It's 2 a.m., and you're doing the thing every parent does: one finger hooked gently inside your baby's lower lip, feeling along the gum for a ridge. She's been up three times tonight. She was up four times last night. There's a little extra drool, maybe, and she chewed on her sleeve at dinner. It must be teeth.
Teething is the great universal explanation of infant sleep. It requires no evidence beyond a bad night and a mouth, and it comes with a built-in expiration date — once the tooth is through, this will end — which makes it one of the most comforting stories a tired parent can tell. The trouble is that the research on teething tells a smaller, quieter story. Teeth do disturb sleep. But they disturb it less than we believe, for far fewer nights than we credit, and the gap between the story and the evidence has a real cost: every night we blame on a tooth is a night we don't look at what's actually going on.
The explanation that fits every bad night
A baby cuts twenty primary teeth, typically starting around six months and finishing with the second molars somewhere past the second birthday. Those eruptions don't arrive on a tidy schedule — they scatter across roughly two years in loose clusters. Which means that at almost any point in your baby's first twenty-four months, a tooth is plausibly "coming."
This is the perfect setup for what psychologists call confirmation bias. When a rough patch of sleep happens to line up with a visible new tooth, the memory sticks: see, it was teething. When a rough patch resolves and no tooth appears, nobody records the miss. Over months, parents accumulate a private dataset made almost entirely of hits, because the misses were never counted.
We come by this honestly. The belief that teething causes serious harm is centuries old — in Victorian England, "teething" was routinely registered as an official cause of infant death, standing in for fevers and illnesses no one could yet diagnose. The modern version is gentler but structurally identical: teething as the label we reach for when a baby is struggling and we don't know why.
What happens when researchers actually count
The most useful teething studies are prospective — they don't ask parents to remember backward from a bad week, they log symptoms forward, every day, and then check which days actually fell near an eruption. The best known of these was run at the Cleveland Clinic and published in Pediatrics in 2000: researchers followed well over a hundred infants and toddlers for months, with daily symptom records and regular checks of the gums, so eruption dates were pinned down by examination rather than parental guess.
What they found reshapes the story. Some symptoms — drooling, gum-rubbing, biting, mild irritability, and yes, disturbed sleep — were modestly more common in a narrow window around eruption, roughly the few days before a tooth surfaced through the day or two after. But no single symptom showed up in even a majority of children, no combination of symptoms reliably predicted that a tooth was coming, and the severe things folklore attributes to teething — high fever, diarrhea, inconsolable illness — were not linked to eruption at all. Later prospective work has found, at most, a slight bump in body temperature right around the day a tooth breaks through. Not a true fever. A baby with a genuinely high temperature isn't teething; she's sick, and deserves to be checked.
So the honest summary is this: teething is real, and it can genuinely cost you a few rough nights per tooth. What it cannot do is explain three weeks of fractured sleep. Teeth move too fast for that.
Why the timing is so misleading
Here's the deeper reason the teething story survives contact with the evidence: the teething years and the sleep-renovation years are the same years.
Between six months and two years, a baby's sleep is being structurally rebuilt on an almost rolling basis. Naps consolidate and drop — three to two somewhere around eight months, two to one early in the second year — and each transition scrambles nights for a while, because daytime sleep pressure and bedtime are suddenly misaligned. Motor milestones arrive in the same window, and babies notoriously practice them in the crib at 2 a.m., pulling to stand in the dark with the dedication of an athlete. Object permanence matures around eight to ten months, and with it comes separation protest: a baby who now knows you still exist on the other side of the door, and objects to the arrangement. Underneath all of it, the circadian system is still consolidating night sleep into one long block.
Every one of these produces night waking that, from the hallway, is indistinguishable from a sore gum. And because a tooth is nearly always somewhere in transit during these months, the eruption and the disruption keep coinciding — the tooth arriving on the scene of a crime it didn't commit.
How to tell teeth from timing
Genuine teething discomfort has a recognizable signature. It's worst in the days just before the tooth surfaces, when the gum is inflamed and taut, and it eases quickly once the edge is through. It shows up in daylight too — a baby who is drooling heavily, gnawing on everything, rubbing her gums, and fussier than usual while awake. And it's brief: think a handful of hard days, not a month.
Schedule problems wear different clothes. Waking that recurs at a consistent clock time, night after night, points to a body-clock or sleep-pressure issue, not pain. Waking that follows short naps or a too-late bedtime, and worsens as the week compounds, points to timing drift. A baby who is wide awake and cheerful at 2 a.m. isn't hurting — that's a split night, a sleep-pressure accounting error. And if the waking outlasts the tooth — the ridge is visibly through, yet nights haven't recovered — the tooth was never the main story. Somewhere in those hard nights, the schedule slipped: a nap moved late, bedtime drifted, a new habit took root. The tooth healed. The timing didn't.
What actually helps on the hard nights
For the real thing, the useful tools are humble. Cold and pressure both quiet the inflamed gum — a chilled (not frozen) teether or a clean, cold washcloth to chew works as well as anything sold for the purpose. Firm gum massage with a clean finger helps for the same reason. For a genuinely miserable night, appropriate pain relief dosed with your pediatrician's guidance is reasonable. Skip amber teething necklaces, which have no plausible mechanism and a well-documented strangulation risk, and skip benzocaine numbing gels, which regulators have explicitly warned against using in infants.
And then do the thing that feels hardest: hold the schedule steady. Comfort your baby warmly at night — a teething night is not the night for new sleep-training resolve — but keep bedtime, wake time, and nap timing where they were. The fastest way to turn a three-day tooth into a three-week problem is to let the whole rhythm of the day dissolve while you're busy soothing, so that when the gum quiets down, the schedule underneath it no longer fits.
Knowing what the tooth can and can't explain
This is, at bottom, a problem of attribution — and attribution is exactly where exhausted parents at 2 a.m. are weakest, and where a running record is strongest. Drowsy learns your baby's actual sleep-pressure rhythm and predicts the next realistic nap and bedtime window, which quietly gives you the thing the teething story can't: a baseline. When a rough patch hits, you can see whether the timing was already drifting before the drool started — or whether sleep was landing right in its windows and this really is just a tooth, three hard days, soon over. Either answer helps, because each one asks something different of you. If you'd like that baseline doing its counting while you do the comforting, Drowsy is at drowsy.lumenlabs.works.