The pediatrician looks up from the chart, friendly and unhurried, and asks the question: "So, how many words is she saying these days?"
And you freeze. You know she says ba. You know there's a sound she makes at the dog that is definitely about the dog. There was a morning last week when she said something that might have been "up," or might have been "cup," or might have been a sneeze. You love this child more than anything, you listen to her all day, and you cannot produce a number.
If that scene is familiar, nothing is wrong with your attention. Memory simply isn't built for this task.
Why a loving parent can't count
A toddler's early vocabulary is slippery in ways that defeat memory. Words arrive as approximations ("baba" for bottle, "nana" for banana, and sometimes for grandma). They show up once, vividly, and then vanish for a week. They get used in one room and not another. Researchers who study early language have long noted how context-bound and unstable first words are.
The human mind is also generous in one direction and stingy in the other. We remember the dramatic moments, like the first clear "no" in the grocery store. We forget the quiet, ordinary ones, like the "bye" that has been part of every departure for a month and has simply become furniture.
So when you're asked to produce a number from memory, you tend to get one of two failures. Either you undercount, because you can only recall the words from the last few days. Or you anxiously overcount, because you're including things you've only wished she'd say. Neither helps you or the person asking.
Recognition is easier than recall
Here is the useful part. Psychologists have known for a long time that recognizing something is far easier than recalling it cold. You can't list every actor in your favorite show from memory, but show you a cast list and you'll nod at every name.
This is the principle behind one of the most widely used tools in child language research, the MacArthur-Bates Communicative Development Inventories. Instead of asking a parent "what words does your child say?" with a blank page, the CDI hands over a long checklist of common early words (animals, foods, body parts, household objects, action words) and asks the parent to mark which ones the child says. Versions exist for roughly the first and second years, and a body of research has found that parent report on this kind of checklist tracks well with what researchers measure directly in the lab. Parents, it turns out, are excellent observers when given the right prompt.
The lesson isn't that you need a clinical form. It's that the method matters more than your memory. A list you can scan beats a blank mind every time.
What counts as a word
The most common reason parents undercount is a rule they've quietly invented: that a word must be clear enough for a stranger to understand. Language researchers don't use that rule. A word generally counts if your child uses it consistently and meaningfully for the same thing or purpose, even if it's pronounced in their own way.
So "ba" for ball counts, as long as it's reliably about balls (and ideally not also about bottles, bubbles, and the moon). "Mo" for more counts. Animal sounds used consistently for an animal are usually counted by researchers as early words too. A word she only repeats right after you say it, with no sign she means anything by it, is shakier territory. You're looking for use that is spontaneous and has a point.
This matters because the number is less important than the pattern behind it: is she using sounds on purpose to refer to things and get things? That's the heart of early language, and a word list captures it far better than a vague feeling.
Why the number actually matters
In research on late talkers, one commonly used screening approach looks at a child around age two. The Language Development Survey, a parent-report tool developed by Leslie Rescorla, flags concern when a child around 24 months has fewer than 50 words, or isn't yet combining words. That is a screening signal, not a diagnosis, and plenty of children who start slowly catch up. But it shows why a real count is useful: it turns "I'm worried" into something specific that a professional can work with.
And the range of normal is enormous. Children at the same age can differ by dozens or even hundreds of words and both be perfectly typical. A list doesn't exist to rank your child against anyone. It exists to show her own trajectory: whether the words are growing, whether new kinds of words (actions, descriptions) are appearing, and whether anything has stalled.
Trends are what a pediatrician or speech-language pathologist can actually use. "She had about eight words in July and about twenty now" tells a very different story than "she doesn't say much."
Your next moves
This takes a few minutes to set up and a few seconds a day to maintain.
- Start a notes-app list tonight. Open a note on your phone titled with your child's name and write down every word you can think of, approximations included. Next to unusual ones, add what they mean ("nana = banana"). Don't aim for complete; aim for started.
- Add a date to each new word. When a new one appears, jot it down with the date, even just "Nov 12." Within a month you'll have a growth curve instead of a feeling.
- Do a recognition pass once a week. Sunday evening, scan a list of common first words (a children's word-book index works well, or any CDI-style list you can find online) and ask yourself of each one: have I heard this lately? Your memory will surface words it would never have recalled cold.
- Include the non-word communication. Make a second short list for gestures and consistent sounds that mean something: pointing, waving, a particular "eh!" that means I want that. These are language too, and they predict the words that come next.
- Bring the list to the next appointment. Hand it over or read it aloud. If you're worried, you now have something concrete to say, and a professional has something to act on. If you're not worried, you've given yourself a record you'll treasure.
Two gentle rules keep this from becoming another source of stress. First, track words you notice, never words you drill. The point is to observe, not to test her. Second, if the list feels discouraging on a given week, put it down. Language grows in bursts and plateaus, and a list is a window, not a verdict.
A record you didn't know you'd want
There's a quieter reason to keep the list, one that has nothing to do with screening. These words are going to vanish. Not the language; the sound of it. The way she says "bubbles" as "bu-bu" will give way to the correct word in a few months, and you will never hear it again. The list is where "nana means banana and also grandma and also, sometimes, the moon" gets to live on.
We photograph their faces because we know faces change. Words change just as fast, and almost no one writes them down.
Where Acorn fits
The words on your list grow fastest when they're met with short, regular, undistracted moments of attention. That's the idea behind Acorn, which offers a daily three-minute first-words session for one- to three-year-olds, with no ads and no upsells. It's a small ritual that gives you something worth adding to the list, and a reason to notice. If you'd like to try it, you can find it at acorn.lumenlabs.works. But the notes app is free, and the best first move is the one you can make tonight.