Continue the Task 3 scene, but change the speaking job
Task 4 shows the same illustration used in Task 3 and asks what you think will happen next. You have 30 seconds to prepare and 60 seconds to speak. The official study pack advises realistic imagination: predictions should grow from what people are doing, how they may feel, and the surrounding people and objects.
You are not asked to describe the scene a second time. A short reference to a visible clue is useful only when it supports a prediction. “The man is filling a watering can” becomes task-shaped when you add, “so he will probably carry it to the dry vegetable bed.”
Inference, development, probability language, and coherence
Content/Coherence includes several relevant predictions that are connected to supporting clues and organized logically. Vocabulary includes accurate verbs for action and a range of future and probability expressions. Listenability includes clear delivery, controlled grammar, and rhythm that lets the listener follow each clue-and-outcome pair. Task Fulfillment includes predicting what happens next, not simply retelling the present scene.
The task does not reward wild originality by itself. “The garden will be hit by a tornado” is possible in the broadest sense, but nothing in the demonstration scene supports it. A modest prediction can be stronger when the visual evidence and consequence are precise.
Match certainty to the clue. Someone reaching for a door is likely to open it; a person looking toward a bus may board it. This variation makes your reasoning more accurate and gives your language a natural range.
Prepare C → P → R pairs
- Clue.Select an unfinished action, a visible problem, an interaction, or an object ready to be used.
- Prediction.State the next likely action with language such as will probably, is likely to, may, might, or looks ready to.
- Result.Add a practical consequence or feeling where the evidence supports it. This develops the prediction instead of producing a list.
- Sequence.Link three or four pairs in a visual or causal order. Use “meanwhile” for a parallel event and “after that” for a later step.
For the garden scene, notes could be: can full → dry bed → plants watered; rake handed → soil loosened; bag carried → empty plot filled; seedlings → planted. The arrows hold the reasoning while keeping the notes short.
Extend the same original garden scene
Scene carried forward from Task 3: At a community garden, a man fills a watering can, children water a raised bed, a volunteer hands a rake to a new gardener, two people carry soil toward an empty plot, and a woman studies seedling trays.
30-second notes: water full → dry bed; rake → loosen soil; soil → empty plot; seedlings → choose/plant; group → plots ready.
Model response:
Several activities will probably move into their next stage. Once the man's watering can is full, he is likely to carry it to the dry section that the adult is showing the children. The children may then help him water that area, and they will probably be pleased when the whole bed is finished.
Near the shed, the new gardener has just received a rake, so she will likely begin loosening the soil in her plot. Meanwhile, the two people carrying the heavy bag will probably pour the soil into the empty bed and spread it out together.
Finally, the woman beside the seedling trays may choose the healthiest plants and bring them to one of the prepared plots. By the end of the morning, several beds should be ready for the new growing season.
Why it works: Each outcome is tied to a visible unfinished action. The response uses different probability levels and connects separate predictions into the shared progress of the garden.
What weakens a prediction response
- Describing for most of the minute. Present action is repeated without moving to what happens next.
- Making unsupported leaps. The future event has no visible person, object, or situation behind it.
- Using only “will.” Every prediction sounds equally certain even when the evidence differs.
- Listing outcomes without reasons. The listener cannot see how the prediction was formed.
- Repeating one action. Several sentences predict that people will continue exactly what they are already doing.
- Ignoring consequences. An action is predicted but never developed into a result, reaction, or next stage.
- Inventing names or private motives. Unsupported backstory consumes time and weakens relevance.
Record three evidence chains
Return to the garden scene. Prepare three C → P → R chains in 30 seconds. Record for 60 seconds and use at least three different forms of probability language.
Reveal the self-check
Write down each prediction from your recording. Beside it, identify the exact visible clue and the result you explained. If a row lacks a clue, replace the prediction. If a row lacks a result, add one practical consequence rather than another unrelated future action.
Check whether the future follows from the scene
- Did I move beyond present description quickly?
- Does every prediction connect to a visible clue?
- Did I make several distinct predictions?
- Did I develop predictions with likely results or feelings?
- Did I vary probability language appropriately?
- Did I organize events by place, time, or cause?
- Did I avoid unsupported names, motives, and dramatic events?
- Was my delivery clear and complete within 60 seconds?