What Is Vocabulary | The Cohort Model — How Listeners Recognise Spoken Words as Speech Unfolds

COHORT MODEL · SPOKEN WORD RECOGNITION · LEXICAL COHORT · UNIQUENESS POINT · AUDITORY VOCABULARY · LEXICAL COMPETITION · SPEECH PERCEPTION

The Cohort Model explains spoken-word recognition as a process of activating and narrowing lexical candidates while speech unfolds. Hearing the beginning of a word activates a cohort of plausible words; later sounds remove competitors until the listener can identify the intended lexical item.

This model is especially useful for vocabulary because spoken recognition is temporal. A reader can inspect a whole printed word, but a listener must make decisions from partial acoustic evidence. Word frequency, phonological similarity, context and the point at which candidates diverge all influence recognition.

This guide explains lexical cohorts, candidate activation, uniqueness points, spoken-word competition, acoustic mismatch, frequency, context, segmentation and second-language listening. Existing eduKateSG Lexical Uniqueness Point and Word Recognition owners remain untouched.

Spoken words unfold over time

Listeners do not receive a complete spoken word at once. Each phoneme changes the set of lexical candidates compatible with the signal.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Initial cohort

In the classic Cohort Model, word onset activates a cohort of candidates sharing the beginning of the acoustic form.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Candidate elimination

As additional speech arrives, candidates inconsistent with the signal are eliminated or lose activation.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Uniqueness point

Recognition can become possible before word offset when only one lexical candidate remains compatible with the input.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Frequency

More frequent candidates can be stronger competitors, especially before the signal distinguishes them.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Context

Later versions of cohort-style accounts allow sentence and discourse information to influence selection among activated candidates.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Mispronunciation

A mismatch can disrupt activation, but real listeners tolerate variation, accent and noise more flexibly than a rigid onset filter would predict.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Continuous speech

Word boundaries are not reliably marked by silence, so lexical candidates help listeners segment the speech stream.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Neighbour competition

Words sharing long initial sequences compete longer than words diverging early.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Vocabulary size

A larger lexicon creates more possible competitors, but richer lexical knowledge also improves discrimination and prediction.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Second-language listening

L2 listeners may activate candidates according to both target-language and first-language phonological expectations.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Top-down knowledge

Meaning, syntax and topic can bias which cohort member remains plausible.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Bottom-up evidence

Acoustic detail constrains the cohort continuously; strong context cannot license a word that is incompatible with the signal indefinitely.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Recognition versus comprehension

Identifying the lexical form is only one stage. The listener must still select the intended sense and integrate it into the sentence.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Model limitations

The original Cohort Model is historically important but later spoken-word-recognition theories refine assumptions about activation, mismatch and feedback.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

Teaching

Listening instruction can train learners to notice early discriminating sounds, lexical competitors and uniqueness points rather than waiting for the entire sentence.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

AI-era speech

Automatic speech recognition also manages competing hypotheses over time, but computational architectures should not be equated directly with human cohort processing.

For learners, the practical question is which competitors remain alive at each moment and what acoustic or contextual evidence finally separates the target from them.

18. Cohort casebook A

1. Cohort case — captain

Onset cohort: begin with the first sound of captain and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put captain inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

2. Cohort case — captive

Onset cohort: begin with the first sound of captive and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put captive inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

3. Cohort case — candle

Onset cohort: begin with the first sound of candle and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put candle inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

4. Cohort case — candy

Onset cohort: begin with the first sound of candy and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put candy inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

5. Cohort case — garden

Onset cohort: begin with the first sound of garden and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put garden inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

6. Cohort case — garlic

Onset cohort: begin with the first sound of garlic and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put garlic inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

7. Cohort case — garment

Onset cohort: begin with the first sound of garment and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put garment inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

8. Cohort case — market

Onset cohort: begin with the first sound of market and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put market inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

9. Cohort case — marble

Onset cohort: begin with the first sound of marble and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put marble inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

10. Cohort case — margin

Onset cohort: begin with the first sound of margin and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put margin inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

11. Cohort case — teacher

Onset cohort: begin with the first sound of teacher and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put teacher inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

12. Cohort case — teacup

Onset cohort: begin with the first sound of teacup and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put teacup inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

13. Cohort case — teammate

Onset cohort: begin with the first sound of teammate and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put teammate inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

14. Cohort case — student

Onset cohort: begin with the first sound of student and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put student inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

15. Cohort case — studio

Onset cohort: begin with the first sound of studio and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put studio inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

16. Cohort case — stupid

Onset cohort: begin with the first sound of stupid and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put stupid inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

17. Cohort case — doctor

Onset cohort: begin with the first sound of doctor and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put doctor inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

18. Cohort case — document

Onset cohort: begin with the first sound of document and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put document inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

19. Cohort case — dolphin

Onset cohort: begin with the first sound of dolphin and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put dolphin inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

20. Cohort case — river

Onset cohort: begin with the first sound of river and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put river inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

19. Cohort casebook B

21. Cohort case — ribbon

Onset cohort: begin with the first sound of ribbon and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put ribbon inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

22. Cohort case — riddle

Onset cohort: begin with the first sound of riddle and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put riddle inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

23. Cohort case — mountain

Onset cohort: begin with the first sound of mountain and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put mountain inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

24. Cohort case — mouse

Onset cohort: begin with the first sound of mouse and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put mouse inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

25. Cohort case — mouth

Onset cohort: begin with the first sound of mouth and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put mouth inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

26. Cohort case — evidence

Onset cohort: begin with the first sound of evidence and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put evidence inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

27. Cohort case — event

Onset cohort: begin with the first sound of event and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put event inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

28. Cohort case — every

Onset cohort: begin with the first sound of every and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put every inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

29. Cohort case — system

Onset cohort: begin with the first sound of system and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put system inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

30. Cohort case — sister

Onset cohort: begin with the first sound of sister and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put sister inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

31. Cohort case — signal

Onset cohort: begin with the first sound of signal and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put signal inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

32. Cohort case — significant

Onset cohort: begin with the first sound of significant and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put significant inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

33. Cohort case — climate

Onset cohort: begin with the first sound of climate and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put climate inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

34. Cohort case — client

Onset cohort: begin with the first sound of client and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put client inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

35. Cohort case — climb

Onset cohort: begin with the first sound of climb and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put climb inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

36. Cohort case — policy

Onset cohort: begin with the first sound of policy and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put policy inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

37. Cohort case — police

Onset cohort: begin with the first sound of police and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put police inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

38. Cohort case — polish

Onset cohort: begin with the first sound of polish and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put polish inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

39. Cohort case — contract

Onset cohort: begin with the first sound of contract and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put contract inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

40. Cohort case — contrast

Onset cohort: begin with the first sound of contrast and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put contrast inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

20. Cohort casebook C

41. Cohort case — control

Onset cohort: begin with the first sound of control and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put control inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

42. Cohort case — music

Onset cohort: begin with the first sound of music and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put music inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

43. Cohort case — museum

Onset cohort: begin with the first sound of museum and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put museum inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

44. Cohort case — mutual

Onset cohort: begin with the first sound of mutual and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put mutual inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

45. Cohort case — analysis

Onset cohort: begin with the first sound of analysis and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put analysis inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

46. Cohort case — analyst

Onset cohort: begin with the first sound of analyst and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put analyst inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

47. Cohort case — analogy

Onset cohort: begin with the first sound of analogy and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put analogy inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

48. Cohort case — predict

Onset cohort: begin with the first sound of predict and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put predict inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

49. Cohort case — prepare

Onset cohort: begin with the first sound of prepare and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put prepare inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

50. Cohort case — prevent

Onset cohort: begin with the first sound of prevent and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put prevent inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

51. Cohort case — reluctant

Onset cohort: begin with the first sound of reluctant and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put reluctant inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

52. Cohort case — relevant

Onset cohort: begin with the first sound of relevant and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put relevant inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

53. Cohort case — reliable

Onset cohort: begin with the first sound of reliable and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put reliable inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

54. Cohort case — meticulous

Onset cohort: begin with the first sound of meticulous and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put meticulous inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

55. Cohort case — metal

Onset cohort: begin with the first sound of metal and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put metal inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

56. Cohort case — method

Onset cohort: begin with the first sound of method and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put method inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

57. Cohort case — resilient

Onset cohort: begin with the first sound of resilient and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put resilient inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

58. Cohort case — resident

Onset cohort: begin with the first sound of resident and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put resident inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

59. Cohort case — result

Onset cohort: begin with the first sound of result and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put result inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

60. Cohort case — ubiquitous

Onset cohort: begin with the first sound of ubiquitous and list several real words that could still match. This demonstrates why recognition cannot be complete at word onset.

Divergence: reveal the next sound one unit at a time and remove candidates that no longer match. Mark the earliest point where the target becomes uniquely identifiable within the learner’s known vocabulary.

Frequency pressure: place a more frequent competitor in the cohort. Before acoustic divergence, the frequent candidate may attract stronger expectation even when it is not the eventual target.

Context pressure: put ubiquitous inside a sentence that strongly favours one cohort member. Compare contextual expectation with bottom-up acoustic evidence and explain how both sources constrain recognition.

Noise test: mask one early phoneme or alter pronunciation slightly. Decide whether later evidence can recover the target and what this reveals about the limits of a strict onset-only account.

Teaching transfer: practise the target beside its closest onset competitors. Learners should identify the discriminating sound quickly enough to avoid carrying the wrong candidate too far into the sentence.

21. Final model

The Cohort Model made a foundational claim: spoken vocabulary recognition is incremental. The listener activates possibilities before the word finishes and narrows them as evidence accumulates.

Modern accounts modify details, but the educational insight remains powerful. Fluent listening depends on fast competition and elimination among lexical candidates, not on waiting passively for complete words.

22. eduKateSG routes

Lexical Uniqueness Point · Word Recognition · Phonological Neighbourhoods · Vocabulary Master Router

23. Extended cohort laboratory

1. captain

Construct a four-word onset cohort for captain. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

2. captive

Construct a four-word onset cohort for captive. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

3. candle

Construct a four-word onset cohort for candle. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

4. candy

Construct a four-word onset cohort for candy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

5. garden

Construct a four-word onset cohort for garden. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

6. garlic

Construct a four-word onset cohort for garlic. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

7. garment

Construct a four-word onset cohort for garment. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

8. market

Construct a four-word onset cohort for market. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

9. marble

Construct a four-word onset cohort for marble. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

10. margin

Construct a four-word onset cohort for margin. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

11. teacher

Construct a four-word onset cohort for teacher. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

12. teacup

Construct a four-word onset cohort for teacup. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

13. teammate

Construct a four-word onset cohort for teammate. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

14. student

Construct a four-word onset cohort for student. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

15. studio

Construct a four-word onset cohort for studio. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

16. stupid

Construct a four-word onset cohort for stupid. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

17. doctor

Construct a four-word onset cohort for doctor. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

18. document

Construct a four-word onset cohort for document. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

19. dolphin

Construct a four-word onset cohort for dolphin. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

20. river

Construct a four-word onset cohort for river. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

21. ribbon

Construct a four-word onset cohort for ribbon. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

22. riddle

Construct a four-word onset cohort for riddle. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

23. mountain

Construct a four-word onset cohort for mountain. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

24. mouse

Construct a four-word onset cohort for mouse. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

25. mouth

Construct a four-word onset cohort for mouth. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

26. evidence

Construct a four-word onset cohort for evidence. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

27. event

Construct a four-word onset cohort for event. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

28. every

Construct a four-word onset cohort for every. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

29. system

Construct a four-word onset cohort for system. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

30. sister

Construct a four-word onset cohort for sister. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

31. signal

Construct a four-word onset cohort for signal. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

32. significant

Construct a four-word onset cohort for significant. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

33. climate

Construct a four-word onset cohort for climate. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

34. client

Construct a four-word onset cohort for client. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

35. climb

Construct a four-word onset cohort for climb. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

36. policy

Construct a four-word onset cohort for policy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

37. police

Construct a four-word onset cohort for police. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

38. polish

Construct a four-word onset cohort for polish. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

39. contract

Construct a four-word onset cohort for contract. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

40. contrast

Construct a four-word onset cohort for contrast. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

41. control

Construct a four-word onset cohort for control. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

42. music

Construct a four-word onset cohort for music. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

43. museum

Construct a four-word onset cohort for museum. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

44. mutual

Construct a four-word onset cohort for mutual. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

45. analysis

Construct a four-word onset cohort for analysis. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

46. analyst

Construct a four-word onset cohort for analyst. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

47. analogy

Construct a four-word onset cohort for analogy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

48. predict

Construct a four-word onset cohort for predict. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

49. prepare

Construct a four-word onset cohort for prepare. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

50. prevent

Construct a four-word onset cohort for prevent. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

51. reluctant

Construct a four-word onset cohort for reluctant. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

52. relevant

Construct a four-word onset cohort for relevant. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

53. reliable

Construct a four-word onset cohort for reliable. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

54. meticulous

Construct a four-word onset cohort for meticulous. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

55. metal

Construct a four-word onset cohort for metal. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

56. method

Construct a four-word onset cohort for method. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

57. resilient

Construct a four-word onset cohort for resilient. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

58. resident

Construct a four-word onset cohort for resident. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

59. result

Construct a four-word onset cohort for result. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

60. ubiquitous

Construct a four-word onset cohort for ubiquitous. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

61. captain

Construct a four-word onset cohort for captain. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

62. captive

Construct a four-word onset cohort for captive. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

63. candle

Construct a four-word onset cohort for candle. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

64. candy

Construct a four-word onset cohort for candy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

65. garden

Construct a four-word onset cohort for garden. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

66. garlic

Construct a four-word onset cohort for garlic. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

67. garment

Construct a four-word onset cohort for garment. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

68. market

Construct a four-word onset cohort for market. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

69. marble

Construct a four-word onset cohort for marble. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

70. margin

Construct a four-word onset cohort for margin. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

71. teacher

Construct a four-word onset cohort for teacher. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

72. teacup

Construct a four-word onset cohort for teacup. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

73. teammate

Construct a four-word onset cohort for teammate. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

74. student

Construct a four-word onset cohort for student. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

75. studio

Construct a four-word onset cohort for studio. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

76. stupid

Construct a four-word onset cohort for stupid. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

77. doctor

Construct a four-word onset cohort for doctor. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

78. document

Construct a four-word onset cohort for document. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

79. dolphin

Construct a four-word onset cohort for dolphin. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

80. river

Construct a four-word onset cohort for river. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

81. ribbon

Construct a four-word onset cohort for ribbon. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

82. riddle

Construct a four-word onset cohort for riddle. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

83. mountain

Construct a four-word onset cohort for mountain. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

84. mouse

Construct a four-word onset cohort for mouse. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

85. mouth

Construct a four-word onset cohort for mouth. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

86. evidence

Construct a four-word onset cohort for evidence. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

87. event

Construct a four-word onset cohort for event. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

88. every

Construct a four-word onset cohort for every. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

89. system

Construct a four-word onset cohort for system. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

90. sister

Construct a four-word onset cohort for sister. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

91. signal

Construct a four-word onset cohort for signal. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

92. significant

Construct a four-word onset cohort for significant. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

93. climate

Construct a four-word onset cohort for climate. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

94. client

Construct a four-word onset cohort for client. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

95. climb

Construct a four-word onset cohort for climb. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

96. policy

Construct a four-word onset cohort for policy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

97. police

Construct a four-word onset cohort for police. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

98. polish

Construct a four-word onset cohort for polish. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

99. contract

Construct a four-word onset cohort for contract. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

100. contrast

Construct a four-word onset cohort for contrast. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

101. control

Construct a four-word onset cohort for control. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

102. music

Construct a four-word onset cohort for music. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

103. museum

Construct a four-word onset cohort for museum. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

104. mutual

Construct a four-word onset cohort for mutual. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

105. analysis

Construct a four-word onset cohort for analysis. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

106. analyst

Construct a four-word onset cohort for analyst. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

107. analogy

Construct a four-word onset cohort for analogy. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

108. predict

Construct a four-word onset cohort for predict. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

109. prepare

Construct a four-word onset cohort for prepare. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.

110. prevent

Construct a four-word onset cohort for prevent. Order the candidates by corpus frequency, then reveal the signal phoneme by phoneme. At each step, record which candidates remain compatible and which should lose activation.

Repeat inside a strongly constraining sentence and inside a neutral sentence. Compare the recognition point and explain whether context changes candidate weighting, acoustic commitment or both. Finish with a delayed listening test using a different speaker so learning is not tied to one voice.