An SLP explains the 50-word rule, why bilingual toddlers get miscounted, and how to tell watchful waiting from a referral. Ages 18–30 months.
“Late talker” is one of those phrases that sits in a parent’s chest at 2am. You’ve counted the words. You’ve watched the cousin who is three months younger narrate her entire day. And somewhere in the back of your mind is a worry you haven’t said out loud yet: did we cause this by doing two languages?
Short answer to that one, before anything else: no. Not possible. I’ll come back to why.
But “late talker” is also a real clinical term with a real definition, and most of what you’ll find online either waves it away — Einstein didn’t talk until he was four! — or treats it as a diagnosis. It’s neither. Here’s what it actually means, and what changes when the child is bilingual.
If your question is the broader one — is my kid roughly where she should be? — start with is my bilingual toddler on track instead. This post is narrower: it’s about one specific pattern, in the 18-to-30-month window, and what to do about it.
It’s a described pattern, not a diagnosis. ASHA — the national body that credentials speech-language pathologists — calls it late language emergence, and the criteria most widely used are:
Fewer than 50 words, and no two-word combinations, by 24 months.
That’s from Rescorla and Paul’s work in the late 80s and early 90s, and it’s still the benchmark (ASHA, Late Language Emergence).
Two things matter in that sentence and both get missed.
It’s “and,” not “or.” A child with 30 words who is combining them — más agua, no bath, mamá zapato — is doing something different from a child with 30 words and no combinations at all. Combining is a structural skill. It’s often the more informative signal.
It’s specifically about expressive language. The classic late talker understands normally. She follows directions, finds the thing you named, brings you the shoe. What’s lagging is output. If comprehension is also behind, that’s a different picture and a faster referral — more on that below.
Late talkers also have, by definition, normal hearing, no cognitive concerns, and no other diagnosis that explains the delay. If any of those are in question, the label doesn’t apply and you’re in a different conversation.
Here’s where I see families — and, honestly, some clinicians — go sideways.
Count across both languages. If your daughter has 30 words in English and 25 in Spanish, she does not have 30 words. She has around 55, and you’d count agua and water once, as one concept, not twice. That’s called conceptual scoring, and it is the correct way to count a bilingual child’s vocabulary.
I cannot tell you how many parents have handed me a number that only counted English. A child who looks under the benchmark in one language routinely clears it when you count the whole child.
Count everything, not just the pretty words. Sounds that consistently mean a thing are words. If bah always means bottle, that’s a word. Animal noises count. Uh-oh counts. Signs count. Words that only grandma understands count. You’re inventorying communication, not diction.
Do this in writing. Nobody can hold a toddler’s vocabulary in their head, and anxiety inflates or deflates it depending on the day. Keep a list on the fridge for two weeks and add words as you notice them. Parents are almost always surprised — usually upward.
I’ve written more on the counting problem in speech delay vs. bilingual difference.
Because this is the question underneath the question, let me be direct about it.
Bilingual exposure does not cause late language emergence. It does not cause language disorders. It does not slow the timeline. Bilingual children hit language milestones on a comparable schedule to monolingual children when you measure them properly — which means conceptually, across both languages.
What bilingualism does do is make a late talker harder to spot, in both directions. Split a smaller vocabulary across two languages and each one looks thinner than it is, so typically-developing kids get over-referred. And in the other direction, a genuine delay gets waved off for a year because someone said “oh, bilingual kids talk later” — which isn’t true, and which costs a child a year of access.
Both errors come from the same place: counting one language and calling it the child. More on the causation question in will being bilingual delay speech.
This is the part where I’d rather be honest than comforting, because you’ll find plenty of sites willing to be comforting.
ASHA’s own summary is that roughly 50 to 70 percent of children with late language emergence catch up to their peers by late preschool or school age. That’s genuinely good news and it’s the most likely outcome.
It also means somewhere between a third and half of them don’t — they go on to have persistent language difficulties that benefit from support. Which is why “most kids grow out of it” is true and is still not a reason to do nothing. The problem is there’s no reliable way to know in advance which group your child is in. You find out by watching, and waiting costs more for the children in the second group than monitoring costs for the children in the first.
That asymmetry is the whole argument for checking rather than reassuring yourself.
If your child fits the late-talker pattern — expressive language behind, comprehension fine, hearing fine, everything else on track — then monitoring is a legitimate plan, not a cop-out. ASHA’s guidance is to re-check at regular intervals, about every six months, and look at the trajectory rather than any single number. A child adding words steadily is telling you something different from a child who has been stuck at the same twenty words since summer.
Get the hearing checked first, before anything else. Not because I think your child can’t hear, but because intermittent hearing loss from repeated ear infections is common, easy to miss, and one of the most treatable things on this list. A child hearing through fluid gets an inconsistent signal — and the language they hear less of, usually the minority language, takes the bigger hit. Before you accept any explanation for a language delay, rule this one out.
Call now, don’t wait, if you see any of these:
That last one belongs on the list. Parent concern is one of the better predictors we have, and I have never once regretted taking it seriously.
An evaluation is not a commitment to therapy. It’s information. And if you do get one, insist it be a real bilingual evaluation — I’ve written about how to find an SLP who can actually do one and how to read the report you get back.
If your child is two, has plenty of words in Spanish at home, and has gone quiet since starting an English-speaking daycare — that is probably not late language emergence. That’s the silent period, a normal and temporary response to being dropped into a new language. Different thing, different timeline, different response from you.
This part doesn’t change based on which category your child is in, which is why I’d start here regardless:
And keep talking to her in the language you’re warmest in. A parent who is relaxed and chatty in Spanish gives a child more usable language than a parent performing careful English. That’s true for a late talker and it’s true for everybody else.
Hi, I’m Lindsey Carleton, MA, CCC-SLP — a licensed speech-language pathologist with more than 11 years of clinical experience, and a fellow toddler mom. I learned Spanish in college and in healthcare work, not at home, so I build for parents who are doing this without being native speakers.
If you want a structured place to start, our free Spanish printables include a first-week plan built around the routines you already have. The 12-month curriculum sequences vocabulary in developmental order across a year.
This post is general information, not clinical advice about your child. If you’re worried, the right move is an evaluation by an SLP who can assess in both of your languages.
This is the part where I’d rather be honest than comforting, because you’ll find plenty of sites willing to be comforting. ASHA’s own summary is that roughly 50 to 70 percent of children with late language emergence catch up to their peers by late preschool or school age. That’s genuinely good news and it’s the most likely outcome. It also means somewhere between a third and half of them don’t — they go on to have persistent language difficulties that benefit from support. Which is why “most kids grow out of it” is true and is still not a reason to do nothing. The problem is there’s no reliable way to know in advance which group your child is in. You find out by watching, and waiting costs more for the children in the second group than monitoring costs for the children in the first. That asymmetry is the whole argument for checking rather than reassuring yourself.
This page is part of Palabra Garden. View all pages: llms.txt · llms-full.txt