multiplication of procedures than in the impossibility, for any single tier, of controlling the whole circulation.

Cross-immersions answer the same logic when they are carried out seriously. A few hours prepared for an official visit teach little. A sufficiently long presence under ordinary conditions makes it possible to understand more of what escapes reports. The reverse movement also has its interest. Having people from the ground take part occasionally in certain stages of decision gives them access to constraints that would otherwise remain foreign to them.

An organisation truly learns when it manages to connect these levels without effacing their differences.

Measurement, language and values

Our instruments of knowledge widen the view while orienting what deserves to be observed. Measuring requires choosing a variable, a unit, a period and a perimeter. A company that tracks its productivity every week while rarely assessing the wear on its teams does not merely hold two different qualities of data. It also grants far greater institutional visibility to one of these phenomena.

The choice of an indicator always expresses an intention. We measure what we wish to track, compare or control. Information left outside the apparatus then risks progressively losing its place in decision-making.

Language works in a similar way. Naming makes thinking possible. Categories help us recognise regularities, compare situations and communicate. They also carve up the world along boundaries that can oversimplify continuous or interlocking phenomena. One and the same person may be described through a medical, administrative, economic or social category. Each brings useful light, without exhausting what it describes.

Scientific disciplines likewise proceed by selection. The biologist, the sociologist, the economist and the psychologist do not pose the same questions when facing a complex phenomenon. Their vocabulary, their methods and their objects of study make certain relations easier to detect. Crossing these approaches increases the richness of the analysis when it allows genuinely complementary information to be confronted.

Values intervene even before the final decision. They influence the choice of the problems studied, the way a desirable outcome is defined, and the importance granted to the different consequences. A measurement can establish that a decision reduces a cost, increases an output or lowers a form of pollution. It does not choose the priority a community should grant to each of these consequences.

This distinction explains why more data does not necessarily put an end to disagreement. Two people may share the same diagnosis and continue to defend different choices because they do not give the same weight to the benefits, the risks or their distribution.

Making these differences visible nevertheless improves the discussion. A conflict becomes more intelligible once we know whether the participants are contesting a fact, a causal link, a timescale or a priority. A numerical error can be corrected by verification. A divergence of values calls for another kind of arbitration.

What we call the real

The distinction between perception and the real necessarily raises a difficulty. No observer possesses a point of view outside the world from which they could directly compare a representation with a reality entirely stripped of all mediation.

We nevertheless have a powerful clue. Some phenomena resist our expectations and our preferences. Observations made by different people converge. Instruments built on distinct principles can measure the same quantity. Predictions are verified under defined conditions. The world thus offers a resistance to our representations and makes it possible to discard some of them.

This resistance takes different forms according to what we study. A physical property can exist independently of any human convention. A currency, a border or a hierarchical office have a different status. Their existence depends on rules, institutions and shared beliefs, while producing very concrete effects. A debt does not grow in nature, but it can durably alter a person's life. A management position rests on a collective organisation, yet it really does give whoever occupies it possibilities of action that others do not have.

Social realities are thus bound up with human activity without being reducible to an individual opinion. An isolated person cannot abolish by will alone a law, a currency or an institution recognised by millions of others. These constructions acquire a stability of their own, while remaining liable to be transformed by collective action.

This particularity also explains why the study of societies cannot reproduce exactly the methods used to study matter. Individuals interpret rules, learn, anticipate the reactions of others and sometimes modify their behaviour when they know they are being observed. The phenomenon studied can evolve along with the knowledge we produce about it.

The existence of these difficulties does not, however, authorise every interpretation. A representation gains in soundness when it explains more observations, withstands confrontation, accounts for contradictions and can be revised when new elements appear.

Seeing enough to act

The search for multiple perspectives meets an obvious practical limit. No person, no organisation and no society can wait until everything is understood before taking a decision. Information remains incomplete, time is short and some uncertainties persist despite analysis.

Caution must then be proportionate to the possible consequences.

An easily reversible decision allows more experimentation. A temporary measure can be tested, evaluated and then corrected. The situation changes when a choice commits resources for several decades, destroys a capacity that is hard to rebuild, or exposes us to consequences whose repair would be very costly. The harder it becomes to turn back, the more reasonable it becomes to widen observation before acting.

This way of deciding avoids two pitfalls. Haste turns a partial representation into a certainty. Waiting for perfect knowledge would condemn us to immobility. Between the two lies a threshold of sufficiency that depends on the gravity of the consequences, on the remaining uncertainty and on the possibility of correcting the trajectory.

A well-designed decision can also produce new knowledge. Intermediate stages, an evaluation after implementation, or the possibility of revisiting certain provisions make it possible to learn from real effects instead of assuming that the initial model had foreseen everything.

This capacity for correction becomes especially precious in complex environments. An error does not always come from a total lack of information. The clues sometimes already existed, scattered among several people, several departments or several disciplines. It was simply that no one held the set needed to recognise their relation.

The quality of our perception then depends as much on the precision of our gaze as on its mobility. Knowing from what position we are observing, understanding what it makes visible, seeking out the information liable to escape it, and accepting to revise our representation when the facts require it: these allow us to approach the real with greater accuracy.

We shall never possess a complete view of it. That limit nevertheless leaves ample room for the progress of knowledge, provided we never confuse the portion that has become visible with the whole of what remains to be understood.