You were there.  That does not mean you saw everything that happened

Imagine a meeting about a project that has begun to go wrong.  A senior manager asks whether the team can still meet its deadline.  One person pauses before answering.  Another points to a figure on the screen that has not changed since the previous meeting.  The manager says, “We need a clear answer by Friday,” and moves to the next item.

Afterwards, one participant says the manager was clearly looking for reassurance and did not want to hear about the risks.  Another thought the question was an invitation to put the risks on the table, and that nobody took it.  A third remembers the pause as the crucial moment: the team had a chance to be honest and hesitated.  Each person can name something that happened in the room.  They disagree about what those things meant, which detail deserves the most weight and who was responsible for what followed.

If the project later fails, those early accounts may harden into histories.  But the first difference between them did not necessarily arise months later, when someone misremembered the meeting.  It may have begun while everyone was still sitting around the table.

We tend to think of perception as the reliable part of the process and memory as the stage at which a story begins to drift.  Yet experience is selective before we attempt to recall it.  What reaches our attention, what we recognise, and what we infer from another person's behaviour are related processes, but they are not the same thing.  We can be sincere witnesses to an event and still have encountered only part of it.

The room contains more than anyone notices

The participants in our meeting were doing different jobs.  Someone responsible for the deadline was listening for a commitment.  Someone who prepared the financial figures was watching how those figures were received.  A person worried about a technical problem was waiting for a chance to raise it without sounding alarmist.  Their attention was shaped by what they needed to do next, and that allocation of attention was useful.  Without it, a meeting would become a flood of unranked gestures, words and interruptions.

Attention can also leave us unaware of something surprisingly visible.  In a well-known study of radiologists, researchers inserted an image of a gorilla into lung scans while experienced observers looked for nodules.  Most did not report the gorilla, even though eye-tracking showed that many had looked at its location.  The finding is striking because expertise did not make the unexpected image impossible to miss.  It also needs to be understood in context: the observers had a demanding, specific search task.  The study does not show that experts usually overlook what matters in their work, or that seeing an unexpected object and recognising its significance are one simple act.

The more general point is that being present is not the same as attending to everything present.  A person may hear the words of a warning while concentrating on the figure that seems to qualify it.  Someone may notice a long silence but miss the glance that explains why another participant stopped speaking.  We cannot infer from a failure to mention a detail that the person deliberately ignored it.  Equally, we should not treat a confident account of what somebody saw as proof that all the relevant information was available to them.

This is not a defect that could be fixed by trying harder to notice everything.  Selecting is how attention works.  In ordinary life it allows us to follow a conversation in a noisy room, drive through traffic and find the particular document we need among dozens.  The danger appears when we mistake the useful selection for a complete record of the scene.  Our account may be accurate about what held our attention and still incomplete about what happened beyond it.

Context helps us make sense of what is there

Selection is only the first part of the story.  The information we do attend to must be interpreted.  A raised voice at a football match means something different from the same volume in a small office.  A pause after “Can we meet the deadline?” may suggest calculation, dissent, anxiety or a search for the right words.  We rarely wait for all possible evidence before deciding what a scene is about.  Usually we could not.

Research on visual perception shows why context cannot simply be treated as contamination.  In experiments on objects within changing scenes, participants' performance depended on whether an object's reappearance matched the movement implied by its surrounding scene.  The research concerns visual expectations in controlled tasks, not motives in a meeting.  It illustrates a modest point with wider relevance: the surrounding situation helps us predict and recognise what we are looking at.  An object isolated from its setting can be harder to identify; the setting can also lead us towards an interpretation that needs checking.

There is an important boundary here.  It would be too easy to say that we simply see what we expect to see.  Expectations can guide attention and influence the interpretation of ambiguous information, but that does not establish that every belief directly alters visual experience.  Researchers still debate how far cognition can penetrate perception itself, and different experiments probe different stages of processing.  The defensible claim is narrower: a person's prior knowledge and current task help determine what they attend to and what sense they make of it.  That is already enough to complicate the idea of a neutral witness.

Think again of the unchanged figure in the meeting.  To the person who produced the figures, it may represent a forecast built on assumptions that are now beginning to fail.  To the manager, it may be the latest agreed number.  Both saw the same cell on the screen.  They did not bring the same information to it, and they may not have realised that the other's interpretation differed.  Nothing in this example requires either person to be foolish or deceptive.  It requires only that a number can acquire meaning from a history not visible on the slide.

People are harder to read than figures

When we move from objects to other people, another layer enters.  A word, expression or gesture is observable.  Intention is usually inferred.  We may be remarkably good at many such inferences, especially with people we know, but familiarity can create its own shortcuts.  A familiar hesitation becomes “what she always does when she disagrees”.  A brief interruption becomes “another attempt to take control”.  The behaviour and the explanation can merge so quickly that we no longer hear the join.

Our opening witnesses may all remember that the manager moved on.  One calls it dismissal.  Another calls it a decision to keep the agenda moving until the team produced more evidence.  The distinction matters because the next action changes with the interpretation.  If you believe you were dismissed, you may stop raising concerns.  If you believe you were asked to return with a clearer case, you may prepare one.  The meaning assigned in the moment can alter the next event, even before anyone has told a retrospective story.

We need to be precise about what this does and does not imply.  Sometimes dismissal is exactly what occurred.  A pattern of interruptions, a written refusal to consider evidence or a later decision to suppress a concern can support that judgement.  The point is to identify which part of the account is directly observed and which part is a claim about another person's purpose.  That distinction gives a serious allegation something firmer to stand on when the evidence supports it.

The same discipline applies in close relationships.  “He did not call” reports a behaviour that may be checked.  “He did not care” proposes an explanation.  It might be right; it might also overlook fear, distraction, incapacity or a competing obligation.  An alternative explanation does not erase the effect of the missed call.  Someone can be hurt by what another person did without being able to know exactly why they did it.  Keeping behaviour, effect and motive separate lets us acknowledge all three without pretending they are interchangeable.

A viewpoint is part of the event

The meeting also looks different depending on where each person sits in its chain of decisions.  A manager may know that a client has threatened to withdraw if the deadline moves.  An analyst may know that the underlying data is weaker than the summary suggests.  A junior team member may have encountered a practical obstacle that never reached the agenda.  None possesses the whole situation.  Each has access to information others may not even know is missing.

This is why “we were all in the room” is less conclusive than it sounds.  Sharing a location does not mean sharing a role, responsibility, background knowledge or cost of speaking.  The person who can end the meeting experiences a pause differently from the person deciding whether to risk being labelled difficult.  Authority does not automatically make one person's perception wrong.  It does affect the conditions under which information becomes visible to them.

A meeting record can help establish what was said and decided, but its form can also obscure the differences.  “Risks discussed” might be accurate while giving no indication that one participant considered the risk unresolved and another considered it closed.  “No objection raised” may mean consensus, or it may mean that nobody found an opening in the time available.  Those possibilities can be investigated; they should not be settled solely by the most convenient reading of the minutes.

The wider systems consequence begins here.  Organisations often believe they are gathering information when they are actually gathering information that can pass through a particular meeting structure.  Questions set the terms of acceptable answers.  Deadlines determine how much uncertainty can be voiced.  Hierarchy affects which interpretations are challenged.  The eventual institutional story may start to take shape while the first discussion is happening, through what gets noticed, labelled and recorded.

Partial is not the same as false

If everyone attends selectively and interprets through context, it is tempting to conclude that there are only versions of reality.  That goes too far.  A deadline was stated or it was not.  A figure changed or it did not.  A warning may have been written down, sent and received.  These facts can matter enormously, even when the full meaning of a silence remains uncertain.  The existence of perspective does not remove the possibility of better and worse accounts.

Nor should we confuse incomplete observation with general incompetence.  Human perception works well enough for us to move through an extraordinarily complicated world, and expertise can make people more sensitive to precisely the patterns they are trained to recognise.  The radiologists' missed gorilla is revealing because it was outside the task they were performing.  A surprising demonstration of selective attention should make us ask what someone was looking for, not dismiss everything they did observe.

That question is useful in an ordinary disagreement too.  What did each person need to know?  What were they watching?  Which details did they agree on before assigning motives?  What information was available to one but not the other?  A disagreement that seemed to require a liar may turn out to contain two partial observations and two competing interpretations.  In another case, one account may be plainly contradicted by the record.  The method does not predetermine the result.

Our witnesses leave the troubled-project meeting with different accounts.  Before asking whose memory later changed, we have to understand that their experiences were never identical to begin with.  Attention selected different details.  Context gave those details different significance.  Each person inferred a meaning from the behaviour of others, then acted on that inference.  The event was real, and so were the limits of any one person's view of it.

What happens to those versions over time is the next part of the story.