The Neurobiology of Birdsong
May 5, 2009 8:24 PM Subscribe
The universal grammar of birdsong is genetically encoded. "A new study, published online in the journal Nature, shows that the songs of isolated zebra finches evolve over multiple generations to resemble those of birds in natural colonies. These findings show that song learning in birds is not purely the product of nurture, but has a strong genetic basis, and suggest that bird song has a universal grammar, or an intrinsic structure which is present at birth."
I've read that bird conversation is fantastically boring. It's all to do with wind speed, wingspans, power-to-weight ratios, and a fair bit about berries.
posted by homunculus at 8:54 PM on May 5, 2009 [4 favorites]
posted by homunculus at 8:54 PM on May 5, 2009 [4 favorites]
nom nom chomsky chomsky nom nom choms
posted by localhuman at 8:57 PM on May 5, 2009 [3 favorites]
posted by localhuman at 8:57 PM on May 5, 2009 [3 favorites]
I know that the song of the caged bird is the same as the song of any other bird.
I know that it takes food and water to keep that bird alive.
And I ask myself, "why?"
A kind of song, the same as any other.
posted by nervousfritz at 9:45 PM on May 5, 2009 [1 favorite]
I know that it takes food and water to keep that bird alive.
And I ask myself, "why?"
A kind of song, the same as any other.
posted by nervousfritz at 9:45 PM on May 5, 2009 [1 favorite]
What I find less compelling about this argument is that we know very little about the syntax of birdsong (as escabeche said), hence we know very little about the combinatorial properties of birdsongs that are "grammatical" or "ungrammatical." I think it is interesting that these birds started to sound more like birds in the wild, which in fact might have something to do with genetically programmed birdsong code that needs to be accessed during some sensitive period. However, that doesn't mean that birds find certain "structures" to be grammatical or ungrammatical. And on a related note, there is human evidence from creole and pigdin language studies that suggest that children overgeneralize in ways that make a pigin more language-like. This has also been found with a native Deaf child who learned American Sign Language from birth from parents who hadn't learned ASL until they were teenagers.
Anyway, I don't think that this study is necessarily in support of Universal Grammar, since it also seems to support the idea that child and young birds alike may pick up on statistical regularities of one sound following another, and will use that information to create a more regular and language-like/birdsong-like system. This doesn't mean that certain rules are hardwired in the brain, but rather that children and birds can pick up on these same rules via exposure to patterns, even if those patterns are irregular, as is the case with the Deaf parents and the song-deprived birds.
posted by joan cusack the second at 9:58 PM on May 5, 2009
Anyway, I don't think that this study is necessarily in support of Universal Grammar, since it also seems to support the idea that child and young birds alike may pick up on statistical regularities of one sound following another, and will use that information to create a more regular and language-like/birdsong-like system. This doesn't mean that certain rules are hardwired in the brain, but rather that children and birds can pick up on these same rules via exposure to patterns, even if those patterns are irregular, as is the case with the Deaf parents and the song-deprived birds.
posted by joan cusack the second at 9:58 PM on May 5, 2009
Wow. Strike both "pigdin" and "pigin" and replace with "pidgin." Pigdin? No more comments after midnight for me.
posted by joan cusack the second at 10:59 PM on May 5, 2009
posted by joan cusack the second at 10:59 PM on May 5, 2009
Does birdsong really have "grammar" as sophisticated as the syntax of human speech?
posted by KokuRyu at 11:19 PM on May 5, 2009
posted by KokuRyu at 11:19 PM on May 5, 2009
Does birdsong really have "grammar" as sophisticated as the syntax of human speech?
Well, they certainly wouldn't be able to tell from this study.
posted by delmoi at 11:45 PM on May 5, 2009
Well, they certainly wouldn't be able to tell from this study.
posted by delmoi at 11:45 PM on May 5, 2009
Does birdsong really have "grammar" as sophisticated as the syntax of human speech?
It probably depends on your measure of sophistication. At least some birds can generate recursive linguistic structures (like phrases in English that can contain other phrases which contain other phrases like this one).
posted by a snickering nuthatch at 11:46 PM on May 5, 2009
It probably depends on your measure of sophistication. At least some birds can generate recursive linguistic structures (like phrases in English that can contain other phrases which contain other phrases like this one).
posted by a snickering nuthatch at 11:46 PM on May 5, 2009
I'll buy this argument when they also find that 1 in every 20 birds is fanatical about birdsong grammar and goes around tweeting the rules at others.
posted by srboisvert at 1:06 AM on May 6, 2009 [3 favorites]
posted by srboisvert at 1:06 AM on May 6, 2009 [3 favorites]
Does birdsong really have "grammar" as sophisticated as the syntax of human speech?
Moreso than Twitter.
posted by rokusan at 4:02 AM on May 6, 2009 [1 favorite]
Moreso than Twitter.
posted by rokusan at 4:02 AM on May 6, 2009 [1 favorite]
It's not so impressive when you realize that 90% of zebra finch grammar boils down to "My tree! Mine! Hey! HEY! Other birds! Fuck you."
posted by Mr. Bad Example at 4:59 AM on May 6, 2009 [7 favorites]
posted by Mr. Bad Example at 4:59 AM on May 6, 2009 [7 favorites]
(And yes, I realize that the Wikipedia article says zebra finches are ground-dwelling and shouldn't be laying claim to trees at all. This is just further proof that zebra finches are tremendous assholes.)
posted by Mr. Bad Example at 5:00 AM on May 6, 2009 [2 favorites]
posted by Mr. Bad Example at 5:00 AM on May 6, 2009 [2 favorites]
There's something fitting about work addressing a nature vs. nurture argument being published in Nature.
posted by kisch mokusch at 7:26 AM on May 6, 2009
posted by kisch mokusch at 7:26 AM on May 6, 2009
Mr. Bad Example: 90% of zebra finch grammar boils down to "My tree! Mine! Hey! HEY! Other birds! Fuck you."
Actually, in my experience, most of it tends to be; "Hey baby, wanna fuck?". Because that is pretty much all zebra finches do.
Except stutter. Recent scientific studies reveal that they do that too, which is pretty neat as it's starting to explain why it happens in humans.
posted by quin at 8:53 AM on May 6, 2009
Actually, in my experience, most of it tends to be; "Hey baby, wanna fuck?". Because that is pretty much all zebra finches do.
Except stutter. Recent scientific studies reveal that they do that too, which is pretty neat as it's starting to explain why it happens in humans.
posted by quin at 8:53 AM on May 6, 2009
An excellent video on the subject of stuttering finches.
posted by quin at 8:58 AM on May 6, 2009 [1 favorite]
posted by quin at 8:58 AM on May 6, 2009 [1 favorite]
I just had this vision of an avian UN where all the birds sit at a circular table and mockingbirds wear headphones and serve as translators.
I think I'll go lie down for a bit.
posted by BitterOldPunk at 9:10 AM on May 6, 2009
I think I'll go lie down for a bit.
posted by BitterOldPunk at 9:10 AM on May 6, 2009
Presumably there'll be a follow-up paper where they attempt to explain why universal grammar doesn't explain this aspect of bird song, and why that explanation doesn't cover this other aspect, and so on until you've got half a dozen really big holes in the original theory, all the while insisting that it's a perfect explanation.
posted by oaf at 9:16 AM on May 6, 2009
posted by oaf at 9:16 AM on May 6, 2009
How on earth can they distinguish between a Universal Grammar and a pattern detection and imitation system where certain types of pattern are just easier to hear or reproduce? Is that what "multigenerational phenotype" means?
posted by Wrinkled Stumpskin at 10:17 AM on May 6, 2009
posted by Wrinkled Stumpskin at 10:17 AM on May 6, 2009
That's really a nice paper.
posted by fantabulous timewaster at 10:43 AM on May 6, 2009
posted by fantabulous timewaster at 10:43 AM on May 6, 2009
How on earth can they distinguish between a Universal Grammar and a pattern detection and imitation system where certain types of pattern are just easier to hear or reproduce?
To the extent that they can (and in fact I don't believe the actual Nature paper tries to), this research does not obviously support a universal grammar for these birds. The words "universal grammar" appear only in the abstract, and that is just as an example of how species-specific constraits can limit cultural diversity. The reason I don't believe the research actually supports any UG-like view of bird-song, is that the songs only evolved towards natural birdsong over the course of 3-4 generations. This is exactly what you'd expect from general constraints on processing/production limiting the kinds of errors the birds made and reproduced. The language cases that they make a parallel to (the spontaneous development of sign language among deaf children in Nicaragua) happened (IIRC) over about one generation, and the results were a full-fledged language within that generation.
The nature paper certainly does make an analogy to UG, but it is really irresponsible to report it as if they were finding UG in zebra finches...on that note, the quote in the post is actually from a blog describing the paper, not the paper itself.
posted by advil at 1:54 PM on May 6, 2009
To the extent that they can (and in fact I don't believe the actual Nature paper tries to), this research does not obviously support a universal grammar for these birds. The words "universal grammar" appear only in the abstract, and that is just as an example of how species-specific constraits can limit cultural diversity. The reason I don't believe the research actually supports any UG-like view of bird-song, is that the songs only evolved towards natural birdsong over the course of 3-4 generations. This is exactly what you'd expect from general constraints on processing/production limiting the kinds of errors the birds made and reproduced. The language cases that they make a parallel to (the spontaneous development of sign language among deaf children in Nicaragua) happened (IIRC) over about one generation, and the results were a full-fledged language within that generation.
The nature paper certainly does make an analogy to UG, but it is really irresponsible to report it as if they were finding UG in zebra finches...on that note, the quote in the post is actually from a blog describing the paper, not the paper itself.
posted by advil at 1:54 PM on May 6, 2009
We can never know if animal communication has the segmental grammatical structure of human language. Never. Because we cannot confirm the most basic property of human language without intersubjective access to the meaning of utterances; we cannot therefore confirm that animal language has the capacity of discrete and combinatorial reference to the external world. Yes, I know vervets "prevaricate" and bees can tell their hive mates where the good flowers are. Pretty thin soup compared to the way we can confirm that human languages we don't understand at first actually refer to the same world as the one we know from our own language.
The obsession with discovering a universal grammar and a purely neurobiological explanation for language (and thus culture and consciousness) knows no bounds. We so want to believe.
posted by fourcheesemac at 4:46 PM on May 6, 2009
The obsession with discovering a universal grammar and a purely neurobiological explanation for language (and thus culture and consciousness) knows no bounds. We so want to believe.
posted by fourcheesemac at 4:46 PM on May 6, 2009
How on earth can they distinguish between a Universal Grammar and a pattern detection and imitation system where certain types of pattern are just easier to hear or reproduce?
Even in human language, we can't. There's a bit of analytic machinery called Optimality Theory, all the rage in phonology these days, which in a nutshell says that the structure of a language is dictated by the tension between the need to use easy patterns and the need to use distinct patterns for different things.
(In a slightly bigger nutshell: Things that are hard to hear or produce are 'marked' and violate 'markedness' constraints, but you can't just turn everything into mumbly mush, because things that are too different from how you learned them violate 'faithfulness' constraints. Every language ranks these constraints in a particular order. When you want to speak, the theory goes, all the things you could say are passed down the line of constraints. Each constraint keeps the candidates that violate it the least, and eliminates the rest, until only one candidate is left. The theory's got some problems, but it offers at least a plausible account for some similarities among human languages, and it's constrained enough that, if it were the Universal Grammar, languages would be learnable.)
/nerd
posted by eritain at 8:05 PM on May 6, 2009 [1 favorite]
Even in human language, we can't. There's a bit of analytic machinery called Optimality Theory, all the rage in phonology these days, which in a nutshell says that the structure of a language is dictated by the tension between the need to use easy patterns and the need to use distinct patterns for different things.
(In a slightly bigger nutshell: Things that are hard to hear or produce are 'marked' and violate 'markedness' constraints, but you can't just turn everything into mumbly mush, because things that are too different from how you learned them violate 'faithfulness' constraints. Every language ranks these constraints in a particular order. When you want to speak, the theory goes, all the things you could say are passed down the line of constraints. Each constraint keeps the candidates that violate it the least, and eliminates the rest, until only one candidate is left. The theory's got some problems, but it offers at least a plausible account for some similarities among human languages, and it's constrained enough that, if it were the Universal Grammar, languages would be learnable.)
/nerd
posted by eritain at 8:05 PM on May 6, 2009 [1 favorite]
« Older The Star Trek Failure Generator | Why you're probably not named Tricia Newer »
This thread has been archived and is closed to new comments
posted by escabeche at 8:28 PM on May 5, 2009