AEvery city has sounds that belong to it: market traders, engines, bells, birds, construction, music and fragments of conversation. Most disappear as soon as they are produced. Sound archivists try to preserve selected examples, arguing that recordings can reveal aspects of urban life that photographs and written documents cannot.

BEarly sound archives were limited by technology. Recording equipment was expensive, fragile and difficult to transport, so many collections focused on formal speeches, music or controlled studio sessions. Everyday street sound was rarely captured systematically. Portable tape recorders changed this in the twentieth century by allowing researchers to record outside studios and institutions.

CField recordings can document social change. A neighbourhood may sound different after a tram line closes, a factory shuts down or a new immigrant community arrives. Linguists also use urban recordings to study accents and language change. However, archivists warn that recordings should not be treated as neutral samples of reality: the person holding the microphone decides where to stand, when to record and what to exclude.

DThis problem becomes more important when archivists build collections intended to represent an entire city. Central districts are easier to record than private homes, industrial zones or communities whose members distrust institutions. As a result, a sound archive may accidentally preserve the voices of people who are easiest to reach rather than those who are most representative.

EMetadata is therefore essential. A recording becomes far more useful when researchers know the date, location, equipment, recorder and circumstances in which it was made. Without this information, future listeners may misinterpret what they hear. A loud crowd, for example, could represent a normal evening market or an unusual political demonstration.

FDigital storage has made large collections easier to create, but preservation remains difficult. Audio formats become obsolete, files can be corrupted, and storage systems require regular migration. A digital archive that is copied once and then ignored may be less secure than expected. Long-term preservation depends on repeated maintenance.

GThere are also ethical questions. Recording people in public spaces may be legally permitted in some places, yet individuals may still object to being preserved in an archive. Voices can reveal identity, age, emotion and sometimes private information. Some projects therefore anonymise recordings, restrict access or seek consent where practical.

HDespite these challenges, sound archives are becoming more important to historians and urban researchers. They capture rhythms, interruptions and background activity that text rarely records. The goal is not to save every sound, which would be impossible, but to build collections transparently enough that future users understand both what was recorded and what was left out.

APeople often assume that expertise guarantees good judgement. Yet experts can disagree sharply even when they have access to the same evidence. This is not always a sign that one side is incompetent. Complex decisions involve uncertainty, different assumptions and different ways of weighing risk.

BStatistician Dr Elena Park studies forecasting tournaments in which specialists make numerical predictions and receive scores for accuracy. She finds that forecasters who update their estimates frequently tend to outperform those who defend an early prediction. Park argues that expertise is most useful when combined with willingness to revise beliefs as new information arrives.

CPsychologist Martin Doyle focuses on overconfidence. In laboratory studies, experienced professionals were sometimes more certain than novices even when their accuracy was similar. Doyle suggests that years of success can create a strong internal sense of familiarity, which people may mistake for evidence that their judgement is reliable in a new situation.

DProfessor Sana Malik examines disagreement between experts from different disciplines. An economist, engineer and public-health specialist may recommend different policies because each model emphasises different consequences. Malik argues that disagreement can be informative: it reveals which assumptions matter most and where evidence is weakest.

EDecision researcher Paulo Mendes studies group discussion. Groups of specialists can improve judgement when members bring genuinely independent information. But if junior members simply repeat the view of a senior expert, apparent agreement can hide a lack of real evaluation. Mendes therefore distinguishes consensus produced by evidence from consensus produced by hierarchy.

FSeveral institutions now use structured methods to improve expert judgement. These include asking experts to estimate probabilities rather than make absolute predictions, recording forecasts before outcomes are known, and comparing performance over time. Such methods can reveal which experts are consistently well calibrated rather than merely persuasive.

GNevertheless, numerical scoring cannot solve every problem. Some policy decisions depend on values as well as facts. Experts may agree on the probability of an event but disagree about how costly the risk would be or which groups should bear the burden. In these cases, disagreement is partly ethical or political rather than technical.

HThe public often finds expert disagreement frustrating because advice appears less certain. Yet hiding disagreement can be more dangerous. When uncertainty is communicated clearly, decision-makers can compare scenarios, identify assumptions and prepare for several possible outcomes.

IThe most reliable expertise may therefore be less about confidence than about method. Good experts know what evidence supports, what remains uncertain and what would cause them to change their minds. Expertise becomes stronger, not weaker, when revision is treated as part of the process rather than as an admission of failure.