A flashcard that reveals its answer immediately can feel wonderfully fast. The learner sees a picture of a kettle, taps, reads “kettle,” and moves on. Twelve cards disappear in a minute. Yet the useful event may never occur: the mind gets no time to search for the English word. A short, deliberate gap before the answer appears turns exposure into retrieval practice.
Six seconds is a sensible starting point.
What happens inside the gap
Retrieval begins with a cue. The cue might be a picture, a Mandarin meaning, a definition, a sound, or a sentence with one blank. During the gap, the learner tries to bring the target form back. Success gives the memory route another use. A failed attempt also provides information, because the answer arrives while the missing word is still the focus of attention.
Compare this with rereading a visible pair five times: the learner may recognise the familiar arrangement, feel that the word is known, and discover during tomorrow's conversation that the English form still will not come forward without the printed answer beside it.
The gap also gives the teacher a diagnostic moment. A quick answer, a hesitant partial form, and a complete blank are three different states. Software can record that difference with confidence buttons or response time, while a partner can simply mark ✓, ~, or 0 on paper. The marks should decide what returns later. They are more informative than the number of cards swiped.
For spoken practice, the learner should answer aloud rather than merely deciding that the word feels familiar. An audible response exposes missing syllables and uncertain stress, and it gives a partner something definite to confirm before the accurate recording or teacher model is played.
Joe Barcroft tested this small timing change with English speakers learning Spanish. Participants first viewed 24 picture and word pairs. During later presentations, 12 pictures appeared six seconds before the Spanish word, leaving time for an attempted retrieval. The other 12 pictures and words appeared together. The retrieval condition produced higher scores immediately, two days later, and one week later, although scores declined over time in both conditions.
The result supports a narrow claim about a cue-answer delay. It does not show that six seconds is ideal for every learner, every word, or every stage of study. The participants were learning 24 concrete Spanish nouns from pictures. Abstract words such as “accountability” need richer cues, and a beginner who has seen a form once may have too little memory to search.
Still, the contrast is useful. A card can look active because it requires a tap. The mental action happens before the tap, while the answer remains hidden.
Match the cue to the skill
Retrieval has a direction. Seeing “kettle” and recalling its meaning practises receptive knowledge. Seeing a kettle and producing the word practises productive knowledge. Learners often need both, though their immediate goal should decide which direction receives more time.
Akifumi Yanagisawa compared receptive retrieval, productive retrieval, and repetition with 18 Japanese university students learning 24 English pseudowords. On the receptive test, receptive retrieval beat repetition and matched productive retrieval. On the productive test, productive retrieval produced the highest scores. The sample was small and the items were invented words paired with illustrations, so a teacher should resist converting the finding into a universal ranking. It does show that practice and later use need to point in the same direction.
A Taipei office worker preparing for a presentation might study “postpone” in two passes. One card shows the word inside “We may postpone the launch until 17 November” and asks for the meaning. Another shows 「延後發表」 plus the meeting context and asks for the English verb. The second card is harder because it requires the form. It also resembles the moment when that speaker must find the word during a meeting.
The same distinction appears in Choc Education's article on building a way back to English. A delay adds one missing ingredient: enough time to travel that route before the card supplies the destination.
Keep the answer close enough
Waiting can become empty frustration. A learner who spends 23 seconds staring at a blank is rehearsing little beyond uncertainty. Set a short limit, reveal the answer, say it once, and schedule another attempt later. The aim is an honest search followed by accurate feedback.
If the first sound appears but the rest will not come, a graded hint can preserve the attempt: reveal the first syllable, play the opening phoneme, or show the number of letters, then allow another three seconds before displaying the complete form.
My preference is a six-second first attempt and a three-second second attempt later in the session. The numbers are a teaching choice, with the first one anchored in Barcroft's design. Learners who retrieve quickly can answer early. Learners who fail still receive the form before attention drifts.
Do not ask for spelling, pronunciation, meaning, and a complete sentence on every card. Those demands compete for limited attention. Choose one response for each pass:
- Hear the word and point to its picture.
- Read the Mandarin meaning and say the English word.
- Read a sentence with a blank and supply the missing form.
- Hear the word, then spell it from memory.
Four formats are listed here because the choice genuinely depends on the target. A concrete noun works well with a picture. A reporting verb such as “acknowledge” needs a sentence that constrains its use. A troublesome sound may need an audio cue. The cue should make the intended answer recoverable without giving the answer away.
Difficulty needs a boundary
Retrieval practice works poorly when almost every attempt fails. The learner first needs a clear encounter with the word, including its sound and meaning. A new card can show the complete pair once or twice. The hidden-answer phase follows after that initial encoding.
Repeated success inside one session also has limits. Tatsuya Nakata studied 98 Japanese learners and 16 English-Japanese word pairs, comparing one, three, five, and seven retrievals. Five and seven produced higher delayed scores than one and three. When the analysis controlled for time on task, one retrieval yielded the largest gain per unit of study time. More attempts can raise total learning while becoming less efficient minute by minute.
That boundary matters for a 30-minute commute. Seven immediate repetitions of eight words consume time that could place a second encounter tomorrow. The study compared retrieval frequency within a session. It does not settle the best schedule across a week.
Use the gap where software usually removes it. Turn off automatic answer reveal. Put the cue on one slide and the answer on the next. In pair work, have the partner silently count to six before helping. For tomorrow morning's review, choose 11 known-but-fragile words, hide each answer, and let the blank screen stay blank long enough for memory to do some work.