Selective interpretation at work

I catalogue smells for a historical research project, although “catalogue” makes the work sound more stable than it is. We cannot put an eighteenth-century bakery odour in a drawer. Instead, I combine written descriptions, chemical traces from objects and experiments with perfumers, then record exactly which choices produced a reconstruction.

My background was in social history, so the laboratory vocabulary initially intimidated me. The useful surprise was that uncertainty became an asset: visitors compare our reconstruction with their own memories and expose assumptions hidden in the archive. The most difficult requests come from institutions that want one authoritative scent for an exhibition.

I can offer a documented interpretation, not a bottled fact. We also publish failed formulas, since an attractive smell that proved unstable can otherwise disappear from the record and be proposed again as though no experiment had occurred. This makes the archive slower to read but much harder to mistake for a catalogue of settled answers.

Because visitors often treat smell as direct memory, I also explain that recognition may be produced by later associations, so emotional force cannot by itself validate the reconstruction.

As an audio describer for theatre, I write the words spoken between lines of dialogue so that blind and partially sighted audience members can follow visual action. The challenge is not to describe everything but to decide what matters before the next actor speaks. I began as a stage manager, and my knowledge of blocking helps, but it also tempts me to notice technical details that an audience does not need.

Rehearsals with users changed my practice: a costume colour may be irrelevant, while the fact that two characters avoid looking at each other can alter an entire scene. The work is precise, yet it depends on leaving enough silence for the performance to remain a performance. Before each performance I compare my script with the director’s visual priorities, then remove descriptions that duplicate what dialogue or sound already communicates.

Restraint is not absence of preparation; it is the result of deciding which information would compete with the scene instead of open it. The script is reviewed again after the first audience, when pauses, laughter and movement reveal whether a description arrived early enough to inform the scene without explaining an effect already understood.

I map unstable cliffs for a coastal authority. People imagine that the job is mainly about dramatic fieldwork, but most decisions emerge from comparing modest changes across photographs, sensors and old survey notes. A single crack can look alarming without indicating movement; a quiet slope may be storing water that matters later.

I trained as a geologist and expected the science to provide firm boundaries. Instead, I spend much of my time explaining ranges and thresholds to planners who need a line on a map. The best maps show where confidence changes rather than disguising uncertainty as a neat edge.

That can frustrate clients, but it prevents a provisional judgement becoming a permanent fact. After storms I revisit places where the model had been least confident, because a correct warning does not validate every boundary and a missed movement may expose an assumption shared by several datasets. The map remains a working claim whose confidence can migrate.

A confidence boundary is retained even when planners demand one line, through notes that specify which evidence could move it and which decision would be affected.

I mediate disputes over shared courtyards in housing blocks. Residents usually arrive with a practical complaint—bins, bicycles, noise—but the object is often carrying a longer history of being ignored. My previous career in restaurant management taught me to stay calm when several people speak at once, though mediation required me to abandon the habit of finding an immediate operational solution.

Sometimes the most useful meeting ends without agreement because the parties have finally distinguished what can be negotiated from what needs a legal decision. I am judged by whether conflict becomes more manageable, not by whether everyone leaves cheerful. That distinction was difficult to explain to the organisation that first employed me.

Written agreements record which issues require housing management or legal advice, preventing mediation from absorbing conflicts it has no authority to settle. That boundary protects the process from appearing successful merely because a difficult dispute has been renamed as communication. When agreement is reached, we record the smallest workable commitment and the condition for revisiting it, preventing a calm final meeting from being mistaken for permanent resolution.