Seeing change: why a conspicuous difference can be missed

Epigraph

It implies withdrawal from some things in order to deal effectively with others.

William James, The Principles of Psychology, Volume I, XI, printed p. 404: clause beginning It implies withdrawal, with its final comma normalized to a period. English text. A historical argument about attention’s selected object.

What this article is about

A change can be easy to distinguish when two versions stand together and nevertheless go unnoticed during an ordinary activity. Someone was reading a caption, answering a question, or following a definite part of a scene. This alone does not establish that they saw nothing, deliberately concealed an observation, or have a particular personal problem. The article examines the questions behind ‘I would notice’. We distinguish a detail’s availability, the current task’s object, comparison between states, and a later report. William James’s historical argument is considered beside two definite studies by Simons and colleagues: replacement of a conversation partner and an unexpected event in a video. Their procedures differ, and results retain sample and condition limits. Voluntary practice examines two versions of a neutral document; it neither measures general powers of observation nor promises to eliminate omissions.

I. Two versions together and two moments in succession

Imagine an editor of a small exhibition catalogue. In the morning she reads the date beneath an image; that evening she checks the artist’s name. Between these activities a colleague changes the date and moves the caption. The editor confirms the name but reports no new date. This scene is invented for analysis, not recorded in a study. When both files are opened together, comparison makes the difference available. That does not establish that the editor had already compared the earlier and later dates. Success with one question does not automatically answer another. First identify her activity, available versions, and required distinction before explaining the omission through her character.

‘Conspicuous’ can describe different properties here. Letters are large; a change occupies much space; a new colour differs from the earlier colour; a correction matters to the organiser. None of these descriptions alone identifies the feature noticed during successive viewing. Importance to one participant does not become another’s task without communication. This does not make every condition equivalent or say that size never matters. It specifies the inference: distinguishability in a definite comparison and reporting during another activity should be considered separately. We can examine whether a difference was available without presenting its availability as an established history of every person’s experience.

When the editor says ‘everything was on screen’, she describes access, which is useful but incomplete as an account of checking. If a colleague answers ‘I saw it immediately’, ask whether he had the old version beside the new one and knew where to look. In this example those conditions can distinguish the later viewing from the first. Subsequent discovery does not make the earlier omission impossible. It supplies a new event with different available information. For a calm discussion, name both tasks: checking the name yesterday and comparing dates today. This clarifies the conversation’s object without claiming discovery of the sole psychological cause of the earlier response.

II. Missing a change and missing an event

The studies consulted ask two related but distinct questions. In change blindness a researcher compares states before and after a change, such as replacing a conversation partner. In inattentional blindness an unexpected event occurs during another task, and observers are later asked whether they noticed it. This need not involve comparing two faces or versions of one object. These names identify research situations rather than diagnoses of readers. ‘Blindness’ in this usage does not describe the condition of someone’s eyes. To understand a result, return the term to its procedure instead of combining every absent response under one ready-made mechanism.

Both tasks can be imagined for the catalogue. In one, a caption is replaced and the editor must identify the difference. In another, an additional mark appears while she checks names, and she is later asked about it. We do not claim these invented activities replicate the published experiments or yield their numbers. They show why a change question includes a relation between states, whereas an unexpected-object question can concern one episode. Even when both answers are ‘I did not notice’, their objects differ. Retaining this distinction prevents transfer of one result’s explanation to another without further grounds, or confusion between comparison and an image’s general recognisability.

The assigned task also differs from a question introduced afterwards. Checking whether a year changed is not identical to reading an interesting page without expecting that question. Knowledge of the forthcoming comparison can belong to the working conditions; it cannot silently be added to the first viewing’s description. A report of later success therefore has a bounded meaning: with a definite question and material, someone found a difference. It describes that activity. It does not establish that the same feature would be noticed under every other goal. Equally, an earlier omission does not establish inability to notice it during a later, differently organised comparison.

III. Replacing a partner in a definite interaction

In Simons and Levin’s 1998 paper the first experiment involved fifteen pedestrians on Cornell University’s campus, with approximate ages from twenty to sixty-five. An experimenter carrying a map asked for directions. After ten to fifteen seconds, others carried a door between the participants. For about one second it obscured the partner, who was replaced by another experimenter. The conversation continued with an identical map copy. The experimenters differed in clothing, voice, and height by approximately five centimetres. After giving directions, pedestrians were asked about an unusual event and, if needed, directly about replacement. The authors report that seven of the fifteen noticed the switch.

The observation matters because the activity was not confined to an image: a pedestrian participated in an ongoing conversation. Yet this natural element does not turn fifteen cases into a description of all conversations. Replacement, door, map, surprise, and questioning constitute definite conditions. The paper discusses possible objections, including map distraction and reporting. One person who initially named nothing unusual later claimed, after a direct question, to have noticed the replacement. Thus a recorded answer and a proposal about earlier inner experience cannot simply be equated. We retain the reported result and procedure limits instead of declaring an incomplete answer proof of the absence of all perception.

In the second experiment the same experimenters dressed as construction workers; twelve younger Cornell students participated. Four reported the switch when asked about something unusual. Three others later claimed noticing but could not accurately describe differences. The authors consider social categorisation and individuating features. This interprets a small comparison of conditions; it does not measure a universal inability of an age or occupational group. Clothing and situation changed alongside how the partner might be understood, and all such factors were not exhaustively studied. The narrower point suffices here: nonreport of replacement occurred during interaction under these conditions, so ‘he stood in front of you’ alone cannot reconstruct the entire detection process.

IV. An unexpected event among basketball passes

Simons and Chabris’s 1999 study asks a different question. The main experiment initially involved 228 observers, almost all students; after thirty-six exclusions, 192 responses were analysed. The paper specifies exclusions for familiarity, losing count, recording problems, unclear response, and one extreme count. There were sixteen conditions, with twelve observers in each. Each watched one of four seventy-five-second videos. Two teams of three passed basketballs, while the observer followed the white or black team. An unexpected woman carrying an umbrella or wearing a gorilla costume walked through the action. In the main experiment this event lasted five seconds, rather than requiring comparison between two versions of a partner.

Conditions differed in both video construction and task. In one version all actors were filmed together; in another, separately recorded images were superimposed transparently. Observers counted either total passes by the selected team or separate bounce and aerial passes. After viewing, questions about the unexpected event became progressively more specific. Across all conditions, fifty-four percent reported noticing and forty-six percent did not. This combines two unexpected events and several tasks; it is not the proportion of people who ‘never see a gorilla’. Even within this study, different conditions cannot be replaced by one famous clip, nor its aggregate percentage transferred to each viewer or everyday life.

The authors report differences between conditions: more noticing with simultaneously filmed actors than transparent superposition, and with one total count than two separate counts. These describe the specified procedure, not a promise that every simplification of work produces the same effect. Discussion includes an additional chest-thumping gorilla video with another event duration and a new group; it is not merged with the main 192 observers. This preserves an important source detail: the familiar image does not identify the experiment’s only arrangement. For the catalogue we can propose a question about what the editor tracked, but obtain neither a ready percentage of missed corrections nor proven benefit from a new editorial technique.

V. A detail’s relevance depends on the question

For the catalogue organiser, a date determines the exhibition’s opening, whereas the editor was assigned to check a name. The same feature has different roles in these tasks. The organiser would err by assuming his date’s importance had already communicated the entire checking procedure. The editor would err by promising that a completed name check certifies every page detail. Before assigning blame to ‘inattention’, establish the agreement: what was to be compared and what result was expected? This reasons about our invented interaction rather than explaining a particular experimental participant. It preserves separate investigation of access, assignment, and the actual report instead of reducing them to one impression.

An unexpected event can seem obvious to someone who already knows what to seek, but that knowledge belongs to the current situation. ‘I saw it immediately’ after its location is indicated does not answer what would happen without the indication. In the video experiment different tasks accompanied one class of unexpected events; team selection and counting method were conditions. We need not claim attention is the sole factor in every omission. It suffices to retain the task when explaining a result. Discussion of conspicuousness then becomes more precise: ask about an object’s properties and the activity during which its definite feature was to be detected and reported.

Importance should also be distinguished from novelty. A new catalogue border can stand out while barely affecting meaning; a familiar date can determine a visitor’s plans. Frame a meaningful transition: is the opening time correct, has the address changed, do caption versions agree? Such selection concerns the document’s purpose. It neither proves the selected detail will necessarily attract attention nor removes other quality questions. Sometimes another participant can receive a separate task; sometimes identifying the unchecked part suffices. In either case, accurately describing the completed activity is more informative than ‘the page was viewed’, which specifies no information actually compared or conclusion the viewing could support.

VI. A historical question about selecting experience

At the opening of his attention chapter, James challenges experience conceived as the mere presence of an outward order before the senses. Interest gives some items emphasis. We use selected paragraphs as historical reasoning about selecting an object, not an explanation of every result in the later experiments. No automatic bridge joins philosophical description to experimental procedure. James supplies neither conversation-partner replacement conditions nor a proportion of observers missing an object. The modern studies likewise do not measure every sentence of his chapter. Their comparison starts with a narrower question: what is the current activity’s object when many things are available, and what can its result establish about that activity?

The epigraph comes from a sentence describing attention through concentration and withdrawal from other objects. It does not advise ignoring a person or important change: it characterises selection in a historical account. Nearby, James says his experience is what he agrees to attend to. Reading this as a contemporary guarantee of voluntary control over every noticed object would introduce a stronger proposition than needed here. A reader can select a catalogue question, but the decision alone does not certify comparison without omissions. Examine the available result instead of treating intention as proof of successful performance, or unsuccessful performance as proof that the intention was insincere.

This reading preserves two tasks. A historical text clarifies the language of experience and purpose; an experiment specifies conditions yielding a definite report. Their roles cannot be exchanged: quotation does not replace procedure, and a response percentage does not exhaust interest’s meaning. The editor might record, ‘My task was checking the name, and I can report that result.’ A date comparison remains separate. We neither justify an error by one philosophical phrase nor infer a permanent personal trait from it. We distinguish the completed check’s limits, the assignment’s meaning, and the possibility of organising another question about the same document while retaining its shared object.

VII. A report and an explanation of an omission

‘Did not report the replacement’ identifies a questioning result. ‘Retained no features of the person’ proposes an explanation of how it arose. The second proposition needs additional grounds. The door-study authors discuss categorisation and distinguishing features, but do not directly observe each pedestrian’s experiential contents. Continuing the conversation is itself ambiguous: the paper notes that some people who noticed the switch still continued giving directions. Similar outward action therefore supplies no ready answer about a difference already noticed. In the editorial scene, confirming the name likewise answers no complete set of questions about date, border, and viewing sequence. Preserve the observed part and keep explanations as separate proposals.

In discussing unexpected events, Simons and Chabris consider the possibility of noticing an object and rapidly forgetting it. They prefer an account involving lack of awareness and provide reasons, while acknowledging difficulty in directly separating the alternatives. This is a substantive boundary, not an inconvenient detail to remove for a vivid story. We can report the responses obtained under the procedure. We should not add that each response identifies an exact inner event or unique neural mechanism. Possible implicit traces of an unnoticed event also remain open in this study. The article therefore does not turn absence of a report into proof of the absence of every subsequent influence.

In ordinary conversation, ask no more of an answer than it states. ‘I do not remember whether the date changed’ leaves a question open. ‘I saw yesterday’s date’ offers a specific recollection that can be compared with a file version. ‘I checked the date against the programme’ identifies another route. These statements differ in object and grounds; they do not measure attentiveness. If a decision needs the information, consult saved versions and separate records. Clarifying a question does not guarantee recovery of a past experience. It helps distinguish a new fact, later recollection, and explanation of an earlier answer, then identify material capable of supporting the required document conclusion.

VIII. An omission is no verdict on a person

A reader may recognise a personal episode and feel embarrassed, but recognition supplies no experimental assessment. The two papers describe definite groups and procedures rather than every individual’s possible abilities. One missed date does not establish permanent unreliability or inability to look carefully. For a shared task, investigate the actual correction and its agreement. The change may matter; repairing it requires no general verdict on personality. This is the proposed inference boundary. It neither declares errors harmless nor removes responsibility for a promised check. It preserves the distinction between required work and an unsupported narrative about the performer’s permanent properties, letting the specific consequence receive appropriate attention.

If the organiser promised visitors an accurate time, he needs date confirmation from an appropriate source. Friendly assurance that ‘I saw everything’ may be insufficiently specific. Ask for the checked item and its grounds without requiring someone to prove a general capacity to notice the obvious. Acknowledging an unchecked part then clarifies work status. It does not concede that the source is false or another person necessarily correct. The document has a definite version, and the report a definite object. Shared checking should join them through an appropriate activity rather than exchange impressions about who is more observant or whose confident statement sounds most persuasive.

The examples are not used to assess health, professional fitness, or a witness’s truthfulness. Those conclusions need other questions, material, and competencies; these procedures supply no ready reader instrument. In the neutral editorial activity, name the result: the date changed, the name agrees, the earlier date comparison is unconfirmed. That is already substantial if those are the participants’ required items. Respect for a person can coexist with correcting a shared document. It requires neither retaining a wrong date nor declaring experience subjective beyond examination. We can preserve the common task and available evidence without making an unsupported conclusion about an observer’s entire life from one detected omission.

IX. Giving comparison an appropriate object

For the catalogue we can propose a separate activity with two versions: open them together, name a field, and record its agreements or differences. This is an authorial organisation of a question, not a technique with effectiveness established by the consulted studies. It provides material that may have been unavailable earlier: simultaneous access to before and after states. Name that changed condition. If a difference is found, the result concerns that field and those saved versions. It guarantees neither discovery of every other difference nor absence of future omissions. The report’s scope can be clear without expanding into an unsupported statement that the entire document has been completely checked.

Sometimes the earlier version is unavailable. ‘What changed?’ then differs from ‘What is written now?’ Read the current date and compare it with the programme; seek revision history elsewhere. No recollection needs to be invented for a complete answer. A saved record may establish the previous value, but this differs from reporting one’s earlier perception. The distinction matters for action: current timing can be confirmed while past experience remains unknown. We do not reconstruct that experience automatically from a record. A shared document check can progress without demanding an account of details the participant does not now remember or is not prepared to assert with appropriate grounds.

Discovery also differs from correction. We can show precisely that two dates differ without knowing which matches the exhibition plan. That requires a scheduling source. Version comparison answers a difference question; it does not independently certify either version’s correctness. Replacing the date because the new one looks more persuasive would move to another conclusion without grounds. Our example can stop at an honest record: a difference was found, and the correct value is still being clarified. This is a substantial intermediate result. Its value lies in making the next question definite, not in promising that every noticed change already leads to a correct decision about it.

X. A missed detail does not erase the world

A report’s limits do not entail that no shared event occurred. The experimenters were actually replaced according to procedure; in the invented catalogue we stipulate two versions. The question concerns an observer’s relation to that change, not creation of an object by confidence. Saved files can differ even if the editor missed the date. A confident difference report can still be checked. These grounds require no promise of complete perception. We can acknowledge an incomplete answer and examine an accessible shared fact. This avoids the opposite exaggeration: lacking an all-encompassing report does not make every assertion equally admissible or render further comparison meaningless.

End the discussion with a named result. A changed date was found; the performed task is known; earlier detection remains unknown; current timing is checked separately. Not everything is explained, but participants can understand their agreement. New evidence can change one part of the record without rewriting the entire narrative about a person. This form of closure is editorial reasoning. Neither James nor the experiments establish its effectiveness for every group. They retain different roles: a historical selection question, results of definite procedures, and our way of limiting inference in an invented activity. Those distinctions let comparison remain useful without a general myth that people see only what they desire.

The title therefore promises no hidden cause for every omission. It invites closer reading of a claim about conspicuousness. What changed? Between which states? What activity was performed? How was the report obtained? What conclusion is needed now? Answers can differ in completeness and grounds. One case needs caption comparison; another leaves a version-history question; a third requires a different source. Every page need not become an endless investigation. Relate work to its purpose and avoid presenting a completed part as the whole. Surprise at a missed difference can remain an occasion for conversation and inquiry without becoming diagnosis, accusation, or proof that no shared world exists.

What this means in practice

Take two existing versions of your own short, neutral document if you want to compare them, perhaps an invitation or a personal drawing’s caption. Do not secretly alter others’ material to test an acquaintance. Choose one definite question, such as the date. Record available versions and whether that field belonged to the earlier task. If that task is unknown, say so rather than reconstruct it from today’s result.

Open the versions together and distinguish the current entry, earlier entry, and discovered difference. If the earlier file is unavailable, limit the conclusion to the current state and another available source. Do not replace document history with invented recollection. ‘I do not remember’ leaves a question open without preventing reading today’s date.

Name the next action’s grounds. Different dates do not establish which is correct; correction may require the original programme. If comparison found nothing, specify the checked field instead of certifying the entire contents. In either case record the result’s limits. Stop when the required question has a sufficient answer; discovering every possible difference is not this small note’s purpose.

Finish with two sentences: what is established about the document and what remains unknown about earlier viewing. This voluntary reading activity measures no general attentiveness, offers no diagnosis, and has no effectiveness against omissions established here. The authors propose it to distinguish tasks, grounds, and reports. There is no requirement to perform a comparison irrelevant to your activity. Keep only a record useful for understanding your own question, without assessing another person from one discovered difference.

Key quotation

My experience is what I agree to attend to.

William James, The Principles of Psychology, I, XI, printed p. 402, complete sentence. The historical proposition retains its discussion of interest; this article infers no voluntary control over every noticed object.

Connections between traditions

James / the door study: a historical question about attention’s object is compared with reporting a changed partner. Quotation does not explain each pedestrian’s inner process.
James / the video study: selection and counting tasks allow a shared question, but philosophical description does not replace sixteen experimental conditions.
Two experimental approaches: replacing an already seen person and introducing an unexpected object share a detection question; state comparison and event reporting remain different procedures.
Historical reasoning / an editorial catalogue: the invented scene clarifies an assignment’s limits; it does not experimentally test James’s chapter.
Research reports / reading practice: distinguishing a recorded answer from an explanation helps frame a question. The consulted experiments do not measure effectiveness of the proposed document comparison.

Primary sources

About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, the unity of traditions, intercultural dialogue, and contemporary spiritual discipline. Instagram: @andrey.zarinsh.

Maria Evilkut is a co-author of United Wisdom. Alongside her research, she seeks to test these ideas in practice and observe how principles of unity and mutual understanding work in real life between people and cultures. Instagram: @ilkutki.