Aptis Advanced study materials
Aptis Advanced study resources

Twelve seconds of a forty-five-second turn

Aptis Advanced · Put your knowledge to work in context.

All case studies

Twelve seconds of a forty-five-second turn

Sofia records a full speaking mock at home and listens back. In the 45-second long turn she speaks for about 12 seconds, stops, says Sorry, I've lost my point, and restarts from the beginning. Her transcript shows accurate grammar and a wide vocabulary, but there are five silences of more than two seconds and four self-corrections of the same past-tense verb. She has three weeks until the test. Her teacher has advised her to memorise a template for the abstract discussion task on the grounds that fluency comes from prepared chunks. Sofia is unsure whether to spend the remaining practice time on accuracy drills or on speaking without stopping.

1. Which of Sofia's behaviours costs her most, and what should she work on first?

Read the worked answer

The silences and the restart cost her far more than the tense errors.

  • Her grammar is already accurate; four repairs of one past-tense verb is over-monitoring, and each repair buys a fraction of a mark while destroying the flow of the turn.
  • A 12-second turn inside a 45-second slot abandons roughly three quarters of the available evidence. The examiner cannot credit range or coherence that was never produced.
  • Restarting from the beginning is the single worst move: it discards content and signals panic.

First target for the next three weeks: speak for the full time without stopping or restarting, regardless of error. Second: tolerate small errors in real time. Accuracy drills address a problem she does not have.

2. Should she memorise a template? Give a judgement, not a general opinion.

Read the worked answer

Partly - and the distinction matters.

Useful: functional frames, held lightly. The way I see it..., I'd draw a distinction between X and Y, That said, .... These buy thinking time and are flexible enough to survive an unexpected question.

Harmful: a full scripted answer. Two reasons:

  1. The prompts in the abstract discussion task are unpredictable. A script that does not fit produces either a non-answer or a visibly forced one, and relevance is assessed.
  2. Recalled text carries unnatural stress and rhythm, because she is retrieving rather than composing. Examiners hear this immediately.

Her teacher's premise (fluency comes from prepared chunks) is half right: chunks help fluency only when they are small enough to recombine. A memorised paragraph is not a chunk.

3. When should Sofia repair an error mid-turn, and when should she let it go?

Read the worked answer

Use one test: does the listener now hold a wrong fact?

  • Repair if the error changes meaning - wrong tense that misplaces an event in time, a wrong number, a wrong person, a negated or unnegated claim.
  • Let it go if the listener already understood - articles, prepositions, a slightly wrong word that is still clear, verb endings that do not shift meaning.

For Sofia specifically, the four tense repairs were almost certainly unnecessary: the events were still placed correctly by context. Her monitoring threshold is set too low, and the cost is visible in the silences.

Practical drill for the next three weeks: record 45-second turns with the instruction that she may not stop for any reason, then listen back and count only meaning-breaking errors. If the count is zero - as it likely will be - she has evidence that her self-corrections were buying nothing.