Grammar Girl here.
I'm Mignon Fogarty.
And today I am here with Maria Lanconsolet from the Department of General Psychology at the University of Padua in Italy.
Maria, Dr. Lanconsolet, welcome to the Grammar Girl podcast.
Hi, thank you for having me.
We don't normally talk about chicken research on the Grammar Girl podcast, but you have done such fascinating work that in baby chicks.
But that is just incredibly applicable to linguistics and understanding how human language developed.
Some of our listeners may be familiar with the Boba Kiki effect.
It's something that linguistics students usually learn early in their career, but not everyone is going to know it.
So can you start by sort of giving us that background on what it is you studied?
Yes, of course.
The effect is also studied in psychology.
I am a psychology background.
And it's the idea that if you see a very, very round like bubbly shape and a spiky shape like a pointy shape, and you have to name them and I give you two options, like booba or kiki, you are more likely to say that the round shape is a booba and the spiky shape is a kiki.
Even though you have no apparent reason to do so.
You have not been told the name.
You've never encountered the shape before, yet you have the intuition that round is booba and spiky is kiki.
And this is known as the booba-kiki effect.
And this isn't just in English, right?
This is universal?
Yes, it works in different cultures.
It works in different orthographic systems, linguistic systems.
And even in pre-verbal infants.
It has been found that, before language acquisition and language mastery, infants already show the association between the sound and the shape.
So babies do it?
Yeah, yeah.
I think the youngest age that was tested was four months old.
Wow.
What is the theory about why humans have this association between sounds and shapes?
There are different theories.
So on the one hand there is this idea of us learning the association because, for example, in different writing systems you do have some regularities like the O is always rounded and the A sound is always kind of rounded, whereas the E sound is more of a spiky, or the I, in English, is more spiky.
Or we have this kind of feedback from our muscles.
Like when you pronounce the sounds, your mouth, your lips get more in a round or in a spiky shape.
So there is this idea that it can be learned.
But this would not explain the infants, the babies.
So another possibility is that we are predisposed to this kind of association.
So we have these intuition.
It's like an early and spontaneous association that comes to mind and possibly reflects natural rules.
There is this one study that shows that if an object hits a surface and rolls on it, the sound that is produced depends on the roundness or spikiness of the object and resembles the booba and kiki sounds.
So in nature there is this physical rule for which objects that are rounded and roll on the surface are more in a booba kind of sound and spiky objects are more in a kiki kind of sound.
So we get that from the environment.
Is this something that primates do too?
There is some literature that tried to test the primates, the chimpanzees, bonobos and, I think, gorillas.
The results were not very strong, so they failed to report the effect.
At the same time, however, these animals were adults and were highly trained to language.
And if we think about the boobakiki effect as a predisposition, it makes total sense that experience can change it.
Predispositions are useful as long as you don't have experience that change them, because then you have a rule that can take the place of the predisposition.
So the studies on AIDS are kind of in between.
So we know that they can respond to the stimuli, but we can still not get rid of the whole experience.
So does it almost seem like learning language made them maybe less likely to have this effect?
That's one possibility.
I wouldn't say that's exactly that, because otherwise you wouldn't observe that in adult humans, because we are also adults, the same way they are.
But we can still have the spontaneous learning response when we are asked what is booba and what is kiki.
With animals it's a bit more difficult because in order to get the response from the animals, you have to train them.
You cannot just ask them, which one do you feel is booba?
You have to find a way to get the response out of them.
And usually what you do is to ask them to associate.
So to show them some shapes and some sounds and to ask them how do they pair.
And possibly in that process, because the animals usually work for a reward, they usually work for food.
They might want to try to get that.
So they might kind of not base the response on the intuition they have but try to get the most correct answer they can so that basically they get the food.
Yeah, I will work for food.
Absolutely.
So given that it wasn't a clear effect in primates, what made you think it might be in chickens?
Well, I've worked with chickens for a long time and we do know that they share some cognitive abilities with humans, with babies.
And they are precautional species, which means that we can buy the eggs and hatch the eggs in the lab, having a full control on their experience.
So basically, what you get is a very, very young animal, similar to the newborns, but without all the noise that comes from the experience.
Because even an infant, even a human infant, four months old, they don't speak, but they had plenty of experience.
They had their parents talking to them.
They showed the mouth.
They showed the toys.
They hear the sounds.
With a baby chicken, we can have an animal that is kind of a blank.
It's naive to that kind of experience.
So we can see if the association is in place at the very, very early stage and before dedicated experience.
What do chickens share with baby humans?
Well, they are a very good model in comparative psychology.
We do know that they share, for example, some perceptual priors.
For example, the ability to recognize a biological kind of movement versus a non-biological random kind of movement.
They respond to faces.
They can process questions.
They can respond to some visual illusions.
So they have kind of a very similar way of processing the environment around them.
And that's because possibly they have similar needs.
If we think about biological movement and faces, even if they are very different, they are two different oxides, a bird and a mammal.
But the need to find the caregiver, the need to find a conspecific, the need to find food is shared.
They all have the same needs.
So it makes sense that they have mechanisms that help them cope to that.
Interesting.
OK, so let's talk about the experiments you did.
First, you started with, I think, three day old chicks.
Yes, we first started with these three day old chicks because we wanted to use kind of a standard pattern with the animals.
That is again work for food.
So what we did was to train baby chicks to go behind one panel and avoid a second one.
The correct panel was fly by one shape that had both round and spiky edges, which means that they don't have any information about absolute roundness or absolute spikiness.
They just know that they have to go where the shape is.
And the other funnel, the one that doesn't have food behind, is blank.
So basically they learn one very simple rule that is go where the shape is and you get food.
So they go behind the panel, they get their food and they're happy, and they can learn that quite fast.
It takes like 15 to 20 minutes to teach a chick to do that.
So they don't even get that much of an extensive experience.
And crucially, they never hear any sound up to this point.
So then we test them and the test is similar to the training.
So again, they have two panels, but the problem is that now both panels have a shape on it.
In one case it's fully round and in one case it's fully spiky.
So the rule that they previously acquired is no longer tenable, because the rule go to the shape is doesn't work.
Shape is on both sides.
And that's in this situation of uncertainty.
We thought if they had a prior, if they have an idea of how what's more likely to be, they might resort on that to answer.
Because remember, they are motivated to give the correct answer because they want the food.
And that's when we present them the sound for the first time.
It's either a repetition of the booba sound, like a booba, booba, or of the kiki sound.
And we hypothesize that if they do resource from these priors and this intuition, they will be more likely to go to the round shape when they hear the booba sound and to the spiky shape when they hear the kiki sound.
And that's exactly what happened.
Wow.
Where was the food behind the spiky and the booba and the kiki?
We decided not to give them food at this point.
Oh, interesting.
So no food at all at that point?
That was a risky move, to be fair, because the risk is that they will learn that food is not there and they will stop responding.
That's why we did only 24 trials per chick, because we didn't want to give them time to learn that, even if they keep trying, the food is not there.
So we had a limited amount of trials.
But we also didn't want to strengthen association, to give them the feeling of a correct answer.
So we have no food test.
Yeah.
Wow.
So, with no food reward, they still learned that the Kiki was the spiky one and the Google was the round one and would go investigate.
Fascinating.
Okay.
And so then not satisfied with just your three-day-old.
You went to even younger chicks.
So what was the next experiment?
The next experiment.
We wanted to test chicks as young as possible, because three days old chicks?
Of course they were naive to the sounds.
They were naive to the shapes, but they had life.
Like they had food before, they interacted with other chicks, they had quite a rich social life.
So they might have picked up some information, some multimodal information that we were not aware of.
There is always this risk.
And if you want to tackle into the ontogeny of a mechanism, you always need to account for that kind of situation.
So we wanted to test newly hatched chicks, like taking them out of the egg and in the first days of life test them, which was a bit tricky because that young chicks, they are not that into food.
They still have nutrients from the egg and they still need to learn what they like, what they don't like.
So it's really difficult to motivate them to work for food.
But one thing that they do have is this mechanism for which they learn the stimuli they see, they first see.
So what we did was to put them in front of a TV screen, on a monitor, and they would see the one ambiguous shape, let's call it, the one with both round and spiky edges, like moving on the screen and like moving on one side.
Like moving on the other side.
So it was kind of interesting to them.
And what we know is that when chicks are in this kind of situation, where they have one salient stimulus, they tend to learn the feature of that stimulus and to prefer that stimulus over another one.
Remember they're very young so they like safe and known stimulus kind of environments and stimuli.
And once they learn that so they were exposed to that thing for one hour then we test them, giving them a choice between two alternatives the fully round and the fully spiky shape while presenting the sound.
Again, either a repetition of the booba or of the kiki.
At this point, they again have a choice, and both shapes have some features that they had previously encountered because they were part of the ambiguous shape.
But the sound could be congruent or incongruent with either.
If it's a repetition of booba, that will kind of lead you toward the round shape and kick you toward the spiky shape, if they do have that sort of predisposition.
And in fact we found that they would spend way longer in proximity of the round shape when hearing the sound booba and the spiky shape when hearing kiki.
Amazing.
So this was published in a really good journal.
And.
But you know it was written up in, you know, Scientific American and big popular science magazines.
It's really, you know, important research.
What do you feel like it changes about the way we think about the booba and the kiki effect?
It's a really fascinating experience for me because I am a comparative psychologist and what I usually did them.
My field of work has this idea that living beings face similar challenges and because of that might have evolved similar mechanisms.
But Language is unique to humans.
So I don't think that we are looking at language here.
We're looking at something different.
And what could that be?
It could be something that comes before language.
It could be some prior.
That helps us understanding how the environment works, how multimodal happens.
Because of course in nature nothing happens in one modality.
Objects have a visual consistency, but they also produce sounds, they also produce smells.
So priors, that helps us making sense of all this.
Multimodal information can be useful for different species.
And it is possible then on top of that, our species kind of exploited this kind of disposition and built language.
So it might be like the starting point and a specialization in our species led to language, but the effect itself, the buba kiki association, pre-exists language.
So I think previously there was sort of an idea that it was part of language or like tied to language, but you've showed that it's pre-language.
It's maybe necessary but not sufficient for language.
What would be the advantage, in the wild, for a chicken to be able to make these associations between sounds and shapes?
Well, if we think about the example I gave of the object rolling on the surface, this means that you don't have to see the object.
It's enough for you to hear that and you have an expectation of what's coming.
So let's not focus on the buba kiki.
Let's think of a different association.
If you hear a high-pitched sound, you can expect a smaller object than a lower pitch sound.
And that's, again, a physical rule because the vocal tract bounces.
If you are small, you are bound to produce higher pitch sounds.
And if you are in the jungle and you hear a sound, it's important for you to know whether it's a large animal coming, possibly a predator, or a small animal coming.
So, evolutionarily speaking, you get an advantage from being able to predict the most likely outcome from just one information, rather than waiting to collect the whole multisensory kind of visual acoustic olfactory, etc.
So if you learn that a booba-like sound is related to something big like a hippopotamus, and then you hear that booba coming for you through the jungle, you know to run.
Exactly.
Okay.
And so what are the next steps for your research?
Are you going to look for this effect in different animals?
Are you going to look for different kinds of things in chickens?
What's next?
Yeah, we would be very curious to know what happens in primates that are not trained to language.
Because of course, it's very unlikely that the same ability evolved in birds and then in humans.
And The other non-human primates would be like kind of cut out from this sort of continuity.
So one possibility that we're really interested in looking at a non-language trained primate to see whether they have the spontaneous tendency to do the same match, which I would expect could be the case.
And also, as we said, in humans.
We do find the effect, irrespective of language, irrespective of culture.
Nonetheless, it is possible that experience has a role in shaping, like in strengthening or weakening the association.
So it will be nice to see, now that we know that there is a predisposition, to see how this predisposition interacts with life events.
Fascinating.
Well, Dr Maria Lacansole from the University of Padua in Italy, you know, I think that this research probably is going to be taught in like every introductory linguistics class from now on.
So thank you so much for joining us and explaining how you did the work.
Thank you.
Thank you very much for having me.
Thank you.
