We got a prompt from Hannah this week. She and Daniel's son Ezra is about a year and a month old, and he's discovered the joy of launching food off his high chair. Hannah says it started as what looked like accidents or maybe experimenting, but now it's deliberate — even when he's clearly hungry, he throws his food on the floor instead of eating it. They pick it up, he throws it back down. She's asking: is there a developmental reason for this? How do they get enough food into him? And should they keep picking it up, or is that reinforcing the whole thing?
This is one of those parenting moments where the behavior looks completely irrational from the outside — hungry baby rejects food by hurling it across the room — but from inside that twelve-to-fifteen-month-old brain, it makes perfect sense. And I say this as someone who spent years in pediatrics watching parents exhaust themselves trying to solve the wrong problem.
The wrong problem being "how do I stop the throwing" instead of "what is the throwing actually for."
And once you understand what it's for, the strategy changes completely. So I think we need to start with what's actually happening in that twelve-to-fifteen-month-old brain, because this looks like defiance, but it's something much more interesting. I remember one family I worked with — their daughter was about fourteen months, and they came in convinced she had some kind of behavioral disorder because she was throwing everything. Food, toys, shoes, the remote control. The mom was in tears. And I asked them one question: when she throws something, what does she do immediately after? They looked at each other and said, "She watches it fall. Every time. She leans over and tracks it all the way down." And I said, congratulations, your daughter is running a research program.
That's such a great reframe. She's not a tiny vandal, she's a tiny physicist. Let's start with the cognitive machinery. What is Ezra actually learning when he drops that piece of banana?
He's running a physics lab. Around eight to twelve months, babies enter what Piaget called the tertiary circular reactions sub-stage — which is a clunky name for something remarkable. They start intentionally varying their actions to see what happens. It's not random flailing anymore, and it's not just repeating something that accidentally produced an interesting result. They're systematically experimenting. Think of it like this: an adult scientist doesn't just drop one ball and say "gravity, check, done." She drops balls of different weights, from different heights, onto different surfaces. She's mapping the boundaries of the phenomenon. That's exactly what Ezra is doing, except his lab bench is a high chair tray.
So the first time the banana goes over the side, maybe it's an accident. But then he thinks... wait. What just happened?
And the next time, he's watching. He drops the banana and tracks it all the way to the floor. Then he tries the spoon. Does the spoon fall the same way? Does it make a different sound? Then the cup. Then the banana again but from higher up. He's varying one thing at a time — the object, the height, the angle — and cataloguing the results. And here's what's wild: he's not just learning that things fall. He's learning that all things fall, regardless of what they are. A banana and a spoon and a cup all obey the same rule. That's a massive cognitive leap — abstracting a universal principle from specific instances.
It's the scientific method in its purest form, before anyone taught him what a hypothesis is.
Form hypothesis, test, observe, revise, repeat. And the thing that strikes me about Hannah's description is that he's hungry and still throwing. That feels like the puzzle at the center of this.
It is, and it's where most parents get stuck. Because we think: if he's hungry, he should eat. But eating is a low-stimulation activity. You're sitting there, you're chewing, it's the same texture you've experienced before. Throwing is high-stimulation — it's a full sensory event. Sound, trajectory, splat pattern, parental reaction. For a brain that's wired to learn as fast as possible during this window, the learning drive temporarily overrides the hunger drive. Think of it like an adult who's deeply absorbed in a project and forgets to eat lunch. The hunger is there, but the cognitive engagement is more compelling in the moment.
So the hunger is real, but the developmental urgency of "figure out how gravity works" is more pressing in that moment. It's not that he's not hungry — it's that his brain has ranked "complete the physics experiment" above "consume calories."
There's research from the UC Berkeley Infant Cognition Lab — this was twenty nineteen — where they gave twelve-to-fourteen-month-olds a choice between dropping an object and receiving a new toy. The babies chose the dropping. The action itself was intrinsically rewarding. They weren't doing it to get something else; the experiment was the point. The researchers basically said, "Here's a novel, interesting toy as a reward for participating." And the babies were like, "No thanks, I'm busy with gravity."
That's a good study. "Here's a new toy." "No thanks, I'm busy with gravity." It really drives home that this isn't about getting something — the dropping IS the thing.
And this is different from what younger babies do. At six to nine months, babies drop things to test object permanence — does the thing still exist when I can't see it? By twelve months, object permanence is fully established. Ezra knows the food still exists on the floor. He's not checking whether it vanished. He's testing variables. It's the difference between "does the thing disappear when I let go?" and "what happens to the thing when I let go in different ways?"
The distinction being: younger baby drops the spoon and looks surprised every time, older baby drops the spoon and looks... satisfied. Confirmed. Hypothesis holds.
The gravity game peaks between ten and fifteen months, and babies will drop objects hundreds of times during this window. Food is simply what's available at mealtime. If you gave him blocks during meals, he'd throw blocks. It's not about the food. And this is actually an important point for Hannah's sanity — Ezra doesn't hate her cooking. He's not rejecting the meal she prepared. The food is just the experimental material that happens to be in front of him.
But the food is what Hannah needs to get into him. So there's a practical problem layered on top of the developmental one. And I can imagine how maddening that is — you spend time preparing something nutritious, you're worried about whether he's eating enough, and he's treating it like laboratory supplies.
Right. And that's where we need to talk about the other two things happening simultaneously. Because it's not just physics experiments. If it were only physics, you could just give him blocks and be done with it. But there are two other developmental drivers at work here, and they're both social.
Okay, so we've got the cognitive piece — he's a tiny researcher. What else?
Communication. At twelve to fifteen months, babies are right on the edge of language. They understand more words than they can produce, and they're frustrated by the gap. Throwing food becomes a gesture. It can mean "I'm done," "I don't want that one," or simply "look at me." And here's the thing about the picking-up response — it's a powerful social interaction. It's not just cleaning. It's a whole back-and-forth routine.
So Hannah bending down to retrieve the banana isn't just retrieving the banana. It's a whole conversation. It's "I throw, you fetch, I throw again." That's a game with rules and roles.
Research shows that babies as young as ten months will repeat actions that reliably produce a parental response, and the valence of the response — positive or negative — barely matters. If every time I drop the food, Mom looks at me, talks to me, moves toward me, reaches down... I have just controlled another human being's behavior with a single action. That's intoxicating for a one-year-old. They spend so much of their day being passive — being picked up, being put down, being fed, being dressed. The moment they discover they can make a grown adult do something on command? That's a revelation.
The attention economy of mealtime. And I think most parents don't realize that even negative attention is attention. The exasperated sigh, the "Ezra, please stop" — that's still engagement.
It absolutely is. And the third piece is autonomy. Around twelve to fifteen months, babies discover they can cause specific, predictable reactions in caregivers. They have very little control over their world — someone else decides when they eat, when they sleep, where they go. Throwing food is one of the few actions where the baby has a hundred percent control over the parent's behavior. The parent must respond. You can't not pick it up, or not react, or not notice. The baby has initiated something and you're now part of it. It's one of the earliest experiences of agency they have.
So we've got three drivers layered on top of each other. Experimentation, communication, and autonomy. And Hannah's dealing with all three at once while just trying to get some banana into her kid. That's a lot to navigate during what's supposed to be a simple meal.
And that's why the response has to address all three, not just one. If you only address the experimentation by giving him more things to drop, but you're still providing a huge social reward every time he throws food, the behavior continues. If you go cold and unresponsive but don't give him an outlet for the dropping urge, he'll find another way to run the experiment — maybe by throwing things from his stroller, or off the changing table. You have to close all the loops simultaneously.
So let's get practical. What do Hannah and Daniel actually do at the next meal? Walk me through it.
Strategy one is what I'd call the no-reaction pickup. You pick up the food without eye contact, without a single word, and you place it out of reach — not back on the tray. You're a robot. The food vanished from the floor and nothing interesting happened. This removes the social reward entirely. The physical act of picking up isn't the reinforcer; it's the eye contact, the sigh, the "Ezra, please eat your banana," the whole parental performance.
The parental performance. That's... uncomfortably accurate. I can see myself doing exactly that — the whole theatrical production of "oh no, not again, look at this mess."
I've been there. We all have. And the thing is, from the baby's perspective, that performance is riveting. It's better than television. Strategy two is the one-chance rule. You offer the food once. If it gets thrown, the meal is over. Not in a punitive way — you're not angry, you're not storming off. You just calmly remove the food and wait twenty to thirty minutes before offering again. What this teaches is that throwing equals the end of the eating opportunity, not the beginning of a game. The message is: food is for eating. If you're not eating, we're done here.
And the hungry baby who just lost his meal... does he figure this out? Because my instinct as a parent would be panic — oh no, he didn't eat, he's going to starve, I need to try again immediately.
Surprisingly fast. Most parents who try this report a dramatic drop in throwing within three to five days. The key is consistency. If throwing sometimes ends the meal and sometimes gets a laugh because it was a particularly good throw, the baby learns that the outcome is variable — and variable rewards are the stickiest kind. That's literally how slot machines work. You don't win every time, but you win often enough to keep pulling the lever.
So intermittent reinforcement is the enemy here. The occasional chuckle at a well-aimed carrot stick is actually making the behavior harder to extinguish than if you were consistently stern.
It's always the enemy in behavior extinction. And parents are human — sometimes that carrot stick lands in a funny way and you can't help but laugh. But that laugh buys you another week of throwing. Strategy three is where you address the experimentation drive directly. You increase safe dropping opportunities outside mealtime. Stack blocks and let him knock them down. Drop balls into containers. Play gravity games with intent, where you're sitting with him and you're both dropping things and watching what happens. You're draining the experiment battery so that when mealtime arrives, the urge has already been satisfied.
This is the "take them to the park so they'll sit still at dinner" principle, applied to physics. Give them an appropriate outlet for the drive, and the drive doesn't need to express itself inappropriately.
And it works. Give him a dedicated dropping activity right before meals. Fill a container with soft objects and let him dump it out, fill it, dump it. Ten minutes of that, then straight into the high chair. You've scratched the itch. The experiment battery, as you put it, is partially drained.
What about the nutritional side? Hannah's worried he's not getting enough food. And I think that's the anxiety that drives a lot of the counterproductive responses — the fear that if you end the meal, you're depriving your child of nutrition.
This is where meal structure matters. In the first five minutes of the meal, the baby is hungriest and least likely to throw. That's your window for the high-calorie, nutrient-dense foods — the avocado, the full-fat yogurt, the eggs. Get the good stuff in early. Then, when the throwing phase of the meal begins, you switch to lower-stakes foods. Things where if half of it ends up on the floor, it's not a nutritional crisis. Think Cheerios, soft cooked peas, small pieces of toast. The stuff that's nice to have but not essential.
Front-load the calories, back-load the experiments. So even if the last two-thirds of the meal end up on the floor, the important nutrition already happened.
Trust that a hungry baby will eat. If Ezra is truly hungry, and the food is offered without the game component, he will eat. The throwing phase is not about rejecting nutrition — it's about prioritizing learning. When the learning opportunity is removed, the hunger reasserts itself. I've seen this happen in real time. Parents remove the social reinforcement, and within a few days the baby sits down and just... eats. Because the food is no longer a prop in a game.
How long does this phase typically last? Because I think knowing there's an endpoint helps parents endure the chaos.
It peaks between twelve and fifteen months and usually resolves by eighteen months. The resolution coincides with language development — suddenly the baby can say "all done" or shake his head, and the need to communicate by throwing drops away. If it persists past twenty-four months with the same intensity, that's worth discussing with a pediatrician, because it may indicate something else going on. But for the vast majority of kids, this is a self-limiting phase.
Ezra's right in the thick of it at thirteen months. This is the peak.
This is the peak. And knowing that helps, I think. It's not a permanent behavior problem — it's a developmental phase with a predictable arc. You're not raising a child who will be throwing food in restaurants at age twelve. This is a very specific window.
I want to go back to something you said about the difference between throwing to experiment and throwing to communicate. How would Hannah tell which one Ezra is doing in a given moment? Because I imagine the response might need to be slightly different depending on what's driving it.
The experimenter varies the objects and watches the trajectory. He's not looking at you — he's looking at the falling food. He's tracking the arc, watching the landing, maybe leaning over to see the splat pattern. The communicator looks at you before or during the throw. He's checking whether you're watching. And if the food is picked up and not returned, the experimenter moves on to something else, while the communicator might cry or fuss — because the point was the interaction, and the interaction didn't happen.
If Ezra is staring at the banana as it arcs toward the floor, he's doing science. If he's staring at Hannah while he does it, he's doing... social science.
That's actually a perfect way to put it. And most babies cycle between both modes within a single meal. It starts as experimentation, then they notice the parental response, and it shifts into a social game. The first throw is physics, the third throw is sociology. They're learning that the same action can serve different functions depending on context.
Which is why the no-reaction pickup addresses both. Whether he's looking at the food or looking at you, nothing interesting happens either way. The physics experiment yields no new data because the food just disappears, and the social experiment yields no engagement because you're a robot.
Right. You've closed both loops. The food doesn't return to the tray for further testing, and you don't provide the social feedback he's checking for. It's elegantly simple when you understand what's actually happening.
Let me pull all of this together into some concrete takeaways. Because Hannah asked a practical question and she deserves a practical answer.
Go for it.
First, this is not defiance or misbehavior. It's a developmentally appropriate science experiment combined with emerging communication and autonomy. Reframing it that way actually changes how you respond — you're not disciplining, you're redirecting a learning drive. And I think that reframe alone reduces parental frustration significantly.
Second, the combination of the no-reaction pickup and the one-chance rule addresses both the experimentation and the social reinforcement. Pick up without engagement, place out of reach, and if the throwing continues, end the meal calmly and try again in twenty to thirty minutes. No drama, no lecture, just a calm removal of the opportunity.
Third, satisfy the dropping urge outside mealtime. Give Ezra dedicated gravity-play activities where he can drop, dump, and knock things over to his heart's content. Drain that experiment battery before he gets in the high chair. Ten minutes of dropping play right before a meal can make a noticeable difference.
Fourth, trust that a hungry baby will eat. Front-load the nutrient-dense foods in the first five minutes, switch to lower-stakes foods for the throwing window, and know that this phase peaks around now and typically resolves by eighteen months as language kicks in. You're playing the long game.
One thing I'd add — and this is more for Hannah's sanity than Ezra's nutrition — is that the amount of food a one-year-old actually needs is probably less than most parents think. Their growth has slowed dramatically since the first six months. A toddler can thrive on what looks like three crackers and a grape. The portion sizes we imagine are often based on adult proportions, not toddler metabolism.
That's a good clinical point. The portion sizes we imagine toddlers need are often wildly inflated. If Ezra is growing along his curve and has energy, he's getting enough, even if half the meal ends up on the floor. Pediatricians care about the growth chart, not the floor-to-mouth ratio at any given meal.
That's the science of the flying food. But it leaves me with one more question.
What's that?
What happens when the same throwing behavior shows up with toys, not just food? Because the developmental driver is the same, but the stakes are different — nobody's worried about a toy deficit. Does the same approach work, or does it change when nutrition isn't on the line?
Same driver, but the social dynamic shifts because parents are usually less anxious about thrown blocks than thrown avocado. The urgency is lower, which means parents often handle it more naturally — they just stop picking up the blocks, and the behavior fades. Might be worth a follow-up to dig into the toy-throwing dynamic specifically, because there are some interesting differences in how the social reinforcement plays out.
Hannah, if you're listening — Ezra is building a theory of physics, cause and effect, and social dynamics all at once, while you're just trying to get him to eat. Both goals are valid. The phase ends. The banana can be cleaned. And one day you'll be sitting across from him at a restaurant, watching him use a fork properly, and you'll almost miss the days when dinner came with a floor show.
If you have a weird prompt about why your baby does something inexplicable — or anything else, really — send it to prompts at my weird prompts dot com. We turn parenting mysteries into science.
Thanks to our producer Hilbert Flumingtop. This has been My Weird Prompts. We'll be back soon.
