What Is Relevance? Attention, Context, Information and Cognitive Art

You are trying to catch a train.

The station has hundreds of facts.

the colour of the ceiling,
the number of tiles on the floor,
the advertisement beside the escalator,
the history of the station,
the temperature of the platform,
the departure time of your train.

All may be true.

Only one may matter right now.

This is the problem of relevance.

Reality contains more information than any mind can process.

Thinking therefore depends not only on acquiring information.

It depends on refusing information that does not help the current job.

Quick Read

Relevance is the relationship between information and a current question, goal, decision, interpretation or state.

Information is relevant when processing it can materially change:

what we believe,
what we predict,
what we decide,
what we attend to next,
which action becomes appropriate.

Relevance is therefore not a property of information in isolation.

It depends on the receiver and the job.

One-sentence answer: Relevance is the degree to which information can change the answer, model, prediction or action that matters in the current context.

True Does Not Mean Relevant

The station ceiling is white.

True.

Your train leaves in two minutes.

Also true.

For deciding whether to run, only one is relevant.

This is a basic but powerful distinction.

Truth concerns whether a statement fits reality.

Relevance concerns whether the statement bears on the current question.

Good reasoning needs both.

A perfectly true answer to the wrong question is still a bad answer.

Relevance Is Not Salience

Salience asks:

What stands out?

Relevance asks:

What matters to this job?

A flashing advertisement can be highly salient and completely irrelevant to your train.

A small platform number can be visually quiet and highly relevant.

Attention is vulnerable when salience and relevance diverge.

The loudest signal wins perception while the quiet signal controls the outcome.

Relevance Is Not Importance

Climate change is important.

It may not be relevant to deciding whether your algebraic sign is wrong in line three.

Your lunch order is not globally important.

It is relevant when the waiter asks what you want.

Importance is broad value.

Relevance is task-relative usefulness.

Relevance Is Not Evidence Strength

A laboratory test may be highly reliable.

But if it measures something unrelated to the hypothesis, it is weak evidence for that hypothesis.

Evidence must first be relevant before reliability can matter.

Then we ask how trustworthy and discriminating it is.

This gives a useful sequence:

Does it bear on the question? → Is it reliable? → How much should it move belief?

A Question Creates a Relevance Field

Ask:

Why did the plant die?

Now some variables become candidates:

water,
light,
temperature,
soil,
disease,
root damage.

The colour of the pot may remain irrelevant.

Change the question:

Why does this pot absorb more heat in sunlight?

Now colour may matter.

The world did not change.

The question changed the relevance field.

Goals Change Relevance

At a supermarket, a nutritionist notices ingredients.

A designer notices packaging.

A parent notices price and allergies.

A logistics manager notices weight and shelf life.

Same object.

Different goals.

Different relevant features.

Context Changes Relevance

One sentence:

It is cold.

At home, it may be a request to close the window.

In a science experiment, it may be a measurement problem.

During a medical consultation, it may be a symptom report.

Context changes which consequences the information has.

One Influential Theory of Relevance

Dan Sperber and Deirdre Wilson developed Relevance Theory, especially in relation to cognition, pragmatics and communication.

Their influential account proposes that cognitive systems tend to allocate attention toward inputs expected to yield useful cognitive effects relative to the effort required to process them.

In their précis of Relevance: Communication and Cognition, they connect relevance with cognitive effects and processing effort. Later discussions of the theory similarly describe relevant inputs as ones that can produce consequences such as new implications, confirmation, correction or useful reorganisation of knowledge.

Cognitive Art does not claim that Relevance Theory is the one settled scientific account of every relevance judgement.

It uses the broader reader-facing problem:

Which information deserves processing because it can change the current cognitive job?

Cognitive Effect: What Changes Because I Processed This?

A useful way to test relevance is to ask what changes after the information enters.

Does a hypothesis become stronger?
Does an assumption become weaker?
Does a new inference become possible?
Does a decision change?
Does an error become visible?

If nothing important changes, the information may have low relevance to the current job even if it is fascinating.

Processing Cost Matters

Imagine two sources contain the same answer.

One requires forty minutes of dense reading.

The other gives the critical fact clearly in two minutes.

If both are equally reliable for the question, the second may have greater practical relevance per unit of cognitive effort.

This does not mean “easy is always better.”

Sometimes difficult processing produces larger cognitive effects.

The point is that cognition must trade information gain against processing cost.

The Relevance Bottleneck

Modern information environments create a peculiar problem.

The shortage is no longer always information.

It is relevance.

Search engines return thousands of pages.

Social media returns endless updates.

Dashboards return hundreds of metrics.

The cognitive job becomes filtering.

Search Is a Relevance Machine

A search query is a compressed relevance instruction.

“Secondary 3 algebra factorisation common errors” tells the retrieval system which region of the information space matters.

Weak query:

Math help.

Broad relevance field.

Strong query:

Why does my student expand (x − 3)² as x² − 9?

Narrower field.

Better chance of retrieving information that changes the action.

Search Failure Can Be Relevance Failure

The answer exists.

You still do not find it.

Perhaps:

the query was too broad,
the wrong variable was treated as important,
salient results displaced relevant ones,
the desired answer type was never specified.

Retrieval quality depends on relevance specification.

Relevance in Reading

A passage contains twenty details.

The question asks:

Why did the character leave?

Strong readers search for details bearing on motive.

Weak readers may retrieve memorable but irrelevant details.

Reading comprehension is not only extracting information.

It is selecting information relevant to the inferential job.

Relevance in Writing

A student has excellent knowledge.

The essay still scores poorly.

Why?

The knowledge is not connected to the question.

One of the most important writing questions is:

What work is this sentence doing for my claim?

If the answer is unclear, relevance is weak.

The Knowledge-Dump Error

Students often mistake amount of knowledge for quality of answer.

Everything remembered enters the page.

The marker must search for the answer inside the information.

World-class writing reverses the burden.

Only knowledge serving the current claim survives selection.

Relevance in Mathematics

A word problem contains:

numbers,
units,
objects,
conditions,
relationships.

Which information enters the equation?

Some details are decorative.

Some define constraints.

Some identify the target.

Mathematical modelling begins by deciding what is relevant enough to represent.

The Distractor Problem

An examination includes a number that is not needed.

Some students feel compelled to use it.

Why?

Because presence is mistaken for relevance.

Relevance literacy allows a student to say:

This information is given, but it does not change the mathematical state I need.

Relevance in Science

A variable can be measured perfectly and still not belong in the causal explanation.

Science asks which variables bear on the phenomenon and which can be controlled, ignored or treated as noise.

Experimental design is partly relevance engineering.

Include variables that can discriminate hypotheses.

Exclude distractions that increase complexity without increasing explanatory power.

Relevance in Diagnosis

A system fails.

Hundreds of facts are available.

Good diagnosis asks which observation changes the ranking among candidate causes.

A fact that all hypotheses predict has low discriminating relevance.

A fact that strongly separates two hypotheses has high diagnostic relevance.

Relevance and Value of Information

Not every unknown deserves to be resolved.

Suppose you can purchase an expensive report.

If none of its possible findings would change the decision, the information has little decision value.

Relevance can therefore be tested counterfactually:

If I learned the answer, what would I do differently?

If the answer is “nothing,” investigate why you are collecting it.

Relevance in Education

A learner asks:

What should I revise?

“Everything” is technically safe and practically weak.

Relevant revision depends on:

exam weighting,
current mastery,
prerequisite structure,
time remaining,
likely transfer value.

Relevance turns study from accumulation into allocation.

The Teacher’s Relevance Problem

Teachers know more than learners need at one moment.

The temptation is to explain everything.

But instruction should ask:

What distinction must become visible now?
What background can wait?
What example will change the learner’s representation?

Teaching quality depends partly on delivering the next relevant structure rather than the entire expert library.

Relevance in Organisations

A dashboard contains fifty metrics.

Which three should the meeting discuss?

The answer depends on the decision being made.

A metric can be important historically and irrelevant to today’s bottleneck.

Organisations drown when reporting systems preserve information without preserving relevance.

Relevance Can Change Mid-Problem

At the start, one variable seems irrelevant.

New evidence arrives.

Now it becomes central.

Relevance is state-dependent.

Good thinkers can reopen discarded information when the problem state changes.

The Premature Irrelevance Error

Discarding information too early can be as dangerous as keeping everything.

A strange measurement looks unrelated.

Later it reveals the hidden mechanism.

Relevance judgements should therefore carry confidence.

“Probably irrelevant” is safer than “impossible to matter” when uncertainty remains high.

The Relevance–Salience Conflict

Breaking news is salient.

Your long-term financial plan may depend more on a quiet interest-rate assumption.

A dramatic student error is salient.

A recurring small misconception may be more relevant to future performance.

Mature attention asks not only what captured me but what deserves me.

Relevance and Assumption

Assumptions determine which information looks relevant.

If you assume the problem is financial, cost data enters first.

If the real problem is reliability, failure data may matter more.

A wrong assumption can build the wrong relevance filter.

Relevance and Perspective

Different observers occupy different jobs.

The engineer asks whether the bridge is safe.

The commuter asks whether it is open.

The historian asks why it was built there.

Different perspectives change relevant variables without changing the bridge.

Relevance and Resolution

Once relevant information is located, how much detail should be preserved?

That is a resolution problem.

Relevance selects the region.

Resolution selects the grain.

Relevance and Causality

A cause is relevant to an effect because changing the cause can change the outcome under appropriate conditions.

But not every correlated fact is causally relevant.

The next Cognitive Art article will separate temporal order, association and causal relevance more carefully.

The Relevance Audit

What exact question am I answering?
What decision or prediction depends on the answer?
Which information would change that answer?
Which facts are true but irrelevant?
What is merely salient?
What assumptions are shaping my relevance filter?
Which evidence discriminates among alternatives?
What information costs more to process than it is worth?
Has the problem state changed enough to reopen discarded information?
What one missing fact would have the greatest decision value?

A Practical Exercise: Ten Facts, Three Relevant

Take a news article, textbook page or meeting document.

Write one question.

Then extract ten facts.

Choose the three that most change the answer.

Explain why the other seven do not.

This trains relevance rather than memory volume.

A Practical Exercise: Salient or Relevant?

For one day, notice what captures attention.

Label each:

salient and relevant,
salient but irrelevant,
quiet but relevant.

The third category is where important work often hides.

A Practical Exercise: The Decision-Change Test

Before seeking more information, ask:

If this turns out A, what will I do? If it turns out B, what will I do?

If both produce the same action, the information may not deserve immediate search.

A Primary-to-Adult Progression in Relevance

Primary: answer the question asked

Children learn that not every interesting detail belongs in every answer.

Lower secondary: distinguish signal from distractor

Students identify which numbers, textual details, variables and historical facts bear on the task.

Upper secondary: use relevance to control evidence and argument

Learners evaluate whether evidence truly supports a claim and whether model variables change explanatory or predictive power.

Adulthood: allocate finite attention

Professional reasoning treats time, attention and information processing as scarce resources and directs them toward facts with the highest decision value.

Five Relevance Failures

1. Truth Dump

Many true facts are supplied without connection to the actual question.

2. Salience Capture

The most vivid information receives weight that belongs to quieter but more diagnostic information.

3. Assumption-Filtered Relevance

A wrong framing makes the right evidence look irrelevant.

4. Processing Waste

Cognitive effort is spent on information that cannot change belief or action.

5. Premature Discard

A weakly understood signal is labelled irrelevant before the problem state is clear.

Frequently Asked Questions

Is relevance the same as importance?

No. Importance concerns broad value or consequence. Relevance concerns whether information bears on the current question, goal or decision.

Is relevance the same as salience?

No. Salience concerns what stands out. Relevance concerns what matters to the current cognitive job. Something can be salient but irrelevant, or quiet but highly relevant.

Is relevance objective or subjective?

It is relational. It depends on a question, goal, receiver and state. But it is not arbitrary: evidence can show that some information changes the correct inference or action while other information does not.

What is Relevance Theory?

It is an influential framework developed by Dan Sperber and Deirdre Wilson, especially in cognition and communication, connecting relevance with cognitive effects and processing effort. It is one theory of relevance, not a universal definition of every relevance judgement.

How do I know whether information is relevant?

Ask whether it can materially change your answer, prediction, decision, hypothesis ranking or next action. If it cannot, it may have low relevance to the current task.

Can relevance change?

Yes. New goals, context, evidence and state changes can make previously irrelevant information relevant and vice versa.

Research Notes and Further Reading

For the classic relevance-theoretic account, see Sperber and Wilson’s Précis of Relevance: Communication and Cognition and the later discussion of testing the cognitive and communicative principles of relevance.

For a broader philosophical overview of pragmatics and relevance theory’s place within it, see the Stanford Encyclopedia of Philosophy entry on Pragmatics.

Cognitive Art deliberately uses a broader everyday-scientific bridge: relevance is the relation between information and the current cognitive job. This does not imply that all disciplines formalise relevance in the same way.

Final Thought: Information Is Cheap; Attention Is Not

The station contains thousands of facts.