Three Sound Features, One Timetable: Prioritising in a Crowded Oral English Slot
At a junior high school in Cape Coast, a Form Two class of 48 learners has three 35-minute English periods a week, no language laboratory and one chalkboard. A diagnostic read-aloud of 24 sentences showed that 31 learners said 'three' as /triː/ and 'think' as /tɪŋk/; 27 merged /l/ and /r/, so 'play' became 'pray' and 'collect' became 'correct'; and 19 dropped the final consonants in 'asked' and 'walked'. In a dictation the same learners wrote 'tink' for 'think' and 'pray' for 'play'. Their first language is Fante, and their teacher, Mr Adjei, speaks Twi and English. A colleague proposes fifteen minutes of daily minimal-pair drilling on all three features for the whole term, arguing that practice makes perfect. Mr Adjei has thirteen weeks before the mock examination and is uneasy about the time cost and about embarrassing learners who are otherwise fluent and willing speakers.
1. Mr Adjei cannot teach all three features with equal weight. Using the diagnostic evidence, decide a priority order and say what the evidence does and does not establish.
Read the worked answer
What the evidence establishes. Three different processes are bundled together here, and they do not deserve equal treatment.
- /θ/ becoming [t] is a segmental substitution. If Fante has no /θ/, learners substitute the nearest available stop; the substitution is systematic, not careless.
- The /l/ – /r/ merger is also segmental, but it is lexically unpredictable: the learner must learn which word takes which sound, so it needs word-level work, not only sound work.
- Loss of final consonants in 'asked' and 'walked' is a syllable-structure process (cluster simplification). It is the most damaging of the three, because the lost segment carries the past-tense morpheme.
The dictation adds a fourth finding: the learners spell what they say. This is not purely phonology. Pronunciation and orthography are feeding each other, and spelling has to be taught alongside.
Priority order. First, final clusters carrying -ed, because the loss is grammatical rather than merely accentual and it surfaces in writing and in grammar questions. Second, /l/ – /r/, because functional load is high - play/pray, collect/correct, glass/grass, lice/rice - and real misunderstanding occurs. Third, /θ/, which rarely blocks intelligibility in context and functions more as a prestige marker than a comprehension barrier; it deserves short, low-stakes attention, not a term of drills.
What the evidence does not establish. A 24-sentence read-aloud is a small, formal sample. It may overstate the problem, because learners often produce the target more accurately in connected, meaningful speech. Mr Adjei should confirm the picture with five minutes of observed conversation before committing thirteen weeks.
2. Design the thirteen-week intervention within the existing timetable. What happens in weeks 1 to 4, and how would you know by week 8 whether it is working?
Read the worked answer
Budget. Three 35-minute periods give 105 minutes. Taking six minutes from each period yields about eighteen minutes a week for focused pronunciation without sacrificing the oral work it must live inside - roughly half the colleague's proposal, and about four hours of deliberate work across the term.
Weeks 1-4.
- Week 1: no drills yet. Establish that accent is not the same as error, model the three features yourself, and let learners notice them in your speech and in short recordings of educated Ghanaian speakers. Agree the classroom rule: we work on sounds, we do not laugh at people.
- Week 2: past-tense endings. Use fifteen verbs the class already needs for narrative writing. Learners clap the syllables in 'asked' and 'walked', then play a paired 'yesterday' game in which one says a verb and the partner uses it with 'yesterday', so the cluster appears inside meaning rather than in a list.
- Week 3: /l/ and /r/. Teach eight high-value pairs (play/pray, collect/correct, glass/grass, lice/rice). Use a 'find your pair' game and a sentence dictation, since the dictation evidence showed learners spell what they say.
- Week 4: /θ/ in ten frequent words (think, three, thank, both, month), modelled in short chunks and chants - five minutes only.
Weeks 5-13. Recycle the targets in the first two minutes of every lesson inside other activities: a sentence of the day, reading aloud, role play. New content stops; the aim is transfer, not novelty.
Checking at week 8. Re-test the same ten items in a fresh read-aloud, but weight the judgement towards a two-minute paired conversation, because that is where the skill has to appear.
3. A colleague says the daily drilling plan is better because it is measurable. How should Mr Adjei assess progress without a laboratory, and how should he handle learners' embarrassment?
Read the worked answer
Assessment without a laboratory. The colleague's underlying point is fair: a plan you cannot check is a wish. Mr Adjei can check cheaply.
- Keep a tally sheet with the three features and each week observe five learners on a three-point scale (uses the target consistently / sometimes / not yet) during ordinary group talk, not during drills. Over thirteen weeks every learner is seen at least three times.
- Use a paired dictation: one learner says a sentence, the other writes it. This tests the sound and the spelling link at once, needs no equipment, and shows whether 'asked' is being heard as 'ask'.
- Judge intelligibility rather than accent. If a partner understands 'pray' as pray in context, the merger is not yet a communication problem and should not be marked as one.
- Record a two-minute conversation in week 1 and week 13 if any phone is available; otherwise run the same short role play and score it live against the tally sheet.
Handling embarrassment. Never correct an individual's sound in front of the class. Correct the class - 'we are working on this ending' - and let individuals self-correct with a hand signal. Use choral and pair work so nobody is exposed alone. Say plainly that many educated Ghanaian speakers use [t] for /θ/ and that this is a variety, not a defect, while explaining why formal speaking and the examination reward the other form. Keep these features out of oral marks for the first term. Praise fluent speakers publicly for the content of what they say, so that pronunciation work is not read as a verdict on their English.