To Stereotype or Not

Lilac-crowned Parrot, a species with relatively unstereotyped vocalizations. Arizona-Sonora Desert Museum, Tucson, AZ, 5/13/2009. Photo by DrStarbuck (Creative Commons 2.0).

Lately I’ve been thinking a lot about stereotype. Not the racial or ethnic kind, but the degree to which a bird’s song is the same each time it’s sung. In a couple of recent blog posts, I showed that Violet-green Swallows and Evening Grosbeaks produce stereotyped vocalizations in complex strings, and therefore I argued that they were exhibiting true singing behavior. Shortly afterwards, Andrew Spencer asked me an intriguing question: why does a vocalization have to be stereotyped in order to qualify as a song? Are there any birds that sing non-stereotyped songs?

I can’t find an answer in the scientific literature.  Among researchers, the definition of bird “song” (as opposed to “call”) has been pretty controversial over the years, but the question of stereotypy isn’t the controversial part.  Most authors simply take for granted that bird songs should be stereotyped — or “crystallized,” in the terminology often used for learned songs.

And perhaps for good reason: I haven’t been able to find any clear-cut examples of unstereotyped courtship song.  Sure, there are the “subsongs” of young, inexperienced males, and birds may sing variable songs at different times of year — but it seems that when an adult male really needs to impress a female, in almost any species, stereotype becomes important.

This may be because females, at least in some species, prefer stereotyped songs.  In the Zebra Finch, for example, the degree of stereotypy varies depending on what the male is doing. When he’s by himself, each song is different from the last, albeit in very subtle and minor ways. But when a female shows up nearby and the male’s got courtship on his mind, he begins to sing directly to her, and those subtle differences go away: each song now becomes a precise copy of the last. Woolley and Doupe (2008) showed that females prefer this stereotyped courtship song over the more variable undirected kind — meaning that females are listening not just to how males produce their songs, but how they reproduce them.

Is stereotypy easier or harder?

Which demands more of the singer—stereotyped or non-stereotyped vocalizations?  My guess would be the former; it seems intuitive that having to produce a sound multiple times in precisely the same way, without any tolerance for variation, should be the more difficult task. But when it comes to demands on the listener, things might be the other way around. Recognizing a sound is likely to be easier when the sound only takes one form. If the singer is trying to send the same message in a slightly different way each time, it actually requires the listener to recognize a class of similar sounds rather than a single sound – presumably a more difficult cognitive challenge.

Let’s consider an example from human speech.  A common word like “birthday” is not going to sound exactly the same every time you say it. Sometimes you might say it quickly, sometimes slowly; sometimes while laughing or when your nose is stuffed up; sometimes with your voice rising toward the end of a question. Sometimes you might even sing it. In all these cases, the spectrogram of your voice saying “birthday” is going to look different—and yet, every single time, our brains reliably classify these sounds as renditions of the the same word. That takes some mental firepower—it’s difficult for computers to do, which is why speech transcription software remains imperfect.

Perhaps it’s no coincidence that the best examples I could find of birds that rarely give stereotyped vocalizations are also the most intelligent birds: parrots and ravens.  Nothing that these birds say in the wild meets the classic definitions of bird “song,” but they are apparently lifelong vocal learners — ravens, like parrots, can be taught to speak in captivity — and they apparently recognize other individuals of their species by voice, even though no two utterances are likely to be precisely the same.  Compare these calls from a Lilac-crowned Parrot in Mexico:

Four examples of calls from one Lilac-crowned Parrot; they sound identical but vary slightly on the spectrogram. Rancho Santa Barbara, Sonora, Mexico, 6/24/2010.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Here’s an unedited sample of consecutive calls from the same bird, showing how different types of calls grade into one another:

Sequence of Lilac-crowned Parrot calls, showing variation. Same bird as above.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Even these parrots have the ability, and apparently the need, to reproduce almost exactly the same call over and over, even if they don’t do it with quite the same precision as most other birds.  Is this lack of perfectionism a sign of intelligence?  Or merely a sign that they’re not in an amorous mood?  Are there any species in which the females are attracted to vocal innovators instead of virtuoso self-imitators?

At the moment, nobody really knows.

3 comments to To Stereotype or Not

  • Hi Nathan,

    I’d be interested in your comments on this study: Kroodsma and Parker (1977) recorded 4654 “phrases” or “song units” from a singing Brown Thrasher (in 113 minutes). They did a detailed comparison of every hundredth phrase throughout the recording to arrive at an estimate that this individual bird sang 1,805 distinctly different song units (yes, nearly two thousand!). There was no way of estimating the total repertoire size for this male, and similarly they could not determine whether the repertoire was stereotyped or the bird was continually improvising new song units.

    I don’t know if there has been any follow-up to this study, but it raises a lot of interesting questions.

    Kroodsma, D. E. and L. D. Parker. 1977. Vocal virtuosity in the Brown Thrasher. Auk 94:783-785. http://elibrary.unm.edu/sora/Auk/v094n04/p0783-p0785.pdf

  • Nathan Pieplow

    Very interesting, David. I knew of this study but had never actually read it before. To clarify for those who haven’t read it: Kroodsma and Parker compared each hundredth song phrase to all the song phrases on the entire recording to see whether and how often they were repeated. 20 of the 45 song phrases were never repeated by the bird in the two hours of singing they recorded. The 2 song phrases that the bird used most were repeated only 7 times each in those two hours.

    I hadn’t closely examined the level of stereotypy in mimid songs before, but it looks like it’s fairly low in Brown Thrasher — slightly lower than in the Lilac-crowned Parrot above. Even when the Brown Thrasher repeats the same thing twice in a row (as it almost always does), those consecutive renditions often vary slightly. On the other hand, Kroodsma & Parker’s thrasher repeated some phrases up to about 45 minutes apart, so it definitely had to “remember” them, meaning it’s not a full-time innovator, if it innovates at all.

    This is definitely worthy of more study.

  • Dan Lane

    In the case of the parrots and corvids, I think the stereotypy is perhaps more a means to have a species-specific vocalization (perhaps best considered a “call” rather than a “song”) that can be used to identify the vocalist without much question (especially in places where more than one species in that family, and particularly that genus, are present). “Songs” (for pairbonding purposes… I suspect it is rare to have a “song” used for territorial purposes in these two groups) in those groups appear to be unconventional at best. Are the “whisper songs” of corvids important for pair bonding (if so, why do they seem to do it most when the vocalist is alone)? The caterwauling of parrots? Of course, trying to apply human concepts such as “song” and “call” to all birds (much less other organisms) may be a flawed endeavor.

    In the case of the thrasher songs, I would argue that individual elements of the song are perhaps less important to the listeners than the overall pattern of the song (in the case of Brown Thrashers, the paired identical notes, for example). This is the species-specific information that is necessary to identify the singer to a species, and perhaps is what is looked for by the females who make the selections that drive the evolution of the vocalizations. Of course, in pointing out that the paired notes in a strophe of Brown Thrasher song are identical (or virtually so) may prove the stereotypy rule Nathan has laid out! Even if a note is invented and used only once in a thrasher’s singing career, it is in fact sung twice in quick succession, so the ability to reproduce it with fidelity is still important. Perhaps with other thrashers that do not have paired identical notes, this is not an issue, but some other aspect is?

    Good listening