Updated on
September 30, 2026
Exploratory Talk: Examples, Reflection Planner and PDF
Teach exploratory talk that tests ideas and uses evidence. Try the reflection planner and download an observation sheet for useful, inclusive feedback.

Exploratory talk is classroom dialogue in which learners test ideas aloud, give reasons and consider alternatives. It makes thinking available for others to question. The aim is a stronger explanation or decision, rather than agreement or a polished performance.
In a Year 5 science lesson, one learner might say, "The shadow might get longer because the light is lower," while a partner asks, "What evidence from our results supports that?" That small exchange matters because it moves talk beyond polite agreement. Teachers need to teach the ground rules, sentence stems, and listening routines that make this kind of tentative language rigorous, inclusive, and useful across subjects.
Mercer (1995) describes exploratory talk as dialogue that makes reasoning visible. Learners ask questions, listen to alternatives, give reasons, and are willing to change their minds. Littleton and Mercer (2013) link this kind of dialogue to classrooms where knowledge is built together, rather than given as finished answers.

Surface participation
“It is condensation.” The group agrees and moves on.
Reasoning to explore
“I think water comes from the air because the glass is cold. How could we test that?”
Follow-up check. Ask what evidence would support or challenge the explanation.
Show the class both responses. Ask which one could be tested, and what test the group could run.
Mercer (1995) explains why exploratory talk depends on taught routines, not on putting learners into groups and hoping discussion improves. Learners need to know how to offer a reason, ask for evidence, invite a quieter partner in, and disagree without closing the conversation down.
Exploratory talk works best after teachers have taught enough subject knowledge for learners to use. Learners need the vocabulary, facts, and worked examples first. Without this, confident speakers can dominate, while others repeat gaps in their knowledge. This is why dialogic teaching should sit alongside explicit instruction: Alexander (2020) stresses purposeful classroom dialogue, while Tricot and Sweller (2014) warn that domain-specific knowledge is central to successful problem solving.
Once learners have something worth discussing, talk helps them compare explanations and spot weak evidence. It also helps them move from first thoughts to better supported claims. Mercer, Wegerif and Dawes (1999) found gains in general reasoning after a Thinking Together intervention, but far transfer into history, maths, or science depends on the subject knowledge learners bring to the conversation. In 2026, this habit still matters when learners use AI tools: tentative language such as might, could, perhaps, and because helps them resist fluent but overconfident answers and keep claims open to evidence.
Vygotsky (1978) argued that learners first build knowledge through social interaction. They then internalise that learning as individual thought. Exploratory talk applies this idea in classroom dialogue. Learners try out explanations with peers, hear other views, and use language to reason, not just to report answers.
Exploratory talk uses classroom discussion to help learners understand (Mercer, 2000). Teachers can support it with thinking strategies. Mercer (2000) identified three types of talk: disputational, cumulative, and exploratory. In pairs, learners use exploratory talk to test and develop ideas.
Exploratory talk combines building on ideas with evaluating them. This means they build on ideas, but also test them. Mercer (2000) states that structured talk creates "interthinking". Learners use language to reason together and solve problems (Mercer, 2000; Littleton & Mercer, 2013).
Plan the conditions for talk before the task begins (Mercer & Littleton, 2007). Seat learners so they can hear each other, teach ground rules in plain language, and model teacher talk moves such as revoicing, prompting for evidence, inviting elaboration, and asking a learner to compare two explanations. Ground rules should value reasons and respectful challenge, but they should not require eye contact, rapid turn-taking, or standardised speech as proof of attention. This matters for EAL learners, SEND learners, neurodivergent learners, and learners who use regional or community dialects.
123Model one numbered step with a single group first, such as inviting a reason and then a question that tests it.
Start with short paired tasks where learners discuss a curriculum question they are ready to answer. Structure each round with prompts such as "What evidence supports this view?", "Can you build on that idea?", "What might someone who disagrees say?", and "Which word in the source makes you think that?" Increase complexity only when learners can give reasons, listen back, and revise an idea without the teacher carrying the discussion.
Mercer (1995) suggests that teachers model the language of reasoning before expecting learners to use it independently. After discussion, ask groups to identify one sentence that moved the thinking on, one challenge that improved the answer, and one voice that needs more space next time. Littleton and Mercer (2013) link this reflection with stronger collaboration.
Choose a task with a reason to compare answers. Learners need enough knowledge to offer more than guesses. A task with only one obvious response may be useful for checking recall, but it gives little room for exploratory discussion.
Classroom practice · Say It
| Use a partner’s response to revise a piece of writing | ||
|---|---|---|
| Starting point | Try in your lesson | Check the boundary |
| Ask partners to identify one claim that needs a clearer reason or example. Let the writer explain a choice, consider the response and decide what to revise. | The school post describes Year 3 / Grade 2 editing practice. It does not report attainment gains. Ask what changed and why, rather than counting comments. |
In a lesson on dissolving, show results from a fair test using water at different temperatures. Give pairs two claims: “Warm water dissolves the sugar faster” and “Warm water can dissolve any amount of sugar.” Ask them to decide which claim the results support. Keep the quantities, temperatures and timing visible so the discussion stays anchored in the evidence.
Give quiet thinking time before talk. Learners mark a result they would use and write or prepare a short explanation. One partner explains the evidence. The other asks, “Does that show how fast it dissolved, or how much could dissolve?” Partners can change their first answer. They should record what made them change it.
Listen for a distinction between rate and amount. If a group confuses them, refer it back to what the test measured. Do not treat a fluent explanation as correct simply because it sounds convincing. Finish with an individual question using a different result. That shows whether each learner can use the distinction independently.
Show two worked solutions to a fraction problem. One has an error that learners have already been taught to recognise. Ask, “Which step would you question, and why?” Give them a common representation, such as a fraction bar, so they can refer to something shared rather than relying on gestures or verbal fluency.
Ask the first partner to identify a step. Ask the second to justify keeping or changing it. Useful prompts include “Can you show that on the bar?” and “Would your method still work for this example?” The aim is a reasoned decision. Agreement without an explanation is not enough.
After the discussion, ask each learner to solve a similar problem alone. Compare the reasoning, not just the final number. If both partners repeat the same misconception, reteach the relevant step. More discussion will not repair missing knowledge by itself.
Choose a short passage with two plausible readings of a character's action. Give learners time to find one line for each interpretation. Ask, “Which interpretation fits the passage better, and what might challenge it?” Keep the text open throughout the task.
A learner might say, “She might be protecting her friend because she does not tell the teacher.” A partner could answer, “That fits this line, but could she also be worried about getting into trouble?” The teacher can ask what evidence would distinguish those readings. Tentative language is useful when it marks uncertainty about an interpretation. It should not turn a known fact into a guess.
Close with a short written judgement using one supporting line and one alternative explanation. Learners can retain different interpretations when both are supported. The aim is to make the reasoning clearer, not to make every group reach the same view.
A short Turn and Talk routine is useful for rehearsal. For a longer task, choose among the discussion structures according to the question and time available. The structure supports the reasoning; it does not substitute for it.
Ask for evidence that could challenge the current answer. Model a respectful question, such as “What would make that explanation less likely?” Do not manufacture disagreement about established facts. Where a task has a settled answer, learners can compare the reasoning behind two methods instead.
Teacher voices · Say It
the vital importance of the talk tasks
@BarbaraBleiman
Choose an explanation, interpretation or decision that learners can examine. A routine will not repair a vague task. Supply the knowledge and evidence they need to discuss it.
Start with preparation before public talk. A brief Everybody Writes task can help learners hold on to an idea. Offer another recording method where writing creates a barrier. Invite contributions through an agreed route instead of unexpectedly calling on a learner who needs a different communication arrangement.
Return to the subject knowledge. Ask learners to point to a result, line or worked example. A sentence stem does not make a claim accurate. Correct the misconception plainly, then offer a new chance to apply the corrected idea.
Pause and restate the shared question. Ask each speaker to explain the other's claim before responding. If the content is emotionally charged or a learner is distressed, stop the activity and follow the school's support arrangements. The aim is learning, not endurance under speaking pressure.
Accountable Talk offers a related way to focus on reasons and accurate knowledge. Questions by purpose can help you choose a follow-up that clarifies, checks or challenges a claim.
Sample a short exchange and record the question being discussed. Note a claim, the reason offered and what another learner did with it. Did they clarify, challenge, add evidence or revise the answer? Record whose ideas were taken up, including contributions made through writing, AAC or a partner.
Mini app · Say It
Choose one moment from a discussion. Build a small teaching experiment for your next lesson. This is an original reflection aid, not an assessment score.
The planner gives you a question to try and something specific to notice when you revisit the talk.
Quiet preparation can support thinking. Use agreed spoken, written, drawn or partner-supported contributions. A short observation cannot show what every learner understands.
Do not grade a learner from one visit. The topic, task, language demands and communication needs change what you can observe. Compare samples over time. A useful coaching question is, “What could we change so more learners can contribute to this particular reasoning task?”
An original reflection aid can help teachers plan and notice these behaviours. It is a formative prompt, not a validated oracy assessment or a diagnostic measure. Voice 21's framework offers a broader view of communication. Do not reduce that framework to counting spoken turns.
For the wider teaching approach, see the Dialogic Teaching guide. Wait time makes room to prepare a response. The introvert and extrovert guide cautions against treating quietness as a fixed ability label.
The evidence for exploratory talk comes mostly from classroom studies in which learners were taught ground rules for reasoning together. The studies are small, so they support teaching the routines carefully, not assuming that any group talk will help.
| Evidence | Useful finding | Limit for your classroom |
|---|---|---|
| Mercer, Wegerif and Dawes, 1999 | Teaching ground rules changed the quality of primary learners' talk and aspects of reasoning. | The study was small. It supports teaching the ground rules, not every sentence stem or group format. |
| Mercer and Sams, 2006 | Explicit language teaching supported reasoning in small-group maths work. | The gains came from a taught programme. Putting learners in groups without that teaching is a different thing. |
| Mercer and Littleton, 2007 | Brings the ground-rules studies together and explains how shared reasoning can be taught. | A synthesis of earlier work, not a new trial. Use it to plan the teaching, not to predict a result. |
For larger-scale evidence on classroom dialogue, including the EEF trial, see the Dialogic Teaching guide.
Before the next group task, write a question with more than one plausible answer, so there is something to reason about.
Tentative language leaves an idea open to revision. A learner might say “One possibility is…” before giving a reason. It helps when an interpretation is uncertain, but it does not replace evidence. Secure facts can still be stated clearly.
Exploratory talk examines reasons and evidence together. Disputational talk relies more on competing assertions or disagreement without developing the reasoning. An exchange can contain both, so use the distinction to reflect on the talk rather than label a learner.
It is talk in which learners share reasons, consider alternatives and test ideas together. Reasoning is made visible so others can question it. Mercer and colleagues distinguish this from uncritical agreement and from disputes that offer little shared reasoning.
Cumulative talk builds a shared response through agreement and additions. Exploratory talk also evaluates contributions. A learner might ask for evidence or revise an answer. Agreement can be productive, but it is not by itself evidence that the reasoning has been tested.
The early Thinking Together studies involved primary learners. Model a brief exchange, use familiar content and teach one ground rule at a time. Adapt the task and communication support rather than expecting an adult-style debate.
Use repeated observations of reasons, evidence, questions and responses, alongside an individual check of subject understanding. Do not use talk frequency, eye contact or accent as proxies for reasoning. A teacher-created checklist is a formative aid, not a validated assessment scale.
The PDF sheet helps you observe one group: notice the talk, reflect on it, then plan one next step. It is a formative aid, not a validated assessment scale.
Watch one group this week with the observation sheet. Note one moment that moved the thinking on, then plan the next step from it.