Popular in Psychology

Books

City Time: On Being Sentenced to Rikers Island

Independent Scholar David A. Campbell

Most Cited

The adapted mind

Dalhousie University Jerome Barkow

Experimental Psychology

Time to investigate human visual shape perception

University of Montreal Martin Arguin

Most Cited

James S. Jackson and the program for research on Black Americans

Harvard University David R. Williams

Psychology

Violence and Mental Illness: Rethinking Risk Factors and Enhancing Public Safety

Duke University Eric Elbogen

Cognitive Psychology

Cognitive remediation – where are we now and what should we do next?

King's College London Til Wykes

Who’s a good boy?! Dogs prefer naturalistic dog-directed speech

View more insights in Social Sciences - Psychology - Language and Memory

View more figures from University of York

View the page of Alex Benjamin

Read the Study

Infant-directed speech (IDS) is a special speech register thought to aid language acquisition and improve affiliation in human infants. Although IDS shares some of its properties with dog-directed speech (DDS), it is unclear whether the production of DDS is functional, or simply an overgeneralisation of IDS within Western cultures. Alex Benjamin discusses.

Image courtesy of interviewee. April 9, 2024

Generating Workspace

Export as PDF

This video video discusses the concept of "infant-directed speech" (IDS), also known colloquially as "baby talk," which is characterized by higher pitch, exaggerated emotions, and distinct pitch contours. This video video style of communication is commonly used by caregivers when speaking to infants to help them learn language and establish emotional bonds. Similar speech patterns are noted in "dog-directed speech," but while both involve some emotional changes and increased pitch, they differ in the articulation of vowel sounds — IDS features more hyper-articulation than speech directed at dogs, which is closer to adult-directed speech in that regard.

Research conducted in this area explores whether dog-directed speech serves a functional purpose in enhancing the bond between humans and dogs or simply stems from the human tendency to use baby-like speech for all infant-like beings. Initial experiments indicated that dogs show a preference for dog-directed speech, but follow-up studies suggest this preference might hinge on the interaction of both prosody (the musicality of speech) and vocabulary.

The researchers have plans for future studies that would involve puppies to investigate whether these preferences are innate or learned, aiming to better understand the factors that influence how dogs respond to different types of speech. The findings imply that using dog-directed speech could help dogs feel more comfortable and engaged during interactions, thereby enhancing the human-dog relationship.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.105 --> 00:02.165
Infant directed speech is the type of speech

00:02.165 --> 00:04.605
that we use when we're talking to an infant, surprisingly.

00:04.945 --> 00:07.565
Um, so this is often referred to as things like mothers

00:07.785 --> 00:10.525
or, baby talk often is a colloquial term.

00:10.625 --> 00:13.685
So when you use a slightly higher pitch voice,

00:13.705 --> 00:16.445
you often exaggerate the emotions in your voice,

00:16.785 --> 00:19.365
and you increase the, the pitch contours.

00:19.545 --> 00:21.485
So you might go up really, really high

00:21.485 --> 00:23.805
and then come down really, really low to sort of tell a,

00:23.925 --> 00:25.045
a sort of emotional story.

00:25.545 --> 00:27.685
Um, and we all, most of us tend to do this.

00:27.685 --> 00:29.285
There's evidence for this in,

00:29.355 --> 00:30.405
lots of different languages.

00:30.545 --> 00:32.045
Uh, lots of different cultures use it.

00:32.425 --> 00:36.085
Um, mainly tends to be caregivers, siblings, parents,

00:36.505 --> 00:39.405
any, anybody sort of like interacting

00:39.405 --> 00:40.965
with the baby on a social level.

00:41.545 --> 00:45.285
Um, and so you might say, the same kind of things,

00:45.355 --> 00:46.365
that you would to a baby.

00:46.505 --> 00:49.205
You might also find yourself talking to an animal.

00:49.305 --> 00:51.325
So lots of people talk to their dogs

00:51.325 --> 00:53.805
and well, they think is a similar speech register

00:53.905 --> 00:55.405
to the way that they talk to babies.

00:55.825 --> 00:56.885
So you might just say,

00:56.905 --> 00:59.885
I'm using baby voice when I talk to my dog.

01:00.305 --> 01:01.925
Um, things like, who's a good boy?

01:01.925 --> 01:03.045
Would you like to go for a walk?

01:03.285 --> 01:04.725
'cause you have that slightly higher pitch

01:04.745 --> 01:07.325
and you have the, the, the pitch contours there

01:07.345 --> 01:09.805
and you use lots of short sentences and things like that.

01:10.425 --> 01:12.325
Um, but what we've actually found,

01:12.345 --> 01:13.685
or what, what the sort

01:13.685 --> 01:17.005
of research in general has found is that, the type

01:17.005 --> 01:18.565
of speech that you use with an infant

01:18.945 --> 01:20.085
and the type of speech that you use

01:20.085 --> 01:21.485
with a dog is subtly different.

01:22.065 --> 01:24.925
So with infants, you have higher pitch, higher,

01:25.265 --> 01:27.045
a affect.

01:27.105 --> 01:29.605
So emotional content, as well as those

01:29.605 --> 01:31.965
that hyper articulation of the vowel sounds

01:31.965 --> 01:33.645
that we see when we're trying

01:33.925 --> 01:35.445
to help somebody understand something

01:35.445 --> 01:37.485
or when we're trying to teach somebody a new word.

01:37.785 --> 01:41.045
So you might say, do you want to go out

01:41.595 --> 01:43.045
outside to a baby?

01:43.045 --> 01:45.125
Whereas we tend not to do that with a dog.

01:45.585 --> 01:47.525
So, with pet directed speech

01:47.525 --> 01:51.285
or dog directed speech, you have the increase in pitch

01:51.385 --> 01:53.885
and you have the increase in pitch contours,

01:54.065 --> 01:55.805
and you have the affect changes,

01:55.985 --> 01:58.285
but you don't have that hyper articulation.

01:58.355 --> 02:00.805
That seems to be something that is specific

02:00.825 --> 02:02.045
to infant directed speech.

02:02.185 --> 02:04.765
So they're all special speech registers that we use

02:04.765 --> 02:06.045
with different listeners,

02:06.045 --> 02:07.965
but they have subtle differences that we might not even know

02:07.965 --> 02:09.405
that we're doing depending on the audience.

02:09.925 --> 02:11.765
I actually started some of this research when I was an

02:11.765 --> 02:13.045
undergraduate student, so it was nearly,

02:13.185 --> 02:14.485
10 years ago now.

02:14.585 --> 02:17.325
Um, but I was interested, there was a lot

02:17.325 --> 02:20.285
of literature showing this type of, information

02:20.285 --> 02:22.285
that I've mentioned about the, the speech itself.

02:22.465 --> 02:24.565
So what is it, how do we use it?

02:25.025 --> 02:27.885
Um, what does it look like in terms of the contours

02:27.885 --> 02:30.765
and the, the, you know, the, the pitch

02:30.765 --> 02:32.645
and the, formants

02:32.645 --> 02:34.845
and all of that sort of stuff about the actual speech.

02:35.545 --> 02:38.605
Um, and there was a lot of evidence to show that this kind

02:38.605 --> 02:40.925
of speech is useful when you're talking to an infant,

02:41.275 --> 02:45.885
both in terms of, helping a baby learn a language.

02:46.105 --> 02:49.845
So helping them learn their native language usually,

02:49.905 --> 02:51.485
by exaggerating those vowel sounds,

02:51.825 --> 02:54.845
but also it increases the social,

02:55.365 --> 02:57.085
relationship with, with an infant.

02:57.085 --> 02:58.285
So it allows an infant to know

02:58.285 --> 03:00.245
that this is a safe social partner for me.

03:00.435 --> 03:02.205
This is someone that's going to provide for me.

03:02.585 --> 03:05.365
And it increases that bond usually at the beginning

03:05.365 --> 03:06.765
between the mother and the, the infant,

03:06.765 --> 03:08.565
but also with the other extended family.

03:09.025 --> 03:10.525
And so there was lots of evidence of

03:10.675 --> 03:12.605
what dog directed speech was,

03:13.105 --> 03:14.365
and there was lots of evidence for

03:14.365 --> 03:16.725
what infant directed speech would be useful for.

03:17.105 --> 03:18.645
So what I wanted to find out is,

03:19.445 --> 03:21.965
although we know that we use dog directed speech with dogs,

03:22.475 --> 03:24.925
does it actually have a functional,

03:25.875 --> 03:26.965
sort of driver to it?

03:27.025 --> 03:29.445
Are we doing it for a good reason for something

03:29.445 --> 03:32.165
that will increase the bond between the dog and the human,

03:32.305 --> 03:34.325
or to increase their attention to what we're saying?

03:34.785 --> 03:36.965
Or is it simply an overgeneralization

03:37.505 --> 03:38.685
of infant directed speech?

03:38.745 --> 03:41.925
So especially in a western culture where we tend to think

03:41.925 --> 03:44.725
of dogs as part of the family, often referred to

03:44.725 --> 03:45.805
as things like fur babies.

03:46.545 --> 03:48.645
Um, and we are, you know, we're a,

03:48.645 --> 03:49.805
we're a nation of dog lovers.

03:49.945 --> 03:52.445
Um, and so lots of people use this kind of speech,

03:52.445 --> 03:55.365
but is that just because we think of them like our babies,

03:55.985 --> 03:58.125
or is there something actually in it for the dog

03:58.125 --> 03:59.405
and for our relationship as well?

03:59.465 --> 04:01.805
So that's how we initially started.

04:02.145 --> 04:05.245
Um, this kind of research was to find out is there some kind

04:05.245 --> 04:06.685
of function to that kind of speech.

04:07.065 --> 04:08.485
So the first paper that I read

04:08.485 --> 04:10.365
and was really grabbed by,

04:10.585 --> 04:12.245
in this area was by burner metal.

04:12.425 --> 04:15.445
And they were, able to record people speaking

04:15.475 --> 04:18.445
with animals, dogs mainly, but also cats.

04:19.065 --> 04:20.925
And that's where they found that,

04:20.995 --> 04:24.045
when recording infant directed speech, dog directed speech

04:24.045 --> 04:27.125
and adult directed speech, they produced these really nice,

04:27.265 --> 04:29.565
hyper articulation triangles.

04:29.905 --> 04:32.245
So for the infant directed speech, they were able

04:32.245 --> 04:34.165
to see these big triangle differences.

04:34.265 --> 04:37.245
So the, the, the vowel hyper articulation was, was large.

04:37.265 --> 04:39.845
So there was a big difference between sort of e sounds,

04:40.125 --> 04:43.405
r sounds and oo sounds in infant directed speech.

04:43.985 --> 04:46.245
Um, but that when people were talking to animals,

04:46.435 --> 04:49.045
they didn't show that hyper articulation triangle.

04:49.385 --> 04:51.845
Um, and they didn't in adult directed speech either.

04:51.985 --> 04:54.445
So in terms of the vowel hyper articulation,

04:54.985 --> 04:57.285
dog directed speech is more sim more close

04:57.345 --> 04:59.965
to adult directed speech than infant directed speech.

05:00.305 --> 05:01.405
But in terms of pitch

05:01.405 --> 05:05.525
and affect, it's more, close to infant directed speech.

05:05.825 --> 05:06.805
So that was one of the first

05:06.805 --> 05:08.005
things that kind of caught my attention.

05:08.385 --> 05:11.525
Um, and I wanted to know, well, is this something

05:11.525 --> 05:13.325
that we use to teach people language?

05:13.545 --> 05:15.485
And if so, why do we use it with a dog

05:15.555 --> 05:16.765
that can't learn a language?

05:17.425 --> 05:20.525
And my research kind of took me through, some more,

05:21.445 --> 05:22.485
evidence from Burnham's lab.

05:22.625 --> 05:25.085
Um, but Zoo Etal also looked at, well,

05:25.085 --> 05:27.485
what if we throw a parrot into the mix?

05:28.105 --> 05:30.285
So parrots are often thought of

05:30.385 --> 05:31.805
as linguistic animals, right?

05:31.805 --> 05:33.965
You know, we can, we can hear examples of,

05:33.985 --> 05:35.365
of parrots talking online.

05:35.985 --> 05:38.245
Um, lots of people will be familiar with this idea

05:38.245 --> 05:39.605
of parrots being talkative.

05:40.185 --> 05:44.165
Um, and although they can learn to make speech sounds,

05:44.415 --> 05:46.885
we're not sure whether they can, you know, learn

05:46.885 --> 05:49.085
to actually talk in the way that humans can.

05:49.545 --> 05:52.045
And so that provides a kind of middle ground

05:52.115 --> 05:56.525
between an infant, a human infant, and a dog who can't learn

05:56.745 --> 05:59.845
Any kind of speech or, or, or sort of, talking.

06:00.265 --> 06:04.005
And then this, this third sort of, audience member.

06:04.505 --> 06:08.085
And they found that hyper articulation was highest

06:08.305 --> 06:12.165
for infants, but then was found in parrot directed speech,

06:12.225 --> 06:13.725
but not in dog directed speech.

06:14.185 --> 06:16.485
So then it started to look like, well, okay,

06:16.485 --> 06:19.085
maybe this is something that we are modulating

06:19.795 --> 06:23.405
unconsciously, perhaps depending on the linguistic capacity

06:23.745 --> 06:24.765
of the listener.

06:24.905 --> 06:26.245
So whether we're talking to someone

06:26.245 --> 06:27.645
that we think can learn language

06:27.905 --> 06:30.085
or whether we're talking to somebody that we know

06:30.815 --> 06:32.885
can't learn language, so the, the the dog.

06:33.305 --> 06:36.605
So, that's where I started kind of,

06:37.345 --> 06:38.965
my undergraduate research and,

06:38.985 --> 06:41.125
and that led onto my, my PhD work as well.

06:41.185 --> 06:43.325
So I think that's where the sort

06:43.325 --> 06:46.365
of the arc came from was kind of, well, we know we do it,

06:46.905 --> 06:48.125
but why do we do it?

06:48.505 --> 06:51.165
Um, and so I, I ended up kind of doing this,

06:51.165 --> 06:55.165
this first experiment, which was the, the initial sort

06:55.165 --> 06:56.765
of starting point for finding out why

06:56.785 --> 06:58.125
we actually use this with dogs.

06:58.625 --> 07:00.125
The first experiment,

07:00.155 --> 07:03.925
that I initiated when I was an undergraduate was,

07:04.065 --> 07:06.885
all about do dogs just show a, a preference?

07:06.985 --> 07:08.685
So we weren't quite sure what to expect here

07:08.685 --> 07:09.885
because they'd like, as I said

07:09.885 --> 07:13.445
before, they had already shown that humans do it

07:13.665 --> 07:16.685
and that, you know, we think dogs like it, you know,

07:17.265 --> 07:18.725
we don't, you know, sort of anecdotally

07:18.785 --> 07:20.045
why, why else would we do it?

07:20.105 --> 07:23.045
But at this point, there was no real evidence to suggest

07:23.045 --> 07:24.285
that dogs had a preference.

07:24.865 --> 07:28.645
Um, and so we set off to look at adult dogs only,

07:29.705 --> 07:33.725
and they were pet dogs, so they had experience of, of,

07:34.225 --> 07:37.685
sort of expected, dog directed speech in the,

07:37.705 --> 07:39.045
in their sort of lifetime.

07:39.785 --> 07:43.965
Um, and we recorded, people speaking to dogs

07:44.145 --> 07:45.685
and people speaking to other adults.

07:46.065 --> 07:48.765
And so the dog stimulus was something like,

07:48.765 --> 07:50.485
you're such a good boy, would you like,

07:50.485 --> 07:51.605
it was, it was dog.

07:51.605 --> 07:53.845
So we had gender neutral, terms so

07:53.845 --> 07:55.125
that if the dog was a boy,

07:55.125 --> 07:56.885
it wouldn't be confused if we were saying girl.

07:57.305 --> 07:58.885
Um, but the stimuli were things like,

07:59.505 --> 08:00.725
would you like to go for a walk?

08:00.725 --> 08:03.085
Shall we go outside? You are such a good dog. Yes you are.

08:03.185 --> 08:05.245
You're such a good dog. That kind of thing.

08:05.625 --> 08:08.165
And it was recorded in a naturalistic,

08:08.165 --> 08:09.365
interaction with the dog.

08:09.745 --> 08:12.405
Um, and then the adult directed speech was just a

08:12.405 --> 08:13.725
conversation between two adults.

08:13.785 --> 08:16.125
So things like, I went to the cinema last night,

08:16.225 --> 08:17.765
saw a really good film, it was really funny.

08:17.925 --> 08:19.485
I think you should go and see it. And

08:19.485 --> 08:22.285
so we had those examples of speech, to keep consistent

08:22.285 --> 08:23.445
to cross the experiment.

08:23.705 --> 08:27.045
And, we then took those two, the dogs

08:27.145 --> 08:31.005
and we had them come into the, the room, the two people

08:31.005 --> 08:33.205
that were acting as the, the speakers,

08:33.205 --> 08:34.285
would be sitting there already.

08:34.905 --> 08:38.085
Um, and they would hear this speech from, from both people.

08:38.345 --> 08:40.045
And then we were measuring

08:40.065 --> 08:41.525
how long they spent looking at

08:41.525 --> 08:42.725
each person while they were talking.

08:43.465 --> 08:46.165
Um, and that was our sort of measure of attention,

08:46.305 --> 08:47.405
was how long they spent looking.

08:47.865 --> 08:50.445
And then we also allowed them to, to,

08:50.665 --> 08:52.565
be off lead at the end of the stimulus.

08:53.185 --> 08:55.765
And that allowed us to see how much do they try

08:55.765 --> 08:56.965
and interact with each of the people.

08:57.105 --> 08:59.285
So one of them had been using dog directed speech,

08:59.285 --> 09:01.285
and one of them had been using adult directed speech,

09:01.745 --> 09:03.685
and we wanted to know, well, do they prefer the one

09:03.685 --> 09:05.245
that was going, you're such a good dog.

09:05.865 --> 09:09.405
Um, and so we found, I'll just talk a little bit about the,

09:09.405 --> 09:13.365
the, the findings from that was that dogs did both prefer

09:13.585 --> 09:15.965
to look towards the person using dog directed speech

09:16.825 --> 09:19.045
and, prefer to spend time with

09:19.045 --> 09:20.965
that person when they were let off the lead.

09:21.305 --> 09:25.045
So from that, we concluded really that the,

09:26.485 --> 09:28.785
it is, it is a preference the for the dogs.

09:28.785 --> 09:30.545
So dogs who have heard this kind of speech

09:30.565 --> 09:34.425
before do seem to show some kind of preferences towards,

09:34.445 --> 09:36.385
dog directed speech over adult directed speech.

09:36.525 --> 09:37.865
For those of you who are thinking, yeah,

09:37.885 --> 09:40.465
but obviously the dogs are gonna prefer to go

09:40.465 --> 09:42.305
to somebody going, yeah, is that your good boy?

09:42.305 --> 09:43.785
Don't wanna go outside? That kind of thing.

09:43.785 --> 09:44.985
If you are using that kind of voice,

09:45.325 --> 09:47.105
why would they not go to that person?

09:47.725 --> 09:49.945
Um, because they might recognize the words.

09:50.365 --> 09:52.505
So they might say, well, good boy is something

09:52.505 --> 09:54.505
that often comes with treats or affection,

09:54.525 --> 09:56.745
or, if you say, do you wanna go outside?

09:56.815 --> 09:58.905
They get excited about it. They know the words.

09:59.565 --> 10:03.265
So from that first experiment, we can't conclude whether

10:03.375 --> 10:06.385
that preference was due to the words

10:06.385 --> 10:07.945
that we'd be using the vocabulary

10:08.405 --> 10:11.625
or whether it was the sound of the speech.

10:11.725 --> 10:13.065
So the sing songy voice

10:13.085 --> 10:14.225
and the, the kind of,

10:14.225 --> 10:15.545
pitch contours and things like that.

10:16.365 --> 10:18.905
So that's when we did experiment two.

10:19.005 --> 10:20.865
So we rerecorded the stimuli,

10:21.005 --> 10:22.625
but this time we swapped the words.

10:23.045 --> 10:25.625
So we wrote out a transcript of the scripts

10:25.625 --> 10:26.865
that we used in the first experiment,

10:27.365 --> 10:31.585
and then we, had people talk to a dog,

10:31.645 --> 10:34.585
but say the words that were from the adult directed speech.

10:34.645 --> 10:35.745
So now we had stimuli

10:35.745 --> 10:38.145
that sounded like we did the cinema last night.

10:38.205 --> 10:41.865
So a really good film. And would you like to go outside?

10:42.045 --> 10:44.865
You're such a good dog. Yes you are. You're such a good dog.

10:45.045 --> 10:47.945
Yes, you are, et cetera, which is actually harder

10:47.965 --> 10:49.665
to do than the first one in some ways.

10:49.965 --> 10:53.345
Um, so you can practice yourselves at home, I'm sure.

10:53.765 --> 10:55.785
Um, so then we took this new stimuli back.

10:55.885 --> 10:58.705
So this was the, this, crossed over stimuli.

10:58.805 --> 11:02.425
Um, and we had them, list do the exact same experiment,

11:02.425 --> 11:05.345
experiment again with, in the same place

11:05.405 --> 11:07.825
and the same sort of dogs, but different individual dogs.

11:08.525 --> 11:12.265
And in this experiment, we actually found that we lost

11:12.815 --> 11:16.545
that preference towards the, the dog directed speech.

11:16.565 --> 11:18.705
It was no longer really obvious that they liked the one

11:18.705 --> 11:20.305
that was going, you're going to similar.

11:20.565 --> 11:21.705
No, a really good film.

11:22.245 --> 11:25.145
And so we from that could conclude,

11:25.205 --> 11:27.265
but they also didn't prefer the, the person

11:27.265 --> 11:28.785
that was saying, do you want to go outside?

11:28.785 --> 11:33.505
You're a good dog. So it was neither the, prosody,

11:33.845 --> 11:35.585
the singsongs of the stimulus

11:36.165 --> 11:38.945
nor the vocabulary that was important in

11:38.945 --> 11:39.985
that first experiment.

11:40.445 --> 11:42.785
So really this was a kind of sanity check for us

11:42.785 --> 11:45.345
that it was some combination of the words

11:45.685 --> 11:48.145
and the way that we said them that was important for driving

11:48.175 --> 11:50.345
that preference that we saw in the first experiment.

11:50.735 --> 11:54.445
Dogs are able to notice differences in vocabulary

11:54.505 --> 11:58.165
and prosody, and that they use both of those cues to decide

11:58.425 --> 12:01.085
who to pay attention to, who to spend time with.

12:01.545 --> 12:03.885
And so we're able to suggest that just like

12:03.885 --> 12:05.485
with the infants,

12:05.625 --> 12:06.845
and infant directed speech,

12:06.915 --> 12:09.805
that dog directed speech has some function for our dogs.

12:09.865 --> 12:13.285
It allows them to get some information about who's likely

12:13.345 --> 12:15.885
to give me a treat or a belly rub, who's likely to take me

12:15.885 --> 12:18.005
for a walk, who do I want to spend time with?

12:18.465 --> 12:20.045
As well as when you're training a dog,

12:20.045 --> 12:22.965
perhaps thinking about, you know, using this kind

12:22.965 --> 12:24.805
of speech to grab their attention if you want them

12:24.805 --> 12:26.845
to pay attention to you, learning a new skill.

12:27.345 --> 12:30.165
So that was the sort of overall,

12:30.395 --> 12:31.925
conclusions that we were able to make.

12:32.145 --> 12:34.885
But it was one of the things that's important

12:34.905 --> 12:36.245
to recognize here is

12:36.245 --> 12:38.925
that these dogs had already had experience

12:38.925 --> 12:41.525
of dog directed speech, you know, with their own owners

12:41.985 --> 12:44.525
and with the staff at the facility that we were working at.

12:45.105 --> 12:49.685
Um, and so from these experiments, it's not possible for us

12:49.785 --> 12:53.925
to determine whether this is an innate ability,

12:54.355 --> 12:55.805
like it seems to be with humans.

12:56.625 --> 12:59.325
Um, and not simply that these dogs learned.

12:59.515 --> 13:01.445
When someone talks to me in that silly voice

13:01.675 --> 13:04.085
with those words, I might get something good,

13:04.305 --> 13:06.285
so I should pay attention and I like that person.

13:07.385 --> 13:11.325
So it's possible, to say that it's likely

13:11.385 --> 13:15.125
to be functional in adult dogs, at least dogs

13:15.125 --> 13:17.165
that have had experience with dog directed speech.

13:17.585 --> 13:19.885
And it's also possible to say that that, that from this,

13:19.885 --> 13:22.725
that it also, it works with famili, unfamiliar people.

13:23.185 --> 13:26.725
So these dogs had never met the people involved in the

13:26.725 --> 13:27.925
experiment until that moment

13:27.925 --> 13:29.685
that they started the experiment.

13:30.185 --> 13:33.965
And so it seems to be, useful for us to be able

13:33.965 --> 13:36.045
to take away from that, that when you are interacting

13:36.045 --> 13:40.205
with an unfamiliar dog, say a dog wanders into your driveway

13:40.225 --> 13:42.565
and isn't, is without an owner,

13:42.905 --> 13:45.365
or if you are a vet technician

13:45.385 --> 13:47.285
or somebody that works at a vet,

13:48.145 --> 13:50.045
and meets a new dog for the first time,

13:50.395 --> 13:52.445
that using these kinds of speech registers

13:52.445 --> 13:53.685
with them might be useful

13:53.705 --> 13:56.965
for making them feel more comfortable, making them see

13:56.965 --> 13:58.245
that you are not a threat to them.

13:58.545 --> 14:00.805
Um, and so there's lots of things that we can take away,

14:01.345 --> 14:03.485
but we still haven't quite answered that question of,

14:03.505 --> 14:04.925
you know, is this something they've learned

14:05.025 --> 14:07.445
or is it something that they innately have?

14:07.825 --> 14:09.725
And so one of the things that we are currently working on

14:09.785 --> 14:12.165
is, getting puppies,

14:12.225 --> 14:14.245
before they have had that experience

14:14.275 --> 14:17.885
that dog directed speech equals fun stuff or tasty treats.

14:18.425 --> 14:21.285
So, getting dogs, you know, really,

14:21.325 --> 14:22.965
really early on in the,

14:23.585 --> 14:26.605
in the litter when they have just opened their ears

14:26.605 --> 14:28.845
and opened their eyes and they're able to show some kind

14:28.845 --> 14:31.285
of preference and using the same stimuli to try

14:31.285 --> 14:34.605
and find out if they show, the same sort of preference.

14:35.185 --> 14:37.245
Um, and I think it might be worth mentioning here

14:37.245 --> 14:39.245
that there are some, some other papers, that have,

14:39.525 --> 14:42.005
that have been, slightly more recent than the ones I

14:42.005 --> 14:44.605
mentioned, to do with the background,

14:44.635 --> 14:49.325
that have shown that when the speech type is when the,

14:49.465 --> 14:52.085
the vocabulary is just adult directed speech.

14:52.705 --> 14:56.045
Um, but you use dog directed, sounds, so you,

14:56.145 --> 14:58.645
you say it in a sing-song way the,

14:59.165 --> 15:01.925
puppies showed a preference, but the adult dogs didn't.

15:02.625 --> 15:04.445
Um, and I talk about this more in the paper.

15:04.445 --> 15:08.285
This is a, a, a finding by Ben Adre et al, from

15:09.125 --> 15:11.405
a few years back, and they find

15:11.405 --> 15:14.005
that there is no preference in adult dogs when there's no

15:14.005 --> 15:16.645
human present and when the vocabulary isn't there.

15:17.105 --> 15:19.925
And so some of the future research that we're,

15:20.495 --> 15:23.205
initiating and, and and looking at doing is to try

15:23.205 --> 15:24.805
and find out this idea of

15:24.955 --> 15:27.045
what factors are important for the dog.

15:27.465 --> 15:30.045
Uh, if there needs to be a human there, does the human need

15:30.045 --> 15:32.565
to be looking at them, for example, is eye gaze important?

15:33.265 --> 15:36.485
Um, does it matter if the person is female or male?

15:37.025 --> 15:38.485
Um, if the owner is female

15:38.585 --> 15:41.325
or male, their experience of the language, for example,

15:41.795 --> 15:45.125
what are the specific, variables within that speech

15:45.305 --> 15:47.765
or within that interaction that are important for the dog?

15:47.865 --> 15:50.645
So, puppies, can they do it?

15:50.825 --> 15:53.245
And what other variables might be important is currently

15:53.245 --> 15:53.725
what we're looking at.

00:00 - Infant directed speech is the type of speech that we use when we're talking to an infant, surprisingly. so this is often referred to as things like mothers or, baby talk often is a colloquial term. So when you use a slightly higher pitch voice, you often exaggerate the emotions in your voice,

00:16 - and you increase the, the pitch contours. So you might go up really, really high and then come down really, really low to sort of tell a, a sort of emotional story. and we all, most of us tend to do this. There's evidence for this in, lots of different languages. lots of different cultures use it. mainly tends to be caregivers, siblings, parents, any, anybody sort of like interacting

00:39 - with the baby on a social level. and so you might say, the same kind of things, that you would to a baby. You might also find yourself talking to an animal. So lots of people talk to their dogs and well, they think is a similar speech register to the way that they talk to babies. So you might just say, I'm using baby voice when I talk to my dog. things like, who's a good boy?

01:01 - Would you like to go for a walk? 'cause you have that slightly higher pitch and you have the, the, the pitch contours there and you use lots of short sentences and things like that. but what we've actually found, or what, what the sort of research in general has found is that, the type of speech that you use with an infant and the type of speech that you use with a dog is subtly different.

01:22 - So with infants, you have higher pitch, higher, a affect. So emotional content, as well as those that hyper articulation of the vowel sounds that we see when we're trying to help somebody understand something or when we're trying to teach somebody a new word. So you might say, do you want to go out outside to a baby? Whereas we tend not to do that with a dog.

01:45 - So, with pet directed speech or dog directed speech, you have the increase in pitch and you have the increase in pitch contours, and you have the affect changes, but you don't have that hyper articulation. That seems to be something that is specific to infant directed speech. So they're all special speech registers that we use with different listeners, but they have subtle differences that we might not even know

02:07 - that we're doing depending on the audience. I actually started some of this research when I was an undergraduate student, so it was nearly, 10 years ago now. but I was interested, there was a lot of literature showing this type of, information that I've mentioned about the, the speech itself. So what is it, how do we use it? what does it look like in terms of the contours and the, the, you know, the, the pitch

02:30 - and the, formants and all of that sort of stuff about the actual speech. and there was a lot of evidence to show that this kind of speech is useful when you're talking to an infant, both in terms of, helping a baby learn a language. So helping them learn their native language usually, by exaggerating those vowel sounds, but also it increases the social, relationship with, with an infant. So it allows an infant to know

02:58 - that this is a safe social partner for me. This is someone that's going to provide for me. And it increases that bond usually at the beginning between the mother and the, the infant, but also with the other extended family. And so there was lots of evidence of what dog directed speech was, and there was lots of evidence for what infant directed speech would be useful for. So what I wanted to find out is,

03:19 - although we know that we use dog directed speech with dogs, does it actually have a functional, sort of driver to it? Are we doing it for a good reason for something that will increase the bond between the dog and the human, or to increase their attention to what we're saying? Or is it simply an overgeneralization of infant directed speech? So especially in a western culture where we tend to think of dogs as part of the family, often referred to

03:44 - as things like fur babies. and we are, you know, we're a, we're a nation of dog lovers. and so lots of people use this kind of speech, but is that just because we think of them like our babies, or is there something actually in it for the dog and for our relationship as well? So that's how we initially started. this kind of research was to find out is there some kind of function to that kind of speech.

04:07 - So the first paper that I read and was really grabbed by, in this area was by burner metal. And they were, able to record people speaking with animals, dogs mainly, but also cats. And that's where they found that, when recording infant directed speech, dog directed speech and adult directed speech, they produced these really nice, hyper articulation triangles. So for the infant directed speech, they were able

04:32 - to see these big triangle differences. So the, the, the vowel hyper articulation was, was large. So there was a big difference between sort of e sounds, r sounds and oo sounds in infant directed speech. but that when people were talking to animals, they didn't show that hyper articulation triangle. and they didn't in adult directed speech either. So in terms of the vowel hyper articulation, dog directed speech is more sim more close to adult directed speech than infant directed speech.

05:00 - But in terms of pitch and affect, it's more, close to infant directed speech. So that was one of the first things that kind of caught my attention. and I wanted to know, well, is this something that we use to teach people language? And if so, why do we use it with a dog that can't learn a language? And my research kind of took me through, some more, evidence from Burnham's lab.

05:22 - but Zoo Etal also looked at, well, what if we throw a parrot into the mix? So parrots are often thought of as linguistic animals, right? You know, we can, we can hear examples of, of parrots talking online. lots of people will be familiar with this idea of parrots being talkative. and although they can learn to make speech sounds, we're not sure whether they can, you know, learn

05:46 - to actually talk in the way that humans can. And so that provides a kind of middle ground between an infant, a human infant, and a dog who can't learn Any kind of speech or, or, or sort of, talking. And then this, this third sort of, audience member. And they found that hyper articulation was highest for infants, but then was found in parrot directed speech, but not in dog directed speech. So then it started to look like, well, okay, maybe this is something that we are modulating

06:19 - unconsciously, perhaps depending on the linguistic capacity of the listener. So whether we're talking to someone that we think can learn language or whether we're talking to somebody that we know can't learn language, so the, the the dog. So, that's where I started kind of, my undergraduate research and, and that led onto my, my PhD work as well. So I think that's where the sort

06:43 - of the arc came from was kind of, well, we know we do it, but why do we do it? and so I, I ended up kind of doing this, this first experiment, which was the, the initial sort of starting point for finding out why we actually use this with dogs. The first experiment, that I initiated when I was an undergraduate was, all about do dogs just show a, a preference? So we weren't quite sure what to expect here

07:08 - because they'd like, as I said before, they had already shown that humans do it and that, you know, we think dogs like it, you know, we don't, you know, sort of anecdotally why, why else would we do it? But at this point, there was no real evidence to suggest that dogs had a preference. and so we set off to look at adult dogs only, and they were pet dogs, so they had experience of, of, sort of expected, dog directed speech in the,

07:37 - in their sort of lifetime. and we recorded, people speaking to dogs and people speaking to other adults. And so the dog stimulus was something like, you're such a good boy, would you like, it was, it was dog. So we had gender neutral, terms so that if the dog was a boy, it wouldn't be confused if we were saying girl. but the stimuli were things like,

07:59 - would you like to go for a walk? Shall we go outside? You are such a good dog. Yes you are. You're such a good dog. That kind of thing. And it was recorded in a naturalistic, interaction with the dog. and then the adult directed speech was just a conversation between two adults. So things like, I went to the cinema last night, saw a really good film, it was really funny. I think you should go and see it. And

08:19 - so we had those examples of speech, to keep consistent to cross the experiment. And, we then took those two, the dogs and we had them come into the, the room, the two people that were acting as the, the speakers, would be sitting there already. and they would hear this speech from, from both people. And then we were measuring how long they spent looking at each person while they were talking.

08:43 - and that was our sort of measure of attention, was how long they spent looking. And then we also allowed them to, to, be off lead at the end of the stimulus. And that allowed us to see how much do they try and interact with each of the people. So one of them had been using dog directed speech, and one of them had been using adult directed speech, and we wanted to know, well, do they prefer the one that was going, you're such a good dog.

09:05 - and so we found, I'll just talk a little bit about the, the, the findings from that was that dogs did both prefer to look towards the person using dog directed speech and, prefer to spend time with that person when they were let off the lead. So from that, we concluded really that the, it is, it is a preference the for the dogs. So dogs who have heard this kind of speech before do seem to show some kind of preferences towards, dog directed speech over adult directed speech.

09:36 - For those of you who are thinking, yeah, but obviously the dogs are gonna prefer to go to somebody going, yeah, is that your good boy? Don't wanna go outside? That kind of thing. If you are using that kind of voice, why would they not go to that person? because they might recognize the words. So they might say, well, good boy is something that often comes with treats or affection, or, if you say, do you wanna go outside?

09:56 - They get excited about it. They know the words. So from that first experiment, we can't conclude whether that preference was due to the words that we'd be using the vocabulary or whether it was the sound of the speech. So the sing songy voice and the, the kind of, pitch contours and things like that. So that's when we did experiment two. So we rerecorded the stimuli,

10:21 - but this time we swapped the words. So we wrote out a transcript of the scripts that we used in the first experiment, and then we, had people talk to a dog, but say the words that were from the adult directed speech. So now we had stimuli that sounded like we did the cinema last night. So a really good film. And would you like to go outside? You're such a good dog. Yes you are. You're such a good dog. Yes, you are, et cetera, which is actually harder

10:47 - to do than the first one in some ways. so you can practice yourselves at home, I'm sure. so then we took this new stimuli back. So this was the, this, crossed over stimuli. and we had them, list do the exact same experiment, experiment again with, in the same place and the same sort of dogs, but different individual dogs. And in this experiment, we actually found that we lost that preference towards the, the dog directed speech. It was no longer really obvious that they liked the one

11:18 - that was going, you're going to similar. No, a really good film. And so we from that could conclude, but they also didn't prefer the, the person that was saying, do you want to go outside? You're a good dog. So it was neither the, prosody, the singsongs of the stimulus nor the vocabulary that was important in that first experiment. So really this was a kind of sanity check for us

11:42 - that it was some combination of the words and the way that we said them that was important for driving that preference that we saw in the first experiment. Dogs are able to notice differences in vocabulary and prosody, and that they use both of those cues to decide who to pay attention to, who to spend time with. And so we're able to suggest that just like with the infants, and infant directed speech, that dog directed speech has some function for our dogs.

12:09 - It allows them to get some information about who's likely to give me a treat or a belly rub, who's likely to take me for a walk, who do I want to spend time with? As well as when you're training a dog, perhaps thinking about, you know, using this kind of speech to grab their attention if you want them to pay attention to you, learning a new skill. So that was the sort of overall, conclusions that we were able to make. But it was one of the things that's important

12:34 - to recognize here is that these dogs had already had experience of dog directed speech, you know, with their own owners and with the staff at the facility that we were working at. and so from these experiments, it's not possible for us to determine whether this is an innate ability, like it seems to be with humans. and not simply that these dogs learned. When someone talks to me in that silly voice with those words, I might get something good,

13:04 - so I should pay attention and I like that person. So it's possible, to say that it's likely to be functional in adult dogs, at least dogs that have had experience with dog directed speech. And it's also possible to say that that, that from this, that it also, it works with famili, unfamiliar people. So these dogs had never met the people involved in the experiment until that moment that they started the experiment. And so it seems to be, useful for us to be able

13:33 - to take away from that, that when you are interacting with an unfamiliar dog, say a dog wanders into your driveway and isn't, is without an owner, or if you are a vet technician or somebody that works at a vet, and meets a new dog for the first time, that using these kinds of speech registers with them might be useful for making them feel more comfortable, making them see that you are not a threat to them.

13:58 - and so there's lots of things that we can take away, but we still haven't quite answered that question of, you know, is this something they've learned or is it something that they innately have? And so one of the things that we are currently working on is, getting puppies, before they have had that experience that dog directed speech equals fun stuff or tasty treats. So, getting dogs, you know, really, really early on in the,

14:23 - in the litter when they have just opened their ears and opened their eyes and they're able to show some kind of preference and using the same stimuli to try and find out if they show, the same sort of preference. and I think it might be worth mentioning here that there are some, some other papers, that have, that have been, slightly more recent than the ones I mentioned, to do with the background, that have shown that when the speech type is when the, the vocabulary is just adult directed speech.

14:52 - but you use dog directed, sounds, so you, you say it in a sing-song way the, puppies showed a preference, but the adult dogs didn't. and I talk about this more in the paper. This is a, a, a finding by Ben Adre et al, from a few years back, and they find that there is no preference in adult dogs when there's no human present and when the vocabulary isn't there. And so some of the future research that we're, initiating and, and and looking at doing is to try

15:23 - and find out this idea of what factors are important for the dog. if there needs to be a human there, does the human need to be looking at them, for example, is eye gaze important? does it matter if the person is female or male? if the owner is female or male, their experience of the language, for example, what are the specific, variables within that speech or within that interaction that are important for the dog? So, puppies, can they do it?

15:50 - And what other variables might be important is currently what we're looking at.

Faculti AI can make mistakes