Skip to content
NatureHQ

Birdsmechanism

How birds learn to sing

A young bird holding a perfect memory of its father’s song, unable to hear its own attempts, sings as badly as one that never heard anything at all.

A young songbird listens first and sings later — sometimes months later. It memorises an adult’s song during a window early in life, then practises through a long phase of variable, disorganised sound, listening to itself and closing the gap. Take away either the model or its own voice and the song comes out wrong.

Song learning is the best-understood case of a complex motor skill being acquired by practice anywhere in biology, and the reason is that it can be taken apart. A young bird’s song is a claim about its history — it sounds like this because of something it heard — and history is exactly what watching an animal cannot give you. So the field proceeded by removal: take away the tutor, take away the bird’s own voice, feed it a wrong echo, swap the eggs between nests, and see which part of the song fails. What emerged is a sequence with two distinct halves. In the sensory phase the bird listens and stores; it is not singing at all, and in some species the store sits unused for months. In the sensorimotor phase it produces subsong — rambling, quiet, structurally almost nothing — and then narrows, gradually and syllable by syllable, until the output matches what it memorised. Then, in many species, the song crystallises and is fixed for life. Each half has its own failure mode, and the two look identical from outside: a bird that never heard a model and a bird that heard one but cannot hear itself both end up singing badly. That symmetry is the finding, and it is why hearing yourself turns out to be as much a part of learning as hearing anyone else.

Developed coverage · 80% complete · reviewed 2026-09-02

What this page covers

The developmental sequence described here is a songbird one, worked out in a handful of laboratory and field species — chaffinches, sparrows, canaries and above all the zebra finch. Parrots and hummingbirds learn too, by routes that are less well characterised.

Often confused with: Vocal learning in general, which is the wider capacity across birds and mammals; Human language acquisition, which shares a developmental shape and very little else

Quick facts

Listening comes first
The sensory phase can end months before the bird sings a note
Practice is required
A bird that cannot hear itself never matches the model
The filter is innate
Young birds copy their own species and ignore the rest
Sensitive period
Wider with a live tutor than with a loudspeaker

Listen, then practise, then stop

Three phases, with different things going wrong in each.

Diagram

From silence to a fixed adult song

The lower curve is variability — what changes over development is how much the bird’s output wanders.

How a young songbird gets from silence to a songSensory phaseIt listens, andmemorises. It isnot singing yet.Sensorimotor phaseSubsong, then plasticsong. It practises,hears itself, adjusts.CrystallisedVariability drops.The adult song isset for life.hatchingadultVariability of what the bird producesIt narrows gradually, syllable by syllable — not all at once.
The same explanation in words

Three labelled boxes running left to right along a timeline arrow from hatching to adult. First, the sensory phase: the bird listens and memorises, and is not singing at all yet. Second, the sensorimotor phase: subsong, then plastic song — it practises, hears itself, and adjusts. Third, crystallised: variability collapses and the adult song is set, in many species for life. Below the timeline, a curve labelled "variability of what the bird produces" starts high on the left and falls steadily, flattening out near the right. A note says it narrows gradually and syllable by syllable rather than all at once — different parts of the song reach their targets at different times.

The gap between the two active phases is the part that surprises people. A white-crowned sparrow memorises song in its first summer and may not produce anything resembling it until the following spring. Whatever is stored in between is not being rehearsed; it is simply being kept. That is also what makes the sensitive period measurable at all — the listening and the singing are far enough apart that the two can be manipulated independently.

The rambling, unstructured noise a young bird makes before it can sing is practice, not communication — and it narrows gradually into song, one piece at a time.

Well supported

Good evidence backs this, though some details remain open.

Subsong is characterised by high acoustic variability, weak spectral structure and low amplitude, and is produced in contexts lacking an apparent receiver. Continuous developmental recording shows progressive, syllable-specific reduction of variability towards tutor-song targets rather than uniform improvement of the whole song.

Who this applies to
Measured in detail in the zebra finch; described qualitatively across songbirds.
Studied in
Taeniopygia guttata, Passeri
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

The developmental trajectory is measured precisely in one laboratory species. The general description is well attested; the quantitative claim about syllable-specific convergence rests on far fewer species.

How far it can be extended

A variable early phase preceding stereotyped song is described throughout the songbird literature, though the fine trajectory has been measured in few species.

Caveats

  • That no receiver is apparent does not establish that subsong has no social function.
  • The zebra finch is a closed-ended learner with a short song, and its trajectory may not be representative.
  • Similarity to a tutor song is an acoustic measure, not a statement about what the bird is aiming at.

Still unanswered

  • Whether subsong in species with very large repertoires follows the same syllable-by-syllable pattern.

Last reviewed 2026-09-02

The evidence (2 studies)

How we know

Recording every sound a young bird makes, for months

Does a young bird’s song improve as a whole, or does it get there some other way?

Young male zebra finches were housed with controlled tutor exposure and recorded continuously and automatically from their first vocalisations until their song had settled — not sampled occasionally, but every sound, for months. The resulting archive was analysed with automatic acoustic measures that score how close each produced sound is to the corresponding part of the tutor song, so the whole developmental trajectory could be plotted rather than inferred from before-and-after snapshots.

What happened

Song did not sharpen uniformly. It began as highly variable, unstructured sound, and individual syllables then differentiated out of it at different times and at different rates, each converging separately on its part of the model.

What it shows

Song learning is a gradual reduction of variability, unit by unit, rather than a switch from noise to song. That is what makes the comparison with human babbling more than a metaphor: in both cases the developmental signature is a slow narrowing of what is produced, with different pieces of the eventual repertoire arriving at different times.

What it does not show

The acoustic measure scores distance from the tutor song, which is not necessarily what the bird is aiming at. And the zebra finch is a closed-ended learner with a short, stereotyped song kept in a cage with one model — a wild bird hears many singers and the trajectory may look nothing like this. Nothing here shows that birds and babies do the same thing for the same reason.

The controls — what makes this evidence rather than a story
  • Continuous recording, which removes the sampling bias in comparing "early song" with "late song".
  • A single known tutor song, so the target is defined rather than reconstructed.
  • Automated similarity scoring applied identically across development, rather than a listener’s judgement at each stage.

From Dynamics of the vocal imitation process: how a zebra finch learns its song

Words used here
Subsong
The quiet, rambling, structurally weak sound a young bird makes before it can sing. Highly variable, produced with no obvious audience, and the direct ancestor of the adult song.
Plastic song
The stage between subsong and the finished article: recognisable, still variable, still being adjusted.
Crystallisation
The point at which variability collapses and the song settles into its adult form. In closed-ended learners it does not change again.

Two ways to break it, and they look the same

The pair of experiments that separate memorising a song from being able to produce it.

Diagram

Copying, and where it can go wrong

Schematic. The dashed juvenile traces stand for variability.

Copying, and where it can go wrongTutorJuvenile, practisingCrystallised adultTwo things have to happen, and either can be removed:Take away the tutorThe bird still sings, and itssong has no real structure.Take away its own voiceIt has the memory and noway to aim at it. Just as bad.Schematic. The dashed juvenile traces stand for variability.
The same explanation in words

Three spectrogram panels in a row, joined by arrows. The first, labelled tutor, shows two clean traces: an arch and a zigzag. The second, labelled juvenile practising, shows the same two shapes drawn as thin dashed lines that wobble off the original path — the right idea, badly executed. The third, labelled crystallised adult, shows the two shapes clean again and matching the tutor. Below, two boxes set out the two ways this can fail. Take away the tutor: the bird still sings, and the song has no species-typical structure. Take away its own voice: it holds the memory and has no way to aim at it, and the result is just as bad.

How we know

Taking away a bird’s ability to hear itself

Is hearing the tutor enough, or does a bird also need to hear itself in order to learn the song?

White-crowned sparrows were deafened surgically at three different points in their development. One group lost hearing before they had heard any song at all. A second group lost it after they had heard a tutor, but before they had started to sing themselves — so they carried the memory and could not hear their own attempts. A third group lost it only after their adult song had already settled into its final form. Every bird’s subsequent song was recorded and compared with normal song and with the tutor material it had been given.

What happened

Birds deafened before they sang produced disorganised song, whether or not they had heard a tutor first. Birds deafened after their song had crystallised carried on singing the song they already had.

What it shows

Learning a song is two processes, not one. Hearing the tutor lays down a target. Hearing yourself is how you get your own output onto that target, by trial and comparison — and a bird holding a perfect memory with no way to check its own attempts ends up no better off than one that never heard anything. Once the match is made, the motor pattern runs without needing to be checked.

What it does not show

Deafening removes all hearing, not just the sound of the bird’s own voice, so it cannot separate self-monitoring from every other use of the ears. It does not show that crystallised song is truly independent of feedback either — later work in which adults heard their own song at a delay found that song degraded over weeks, which means it is being actively maintained rather than simply stored. And it is one species; in open-ended learners the settling never fully happens.

The controls — what makes this evidence rather than a story
  • Three timings, which is what turns one operation into an experiment: the same procedure at different developmental moments produces different outcomes.
  • The middle group is the critical one — it holds the memory constant and removes only the feedback.
  • Birds tutored and left hearing, as the normal-development comparison.

From The role of auditory feedback in the control of vocalization in the white-crowned sparrow

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)

For thirty years the reading of that result was that a crystallised song becomes a fixed motor programme — the bird no longer needs to hear itself because the song is simply stored. Then somebody gave adult birds a slightly wrong echo instead of no echo, and the song came apart over weeks. Removing the feedback is survivable; corrupting it is not. A finished song turns out to be actively held in place rather than filed away.

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

Why a bird in a noisy wood copies almost nothing

A young songbird hears dozens of species and learns one. The filter is on the way in.

How we know

Building a tutor song out of two species’ syllables

When a young bird ignores the other species singing around it, is it filtering on the sounds themselves or on how they are arranged?

Hand-raised swamp sparrows were tutored with songs that had been assembled rather than recorded. Syllables were taken from swamp sparrow song and from song sparrow song, and each set was arranged both in the swamp sparrow’s temporal pattern and in the song sparrow’s. That gives four combinations, so the syllable material and the arrangement can be varied independently — a young bird hearing its own species’ syllables in the wrong arrangement, or another species’ syllables in the right one.

What happened

The birds learned their own species’ syllables and largely ignored the other species’, and the arrangement they were embedded in made little difference.

What it shows

The filter is on the acoustic material rather than on the structure. A young songbird arrives with a bias about which sounds are worth memorising, which is why a bird in a wood full of other species copies almost none of them. Vocal learning is not a blank slate — culture chooses among variants that biology has already made eligible.

What it does not show

It does not show that the bias is absolute. Tutoring by a live bird rather than a loudspeaker can override it, and mimics such as starlings and lyrebirds evidently have very little of it. It also does not identify which acoustic features the filter is tuned to, and it involves two closely related sparrows, so how general the specific filter is remains open.

The controls — what makes this evidence rather than a story
  • Crossing syllable type with temporal pattern, so neither can explain a preference on its own.
  • Hand-raised birds with no other acoustic experience, so the tutor tapes are the entire input.
  • Normal songs of both species included, establishing what full copying looks like.

From Selective vocal learning in a sparrow

A young bird in a wood full of singers copies almost none of them: it arrives with a bias about which sounds are worth memorising.

Well supported

Good evidence backs this, though some details remain open.

Song learning in several songbird species is guided by an innate perceptual predisposition favouring conspecific acoustic material. In tutoring experiments with constructed songs, learners select conspecific syllables largely independently of the temporal pattern in which they are presented.

Who this applies to
Demonstrated in a small number of sparrows; the strength of the bias varies greatly across songbirds and is nearly absent in mimics.Do not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Melospiza georgiana, Melospiza melodia, Zonotrichia leucophrys
Why we rate it this way, and what the caveats are
Well supportedHigh confidence

The crossed design — syllable type against temporal pattern — is a strong test, and the result replicates. The confidence attaches to the species tested rather than to songbirds as a whole.

How far it can be extended

Starlings, lyrebirds and mockingbirds copy heterospecific and non-biological sounds extensively, so a strong species filter is a feature of some songbirds rather than of songbirds.

Caveats

  • Tape tutoring understates what a bird will learn; social context weakens the filter substantially.
  • The acoustic features the bias is tuned to have not been isolated.
  • Vocal mimics demonstrate that the filter can be nearly absent within the same clade.

Still unanswered

  • How a predisposition tuned to species-typical sound coexists with mimicry in close relatives.

Last reviewed 2026-09-02

The evidence (3 studies)

The filter is not absolute, and the way that was discovered is instructive. Every published estimate of the sensitive period had been made with a loudspeaker, and when somebody put a live bird in the next cage instead, young sparrows learned at ages where tapes produced nothing — and some of them copied the wrong species. A deadline measured with a recording is partly a fact about recordings. The general lesson goes well beyond birds: an impoverished stimulus makes a learner look more constrained than it is.

How we know

Replacing the loudspeaker with an actual bird

Is the sensitive period for song learning a property of the bird, or a property of the impoverished way it has been tested?

Young sparrows were tutored either with tape recordings — the standard method behind every published estimate of the sensitive period — or by a live bird in an adjacent cage, including in some cases a live bird of a different species. Crucially, tutoring was given both inside the window in which tape tutoring is known to work and well outside it, at ages at which taped song reliably produces nothing.

What happened

Birds with live tutors learned at ages where taped song produced no learning at all, and some of them copied the song of the other species — an outcome that tape tutoring essentially never produces.

What it shows

A good deal of what looked like a hard developmental deadline was an artefact of the stimulus. Put a real animal in the room and the window widens and the species bias weakens. The general lesson reaches well beyond birds: an impoverished stimulus makes a learner look more constrained than it is, and constraints measured that way are upper bounds on plasticity rather than descriptions of it.

What it does not show

It does not identify what a live tutor actually supplies. Attention, a response contingent on what the pupil does, something visible, or simply being more interesting are all live possibilities and this design separates none of them. Nor does it abolish the sensitive period — the window moves and widens, it does not disappear — and cross-species copying under live tutoring remains a minority outcome rather than the normal one.

The controls — what makes this evidence rather than a story
  • Tape tutoring run in parallel as the comparison, so the difference is the tutor rather than the birds or the year.
  • Ages deliberately chosen outside the established window, where the expected result is no learning.
  • A live heterospecific tutor, which tests the species filter and the timing at once.

From Social interaction, sensitive phases and the song template hypothesis in the white-crowned sparrow

Is subsong really like a baby babbling?

Closer than it has any right to be, and not in the way people mean.

The short answer

Do young birds babble the way human babies do?

The developmental resemblance is real and specific: both start with highly variable sound produced with no apparent audience, and both narrow gradually towards a model heard earlier. What they do not share is anything to do with language.

The features that line up are the ones about acquiring a motor skill. A perceptual phase before production; a long variable practice phase; dependence on hearing yourself, which is why a person deafened after learning to speak keeps their speech and a bird deafened after crystallisation keeps its song; and a sensitive period. Continuous recording of young zebra finches showed the variability narrowing syllable by syllable rather than the song improving as a whole, which is what infant babbling does too. Then the differences, which are not details. A human infant is babbling towards an open system of contrasting units used to say new things; a young songbird is converging on one song, and in many species will never change it again. And the two lineages are separated by hundreds of millions of years, so whatever this is, it is convergence rather than inheritance. The comparison is a good one about learning and a bad one about language.

Check it for yourself

Young birds do something genuinely like babbling — variable practice that narrows towards a model heard earlier — and that similarity is about how a motor skill is acquired, not about language.

Well supported

Good evidence backs this, though some details remain open.

Songbird song development and human speech acquisition share several features: a perceptual phase preceding production, a variable practice phase, dependence on auditory feedback, and a sensitive period. They do not share the properties that distinguish language, including reference to absent entities and open-ended combinatorial semantics.

Who this applies to
A comparison between two vocal-learning lineages that are not closely related and did not inherit the trait from a common ancestor.Do not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Passeri, Homo sapiens
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

The developmental parallels are well documented on both sides. What remains genuinely uncertain is how deep the resemblance goes at the level of mechanism, and confident statements in either direction outrun the evidence.

How far it can be extended

The similarity is convergent. Reading it as homology, or as evidence that song is a form of language, is the specific error this claim exists to block.

Caveats

  • Song is a single performance learned once in many species; speech is an open system used to say new things, and no amount of developmental similarity closes that gap.
  • Human infants babble towards a system of contrasting units; a young bird converges on one song.
  • The comparison is convergent rather than homologous — the two lineages are separated by hundreds of millions of years.

Still unanswered

  • Whether the shared developmental structure reflects a shared computational solution or two different solutions that look alike from outside.

Last reviewed 2026-09-02

The evidence (3 studies)

What this looks like outside a laboratory

In a cage a young bird is a recording device. In a wood it is choosing who to sound like.

Almost everything above was established by taking things away, which is the right way to find out what is necessary and a poor way to find out what normally happens. Field studies give a different picture: young birds often learn after settling on a territory, from the specific neighbours they will spend their lives disputing boundaries with, rather than absorbing whatever they happened to hear as nestlings. That reframes dialects too — sharing song with your neighbours is not an accident of living near them, it is plausibly the point.

And a consequence: the song carries a record of how the bird’s first weeks went.

A bird that was underfed as a nestling sings a measurably poorer song as an adult, and cannot make it up later.

Well supported

Good evidence backs this, though some details remain open.

Nutritional stress during the sensitive period for song learning produces degraded song copying in adulthood and, in several species, reduced volume of song-control nuclei. The effect persists after the stressor is removed, which is the basis of the developmental stress hypothesis for song as an honest signal.

Who this applies to
Demonstrated in several songbird species by nutritional manipulation during development.
Studied in
Passeri
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

The effect of early stress on song is well replicated. The further step — that receivers use song to assess developmental history — is much less well evidenced, and the popular version takes it for granted.

How far it can be extended

Independent nutritional-restriction experiments in different songbird species produce the same direction of effect on song quality.

Caveats

  • Evidence that stress degrades song is stronger than evidence that females attend to the difference.
  • Effect size and even direction vary between species and between the specific song features measured.
  • Laboratory nutritional restriction is a blunt stand-in for the range of stressors a wild nestling meets.

Still unanswered

  • Whether receivers discriminate on the features that early stress actually affects, rather than on features that happen to be easy to measure.

Last reviewed 2026-09-02

The evidence (2 studies)

Where this leads

What is still unsettled

  • What is a young bird actually comparing its own output against?

    Why it matters: The whole account rests on a stored target and an error signal, and neither has been observed directly. Whether the target is a stored copy of the tutor song or something transformed from it changes what the circuit is doing.

    What would settle it: Recording from the auditory-motor comparison during learning in a way that distinguishes a stored acoustic memory from a motor target.

  • What does a live tutor supply that a loudspeaker does not?

    Why it matters: It is the difference between a sensitive period being a developmental fact and being an artefact of how it was measured, and the same question arises in every study of social learning.

    What would settle it: Tutoring designs that add back one component at a time — contingent response, visual presence, movement — to a taped song.

  • Does the syllable-by-syllable trajectory hold in species with large, open-ended repertoires?

    Why it matters: The precise developmental picture comes from one closed-ended species with a short song. A canary that learns new song every year may not be doing the same thing more often.

    What would settle it: Continuous developmental recording of an open-ended learner, which is technically harder and has rarely been attempted.

Claims about this, checked

Things people have heard, and what the evidence actually supports.

The research behind this page

13 studies, newest first. Each one has a page explaining what it found and what it could not show.

2019Science

Evolution of vocal learning and spoken language

Vocal production learning is documented in a small number of distantly related lineages, and the independently evolved brain circuits supporting it share a recurring feature: a direct projection from forebrain to the motor neurons controlling the vocal organ, which is absent in non-learners.

2009Nature

De novo establishment of wild-type song culture in the zebra finch

Song moved towards wild-type structure at every generation, and within three to four generations the colony was singing song statistically indistinguishable from wild-type on the measured features.

2008Cambridge University Press

Bird Song: Biological Themes and Variations

Song serves two principal functions — repelling rivals and attracting mates — supported by playback and removal experiments.

2005Trends in Ecology & Evolution

Functional aspects of song learning in songbirds

Field studies show young birds learning selectively from the neighbours they will actually compete with, often after settling on a territory, which is later and more socially targeted than the tape-tutoring literature implied.

2001Science

Dynamics of the vocal imitation process: how a zebra finch learns its song

Song development proceeds from highly variable, unstructured subsong through progressive differentiation of individual syllables, with different syllables converging on their targets at different times and rates rather than the song improving as a whole.

1999Nature

Decrystallization of adult birdsong by perturbation of auditory feedback

Song progressively degraded — syllables were lost, added and distorted — over weeks of perturbed feedback, and largely recovered towards the original song once normal feedback was restored.

1998American Zoologist

Song learning, early nutrition and sexual selection in songbirds

Birds subjected to nutritional stress during the song-learning period produced poorer copies and, in some species, had smaller song-control brain regions, with the deficit still detectable in adulthood.

1984Animal Behaviour

Social interaction, sensitive phases and the song template hypothesis in the white-crowned sparrow

Birds tutored by live models learned songs at ages at which tape tutoring no longer worked, and some learned the song of the other species when the tutor was live — an outcome tape tutoring essentially never produced.

1977Science

Selective vocal learning in a sparrow

The young birds learned syllables of their own species and largely ignored the other species’ syllables, regardless of which temporal pattern the syllables were embedded in.

1976Journal of Comparative Neurology

Central control of song in the canary, Serinus canarius

A specific interconnected set of forebrain regions was found to be necessary for normal song, with lesions producing loss or severe degradation of song, and the pathway traced from forebrain to the motor neurons controlling the syrinx.

1970Journal of Comparative and Physiological Psychology

A comparative approach to vocal learning: song development in white-crowned sparrows

Birds tutored during a sensitive period early in life reproduced that dialect as adults.

1965Zeitschrift für Tierpsychologie

The role of auditory feedback in the control of vocalization in the white-crowned sparrow

Birds deafened before singing produced highly abnormal, unstructured song even when they had already heard a tutor.

1958Ibis

The learning of song patterns by birds, with especial reference to the song of the chaffinch Fringilla coelebs

Isolated birds produced song of roughly normal length and pitch range but without the phrase structure and terminal flourish of wild song.

This page is a stop on a longer route

A guided journey reads several subjects in a deliberate order, with an argument for why one follows another. You can join in the middle.

Where to go from here

Each of these follows from something on this page — a relationship in the evidence, a claim people ask about, or the next mechanism along.

How complete this page is, and what it is still missing

NatureHQ publishes its own gaps. This page is at 80% completeness against what we would call a finished subject, and was last reviewed on 2026-09-02. It carries 7 claims and answers 6 mapped search questions.

  • no research from the last few years is attached — check for newer work
  • The neurobiology is summarised on the vocal learning page rather than developed here; the basal-ganglia circuit involved in vocal variability is not covered at all.
  • Hormonal control of song development and seasonal singing is absent.
  • Almost everything here is songbird; how parrots and hummingbirds acquire their vocalisations is much less well described and is not attempted.