It happens the same way most nights. The last feed of the day, the lights low, and somewhere between the third and fourth minute your baby's jaw slows, the sucking turns fluttery, and the eyelids give up. You sit very still. You count to a hundred. You attempt the transfer — that slow-motion lowering into the crib that every parent performs like bomb disposal — and either it works, or the eyes snap open and you start again.
And at some point, usually around a growth spurt or a rough week of night wakings, someone tells you the whole arrangement is a mistake. You're creating a bad habit. They'll never learn to fall asleep on their own. You'll be feeding to sleep until kindergarten.
It's worth slowing down here, because the science underneath this warning is more interesting — and more forgiving — than the warning itself. Feeding to sleep isn't a moral failing or a trap. It's a sleep-onset association, and understanding what that term actually means will tell you, far better than any relative or forum thread, whether yours is a problem worth solving.
What a Sleep-Onset Association Actually Is
A sleep-onset association is simply the set of conditions present when a person falls asleep. You have them too. Maybe you sleep on your left side, with a particular pillow, in the dark, with the door closed. Those conditions become part of how your brain recognizes that sleep is available.
Here's the part that matters: sleep is not one continuous event. Everyone — adults and babies alike — cycles through lighter and deeper stages all night, surfacing briefly between cycles. In those brief arousals, the brain runs a quick environmental check. If everything matches the conditions under which you fell asleep, you sink back down without ever forming a memory of waking. If something has changed — your pillow fell off the bed, the light is on, the sound stopped — you're more likely to wake fully and notice.
For a baby who fell asleep feeding, in arms, warm against a body, the middle-of-the-night check can come back with a startling result: everything is different. No milk, no arms, no warmth. Alone in a crib. The baby isn't waking up more than other babies. The baby is waking up the same amount and finding the world rearranged.
Every Baby Wakes at Night — the Question Is What Happens Next
This distinction was made vivid by the sleep researcher Thomas Anders, who spent years filming infants overnight with time-lapse video in their own cribs. What the footage showed upended the way parents talk about "sleeping through the night." Essentially all of the babies woke during the night — briefly surfacing between sleep cycles, exactly as sleep architecture predicts. The difference was what happened next. Some babies, whom Anders called self-soothers, stirred, maybe looked around, and drifted back down on their own. Others, the signalers, cried out for a parent.
So "sleeping through the night" was never really about sleeping through the night. It was about what a baby does at the wakings that were always going to happen. And one of the strongest predictors of which group a baby lands in is whether the conditions at those 2 a.m. surfacings match the conditions at 7 p.m. sleep onset.
This reframe matters because it takes the question out of the realm of habit and character — is my baby spoiled, am I too soft — and puts it where it belongs: in the mechanics of association and arousal.
Why Feeding Is So Powerfully Sleep-Inducing
It's also worth saying plainly: feeding to sleep works because it is supposed to work. This is not a hack you invented; it's biology operating as designed.
Suckling and the arrival of milk in the gut trigger the release of cholecystokinin, a hormone that promotes satiety and drowsiness in infants — one reason babies so reliably go glassy-eyed mid-feed. The warmth, the containment, the heartbeat, the rhythmic sucking all engage the same calming systems that make contact naps and babywearing so effective. And breast milk itself changes across the day: evening and night milk carries more melatonin and more tryptophan (a precursor the body uses on the way to making melatonin) than morning milk, a phenomenon researchers call chrononutrition. Milk at bedtime is, quite literally, formulated differently than milk at breakfast.
In other words, an infant falling asleep while feeding is one of the most ancient, over-determined outcomes in human caregiving. The guilt attached to it is very new. The mechanism is very old.
When Feeding to Sleep Is a Problem — and When It Isn't
Here is the honest answer to the question in the title: feeding to sleep is a problem when it creates night wakings your family can't sustain, and it isn't when it doesn't.
Plenty of babies feed to sleep at bedtime and then resettle themselves at night just fine. If that's your house, there is nothing to fix. Pediatric sleep medicine doesn't classify a sleep-onset association as a disorder because it exists; the clinical concern arises only when the association produces frequent, prolonged wakings that require an adult to recreate the original conditions — a feed, a rock, a rescue — at every surfacing, night after night.
The practical test isn't ideological, it's arithmetic. If your baby feeds to sleep and wakes once to eat, that may simply be a baby who needs a night feed. If your baby needs to be fed back down at every sleep-cycle boundary — every 45 minutes to two hours, all night — the feeding is no longer about hunger. It has become the password to sleep, and the password expires every cycle.
Loosening the Association Without Breaking It
If you land in the second camp and want change, the good news is that associations respond to gradual revision, not just abrupt removal.
The most common gentle approach is sequence-shuffling: move the feed earlier in the bedtime routine, so it's followed by something else — a book, a song, the sleep sack — before the crib. The feed still happens; it just stops being the final step. The goal is for the last thing your baby experiences before sleep to be something that will still be true at 2 a.m.: the crib, the dark, the white noise.
A smaller intermediate step is interrupting the drift: when the sucking turns fluttery and the eyes are closing, unlatch or remove the bottle, so the final descent into sleep happens without milk in the mouth. Babies often protest for a few seconds, then finish falling asleep. Over days and weeks, the association loosens — the baby learns to cross the border into sleep without a feed as the escort, which is precisely the skill they need at night.
None of this has to happen on anyone else's timeline. Associations aren't cement. They're paths worn by repetition, and repetition can wear new ones.
Timing Is the Quiet Variable
There's one more piece, and it's the one most advice skips: changing a sleep association is dramatically easier when sleep pressure is on your side. A baby put down at the exact moment their body is ready for sleep will tolerate a revised routine, a missing feed, a new final step — because the biological pull toward sleep is doing most of the work. The same baby put down twenty minutes early has no such pull, and every prop you remove leaves a gap that protest rushes in to fill. Undertired babies fight change; well-timed babies barely notice it.
That's the quiet logic behind Drowsy. It uses AI to learn your baby's individual rhythm and predicts the precise window when sleep pressure peaks — so when you're ready to move the feed earlier, or attempt the crib without the full escort service, you're doing it at the moment your baby's biology is most cooperative. You bring the patience; it brings the timing. If you're renegotiating bedtime and want the odds tilted in your favor, you can try it at drowsy.lumenlabs.works.