0:02
So, experiments have been conducted to show that kids actually, if you use
a nonsense word like let's use the koba to measure this or if you say, let's use my
uncles' to measure this, and let's use the one with a sticker to measure this.
Kids will remember even though you've only said that one time and
it's in the middle of play, and you don't really say, okay,
now it's time to remember our word, remember the word koba.
You're just interacting, playing with a bunch of toys and you say, oh yeah,
hand me the koba, or can you give me the toma.
And then because the kids see the direction that you're looking,
because the kids see an object they've never seen before, that's novel and
they don't know the label for it.
And because they see you interacting with the object and calling it something.
They then infer that the word koba must be the label for this object.
But likewise, they don't necessarily have to have a new label.
It could also be that they could map on an older term like
a word like uncle that they already know.
But you say, can you hand me my uncles to measure this, and
then they would also infer that that must be how you refer to that object.
So what's interesting is, when you come back, and you have
only done this sort of surreptitiously, three weeks later, kids and
even adults will remember the label for novel objects, even though they haven't
had any more practice and you didn't really practice with them to start with.
So, that's what fast mapping is.
Kids infer that using your gaze or using your
attention and assessing whether they know the name of an object or not already,
they will infer, through the principle of exclusion, that when you say koba,
you must be talking about this thing I've never seen before.
And, amazingly, with no further practice or repetition, they can remember that.
1:58
So, Rico was the first dog to show a very similar capability.
Now this is very surprising.
The reason it's surprising is because great apes, in the 70s and
in the 80s there were a number of projects that focused for
over a decade each, on trying to teach language to great apes.
Now great apes learned many words, hundreds of words.
But there was never any evidence that great apes used the same process of
fast mapping, to learn the labels of novel objects they didn’t know before.
Maybe future research will show it, but
at least right now dogs are the champions when it comes to fast mapping.
And it was because of Rico that my colleague at the Max Planck Institute,
Juliane Kaminski, discovered was capable of fast mapping and
she really shocked a lot of people, including me.
2:51
Okay, so what did Rico do?
Well, Rico was really an interesting dog, basically her owner
contacted Juliane Kaminski, who's now at Portsmouth University.
And said, I have this really interesting dog, who knows 200 words,
are you interested in studying the dog?
And Julianne wanted to go see Rico and she did.
And when she got there she heard the story of how the owner taught the dog words.
And how the owner taught the dog words was,
she would put out different objects that Rico already knew the label for, so
imagine a lot of different balls of different colors and sizes and shapes.
And Rico knows the pink ball and the orange ball, but
then the owner would say, well, bring me the yellow ball.
Well, Rico knew the label for all the other objects, and
there was just one ball that Rico didn't know the label for.
And that was the yellow ball, and so
Rico would then bring back the yellow ball without any further instructions.
So, that's the same type of inference kids are using.
I know that that can't be a koba because I know all these other things, so
they must be referring to this thing I've never seen before.
That type of inference is what kids are using, and
the owner suggested that that's what Rico was doing.
So Jillian came up with an experiment or
series of experiments to investigate Rico and see what was going on.
Is it really that Rico learned 200 words like
a small child learns words, using the principle of exclusion?
So she, first of all, tested the ability of Rico to
actually recognize the 200 different toys and 20 different sets of 10 toys.
And sure enough, Rico knew the labels of all the different toys that
Rico had played with before, and then what she did though, and the critical test that
really shocked everybody, was she tested his ability to learn novel labels for
novel toys, when eight of the toys presented were old.
So they put ten toys in a room.
Eight of the toys Rico knew before and hand seen many times and
knew the labels for, but two of those toys Rico had never seen before.
Then they went in a separate room.
They would ask Rico to bring back first,
two toys that Rico already knew and had interacted with many times before.
Now, they did that to rule out the possibility
that Rico was just attracted to the new toy and will bring that back first.
In fact, Rico had to inhibit his interest or tendency to bring back some novel toy,
5:22
cuz Rico was asked to bring back two known toys first.
So once he had successfully brought back two known toys,
then they would say a novel word.
They would say the equivalent of bring me the koba,
and Rico had to go in the room with the remaining eight toys and
find one of the two toys that was new and bring it back.
And that's what Rico did.
And then Rico was tested four weeks later, and also ten minutes after the initial
exposure to these new labels and new toys to see what Rico remembered.
And amazingly, ten minutes later, it was very easy for Rico to remember.
And even four weeks later, even though they only repeated the test once or twice,
and Rico only had exposure to this on the first day, once or twice.
A few weeks later he remembered many of those labels, just like small kids do.
And then what was even more remarkable was that Rico was tested not just for
his ability to use a spoken reference to the word.
Sorry, a spoken reference to the object, but Rico was shown a picture,
or was shown a replica of the object that he was supposed to go and retrieve.
Now this is something his owner had never done before.
So if he was supposed to go get a toy that looked like a hot dog,
he was shown a little miniature hot dog, and say, go get this.
And then in the study where he was examined for his ability to look at
pictures he was shown a photo of a hotdog and said go get this.
What's depicted in that picture?
And amazingly, Rico and two other dogs that Julian Kamensky had identified
subsequently were very good at going and retrieving these objects.
They weren't perfect, and there were some individual variability.
But it was amazing they could do it at all,
Brian HareAssociate Professor
