A-Level Revision Aid: This resource is designed to support your revision. Content may contain errors. If you find a discrepancy with your class teaching, your teacher is correct — please let us know at alevelrevise@scott.scottrix.co.uk.
Child Language Acquisition
Year 1 / ASYear 2 / A-LevelAQA, OCR, WJEC/Eduqas
A* revision for AQA 7702/1 Section B and OCR H470/02 Section A: stages of spoken development, phonology, virtuous errors, theory evaluation, reading and writing development, and the data-analysis method that secures top-band marks.
🎯 Learning Objectives
Sequence the stages of spoken development - pre-verbal, holophrastic, two-word, telegraphic, post-telegraphic - and label the stage in unseen data from linguistic evidence, not from age alone (AO1).
Analyse phonological development (substitution, deletion, assimilation, cluster reduction) and link phonological processes to lexical and grammatical growth (AO1).
Categorise virtuous errors - overextension, underextension, overgeneralisation - and explain why they evidence rule formation rather than failed imitation (AO1/AO2).
Evaluate theories critically: Chomsky's LAD and its critiques, Skinner's behaviourism, Piaget's cognitive model, Vygotsky's ZPD, Bruner's LASS and Halliday's functions (AO2).
Analyse reading and writing development (emergent literacy, Kroll's phases, Britton's modes) and compare spoken and written trajectories (AO1/AO2).
Apply the data-analysis method for AQA 7702/1 Section B (30 marks, AO1 15 + AO2 15) and OCR H470/02 Section A (20 marks, AO1 10 + AO2 10) within exam timing.
📋 Key Definitions and Core Concepts
Holophrastic stage: The single-word stage (roughly 12-18 months): one word may express a whole proposition (a holophrase); Nelson (1973) found over 60% of first words are nominals.
Two-word stage: From roughly 18-24 months: subject-verb and verb-object pairs ('daddy gone') with emerging word order; Braine analysed it as pivot + open-class grammar, with function words still absent.
Telegraphic stage: From roughly 24-36 months: content words only, with articles, auxiliaries, prepositions and inflections missing; measured by rising mean length of utterance (MLU - Brown, 1973).
Post-telegraphic stage: From roughly 36 months: grammatical function words acquired; remaining errors concentrate on inflection and agreement ('goed', 'sheeps'); subordinate clauses and auxiliary questions consolidate.
Virtuous error: A systematic, logical error produced by applying a rule too generally ('goed', 'foots'); crucially absent from adult input, so it evidences grammar construction rather than imitation (links to Chomsky).
Overextension: Applying a word beyond its adult meaning ('dog' for all four-legged animals); Rescorla (1980) distinguishes categorical, analogical and statement overextensions.
Underextension: Restricting a word to a subset of its adult meaning ('dog' used only for the family pet); shows meaning-building in progress in the opposite direction to overextension.
LAD: Chomsky's Language Acquisition Device: an innate faculty containing universal grammar; children exposed to imperfect, 'degenerate' input still acquire grammar rapidly - the poverty-of-stimulus argument.
ZPD: Vygotsky's zone of proximal development: the gap between what a child can do alone and what they can do with a more knowledgeable other; scaffolding (Wood, Bruner & Ross, 1976) bridges it.
Child-directed speech (CDS): Caregiver speech adapted for the child - higher pitch, exaggerated intonation, simplified lexis, recasts, expansions, question loops; the input side of Bruner's Language Acquisition Support System (LASS).
📌 Key Points
AQA 7702/1 Section B = 30 marks: a discursive essay, choice of two questions, data may be spoken, written or multimodal (June 2024: a child transcript and children's writing).
The June 2024 mark scheme splits Section B as AO1 15 + AO2 15: half data analysis, half critical evaluation of theories and the given statement - theory description alone cannot pass.
OCR H470/02 Section A = 20 marks (AO1 10 + AO2 10): the language development stage of a child aged 0-7, analysing phonology, grammar and meaning, with the IPA provided in the paper.
Label stages from evidence - utterance length, grammatical completeness, phonological processes - never from the child's age; mixed-stage data is normal and expected.
Theories work best in evaluation pairs: nature (Chomsky) vs nurture (Skinner); cognition-first (Piaget) vs interaction-first (Vygotsky/Bruner); functions (Halliday) cut across all four.
Virtuous errors are the single strongest classroom argument: systematic, rule-governed errors absent from adult input cannot be products of imitation alone.
Stages of spoken development
Both boards anchor this topic in data. AQA's Section B discursive essay (7702/1, 30 marks) supplies a Data Set and asks you to evaluate a given statement against it; OCR's H470/02 Section A (20 marks) asks you to examine the language development stage of the child-participant in a transcript. Either way, the stages are the spine - but treat them as descriptive guides, not universal checkpoints, because a top answer labels the stage from linguistic evidence and acknowledges overlap.
Stage
Rough age
Features
Exam anchor
Pre-verbal
0-12 months
Vegetative sounds, then cooing (4-7 months), then babbling - reduplicated strings ('mamama'); Jakobson argued babbling is universal across languages.
Proto-words with consistent meaning ('baba' for bottle) mark the transition into lexical development.
Holophrastic
12-18 months
Single-word utterances where one word may express a whole proposition; Nelson (1973): 60%+ nominals, then action words, modifiers, personal-social words.
Overextension and underextension typically begin here.
Two-word
18-24 months
Subject-verb and verb-object pairs ('daddy gone', 'want juice'); emerging word order; Braine's pivot + open-class grammar; function words absent.
Word-order consistency is evidence of grammar construction - useful against pure imitation claims.
Telegraphic
24-36 months
Content words only: articles, auxiliaries, prepositions and inflections missing, like a telegram; MLU rises (Brown, 1973).
Name which grammatical items are missing - precision beats the label.
Post-telegraphic
36 months+
Function words acquired; errors concentrate on inflection and agreement; subordinate clauses, auxiliary questions and consolidated negatives appear.
Link to Bellugi's negative and question stages for the grammar evidence.
Deploy stages as evidence, not assertion: 'the utterances show telegraphic features - auxiliaries and articles are systematically absent - so the child sits at the telegraphic/post-telegraphic boundary' beats 'the child is 2, so telegraphic'.
Mixed-stage data is normal: performance varies by context, task and fatigue; the transcript may show post-telegraphic syntax in one turn and holophrastic nouns in another - say so, and use it (the variation itself is analysable).
Metrics: MLU (mean length of utterance) and the percentage of grammatically complete utterances are quick quantitative hooks that earn AO1 credit in a discursive essay.
Where the marks are: in AQA's discursive essay the stage labels are scaffolding for the evaluation - the examiner rewards what you do with them: evidence of rule construction, of caregiver scaffolding, of the gap between perception and production. In OCR's Section A the stage is the question itself, so the opening sentence should commit to a stage and the closing paragraph should return to it.
Phonological, lexical and grammatical development
OCR's Section A question explicitly requires phonology, grammar and meaning - three levels, one answer - so have the detail ready at each.
Phonological development: the phonemic inventory grows from vowels and plosives to fricatives and affricates; late-acquired sounds (/θ/ in thumb, /ʃ/ in shoe) drive the classic processes: substitution ('fumb' for thumb), deletion ('ca' for cat), assimilation ('gog' for dog), consonant cluster reduction ('tee' for tree) and addition. The processes are systematic, which is the point: the child is operating a rule-governed system, not producing random error. Berko & Brown (1960)'s fis phenomenon - a child says 'fis' for fish but rejects an adult's 'fis' - shows perception running ahead of production.
Lexical development: Nelson's category data (nominals dominant) at first-word stage; then semantic organisation - overextension (categorical: 'apple' for all round fruit; analogical: 'ball' for the moon; statement-type tied to context - Rescorla, 1980) and underextension ('dog' only for the family pet). Clark's (1973) semantic feature hypothesis explains the direction of error: meanings are built feature by feature. Note the comprehension-production gap: children understand more than they can say.
Grammatical development: Brown (1973) tracked 14 grammatical morphemes in acquisition order - progressive -ing first, then articles, then regular past -ed (with overgeneralisation 'goed'), then third-person -s. Bellugi (with Brown) staged negatives (external 'no' + X, then external 'don't'/'can't', then internal negation 'I don't want') and questions (intonation-only, then wh-words without inversion, then auxiliary inversion). These sequences are ready-made analytic frames for any transcript.
Pragmatics: turn-taking learned through CDS games and routines; politeness formulas ('please', 'sorry') early; Gricean implicature, irony and audience design much later, through the school years - a useful contrast when a question targets what children cannot yet do.
Child-directed speech (CDS) threads through all three levels: higher pitch, exaggerated intonation, simplified lexis, slower tempo, recasts and expansions, question loops. In AQA's discursive essay CDS is usually the fastest route to the interaction side of the evaluation - but always tie the feature to a pattern in the data.
Theories and functions: evaluate, don't describe
Theory earns its AO2 marks only when evaluated against the data: give the claim, the support, the challenge and a judgement. Memorise each theory as a pair of arguments, not a paragraph of description.
Chomsky - LAD (1965 onwards): an innate language faculty with universal grammar; the poverty-of-stimulus argument holds that adult input is imperfect and incomplete, yet children acquire grammar rapidly and uniformly. Evidence for: virtuous errors ('goed', 'foots') are never heard in adult speech, so they cannot be imitated; milestones are strikingly uniform across languages. Critiques: the LAD is a black box - no mechanism is specified; it underestimates input quality (CDS is more structured than Chomsky assumed); it has little to say about reading and writing development; and it treats grammar as the whole of acquisition, sidelining pragmatics and functions.
Skinner - behaviourism (1957): language learned through imitation, reinforcement and habit formation. Evidence for: children demonstrably copy intonation contours, lexical items and CDS recasts. Critiques: the decisive one is Brown & Hanlon (1970) - caregivers respond to the truth value of utterances, not their grammatical form, so 'doggy bringed it' passes uncorrected while a true but ungrammatical utterance is approved; reinforcement cannot explain rule formation.
Piaget - cognitive first: language maps onto cognitive development - object permanence (around 8 months) precedes naming, so thought precedes speech. Critiques: some lexical and pragmatic competence appears to precede the relevant cognitive stage, and the model underplays social interaction.
Vygotsky - ZPD and the more knowledgeable other: development happens in the gap between independent and supported performance; scaffolding (Wood, Bruner & Ross, 1976) is the mechanism, and Bruner's LASS supplies the formats - routines, games, picture-book reading - in which CDS does its work. Evidence for: recasts and expansions in caregiver data; turn-taking skills built through CDS games. Critique: less explanatory for phonological detail, which develops with striking independence from input style.
Halliday - functions (1975): seven functions of early language - instrumental (I want), regulatory (do as I say), interactional (me and you), personal (here I come), heuristic (tell me why), imaginative (let's pretend), representational (informative) - later regrouped into the mathetic (learning) and pragmatic (doing) phases. In a discursive essay, categorise data utterances by function and evaluate which dominate: infant speech is often more interactional and heuristic than the 'get things' stereotype suggests.
The virtuous-error argument - why it kills pure imitation
If language were learned by imitation alone, every child error would be a copy of adult speech. But 'goed', 'foots' and 'fis' never appear in adult input - the child has constructed a rule (add -ed for past tense; final devoicing) and applied it generally. That construction is exactly what Chomsky's LAD predicts and Skinner's reinforcement loop cannot explain. It is the highest-value three-line argument in this topic: name the error, state the rule, name the theory it supports.
Reading and writing development, and the data-analysis method
AQA's Section B can set written or multimodal data (June 2024's Question 5 asked about writing development), so the written strand is examinable, not optional.
Reading: emergent literacy begins with environmental print and grapheme-phoneme awareness; the phonics (decoding) vs look-and-say (whole-word) debate frames instruction, and decoding typically precedes comprehension - Chall's stages run from prereading through initial decoding and fluency to reading for learning and multiplicity. For data analysis, note which words a child reads accurately and which they guess from shape or context.
Writing: Kroll's (1981) phases - preparation (motor control, letter formation), consolidation (writing as they speak, strings of clauses), differentiation (genre awareness, punctuation), integration (audience and purpose control) - plus Britton's expressive → poetic → transactional progression. Invented spellings ('becos', 'wos') are orthographic virtuous errors: phonetic logic applied to spelling, evidence of systematic hypothesis-testing rather than carelessness.
Multimodal development: children compose with emoji, capitals and formatting in messages - graphology as meaning-making, and a modern contrast to Kroll's print-based phases.
The data-analysis method (AQA 7702/1 Section B, 30 marks = AO1 15 + AO2 15):
Read the data twice. Note mode, context and the child's age - then set the age aside and identify the stage(s) from evidence: utterance length, grammatical completeness, phonological processes.
Collect AO1 evidence level by level: phonology (use the transcription key; OCR supplies the IPA), lexis/semantics (Nelson's categories, over/underextension), grammar (Bellugi's negatives and questions, inflections, MLU), pragmatics (turn-taking, Halliday's functions, caregiver scaffolding visible in the data).
Evaluate (AO2): match theories to patterns - CDS features → Vygotsky/Bruner; virtuous errors → Chomsky vs Skinner; functions in utterances → Halliday - and weigh both sides of the given statement with counter-examples.
Write discursively: balanced paragraphs, each theory point anchored in quoted data ('Referring to Data Set 1 in detail' is a literal instruction), and an explicit judgement in the conclusion.
OCR H470/02 Section A (20 marks = AO1 10 + AO2 10): the question names what it wants - the stage, then phonology, grammar and meaning, with theories supporting interpretation. Roughly 40 minutes: identify the stage with evidence in your opening, then one paragraph per level, theory woven into each, terminology throughout. OCR also expects you to read the phonemic transcription actively: pauses, stressed syllables and simultaneous speech in the key are analysis opportunities, not decoration.
💡 Worked Example
Exam-style: 'Adults play the most important role in children's language development.' Referring to Data Set 1 in detail, and to relevant ideas from language study, evaluate this view of children's language development. [30 marks] (AQA 7702/1, June 2024, Question 4 - Section B)
Mark-scheme logic: Mark-scheme logic (AQA 7702/1 Section B: 30 marks = AO1 15 + AO2 15, per the June 2024 mark scheme):<br><br><strong>AO1 - analysis of the data (15):</strong> open by identifying the stage(s) evidenced in Data Set 1 from linguistic evidence - utterance length, grammatical completeness, phonological processes - not from age alone. Then analyse systematically: phonology (substitutions, deletions, cluster reduction read against the transcription key - what the child can and cannot produce), lexis/semantics (Nelson's categories, overextension/underextension, concrete nominals dominant), grammar (negatives and questions against Bellugi's stages, missing auxiliaries and articles, overgeneralisation such as 'goed'), pragmatics (turn-taking, Halliday's functions, caregiver recasts and expansions visible in the data). Every point anchored in a precise quotation from the data set.<br><br><strong>AO2 - critical evaluation of the view (15):</strong> evaluate the statement, don't describe it. <em>Support:</em> child-directed speech and scaffolding - Vygotsky's ZPD and the more knowledgeable other, Wood, Bruner & Ross (1976), Bruner's LASS - with recasts and expansions shaping development in the data. <em>Challenge:</em> Chomsky's LAD and the poverty of stimulus (grammar acquired from imperfect input), virtuous errors absent from adult speech ('goed', Berko & Brown's 1960 'fis'), Brown & Hanlon's (1970) finding that caregivers correct truth value rather than grammatical form, and cross-cultural evidence that acquisition proceeds on schedule where CDS is minimal. <em>Judgement:</em> interaction clearly matters, but it cannot explain rule formation alone - a balanced conclusion weighing both sides.<br><br><strong>Discursive craft:</strong> both sides engaged, theories evaluated against data patterns rather than bolted on, explicit conclusion. Question 5 (the writing-development option) works identically: analyse the children's writing samples against Kroll's phases and emergent-literacy features (AO1), then evaluate the purpose-driven-writing claim against genre, audience and instruction research (AO2).
❓ Practice Questions & Mark Schemes
Using the appropriate terminology to explain your findings, examine the language development stage of the child-participant as evidenced in the transcript. You should identify and analyse the phonology, grammar and meaning of his utterances. Use your knowledge of theories and concepts of child language acquisition to support your answer. [20 marks] (OCR H470/02, June 2024, Question 1)
Answer: Method: 40 minutes; the question names three levels plus stage identification, so plan one paragraph per level with theory woven through (AO1 10 + AO2 10).<br><br>Model hits: (1) <strong>Stage identification with evidence:</strong> utterance length, grammatical completeness and phonology locate the child at a stage boundary - state it in your opening sentence. (2) <strong>Phonology:</strong> substitution, deletion, assimilation and cluster reduction with transcribed evidence (the IPA is provided - use it); note late-acquired sounds driving the processes; systematicity argument - rule-governed, not random. (3) <strong>Grammar:</strong> missing function words (telegraphic features), morpheme development (Brown's order), negatives and questions against Bellugi's stages. (4) <strong>Meaning:</strong> lexical semantics - overextension/underextension, Nelson's categories; Halliday's functions for what the utterances <em>do</em>. (5) <strong>Theory support (AO2):</strong> virtuous errors point to Chomsky's LAD vs Skinner's imitation; CDS features in the transcript point to Vygotsky/Bruner. Terminology throughout - the question explicitly rewards it.
'Writing with a sense of purpose is the most important skill to develop when learning to write.' Referring to Data Set 2 in detail, and to relevant ideas from language study, evaluate this view of children's language development. [30 marks] (AQA 7702/1, June 2024, Question 5)
Answer: Mark-scheme logic (AO1 15 + AO2 15):<br><br><strong>AO1 - the data:</strong> analyse the children's writing samples: emergent features (letter formation, letter-like forms, invented spellings such as 'becos' - orthographic virtuous errors), structural features (strings of clauses suggesting Kroll's consolidation phase, or genre-appropriate organisation suggesting differentiation), punctuation, paragraphing and graphological choices - each anchored in quoted extracts.<br><br><strong>AO2 - evaluation:</strong> <em>Support the view:</em> writing for real audiences and purposes drives genre awareness - Kroll's differentiation and integration phases; Britton's expressive to transactional progression is purpose-driven; purpose explains vocabulary and syntax choices in the samples. <em>Challenge it:</em> purpose cannot operate before the basics - motor skills, grapheme-phoneme correspondence (Chall's early reading stages) and syntax must precede it; phonics and decoding underpin everything a child writes; some children write accurately for no evident purpose. <em>Judgement:</em> purpose accelerates the later phases but sits on top of foundational skills - weigh both.<br><br><strong>Discursive craft:</strong> balanced paragraphs, every theory point tied to Data Set 2, explicit conclusion - the AO1/AO2 balance is what separates top-band essays from theory lists.
'Children learn to talk through imitation.' Referring to data and relevant ideas from language study, evaluate this view of children's language development. [30 marks] (AQA 7702/1 Section B style)
Answer: Mark-scheme logic (AO1 15 + AO2 15):<br><br><strong>AO2 frame - the view is Skinner's behaviourism</strong> (imitation + reinforcement + habit formation).<br><br><strong>AO1 - data evidence:</strong> quote utterances that match caregiver recasts (imitation demonstrably occurs - intonation contours, lexical items) and utterances the child could never have heard (virtuous errors: 'goed', 'foots', 'sheeps') - the contrast is your evidence engine.<br><br><strong>AO2 - evaluation:</strong> <em>Challenge:</em> virtuous errors are systematic and rule-governed, so they evidence grammar construction (Chomsky's LAD, poverty of stimulus); Brown & Hanlon (1970) - caregivers correct truth value, not grammatical form, so reinforcement cannot teach syntax; milestones are rapid and uniform across languages, faster than imitation predicts. <em>Support:</em> CDS recasts and expansions shape development (Vygotsky/Bruner - interaction matters); imitation of prosody and vocabulary is real. <em>Judgement:</em> imitation is a partial mechanism for lexis and prosody, but rule formation requires an innate or constructive capacity - conclude with the virtuous-error argument.<br><br><strong>Craft:</strong> both sides in the same paragraphs where possible; theory evaluated against quoted data throughout.
Examine the phonological development evidenced in the transcript, using appropriate terminology, and support your answer with theories and concepts of child language acquisition. [20 marks] (OCR H470/02 Section A style)
Answer: Method: one level done in depth beats breadth; 20 marks = AO1 10 + AO2 10.<br><br>Model hits: (1) <strong>Processes with evidence:</strong> substitution ('fumb' for thumb), deletion ('ca' for cat), assimilation ('gog' for dog), cluster reduction ('tee' for tree), addition - each quoted from the transcript with accurate terminology. (2) <strong>Inventory gaps:</strong> late-acquired phonemes (the <em>th</em> and <em>sh</em> sounds) explain which substitutions appear - link inventory to process. (3) <strong>Systematicity:</strong> the processes are consistent across the transcript - a rule-governed system under construction, not random error. (4) <strong>Perception vs production:</strong> the fis phenomenon (Berko & Brown, 1960) - children perceive contrasts they cannot yet produce. (5) <strong>Theory (AO2):</strong> universal babbling to first words (Jakobson); the systematicity argument supports Chomsky's rule-based acquisition; CDS phonology - higher pitch, exaggerated intonation - as interactional support (Vygotsky/Bruner). Use the provided IPA symbols precisely; OCR rewards terminology explicitly.
'Children's early language is used mainly to get things they want.' Referring to data in detail, and to relevant ideas from language study, evaluate this view of children's language development. [30 marks] (AQA 7702/1 Section B style)
Answer: Mark-scheme logic (AO1 15 + AO2 15):<br><br><strong>AO2 frame - the view maps onto Halliday's instrumental and regulatory functions</strong> (language as doing).<br><br><strong>AO1 - data evidence:</strong> categorise utterances by function - instrumental ('want juice'), regulatory ('stop it') - but count the others: interactional ('hello'), personal ('me big'), heuristic ('what that?'), imaginative (play talk), representational (naming) - and quote each category you find. The tally is your evidence.<br><br><strong>AO2 - evaluation:</strong> <em>Support:</em> instrumental/regulatory utterances can dominate at the holophrastic stage; caregivers respond fastest to requests (contingent interaction - Vygotsky's ZPD in action). <em>Challenge:</em> interactional and heuristic utterances usually outnumber requests in transcripts (as in Halliday's own data on his son Nigel); pragmatic development moves from request-centred routines to turn-taking, politeness and eventually implicature; the mathetic/pragmatic split shows doing and learning are intertwined. <em>Judgement:</em> 'getting things' is one function among seven - early language is as much social and heuristic as instrumental.<br><br><strong>Craft:</strong> balanced, data-anchored evaluation with an explicit conclusion - the discursive register AQA's Section B mark scheme rewards.
amazon.co.uk
