You throw a ball at a wall.
It comes back lower than you expected.
You throw again.
This time you change the angle.
The wall did not explain anything.
It simply returned information about what your action produced.
That return is feedback.
Human beings learn by acting into a world that answers back.
Quick Read
Feedback is information returned after an action, prediction, attempt or state change that helps a system evaluate what happened and adjust what it does next.
Feedback can come from:
- the physical world,
- a teacher,
- a score,
- a sensor,
- a conversation partner,
- a failed prediction,
- the consequences of our own action.
One-sentence answer: Feedback is returned information about the consequences or quality of an action that can change the next action, model or state.
Feedback Is a Return Path
Without feedback, action becomes open-loop.
You act.
Nothing comes back.
You do not know whether the system responded as expected.
With feedback, the loop closes:
state → action → outcome → feedback → updated state → next action.
The return path turns action into learning.
Feedback Is Not the Same as Correction
A thermometer says 39°C.
That is feedback about temperature.
It does not correct the temperature.
A teacher writes:
Your explanation does not connect the evidence to the claim.
That is feedback.
The student still has to revise.
Feedback informs correction.
It is not correction itself.
Feedback Is Not the Same as Praise
“Good job” can feel encouraging.
But what should the learner do differently next?
Nothing is specified.
Useful feedback contains information about performance, process, target or next step.
Praise can coexist with feedback.
It should not replace it.
Feedback Needs a Reference
“You wrote 600 words.”
Good or bad?
Depends on the target.
If the task asked for 500–700 words, fine.
If the task asked for 150 words, the same observation means something else.
Feedback compares actual performance with:
- a goal,
- a prediction,
- a standard,
- a prior state,
- an expected trajectory.
The World Gives Feedback Whether We Ask for It or Not
A chair collapses.
An argument persuades nobody.
A route takes twice as long as expected.
A plant wilts.
Reality returns consequences.
The difficulty is interpretation.
The chair may have failed because of poor material, poor joint design, misuse or accumulated wear.
Feedback says something happened.
Diagnosis explains why.
Feedback Can Be Immediate
Touch a hot surface.
Immediate.
Miss a basketball shot.
Immediate.
Type a wrong password.
Immediate.
Immediate feedback can support rapid adjustment when the task is tightly coupled to observable consequences.
Feedback Can Be Delayed
A long-term investment.
A policy.
A curriculum change.
A student’s study strategy.
Consequences may return weeks, months or years later.
Delay creates an attribution problem.
Which earlier action caused the current result?
Long feedback loops make learning harder because many other variables can change before the outcome arrives.
Fast Feedback Is Not Always Better
If a learner receives the answer instantly after every tiny difficulty, they may never attempt retrieval, diagnosis or self-correction.
If feedback is delayed too long, the original state may be forgotten.
Timing should match the learning job.
Valerie Shute’s influential review Focus on Formative Feedback emphasises that feedback effectiveness depends on features such as timing, specificity, task and learner characteristics rather than one universal delivery rule.
Verification Feedback
Correct.
Incorrect.
Score: 7/10.
This is verification.
It tells the learner whether performance met a criterion.
It may not explain what to change.
Elaborated Feedback
“Your answer is wrong because you treated total distance as one leg of the triangle. Redraw the diagram and label the horizontal and vertical components separately.”
Now the feedback carries mechanism and next action.
More information is not automatically better, but useful elaboration can target the exact representation that needs repair.
Good Feedback Reduces the Right Uncertainty
A student knows the answer is wrong.
What do they not know?
- why it is wrong,
- where it went wrong,
- what the correct principle is,
- what to do next.
Feedback is useful when it reduces uncertainty at the level needed for the next transition.
Feedback and Self-Regulation
Deborah Butler and Philip Winne’s classic synthesis Feedback and Self-Regulated Learning places monitoring at the heart of self-regulation.
The learner sets a goal.
Acts.
Monitors progress.
Generates internal feedback.
Receives external feedback.
Adjusts strategy.
Self-regulated learning is a feedback architecture.
Internal Feedback
You write a sentence and think:
That sounds wrong.
No teacher has spoken.
Your own monitoring system compared output with an internal standard.
Expertise increases the quality of internal feedback because the expert owns richer standards and error detectors.
External Feedback
A teacher points out that your essay evidence is irrelevant.
A coach shows that your body position is reducing power.
A customer says the interface is confusing.
External feedback can reveal mismatches invisible to the actor.
This is why strong systems combine internal monitoring with outside measurement.
Feedback Can Be Wrong
A teacher can misdiagnose.
A sensor can drift.
An audience can dislike something for reasons unrelated to quality.
A market can reward bad behaviour temporarily.
Feedback is evidence.
Evidence must still be evaluated.
Feedback Quality Has Dimensions
- Accuracy: is it correct?
- Relevance: does it bear on the target?
- Specificity: does it locate the issue?
- Timing: does it arrive while revision is possible?
- Actionability: can the receiver do something with it?
- Credibility: is the source trustworthy?
Feedback can be abundant and still useless if it fails these dimensions.
The “Too Much Feedback” Problem
A page returns twenty comments.
The student fixes none.
Why?
Feedback has become another cognitive load.
Good feedback often needs priority.
Fix the load-bearing error first.
Then the next one.
Feedback Must Match the Level of Error
Sentence-level feedback cannot repair a thesis-level problem.
Arithmetic feedback cannot repair a modelling error.
Motivational feedback cannot replace missing knowledge.
Hierarchy matters.
Feedback should enter at the level where the representation diverged.
The Next-Move Test
After giving feedback, ask:
Does the receiver now know what to do differently?
If not, you may have delivered evaluation rather than feedback.
Feedback in Mathematics
“Wrong answer” is low-resolution feedback.
“Your equation is correct until line three, where dividing by x loses the x = 0 case” is high-resolution feedback.
The second identifies the transition that broke the invariant.
Mathematical feedback becomes powerful when it connects:
error → violated principle → corrected representation → next check.
Feedback in English
“Use better vocabulary” is vague.
“Your word choice is accurate, but the paragraph lacks a causal link between the example and your thesis” targets structure.
Strong English feedback distinguishes language surface from argument architecture.
Feedback in Science
A model predicts one result.
The experiment returns another.
The world has given feedback.
But science must decide whether the mismatch came from:
- measurement error,
- random variation,
- wrong assumption,
- wrong model,
- changed state.
Scientific feedback is interpretation under uncertainty.
Feedback in Organisations
A dashboard reports sales.
A customer service system reports complaints.
Employees report process friction.
These are feedback channels.
An organisation fails when the channel exists but nobody can safely act on the signal.
Silenced Feedback Creates Beautiful Failure
If people are punished for reporting problems, reports improve while reality worsens.
The dashboard becomes green because the sensors became afraid.
Feedback systems require psychological and institutional permission to carry bad news.
Positive and Negative Feedback: A Systems Distinction
In systems language, negative feedback often refers to feedback that counteracts deviation and stabilises a system.
A thermostat notices temperature below target and adds heat.
Positive feedback amplifies change.
More of something creates conditions for still more.
These technical meanings are not the same as “good feedback” and “bad feedback.”
Reader literacy requires keeping the meanings separate.
Feedback and Error
Error is the mismatch.
Feedback is information returned about that mismatch.
A system can make an error without receiving useful feedback.
A student answers incorrectly but never sees the mark scheme.
The error occurred.
The learning loop stayed open.
Feedback and Evidence
Feedback is evidence about consequences.
One failed attempt is evidence.
Repeated failure under the same condition is stronger evidence.
Independent feedback channels reduce the chance that one faulty sensor defines the state.
Feedback and Confidence
Feedback calibrates confidence.
Without outcome information, overconfidence can persist.
A prediction ledger works because it closes the loop between forecast and reality.
Feedback and Revision
Feedback does not matter until it enters revision.
The next article in this batch asks exactly that question:
How should a mind change its model without throwing everything away?
The Feedback Audit
- What action or prediction produced this feedback?
- What reference or target is being used?
- Is the signal accurate?
- Is it specific enough to locate the issue?
- Does it arrive while revision is still possible?
- What uncertainty does it reduce?
- What is the highest-leverage next move?
- Is the receiver able and permitted to act on it?
- What would confirm that the adjustment worked?
- Does the loop close again?
A Practical Exercise: Turn Evaluation Into Feedback
Start with:
This essay is weak.
Now transform it:
- What is weak?
- Where?
- Relative to what criterion?
- What should change?
- How will the learner test the revision?
Feedback becomes useful when the next state is imaginable.
A Practical Exercise: Close One Personal Loop
Choose one repeated behaviour.
Studying.
Exercise.
Writing.
Budgeting.
Define:
- target,
- action,
- observable outcome,
- review time,
- adjustment rule.
You have built a feedback loop.
A Primary-to-Adult Progression in Feedback
Primary: know what happened
Children learn to compare attempt with target and make simple corrections.
Lower secondary: use feedback diagnostically
Students identify whether feedback concerns knowledge, method, evidence, execution or checking.
Upper secondary: generate internal feedback
Learners increasingly evaluate their own work using standards, counterexamples, estimates and independent checks.
Adulthood: design return paths
Professional systems need measurements, audits, user reports, monitoring and safe escalation so action never becomes detached from consequence.
Five Feedback Failures
1. Feedback Without Reference
The receiver hears a judgement but cannot see the criterion.
2. Feedback Flood
Too many comments overwhelm the next move.
3. Feedback Delay
The signal arrives after the relevant state has disappeared or the decision is irreversible.
4. Wrong-Level Feedback
Surface corrections are applied to structural errors.
5. Feedback Without Revision
Information returns, everybody nods, and nothing changes.
Frequently Asked Questions
What is feedback in simple terms?
It is information returned after an action or attempt that tells you something about the result and can guide the next attempt.
Is feedback always external?
No. Learners and experts generate internal feedback by monitoring their own performance against standards and expectations.
Is faster feedback always better?
No. Timing depends on the learning objective. Immediate feedback can support correction, while delayed feedback can sometimes allow productive retrieval and self-monitoring. Excessive delay can make attribution difficult.
What makes feedback actionable?
It identifies the relevant gap, connects it to a criterion and gives enough information for the receiver to make a useful next move.
Can feedback harm learning?
Yes. Inaccurate, excessive, poorly timed or person-focused feedback can distract, demotivate or create dependence. Feedback effectiveness depends on design and context.
Why is feedback important for self-regulated learning?
Because learners need information about progress toward goals in order to adjust strategies, effort and attention.
Research Notes and Further Reading
Valerie Shute’s Focus on Formative Feedback remains a major review of feedback design, including specificity, timing, learner characteristics and task effects.
Deborah Butler and Philip Winne’s Feedback and Self-Regulated Learning: A Theoretical Synthesis places monitoring and feedback at the centre of self-regulation. A newer meta-analysis of digitally delivered instructional feedback, A Meta-Analysis of the Effects of Context, Content, and Task Factors of Digitally Delivered Instructional Feedback on Learning Performance, reports an overall positive effect while also showing substantial variation by feedback design and context.
The reader-facing Cognitive Art definition is broader than instructional feedback. It includes physical, social and system return signals while keeping technical uses distinct.
Final Thought: The World Must Be Allowed to Answer Back
You throw the ball.
The wall returns it.
You adjust.
That tiny loop contains something enormous.
Every useful model needs a return path to reality.
Every learner needs a return path from attempt to consequence.
Every organisation needs a return path from policy to outcome.
Every confident mind needs a return path from belief to evidence.
Without feedback, projection becomes fantasy.
With feedback, action becomes revisable.
And revisability is how intelligence keeps contact with the world.