Skip to content
NatureHQ

Claim check

Can birds learn new songs throughout their lives?

Mostly supportedThe core is right; the usual phrasing overstates part of it.

Some do and many do not, and the difference is a real distinction in the field. A zebra finch learns once and sings the same song for life; a canary rebuilds its repertoire every year; a hummingbird at a lek keeps adjusting to match its neighbours.

The claim as it circulates

“Birds keep picking up new songs as they go along.”

Where you may have met it: Popular writing about mimics such as starlings and mockingbirds; Discussions of pet birds acquiring new sounds

Taking it apart

What was claimed
That birds are lifelong learners of song, continuing to acquire new material as adults.
What was actually observed
Species divide into closed-ended learners, which crystallise a song in their first year and do not change it, and open-ended learners, which continue to modify or add song as adults. Canaries and starlings are open-ended; zebra finches are the standard closed-ended model. Marked long-billed hermits — a lekking hummingbird — were followed across years and were found to change their songs in ways that tracked shifts in the locally shared song type rather than drifting idiosyncratically. Separately, adult zebra finches given a delayed echo of their own song lost it over weeks and recovered it when normal hearing returned, so even a crystallised song is actively maintained rather than fixed.
What the evidence supports
That lifelong vocal learning is real in a substantial set of species and absent in others, and that the closed/open distinction is one of the more useful divisions in the field. It also supports a subtler point: crystallised does not mean stored — the song is held in place by a continuing comparison, which is why corrupting the feedback destroys it.
What it does not support
It does not support treating lifelong learning as the general condition. Most of what is known about song learning comes from a closed-ended species, which is itself worth knowing, and generalising in either direction from any single species is what gets this question wrong.

How we know

Giving an adult bird a slightly wrong echo of itself

Is an adult bird’s finished song a fixed motor programme, or is something still holding it in place?

Adult male zebra finches whose song had been stable for months were fitted with a system that detected when they sang and played their own song back to them a fraction of a second late. The bird therefore heard itself, but out of step with what it was doing — the error signal it normally uses to check its output was made systematically wrong rather than absent. Song was recorded throughout, and again after the system was switched off.

What happened

Over weeks, the song came apart — syllables were dropped, added and distorted. When normal hearing was restored, the birds largely reassembled their original song.

What it shows

A crystallised song is not stored and replayed; it is actively held in place by a continuing comparison between what the bird produces and a target it still carries. Deafening leaves that comparison silent, which is survivable. Feeding it a wrong answer is far more destructive. The recovery is the striking half: after months of singing something else, the target was still there.

What it does not show

Delayed playback has no natural equivalent, and what the bird treats as the mismatch is not measured directly. Recovery was substantial rather than complete in every bird. One closed-ended species, and open-ended learners plausibly maintain song by a different arrangement altogether.

The controls — what makes this evidence rather than a story
  • Adults with fully stabilised song, so any change is a change to something that had stopped changing.
  • The perturbation distorts feedback rather than removing it, which is what distinguishes this from a deafening experiment.
  • Recovery measured after the perturbation ended, testing whether the original target survived.

From Decrystallization of adult birdsong by perturbation of auditory feedback

Related

The claims underneath

Each one carries its own evidence, scope and caveats. Expand any of them to reach the studies.

Hummingbirds are the third group of birds that learn their sounds — and at least one species carries on changing its song throughout adult life to match its neighbours.

Well supported

Good evidence backs this, though some details remain open.

Long-billed hermits at leks modify song after the first year in ways that track changes in the locally shared song type, indicating open-ended vocal production learning in a trochilid.

Who this applies to
One lekking hummingbird; the family is large and vocal behaviour varies widely across it.
Studied in
Phaethornis longirostris
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

Marked individuals followed across years, with changes tracking the neighbourhood rather than drifting idiosyncratically. Observational, and one species.

Caveats

  • Observational — no tutoring, isolation or removal experiment has been run in this species.
  • Marked-individual field studies give modest sample sizes.
  • Lekking, with stable acoustic neighbours across years, is an unusual social arrangement.

Still unanswered

  • How widespread open-ended learning is across the hummingbird family, most of which has never been examined.

Last reviewed 2026-09-02

The evidence (2 studies)

Hearing the tutor is not enough. A young bird also has to hear its own attempts, and one that cannot ends up singing as badly as a bird that never heard anything.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

Song development requires auditory feedback during the sensorimotor phase. Birds deafened after tutor exposure but before song production develop severely abnormal song, whereas birds deafened after crystallisation retain their existing song — though adult song is subsequently shown to be actively maintained by feedback rather than fully independent of it.

Who this applies to
Songbirds, from deafening and feedback-perturbation experiments in a small number of laboratory species.
Studied in
Passeri
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

Two independent manipulations — removing feedback and corrupting it — converge on the same conclusion, and the second produces a reversible effect, which is a stronger form of evidence than the first.

How far it can be extended

Demonstrated in white-crowned sparrows and zebra finches by different methods, and consistent with the shared organisation of the song-control pathway across songbirds.

Caveats

  • Deafening removes all hearing, not only the sound of the bird’s own voice.
  • The degree to which crystallised song becomes feedback-independent varies, and in open-ended learners it does not fully happen.
  • What the bird treats as a mismatch is inferred from behaviour, not measured.

Still unanswered

  • Whether the target a bird compares itself against is the memorised tutor song itself or a transformed representation of it.

Last reviewed 2026-09-02

The evidence (3 studies)

Go deeper

Vocal learning

Learning what a sound means is common. Learning to make a sound you were not born able to make is rare enough to list — and humans are on a list with parrots and whales.

Related things people ask

Last reviewed 2026-09-02