Kind: captions
Language: en
A song sparrow will have like
or
wait a minute. You speak bird.
>> I tried a whistleb bird. Yes.
[Music]
>> Eric, welcome to Particles of Thought.
You do some fascinating research
>> that involves understanding how the
brain works, understanding how language
works, and you study nonhuman animals to
figure out how humans work.
>> Yep.
>> Yeah. So, go into that a little bit for
me. Yeah.
>> Yeah. So, the brain does many
fascinating things and humans have one
fascinating thing that the brain does
that you don't find in many other
species and that is uh language,
particularly spoken language.
And um uh one of the components that
give rise to spoken language is the
ability to imitate sounds.
>> Uh the words I'm saying now are imitated
sounds. And guess what? There are very
few species that have it, but they're
out there.
>> Parrots is one of them. Many people know
that parrots can imitate. Songirds as
well, hummingbirds,
>> right?
>> Amongst birds and amongst mammals
besides just humans, we got dolphins,
whales.
>> Yeah. uh sea lions and uh bats.
>> Oh wow.
>> And some say elephants as well. And and
then that that's it that's been found so
far.
>> That's a bigger list than I think most
people would appreciate.
>> That's right. Yep.
>> So all right. So we have for example
dogs that seem to understand certain
vocalizations that we make. Yes.
>> But we don't ever expect for them to be
able to vocalize themselves. Right. So
is is it completely two different areas
of the brain?
>> Yes. See this is what a lot of actually
not even people in the public but
academic scientists don't get you know
that that when we talk about language
uh as I was saying earlier there are
multiple components and a lot of people
think all those components if even if
they think about them they think that
they're all unique to humans
>> they're not
>> and so the ability to understand spoken
language actually is more widespread
>> than people think.
>> Yeah. Um, and you just got to go to your
pet animals, particularly dogs,
>> right?
>> Dogs can understand the word sit.
>> Yeah.
>> Or ci in Spanish or in Japanese. Dogs
aren't born to learn the to understand
Japanese or English.
>> Right. Right. Right.
>> Um, or even whole sentences. Get the
newspaper, get my ball, you know.
>> Wow.
>> Want to go outside, want to be petted.
>> Yeah.
>> They understand all these words. Several
hundred, you know, words. interesting.
>> Um, but they can't say it,
>> right?
>> You know, the best you can get out of a
dog is trying to modify its vocal tract
to say,
you know, right?
>> And that's that's that's good,
>> you know, and I would consider this like
a semicontinuum of vocal learning. They
have very limited ability,
>> but they did what you you brought up
earlier, which is imitating sound,
right?
>> Yeah. Yeah. getting close to to saying u
uh something that sounds like an
imitated sound but not full-blown
imitation like we see in humans and
parrots and song birds. Right. Right.
>> So So you asked about the brain
pathways. Yes. Right. And and when you
look in the brain,
>> the brain areas involved in
understanding speech
>> that you know you're hearing brain
circuits
>> is separate from the brain pathway
producing that speech.
>> I see. And when we look in the brain of
various different species, we humans,
parrots and songirds
have brain areas for both hearing and
processing speech-like sounds and
producing them.
>> But dogs only have the brain areas for
hearing them.
>> Oh.
>> They don't have the brain pathway that
goes to the larynx to control the
imitation of those sounds.
>> Wow. Okay. So, they're just missing it.
>> They're missing it. Or it could be very
rudimentary. We think we found a very
rudimentary
>> uh circuit in mice. So I wouldn't be
surprised if it's there in dogs, but
it's it's it's not this full-blown
advanced circuit that we find in uh
humans and the other vocal learning
species.
>> Okay. So that brings me to another
research study
>> that I heard I think a year ago. Uh I
don't think the result was a year ago. I
think I heard someone speaking of it a
year ago. And what that was is that it
was originally thought that chimpanzees
don't have the mechanical equipment for
producing speech, but then some study
found well actually they do. They just
don't have the brain for
>> That's right. That's right. Yes. Okay.
Yeah. Yeah. This has been a debate over
a number of years. Uh and it's it
includes you know what what what in the
in the biology in the brain in the
muscles and so forth makes speech
special,
>> right? And there are a number of
different hypothesis and this is one of
them
>> that there's a difference in the
musculature
>> of the larynx or that it's more
descended in humans
>> and allowing greater air space in the
vocal tract to produce a greater variety
of sounds.
>> A lot of this has been proven not to be
the case.
>> Ah
>> all right. So um a colleague of mine
Tukuma Fitch
>> uh blows air through post-mortem
you know larynxes of humans
>> of chimpanzees of monkeys
>> and you get similar kinds of sounds.
>> Oh wow.
>> That you get in speech when you you know
phonms we call them different kinds of
phonms that you get when you blow air
even through a postmortem larynx.
>> Wow. But um but when it's inside the
living animal of a of a monkey, you you
can't get the speech sounds out of it.
And why? Because it's not in the
muscles, it's in the brain that makes
that difference between being able to
speak
>> and not speak.
>> So is it the case that you know you have
birds, you have mammals, you know, a
different part of the family tree. So
are the same
outcomes being achieved by different
brain areas or is it that
>> there's some ancient ancestor that
predates birds and mammals that had this
particular equipment and now uh it's
being used in different ways to get the
the the
>> well well discovering these gene
specializations uh where the regulation
is different helped us answer some of
the questions that you just asked.
>> Okay. And putting it all together, the
story that you know I've come up with is
that um
all vertebrates have the ability to
learn how to move.
And what happened in us humans and the
song learning birds, dolphins I believe
as well is that these brain pathways
that control learned movement of the
body are are duplicating themselves to
control the hands, the chest, the feet
and so forth during embryionic
development. And in us humans and the
vocal learning birds, the brain pathway
duplicated one more time
>> and now got hooked up to the muscles of
the larynx and the jaw and so forth to
control uh sound production to learn how
to imitate sounds. And we already have
auditory input from the hearing pathway
going into the movement pathway. If you
got to you're going to learn how to
move, you got to learn to move to sounds
that you hear, learn to move away from
sounds and so forth,
>> right? And so so this auditory input is
already happening.
>> Yeah.
>> Uh the ability to understand complex
sounds is already there.
>> Yeah.
>> All right. I think consciousness is
already there.
>> And yes, language evolved more recently.
Spoken language. Uh so uh we call this
bra uh brain evolution by brain pathway
duplication.
>> Oh
>> like gene duplication. Yeah. A whole
brain pathway duplicates, connects to
the vocal organs.
>> Wow.
>> And now you get spoken like speech
basically.
>> Holy cow. Talking about taking something
that already exists and
>> Yeah. Yeah.
>> And so this happened multiple times. Uh
and when each time that it happened, a
certain set of genes change in their
regulation
>> in humans and in the song learning
birds. We call that convergent
evolution. Right.
>> All right. Because their closest
relatives like us for chimpanzees,
we are their closest relatives. I mean,
they're our closest relatives too, you
know. So, chimpanzees living surviving
surviving relives.
>> Yes, that's right.
>> Um, let's say what we call subassene
birds for songirds or falcons for
parrots. uh none of them have these
brain pathways or this gene expression
specializations
>> that we see in uh in the vocal learning
birds. So we think the whole brain
pathway duplicated the gene
specializations uh then were evolved uh
in a convergent manner.
>> I see.
>> And so it suggests that if vocal
learning and spoken language were to
evolve another half a million years from
now, say in a crocodile
>> Yeah.
>> Right.
>> Yeah. or a chimpanzeee,
>> right?
>> I could tell you what the brain pathways
are going to look like and I can tell
you which genes are going to change.
>> You It's predictable.
>> It's predictable.
>> Wow.
>> That's right.
>> Because you know what the the the
rough material they have to start with
is already.
>> That's right. Well, because we we've
already seen it multiple times in
multiple spec lineages of species. um
that there seems to be a basic principle
of a a fundamental set of genes
>> that's going to that needs to change in
order to get vocal limitation.
>> Right.
>> All right. If and in in the songirds
that was like 30 million years ago.
>> Yeah.
>> In parrots that was 50 million years
ago.
>> In humans it was the most at a million
years ago. Okay. with I think with some
common ancestors with Neanderthal and
Denisovven archaic humans.
>> So it would have been Hobo erectus then
that that had language first.
>> Um
I we we might go that far back that
would be a prediction.
>> Okay.
>> But certainly you know uh uh the
hominids as we call them.
>> Right. Right.
>> Uh which are modern humans and archaic
humans.
>> Right. Right. Right. Wow. So you brought
up this idea between hearing M
>> speech and motion.
>> Mhm.
>> So this is something my niece was born
in 1984 and I was in high school
>> and you know what in rap was new.
>> Y
>> so I used to sit around
>> beating out beats and my little niece as
soon as she could sit up she would bob
and dance to the little beats I was
making. And at the time I was a high
school musician as well. It occurred to
me I was like what is this thing with
music? Why is it that we respond
in such a way? We get emotional. Our
bodies move almost without us thinking
about it to these sounds. Like what the
hell is going on? Why? Because it seems
like a lot of species are completely
they don't respond that way. Yeah.
Right. Right.
>> Yep.
>> Have you guys figured that out? What the
>> Yeah. Uh what one of the remarkable
things discoveries in the last you know
10 to 15 years is that um it's been
found that only vocal learning species
can learn how to dance
>> and when I say dance I mean rhythmically
to a beat of sound in music. Uh and why
is that the case?
>> Yeah. Um once uh I in order to evolve
the ability to imitate sounds, you need
the auditory input going through your
ears to have rapid integration with the
movement pathway that's controlling your
muscles to produce those sounds.
>> All right? You need that tight auditory
motor integration as we call it. I think
that once that tight integration
occurred for the vocal organs, it
contaminated the rest of the movement
pathway.
>> Wow.
>> To now process sound
uh in a way that controls movement or
influences movement of the other organs.
>> Wow.
>> Of the other body parts or muscles that
control the body, the the arms, the
legs. And so yes,
you know, you know, your little niece,
right, is has something special going on
that they find in these humans and uh
and other vocal learning species. And so
I think it's a side effect of having
vocal learning. The ability to dance
came about because of our ability to
speak.
>> Wow.
>> Yeah.
>> That is so deep. And what's what's
remarkable
uh is that there are several kinds of
abilities that came along either for the
ride or that are evolved that are
correlated with our ability to speak.
>> And that's one of them is the ability to
dance synchronously to a beatum music.
>> Another is problem solving.
>> Interesting.
>> We found that the more advanced vocal
learning abilities you find, let's say
in the songird species, the better you
are at problem solving.
All right.
>> Interesting.
>> And so there's some cognitive uh ability
there. And putting all this together uh
these abilities that are special in
vocal learners, I call it the uh vocal
learning cognitive complex
>> and I'm including dancing in that
cognitive complex.
>> Wow. Wow.
This podcast is from the producers of
Nova. Nova is supported by Carile
Companies, a manufacturer of innovative
building envelope systems. With
buildings responsible for over a third
of total energy use, Carile's
energyefficient solutions are built to
reduce strain on the energy grid. For
example, Carile's Ultra Touch Denim
Insulation made from sustainable
recycled cotton fibers delivers energy
efficiency while being safe to handle
and easy to install. Made with 80%
recycled denim, Ultra Touch diverts
nearly 20 million pounds of textile
waste from landfills each year.
Operating across North America, Carile
is working towards a more sustainable
future. Learn more at carile.com.
>> Okay, so if you have if I give you a
random genome
>> and I say, can this species do vocal
learning? Can this species do auditory
learning? Can this species produce a
range of sounds?
>> Could you tell all of that just from the
genome?
>> We're trying to do such tests now.
>> Cuz here's where this is going, man. I'm
like, you know, we got these other human
species. Haididerberges, Dennisovven,
Florences, Deanderthals,
>> you know, you have access to their DNA.
So, you might be able to know whether or
not they had speech.
>> That's right.
>> And
even though it's make believe of getting
the DNA out of the mosquito and the
amber.
>> Yep. What if there were like dinosaur
species that could have similar
properties because birds descended from
them?
>> That's right. That's right. So, so to be
able to to ask such questions uh we have
to have the genetic data of all these
species.
>> Right.
>> And so I'm also help leading several
large scale projects
>> to sequence the genomes of all
vertebrate species on the planet like
all 70,000 and eventually sequence the
genomes of all 1.8 8 million ukareotic
species that have
>> Oh, wow.
>> nuclear membrane.
>> Right. Right.
>> And uh and from that those genomes,
we're going to be interrogating them. Do
do they have some of the same genetic
differences we see in vocal learning
species,
>> right?
>> Um right
>> to and so far we found that uh genetic
differences in humans and some songirds
that are associated with vocal learning,
we find them in the endol sequences.
>> Oh,
>> okay. So I I that's what makes me say
archaic humans, Neanderthalss,
uh probably had spoken language.
>> Wow. Wow. So, you know, that's you know,
when you when you branched up that the
the idea of dance coming from language,
it made me think, you know, cuz
sometimes,
>> you know, we have the
>> genetic story and we have the archa
archaeological story. And I remember
there being some statement of there
being a flute that was found and thought
to be Neanderthal, an ancient flute. So
does that equal if you have music,
>> then that means that you actually had
speech.
>> Yeah.
>> Yeah. So, so let me answer all those
questions I'm going to keep in my mind.
One is the music speech.
>> Yeah. I and a number of other people
think that vocal learning in humans and
somber other species evolved first for
singing
>> and for singing for the purposes of mate
attraction
>> territorial defense keeping a group of
of individuals in a cohesive
communicating manner and then later
became used for things like we're doing
now
>> right
>> like abstract uh communication
>> so singing preceded
speaking.
>> I believe singing preceded speaking and
that the same brain areas that control
speaking is used for singing as well.
>> The right side is more dominant for
singing, the left side is more that's my
left, sorry. The left side is more
dominant for speech and the right side
more dominant for singing.
>> Interesting.
>> So, so that was one question. The other
asked earlier about the dinosaurs,
right? And
>> and so I mentioned Neanthal genomic
sequences,
>> right? Yeah. Um, DNA in fossils or let's
say amber or let's or something frozen.
Yeah. You know, uh, in the t frozen
tundra up north,
>> uh, only lasts for a certain number of
years.
>> Yeah.
>> To go back to the time of dinosaurs, you
know, no one's found DNA yet.
>> Yeah. I I I know that it hasn't,
>> but let's say theoretically one day it
happens. Yeah.
>> And and we sequence that DNA.
>> Yeah.
>> Yes. I think we'll be able to tell you
whether that dinosaur was a vocal
learner or not. And is it possible even
without it? I the answer is yes.
Songirds are dinosaurs. Parrots are
dinosaurs. All birds are the only
surviving dinosaurs that survived the
last mass extinction. So if the bird
dinosaurs had vocal learning abilities,
why not some of the others?
>> Right. Right. Right.
>> Tyrannosaurus. I I that would be a
wonderful world, right? We went from
these uh depictions of dinosaurs that
were basically like, "Hey, let's just
wrap skin around the skeleton." That's
what they are. Then eventually, you
know, they're like, "Oh, maybe they had
fat." Then it's like, "Oh, maybe they
had colors and feathers." Right.
>> And now they're singing.
>> That's right. That is
>> That is amazing.
>> Yeah. We we humans tend to denigrate
other species whether living or ancient
species, extinct ones. Uh we we tend to
denigrate them to put ourselves at a
higher status, right?
>> But when you actually study them, you
start to learn that there many other
species that are smart, that are
good-looking, right?
>> That are, you know, have
>> abilities even beyond humans.
>> Yeah.
>> Uh and and that there's a greater
diversity out there than we appreciate.
And you can and if you don't come to
that realization, you're hurting your
ability to actually learn something the
truth
>> right
>> about nonhuman species.
>> You got to be completely open-minded,
right? You have to be completely
openminded, let the data do the talking,
and you know, hopefully you can ask
great questions.
>> Yep.
>> We now have the ability to edit genes.
>> Can you engineer speech into species or
vocal learning into species?
>> This is exactly what we're trying. Yes.
So trying it for several reasons. I
mean, of course, there's the cool
factor. Can you can you ap
Yeah.
>> Yeah. I'm not sure if that'll happen in
my lifetime, but
theoretically possible.
>> Yeah.
>> It's theoretically possible. Yes.
>> Right. Right. Um and so even even a
number of years ago when we started
coming up with hypotheses about how
could song pathways and song birds in
human speech areas uh convergently
evolve to function in a similar way with
a similar set of genetic changes and
what are the function of these genes in
vocal learning.
>> So um we would love to genetically
modify these genes in a human
>> and test what they do. Yeah,
>> that but that's that's tricky and
unethical to certain degrees.
>> Um, so and we would love to do it in a
songird as well, but the genetic tools
to manipulate genes in song birds is not
as advanced as we can do in mice.
>> I see. And so what we're trying to do
and what we are doing is taking a gene
variance that we find in humans that is
either unique to humans or unique to
vocal learning species and gene editing
them into the mouse genome. What are we
looking for? We're looking for uh
changes in the vocalizations in two
characteristics.
>> Okay. Um, our words are made up of
phonms. A, a, o, u, and we sequence
those phonms together to make words.
>> And then we sequence those words
together to make sentences.
>> Right?
>> We call those sequences and the rules in
which they're based syntax.
>> Got it?
>> Okay. Well, guess what? These mice and
other species, they have individual
phonms or syllables is another name
>> uh that we call them. And what is
learned is the sequences.
>> Okay. And you can change around the
sequences. Sometimes those sequences are
innate.
>> Yeah.
>> Uh and us, we can actually have learned
sequences of sounds to make words and
make sentences.
>> Right.
>> And so what we're looking for is changes
to those sequences
>> as well as to the individual structure
of the phonms.
>> I see. So let me ask you a question
about the sequences right quick. So if
there's a particular mouse that makes
vocalizations and you record that, is it
the case that you find repetitions of
the same sequence?
>> Yes.
>> Oh
>> yeah. Yeah. There. And they're
repetitions of the same sequences that
are innate.
>> So every other mouse will share those.
>> Every other mouse will share them.
>> Oh.
>> And what we're trying to do is to see if
we can get the mouse to learn new
sequences or change the acoustic
structure of each syllable within the
sequence. So is it is it incredibly
subtle where it's not obvious to the
ear? You have to do some take the
waveform of the recording and
>> the changes we're seeing thus far like
we're changing one gene at a time. It's
it's more in the subtle side.
>> Yeah.
>> But in the direction one would predict.
Ah
>> so like uh we recently published on a
working with Bob Darnell at Rockefeller
recently published on a study where a
nova one gene it's a what's it's a gene
that controls splicing
>> of cutting up RNA molecules and reputing
them back together. Um there's a human
variant that you don't even find in
Neandithol and we put this human variant
in the mouse genome and these mice start
producing more complex syllables. Oh,
interesting. Interesting.
>> Another gene called plexen
A1. It's a gene that controls
connectivity in the brain. Uh, and we we
>> connectivity between neurons
>> between neurons. And this gene actually
is turned down in the human speech motor
cortex.
>> All right. It's not one that's turned
up.
>> It's turned down. And when we when it
gets turned down in a in a
counterintuitive fashion,
it allows certain connections to form
from the speech brain areas to the areas
that control the vocal organ.
>> Uh so we we call it a loss of function
of the gene causes a gain of function in
the behavior.
>> Wow.
>> All right. It allows a certain
connection to form. We can see this
connection form in mice.
>> Yeah. And these mice too are producing
more complex sequences of vocalizations.
>> Wow. Wow.
>> We have not yet seen or thoroughly
tested can these mice imitate sounds.
That's our next step. But I do think
we're going to have to manipulate
multiple genes to get imitation.
>> Wow.
>> But that's also theoretically possible.
>> Geez. So do you have models that
basically you can uh play with the
various genes and predict the behavioral
output? Yeah, we have computer
algorithms that we developed that look
at the regulation of these several
hundred genes. If we were to tweak one
in one direction to or tweak it in
another direction, we can make
predictions.
>> Well, this takes us to a obvious
direction. Are we moving toward a future
where we have talking pets?
>> Is that
>> Yeah, someone asked me about that
recently. Another scientist.
>> Oh yeah.
>> You know, thinking about can we actually
do that? Yeah.
>> Right. Yes. eventually. Not today. But
>> yeah, and do people want to know what
their pets are thinking?
>> Oh, yeah.
>> You see, I I do think I do think once
you have the ability to imitate sounds,
>> right,
>> you have this inner speech
>> in your brain,
>> right?
>> And I I believe that the inner speech
brain circuit um is the same that's
being used to produce the sounds. M
>> uh so and and that is separate from the
auditory circuit that's hearing that
speech.
>> All right. So let me let me I got to
throw this in. You say you believe. So
does that mean that no one has stuck a
brain in a scanner and saw what lights
up when you're
>> No, that means maybe it's my cautious
scientific self is that this there's
debate.
>> Yeah, I see. If your consciousness, your
inner speech brain areas are um the same
as what's used to speak the sounds.
>> Is the debate based on data?
>> Yeah, debates based on data. Okay.
>> And I strongly favor, you know, the the
human fMRI studies
>> that show that the brain regions that
control speech production is the same
brain regions that is being lit up when
you're actually thinking in speech.
>> Right. Right. Yeah. Yeah. Yeah. Yeah.
>> And so, so you asked about our pet
animals, right?
>> Right. Yeah.
>> Um, the pet animals have the brain areas
that hear speech but not produce them.
>> Oh, that's right.
>> So, my I I do think if we can get our
pet animals to to speak through genetic
manipulations,
>> right,
>> uh that will both allow them to, you
know, to say what they've been thinking
in the hearing pathway, right? you know,
um, but it'll even give them a great
greater ability now to have inner
speech.
>> Oh, interesting.
>> And what we call conscious what your So,
if you want to know what your pets are
thinking by giving this them this
ability, you're giving them new
thoughts.
>> That's right.
>> Yeah. So, it's not like you're getting
that same animal. It's basically a new
animal now.
>> That's a possibility. That's a
possibility. That's right. Yes. Yes.
>> Wow. So, let's talk about this then. If
we could not speak and then we could
speak that pathway from not speaking to
speaking. So you say it goes from song
>> into speech partic potentially.
>> So how do how do you imagine that
playing out like like so so if we're
>> singing for mating purposes for you know
that's not the same as go get some
water. Right. Right. So, does it start
out as
go get some water, you know, like
>> it's like it's like look at me,
>> right? Right. That's where it starts,
right? And look at
>> I'm healthy and sexy.
>> Right. But how does it switch over to
clean your room, right?
>> Right. Yes. Yes.
>> Yeah. Like how do we imagine this?
Because again, when you think about
things that happen evolutionarily, you
think that there's some driver. Either
it's present preventing something bad
from happening or making something good
happen. Right. Yeah. So, how does speech
fit into there? Because
>> singing for reproduction is directly on
the evolutionary pathway. It's
reproduction, right?
>> That's right. Yeah. That's right. Yes.
>> Yeah. But speech is different. It's it's
one of those things that might I I would
imagine give you a slight advantage,
>> right? Right.
>> At the start.
>> Yeah. Yeah. You're kind of asking the
question, how do you go from like
Jennifer Lopez and Ricky Martin to
Einstein? Right. Uh
>> I don't get that analogy. What do we
>> Because because you have these these
musicians
>> Oh, right.
>> who are like uh you know fantastic
singers,
>> right? Right.
>> Uh who, you know, have let's say sex
appeal,
>> right?
>> Right. where the your that ancient um
ability to imitate sounds uh then gets
translated into now communicating
complex abstract ideas
>> like the theory of relativity.
>> Yeah.
>> So so that's the peril I'm trying to
bring out. That's
>> I hope those people can appreciate that.
Yeah.
>> So, um I I would say that uh
what is selecting for this ability to uh
imitate sounds and pass on this
information or these sounds culturally
from one generation to the next is uh
mate attraction is a strong selective
factor in biology. That's survival of
the species.
>> Yeah.
>> All right. Or survival of an individual
or population of individuals.
>> Right. Right.
>> And so anything that can help uh being
selected as a main
>> Yeah.
>> is going to be strongly selected for,
>> right?
>> And if imitation of sounds and
increasing the variety of sounds you
produce, you know, the like the more
information you have,
>> the more attractive you are.
>> Right.
>> Yeah.
>> And think about it. The more intelligent
you seem, the more attractive you you
>> Ain't that the truth?
>> Yes.
>> No, but it's real. That's a real
phenomenon. And so that's what's
selecting this vocal learning singing
ability.
>> Um and then after that uh just like the
dancing itself,
>> yeah,
>> you start to start to use this new
ability in new ways,
>> right?
>> Like for communicating concepts,
communicating information and passing
that information culturally from one
generation to the next.
>> Yeah. Yeah.
>> And and it does have a flaw. Uh and that
that flaw of the cultural transmission
of of vocalizations as you go further
and further away in geographic distance,
you start to get dialects. Right.
>> You start to get new languages, new
songs. Yeah.
>> Until you don't understand each other
anymore.
>> That's right.
>> This happens in all vocal learners.
>> Ah, all right. So,
>> all species of vocal learners.
>> All species of vocal learners. Yeah.
>> Yeah. as as as they move away from their
origin territory.
>> Yeah.
>> Their their repertoire starts to change.
>> Right. Right. Right.
>> Until they don't understand each other
anymore.
>> Wow. Yeah. So, what I I read a book many
years ago that basically said you don't
learn language. Language is an instinct.
>> Mhm.
>> Yeah. That's a little ex exaggeration.
Okay. All right. And and I've heard
people say that.
>> Yeah. Uh um so we have an instinct, we
have an innate ability to learn how to
imitate sounds, but the the actual
sounds that we imitate, the actual
words, the syntax, and the grammar is
something that is learned.
>> Got it.
>> And passed on culturally.
>> Got it. Got it. So what are they trying
to what are they getting at by saying
it's an instinct that
>> there there there's a genetic component
to to learn how to imitate sounds.
>> Okay.
>> All right. You you need the genetic
circuit I mean the genetic molecular
pathway right the genes that I talked
about earlier that are specialized in
their up or down regulation
>> that is genetically controlled if you
don't have that you won't learn how to
imitate sounds
>> got it
>> but the actual acoustic structure to the
sounds
>> like sit
>> right
>> ci say
>> that is that is learned it's not
instinctual
>> hey everyone if you're loving this
podcast please go ahead and like us or
leave a comment. And also make sure to
subscribe so you never miss an episode.
Your support means everything and helps
us reach more curious minds like yours.
Now back to the show.
>> So we know what learning language is
like or coming to speak is like in
humans, right? There's you mentioned
there's the instinctual component that's
in our genes and our hardware in our
brains. But then we have to learn the
actual vocalizations of our local
language, the accent of our local
culture. Um, so do the other animals go
through a similar process? Do they have
like coochie coochie coo baby talk and
then eventually they they're
>> Yes. So, so one another characteristic
you find in all vocal learners is they
go through a developmental period um
after hatching or birth uh that goes
through these critical periods in which
they learn how to imitate uh sounds.
>> Um in humans we start out what we call
babbling,
>> right? Uh and then we go through a play
phase where we have like we call it in
songirds like plastic song where um
we're starting to imitate but we don't
do it as well.
>> And then when we become a full-blown
adult we crystallize onto a particular
set of um vocalizations that are part of
our repertoire
>> that are full-blown imitations of what
we hear.
>> Interesting.
>> So you don't see this in a dog or
chimpanzeee.
uh they they're there they're there
small modifications to their
vocalizations over time but not some big
ones
>> I see
>> uh that that go you know through
juvenile development to adulthood.
>> Yeah. They don't go from ooh ooh ooh ah
ah ah to hello good day.
>> No that's right. That's right. Yes. Yes.
>> Yeah. It's more ooh ooh ah ah ah.
>> Right. That's right.
>> Yeah. Yeah. So birds though they're very
complex in in in their vocalizations. So
is does that mean that they and also I
would say the same thing with whales,
right? They're very complex.
>> That's right.
>> So they would have these critical
junctures.
>> That's right. These critical junctures
and they have to be exposed to your own
species. So
>> So what's also characteristic of all
vocal learners is that they're taken
care of. They're br um precosial, right?
where um they you know the young have to
be taken care of by the parents whereas
like chickens they can start walking
right away they grow up with parents for
a short period of time and then they're
independent. Yeah.
>> All vocal learning species aren't
independent until later in life.
>> Interesting.
>> And this is because you got to give them
time to go through these critical
periods and learn how to imitate the
particular repertoire in which they're
exposed to. So it seems like there
actually is some sort of communication
going on.
>> Yep.
>> That's similarish to the language we
have, right?
>> So how close have we gotten to
interpreting that? And do we see a day
where say for example with AI perhaps we
could actually you know if we have
videos of what they're doing and
recordings of what they're saying as
they're doing it could we somehow use
the equivalent of machine learning
techniques to actually understand what
they're saying as they speak in their
own way.
>> Yeah. So um I uh actually were working
with a group of people to do just this.
Well, you're already doing that.
>> Yeah, we're going to start. You know,
we're collecting up sounds of a whole
repertoire of a species, sets of
vocalizations from different parts of
the world where the species lives
>> and then try to use AI tools to try to
infer uh is there grammar to it? Is
there is a fundamental syntax? Um
>> is that not limiting? Like is because
that's how we think of our speech. Is it
is it
>> is that the right
>> Well well I would say searching for
patterns.
>> Got it.
>> Searching for patterns.
>> Right. Right. Right.
>> And in searching for those patterns, you
know, we might find some principled
rules
>> uh that you know hopefully are unbiased.
>> Right. Right. And do you have that
metadata of the circumstances that these
vocalizations match?
>> Yeah. Yeah. So so there it it can be
done with AI tools. You need a lot of
data,
>> right? That's the way AI tools uh work.
>> We have data from all over the world.
>> From animals in the wild, animals in
captivity.
>> Interesting.
>> Yeah. Yeah. Yeah. So you you got a lot
of data. You
>> we have a lot of data. Yes. Yes. And we
need to generate more data as well. But
>> uh so so the so the point being that um
we humans
think that other animals aren't so smart
because what they're saying doesn't make
sense to us.
>> Right.
>> But they could be communicating in ways
that we just don't realize that is more
complex
>> than what it seems like to the ear.
>> Right. Right. Uh,
>> and there's some species that have like
thousands of songs, right?
>> Oh, wow.
>> And there's some that just have one.
>> Yeah.
>> Right. Okay.
>> That are learned,
>> right?
>> And so, people have avoided those
species that sing thousands of different
kinds of songs.
>> Uh, or even parrots. You know, some
parrots have these warblelike song
vocalizations that seem to go on
endlessly
>> uh with various different combinations.
And so there's more complexity out there
to the vocal learning animal world,
>> right? We So it sounds like we haven't
really even
gotten deep into the data that could we
could be exact.
>> That's right.
>> Yeah. Ju just think about hearing a
totally foreign language to your your uh
you know perception one day.
>> Yeah.
>> Do you think you're going to understand
what's going on?
>> No. No.
>> No.
>> No. Not at all. And that's and that's
with my own species.
>> That's right. That's right. So much less
another species. So, right.
>> So, I just think we don't have the the
computational or other kind of tools yet
to do this, but with AI, it's maybe
coming close.
>> Yeah. Yeah. Yeah. So, you know how
there's a this recent study of the last
universal common ancestor
>> where they were like, "Oh, okay. It
lived at this time. It was a anorobic
bacteria. It already had an immune
system."
>> Just by looking at what's common
>> among various species.
>> Yep. Has there been a study where we
look at vocal learning species and
non-vocal learning species to identify,
you know, when perhaps these vocal
learning genes showed up in the animal
genome?
>> Yeah, we have some preliminary data from
already sequenced genomes
>> uh to indicate and and a fi a family
tree of species like of birds to
indicate that vocal learning showed up
somewhere around 30 million years ago in
song birds. Oh,
>> and about 50 million years ago in
parrots.
>> Oh, yeah.
>> U because we sequenced different parrot
species. And the cockapo from New
Zealand,
>> only 200 animals left on the planet, by
the way.
>> Oh, boy.
>> Endangered species of flightless parrots
>> are at the base of the family tree of
parrots.
>> Oh, interesting.
>> And we find uh in their genome
>> genetic changes we see in other parrot
species uh that learn how to imitate
sounds. Not as advanced as the other
parent species but but still there right
>> but we don't find them in their closest
relatives who are the falcons.
>> Ah
>> or what we call also another group of
who are non vocal learning.
>> So you know when they split we know when
they split. You can Yeah. Wow. Yeah.
>> Yeah. What about like in in mammals? Is
there a
>> Yeah. So
>> earliest known case?
>> Uh well because we're finding some of
these genetic changes in Neanderthalss,
>> right? uh you know we're saying at least
uh these genetic changes associated with
speech went back 500,000 years ago
>> in humans but with in whales it has to
be
>> in in whales I think it's going to be in
the 50 million year range.
>> Wow.
>> Yeah.
>> And these are all convergent evolution.
>> And what what's all convergent and
what's interesting is that all these
vocal learning species evolved after the
last mass extinction.
>> Oh interesting.
>> Yeah. Okay. So there I think there's
something that after the last mass
extinction 65 million years ago, there's
something ecological niches that opened
up that influence greater diversity and
speciation that may have selected for
vocal learning. When we humans are just
the last ones on the planet to evolve
vocal,
>> we're always late, man. You know, when
they when they divide the earth in the
24 hours, you know, we show up at the
very end, right?
>> We take credit for everything.
>> Well, we're different, you know. We're
different, you know. I guess that would
be a great question, like what is the
newest species to evolve on Earth,
right? It ain't us. But that's the other
question I have, right? If you look at
birds, they're super ancient.
>> If you look at humans, we've been
rapidly evolving for, you know, a few
million years. You take Homo erectus,
right? Their brain size doubles. And I
guess a question I have is that brain
size doubling,
could language be a driver of that or a
result of that?
>> Yeah. So um ju just to put uh make this
comparison more equal. So birds birds
are part of an order that is pretty old.
>> Yeah.
>> You know 100 150 million or more years
ago.
>> Yeah.
>> All right. Are the are the first birds.
uh primates to which humans belong is
the equivalent order we would call it.
Primates are more recent.
>> Yeah.
>> All right. Roughly around the time of
the last mass extinction. So 60 million
years.
>> So birds been on this planet two times
or more as primates.
>> Yeah.
>> And then amongst the primates,
>> right?
>> Um voc we vocal learning primates us
humans
>> is let's say I said somewhere in 500,000
to a million years ago.
>> Yeah. That's that's when we evolved this
ability I believe. Yeah.
>> Uh but birds you know at least among
songirds is 30 million years ago. So 30
times as long.
>> Wow.
>> Yeah. So so your question was
>> before I get lost in that sunshine.
>> It it led to some good stuff whatever it
was. Right. Well, I guess the question I
had is is in the one case, you know, I
don't know how much birds have changed
over the
the time period that they've been
burned, right? But humans have been
rapidly evolving. And one of the things
that happened during Homo erectus, which
preceded this all of us homo cousins
popping out, was their brain size
doubled over this short period of time.
So, I'm wondering if if language was a
driver of that, a result.
>> I think so. Yeah.
>> And and and here is, you know, my
explanation for it is that we humans do
have a larger brain size relative to
body size
>> um than other not only primates and
other mammals.
>> And uh and why is that the case? Some
people argue, well, it's because of
greater intelligence. We got to you have
you need more brain to have greater
intelligence and so forth. I I do think
that the speech brain circuits take up
space.
>> All right? and they evolved out of this
motor learning circuit
>> uh that controls movement,
>> right?
>> Um I don't think we lost brain tissue
because of this new brain circuit. We
just added to it and increased our skull
size a little bit more.
>> Uh it's not the only thing that I think
is taking up greater space. I do think
there are other parts of the human brain
that take up more space, but the speech
circuit I think is the biggest.
>> Right. Right. And I and I would imagine
that if it turns out to be
evolutionarily
advantageous then it would continue
along that path and perhaps get bigger
and bigger.
>> That's I think so. Yeah. Um but there's
there's only uh so much size that your
body can accommodate, you know, for a
heavy head,
>> right?
>> All right.
>> The bigger the brain, the bigger the
skull, the more skull you got to carry
around. Well, you know, now that we're
all like, you know, bending over and you
know how they say that that makes your
neck have to support more. Our brains
have to shrink now to accommodate.
>> That's right. That's right. So, our
smartphone so your head doesn't fall
down too much. Yes.
>> Now, that's humans,
>> right?
>> Okay. What about birds,
>> right? Well, it turns out K colleague my
her her Lano um she from Brazil she uh
found that um the brains of vocal
learning birds
>> like song birds and parrots are not only
bigger relative to body size like humans
>> uh but the difference in body size to
brain size is not as big as humans. How
did they compensate for that? They have
a doubling of neurons. They fit more
neurons
uh into a smaller brain space by
increasing the density.
>> So they have a more dense brain.
>> They have a more dense brain.
>> Wow.
>> More dense than humans.
>> Wow.
>> More dense than other birds. Yeah.
>> So So the total number of neurons in a
parrot brain is the same as a monkey
five times its brain size.
>> Why are you guys keeping all this stuff
secret? This is so cool. like I'm like
how am I hearing all of this for the
first time today? This is so amazing.
>> Yeah, you get some cool stuff out of
these birds. That's why it keeps me
going and studying them.
>> So, so I I do think that um this has
allowed parrots and songirds to now have
this new vocal learning system which I
can physically see with my own eyes when
I look at those brain areas. Take up a
lot of space.
>> Wow.
>> In the brain.
>> Wow.
>> Yeah. Yeah. And is that the same thing
uh with the the whales? Do they also
have uh do they have denser brains? I
know they have big brains. I know the
sperm whale has the biggest brain on the
planet.
>> That's right. Yeah,
>> that's right. So they and the and they
have big brains because they have big
bodies too, right?
>> You know, but they do have large brains
relative body size. Yeah.
>> But uh but are they extra?
>> We don't we don't know that answer yet.
I have somebody in my lab trying to
figure that out.
>> So I guess you need a comparison. Are
there non uh vocal learner whales that
exist? I just
>> No, we we we so far they look like
they're all vocal learners, right?
>> There are some that have producing more
simples vocalization, some producing
more complex.
>> Yeah.
>> But when you look at them and study
them, you know, purposely for this
trait, you find that they're vocal
learners.
>> All right. So, this is a this is off the
beaten path
>> but on the path.
>> We're talking vocal learning, but we
humans use symbolic learning, writing,
right? We use writing as well. So, is
there anything similar to that? And and
and is there a um the brain circuitry
that allows for that? Is that built It
has to be built off of something that
was
>> doing something else or pre-existed. And
so, the question would be, do the
questions, do we see that in any other
species? And do we see the circuitry for
that in any other species? I'm glad you
asked about the circuitry because I
don't like to give answers to fancy or
advanced behaviors unless I know there's
an underlying biological explanation in
the brain, right?
>> Otherwise, we could be making things up,
>> right?
>> Uh so far, there isn't any evidence that
other species
can write in the way humans do. They can
draw. You can get Well, let's say
non-vocaling species. Yeah.
>> You can get chimps and other primates to
push buttons on a screen.
>> Yeah. uh to spell out something. uh but
um the actual writing with the hand and
so forth
>> uh no and so I do think that is a unique
human characteristic that did come about
as a result of having spoken language
>> uh what's going on there uh in the brain
uh based upon now this is not work we've
did but this is work that I pulled
together from the literature
>> let me let me let me broaden it a little
bit right because there is that some
group that use knots as a written
language right and so you see that Birds
make these very complex structures like
bower birds. Yeah, that's right.
>> And different. Yeah. So, it seems like
even if it's not writing with your hand,
there are other ways to convey language
symbolically.
>> Yes. Right.
>> Yes. That's right. That's right. And I
think some symbolic communication like
sign gesturing.
>> Yeah.
>> Is more doable and doesn't require a
speech pathway to do it.
>> Right.
>> All right. So, uh, Koko, who is a
gorilla raised with humans for like 39
years,
>> was able to learn how to produce signed
communication in English.
>> Yeah.
>> All right. For hundreds of words, some
say thousands.
>> Wow.
>> All right. Whereas she was not able to
speak a single word. She can cough.
>> She can voluntarily blow air through the
vocal tract, but not say a single word
that she could sign.
>> Wow. So that ability to communicate
symbolically
is already there before speech evolves.
>> Ah
>> and speech is an expression of a
symbolic communication a