If you don't know Bunny,

she's a dogfluencer with a book.

And while she didn't write it, Bunny does communicate,

with the help of these buttons.

- Mom. Love you.

- These buttons not only launched

a ton of other dogs to fame,

they divided the internet into two camps.

The, of course dogs can talk group,

and the, this is faker than "Lassie" group.

- Lassie!

- People have fantasized

about talking to animals for a long time,

and we constantly pretend that our pets are people.

So when I see something like this -

- [Button] Mom is sick.

- it feels like we might be reading

a little too into things.

Are dogs actually capable of learning language

or is this just more social media fluff?

(eye chimes)

Luckily, a UC San Diego cognitive science team

is trying to figure that out.

They've risked their reputations

to study this viral phenomenon by reopening the messy

and complicated history of animal communication studies.

But finally, science is weighing in

on the great TikTok talking dogs debate.

(bright music)

It's no secret that dogs can understand verbal commands,

sit, shake, spin around.

This good girl could identify over a thousand toys by name,

but these buttons are something entirely different.

They're trying to let our dogs talk back to us.

- [Button] Water, water.

- For some people you want to understand more

about what's going through their mind,

what they're experiencing, their emotions.

- This is Federico Rossano,

a cognitive science professor at UC San Diego.

He's the leader of the Animal Communication Project,

which is more controversial than it sounds.

A few years ago,

Federico's colleague sent him videos of Bunny

and told him he should study this extraordinary dog.

- Familiar with the history of the animal language studies

and all the pushback that the scientific community

and the public had provided,

I did not want to deal with this.

- But Federico eventually changed his mind.

Before we understand why,

let's talk about how these buttons work.

Meet Donut, one of our office dogs.

Donut has a lot to say,

so we're gonna try to teach him how to use these buttons.

Training on the buttons is typically done by a pet's owner.

You start off easy, teaching a familiar word like toy

or treat through repetition.

It's harder than it looks,

but eventually Donut may start to request his favorite food

by pressing the button.

Federico's data shows that the median number of buttons

a dog has is nine.

For most, that's all they need.

But like with humans,

- Psammophile?

- [Examiner] That is correct. (audience applauding)

- there are a handful of high achievers,

dogs with around a hundred buttons,

including emotions and abstract concepts.

One day, Donut, one day.

The concept isn't all that different

from teaching your toddler new words.

- Bunny, very early on,

had a phase in which she would constantly talk about poop.

My 2-year-old is going through that exact same phase,

and it's all about toots and poop jokes.

Now you can look at this and say, "Whatever,"

but when you have several dogs that go through a poop phase

where they just only talk about poop,

and you have a 2-year-old that goes through the same phase,

it makes you wonder.

- But before we try to study

our dog's sense of humor,

Federico's team needed to do some basic research

to figure out what's going on here.

So the team has collaborated with FluentPet,

one of the many button-making companies,

to analyze their massive trove of user-submitted data.

Critics have accused this research

of being nothing but a way to sell buttons,

but Federico isn't promoting any products.

- We are getting no money out of this,

absolutely no money out of this.

We're getting a lot of grief in many,

in many ways from the scientific community,

but we do deeply believe that because people are doing it,

somebody should study it

and it should be studied

in the most systematic way possible.

- So they're attempting to figure out if

and how these dogs are communicating with us.

With over 10,000 button-pushing pups from 47 countries,

his group has created the largest citizen science

animal communication study, ever.

Having this data is what changed Federico's mind

because it's so different

from how things were done in the past.

- [Narrator] Why should two comparative psychologists

suddenly wanna start teaching sign language to a chimpanzee?

In the 1960s and 70s, researchers began doing experiments

that felt more like stunts than science.

One built a giant tank in a house

to live alongside a dolphin

and teach them how to mimic human speech.

Spoiler, it didn't work.

Then there's Nim Chimpsky

who was kind of taught sign language.

Scientists named him after the famous linguist,

Noam Chomsky, poking fun at his view that no other species

but humans were wired for language.

But the study had so many scientific and ethical problems.

First off, Nim was raised like a baby

alongside a human family.

Nim worked with 60 different trainers,

and most of them didn't actually know sign language.

They had him smoke weed and drink alcohol.

One of his handlers even breastfed him.

All of this was traumatic for Nim

who ended up with severe emotional issues

and died at 26, about half of a chimp's life expectancy.

Institutions eventually valued animal welfare more,

and by the 1980s,

major funding for this kind of research vanished.

Since then, it's been scientifically taboo,

but the UC San Diego research is different.

Dogs have co-evolved with people for thousands of years.

Part of what convinced Federico to start this project

was that it wasn't harmful to the animals.

The dogs could stay in their natural habitat, our homes,

and be trained by their owners.

And scientifically,

the project's gigantic sample size is unprecedented.

Most other animal communication studies

focus on just one animal,

and they've all been plagued by this bias

known as the Clever Hans effect.

Before there was Bunny,

our great-grandparents had Clever Hans.

In the early 1900s,

an amateur horse trainer had a brilliant idea.

He would teach his horse, nicknamed Clever Hans,

to do incredible things like math.

- Hans, what is two times two?

(hoof banging)

- People loved Hans.

He was covered in "The New York Times,"

and "Scientific American,"

but his abilities were eventually debunked.

It turns out that Hans

would read his trainer's body language.

- Hans, what is one plus one?

(hoof banging)

- It's like if a dog's owner

pushes the button outside

and naturally points to the door.

If the dog goes to the door, what are they responding to,

the button or the point?

How do we know that's not happening here?

Federico needed to rule out the Clever Hans effect.

So the research team went into the homes of 23 dogs

and conducted a simple experiment.

The study's lead researcher, Amalia Bastos, wore headphones

and pressed a covered button so that she didn't know

what word was being pressed, this way, there was no cueing.

Then she just stood there.

What they found was that the dogs responded

to the word produced by the button,

like grabbing a toy when the button, play, was pressed.

Of course, not every time,

but enough to show that the dogs can respond

solely to the button, ruling out the Clever Hans effect.

- And so it kind of became a green light for us to move

to the next study that was looking into

how the dogs are actually using these buttons.

- Federico's team selected 152 dogs

who use the buttons regularly,

with more than 260,000 buttons pushed over 21 months.

They analyzed multi-sequence button presses,

logged through an app,

and when available, video submitted by the owners.

One of the biggest findings from this study is

that the dogs initiated the majority of these interactions,

suggesting it's a useful tool for them to communicate.

And while a few of the dogs

seemed to just be randomly smashing buttons like this,

a majority of them were using them intentionally.

They were significantly more likely compared

to the button-mashing dogs

to communicate something like

"want outside" or "want treat."

These dogs also push certain buttons

more often than their owners.

- One of my favorite is a button that the dog tends to press

significantly more than chance is

"now" while the button that the owner tends to press

significantly more than chance is "later," (chuckling)

so clearly they're not matching

what the owner is pushing here.

- This research is a huge deal

because it shows that dogs are communicating

with these buttons,

and this discovery opens up so many new questions like,

are dogs capable of learning language?

Federico's team can now study this fascinating concept

called productivity,

where you combine two words to refer to something new.

- One of the dogs, for example, went to the soundboard,

the human was distracted, went to the soundboard

and pushed squeaky car.

They have a word for squeaky, the toy that makes noise,

car is something they have,

and then when they went to the human looking at it

and the human was like, "What is that?"

And then they looked outside the window,

and there's an ambulance,

oh yeah, an ambulance, squeaky car.

And all this, of course, is very social

and tells us that these animals might have

a deeper inner life than we give them credit for.

- We can learn a lot from studying animals

in their natural habitats, and it isn't just dogs.

Take Annie, a peregrine falcon

who found a prime piece of real estate

here in the big city.

Thanks to some well-placed webcams,

she and her family have became beloved Berkeley celebrities.

But why did she raise her chicks here with all the people

and noise and traffic instead of the typical cliffside nest?

We've got a video all about that

and how these falcons beat extinction.

And if you want more science

and research breakthroughs, be sure to subscribe.