Daniel sent us this one about Ezra — who's now eleven months old — and a new habit that's driving him and Hannah up the wall. Every five seconds or so, Ezra lets out this brief, piercing scream. Not crying, not distressed, just... He says they've gathered this is actually pretty common at this age, and he's wondering if it's connected to getting ready to talk, maybe trying out the vocal cords. So the question is: what's with the random screaming thing?
The short answer is — it's vocal practice. But the mechanism behind it is genuinely fascinating, and most parents never get it explained properly. They just get told "oh, it's a phase," which is true but completely unsatisfying.
"It's a phase" is the parenting equivalent of "thoughts and prayers.
So let's actually unpack what's happening. First, we need to define the behavior precisely, because not all screams are created equal. What Ezra's doing — and what about seventy percent of infants between ten and fourteen months do, according to a twenty twenty-one longitudinal study in Infant Behavior and Development — is a very specific acoustic event. It's brief, typically half a second to one and a half seconds. The fundamental frequency sits around four hundred to six hundred hertz, with harmonics that extend up to four kilohertz. The onset is abrupt — there's no gradual build-up — and the offset is equally sudden. It's not rhythmic. It happens at irregular intervals, often when the baby is otherwise content or only mildly frustrated.
It's a very particular signal. How do you distinguish it from, say, a pain cry or a colic cry? Because I imagine a lot of parents hear any loud sound from their baby and their brain just goes "red alert, all hands on deck.
That's the problem, right? The parental alarm system isn't calibrated to make these distinctions. It's calibrated to respond to anything loud. But there are three acoustic markers that give the practice scream away. First, the spectral peak. A pain cry has a characteristic pattern where the fundamental frequency drops and then rises sharply — it has what researchers call a "tremolo" quality, this warbling instability. Think of the difference between a steady trumpet note and a note with vibrato. The practice scream is spectrally cleaner. Second, the onset abruptness. Pain cries have a slightly longer rise time because the infant is inhaling first, building up pressure. The practice scream is more like a percussive burst — imagine someone striking a piano key versus pressing it gradually. Pain cries and colic cries tend to be longer and more rhythmic — they come in bouts with predictable patterns. The practice scream is short, isolated, almost staccato.
If you're a parent hearing this at three in the morning, you can actually do a quick acoustic triage in your head. "Was that a half-second burst with a clean tone and an abrupt stop? Okay, probably not a crisis.
And the most reliable tell is what happens immediately after. The practice scream is often followed by a pause where the baby is clearly waiting — listening, even. There's a twenty twenty-two observational study in Infancy that documented this "scream-and-pause" behavior emerging reliably between ten and twelve months. The baby screams, then goes quiet, then looks at the caregiver. That's not distress. That's an experiment. It's the exact same sequence you see when an adult tries out a new piece of equipment — they push the button, then they stop and assess what just happened.
What's actually happening in the brain and the vocal apparatus to produce this? Because that's where this gets interesting.
This is where we go under the hood. At eleven months, there's a convergence of three neurological developments. The first is in the perisylvian cortex — that's the region that includes what will become Broca's area, the speech production center. Around this age, that region is undergoing rapid synaptogenesis. New connections are forming at an extraordinary rate. The infant is gaining volitional control over the laryngeal muscles for the first time.
Volitional meaning intentional, as opposed to reflexive.
Newborn cries are largely reflexive — they're controlled by the brainstem and limbic system. But around ten to eleven months, cortical control starts coming online. The problem is, it's coming online without the inhibitory circuitry that would allow fine control. So you get what neuroscientists call "open-loop" vocal bursts. The baby decides to make a sound, sends a motor command, and the result is this full-throttle, maximal-effort output because there's no modulation yet.
It's like if you gave someone a sports car but the brake pedal isn't connected yet.
That's exactly the analogy. The accelerator works great. The steering, not so much. More specifically, what's developing is the corticobulbar tract — that's the neural pathway that runs from the motor cortex down to the brainstem nuclei that control the muscles of the larynx, pharynx, and soft palate. At six months, that tract is only about twenty percent myelinated. By twelve months, it's at fifty percent. Myelination is what allows signals to travel quickly and precisely. At fifty percent, you've got enough insulation to send a strong signal, but not enough for nuanced control.
I want to pause on that myelination number because it's doing a lot of work here. Can you unpack what's actually happening at the cellular level when a tract goes from twenty to fifty percent myelinated? What changes in terms of signal transmission?
Myelin is the fatty insulation that wraps around axons — the long projections of neurons that carry electrical signals. Without myelin, signals propagate at maybe half a meter per second. With full myelination, they can hit a hundred and twenty meters per second. At fifty percent, you're somewhere in the middle — fast enough to get a signal through reliably, but not fast enough for the kind of precisely timed, coordinated firing patterns that produce nuanced motor control. It's the difference between being able to slam a door and being able to close it quietly. The corticobulbar tract at eleven months can slam the door. It cannot yet close it quietly.
The baby is essentially calibrating the instrument.
And this is where the songbird analogy is so useful. Songbirds go through a phase called subsong — it's this quiet, rambling, highly variable vocalization that juvenile birds produce while they're learning to sing. They're not trying to produce a specific song yet. They're exploring the full acoustic space of their vocal apparatus. They produce sounds at the extremes of their range, they listen to the result, and they gradually refine. The infant scream is the human equivalent of subsong.
The "screech-and-listen" phase.
And the reason they're drawn to these extreme sounds is the acoustic feedback loop. High-amplitude, harmonically rich sounds produce the strongest auditory feedback. When a baby produces a sound that's loud and spectrally complex, they get a very clear signal back through their own auditory system. It's more informative than a quiet coo. They're essentially running a system diagnostic at maximum gain.
Which also explains why they do it more when people are around.
There was a twenty nineteen study in PNAS that showed ten-to-twelve-month-olds produce significantly more non-cry vocalizations when caregivers are present versus when they're alone. It's not just the auditory feedback from their own voice — it's the social feedback. The parent's startle response is incredibly reinforcing.
Let's talk about that startle response, because I think this is where parents get tangled up emotionally.
The human startle reflex is triggered by sounds with a rapid onset and a frequency between about two and five kilohertz. An eleven-month-old's practice scream, with its abrupt attack and harmonics extending to four kilohertz, is practically engineered to trigger that reflex. It's not subtle. The parent's head whips around, eyes widen, posture changes — all within about two hundred milliseconds. For the baby, this is fascinating. They did something, and the big person did something back. That's cause and effect in its purest form.
Parents often interpret this as manipulation. "He's doing it to get a reaction out of me.
Which is a misconception worth addressing directly. At eleven months, theory of mind — the ability to understand that other people have different thoughts and intentions — is not sufficiently developed for intentional manipulation. The baby isn't thinking "if I scream, I can make dad look at me." They're discovering that screaming produces an interesting result, and they repeat it because the result is interesting. It's exploratory, not strategic.
It's the difference between "I wonder what this button does" and "I'm going to push this button to annoy you.
And the button is extremely satisfying to push, because it produces multiple layers of feedback simultaneously — the physical sensation of maximal vocal fold vibration, the loud auditory signal, and the dramatic social response. It's a triple reinforcement loop.
Here's what I think trips parents up. If the baby isn't being strategic, why does it feel so personal? Why does that scream land in a parent's nervous system like an accusation?
That's a really important question, and it gets at something that's more about adult psychology than infant development. The sound triggers your startle reflex before your prefrontal cortex even gets a chance to evaluate what's happening. You're already in a heightened state by the time you're consciously processing the event. And in that heightened state, your brain looks for meaning. It's the same mechanism that makes a car backfire feel like an attack for a split second before you realize what it was. The difference is, the car backfire happens once. The baby screams every five seconds for weeks. You're living in a state of repeated micro-startles, and that wears down your ability to respond thoughtfully.
The parent's emotional reaction is real and physiologically grounded. It's not that they're being oversensitive.
Not at all. They're having a completely normal response to a stimulus that is, as we'll discuss, literally engineered by evolution to be impossible to ignore. The reframing isn't about telling parents they shouldn't feel what they feel. It's about giving their prefrontal cortex enough information to override the startle response once it's happened.
We've got the neurological substrate, the motor calibration, the acoustic feedback, and the social reinforcement. What about the evolutionary angle? Because a sound that reliably triggers a startle response in caregivers didn't evolve by accident.
This is where the signal becomes fascinating from an evolutionary perspective. The acoustic structure of the infant scream — rapid onset, high fundamental frequency, nonlinear phenomena like subharmonics and deterministic chaos — is what biologists call an "honest signal." It's designed to be impossible to ignore. But here's the key: it's not necessarily a distress signal. It's an "I am here and I am alive" signal.
A vitality broadcast.
In ancestral environments, an infant who was quiet for too long was an infant at risk. Periodic loud vocalizations served as a kind of acoustic heartbeat — proof of presence and vigor. The scream's structure exploits the mammalian auditory system's built-in alerting mechanisms. It's the same acoustic niche occupied by alarm calls in other primates. But in human infants, it's been co-opted for a developmental purpose.
The thing that makes it so grating — that piercing quality — is actually the point. It's not a bug.
It's the feature. And it connects directly to language readiness in a way that's not immediately obvious. Around eleven months, infants are entering the canonical babbling stage — that's when they start producing reduplicated syllables like "bababa" and "dadada." That's one vocal register. The scream represents a different register entirely — one that practices sustained phonation, breath control, and maximum vocal output. These are all skills that will eventually be necessary for speech.
It's not that the scream is a direct precursor to words. It's that the scream and the babble are two different practice modes for two different aspects of vocal control.
And this addresses another common misconception — the idea that screaming means the baby is about to start talking. Canonical babbling and gesture use are actually stronger predictors of imminent speech than screaming is. The scream is practicing the hardware. The babble is practicing the software.
That hardware-software distinction is really helpful. Can you give me a concrete example of what the hardware practice is accomplishing that the babble isn't?
Think about what's required to produce a word like "ball." You need to coordinate breath support, vocal fold adduction, and articulatory shaping — tongue, lips, jaw — all in a precisely timed sequence. The scream is practicing the first two pieces in isolation. It's building respiratory strength and laryngeal control without the articulatory complexity. The babble, meanwhile, is practicing the articulatory transitions — moving from one consonant-vowel combination to another — without necessarily demanding the same respiratory power. They're complementary drills. It's like a pianist practicing scales with the left hand and arpeggios with the right before putting them together.
Which brings us to the practical question. What do you actually do about this at three in the morning?
The first and most important intervention is reframing. When parents understand that the scream is vocal practice rather than distress, their anxiety drops significantly. That alone changes the dynamic. If you're not panicking, you're not providing that dramatic startle response that reinforces the behavior.
Step one is essentially cognitive behavioral therapy for the parent.
Step two is what researchers call contingent responding. You want to reward quieter vocalizations and not punish the scream, but simply not reinforce it. When the baby produces a softer sound — a coo, a babble — respond immediately with attention and engagement. When the scream happens, maintain a neutral expression and respond with a quiet, modulated voice. You're essentially demonstrating the vocal range you want to hear.
Modeling the dynamic range. Like showing someone how to use the volume knob instead of just the on-off switch.
That's the musician analogy again, and it's apt. You can actually do structured "vocal play" sessions — five minutes of reciprocal sound-making where you mimic the baby's sounds and then introduce new ones. Start with their sound, then drop to a whisper, then go back to normal volume. You're showing them the full range of what the instrument can do.
If that's not practical because you're losing your mind?
Ear protection is not silly. The scream can reach a hundred and ten decibels — that's above the threshold for hearing damage with prolonged exposure. Musicians' earplugs, the kind that reduce volume without muffling everything, are a completely reasonable intervention during high-scream periods. You're not a bad parent for protecting your hearing.
There's also the "scream-safe" space concept. Designate a time and place where screaming is fine — outside during play, or during a specific part of the daily routine. The behavior gets channeled rather than suppressed.
That's actually more effective than trying to eliminate it entirely, because the developmental drive behind the behavior is real. The baby needs to do this. The goal isn't to stop the vocal exploration — it's to help them do it in a way that doesn't destroy the household's sanity.
I want to dig into the contingent responding piece, because I think some parents hear "don't reinforce the scream" and interpret that as "ignore your baby when they scream," which feels wrong. How do you respond without reinforcing?
This is a crucial distinction. You don't ignore the baby. You respond, but you respond in a specific way. The key is the quality of your response. When the baby screams, you turn toward them, you maintain a neutral but present expression, and you say something in a quiet, calm voice — "I hear you. That's a big sound." You're acknowledging the baby's existence and the communication attempt without providing the dramatic, high-arousal reaction that the scream is designed to elicit. You're essentially saying, "I'm here, and the way to engage me is with a quieter voice." Then, the moment they produce any softer sound — even a grunt or a breathy vocalization — you light up. Big smile, animated voice, full attention. The contrast is what teaches.
You're not withdrawing. You're redirecting the volume of the interaction.
And this works because babies are exquisitely sensitive to social contingency at this age. They notice when their actions produce predictable responses. If screaming consistently gets a calm, quiet response and babbling consistently gets an enthusiastic, engaging response, they'll shift their vocal behavior. Not overnight, but reliably within a couple of weeks.
Let's talk about when this crosses the line from normal to concerning. What are the red flags?
Three things to watch for. First, if the scream is consistently accompanied by signs of pain — arching the back, pulling at the ears, fever — that's not vocal practice. That needs a pediatrician. Second, if the scream replaces all other vocalizations. If a baby who was previously babbling stops babbling entirely and only screams, that's a regression that warrants investigation. Third, if the scream has a different acoustic quality — if it's longer, more rhythmic, has that tremolo quality I mentioned — that could indicate distress.
The "scream diary" method is actually useful here.
Log the time, the context, and the baby's affect for three days. If more than eighty percent of the screams occur in non-distress contexts — baby is fed, dry, comfortable, and the scream is followed by neutral or positive affect — it's almost certainly vocal practice. If you're seeing a different pattern, you have data to bring to the pediatrician rather than just a vague sense that something's off.
The "scream-and-smile" pattern is the clincher, right? Baby screams, then looks at you with a grin.
That's the clearest possible signal that this is a social bid, not distress. A baby in pain doesn't follow their cry with a grin. That's a baby who just discovered they can make a very interesting sound and wants to share the discovery.
Which is actually kind of delightful when you think about it. They've figured out how to do something new and they're showing you.
It's the same impulse that makes a three-year-old say "look at me" fifty times a day. The eleven-month-old just has a cruder tool for getting your attention.
To pull this together into something actionable — we've got a three-part protocol. One, do the scream diary for three days to confirm it's practice, not distress. Two, implement structured vocal play — five minutes of reciprocal sound-making where you model the full volume range. Three, protect your own ears and sanity — musicians' earplugs, scream-safe times and spaces, and contingent responding that rewards quieter sounds.
The contingent responding piece is worth elaborating. The research shows that when parents consistently respond more quickly and more positively to quiet vocalizations than to screams, the proportion of screams in the baby's vocal repertoire drops significantly within about two weeks. It's not suppression. It's shaping. The baby isn't learning not to scream — they're learning that quieter sounds work better for getting social interaction.
Which is actually a more sophisticated lesson. They're learning that volume modulation has social consequences.
That's a useful thing to learn at eleven months, because it's foundational for all the pragmatic aspects of communication later on. Turn-taking, volume regulation, reading your audience — it all starts here.
This random screaming phase isn't random at all. It's a predictable, well-documented developmental milestone that serves multiple functions simultaneously — motor calibration, acoustic exploration, social learning, and the foundations of intentional communication.
The fact that it's so universally reported across cultures, with very similar acoustic profiles, tells us this is deeply biologically driven. A twenty twenty-three study in Developmental Science found that eleven-month-olds produce an average of twelve to eighteen non-cry vocalizations per hour during wakeful periods, with screams comprising about fifteen percent of those sounds. That's consistent across the samples they looked at.
Fifteen percent is not trivial. That's a significant portion of the vocal budget.
And it peaks right around eleven to twelve months, then gradually declines as more modulated vocalizations take over. By eighteen months, most children have moved through it. The scream gets replaced by words, by more controlled vocal play, by the expanding repertoire of intentional communication.
Which raises an interesting open question. Does the intensity or duration of the scream phase predict anything about later language development? Are early, enthusiastic screamers more verbally fluent later on?
The research isn't settled on that. Some studies suggest a weak positive correlation between early vocal exploration — including screaming — and later vocabulary size. But the stronger predictors remain things like joint attention, gesture use, and the amount of contingent verbal interaction with caregivers. The scream phase is probably necessary but not sufficient. Every typically developing child goes through some version of it, but how much they scream doesn't seem to be a strong predictor of how early or how well they'll talk.
Parents shouldn't read too much into it either way. You can't look at a particularly screamy eleven-month-old and think, "Ah, future orator.
It's a milestone, not a metric. Though I will say, anecdotally, there are plenty of parents who swear their extremely vocal infants grew up to be extremely verbal children. Whether that's actual prediction or just a satisfying narrative is hard to untangle.
Let's talk about the technology angle for a moment. Baby monitors are getting increasingly sophisticated with AI-driven cry classification. Some of them claim to distinguish between hunger cries, pain cries, and "other" vocalizations. How well do they actually handle the practice scream?
It's an active area of research and development. The challenge is that the practice scream's acoustic profile sits in a kind of ambiguous space. It has some features that overlap with distress — the high fundamental frequency, the abrupt onset — but lacks others, like the rhythmic patterning and the frequency instability. The better classification systems are starting to incorporate temporal features and spectral flux analysis to make the distinction. But we're not quite at the point where a monitor can reliably tell a parent "this scream is developmental, not distress.
Which is probably fine, because the parent should still check. But it would be nice to get a probability score. "Eighty-seven percent chance this is vocal practice, but hey, maybe go look anyway.
That's the direction the technology is heading. And the training datasets are getting better. The twenty twenty-two Infancy study I mentioned actually contributed to one of the larger acoustic databases that some of these companies are using.
I'm curious about the false positive problem. If a monitor misclassifies a genuine distress cry as a practice scream, that's potentially dangerous. How do developers handle that?
They bias toward sensitivity. It's the same principle as any screening tool — you'd rather have false alarms than missed detections. Most systems are designed to flag anything ambiguous as "check on baby." The practice scream classification is more of a secondary feature — it's not telling you to ignore the sound, it's providing additional context after the fact. "Your baby vocalized twelve times in the last hour. Based on acoustic analysis, eight of those appear to be practice vocalizations." That's useful information for a parent trying to understand patterns over time, even if it's not reliable enough for real-time decision-making.
Before we move to takeaways, I want to zoom out for a second. There's something almost philosophical about this phase. The baby is discovering their own voice — literally discovering that they can make sound intentionally. And the first thing they do with that power is push it to the absolute limit. Maximum volume, maximum intensity. It's like the vocal equivalent of a baby throwing food off the high chair to see what happens.
It's the same exploratory impulse. What are the boundaries of this thing I can do? What happens when I push here? What about here? The scream is just the auditory version of the food-dropping experiment. And both are early science. Hypothesis, test, observe result, repeat.
"If I scream, dad's head turns. Let's run that again.
And the repetition isn't mindless — it's the foundation of learning. They're establishing the reliability of the effect. Does it work every time? Does it work on different people? Does it work in different rooms? They're running a systematic investigation.
The baby is a tiny empiricist.
Always has been. The entire first year of life is one long experiment in cause and effect, and the vocal apparatus is just the latest laboratory. And here's the thing that makes it clever from a developmental standpoint — the baby isn't just learning about sound production. They're learning about social causality. "My action changes your behavior." That's a profound discovery, and it's the foundation of all communication. Before you can use words to request something, you have to understand that your vocalizations can affect other people. The scream phase is where that understanding gets wired in.
In a way, the scream is the first conversation. It's crude, it's loud, but it's bidirectional.
The baby vocalizes, the caregiver responds, the baby observes the response. That's the basic structure of a conversational turn. The content comes later. The structure is being built right now, at eleven months, in these brief, piercing exchanges.
Let's turn this into a practical playbook. Three things parents can do starting tonight.
First, the scream diary. Three days, log every scream — time, context, what happened right before, what happened right after, and the baby's expression. If you see that "scream and smile" pattern or if most screams happen when the baby is otherwise content, you can relax. You're dealing with vocal practice, not distress.
Second, the vocal play routine. Five minutes a day, ideally at the same time so it becomes a predictable ritual. Sit face to face with the baby, mimic whatever sound they make, then introduce a new one — quieter, louder, different pitch. Make it a game. The goal is to show them the full range of what voices can do.
Third, manage your own response. Ear protection if you need it. Contingent responding — quick, engaged attention for quiet sounds, calm and neutral for screams. And designate scream-friendly times and places so the behavior has an outlet.
The key insight behind all three is the same: this is not a problem to be solved. It's a developmental milestone to be navigated. The scream is not a sign that something is wrong. It's a sign that something is coming online.
That reframe alone — from "my baby is screaming at me" to "my baby is practicing their vocal instrument" — is therapeutic for exhausted parents.
I want to add one thing about the ear protection, because I think parents feel guilty about it. The practice scream can hit a hundred and ten decibels. At that level, the National Institute for Occupational Safety and Health recommends a maximum exposure of about thirty seconds per day. If your baby is screaming every five seconds for an hour, you're way past that threshold. Protecting your hearing isn't selfish — it's preventing cumulative damage that could affect your ability to hear your child's actual distress signals later.
That's an excellent point. And musicians' earplugs are designed to reduce volume without distorting the frequency spectrum. You'll still hear the scream. You'll still hear the difference between a practice scream and a pain cry. You're just taking the edge off something that is literally, measurably, dangerously loud.
"Dangerously loud" is not hyperbole here. We're talking about sound pressure levels comparable to a rock concert or a chainsaw. The fact that it comes out of something so small doesn't make it less damaging to your cochlea.
The intermittent nature of it actually makes it worse in some ways, because your acoustic reflex — the tiny muscle in your middle ear that tightens to dampen loud sounds — doesn't have time to engage fully before each scream hits. The acoustic reflex has a latency of about twenty-five to fifty milliseconds. A practice scream with its abrupt onset is already at peak amplitude before the reflex can kick in. So your ear is essentially defenseless for the first part of every single scream.
Earplugs are evidence-based parenting. There's your headline.
There's your headline.
Before we wrap, let's zoom out and ask the bigger question. What does this phase tell us about the trajectory of language development as a whole?
I think it tells us that language isn't a switch that flips. It's a gradual assembly of components — breath control, phonation, articulation, social intention, turn-taking, volume modulation. The scream phase is the point where breath control and phonation come under voluntary control, but before the other pieces are in place. It's a window into the construction process.
It's a reminder that development is messy. It's not a smooth upward curve. It's fits and starts, regressions and breakthroughs, phases that seem pointless until you understand what's being assembled underneath.
The scream is the sound of the scaffolding going up. It's not pretty, but it's necessary.
Now: Hilbert's daily fun fact.
Hilbert: The word "fives" in the sport of Eton fives — a handball variant played at Eton College — derives from the French "fève," meaning bean. This traces back to the 1720s, when French traders in Mali observed local children playing a hand-slapping game with dried beans, and the term migrated through merchant slang into English public-school sporting vocabulary.
...Beans.
From Mali to Eton via a legume. I did not see that coming.
This has been My Weird Prompts. Thanks to Hilbert Flumingtop for producing.
If you want more episodes, visit myweirdprompts.com or find us on Spotify. We'll be back next week.
