You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Grammatical development describes how children acquire the rules for combining words into phrases, clauses and sentences. It has two branches, and you must keep them distinct in analysis:
-s, past tense -ed, progressive -ing and possessive 's.Grammar is arguably the single most fertile area for child-language data analysis, because grammatical errors are so systematic and so theoretically loaded. A child's "virtuous errors" in particular are among the strongest pieces of evidence in the whole nature–nurture debate, which is why this topic recurs in AQA Paper 1 Section B. Two further pieces of vocabulary are worth fixing now. Content (lexical) words — nouns, main verbs, adjectives, adverbs — carry the main meaning; function (grammatical) words — articles, prepositions, auxiliary verbs, pronouns — express grammatical relationships. The earliest grammar consists almost entirely of content words, with function words and inflections added later; tracking that shift is central to assessing a child's stage.
Before tracking grammar you need the unit of morphology: the morpheme, the smallest meaningful unit of language. Cat is one morpheme; cats is two (cat + plural -s); unhappiness is three (un- + happy + -ness). Morphemes come in two basic kinds:
-s, the past-tense -ed, the prefix un-.A further distinction is crucial for child language:
| Type of bound morpheme | What it does | Examples | Note |
|---|---|---|---|
| Inflectional | Marks a grammatical category without changing the word class | plural -s, past -ed, progressive -ing, possessive 's, comparative -er | English has only a handful; these are the morphemes Brown tracked |
| Derivational | Creates a new word, often changing word class | -ness (happy→happiness), -ly (quick→quickly), un- (do→undo) | Acquired more gradually and later |
Children's grammatical development is, to a large extent, the story of acquiring the inflectional system — and their over-application of it (saying goed, mouses) is the headline evidence we examine below. A subtle but examinable point is that a single inflectional morpheme can have several spoken forms, called allomorphs: the regular plural is pronounced /s/ after a voiceless sound (cats), /z/ after a voiced one (dogs) and /ɪz/ after a sibilant (horses). Remarkably, children master this conditioning very early — evidence that what they acquire is an abstract rule, not a list.
Children begin with single-word utterances (holophrases) around 12–18 months. The move to two-word utterances, typically around 18–24 months, is a watershed: for the first time the child is combining words, and the consistent ordering of those combinations shows an emerging grasp of syntax.
Crucially, two-word utterances are not random pairs. Roger Brown and others showed that they encode a small, recurring set of semantic relations — and that children reliably put the words in a meaningful order (Daddy kick, not usually kick Daddy for the agent meaning), which is why word order is taken as the first sign of true syntax:
| Semantic relation | Example | Intended meaning |
|---|---|---|
| Agent + action | Daddy kick | Daddy is kicking |
| Action + object | Kick ball | (I am) kicking the ball |
| Agent + object | Mummy sock | Mummy is handling the sock |
| Entity + attribute | Big doggy | The dog is big |
| Entity + location | Teddy bed | Teddy is on the bed |
| Possessor + possessed | My cup | That is my cup |
| Recurrence | More juice | I want more juice |
| Negation | No bed | I don't want to go to bed |
Key Definition: Two-word stage — the period (approximately 18–24 months) in which children combine two words to express a limited set of semantic relations. Although grammatical markers (articles, inflections, auxiliaries) are absent, the consistent word order reveals an emerging syntactic system.
From roughly two years, utterances lengthen into telegraphic speech — multi-word strings that keep the high-information content words but drop the function words and inflections, exactly as an old telegram, charged by the word, would omit everything inessential.
| Adult form | Telegraphic form |
|---|---|
| "I want more milk, please" | More milk |
| "Daddy is kicking the ball" | Daddy kick ball |
| "The cat is sitting on the mat" | Cat sit mat |
| "I don't want to go to bed" | No go bed |
The systematic retention of content words and omission of function words is itself evidence of grammatical organisation: the child is not dropping words at random but preserving precisely the elements that carry referential meaning while the grammatical scaffolding is still being built.
Martin Braine (1963) offered an early structural description, pivot grammar. He observed that some words behave as pivots — a small, frequent set fixed in one position (more ___, all-gone ___, ___ off) — combined with a large class of open words:
more, no, all gone).milk, juice, sock).more milk, more juice, more biscuit — more is the pivot in initial position.Pivot grammar is worth knowing but is now regarded as descriptively too thin: it captures the positional regularity but misses the meaning relations (agent, location, possession) that a richer analysis reveals, and some children's data do not fit the pivot/open split neatly. Treat it as a historically important first attempt rather than the last word.
From around 2–3 years, utterances lengthen and the missing grammatical machinery starts to appear. The landmark study is Roger Brown (1973), a longitudinal investigation of three American children he gave the pseudonyms Adam, Eve and Sarah. Brown made two enduring contributions: a measure of grammatical maturity, and a discovery about the order in which grammatical morphemes are acquired.
Key Definition: Mean Length of Utterance (MLU) — a measure of grammatical development obtained by dividing the total number of morphemes in a speech sample by the number of utterances. Because morphemes (not words) are counted, cats scores 2 (cat + plural
-s) and walked scores 2 (walk + past-ed). A rising MLU indicates growing grammatical complexity.
Brown used MLU to define five stages of development. Note that the ages are approximate and vary considerably between children — MLU tracks grammatical level better than chronological age does, which is precisely why Brown preferred it:
| Brown's stage | MLU (approx.) | Age (approx., variable) | Key features |
|---|---|---|---|
| I | 1.0–2.0 | 12–26 months | Single words and two-word combinations; telegraphic |
| II | 2.0–2.5 | 27–30 months | First grammatical morphemes (-ing, plural -s, in, on) |
| III | 2.5–3.0 | 31–34 months | Simple clause structures; emerging questions and negatives |
| IV | 3.0–3.75 | 35–40 months | Embedding; developing auxiliary verbs |
| V | 3.75–4.5 | 41–46 months | Coordination and subordination; complex clauses |
Brown's most cited finding is that the 14 grammatical morphemes he tracked are acquired in a strikingly consistent order across children — an order that does not match the frequency with which the morphemes occur in parental speech, which is a serious problem for any purely input-driven (behaviourist) account. The general progression:
| Order | Morpheme | Example |
|---|---|---|
| 1 | Present progressive -ing | Daddy running |
| 2 | Preposition in | in box |
| 3 | Preposition on | on table |
| 4 | Regular plural -s | two dogs |
| 5 | Irregular past tense | went, came |
| 6 | Possessive 's | Mummy's hat |
| 7 | Uncontractible copula be | He was here |
| 8 | Articles a / the | a dog, the cat |
| 9 | Regular past tense -ed | walked |
| 10 | Third-person regular -s | He runs |
| 11 | Third-person irregular | He does, She has |
| 12 | Uncontractible auxiliary be | He was running |
| 13 | Contractible copula be | He's good |
| 14 | Contractible auxiliary be | He's running |
The order is best explained by grammatical and semantic complexity, not frequency: simpler, more concrete notions (an ongoing action -ing; basic location in/on) precede semantically subtler ones, and the contractible forms come last because they require the most grammatical control.
Brown's finding is one of the most powerful single pieces of evidence in the whole field, and it is worth being explicit about why. He checked whether the acquisition order tracked the frequency of each morpheme in the parents' speech — the prediction a behaviourist would make, since more-heard, more-reinforced forms should be learned first. It did not. The morphemes are not acquired in frequency order; they are acquired in an order that matches their grammatical and semantic complexity, and that order is consistent across children who hear very different input. This decouples acquisition from input frequency and points to the child following an internal developmental schedule — strong support for a maturational, nativist-leaning view, and a serious problem for strict behaviourism. A balanced evaluation notes the limits: Brown studied only three children acquiring English, so we should be cautious about claims of strict universality across languages, and richly inflected languages (such as Turkish or Finnish) raise their own questions about how morphology is acquired. Still, the core result — order driven by complexity, not frequency — is robust and replicated.
MLU is a useful, objective index but it is not infallible, and a sophisticated answer flags both sides:
| Strength of MLU | Limitation of MLU |
|---|---|
| Objective and quantifiable; allows children to be compared on a common scale | Becomes unreliable above about 4.0, where utterance length stops tracking real complexity |
| Counts morphemes (not just words), so it captures inflectional growth | Two utterances with the same MLU can differ greatly in genuine grammatical sophistication |
| Tracks development better than chronological age | Sensitive to the sample: mood, task and topic all affect how a child talks on the day |
The single most theoretically important feature of children's grammar is overgeneralisation — applying a regular rule to a form that is irregular, producing a word no adult uses.
| Type | Adult form | Child's overgeneralisation |
|---|---|---|
| Past tense | went | goed |
| Past tense | ran | runned |
| Past tense | swam | swimmed |
| Plural | mice | mouses |
| Plural | sheep | sheeps |
| Plural | feet | foots / feets |
| Comparative | better | gooder / more better |
Key Definition: Overgeneralisation (virtuous error) — the over-application of a regular grammatical rule to an irregular form (
goedfor went,mousesfor mice). It is called virtuous because, although the output is non-standard, it demonstrates that the child has internalised a productive rule rather than memorising individual forms.
These errors are decisive evidence against behaviourism and for an active, rule-forming child (the nativist position): a child cannot have imitated goed, because no one says it, and it was certainly never reinforced — so it must be generated by a rule the child has abstracted for themselves.
Overgeneralisation produces a famous U-shaped curve in the accuracy of irregular forms — accuracy starts high, dips, then recovers:
went), but only as an unanalysed, memorised whole.-ed rule and over-applies it, now producing goed and even self-correcting back to the error — accuracy on irregulars temporarily falls.walked and went).The temporary regression in Stage 2 is the give-away: an imitation-based learner could only get better, never worse. The dip happens precisely because the child has formed a rule and is applying it too widely — progress disguised as error.
Jean Berko (1958) designed an elegant experiment — the wug test — to prove that children possess productive grammatical knowledge rather than a memorised inventory. Shown a picture of an invented bird-like creature, the child hears:
"This is a wug. Now there are two of them. There are two ___."
Even pre-school children reliably supply wugs, correctly applying the regular plural to a word they have never heard. Because wug is novel, the child cannot be retrieving a stored plural; they must be applying a rule. Berko also probed other inflections with further nonsense words:
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.