Skip to content
NatureHQ

Birdsphenomenon

Birdsong dialects

A sparrow’s song can place it within a few kilometres — and a variant that started in western Canada replaced the old one across a continent in twenty years.

Birds in one valley can sing a consistently different version of the same species song from birds a few kilometres away, and the difference is learned rather than inherited. A bird raised away from its home population does not develop the local variant.

A dialect is the visible output of vocal learning running in a population. If each generation copies the birds it grows up among, local versions drift apart and then persist, and after a while a recording can be assigned to a place. That is a claim about cultural transmission, and it needs the two obvious alternatives excluded: that each bird works out the same solution independently, and that the difference is genetic. Both were ruled out early and by different means — the boundaries are far finer-grained than any plausible population structure, and birds taken from a dialect area and raised without local song simply do not develop the local variant. What has changed since is the timescale. Song culture was assumed to be slow, and it turns out to do both things: a variant of the white-throated sparrow’s song appeared in western Canada and replaced the traditional ending across the continent in about twenty years, while model fitting on swamp sparrow song implies that some song types have persisted essentially unchanged for centuries. The difference is not random. It depends on whether young birds copy what they hear or copy what most birds around them are doing — and conformity, it turns out, is what freezes a tradition. The word "dialect" is worth handling carefully. It is borrowed from human language and brings assumptions with it: that a dialect marks who you are, that speakers of different ones understand each other with effort. Neither is established for birds, and both keep getting assumed.

Developed coverage · 63% complete · reviewed 2026-09-02

What this page covers

Sharply bounded dialects are best documented in a few songbirds — white-crowned and white-throated sparrows above all — and in at least one parrot. Most songbirds show gradual geographic variation rather than crisp boundaries.

Often confused with: Human dialects, which the word is borrowed from and which carry implications about identity and mutual understanding that the bird case does not support; Genetic differences between populations, which dialects are specifically not

Quick facts

Learned, not inherited
Birds raised away from a dialect area do not develop its song
How fast culture can move
A continent-wide song replacement in about twenty years
How slowly it can move
Some song types modelled as stable for centuries
Not only songbirds
Regional dialects also documented in a parrot’s contact calls

What makes a difference a dialect

Two alternatives have to be excluded before geographic variation means anything.

Diagram

What a dialect boundary is, and what it is not

What a dialect boundary is, and what it is notValley AValley Ba few kilometresIt is: learned. A bird raised elsewhere never sings Valley A.It is not: genetic. The boundaries are finer than the genetics.And “dialect” is borrowed from human language. Nothing here saysthe birds fail to understand each other.
The same explanation in words

Two panels side by side, separated by a dashed line labelled "a few kilometres". The left panel, Valley A, shows three identical arch-shaped song traces. The right panel, Valley B, shows three identical zigzag traces — the same species, a consistently different song. Beneath, two statements. It is learned: a bird raised away from Valley A does not develop the Valley A song. It is not genetic: the boundaries are far finer-grained than the population structure. A final note adds that "dialect" is borrowed from human language, and that nothing here implies the birds fail to understand each other.

Local song variants are learned from local birds, not inherited. Populations a few kilometres apart can sing distinguishably different songs with no genetic difference behind it.

Established

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

Geographic variation in song in several songbird species takes the form of sharply bounded local variants that are acquired by learning from local conspecifics. Birds raised without exposure to the local variant do not develop it, and variant boundaries are finer-grained than the genetic population structure.

Who this applies to
Demonstrated in a small number of well-studied species across two bird groups; geographic song variation is far more widespread than the crisp dialect case.
Studied in
Zonotrichia leucophrys, Zonotrichia albicollis, Amazona auropalliata
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

The combination that rules out the alternatives is present: variants are too fine-grained to track genetics, and laboratory rearing without local models fails to produce them.

How far it can be extended

Learned geographic variation appears independently in songbirds, parrots and cetaceans, which is the pattern expected if it follows from vocal learning rather than from anything species-specific.

Caveats

  • The sharply bounded dialects of white-crowned sparrows are unusually crisp; most songbirds show gradual clinal variation instead.
  • Whether dialect boundaries have any consequence for mating or dispersal is contested and varies by species.
  • "Dialect" is borrowed from human language and carries implications about identity and comprehension that the bird case does not support.

Still unanswered

  • Whether birds treat a foreign dialect as a different signal or merely as an unfamiliar version of the same one.

Last reviewed 2026-09-02

The evidence (3 studies)

The crisp boundaries that made the phenomenon famous are not typical. White-crowned sparrow dialects around San Francisco Bay change over a few kilometres sharply enough to be heard, and that is what made them tractable; most songbirds show gradual clinal variation with no boundary anywhere. Reading the sparrow case as the general one makes dialects sound tidier than they are.

Words used here
Dialect
A version of a species’ song shared by birds in one area and consistently different from the version next door, arising because young birds copy local adults.
Cultural transmission
Behaviour passed on by learning from others rather than through genes. What makes a dialect a dialect.

Song culture can move fast, or barely at all

Two results in closely related sparrows, pointing in opposite directions, and both correct.

A new song ending appeared in western Canada and replaced the old one across the continent in about twenty years — a cultural change tracked as it happened.

Well supported

Good evidence backs this, though some details remain open.

A doublet-ending variant of white-throated sparrow song, first documented in a western Canadian population, spread eastward across most of the sampled Canadian range over approximately two decades, largely replacing the historical triplet ending.

Who this applies to
One North American songbird, documented across most of its Canadian range.
Studied in
Zonotrichia albicollis
Why we rate it this way, and what the caveats are
Well supportedHigh confidence

The replacement is large, documented over two decades and across a continent, using archived and citizen-science recordings that predate the hypothesis being tested.

Caveats

  • Observational. The proposed mechanism — young males meeting the variant on shared wintering grounds — is inference, not demonstration.
  • Recording effort is uneven, and citizen-science archives over-represent accessible places.
  • Why this variant spread rather than remaining local is unknown; nothing shows it is preferred or more effective.

Still unanswered

  • Whether the doublet confers any advantage at all, or whether the sweep is cultural drift amplified by winter mixing.

Last reviewed 2026-09-02

The evidence (2 studies)

The sparrow sweep is worth noticing for how it was documented as much as for what it shows. Nobody set out in 1998 to record a continent; the analysis leans on twenty years of recordings made by amateur birders and deposited in public archives, which is the only reason anyone had the coverage. It is unusually strong evidence about what happened and necessarily weak about why — no experiment can be run on a continent, and the proposed mechanism, young males meeting the new variant on shared wintering grounds, is inference.

Song culture can also do the opposite of changing. Where young birds preferentially copy the commonest version they hear, a song type can persist essentially unchanged for centuries.

Emerging evidence

Real findings exist, but too few or too recent to be settled.

Fitting competing cultural transmission models to a large corpus of wild swamp sparrow song favoured conformist learning over unbiased copying with error, and the fitted models implied persistence times for individual song types on the order of centuries.

Who this applies to
One songbird species with a small, discrete repertoire — the easiest case for this kind of model fitting.
Studied in
Melospiza georgiana
Why we rate it this way, and what the caveats are
Emerging evidenceModerate confidence

The model comparison is careful and the conformity result is convincing on the data. The centuries-long persistence is an output of the fitted model rather than an observation, and inherits its assumptions.

Caveats

  • Persistence over centuries is inferred from a model, not observed; no recordings exist from before the twentieth century.
  • Conformity is inferred from the distribution of variants rather than from watching birds choose.
  • A species with a small discrete repertoire is the most favourable possible case for this method.

Still unanswered

  • What determines whether a song culture conforms and freezes or drifts and sweeps, given that both happen in closely related sparrows.

Last reviewed 2026-09-02

The evidence (2 studies)

Put the two together and the useful conclusion is not that song culture is fast or slow but that its speed is set by a specific, testable property of how learners choose. Copy what you hear and variants drift and spread. Copy what most of your neighbours are doing and the commonest variant is reinforced every generation, and the tradition freezes. Both mechanisms are present in the same genus.

Is a dialect the same as an accent?

The word is borrowed, and it brings assumptions that have not been tested.

The short answer

Do birds have accents?

They have local song variants that are learned from local birds, which is the part of an accent that transfers. What has not been shown is that a bird treats a foreign variant the way a person treats a foreign accent — as the same thing said differently.

Two things get imported with the word and neither is established. The first is comprehension: a human hearing an unfamiliar accent understands it with a little extra effort, and whether a sparrow hearing a foreign dialect recognises it as its own species saying the same thing, or as something else, varies by species and by study. The second is identity — the idea that a dialect marks where you are from and that birds use it that way. There is evidence in some species that males respond more strongly to local song, and evidence in others of no discrimination at all. The honest position is that dialects are real and well evidenced as a cultural phenomenon, and their consequences for the birds are much less settled than the vocabulary implies.

Check it for yourself

The same phenomenon in animals that are not birds

Dialects follow vocal learning wherever it occurs, which is one of the better arguments that the two are connected.

  • Sperm whales: clans sharing the same water use distinguishably different coda repertoires, and the boundaries are social rather than geographic.
  • Orcas: pods maintain distinct call repertoires that persist for decades and track matriline rather than location.
  • Parrots: yellow-naped amazons in Costa Rica have contact-call dialects with boundaries sharp enough to assign a recording to a region.
  • Some bats: colony-specific call features have been described, on a much thinner evidence base than any of the above.

The pattern is the argument. Learned vocal variants that differ between groups turn up in songbirds, parrots and cetaceans — three lineages that acquired vocal learning independently — and in essentially no group that lacks it. That is what you would expect if dialects are a consequence of the learning rather than a separate adaptation, and it is a rare case where a cross-species regularity is doing real explanatory work rather than decorating a list.

The cetacean cases

What is still unsettled

  • Do birds treat a foreign dialect as a different signal, or as the same signal said oddly?

    Why it matters: It decides whether dialects are consequences of learning or things birds actually use, and the whole "accents" framing rests on the answer.

    What would settle it: Systematic playback of local and foreign dialect across many species with the same protocol, rather than one species at a time.

  • Why did the white-throated sparrow doublet spread rather than staying local?

    Why it matters: A continent-wide replacement in twenty years is either selection on something, or drift amplified by winter mixing. Those are very different stories about how vocal culture works.

    What would settle it: Evidence that females prefer the new ending or that it is more effective against rivals — neither of which currently exists.

  • What decides whether a song culture conforms and freezes or drifts and sweeps?

    Why it matters: Both happen in closely related sparrows, so the difference is not deep phylogeny. Something about the species’ social structure or dispersal is doing it, and nobody knows what.

    What would settle it: Fitting the same transmission models across a set of species that differ systematically in dispersal and in when they learn.

Claims about this, checked

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

The research behind this page

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

2020Current Biology

Continent-wide shifts in song dialects of white-throated sparrows

A variant ending in a two-note doublet, first noted in a small western population, spread east across the continent over about twenty years and replaced the historical triplet ending across most of the range surveyed.

2019Animal Behaviour

Culture and cultural evolution in birds: a review of the evidence

Song is by a wide margin the best-evidenced avian culture, with experimental support from tutoring, isolation and population-level tracking.

2018Nature Communications

Cultural conformity generates extremely stable traditions in bird song

The observed pattern was best explained by conformist learning, and the fitted models implied that some song types have persisted essentially unchanged for many hundreds of years.

1996Proceedings of the Royal Society B

Regional dialects in the contact call of a parrot

Calls fell into a small number of regionally distinct types with sharp boundaries between them, and birds within a region shared call structure closely regardless of roost.

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.

1964Science

Culturally transmitted patterns of vocal behavior in sparrows

Populations only a few kilometres apart sang consistently different song variants, sharp enough that a recording could be assigned to a location by ear.

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 63% completeness against what we would call a finished subject, and was last reviewed on 2026-09-02. It carries 3 claims and answers 11 mapped search questions.

  • more experiments could be explained in plain English
  • Whether dialect boundaries affect mate choice or dispersal is summarised as contested rather than reviewed properly; the literature is large and inconsistent.
  • Clinal variation — the common case — gets much less attention here than the crisp dialects that are the exception.
  • The bat dialect literature is mentioned in a single line and is not assessed.