Learning about Language from Birdsong

Songbirds are one of the few known species who learn to speak (or sing) like we do. That makes them the perfect case study to learn about the origins of language in the brain.

Transcript: How do we go from this [audio: baby babbling] to this [audio: toddler speaking]? Language is one of the defining features of being human, but how are we able to acquire it?

Many animals make sounds – lions roar [audio: lion roaring] and frogs croak [audio: frog croaking]. But most don’t learn to make these the same way we learn language. As babies, we listened to people around us speaking and learned how to imitate those sounds. Humans, dolphins, whales, and some bats are the only mammals that learn to vocalize this way. But some species of songbirds also employ this kind of vocal learning.

In fact, songbirds have a lot to teach scientists about vocal learning. By studying how songbirds learn their tunes, [audio: songbird] we are discovering how our own brains are organized and how we acquire language.

Research in songbirds like zebra finches has unearthed important clues about the brain circuits underlying vocal learning. Native to Australia, the zebra finch is a stocky little songbird. With bright orange cheeks and black and white striped feathers adorning his breast, the male zebra finch is an ostentatious entertainer, belting out tunes in hopes of impressing a female. But he wasn’t born with such a beautiful singing voice — he acquired it through learning and practice.

Zebra finches begin their vocal training with indiscriminate chirping. After listening to adult finches, they slowly progress to singing full tunes, just like babies progress from babbling to uttering complete sentences by about 3 and a half years of age. And just like in human babies, there’s a critical window of time during which a young finch must hear adults sing so he can learn to make those same sounds.

Whether it’s human language or birdsong, producing sounds is a motor skill. Anything that requires you to consciously move your muscles is a motor skill, like running or swinging a bat. Even getting out of bed in the morning and then walking downstairs. And motor skills, as we all know, take practice. It’s a process of trial and error that requires sensory feedback from—wait, let’s back up a little bit.

Outside of your brain and spinal cord there are 2 kinds of nerves: motor nerves and sensory nerves. Extending from the spinal cord, motor nerves signal your muscles to move. Sensory nerves, on the other hand, send information about sensation — touch and position to name a few — back to your brain.

Motor skills — like mastering a wicked backhand in tennis, playing the piano without missing a note, or learning how to speak — depend on this kind of sensory feedback so we can adjust and perfect our performance through trial and error.

Key to this process of trial and error is a group of brain structures called the basal gangliaA collection of structures thought to be especially import... More. It receives motor and sensory information from an area of the brain called the cerebral cortex which is responsible for planning movement among many other things. The basal ganglia relays this information back to the cortex and other areas of the brain, forming a circuit. This circuit is important for learning how to execute precise motor behaviors like language or birdsong.

Although scientists once thought this circuit was only involved in song learning during development, they have since discovered another important role: altering the bird’s song performance to hone motor skills as an adult.
For example, when a male zebra finch sings alone, neurons in the circuit seem to fire randomly and the finch launches into a free-form improv session [audio: jazz music].

But if there happen to be a female zebra finch nearby, the neurons in the circuit fire in a regular, patterned way and the finch performs a well-rehearsed tune [audio: guitar music].

From these experiments, scientists discovered this circuit helps us perfect our ability to sing, play guitar, ace our opponents and perform any number of activities.

By revealing how the songbird brain generates these different vocal behaviors, researchers are beginning to get a handle on motor learning throughout the lifespan and how social signals powerfully influence learning in all animals, especially highly social creatures like ourselves. [audio: background chirping and guitar music]

~

Animation by Matt Wimsatt. Illustration by Sean Noah.

Brainfacts.org

Brainfacts.org is a public information initiative of the Kavli Foundation, the Gatsby Charitable Foundation, and the Society for Neuroscience.

Latest posts by Brainfacts.org (see all)

Brainfacts.org

Brainfacts.org is a public information initiative of the Kavli Foundation, the Gatsby Charitable Foundation, and the Society for Neuroscience.

3 thoughts on “Learning about Language from Birdsong

  • August 26, 2017 at 1:05 pm
    Permalink

    Transcript: How do we go from this [audio: baby babbling] to this [audio: toddler speaking]? Language is one of the defining features of being human, but how are we able to acquire it?

    GS: Good question! And, more importantly, decent answers – as far as they go…I’ll humbly offer a few comments below. Well, I’ll offer them in my usual style instead of humbly.
    Transcript: Many animals make sounds – lions roar [audio: lion roaring] and frogs croak [audio: frog croaking]. But most don’t learn to make these the same way we learn language. As babies, we listened to people around us speaking and learned how to imitate those sounds. Humans, dolphins, whales, and some bats are the only mammals that learn to vocalize this way. But some species of songbirds also employ this kind of vocal learning.

    GS: Yes…imitation probably plays a large role but what is going on is complicated. And it would pay – but you don’t ever see it – if the phenomena would be discussed using the technical vocabulary that emerged from the natural science of behavior (i.e., behavior analysis) since the vocabulary emerged via rigorous experimental control and equally rigorous conceptual analyses. So…I’ll tell you what happens…and actually the above paragraph from the transcripts is sort of contradicted later on. Above, imitation is pointed to, and later, “trial-and-error” learning (i.e., operant conditioning) is involved. Further, the song that the young birds hear is not only imitated but it also – speaking technically – functions as a conditioned reinforcer. Here’s how everything hangs together:

    The young bird hears the song early, and doesn’t, to my knowledge, sing at all yet. Later, the bird begins to sing. Now, to the extent that, when the bird FIRST begins to sing, the song is influenced by the original hearings of the song, that is the extent to which imitation ALONE is operating. But the bird also hears itself sing – that is, what the bird sings is a stimulus that is generated, obviously, by its behavior – the heard-song is a CONSEQUENCE of the sung-song (this response-consequence relationship may be eliminated experimentally by deafening the birds after they have heard the model song, but before they begin to sing). And consequences are often important if they modify the behavior that results in those consequences. And in this case, they appear to do just that. That is, the heard song, to the extent it resembles the original song heard, functions as a positive reinforcer which, by definition, makes the behavior that produced it more probable. The birds singing behavior is automatically shaped into its final form by reinforcement from the song produced by the singing behavior itself. This is “just” operant conditioning but the automaticity of the response-consequence contingency is hard, perhaps, to “wrap one’s head around.” In principle, it is no different than the prototypical food-reinforced lever-press; the rat pressed the lever (behavior) and the consequence was the delivery of food. As a result of being exposed to this response-consequence CONTINGENCY, the rat’s rate of lever-pressing increases. In the birdsong example, the rate of responses that produce a song (consequence) that resembles part of the original increase in frequency (the process called “positive reinforcement”) and those that do not result in song that resembles the original decrease in frequency (the process called “extinction”). The result of this process is, of course, that the bird comes to sing a song resembling the original song it heard, and to not sing phrases that are not part of that song.The only unusual thing, process-wise, is how the original song becomes a conditioned reinforcer. It seems that merely being exposed to the original model song establishes it (the song) as a conditioned reinforcer. This is like what happens in classical imprinting in waterfowl – the first moving thing seen during the critical period makes it so reducing the distance between itself (the duck) and the object is a reinforcer.

    Transcript: In fact, songbirds have a lot to teach scientists about vocal learning. By studying how songbirds learn their tunes, [audio: songbird] we are discovering how our own brains are organized and how we acquire language.

    GS: Well, studying the birds’ behavior doesn’t tell you anything about the brain and how the brain comes to mediate the behavioral-level events is, IMO, a science different than the natural science of behavior qua behavior. But studying the brain in relation to the behavior – how the brain mediates the relevant behavioral processes is certainly important. But here a problem rears its head: even the descriptions in the scientific literature of just what the relevant behavioral processes are prove to be as fuzzy as the way they are expressed here in the very brief post to which I am responding. The way I described the acquisition of the song is accurate, and it – so to speak – parses the behavioral processes in the right way. Not somewhat vague references to “imitation” and “trial-and-error learning.” This points to a very big problem in neuroscience – ordinary language (and its resulting folk-psychology) and the cognitive psychology and neuroscience that it eventually spawned do not parse behavior into the “right processes.” The relevant behavioral processes must be understood before one can explain them at the reductionistic level that is physiology.

    Transcript: Research in songbirds like zebra finches has unearthed important clues about the brain circuits underlying vocal learning. Native to Australia, the zebra finch is a stocky little songbird. With bright orange cheeks and black and white striped feathers adorning his breast, the male zebra finch is an ostentatious entertainer, belting out tunes in hopes of impressing a female. But he wasn’t born with such a beautiful singing voice — he acquired it through learning and practice.
    Zebra finches begin their vocal training with indiscriminate chirping. After listening to adult finches, they slowly progress to singing full tunes, just like babies progress from babbling to uttering complete sentences by about 3 and a half years of age. And just like in human babies, there’s a critical window of time during which a young finch must hear adults sing so he can learn to make those same sounds.
    Whether it’s human language or birdsong, producing sounds is a motor skill.

    GS: In particular, it is operant behavior, shaped and maintained by its consequences (by definition).

    Transcript: Anything that requires you to consciously move your muscles is a motor skill, like running or swinging a bat.
    GS: Nothing is gained by using the term “consciously” or “voluntary” for that matter. “Motor skills” are operant behavior and they are “skillful” precisely because the interplay of reinforcement and extinction results in the final form.
    Transcript: Even getting out of bed in the morning and then walking downstairs. And motor skills, as we all know, take practice. It’s a process of trial and error that requires sensory feedback from—wait, let’s back up a little bit.

    GS: Nothing is gained by the use of term “trial-and-error” – it’s ambiguous and vague. I have already discussed how to talk about “skillful” behavior.

    Transcript: Outside of your brain and spinal cord there are 2 kinds of nerves: motor nerves and sensory nerves. Extending from the spinal cord, motor nerves signal your muscles to move. Sensory nerves, on the other hand, send information about sensation — touch and position to name a few — back to your brain.
    Motor skills — like mastering a wicked backhand in tennis, playing the piano without missing a note, or learning how to speak — depend on this kind of sensory feedback so we can adjust and perfect our performance through trial and error.

    GS: Birdsong, like language is complicated operant behavior. Period. I have omitted the drivel about how the brain mediates these behavioral processes. Contrary to popular belief, and fueled by the endless “neurohype” that constantly bombards us, we are not close to answers to questions like this. We can’t even give complete answers concerning very simple behaviors in invertebrates and pointing to a few brain areas and circuits and claiming that that is particularly meaningful is a symptom of the conceptual cesspool that characterizes cognitive psychology and the fields (like much of neuroscience) it has corrupted.

    Cordially,
    Glen

    Reply

Leave a Reply

%d bloggers like this:
Skip to toolbar