A case does not always need to be compared with the whole world. Sometimes it needs to be compared with the right local world.
A student scores 62.
Against the whole school, that score may look ordinary.
Against students taking the same advanced paper under the same time limit, it may be unusually strong.
Against students who made the same first-step error, the mark may be almost irrelevant.
The local comparison set changes the meaning of the case.
That local comparison set is the neighbourhood.
Quick Route
- Distance: how far apart are two cases?
- Neighbourhood: which nearby cases form the relevant local comparison set?
- Density: how crowded is that local set?
- Cluster: does local closeness form a broader group?
- Outlier: is the case unusual globally, locally or both?
Canonical Job
Cognitive Art uses neighbourhood as a reader-facing concept for one job:
Which cases are locally close enough to this case that they should influence how it is interpreted, predicted or acted upon?
This is not a claim that psychology has one universal “neighbourhood module.”
Different fields use neighbourhood ideas differently.
Word recognition has orthographic and phonological neighbourhoods.
Statistics and machine learning use nearest neighbours and local density.
Geography uses spatial neighbourhoods.
Cognitive Art extracts the common structure without pretending the mechanisms are identical.
One-sentence answer: A neighbourhood is the local set of cases close enough to a target under a chosen representation that they become the most relevant immediate reference for similarity, expectation, classification or action.
A Neighbourhood Is Not Context
Context includes the broader information surrounding a case:
- task,
- time,
- role,
- culture,
- prior state.
Neighbourhood is narrower.
It asks which nearby cases in the active representation should form the local comparison set.
Context can change the neighbourhood by changing which dimensions matter.
A Neighbourhood Is Not a Cluster
A cluster is a group-like structure in a space.
A neighbourhood is centred on a target case.
One point can have a neighbourhood even when no clean global cluster exists.
Neighbourhood is local and target-relative.
Cluster is a larger grouping claim.
A Neighbourhood Is Not a Category
Category says:
these cases belong to the same class.
Neighbourhood says:
these cases are currently close enough to matter as local references.
Cases can be neighbours without sharing one formal category.
Neighbourhood Requires Distance
The new Cognitive Art article on distance asks how far apart two cases are under a chosen representation.
Neighbourhood begins once we ask:
Which cases are near enough to count as local?
That requires a boundary or ranking rule.
Nearest five cases?
All cases within a chosen radius?
All cases sharing a key relation?
Different neighbourhood definitions answer different jobs.
There Is No Neighbourhood Without a Metric
People can be neighbours geographically.
Far apart professionally.
Close in values.
Far apart in age.
The neighbourhood depends on the distance relation chosen.
This is why “nearby cases” should never be treated as self-evident.
Local Comparison Can Beat Global Average
Average adult height tells you little about whether a professional basketball centre is unusually tall for their position.
Average company revenue tells you little about whether one branch is unusual relative to branches of the same size and market.
Global averages compress heterogeneous populations.
Local neighbourhoods can restore relevant comparability.
The Wrong-Neighbourhood Error
A Secondary 4 student is compared with all secondary students.
Too broad.
A new startup is compared only with the three famous survivors in its industry.
Too selective.
Neighbourhood choice changes the baseline against which the target is interpreted.
A misleading neighbourhood creates a misleading local reality.
Lexical Neighbourhoods: A Real Cognitive Example
Word-recognition research provides a concrete example of neighbourhood structure.
A written word can have orthographic neighbours—other words differing by a small letter change.
Forster and Shen’s 1996 paper, No Enemies in the Neighborhood, examined how orthographic neighbourhood density influenced lexical decision and semantic categorisation.
Importantly, the effect depended on task.
This gives us a useful boundary:
even when a neighbourhood is objectively definable, its cognitive consequence can depend on what the person is trying to do.
Neighbours Can Help or Compete
A nearby case can help because it provides a useful precedent.
It can hurt because several similar candidates compete for selection.
In lexical recognition, different models make different predictions about competition, activation and facilitation.
Cognitive Art does not collapse those debates.
The broader public lesson is:
the local field around a target can change how easily that target is recognised or interpreted.
Nearest Cases Can Guide Classification
You encounter a new mushroom.
Its nearest remembered cases matter more than distant unrelated species.
This is structurally related to the earlier Cognitive Art article on exemplars.
Exemplar supplies the specific remembered cases.
Neighbourhood organises which of those cases are locally close enough to deserve the most weight.
Local Does Not Mean Small in Physical Space
A conceptual neighbourhood can span continents.
Two scientific papers may be local neighbours in theory space despite being produced decades and thousands of kilometres apart.
Neighbourhood is local in the active representation, not necessarily local on a map.
Neighbourhoods Can Be Asymmetric in Practice
Mathematical nearest-neighbour relations can be defined symmetrically or asymmetrically depending on method.
Human attention can make the practical relationship even less symmetric.
A novice may treat an expert case as far away.
The expert may immediately recognise the novice case as a familiar local variant.
Representation and expertise change neighbourhood membership.
Neighbourhoods Can Change With Learning
Before learning calculus, many problems feel unrelated.
After learning derivatives, questions about slope, rate and optimisation move into one conceptual neighbourhood.
Learning reorganises psychological space.
Old distant cases become neighbours because a new invariant has been learned.
Neighbourhoods Can Change With Perspective
A historian groups events by period.
An economist groups the same events by monetary regime.
A military analyst groups them by logistics.
Perspective changes the dimensions and therefore changes which cases become local neighbours.
Neighbourhood and Density
A neighbourhood can contain three close cases.
Or three thousand.
The local concentration changes.
That is density.
Neighbourhood defines the local region.
Density describes how occupied it is.
Neighbourhood and Outlier
A case can be ordinary globally and strange locally.
The earlier Outlier article already owns this distinction.
Neighbourhood supplies the local comparison field that makes local unusualness visible.
Neighbourhood and Base Rate
Global prevalence can differ sharply from local prevalence.
A disease may be rare nationally but common in one exposed subgroup.
A type of exam error may be uncommon overall but frequent among students using one specific mistaken rule.
The right base rate depends on the right reference neighbourhood.
This does not mean shrinking the comparison group until the preferred answer appears.
Neighbourhood must be justified by mechanism or task relevance.
Operating Envelope
Neighbourhood reasoning is strongest when:
- distance has a defensible meaning,
- the local set is large enough to support inference,
- the neighbourhood is chosen before seeing which group produces the desired conclusion,
- local structure predicts something useful.
It becomes dangerous when local comparison is used to cherry-pick a favourable reference group.
Neighbourhood in Mathematics
Mathematics formalises neighbourhoods in several technical ways, including analysis, topology and computational methods.
Cognitive Art does not claim those definitions are interchangeable.
The shared structural idea is local relation around a target.
Neighbourhood in English
A word sits inside local lexical neighbourhoods.
Change one letter.
Change one sound.
Change one semantic relation.
Different neighbourhoods appear.
This helps explain why some words are easier to confuse, retrieve or associate than others, while effects vary by task and model.
Neighbourhood in Science
A new observation should often be compared with cases under similar:
- temperature,
- population,
- instrument,
- mechanism,
- experimental condition.
Local comparability reduces false conclusions caused by mixing structurally different regimes.
Neighbourhood in Education
A learner answers one question incorrectly.
The useful neighbourhood may be:
- questions with the same topic,
- questions with the same hidden structure,
- questions producing the same first wrong step.
The third neighbourhood can be more diagnostic than the first.
Good teaching chooses neighbours by mechanism, not only chapter label.
Neighbourhood in Organisations
Benchmarking one branch against the entire company can be misleading.
A better local neighbourhood may match:
- customer mix,
- staffing,
- market,
- operating hours,
- scale.
Fair comparison begins with a defensible neighbourhood.
Failure Modes
1. Global-Average Capture
The whole population is used when the local comparison class is more informative.
2. Cherry-Picked Neighbourhood
The local group is defined after seeing which comparison produces the preferred conclusion.
3. Surface Neighbourhood
Cases are treated as neighbours because they look alike while the causal structure differs.
4. Tiny-Neighbourhood Overconfidence
Two nearby cases are treated as enough to establish a reliable local pattern.
5. Frozen Neighbourhood
The comparison set is not updated after learning, population change or regime shift.
Repair Path
- Define the target case.
- Define the job.
- Choose the relevant distance.
- Construct a local comparison set before inspecting the desired outcome.
- Check whether the local set is large and coherent enough.
- Compare local and global conclusions.
- Test whether the neighbourhood predicts different action or explanation.
The Neighbourhood Audit
- Neighbourhood around which target?
- Near according to which distance?
- Which dimensions define local similarity?
- How many neighbours are included?
- Why should these cases be comparable?
- Does local prevalence differ from global prevalence?
- Could another defensible neighbourhood change the conclusion?
- Is this local set a cluster, or merely a target-centred sample?
- Has the neighbourhood changed over time?
- What decision improves because this local frame exists?
A Primary-to-Adult Progression in Neighbourhood Thinking
Primary: compare with nearby examples
Children learn categories through small sets of similar and contrasting cases.
Lower secondary: choose the right comparison set
Students learn that a case can look ordinary globally and unusual locally.
Upper secondary: test local versus global inference
Learners examine subgroup choice, local base rates, sample size and the risk of cherry-picking reference classes.
Adulthood: use local structure without losing the whole system
Professional reasoning moves between local and global views, choosing the narrowest comparison set that preserves enough evidence and mechanism to support action.
Research Notes and Further Reading
Lexical research provides one concrete cognitive use of neighbourhood structure. See Forster and Shen, No Enemies in the Neighborhood: Absence of Inhibitory Neighborhood Effects in Lexical Decision and Semantic Categorization. The study found task-dependent effects of orthographic neighbourhood density, illustrating why local similarity structure and cognitive task cannot be separated casually.
For the broader geometry underlying local similarity, see Roads and Love, Modeling Similarity and Psychological Space (Annual Review of Psychology, 2024).
For a different domain showing that human “neighbourhoods” can have graded and context-sensitive boundaries, see Clare Davies, Places as Fuzzy Locational Categories. Cognitive Art treats this as a useful parallel, not evidence for one universal neighbourhood mechanism.
World Return
A neighbourhood earns its place when local comparison improves prediction or action.
If the selected neighbours repeatedly mislead, the distance or reference class must change.
Local reasoning should be accountable to outcome, not protected by the convenience of a small comparison set.
Final Thought: The Whole Population Can Be Too Far Away to Help
The student still scored 62.
The number did not change.
What changed was the company it kept.
Compared with the wrong group, the score tells the wrong story.
Local intelligence begins by asking not only “What is this case?” but “Which nearby cases deserve to define what normal means here?”