The Vocabulary Spurt — The 50-Word Inflection Point and the Move to Two-Word Speech

Lead

"She only had about ten words last week. Suddenly she's adding new ones every day." From around 18 months, parents sometimes describe this shift. At the same time, a different kind of worry surfaces: "The children her age are already putting two words together — ours is still around twenty words." Both reactions come from the same period.

Language development is said to include a phase of rapid acceleration often called the vocabulary spurt. But this spurt does not happen at the same time, or with the same intensity, for every child. The more accurately documented fact — worked over for decades in research — is that some children accelerate sharply while others do not. That heterogeneity is where the story actually begins.

This article covers what the research says about the inflection point around 50 words, the move from single words to two-word combinations, and how to read a child who appears to be running a little behind.

The 50-Word Mark: An Inflection Point, Not a Finish Line

The vocabulary spurt was first described systematically by Goldfield and Reznick in a 1990 longitudinal study [1]. Starting at 14 months, they tracked the expressive vocabularies of 18 children via maternal diaries, checking every two-and-a-half weeks. In 13 of the 18 children, a period of sharp vocabulary acceleration — a spurt — was identifiable within a three-month window. Roughly three-quarters of the words added during the spurt were nouns, which gave rise to the term "naming spurt." The remaining five children showed no clear acceleration; they added nouns and other word classes in a more gradual, balanced pattern [1].

The spurt, in other words, was characterized by the original researchers as something that occurs in some children and not others. The 50-word figure is an average-level description and a rough marker for where the spurt tends to begin — not a deadline. Bates and colleagues' subsequent reviews reinforced this point, noting that vocabulary growth shows both continuous and stage-like patterns depending on the child, and that individual profiles vary substantially [2].

The normative data most often cited in the research literature comes from the MacArthur-Bates Communicative Development Inventories (CDI) [3]. The US norms describe expressive vocabulary from 8 to 30 months across the 10th, 25th, 50th, 75th, and 90th percentiles. At any given age, the spread between the top and bottom of that range is several times wider than a single number suggests. The CDI norms are the appropriate reference for understanding how much variability is normal — far more informative than a one-line developmental table.

For Japanese families, the Cabinet Office's Children and Families Agency (Kodomo Katei-cho) published a summary of the 2023 Survey on Infant and Child Physical Development in 2024 [4]. By 16–17 months, more than 90% of caregivers report their child says single words; below 12 months, fewer than half do. The same survey noted a slight decline in the proportion of infants showing word production compared with the prior survey [4].

From Single Words to Two-Word Speech: Brown's Stage I

Once vocabulary reaches roughly 50–60 words, children begin combining them. "Mama gone." "Car came." These two-word utterances mark the beginning of syntax.

The classic framework for characterizing this stage is Roger Brown's Stage I, defined as a mean length of utterance (MLU) of 1.5–2.0 morphemes — roughly corresponding to the 15–30-month range [5]. The semantic relations that appear at this stage are limited in number: recurrence ("more milk"), negation ("no nap"), possession ("daddy shoe"), agent + action ("doggy sit") [5]. Stage II, following at an MLU of 2.0–2.5 and roughly 28–36 months, adds early inflectional morphemes — the "-ing" progressive, regular plurals, and so on [5].

A crucial point Brown himself made: the stage definitions are based on MLU, not on age. He introduced MLU precisely as a way to summarize developmental stage in place of chronological age, because the mapping between age and linguistic complexity is loose. The design intention is to assess stage by the length of the child's own utterances, not by the number of months since birth.

Bates, Dale, and Thal's review adds a further complication: even within typical development, vocabulary and grammar can dissociate [2]. A child may have strong vocabulary but delayed grammar, or the reverse. Measuring development against a single index — word count alone, or two-word-sentence onset alone — can obscure these individual profiles.

When the anxiety "shouldn't two-word sentences be here by now?" surfaces, it is worth remembering that Brown's framework would push back against the age frame entirely.

How to Read a Late Talker

Children identified at age two as having fewer than 50 expressive words and no two-word combinations have been described in the research literature as late talkers. One widely used screening instrument is the Language Development Survey (LDS), a parent-report vocabulary checklist developed by Rescorla in 1989 and used to identify expressive vocabulary delay in toddlers [6].

In Rescorla's long-term follow-up studies, many children identified as late talkers at age two reached age-typical levels by school age [7]. However, when outcomes were examined at age 17, the late-talker group scored on average below a socioeconomically matched comparison group on measures of vocabulary, grammar, and verbal memory [7]. Late talkers are sometimes described in shorthand as "they'll catch up, don't worry" — but the longitudinal data present a more complex picture.

The appropriate take-away from this evidence is not "act urgently" but rather: "there is a reason to keep observing and to keep options for support available." Rescorla herself has noted that no predictor currently exists that can reliably identify which individual late talker will close the gap and which will not [8]. In the absence of that predictor, combining population-level trend data with careful individual observation is the most defensible approach.

In Japan, single-word production is checked at the 18-month checkup, a universal pediatric visit. If there is concern at any point, contacting a child health center or pediatrician is low-cost, and consulting unnecessarily costs nothing. Two indicators that provide signal earlier than word count alone: whether comprehension seems intact (does the child respond to their name, to simple requests?) and whether joint attention behaviors are present — pointing, gaze-following, expressed interest in other people. These can be observed before a vocabulary gap is large enough to see clearly.

Practical Application: List Words, Don't Count Them

Counting words daily is tiring and, in most cases, not illuminating. A more sustainable approach is to record new words in order of appearance.

• "Mmm" (said at mealtimes): 15 months
• "Vroom" — first used pointing at a car: 16 months
• "Mama," "Dada" — with a clear addressee, not just sounds: 17 months
• "No" and "more" emerge at 19 months; within days, "more juice" appears

This kind of log does not track vocabulary on a horizontal count-axis. It documents the child's own lexical history on a vertical timeline. Looking back at it later, you can judge for yourself whether a spurt occurred or whether growth was gradual [1,2]. A time-stamped tool like Memori works for this, as does a paper diary — format is not the point. What matters is logging at consistent density, so that rapid periods and slow periods both leave a trace.

One practical note: parent recall of developmental milestones is known to shift in the direction of earlier ages the further from the event the recall takes place. When the 18-month or 3-year checkup asks about first words or when two-word sentences appeared, a contemporaneous note is a more accurate primary source than memory.

Summary

The 50-word mark is a loose inflection point, not a passing grade. A clear spurt appears in some children; gradual growth appears in others [1,2]. Brown's stage framework was designed around MLU, not age, and dissociations between vocabulary and grammar occur even in typical development [5,2]. Many late talkers reach age-typical levels, but a subset show residual differences into later childhood — meaning the evidence supports continued observation, not urgency, and not dismissal [7,8].

List the words that appear rather than counting them. Development reads more richly on a vertical timeline than on a horizontal tally.

References

1. Goldfield BA, Reznick JS. Early lexical acquisition: rate, content, and the vocabulary spurt. J Child Lang. 1990;17(1):171–183. doi:10.1017/S0305000900013167. PMID: 2312640.
2. Bates E, Dale PS, Thal D. Individual differences and their implications for theories of language development. In: Fletcher P, MacWhinney B, eds. The Handbook of Child Language. Oxford: Blackwell; 1995:96–151.
3. Fenson L, Marchman VA, Thal DJ, Dale PS, Reznick JS, Bates E. MacArthur-Bates Communicative Development Inventories: User's Guide and Technical Manual. 2nd ed. Baltimore: Brookes Publishing; 2007.
4. Children and Families Agency, Cabinet Office, Japan. Summary Report of the 2023 Survey on Infant and Child Physical Development. 2024. https://www.cfa.go.jp/policies/boshihoken/r5-nyuuyoujityousa
5. Brown R. A First Language: The Early Stages. Cambridge, MA: Harvard University Press; 1973.
6. Rescorla L. The Language Development Survey: a screening tool for delayed language in toddlers. J Speech Hear Disord. 1989;54(4):587–599. doi:10.1044/jshd.5404.587. PMID: 2811339.
7. Rescorla L. Age 17 language and reading outcomes in late-talking toddlers: support for a dimensional perspective on language delay. J Speech Lang Hear Res. 2009;52(1):16–30. doi:10.1044/1092-4388(2008/07-0171). PMID: 18723598.
8. Rescorla L. Late talkers: do good predictors of outcome exist? Dev Disabil Res Rev. 2011;17(2):141–150. doi:10.1002/ddrr.1108. PMID: 23362033.