Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
This message comes from NPR
0:02
sponsor HubSpot. Imagine growing a
0:04
business with high quality leads,
0:07
fast closing deals, wildly happy
0:09
customers, and more benchmark breaking
0:11
quarters. It's not a miracle.
0:14
It's HubSpot. Visit hubspot.com to
0:16
get started today. This
0:19
message comes from NPR sponsor Synchrony
0:21
Bank, empowering you to tackle your
0:24
savings goals with flexible access to
0:26
your money and no monthly fees
0:28
or minimums. It's never been easier
0:30
to take control of your financial
0:32
future. Go to synchronybank.com/NPR. Remember FDIC.
0:36
You're listening to Shortwave from
0:39
NPR. Hey, Shortwavers. Regina
0:42
Barber here. So when
0:44
I think about whale songs, I think
0:46
this. But
0:50
not this.
1:03
That's a family of sperm whales. Here
1:05
today to tell us more about this
1:07
whale conversation is NPR's climate correspondent, Lauren
1:10
Summer. Hey, Lauren. Welcome back. Hey, Gina.
1:12
So, Lauren, I'm not going to lie.
1:14
This whale chatter kind of sounds like
1:16
bike spokes to me. Yeah,
1:21
I kind of get like Morse code
1:23
combined with microwave popcorn. Mmm, microwave popcorn.
1:26
I need to eat lunch. Okay. So
1:28
hearing these sounds makes me wonder, like,
1:30
are these whales really talking to each
1:32
other? And like, what are they saying?
1:35
Yeah. What is happening there? You're not
1:37
alone in wondering that. It's kind of
1:39
this age old question. Like, what are
1:42
animals saying? Right, of course. Yeah. And
1:44
especially whales because sperm whales have big
1:46
brains. They have close family groups and
1:49
they coordinate a lot. They
1:51
dive together. They hunt together. They
1:53
even babysit for each other. Shane
1:56
Garo, who is a sperm whale biologist who
1:58
has spent years with these whales. He
2:00
says he sees that kind of family
2:02
dynamic all the time. It's hard
2:04
not to see cousins playing while
2:07
chatting to not see mom's hand
2:09
over to a babysitter and Exchange
2:12
a few words before sort of walking out the
2:14
door so to speak to go eat in the
2:16
deep ocean Oh my gosh, I
2:18
can relate to this a lot as like the
2:21
older cousin babysitter and as a mom But
2:24
as someone who has listened to
2:26
like a lot of audio these
2:28
sounds sound pretty complicated to decode
2:30
They have so many clicks. Yeah,
2:33
it sounds pretty messy It's not
2:35
easy for us to figure out
2:37
but that's where computers are coming
2:39
in researchers are hoping artificial Intelligence
2:41
could tease out what the whales are
2:44
saying and as a first step They
2:47
figured out a sort of sperm
2:49
whale alphabet. Oh, okay an
2:51
alphabet So I'm this is starting to
2:53
sound like a language language. Okay, that's the
2:55
tricky word here. It's kind of a tough question There's
2:58
been a very heated debate for
3:00
years about whether animals can have
3:02
language or whether that's something special
3:04
that only we humans can claim So
3:08
today on the show how technology is helping
3:10
us figure out the mysteries of animal communication
3:13
And if we could figure out
3:15
what sperm whales are saying should we try to
3:17
talk to them? I'm Regina
3:19
barber and you're listening to show the
3:21
science podcast from npr Support
3:34
for NPR and the following message
3:36
come from the American Cancer Society.
3:38
Dr. Alpa Patel leads a team
3:41
that researches cancer risk factors, and
3:43
she shares how her team makes
3:45
an impact. We always
3:47
do what we like to think of
3:49
as actionable science. So the
3:51
work that we do makes its way
3:53
to things like nutrition and physical activity
3:56
guidelines for cancer.org, where millions of people
3:58
come each year to learn
4:00
about how they can
4:02
better prevent cancer. So
4:30
Lauren, sperm whales are somewhat famous for being like in
4:32
Moby Dick, but what are their
4:49
lives actually like? Yeah, so
4:51
sperm whales are diverse. They're
4:55
the size of a school bus, they're
4:57
kind of those big foreheads, for lack
4:59
of a better term, foreheads, and
5:02
they spend most of their time diving
5:04
in the deep ocean, searching for their
5:06
favorite food, which is squid. Me too.
5:09
They can go thousands of feet below the
5:11
surface, and so Shane told me they are
5:13
in the dark a lot. So
5:15
sound is everything to sperm whales. In
5:18
the darks of the deep ocean, these are
5:21
places where sunlight never gets to. So
5:24
they navigate their world
5:26
through sound, just like bats
5:28
in the dark sky, they use echolocation. And
5:33
they use sound to stay in touch with
5:35
one another and coordinate with the families in
5:39
which they live. Wow. Yeah, it's
5:41
also just hard to see underwater, right,
5:43
in general. So I can see how
5:45
sound would rule almost like everything. Yeah,
5:47
yeah. And they live in these tight-knit
5:49
groups. They're female led, so
5:52
there's grandmothers, mothers, and daughters. They all
5:54
stay together their entire lives. The
5:56
males get to hang out until they're adolescents, and then they
5:58
have to leave to head out on own. I
6:01
totally would love watching like these
6:03
whales family dynamics. Right, yeah,
6:05
because they live a long time. And
6:08
Shane studies these whales in the Caribbean
6:10
with the Dominica Sperm Whale Project. He
6:13
knows these families and they vocalize
6:15
a lot. They have long exchanges
6:17
with each other.
6:20
It's not rude in sperm
6:22
whale society to talk at the same
6:24
time and overlap one another. It
6:28
sounds like an extended family, like
6:30
loud summer barbecue. Yeah,
6:33
a lot going on. It sounds
6:35
kind of messy, but actually all
6:37
those clicks can be broken up
6:39
into patterns of clicks. And they're
6:41
ones that the whales use over
6:43
and over. They're called cotas. One
6:46
that's really common in the Caribbean
6:48
is the one plus one plus three
6:51
coda, which sounds like this. You know,
6:53
one plus one plus three. Sounds like a
6:55
salsa dance. Yeah, it's cha cha cha. Yeah.
7:02
So what sounds like a whole bunch
7:04
of like clicking is actually like these
7:06
discrete units, these discrete chunks. Yeah, yeah,
7:08
all kind of strung together and family
7:10
groups can have dozens of these different
7:12
cotas. They all have a different number
7:14
of clicks or a different pattern of
7:16
clicks. And
7:21
researchers like Shane, you know, they've been recording
7:23
these for years because they've been trying to
7:26
tease out the patterns and how
7:28
the whales use them. And more
7:30
recently, they teamed up with artificial
7:33
intelligence researchers in a collaboration called
7:35
Project SETI. And their goal
7:37
using that technology is to decode what
7:39
the whales are saying. So
7:42
AI has been learning human languages and
7:44
researchers are trying to like test the limits
7:46
of what it can do. But learning
7:48
whales seems really complicated. For sure. Yes.
7:52
You know, they kind of took this
7:54
first step where they use machine learning to
7:56
analyze more than 9000 recordings of sperm whales.
7:58
Wow. And Daniella who directs
8:01
MIT's computer science and artificial
8:03
intelligence laboratory, she told me
8:05
it was key to use computers because it could
8:07
find clicks that humans couldn't find on their
8:09
own. It really turned out that firmware
8:12
communication was indeed
8:14
not random or simplistic, but
8:17
rather structured in a very
8:19
complex combinatorial manner. They found
8:21
far more variation than researchers
8:23
thought there was. Like Shane says
8:25
sometimes it's the same coda, the same
8:27
set of clicks, but they make it
8:30
slightly longer. So that one plus one
8:32
plus three coda we talked about might be half a
8:34
second long or it might be 1.3 seconds
8:37
long, nearly three times as long, right?
8:40
And whale, the same whale will make short ones
8:42
and long ones and different families will make short
8:44
ones and long ones. Which kind of sounds like this.
8:52
That's subtle, right? Yeah, to us humans.
8:54
But whales actually pick up on these
8:56
differences and they even repeat them back
8:59
to each other. So, you know, sometimes
9:01
it's the tempo of the clicks that's
9:03
different. Sometimes the whales throw in an
9:05
extra click at the end of the
9:07
coda, Daniel says. And this was very
9:09
interesting. We started wondering is this extra
9:11
click sort of like a, the end
9:13
of a sentence or something else. And
9:15
doing this analysis, they identified what they're
9:18
calling the sperm whale phonetic alphabet, which
9:20
catalogs all these variations. It's
9:22
actually making me think like, what if
9:24
they're slang with the different family members?
9:26
Okay. If we're using like the word
9:28
alphabet, that again makes me think of
9:31
language. Like are all these different codas
9:33
different words? Are they like parts of
9:35
words? You know, like the sounds that
9:37
make up an alphabet. Right. And that's
9:39
what's hard to figure out because the
9:41
thing that researchers say is exciting is
9:44
that these codas don't seem to
9:46
be random. They can be
9:48
predicted by machine learning in
9:51
the same way in which you
9:53
might predict the sequence of syllables
9:55
or the sequence of words in
9:57
a sentence. She's saying there could be a
9:59
possible. ability of recombining these
10:01
codas to make meaning. And
10:04
that's something we do as humans in language,
10:06
right? We take sounds that don't really mean
10:09
anything on their own, like short,
10:11
right? Shh, ort becomes short,
10:13
and we combine it to make something
10:15
that has meaning. Okay, but like, just because you
10:17
understand like the rhythm of the cliques doesn't mean
10:20
you might understand like what they mean.
10:23
So could scientists understand what the sperm
10:25
whales mean? Like how could they prove
10:27
sperm whales are conveying complex things like
10:30
language? Yeah, that's really hard to do.
10:33
Shane says they're working on recording sperm whales
10:35
and observing their behavior at the same time
10:37
to kind of build up a data set.
10:39
But you know, it's kind of hard to
10:41
know if you're capturing their world and what's
10:43
important to a sperm whale in that moment. If
10:45
we only ever studied North
10:47
American English-speaking society in the
10:49
dentist's office, first of all,
10:51
we'd walk away with the fact that the key part
10:54
of their communication system is the
10:56
word root canal, right? And
10:58
we'd just be wrong because we
11:00
didn't have a comprehensive picture. Yeah,
11:02
I mean, that's a really good point. And those
11:04
are two words in a dentist's office I'm really
11:06
actually scared of. Yes, very much so.
11:09
And it kind of
11:11
just shows like we're used to looking at things
11:13
in a very human centric way.
11:16
And people have been debating this animal language
11:18
question for a really long time. It
11:22
goes back, you know, 1970s researchers
11:24
were teaching chimps and gorilla sign
11:26
language, you might remember. Yeah. And
11:29
the question was whether they were copying
11:31
us or really using language the way
11:33
we do. And there's a
11:35
lot of other examples like bees, you know, they
11:37
do that special waggle dance in a hive. It
11:40
tells other bees like how far the flowers
11:42
and what direction they are. And I talked
11:44
to Taylor Hirsch about this. She's a researcher
11:46
at the Marine Mammal Institute at Oregon
11:49
State University and she studies sperm whales.
11:51
Some of what they're doing may be
11:53
totally different from our way of communicating.
11:56
And we're probably never going to be able to fully
11:58
grasp those differences. So
12:00
I think there is value in
12:03
seeing as patterns in animal communication
12:05
near patterns in human language. But
12:07
I think it's important to remember that perhaps
12:09
just because we don't find evidence of something
12:11
doesn't mean that that system isn't complex in
12:13
ways that we don't understand. Right.
12:16
Like, can our human brains
12:19
comprehend whatever system these sperm
12:21
whales have worked out for themselves? Like, it
12:23
actually makes me think of, like, a rival,
12:25
that film. Oh, yeah. All
12:27
we really know is how our own
12:29
language works. Right. And that's, I guess,
12:32
where we start. Yeah. And researchers like
12:34
Shane, he agreed that looking for those
12:36
similarities is valuable. You know, when we can
12:38
talk about whales and how important
12:40
their grandmothers are or how important being
12:42
a good neighbor is or learning from
12:44
different cultures is and the importance of
12:46
cultural diversity in society, that really sort
12:49
of resonates with people and can
12:51
drive change in human behavior in
12:54
order to sort of protect the whales. And, you
12:57
know, sperm whales are still coming back
12:59
from commercial whaling where their numbers were
13:01
just decimated. Today, they face threats like
13:03
ship strikes and plastics in the ocean.
13:05
So, Shane said it's important to appreciate
13:07
what they share with us because we
13:10
have such a big impact on their
13:12
world. So if artificial intelligence
13:14
figures out sperm whale language, what's the
13:16
next step? Like, are researchers hoping to
13:18
talk to them? Yeah, it's kind of
13:20
an ethical question, right? Like, do you
13:23
play some sounds back to sperm whales
13:25
to try to say something to them?
13:27
What does that do to them, especially
13:29
if we don't really know exactly what
13:31
we're saying to them? Right. Like, are
13:33
we going to scare them? Like, like,
13:35
I would be a little worried. And
13:38
do we want to like hear what they say back
13:40
to us? Like humans, they might
13:42
not like us. Yeah, we
13:45
don't have a great track record.
13:48
Right. And, you know,
13:50
everyone I talk to in reporting this,
13:52
all these researchers told me they get asked
13:54
a lot of what they would want
13:56
to say to sperm whales. And Taylor
13:59
told me, you know, She's not like
14:01
raring to go on this. There's this
14:03
implicit like do I have the
14:05
right? What gives me the right to say anything
14:07
to them? I Mean
14:10
sperm whales have been communicating with
14:12
each other a lot longer than
14:14
humans have right? They've been doing it
14:16
before humans had language. So clearly
14:19
they've got it figured out on their own Wow,
14:22
Lauren, thank you so much for
14:25
communicating all this to all of us. Thank
14:27
you so much. Yeah, thanks Before
14:31
we head out a quick shout out to our shortwave plus
14:33
listeners We appreciate you and we thank you
14:35
for being a subscriber Here
14:37
we've plus help support our show and if you're a
14:39
regular listener We'd love for you to join so you
14:41
can enjoy the show without sponsor interruptions find
14:44
out more at plus dot NPR org
14:46
slash This
14:49
episode was produced by Rachel Carlson and edited
14:51
by your showrunner Rebecca Ramirez Lauren check
14:53
the fact Patrick Murray and
14:56
Stacy Abbott were the audio engineers Betts
14:58
on it in is our senior director and
15:00
Colin Campbell is our senior vice president of
15:02
podcasting strategy I'm Regina Barber.
15:05
Thanks as always for listening to sure
15:07
with the science podcast This This
15:21
message comes from NPR sponsor,
15:23
Viore, a new perspective on
15:25
performance apparel. Clothing designed with
15:27
premium fabrics, built to move
15:29
in, styled for life. For
15:31
20% off your first purchase,
15:33
go to viore.com/ NPR. Support
15:36
for Npr and the following message
15:38
come from our sponsor Whole Foods
15:40
Market. Whos that? celebratory brunch for
15:42
a less with three Sixty Five
15:44
by Whole Foods Market featuring while
15:46
it happy fines like hold smoked
15:48
Atlantic salmon and. More. Last
15:54
year over 20,000 People joined
15:56
the body electric study to change
15:58
their sedentary. Greenfield lives
16:00
in. Guess What? We saw
16:03
amazing effects. Now you can try
16:05
and br his body. Electric Challenge
16:07
yourself listen to updated a
16:09
new episodes where ever you
16:11
get your podcasts.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More