A hint can rescue the answer and still weaken the evidence that the learner owns the answer.
A teacher says, “What formula should you use?” A worksheet underlines the key word. A tutor points to the line where the error occurred. A learning app reveals the first step. The learner succeeds.
The success is real, but it may belong to a combined system: learner + hint.
Hint dependence occurs when external support repeatedly performs part of the cognitive job that the learner should eventually be able to perform alone — retrieval, cue recognition, method selection, monitoring or recovery.
This article sits beneath How Feedback Works, How Retrieval Works, How Worked Example Fading Works and the existing How Confidence Works | Borrowed Confidence. The narrow question is: when does useful help stop being a bridge and start becoming part of the learner’s dependency?
1. A Hint Can Carry Different Functions
Not all hints do the same work.
- Retrieval hint: reminds the learner of a fact or rule.
- Attention hint: points to the relevant feature.
- Method hint: names the procedure to use.
- Step hint: supplies the next operation.
- Error hint: identifies where the solution went wrong.
- Metacognitive hint: prompts the learner to check, compare or justify.
Dependence risk is highest when the hint repeatedly supplies the very function the assessment later expects the learner to perform independently.
2. Immediate Performance Can Improve Faster Than Learning
If a hint gives the method, the learner can produce a correct answer before the method-selection capability is learned.
This makes supported practice look successful while unsupported transfer remains weak.
The distinction is the same one that appears throughout learning science: performance during instruction is not identical to durable independent capability.
3. Hints Can Become Retrieval Cues
Memory is cue-dependent.
If a learner always retrieves a formula after seeing a colour, phrase or teacher question, that external cue can become part of the retrieval route. Remove it and access slows or fails.
The existing How Retrieval Works | Retrieval Cues article owns the broader cue mechanism.
4. Method Labels Can Hide Classification Weakness
A worksheet titled “Simultaneous Equations by Elimination” has already performed the classification step.
The learner only needs to execute elimination. In a mixed examination, the learner must first recognise that simultaneous equations are present and choose among methods.
Practice organised by method family can therefore build procedural fluency while leaving method selection underdeveloped.
5. Leading Questions Can Become Invisible Scaffolds
Teachers often help through questions: “What do you know?” “Which angle rule applies?” “What should the denominator be?”
These questions are valuable when they teach a self-questioning routine. They create dependence when the learner waits for the teacher to generate the questions forever.
The long-term goal is internalisation: the learner eventually asks the diagnostic question without needing the teacher to say it aloud.
6. Hint Timing Matters
A hint delivered immediately after hesitation can prevent the learner from making a genuine retrieval attempt.
Waiting too long can allow frustration and unproductive guessing to dominate.
A useful timing rule is to preserve a meaningful attempt before intervention, then give the smallest hint that can restart productive reasoning.
7. Hint Size Matters
“Check the sign on line three” carries less solution state than “subtract equation two from equation one.”
Smaller hints preserve more learner ownership when they are sufficient.
A hint ladder can move from broad metacognitive prompt to local cue to partial step to explicit solution only when earlier levels fail.
8. Feedback Contamination Makes Learning Harder to Measure
Once the teacher has supplied enough information to correct the response, the next attempt no longer measures the learner’s original independent state.
This is not a reason to withhold help. It is a reason to separate teaching trials from measurement trials.
The existing How Feedback Works | Feedback Contamination article owns that assessment problem.
9. Help-Seeking Can Be Productive or Avoidant
Requesting help is a legitimate learning strategy.
Productive help-seeking occurs after a real attempt and asks for enough information to continue. Avoidant help-seeking bypasses the reasoning step whenever difficulty appears.
Learning systems should therefore reward good help-seeking without making hint access the fastest route to every answer.
10. Adaptive Systems Can Track Hint Use
Digital tutors can record how many hints were requested, how quickly, at which step and whether performance remains correct when hints are removed later.
Those signals can distinguish a learner who occasionally needs a cue from one whose apparent mastery exists only under continuous support.
Hint count alone is not enough. Some tasks are harder and legitimately require more support. The important pattern is dependence relative to task and later independence.
11. Fading Hints Turns External Questions Into Internal Questions
One useful sequence is:
- teacher asks the full diagnostic question;
- teacher uses a shorter prompt;
- teacher points to a checklist item;
- learner uses the checklist independently;
- checklist disappears;
- learner performs the diagnostic routine internally.
The support succeeds when the learner inherits its function.
12. Unsupported Retesting Is the Proof
The only reliable way to know whether hint dependence has reduced is to remove the hints and retest.
Do this after a delay and with changed surface features. Immediate repetition can be supported by short-term memory of the hint itself.
The existing Feedback Retesting route owns the broader correction-survival question.
13. Worked Example: Mathematics
A tutor always asks, “Can you factorise first?” whenever a quadratic equation appears.
The student becomes excellent at factorising after the prompt and still fails mixed questions because recognition of factorisable structure was never practised independently.
The repair is to fade the method cue and mix quadratic forms so classification becomes part of the learner’s job.
14. Worked Example: Reading
A teacher highlights the sentence containing the answer before every comprehension question.
The learner practises extraction and not information location. Remove the highlighting and comprehension performance falls.
Support should gradually move from highlighted sentence to paragraph cue to independent search.
15. Worked Example: Vocabulary
A learner can use a new word whenever a synonym is provided in brackets.
The synonym acts as a retrieval cue. Later practice should remove it and vary sentence context so the learner retrieves meaning and collocation from the word itself and the surrounding discourse.
16. Worked Example: Sport
A coach shouts tactical instructions during every drill.
Players execute correctly while the coach supplies perception and decision. Match performance remains poor because the cue-detection function never transferred.
The coach can fade instructions and ask players to call the cues themselves.
17. A Hint-Dependence Checklist
- Identify what cognitive function the hint is carrying.
- Preserve a genuine learner attempt before intervention.
- Give the smallest useful hint.
- Use a graded hint ladder rather than jumping to the solution.
- Teach the self-question the hint represents.
- Track how quickly and how often help is requested.
- Fade hints as competence improves.
- Separate teaching trials from independent measurement trials.
- Retest after delay without support.
- Use mixed and changed-context tasks to prove independence.
18. Read the Mechanism Forward, Backward and Sideways
Forward: difficulty → hint → immediate success → repeated external cue → possible dependence → fading → unsupported retrieval. Backward: start from a learner who fails only when help disappears and identify which function the help had been supplying. Sideways: compare learner, teacher and assessment. The teacher sees a successful interaction; the assessment asks whether the same success survives when the interaction is absent.
19. The Civilisation Lesson
Support systems are strongest when they increase capability rather than permanently substitute for it.
Education, rehabilitation, public services and intelligent tools all face the same design question: which assistance should remain available, and which assistance should gradually disappear because the user can now carry the function independently?
A good hint changes the next attempt. A great hint eventually teaches the learner how to generate the hint for themselves.
Continue through How Worked Example Fading Works, Feedback Contamination and the How X Works hub. Next: desirable difficulty — why some effort makes learning stronger while other difficulty simply consumes the learner’s limited capacity.