You say where's your shoe and nothing happens. You say it again, louder, slower, with your eyebrows doing the work. Still nothing. Then — a beat too late, long after you'd already decided they weren't listening — they turn and look at the shoe.
In that gap, most parents feel one of two things. Irritation: she's ignoring me. Or a small cold worry: does he actually understand? Both are wrong, and the truth underneath is stranger and more useful than either. Your toddler is not ignoring you. Your toddler is computing. And the speed of that computation — the raw milliseconds between hearing a sound and knowing what it means — is one of the most quietly predictive things about a young child's language, more telling in some ways than the number of words they can say.
The uncomfortable part is that the gap is real, and it is longer than you think, and most of us fill it with more talking.
Understanding a word is not a lookup. It's a race.
We imagine word recognition like a dictionary: the sound comes in, the meaning comes out. It isn't. In adults and children alike, hearing a word triggers a race between competing candidates. The moment you hear sh-, your mind partially activates shoe, ship, shirt, shovel — and they compete until enough sound has arrived to let one win. Fluent listeners resolve this race so fast it feels like nothing is happening at all.
Toddlers are running the same race with worse odds. Their word forms are less sharply specified. Their vocabularies are new. And in real life, the signal is filthy — a dishwasher, a sibling, a car, your own voice trailing off.
Researchers can watch this race happen. In a method called looking-while-listening, developed by Anne Fernald and colleagues at Stanford, a child sits on a caregiver's lap facing two pictures — a ball and a dog, say. A voice says Where's the ball? Frame-by-frame video shows exactly when the child's eyes move to the correct picture. Not whether. When.
And the developmental picture is remarkable. Younger toddlers, around 15 months, tend to shift their gaze only after the word has finished — they wait for the whole thing, hear "ball," then look. By around 24 months, children are shifting mid-word. They hear ba– and their eyes are already moving. They no longer wait for the end of the word because they no longer need it.
That is the developmental milestone almost no one talks about. Not more words. Faster words.
Why speed matters more than the number
Here is where it stops being a lab curiosity. Fernald and Virginia Marchman found that how quickly a toddler recognizes familiar words predicts how much vocabulary they will gain over the following months. Speed at the earlier age forecasts size at the later age. And in longer follow-ups, processing efficiency in the second year has been linked to language and cognitive outcomes years later, into the school years.
Why would that be? Think about what a slow toddler is spending their resources on. Language arrives in a stream and does not wait. If a child burns most of their working memory decoding ball, there is nothing left over for the three words that came after it. If they decode ball instantly, the rest of the sentence is still there, still available, still learnable. Efficiency doesn't just help you understand the sentence you're in — it buys you the attention to learn from the next one. Fast processing compounds.
There's a corollary that stings a little: the child who seems to "pick things up so easily" often isn't smarter in some fixed sense. They're faster at the front door, so more gets in.
What actually makes a toddler faster
The encouraging finding — and this is the heart of it — is that processing speed is not a fixed trait you're issued at birth. It's built.
Adriana Weisleder and Fernald ran a study that took the well-known link between how much a child is talked to and how many words they know, and looked underneath it. What they found was a chain: children who heard more speech directed at them (not merely overheard around them) became more efficient real-time processors of language, and that efficiency helped explain their larger vocabularies. Talk didn't just deposit words. It tuned the machine that catches words.
The mechanism is repetition doing something more interesting than memorization. Every time a child hears shoe in a slightly different sentence, in a slightly different voice, in a slightly different room, the mental representation of that word gets sharper — more resistant to noise, more distinguishable from shirt, recognizable from a smaller fragment of sound. This is why the twentieth time you say a word is not redundant with the first. The first time built a word. The twentieth time built a fast word.
A few structural details from the research are worth knowing, because they are things you control:
Position in the sentence matters. Words at the end of an utterance — look, a duck — are easier to recognize than words buried mid-sentence. The end of a phrase is prosodically stretched, louder, and followed by silence rather than more competing sound.
Familiar frames help. A predictable carrier phrase (Where's the..., Look at the...) lets the child spend nothing on the frame and everything on the word that matters.
Noise is not neutral. Toddlers are dramatically worse than adults at recognizing words against background sound. The TV playing to no one isn't background for them. It's static across the signal.
And time. The response you're waiting for might be arriving. It's just arriving after you've already started talking over it.
The waiting problem
Adult conversation runs on gaps of roughly a fifth of a second. That is the rhythm in our bones. A silence of two seconds feels, to us, like a small social emergency — like something has gone wrong and we should fix it.
So we fix it. We repeat the question. We rephrase it. We add a word. Every one of those repairs arrives while the child is still mid-race, and each new sentence resets the race with new competitors. The most common way parents accidentally slow their toddler down is by being kind, being helpful, and refusing to shut up.
A toddler needs several seconds — often four, sometimes more — to hear, resolve, retrieve, and organize a response. Not because they're slow children. Because they're new listeners with a full processing load.
The hardest, cheapest, most effective language intervention available to you is silence, held slightly past the point of comfort.
Your next moves
- Count to five in your head after you ask a question. Actually count. Ask where's your cup?, then silently count one-two-three-four-five before saying anything else. You will feel the urge to rescue at about two. Don't. Notice how often the answer arrives at four.
- Put the target word last. Instead of the duck is over there, say look — over there — a duck. Move the word you want them to catch to the end of the sentence, where the silence protects it.
- Pick one predictable frame and reuse it all week. Where's the ___? or I see a ___. The repeated frame is free for their brain to process, so all their effort lands on the new word.
- Kill one source of background sound today. The radio during breakfast, the TV nobody's watching. Toddlers can't filter competing speech the way you can — you're asking them to do sprints in sand.
- Say the same word in three different sentences today, not the same sentence three times. Your shoe. Where's your shoe? Let's put on your shoe. Variation with a constant word is what sharpens a representation.
One more thing
None of this requires an app. It requires a pause you are probably too uncomfortable to hold, a word moved to the end of a sentence, and a radio switched off.
But if you want a place where those conditions already exist — where a word arrives at the end of a phrase, in a quiet frame, with a pause built in on purpose and no one talking over your child while they think — that's what Acorn is. Three minutes a day, one word at a time, with the silence left in. No ads to fill the gap. No upsells. Just the race, run slowly enough to win.
Your toddler isn't ignoring you. They're almost there. Wait one more second.