It’s 3am and you’re running the checklist. Fed twenty minutes ago. Clean diaper. Burped twice. And the whole time you’re listening hard: is this the hungry cry, the tired one, the something’s-wrong one? You’ve probably been told every cry means something specific, and that a good parent would know which. Scientists checked.
In a study of French and German newborns, the melody of their cries matched their parents’ language: French babies’ cries tended to rise at the end, and German babies’ cries tended to fall. Babies hear the rhythm of your speech from inside the womb, and within days of birth they cry in it.
So if a cry can carry an accent, surely it carries the reason?
The most famous claim says every baby in the first three months makes five sounds that work like words: “neh” means hungry, “owh” means tired, “eh” means a burp is coming, and so on. It was featured on Oprah, sold as DVDs and courses, and lives on in cry-translator apps. A proper clinical study was planned with university researchers, but the company went to market without it.
A team in France recorded 24 babies at home, from 2 weeks to 3½ months old, and collected 39,201 cries. For each cry, parents noted what made it stop: feeding meant hunger, rocking or changing meant discomfort, and being picked up meant the baby wanted company. Then the cries went to a computer program and to trained human listeners.
Neither could tell the reasons apart by sound. But the computer could tell which baby was crying and roughly how old it was. A cry doesn’t say why. It says who.
So if an app tells you your baby is “87% hungry,” it’s guessing, the same as you. It’s just more confident about it.
A pain cry, like the one after a vaccination, is rougher than an ordinary cry. The voice cracks and breaks up. When researchers played pain cries and ordinary cries from unfamiliar babies, people without baby experience mostly couldn’t tell them apart, but parents of young babies could. Recognizing pain is learned, not instinct.
The same goes for picking out your own baby’s cry. A 2013 study found fathers were just as good at it as mothers. What mattered was how much time each parent had spent with the baby.
A newborn doesn’t know it’s “hungry” as an idea. It just knows something is off, and it has one way to say so. A smoke alarm doesn’t tell you “toast” or “the oven.” It says come here, now. A cry carries who’s calling and how bad it is, and finding the cause is your job. Running the checklist at 3am isn’t you failing to understand your baby. It’s the system working as designed.
A 2017 review of crying diaries from nearly 9,000 babies found about two hours of crying a day in the first six weeks, easing to about an hour by three months. Countries varied a lot: colic was reported for about 1 in 3 babies in Canada and 1 in 20 in Denmark.
Among the !Kung (Ju/’hoansi) hunter-gatherers in Botswana, babies are carried most of the day and cries are answered within seconds. They started crying just as often as Dutch babies, but each bout was short, so their total crying was about half.
In a Montreal trial, babies whose parents carried them a few extra hours a day cried 43% less at six weeks. But when the same researchers tried extra carrying with babies who already had colic, it made no difference. Answering quickly shortens everyday crying, but it doesn’t cure colic. Colic usually fades by 3 to 4 months, and it isn’t something you caused.
Could a smarter app crack it someday? Nobody knows. A few small studies say a computer can sort cries by cause, but the biggest real-home test found no “why” in the sound. Today, no app can tell you why your baby is crying.
You don’t have to translate the cry. You only have to answer it.
This is educational content, not medical advice. For anything about your own baby, talk to your pediatrician.