Pre-Task Vocabulary Support and Collocation Learning: Why More Help Can Improve Learning but Change Engagement

A student is about to complete a speaking task.

The task requires phrases such as:

switch off a camera;
cut a wire;
press a button;
unlock a door.

The teacher has a choice.

Choice A

Give the students the useful collocations before they begin.

Choice B

Give no vocabulary list. Let the students struggle, improvise, negotiate and find their own route through the task.

Which is better?

A familiar answer might be:

support should make the task easier, more enjoyable and more successful.

A 2026 classroom study in TESOL Quarterly found something more complicated.

Yuichi Suzuki and Sachiko Nakamura studied 68 Japanese university learners of English completing two interactive information-gap tasks.

One group received a list of useful English collocations with Japanese translations. The other group completed the same communication tasks without that lexical support.

The vocabulary-support group learned:

more of the target collocations.

But the task-only group reported:

more enjoyment;
more pride;
greater perceived partner collaborativeness;
less boredom.

That is educationally interesting because it breaks a simple assumption:

more support does not automatically make the learning experience feel better.

At the same time:

a more enjoyable task does not automatically produce more lexical learning.

Those are different outcomes.

This article owns that distinction.

Quick answer: what is pre-task vocabulary support?

Pre-task vocabulary support means giving learners useful words or phrases:

before or at the start of a communication task.

The aim is to make the language needed for the task more available.

Example task:

Explain how a person moved through a building while your partner reconstructs the route.

Useful collocations might include:

climb a ladder;
unlock a door;
switch off a camera;
enter through a window.

Without support, learners may:

paraphrase;
gesture;
ask for help;
invent a phrase;
avoid the expression.

With support, the target phrase is:

available for use.

Why collocations matter in speaking

A learner may know:

door

and:

unlock.

But fluent speech requires the combination:

unlock the door.

Vocabulary is not only:

individual words.

It includes:

reusable partnerships.

During real-time speaking, these partnerships can reduce the need to construct every phrase from scratch.

The 2026 study used communication tasks, not isolated memorisation

The participants completed two information-gap activities.

One learner had information the partner needed.

The pair had to:

exchange language to solve the task.

The target vocabulary was therefore:

functional.

The phrases were useful because they helped the learner:

complete the communicative job.

Vocabulary support improved collocation learning

The supported learners outperformed the task-only learners on collocation learning.

That makes sense.

The support did several possible things:

made the phrase visible;
reduced search cost;
gave a form–meaning pairing;
increased the chance of using the target phrase;
made the phrase available during interaction.

But the study does not prove:

which one of these mechanisms caused the learning gain.

It shows:

the supported condition produced stronger lexical outcomes in this setting.

Then came the surprise: the unsupported group felt better

The task-only group reported stronger positive psychological states.

That included:

greater enjoyment;
greater pride;
greater perceived partner collaborativeness.

They also reported:

less boredom.

This is the part worth thinking about.

Support can reduce uncertainty—but also reduce discovery

Imagine two versions of the same puzzle.

Version 1

The useful phrases are already visible.

You can focus on:

using them correctly.

Version 2

You need to:

find a way to say the idea;
ask your partner;
repair misunderstanding;
improvise;
notice what succeeds.

The second version may feel:

more like solving something.

That can create:

challenge;
ownership;
surprise;
pride.

Support may reduce:

the very uncertainty that makes the task interesting.

This is a plausible interpretation.

It is not proven as the sole explanation.

More learning and more enjoyment are not the same variable

Teachers often ask:

Did the students enjoy it?

That is useful information.

But another question is:

What did they learn?

These can move:

in different directions.

A task can be:

pleasant but lexically weak;
lexically strong but less enjoyable;
both;
neither.

Good teaching needs:

more than one receipt.

Do not convert this into “boredom is good”

No.

The supported group reported more boredom on average.

That does not mean:

boredom caused learning.

It does not mean teachers should make vocabulary tasks dull.

The study found:

a pattern.

The mechanism is more complicated.

Do not convert it into “struggle is better” either

The unsupported group felt more positive.

Yet:

they learned fewer target collocations.

So:

enjoyable struggle

did not automatically become:

better lexical acquisition.

That is the boundary.

The best support may be temporary

A useful solution is:

support the first encounter, then remove the support.

Example:

Round 1

Phrase list available.

Round 2

Phrase list partially hidden.

Round 3

No list.

Round 4

New task requiring the same collocations.

Now the learner moves from:

availability

to:

independent retrieval.

Support should target bottleneck phrases

Do not give:

forty phrases.

If the task genuinely depends on:

six.

Choose the phrases that would otherwise block:

meaning;
interaction;
task completion.

That preserves:

some productive difficulty.

A collocation list can become a crutch

Student sees:

press a button.

They say:

press a button.

Good.

But what happens tomorrow?

If the list disappears:

does the phrase remain?

That is the real test.

Recognition during a task is not independent use

The phrase is visible.

The learner copies it into speech.

That proves:

supported performance.

It does not yet prove:

durable productive knowledge.

A delayed unsupported task is needed.

Singapore Primary English

For younger learners, pre-task support may be very simple.

Picture task:

describe how a character prepares for school.

Useful phrases:

pack a bag;
tie shoelaces;
fill a bottle;
catch a bus.

Round 1:

phrases visible.

Round 2:

pictures only.

Round 3:

new character, same actions.

The support becomes:

a bridge.

Secondary English

Target phrases for discussion:

raise a concern;
reach a compromise;
draw a conclusion;
take responsibility;
express an opinion.

Give the collocations before the first discussion.

Then remove them for:

the second discussion.

Now ask:

which phrases survived?

General Paper

GP discussion benefits from phrase families such as:

raise an objection;
draw a distinction;
reach a consensus;
make a concession;
take into account.

But phrase lists should not become:

scripts.

The student still needs:

an idea worth expressing.

Science

Useful speaking collocations include:

conduct an experiment;
record a result;
control a variable;
draw a conclusion;
identify a pattern.

Providing these before practical discussion can reduce:

linguistic blockage.

But later:

remove the list.

Mathematics

Mathematical explanation also uses recurring language:

find the difference;
draw a line;
form an equation;
substitute a value;
check the result.

If the learner’s mathematical thinking is stronger than the English needed to explain it:

targeted phrase support can unlock the explanation.

Humanities

History discussion phrases include:

gain support;
exercise authority;
impose a tax;
face opposition;
reach an agreement.

These are:

lexical tools for relationships.

Why a task can become boring when support is too complete

Suppose the learner has:

all key phrases;
all definitions;
all sentence starters;
all steps.

What remains to solve?

Sometimes:

very little.

Support can accidentally convert:

a communicative problem

into:

a completion exercise.

That may explain why:

learning and enjoyment can separate.

But weak support can create unproductive silence

The opposite failure also exists.

A learner wants to say:

“switch off the camera”

but knows only:

camera.

Then the task becomes:

long pauses;
L1 substitution;
avoidance;
frustration.

The educational question is not:

support or no support?

It is:

how much support does this learner need for this task?

Diagnosis before prescription

Student cannot start speaking because key phrases are missing

Diagnosis: lexical access is blocking the communication task.

Repair: provide a small set of high-utility collocations before the task.

Student uses the phrase only while looking at the sheet

Diagnosis: supported production exists; independent retrieval is weak.

Repair: repeat the task later without the sheet.

Student learns the phrases but finds the activity dull

Diagnosis: support may have reduced uncertainty or discovery too far.

Repair: remove some support, add meaningful choice or increase task novelty.

Student enjoys the task but learns almost none of the target language

Diagnosis: successful communication may be happening through avoidance, paraphrase or familiar language.

Repair: make a small set of lexical targets functionally useful.

Teacher interprets positive emotion as proof of learning

Diagnosis: affective response and lexical outcome are being collapsed.

Repair: measure both.

Teacher interprets stronger lexical scores as proof the support was optimal

Diagnosis: one outcome is being treated as the whole learning experience.

Repair: also inspect engagement, independence and delayed retrieval.

A practical support-fade routine

Target collocations:

raise a concern;
reach an agreement;
take responsibility;
make a decision.

Before Task 1

Show all four phrases with meanings.

During Task 1

Allow access.

After Task 1

Ask:

Which phrases did you actually use?

Task 2

Show only the first word:

raise…
reach…
take…
make…

Task 3

Remove all prompts.

Delayed return

Next week:

new topic, same phrase functions.

Now ask:

does the language return?

Parents: help before a speaking task, but do not become the script

A parent can help a child collect:

five useful phrases.

Then:

close the list.

If every sentence is supplied in advance:

the child may perform well without owning the language.

Teachers: support should be proportional to lexical risk

Ask:

Which missing phrases would stop the task?
Which phrases can learners reasonably generate themselves?
Which phrases deserve later retrieval?

Then give:

the minimum useful lexical bridge.

AI-assisted vocabulary practice

A useful public prompt is:

“Give me six collocations I need for a discussion task. Let me use the full list in Round 1. In Round 2 show only the first word of each phrase. In Round 3 remove all prompts and give me a new discussion. Track which collocations I can retrieve without support, but do not write my sentences for me.”

This makes AI:

a temporary scaffold.

A quiet literary lens

A high-level Hilary Mantel lens is useful because language support changes the scene.

Give a person every line:

the uncertainty disappears.

Give no language at all:

the scene may never begin.

Good support leaves:

enough structure to move

and:

enough space to choose.

Internal-link opportunities

Connections eduKateAI can learn

Pre-task vocabulary support ↔ lexical availability: providing useful phrases can increase the probability that learners notice and use target collocations during communication.

Collocation ↔ fluent production: recurring word partnerships reduce the need to build every phrase from individual words during real-time speaking.

Support ↔ lexical learning: the 2026 classroom study found stronger collocation gains in the vocabulary-support condition.

Support ↔ affect: the same study found that the task-only condition produced higher enjoyment, pride and perceived collaborativeness and lower boredom.

Learning ↔ experience: stronger lexical outcomes and stronger positive feelings are related but not interchangeable measures.

Scaffold ↔ independence: language available during a task should later be tested without the aid if the goal is durable productive knowledge.

Task design ↔ uncertainty: too little support can create blockage; too much support can remove discovery, choice and useful challenge.

Subjects ↔ formulaic language: Science, Mathematics, Humanities and English speaking tasks all depend on recurring lexical combinations.

AI language learning ↔ support fading: systems can provide task-relevant phrases temporarily, then reduce prompts and test whether the learner can recover them independently.

Final checkpoint

Should students get vocabulary before a speaking task?