- 3 hours ago
Category
📺
TVTranscript
00:19You ever wonder what your dog is thinking?
00:24Water.
00:25Now, some dog owners say they know.
00:28Look.
00:30Thanks to buttons that seem to give dogs the ability to talk.
00:34Oh.
00:36It's like, how is this even possible?
00:40Outside.
00:41Some dogs are saying so much that scientists have decided,
00:44we got to study them.
00:46Button.
00:47Great.
00:47Yeah, so good.
00:48Go get it.
00:50We need to look into this.
00:53They've set up the largest animal communication study in history,
00:58with thousands of dogs from around the world.
01:05What we want to see is where can we take this and how far can dogs go?
01:08We know that our bond with dogs is incredibly tight.
01:12Oh, you're just so good.
01:13But could we really teach them to speak our language?
01:16So you have dogs that can translate on behalf of other dogs to humans.
01:22Correct.
01:22Correct.
01:22That is mind-blowing.
01:24She knows what she's saying.
01:26She's communicating with me.
01:30Couldn't it just be owners reading too much into things?
01:35What's really going on here?
01:38Are they just pushing our buttons?
01:40I love you.
01:42Or can dogs talk?
01:53The whole phenomenon of talking dogs starts with one very special dog.
01:59Check this out.
02:04This is Stella.
02:05It's late, but Stella isn't ready for bed.
02:16Stella appears to be talking.
02:18You want to go to the park?
02:20It's a little late to go to the park.
02:23This video has been sped up a little,
02:25but otherwise nothing's been edited in or out.
02:29Do you want to go for a walk instead?
02:31Do you want to go to the park instead?
02:34Yes.
02:39Walk.
02:41Okay.
02:43Okay.
02:44We'll go for a walk.
02:46Let's get your leash on.
02:49Leash on.
02:52Stella has been taught to use word buttons.
02:56Walk.
02:57Her owner, Christina Hunger, can record different words on each of the buttons.
03:02Stella.
03:02And by using them, Stella became known as the world's first talking dog.
03:09What?
03:10This is amazing.
03:12So this is the board.
03:13This is where it all happened.
03:14This is.
03:14That's right.
03:15How many buttons are there in total?
03:17There's 48 on this board right now.
03:20Like, why would you do this?
03:21Like, what would you make you think you could do something like this?
03:24So I was a speech language pathologist.
03:26I worked with kids every day who used communication devices to talk.
03:30And then when I brought Stella home,
03:32I just saw how much she was already communicating with her gestures,
03:35her nonverbal language, her barks and whines.
03:39And it got me thinking maybe if they had a different way to say words
03:43other than verbal speech that they could say it too.
03:47So, Christina got her some buttons to try.
03:51I wasn't just training her to push a button and giving her a treat.
03:54Never once did that.
03:56I was showing her how to use it in the right context
03:58and seeing if she would learn.
04:01Even when Stella became interested in the buttons,
04:04it took a while before she managed to actually press one.
04:09You're so close, Stella.
04:12But finally, she cracked it.
04:14And the next morning, Christina caught it on camera.
04:18Let's go outside.
04:20I was very excited.
04:24And then when she went outside, she went to the bathroom right away.
04:27She was using a word appropriately and functionally.
04:32I just jumped into gear and found more buttons and taught more concepts.
04:39Can you ask her questions and she'll answer using these?
04:42Is that the way usually it works?
04:44So, normally, she just uses them completely independently.
04:47Like, when she has something that she wants to say,
04:49she'll come over and use her buttons.
04:51Or if I ask, like, what she wants to do,
04:54she'll normally answer with an activity.
04:56If I ask her when she wants to do something,
04:58it'll always be now.
05:00Hi, Stella.
05:03As Stella started to use her buttons more,
05:05Christina began sharing her videos online.
05:09We left our apartment one day
05:12and we set up a camera just to see,
05:14you know, what she would do while we were gone.
05:23I had never heard her made a sound like that.
05:29It was heartbreaking.
05:36Stella blew up online
05:38and her videos turned her into a social media phenomenon.
05:41We're joined now by Christina Hunger,
05:44a speech pathologist in San Diego,
05:46owner of the very clever Stella.
05:49So, Christina, what does Stella say the most?
05:52She definitely says outside the most.
05:54She absolutely loves being outside.
05:57As news of Stella and her word buttons spread,
06:01others so-called talking dogs
06:03have flooded social media
06:05with their seemingly amazing feats of communication.
06:09Mommy!
06:11These dogs seem to be using our words.
06:16But is that enough to say that they're talking?
06:20Do they really understand what they're saying?
06:25Or is their button pushing just random?
06:29I love you too, Cobb.
06:32I love you too.
06:35To find out, I've come to see Federico Rossano,
06:39director of the Comparative Cognition Lab
06:41at the University of California, San Diego.
06:44You were running the largest dog study.
06:46How do you even describe it?
06:48What are you doing?
06:49It's the largest citizen science study
06:52on animal communication ever attempted.
06:54Right.
06:54We have more than 10,000 participants
06:57from 47 countries.
06:59And these participants have been training
07:01their dogs for years.
07:03That's a huge study.
07:04It's a huge study.
07:05But at first, Federico didn't want
07:07to have anything to do with button-pressing dogs.
07:10How can we know that this is actually happening
07:13and there's no cueing, that the owner has not just trained
07:16and then recorded the clip so that, you know,
07:18you can show it.
07:19After all, we know that you could train a pigeon
07:21to open boxes.
07:22So it's like, there's things you could do
07:24if you just do it enough times.
07:25Right.
07:26But despite Federico's concerns about social media,
07:29he saw some clips that he just couldn't ignore.
07:32When you start looking at some of these clips,
07:34you start noticing something that is a little different
07:36from what you expect.
07:38Yeah.
07:38Let me just show you one of these clips.
07:41Mad.
07:43Why are you mad?
07:46Why mad?
07:57Where is your ouch?
08:00Where ouch?
08:02Stranger.
08:04In your ear?
08:07Where is your stranger?
08:13Pa.
08:14In your paw?
08:20Let me see your paw.
08:25Okay, I'm going to put this down.
08:26This is the stranger in her paw.
08:28She's got a mat between her, ow, toes.
08:33So they seem to be able to kind of do some back and forth
08:36with the humans and there seems to be the ability
08:39to put more than one, using more than one word
08:42at the same time in a sequence.
08:44Kind of like a sentence.
08:44To combine ideas, right.
08:46And so when you look at this, it's pretty impressive.
08:49We are seeing from different dogs things that could be
08:53really interesting if true.
09:00The trouble is, past efforts to teach animals to use our
09:04language have been incredibly controversial.
09:07Which color?
09:09That one?
09:10The blue one.
09:11The way it was done was basically a researcher would get
09:16one animal and then train this animal to learn, for example,
09:20sign language or to learn how to use other devices.
09:26Sometimes it looked like something extraordinary.
09:29But if you only have one animal, it's very hard to tell how that
09:32came about or what that really means from the perspective of
09:35cognition and what that means about how they're thinking.
09:39After several studies focused on one or two animals
09:43taken out of their natural habitats and with concern
09:47about researchers over interpreting results,
09:49the entire field of research was called into question.
09:54So when the button dogs emerged on social media,
09:57many of Federico's colleagues thought he'd be crazy to get involved.
10:01I had scientists calling me, saying,
10:05do you know what you're getting yourself into?
10:07You might actually lose your job because of what you're doing.
10:10But this study would be different.
10:13Based on more than just one or two highly trained subjects.
10:19Much, much more.
10:21Thousands and thousands of dogs to help make the study
10:25as legitimate as it could be.
10:29Given its scope, Federico asked Amalia Bastos,
10:33a researcher in animal cognition, to join him.
10:36Where can we take this and how far can dogs go?
10:39We need to actually test things out empirically
10:40to find out what's going on.
10:43Federico and Amalia have been recruiting participants
10:46from across the world.
10:50While some are manually recording data for use in the study,
10:53many participants are using buttons
10:56that automatically record what words are pressed and when.
11:01Every time a button is pressed,
11:03that goes straight into the app
11:05and then we can download through the app
11:07the kind of data that has been collected.
11:15To properly investigate this whole phenomenon,
11:18Federico first needs to check
11:20that the dogs actually understand the words
11:22on the buttons when they hear them.
11:27I thought I might see a study in their lab room,
11:30but it's totally empty.
11:33We actually don't bring dogs into the lab.
11:36Dogs are trained in their own homes
11:37and we go to people's homes to actually collect data.
11:41Why is it important to do a study where the dog lives?
11:44You need to imagine, like,
11:46how do you behave when you go to the doctor
11:48when you are in your office
11:51versus how do you behave when you're home?
11:52And how do you feel comfortable, natural,
11:56not necessarily under pressure?
11:59And I think it creates an environment
12:01where you can actually show your abilities
12:03if this is there.
12:05I want to show you something
12:06that is actually one of the studies
12:07that we've been doing.
12:09This is Makai.
12:11You're so calm.
12:13Federico and Amalia need to test
12:15whether a word generated by a button
12:17means the same thing to a dog
12:19as when that same word is spoken by a person.
12:23Who's having some water?
12:25Because when we say words,
12:26we add all sorts of extra details.
12:29Pitch, volume, body language.
12:31You are the messiest drinker.
12:34But a word from a button is the same every time.
12:38The first study had to be
12:40do the dogs that are trained with these buttons
12:42understand the meaning of the words
12:45that are associated with it?
12:47And can we show that they would respond
12:49not just to their owners,
12:50but they would respond to others
12:52pushing those buttons?
12:53All right.
12:54This is client trial number four.
12:57The experimenter wears headphones
12:59so she can't hear what word button she will press.
13:03And the owners are blindfolded
13:05so no one can give away any clues.
13:10And then you see the dog
13:12goes straight for the toy basket,
13:14retrieves the ball,
13:15like, let's play.
13:20And so it's contextually appropriate
13:22with respect to what button had just been pressed.
13:27It's the simplest possible test.
13:3158 other dogs were also tested.
13:37And the results showed that dogs respond appropriately
13:41to the play and outside buttons
13:43with no other cues,
13:45just as they would respond to those words
13:47spoken directly by a human.
13:52That was critical for us
13:53because if they did not respond
13:55in a way that would suggest
13:56they understand it,
13:57then there's no point
13:58in moving forward
13:59with any other research.
14:02So it's now clear
14:03that dogs can respond
14:05to at least some button words.
14:08The next question is
14:09how well can dogs use them?
14:12But first,
14:14it's massive.
14:15How many words does
14:16the average dog seem to know?
14:19I love you.
14:31I am here at a dog show
14:32and I'm not expecting to find
14:34any talking dogs here,
14:35but it's probably a pretty good place
14:37to understand how they understand us.
14:39Hello.
14:41You're so cute.
14:45I'm here to meet Sophie Jacques,
14:47Associate Professor
14:48of Developmental Psychology
14:50at Dalhousie University in Nova Scotia.
14:54We're at a dog show.
14:55What can a developmental psychologist
14:57learn from studying dogs?
14:59So I study language development in children
15:01and the differences between dogs
15:04and babies in learning language
15:06as well as the similarities
15:07can tell us a lot about
15:08what's going on in both kinds of species.
15:11We talk to dogs in the ways
15:13that are very similar
15:14to how we talk with babies.
15:17When measuring language ability
15:19in infants,
15:20researchers often use word lists
15:23filled out by parents
15:24who report what words
15:25they believe their baby understands.
15:28Sophie produced the same kind of list,
15:30but for doggies.
15:33So this is a list of words
15:35that we developed
15:36to measure the same thing in dogs.
15:38So we figured
15:39that if parents were really good
15:42at picking out what words
15:43they think their babies know,
15:45maybe dog owners would be...
15:46You're telling me this is a list of words
15:48that dogs know?
15:49Yes.
15:50This is a list of words
15:51that dogs know.
15:52It's massive!
15:55In general, dog owners
15:57are reporting that their dogs
15:59know about 80 to 85 words.
16:01That's a huge number of words.
16:03I mean, when you think about
16:04the average one-year-old
16:05or maybe two-year-old,
16:06the list is about the same, isn't it?
16:08Yeah, so a one-year-old
16:10parents report
16:11they respond to about 80 words.
16:13So about the same number of words.
16:14But there are big differences
16:15in the list.
16:16Most of those words are objects.
16:18They're nouns.
16:18For dogs, most of those words
16:20are commands or actions.
16:23So sit, stay, come.
16:25Whereas with babies
16:26it might be bottle, mummy.
16:29And words we think of as objects
16:32a dog might relate to differently.
16:35They'll recognize words like car.
16:37So do you want to go for a car ride?
16:39Now the question is
16:40do they recognize the word car
16:42or mat as an object
16:44or as a place to go?
16:45They hear car and they go to the car.
16:47Yes, that's right.
16:48It's an action word.
16:50Sit.
16:50With so many words
16:51that dogs seem to respond to
16:52Good boy!
16:54How close could they be
16:55to actually using our language?
16:57So what would it take to convince you
16:59that a dog can talk?
17:02Well, a language is more than
17:03knowing a hundred words
17:04or a thousand words.
17:06It's more than just knowing words.
17:08Language has purpose.
17:09It has meaning.
17:10It has structures.
17:12It is a system of symbols
17:14that is organized
17:16in a very precise way.
17:17Right.
17:17Good job.
17:19Just because dogs respond
17:20to lots of different words
17:21in a particular language
17:22Good man.
17:23Ready, ready, ready.
17:24Over!
17:25Woohoo!
17:26Doesn't necessarily mean
17:27they comprehend that language.
17:29That's so good.
17:31And even when
17:32they're pressing buttons correctly
17:33it's hard to know
17:35how much they understand.
17:38Good boy.
17:38You're gonna try not to bark
17:39at stuff today?
17:40Yeah.
17:46Can you go to bed?
17:47This is Rohan
17:48Yeah, good boy.
17:50Amalia's 23-month-old
17:51Australian Shepherd.
17:54Ready?
17:55Ready?
17:56Button.
17:57Great.
17:57Yeah, so good.
17:58Go get it.
18:01Jake.
18:04Love you.
18:05Good boy.
18:06Go to bed.
18:18Good boy.
18:21Good boy.
18:29Good boy.
18:31Good boy.
18:32I could have had a bird chirp
18:33or I could have had, you know,
18:34siren sound.
18:36It could have been literally anything
18:37and he would have performed
18:38the behavior exactly the same way.
18:40Love you.
18:41Love you.
18:42You know, the fact that they're
18:43human language words
18:44is completely arbitrary.
18:47Love you.
18:48It's just associative learning.
18:50Rohan has learned to associate an action,
18:52like pressing a button,
18:54with an outcome,
18:55like getting a treat.
18:58It's known as associative learning
19:00or conditioning.
19:02But it doesn't mean he knows
19:03what the words treats or thanks
19:05or love you actually mean.
19:08Everything in life starts
19:09through associative learning
19:10and actually humans learn
19:11through associative learning.
19:13Good boy.
19:13Associative learning is common
19:14to pretty much all animals.
19:17Freak.
19:18And it doesn't require
19:19any sort of complex cognition
19:20that we should be impressed by.
19:22Great.
19:23You could easily teach
19:25a mouse to press buttons.
19:26Okay, you got that one.
19:28Button.
19:28And that's the big problem.
19:30Just because they can press buttons
19:31that say words
19:32doesn't necessarily mean
19:34that the dogs have
19:34any real understanding
19:35of what those words mean.
19:38Great.
19:39So how much of all this button pressing
19:41could really just be random?
19:46I love you.
19:51I love you.
19:51In principle, the dog
19:52could just be spamming
19:53the soundboard,
19:54just coming there,
19:55touching buttons randomly.
20:00If it's random,
20:01we should stop doing
20:02this research, right?
20:03Why are you looking into
20:04something that's completely random?
20:08Or what if the dogs
20:09are just pressing buttons
20:10that are easy to reach?
20:12What can happen, of course,
20:14is that the dog might
20:15just end up pushing,
20:16for example,
20:17these two buttons
20:18that are close to each other
20:19just because they are
20:20adjacent to each other
20:21and it's just easy for them.
20:24And so it is very important
20:27for us to know
20:28not just which buttons
20:29they have, participants have,
20:31but also how they're laid out
20:33so that we can, for example,
20:34run analysis to see
20:35the dogs seem to be actually
20:37doing something intentionally.
20:39But after two years
20:41of analyzing over
20:4210 million button presses
20:43by thousands of dogs,
20:46Federico is seeing
20:47clear patterns emerge.
20:49What are some of the most
20:50common words
20:51that dogs are saying?
20:52So outside, play and food,
20:54the size of the word
20:56tells you something
20:56and how common are.
20:58Treat, of course, water,
21:01scritches, walk, want.
21:03So part of the idea is
21:04that, you know,
21:05the words that I use the most
21:06are words that make sense
21:07for a dog.
21:08And again,
21:10so some of these things
21:11they don't use a lot.
21:13Other words that might not be
21:14quite in line with what we expect
21:16like training or mad
21:20or friend.
21:21So on the whole,
21:23button use is not random.
21:27If it was random,
21:28you would just have
21:29all the buttons
21:30that you have on the soundboard
21:31have an equal probability
21:32of being pressed.
21:36Okay, some dogs
21:37press an outside button
21:39when they want to go outside.
21:41Should we be impressed?
21:43In principle,
21:44there is no difference
21:46between pushing the buttons
21:47and saying outside
21:48and trying to tell you
21:50by scratching on the door
21:51that they want to go outside.
21:53It's another way
21:54of communicating.
21:56So is it possible
21:58for dogs to do more
21:59than just communicate
21:59in these simplest of ways?
22:01What would it mean
22:02for them to use language?
22:04All animals communicate.
22:06All animals communicate.
22:08All animals communicate
22:08and can understand
22:09communication
22:10but not all communication
22:11is language.
22:13Arik Kirschenbaum
22:14studies animal communication
22:16at the University of Cambridge
22:18in England.
22:19I think really communication
22:20becomes language
22:21when you cross that boundary
22:23from a signal
22:25that manipulates
22:26other individuals
22:27to a system
22:29by which information
22:31is exchanged
22:32between individuals
22:33who understand
22:34that their conversation
22:37partner
22:37is also participating
22:39in that conversation.
22:41Could dogs
22:42and humans
22:42cross that boundary
22:43and engage
22:44in a deeper exchange?
22:48It's hard to imagine
22:49our dogs actually
22:50speaking to us
22:51but we have to remember
22:52that they evolved with us
22:54over the course
22:54of thousands of years
22:56and that would have
22:56changed their behavior
22:57in some pretty
22:58profound ways.
23:02All dogs evolved
23:04from wolves,
23:05their wild ancestors.
23:07More than 10,000 years ago
23:09some wolves were taken in
23:11by humans
23:11and domesticated.
23:13Over generations
23:15spent living with
23:16and around us
23:17they changed in different ways
23:19according to how
23:19we bred them.
23:23We selected traits
23:24dogs that we valued
23:25leading to breeds
23:26that were smaller
23:27and cuter
23:28or faster
23:29and slimmer
23:31or bigger
23:31and stronger.
23:33And of course
23:34over time
23:35we favored dogs
23:36that could communicate
23:37well with us.
23:39Dogs that were good
23:40at understanding
23:41what we're trying to say.
23:44And a small percentage
23:45of dogs worldwide
23:46really do seem to have
23:48an extraordinary grasp
23:49of our words.
23:52take Gaia for instance,
23:54a six-year-old border
23:56collar.
23:57Isabella Ruiz
23:58is her owner.
24:03While most dogs
24:04are good at learning
24:04action words like sit
24:05and stay,
24:07not many seem to recognize
24:08the names of more
24:09than a few objects.
24:11But that's not the case
24:12with Gaia.
24:13Gaia has 215 toys now.
24:17She knows all of them
24:18by name.
24:19It's a lot of toys now
24:21for us to remember.
24:22So we did a catalogue
24:25of the toys
24:25to remember their names.
24:28But Gaia always knows
24:29the toys.
24:30So yeah,
24:30I think her memory
24:31is better than ours.
24:34Isabella enrolled Gaia
24:36in a study called
24:36the Genius Dog Challenge,
24:38run by Claudia Fugatza.
24:41She's traveled here
24:42all the way from Hungary
24:43where she works
24:44at one of the world's
24:45leading canine research
24:46centers.
24:48Gaia is one of the,
24:50what we call the gifted
24:51word learner dogs.
24:52And these are dogs
24:53that have a very special
24:55talent for learning
24:57the name of toys.
24:58And they learn very fast.
25:01Only a few dogs
25:02around the world seem
25:03to have this capacity.
25:06Hello.
25:07Hi, Gaia.
25:08Hello, girl.
25:09Hi.
25:10She's here to put Gaia
25:11to the test.
25:15She wants to find out
25:16whether Gaia's cognitive
25:17skills might go beyond
25:18learning names and recall.
25:21Coffee.
25:21Can Gaia work out
25:22what a toy might be called
25:23based only on how
25:25she plays with it?
25:27Gaia knows that tug
25:29and fetch
25:30are games she can play.
25:32New toys?
25:34And last week,
25:35she was given two new toys.
25:37One to tug
25:38and one to fetch.
25:40So this one,
25:42I play like this.
25:44Isabella hasn't given
25:45the toys any names
25:46and hasn't said the words
25:47tug or fetch
25:48when playing with them.
25:49And this one,
25:50we've been playing like this
25:52but I don't say the name
25:53of the categories,
25:54so yeah.
25:56Is Gaia smart enough
25:57to work out which
25:58is a fetch toy
25:59and which is a tug toy?
26:05So are they able
26:07to sort those items
26:08into the tug
26:09or the fetch category
26:10just based on the way
26:12you play?
26:12Where's the tug?
26:26Where's the tug?
26:28Where's the
26:35Gaia, where's the tug?
26:39Where's Thug?
26:51Good girl!
26:53Good girl!
26:55Well, I have to say, I had no idea how this would go,
26:59but that was really, really exciting.
27:01She was able to assign this toy to the named category
27:05without ever having heard that name paired with that toy.
27:09Totally, I believe this is a big step
27:11in our understanding of dogs' minds.
27:15What Gaia has done seems different
27:17from just simple conditioning or associative learning.
27:21For sure, Gaia could not have associated the name,
27:26fetch or tug, to the specific item
27:28because she has never heard that name
27:31while seeing that toy or while playing with that toy.
27:34This is not possible.
27:35She could not have learned that this way.
27:38It seems some dogs can understand
27:41that a word can refer to a whole category of objects.
27:44It doesn't necessarily have to refer to one specific thing.
27:49Go! Yay!
27:51But that's just words that they're hearing.
27:53The owners are not allowed to speak to the dog,
27:55point or gesture at the dog, nothing at all.
27:58Could some dogs be genius enough
28:01to use their word buttons in totally new contexts?
28:04Oh, hi! Do you need help?
28:18Meet Parker.
28:20She is a beagle cross,
28:22and she is all about using those word buttons.
28:25Water.
28:26So Parker is a bit atypical.
28:29When we first got them,
28:31it took her about 30 seconds, maybe not even,
28:34to press her first button.
28:35And by the end of that kind of evening,
28:38she had pressed at least experimentally all six of them.
28:42Snug, snuggle.
28:44You want snuggle? Come here.
28:46Parker is being tested by Amalia, Federico's colleague,
28:50with something called the impossible task.
28:53It's designed to see if a dog who has a help button on a soundboard
28:57will use it in a situation it has never encountered before.
29:03So with most of the buttons that dogs have,
29:05like food or play or walk,
29:06these are simple associations that are related to one particular context.
29:11Help.
29:12Help is interesting because it's a little bit more general
29:14than a lot of the simpler buttons that these dogs might have.
29:18Sit.
29:19Good girl, wait there.
29:20So the way that the study is set up
29:23is we have this mat on the floor here,
29:25and it's got a Tupperware in the middle.
29:27First, the owners give the dogs opportunities over four different trials
29:30to obtain a food reward from inside the Tupperware,
29:32and at that point, the lid is open.
29:35And at all these stages,
29:36the dog can knock off the lid
29:37and eat the food inside the Tupperware without much difficulty.
29:41Until the very final trial, the fifth trial,
29:43the Tupperware is completely locked,
29:45and that's the point at which the dog might need help to access the food.
29:49The owners are not allowed to speak to their dog,
29:51point or gesture at the dog, look at their dog, nothing at all.
29:57I didn't know what Parker would do in that circumstance,
30:00whether she would use the buttons or not.
30:05I can see she's immediately gone to Sasha,
30:07and she's climbed onto Sasha's back trying to get her attention.
30:10But what we're interested in is can these dogs then make that leap and use a word?
30:16Help! Look!
30:19She presses help, look, and then she looks to Sasha.
30:23Parker had never before been presented with this scenario,
30:25and so had never used the words look or help in this context.
30:30Help! Help! Look! Look!
30:33So, because it's the very first time that they've seen this,
30:36it tells us two things.
30:37It tells us that behavior wasn't trained,
30:38and it tells us that the dog is having to think about what this requires from them
30:44in terms of communication to their owner,
30:46and that they need to generalize a button that they've not used in this context before.
30:50Help! Help! Help!
30:51It seems from this experiment that Parker understands that the word help
30:55can be used in a broad range of scenarios,
30:59revealing a level of cognition beyond just simple association.
31:03Oh, hi! Do you need help?
31:08There you go.
31:12As well as using words in new, untrained ways,
31:15a huge step would be if dogs were using words to do more than just ask for things.
31:22What we're actually interested in is whether dogs might be able to use the soundboard
31:26not just as a vending machine where they're requesting things from humans,
31:29but actually understand it as a communicative device.
31:33So, humans do not just request things. Humans can inform.
31:38It's a thing that humans do all the time.
31:41Toddlers start doing it. They will point at the moon and say moon,
31:44and, you know, they're not requesting the moon. They're just letting you know the moon is there.
31:48Parker uses the buttons to narrate a decent amount.
31:51She has narrated me watering the plants quite a bit.
31:56She'll narrate also things she does sometimes.
32:00Settle.
32:02She's pressed settle before and then just gone into her bed and lie down.
32:08She didn't get anything from me for it. She just pressed settle and then she settled.
32:15Of course, it's impossible to know for sure whether Parker really was commenting on what she was doing.
32:21But other dogs like Copper, a three-and-a-half-year-old Labrador,
32:26also seem to be using buttons to do more than just make requests.
32:30Outside pool.
32:31Outside pool.
32:32Outside pool maybe tomorrow.
32:40Outside pool later.
32:42Outside pool.
32:43Later.
32:47Oh, my goodness. Copper.
32:49Copper.
32:50It looks like Copper was using her buttons to express emotion.
32:55But, of course, we don't know for sure.
32:58An animal could learn that pressing the sad button gets attention.
33:03And once they've learned that, they can manipulate these different concepts in quite complex ways
33:09without understanding there's a semantic meaning behind it.
33:16Outside pool later.
33:21From the data he's receiving from button boards, Federico tells me that, according to their owners,
33:27dogs are using buttons to do a lot more than just ask for things.
33:31This is human responses?
33:32This is what the human tells you when the buttons are pushed, what they think the dog is doing.
33:39So, it's interesting for us because you would expect most of the things to be a request for action.
33:44And you expect some informing or attention-seeking, sharing thoughts and feelings.
33:48Can they talk about emotions?
33:49It happens 5% of the time, but it's happening, and maybe we should look more into it.
33:5650% of button presses were recorded as requests.
33:59But there are other categories where dogs seem to be doing stuff that I find shocking.
34:05And this, to me, is one of the most fascinating ones, speaking for others.
34:11What happens sometimes is that you have multiple learners in their household, so there might be like two dogs or
34:17three dogs.
34:18Okay.
34:18And one is proficient with the soundboard, can use the buttons, and the other one doesn't.
34:22And so you have the one that can use it is actually telling the humans about what the other one
34:28wants.
34:28And so they might tell you, for example, the other one wants water, or the other one needs help.
34:34So you have dogs that can translate on behalf of other dogs to humans.
34:39Correct.
34:39That is mind-blowing.
34:41And that's why it's very exciting for us, because even though it's fair, imagine what it could mean.
34:47For now, these reports are certainly tantalizing, but it's just what owners believe their dogs are communicating.
34:54Knowing for certain what a dog means when it uses words, or what it thinks when it hears one, is
35:00far harder.
35:04And it's got some scientists wondering, maybe we should look in their brains.
35:08So, right now I'm putting on the electrodes.
35:21In Budapest, Hungary, Marianna Boros wants to find out what's happening inside a dog's brain when it hears words it
35:29knows.
35:31For most people, we have a mental representation of things from the external words.
35:37So we have a mental representation of an object, some kind of a memory of that thing.
35:43We can imagine it in a way that we hear the word ball, we will see in our mental imagery
35:49a ball.
35:51Marianna wants to test whether dogs' brains might work similarly.
35:56When a dog hears a word, does it conjure up an image of that thing in its head?
36:02To find out, she's going to study the brain activity of Demi, an 11-year-old Swiss shepherd.
36:11So, right now I'm putting on the electrodes on Demi's head.
36:16This will measure the brain signal.
36:22Demi's owner, Timmy, has brought in some of Demi's toys.
36:26Then Demi hears a message telling her what toy she's about to see.
36:32Demi, can you see?
36:35Sometimes, Timmy shows her that toy.
36:43But sometimes, Timmy shows a different toy.
36:46The idea is that if the dog sees a different object, it should be somehow surprised.
36:51And this surprise effect should be visible in the EEG signal.
36:56Demi, mutatom. Telephone.
37:01What we can see here is that if a dog saw an object that matched the word that he just
37:07heard,
37:08this is the response that we get.
37:09But when we violate their expectations, so we show a different object,
37:14then we get a more positive response in the EEG signal.
37:18Marianna believes that this difference in the signals indicates that dogs have mental representations of objects.
37:26We can claim, based on this study, that when they hear the name of an object,
37:31they activate something, the so-called memory of this object.
37:35So, we think that the dog is actually surprised when it sees a different thing.
37:40But it is surprised because it has an expectation and because it understands the meaning of that word.
37:48That's a qualitatively different way of using language as opposed to associative learning.
37:58So, yes, I would say there's a possibility that they understand words similarly to us.
38:04Of course, this is still an interpretation.
38:08But if dogs do understand words in similar ways to us,
38:13could they start putting them together and using them in more complex ways?
38:19To find out, we need to look at the dogs that use a lot of buttons.
38:32The Holy Grail would be to find dogs that are capable of using their limited words to express new ideas
38:39that they don't have buttons for.
38:41An ability known as productivity.
38:45Linguistic productivity is absolutely central to the way that humans do language.
38:50But what makes humans special is that we can communicate to each other anything, any concept.
38:57So, that cognitive ability is very, very powerful and very, very rare in the animal world.
39:03In order to look at productivity, you need to have dogs that can combine at least two buttons.
39:09As of yesterday, we have 790 dogs in our pool that do multi-button combinations.
39:15Maybe there's going to be 1, 2, 3, 5, 10, 20 that actually have the ability to kind of refer
39:21to things that you don't have a word for or by combining other signals.
39:25We have anecdotal reports from some of them that they seem to be doing it.
39:31So, Stella has a beach button. We used to live right by the beach.
39:35She loved going to the beach. We went almost every day.
39:37Right.
39:38And when it broke, we took it off because we had to fix it.
39:41And she sniffed right where the empty spot was, where her beach button should be.
39:45And then she said, help, water, outside.
39:50Help. Hey.
39:54Okay.
39:56So, she used water and outside to say beach because she didn't have the beach button available to her.
40:02So, she's coming up with new concepts, combining words the way that people do.
40:07Right.
40:07To express ideas.
40:08Exactly.
40:09So, she, it was like her own synonym for beach.
40:11It's, oh, I can't say beach, so I'll say water outside.
40:15That's incredible.
40:16Yeah.
40:17Did you freak out?
40:18Yeah.
40:18Yeah.
40:19Definitely freaked out.
40:21It's impossible to know whether Christina's interpretation is correct.
40:24But clips like these are helpful for Federico to highlight possible areas for future research.
40:31Hey.
40:33Okay.
40:33Hey.
40:34This is the reason I'm doing it, even though people told me I might lose my job.
40:38Um, it's because I've seen things like that.
40:41I've seen animals putting together two or three buttons in ways that look like that were producing sentences.
40:47But, of course, you should not stop there because that thing could have been trained.
40:51Okay.
40:51We'll go to the beach.
40:53It might look like productivity, but we don't know for sure that it is.
40:57I do find interesting that she pressed buttons that are near beach.
41:00Help!
41:01Hey!
41:02I would have loved to know if she would have done the same if, say, beach was on the other
41:07side of the soundboard.
41:09Although Stella had never before used those buttons together, testing what she meant by them after the event is impossible.
41:16Still, it's a really interesting sign.
41:20Oh, are you very excited?
41:24So, can dogs talk?
41:26I think Parker can communicate using English language words.
41:33Do you want to do training?
41:35No.
41:36Okay.
41:38Do you want to do training, then?
41:41If you want to call that talking, great.
41:45I say that Stella talks because she's using words to communicate.
41:50Okay.
41:52The fact that it's not a trained behavior and that it's independent and that she can use words to mean
41:59a lot of different things indicates language, not just a conditioned trick.
42:05It's not surprising that owners with their close relationships and strong communicative bonds might feel that their dogs are talking.
42:13But scientific understanding of their linguistic capabilities is still in its early days.
42:22We're very much at the beginning of what we're doing and we need to be ready to collect data for
42:27years and years.
42:31And to me, the exciting part is that maybe you can see something that nobody else could see before because
42:36we just didn't have access to all this data.
42:40We're really at the start of this journey of understanding animal cognition, animal communication, and where that lies on that
42:49spectrum between language and non-language.
42:53Where?
42:56Where?
43:02Whereby?
43:03Are you asking where you're going? You're going to go see Natalie.
43:08No one believes dogs will ever reach our level with language. That's clear.
43:12Does it matter if the sentence you're saying is grammatical or not as long as you're getting your point across
43:17and you're actually benefiting from that two-way communication?
43:20I don't think so.
43:22Whether dogs are using language or not, with a new tool in their communication kit and access to our words,
43:30it's possible that the close bonds between humans and dogs could become even closer.
43:35...
43:49...
43:50...
43:51...
43:51...
43:51...
43:51...
Comments