0:01

Over the past 40 years, the number

0:03

of people in the United States who

0:05

speak a language other than English at

0:07

home has nearly tripled to 68 million.

0:10

That's about one in every five Americans. And

0:13

in many other parts of the

0:15

world, bilingualism and multilingualism are even

0:17

more common. In fact, more

0:20

than half the world's population speaks more

0:22

than one language. In

0:24

recent years, scientists have begun to explore

0:26

what's going on in the brains of

0:28

people who are bilingual or multilingual.

0:31

And they found evidence that speaking

0:34

multiple languages may have cognitive benefits

0:36

that include improving people's executive

0:38

function abilities and slowing down

0:40

the progression of age-related dementia.

0:43

Researchers are also learning more about how

0:45

the bilingual brain processes language and

0:47

how the languages we speak shape

0:50

the way we think and perceive the

0:52

world. So why

0:54

might speaking multiple languages have

0:56

such far-reaching benefits? How

0:59

does the brain process more than one language at

1:01

a time? How do

1:03

the languages we speak affect the way we see

1:05

the world and the decisions we make? And

1:08

what counts as being bilingual? Are

1:10

people who are fluent in computer code or

1:12

even math or music bilingual? And

1:15

finally, if you're interested in the benefits of learning

1:17

a new language, is it ever too late to

1:20

try? Welcome

1:22

to Speaking of Psychology, the flagship

1:24

podcast of the American Psychological Association

1:26

that examines the links between psychological

1:28

science and everyday life. I'm

1:31

Kim Mills. My

1:35

guest today is Dr. Veerika Marion,

1:37

a professor in the Department of

1:39

Psychology and Communication Sciences and

1:42

Disorders at Northwestern University.

1:45

She's a psychologist and cognitive

1:47

scientist whose research focuses on

1:49

bilingualism and multilingualism. She

1:52

studies the relationship between language and memory,

1:55

as well as how people process spoken

1:57

and written language. Dr. Marion is the Author of the

1:59

book, The American Psychological Association. After of the recent

2:01

book The Power of Language how the

2:03

codes we used to think, speak and

2:05

live transform our minds That marion thank

2:08

you for joining me today. Think people

2:10

having need to be here. You

2:12

open your book with an epigraph.

2:14

To have another language is to

2:16

possess another soul. Why? Did

2:18

you choose that quote? What does it mean to you?

2:21

A suit of clothes decline interests

2:23

like it's all black. People.

2:26

Who speaks To have more

2:28

languages have to experience? That's

2:30

when they you think different

2:32

languages they become slightly different

2:34

vs himself with each language

2:37

bringing to the full friends

2:39

different cultural experiences, different memories

2:41

isn't ethics of their identity.

2:44

A friend. Relationships.

2:46

And in an inner we are

2:48

serving as a filter to which

2:51

they see the world and interact

2:53

with. Themselves. With

2:55

others and with with the world at large. Now.

2:58

You speak multiple languages including

3:00

Romanian, Russian, and English. How

3:02

did your own early experiences of

3:04

language lead you to the

3:06

work that you doing now.

3:08

Yeah so I grew up in the

3:11

family that spoke Romanian at home in

3:13

one of the former Soviet republics way

3:15

a rush and with the official language

3:18

outside the home so I'm interest in

3:20

childcare and damn growth is also million

3:22

invested in parallel and leading school. I

3:24

started to set the English than in

3:27

college friends ah that includes some Spanish.

3:29

I married into a judge family and

3:31

ended up really been exposed to a

3:33

variety of languages which as you mentioned

3:36

earlier, introduction, theory, com and not only

3:38

in Europe. But it needs different parts of the

3:40

world. In interest.

3:44

Me: Much every continent people often

3:46

feet onto a woman was just

3:48

too early childhood and then apply

3:51

addition the languages often out later

3:53

in life related theory Common for

3:55

the human condition to really be

3:58

a bilingual multilingual one. When

4:00

you call someone bilingual, does that

4:03

mean the person grew up speaking

4:05

two languages or can you become

4:07

bilingual later in life by

4:09

studying another language and becoming fluent? In other

4:11

words, I'm looking for a definition. You

4:14

can become bilingual at any age. You

4:16

don't have to grow up with two

4:19

languages to be bilingual. You can learn

4:21

a second language in childhood, in adolescence,

4:23

in adulthood and even in your older age.

4:26

A bilingual is a person who can

4:28

use two languages. A multilingual is a

4:30

person who knows and can use more

4:33

than two languages. The

4:35

definition of bilingualism is broad

4:38

and it depends really

4:40

who you talk to, but it's the

4:44

ability to encode

4:47

our thoughts, our ideas, concepts

4:49

in more than one symbolic system

4:51

and then be able to transmit

4:53

across time and space these

4:56

thoughts, these ideas and have someone else

4:58

decode them. So the more symbolic systems

5:00

you have at your disposal, the more

5:03

languages you know. What

5:05

about people who can understand a second language but

5:07

they can't actually speak it or they speak a

5:09

little of it with great difficulty? I'm thinking of

5:11

this, my mother grew up in

5:13

an Italian family at a time when you weren't

5:15

supposed to speak Italian except in the home and

5:17

they didn't want their kids to speak Italian, but

5:20

the kids understood what the parents were saying. Do

5:23

you get any of the brain benefits of

5:25

being bilingual if you are in that situation?

5:28

Yes. So if you are referring

5:30

to what's known as heritage bilingualism.

5:33

Bilinguals who grew up in a household

5:35

where there is a heritage language that

5:37

their parents or grandparents or their previous

5:40

generations of their families spoke that

5:42

may be different from the dominant

5:44

language outside the home. And

5:48

there are many heritage bilinguals in the United

5:50

States. The

5:53

extent of their fluency in

5:55

the two languages varies

5:57

but they do indeed show many

5:59

of the of knowing

6:01

two languages that children and who

6:03

grew up with two languages, or

6:05

people who are fluent in both

6:08

languages, show as

6:10

well. So it's

6:12

really the type of language

6:14

experience that you have that

6:16

can make your bilingual or multilingual. And there

6:19

are different kinds of bilinguals and multilinguals. Oh,

6:22

let's talk for a minute about some

6:24

of the benefits that you researchers have

6:26

found by studying people who are bilingual

6:28

or multilingual. What are the

6:31

kinds of benefits that they experience in the way

6:33

that they think, in the way that their brain

6:35

functions, and maybe even is structured? Yeah.

6:37

So one of the most striking

6:39

and interesting findings in recent

6:42

years is that people

6:44

who use two or

6:46

more languages show

6:48

a delayed onset of Alzheimer's

6:51

and other types of dementia.

6:53

They are diagnosed, on average, four

6:55

to six years later than monolinguals,

6:58

depending on the study that you

7:00

are looking at. It seems

7:02

that knowing

7:05

two or more languages offers

7:07

some of those protective benefits

7:11

against the cognitive decline

7:13

that's sometimes seen with

7:15

healthy aging and that's

7:17

regularly seen with dementia.

7:20

And not only at the individual level, but

7:22

at the population level, there are now studies

7:26

that show that countries

7:28

in which two or more

7:30

languages are official languages of the country

7:34

have a lower incidence of dementia

7:36

than countries where one language is

7:38

officially linked to the country. And

7:40

there is actually a direct correlation

7:42

between the non-bold languages spoken

7:45

in the country and the incidence of

7:47

Alzheimer's. So it's a really interesting

7:49

new finding showing that the consequences

7:51

of being bilingual or multilingual

7:54

are not only at the level of the

7:56

individual, but also at the level of society.

8:00

the power of language, I organized

8:02

the, the

8:05

book is structured into two parts. The

8:07

first one talks about the self and

8:09

then the second one talks about society

8:11

and how both are changed by knowing

8:13

two or more languages. In

8:15

the case of dimension, what

8:18

seems to be happening is that the

8:22

constant workouts that our brains

8:24

receive by having to juggle

8:26

multiple languages is like

8:29

an exercise for the brain and has

8:31

this positive consequences.

8:35

So one way to

8:37

describe this is imagine

8:39

that you have been taking the same bridge

8:41

or the same road home every day for

8:43

many years throughout your life. You go to

8:45

work, you come home, you take this road,

8:48

you go to

8:50

the store, you go to the pharmacy, wherever you're

8:52

going, you go back and forth, you always take

8:54

the same road. And then one day when you're

8:56

way back, you discover that that road has collapsed.

8:59

If that's the only road that exists to

9:01

your house, then you're in trouble. You won't

9:03

be able to reach your destination. But

9:06

if there are multiple roads that have been

9:08

built over time, if you know those roads,

9:11

then you simply reroute your

9:14

car and you're able to reach your destination

9:16

with no problem. In the same way,

9:18

if there are multiple languages

9:20

that you've used throughout your life to encode

9:24

memories, to encode experiences, to

9:26

learn, to live your life in,

9:29

you can compensate functionally through

9:32

using those languages for

9:36

the anatomical deterioration that your brain

9:38

experiences. So it's not

9:41

that the bilingual or multilingual

9:43

brain doesn't experience deterioration, it

9:45

does. But people who

9:47

don't have a normal language are able to compensate

9:50

functionally and live a

9:53

healthy life as if

9:55

they don't have the same degree

9:57

of impact for longer. to

10:00

six years might not seem like a lot when you

10:02

are, I don't know, 20,

10:05

18, but when you are older, four to

10:07

six years is really a big difference and

10:11

can make a difference from being

10:13

able to play with your grandchildren or

10:15

never knowing them. And

10:17

there is nothing else that we know

10:19

of other than perhaps exercise that

10:22

has a strong effect on,

10:25

a protective effect on cognitive decline

10:27

and Alzheimer's and dementia as a

10:30

safe exercise of bilingualism and

10:32

multilingualism. So it's

10:34

one of the most remarkable recent

10:36

findings in the field. So exercise

10:39

is up there, but what about some of

10:41

these other mentally enriching

10:43

things that we engage in, such

10:45

as maybe having more years of

10:47

education or a challenging hobby? You're

10:49

saying that being multilingual gives you

10:51

greater cognitive benefits than those things?

10:54

Those are all good and you're right that

10:56

a level of education is a plus,

11:01

nutrition is a plus, engaging

11:04

and enriching activities is a plus. All

11:06

of those things are very

11:08

good. But based on my

11:11

review of the literature looking at

11:13

meta-analysis, it seems like that effect

11:15

is a large

11:17

exercise in bilingualism. And

11:20

what's really interesting about bilingualism and multilingualism

11:22

is that you don't have to take

11:25

time out of your life to

11:28

engage in those activities to see

11:30

benefits. So for example, we know

11:32

that engaging in cognitively enriching benefits,

11:35

cognitively enriching activities like sudoku

11:39

or word puzzles or math or just

11:42

any kind of cognitively

11:44

challenging activity is

11:47

beneficial for the brain. And it's

11:49

a good thing to do as we age for our

11:51

cognitive health. But for all

11:53

of those, you have to actively take time out

11:55

of your life to

11:57

do those things, whether it's reading and of course reading is

12:00

wonderful for you, whether it's doing puzzles,

12:02

that's all great. But with being bilingual

12:05

and multilingual, you simply go about your

12:07

life using one language or another. And

12:10

you are constantly giving your brain a workout

12:12

because your brain has to facilitate the language

12:14

you are using and inhibit the language you

12:17

are not using and sort of control the

12:19

languages you are using at any given time.

12:21

For example, you and I right now are

12:23

speaking in English. Our

12:25

brains have to control our output. It would do

12:27

us a little good if I suddenly would switch

12:29

to Spanish or French or Romanian or Russian. So

12:32

I have to make sure that I control those. I

12:35

inhibit those and I facilitate just English.

12:38

And you do this all the time, whether you

12:40

are reading a road sign, you're listening to a

12:42

movie, a song, you're

12:44

speaking with someone, you're

12:46

constantly juggling your many

12:49

languages. So this constant experience,

12:51

juggling different languages, seems

12:54

to be having an effect

12:56

on your executive function that

12:58

later translates to benefits

13:01

beyond just language. And I have to say

13:03

that this area of research, not

13:05

the dementia part, but the

13:07

executive function part is somewhat

13:09

controversial. Most all studies find

13:11

benefits of bilingualism on executive

13:13

function. And

13:15

sometimes people up in arms about that. I

13:18

am not up in arms about that and I can tell

13:20

you why. The reason I'm not up

13:22

in arms about that is because bilingual experience

13:25

varies. There are different kinds of

13:27

bilingualism. Executive function is

13:29

an umbrella term that includes multiple

13:31

components. So I think asking the

13:33

question in this absolute ways,

13:35

does bilingual benefit executive function or more,

13:38

this black or white approach is not

13:40

the right way to look at things.

13:42

A better question would be to ask

13:45

what aspect, what kinds of language experience

13:49

influence what kinds of executive function

13:51

and under what circumstances? It's

13:54

sort of like thinking about exercise. I

13:56

think universally everyone knows that exercise is

13:58

good for you. We agree. that exercise

14:00

is good for your health and for aging. But

14:03

it's not like every single type of exercise

14:05

you do will benefit

14:07

every single type of your health under

14:10

every single amount and

14:12

duration. And there are just

14:14

so many other variables. Some things benefit, some

14:18

variables, others benefit, other variables. So

14:20

taking this absolute work-on-wide approach is

14:22

really not a good

14:25

way to do science. I know

14:27

that we all like simplified yes-no

14:29

answers, but it's a much more

14:31

nuanced, complicated issue. I

14:34

want to talk for a minute about what

14:36

goes on in the brains of people who

14:38

know multiple languages. Are all of

14:40

the languages you know always active or

14:42

is your brain shutting off and sectioning

14:44

things so that you know that we're

14:46

speaking English right now, you're not going

14:48

to slip into Romanian by accident? Yeah,

14:52

we used to think that people switch

14:54

between languages. There was this language

14:56

switch hypothesis where they would believe

14:58

that when you use one language,

15:01

you're using it, then you're done with it, you switch it

15:04

off, you switch the other one on, you use that other

15:06

language. And there's also

15:08

the belief that different languages are represented in

15:10

different parts of the brain. So

15:13

you have like a place for English, a

15:15

place for French, a place for Hebrew, a

15:17

place for Arabic, whatever language you are using.

15:21

We now know that's not the case.

15:23

We know that language

15:25

is distributed through multiple

15:28

areas of the brain, it's a network.

15:31

And we also know that we don't ever

15:34

really switch off our

15:36

languages. They are constantly co-activated

15:38

and running in parallel. So

15:40

our brain is this very

15:44

impressive super organism that processes

15:46

information in parallel. It's all

15:48

time from multiple modalities. And

15:53

we know this through multiple lines of research,

15:55

including research that uses eye tracking that

15:57

shows that people who speak multiple languages. languages,

16:00

when they hear a word in one

16:02

language, they often make eye movements to

16:05

objects in their visual display, in

16:07

their visual scene, that sounds similar

16:09

in another language. So, for example,

16:11

if you are an English speaker

16:13

and I sit you in front

16:15

of a bunch of objects on

16:18

the desk and I tell

16:20

you to pick up a marker while recording

16:22

your eye movement, as you pick

16:24

up the marker, you will often make eye movements

16:26

to other objects like marbles,

16:29

for example, that start with the

16:31

same word onset like marker and

16:33

marbles. But if you speak

16:35

Spanish, you'll make eye

16:37

movements to other objects that start with

16:39

the same onset in your language like

16:42

a butterfly, the Spanish word for butterfly

16:44

is mariposa. If you speak

16:46

Russian, you may make eye movements or you

16:48

do make eye movements to a stamp because

16:50

the Russian word for stamp is marca. So,

16:52

as you hear, marker, you

16:54

make eye movements to a stamp. So,

16:56

people who speak different languages make eye

16:59

movements towards different objects as they hear

17:01

language, different objects

17:03

in their display, in their visual

17:05

scene attract their attention. And

17:08

we are now even finding that they

17:11

remember what they had seen differently

17:13

depending on the languages they speak.

17:15

So, if you ask a

17:17

Russian speaker, they are much more likely to

17:19

remember that there was a stamp on their

17:21

desk when they were looking for a marker

17:24

than an English monolingual speaker. And the Spanish

17:26

speaker is more likely to remember that they

17:29

saw a butterfly flying around when they were

17:31

looking for a marker than if you don't

17:33

speak Spanish. It's really

17:36

interesting because it shows us that the languages

17:38

we speak influence

17:40

what we see, our

17:43

perception of the world around us,

17:45

our perception of reality, and influences what we

17:47

remember later when we think about our environment.

17:52

So, think back to experiences we've had.

17:54

The language is this very

17:57

powerful experience that we...

18:00

filter our world through in our life. If

18:03

you grow up in a bilingual household, at

18:06

what point do you realize that

18:08

there are two different languages being

18:10

spoken instead of one big language

18:12

that encompasses both? That's

18:15

a good question because it taps

18:17

into what's known as metalinguistic ability.

18:20

Metalinguistic ability is the ability to reason

18:23

about language abstractly. And

18:25

we all as adults develop

18:27

it eventually. So adults know

18:29

that you can call something,

18:32

but by any word you want, and

18:34

it doesn't change what this word is, I

18:36

might have a pen,

18:38

and I might call it a pen

18:40

or a ruchka or a

18:42

stilou, or however you might call it, and

18:44

it doesn't change it. But

18:48

bilingual kids develop this ability to

18:51

reason about objects and

18:53

their names in the abstract ways earlier

18:56

than monolingual kids. At an early

18:58

age they understand that the

19:00

world and the words we used to refer to

19:02

the world are not one and the same, that

19:04

you can change the labels and it doesn't change

19:06

the object. So this is

19:09

really a foundational skillful cognitive

19:11

development, being able to read an

19:13

abstractly about our

19:16

environment and the labels we use to refer

19:18

to it. In your book you write

19:20

about how for people

19:22

who are bilingual the experience of

19:24

speaking in their native versus their

19:27

non-native language may be quite different.

19:29

For example, people may make different

19:31

moral decisions or access different memories

19:33

when they're talking in their native

19:36

versus their non-native language. Can

19:38

you talk about that? What are the

19:40

connections among language, memory, and emotion that

19:42

are all at play? This

19:45

is a really rich topic

19:47

because it does make a

19:49

big difference in our decision

19:51

making and the emotions we feel and

19:53

how we think about our identity and

19:55

how we interact with others,

19:58

depending on which language we use. we're speaking

20:00

at any given time. There is

20:02

some evidence suggesting that the native

20:04

language tends to be

20:07

a more emotional language, tends to

20:10

be tied to emotions more

20:12

closely, whereas the

20:14

second language often can be

20:16

a more logical language, or

20:19

lead to people making decisions

20:23

that are more logical and

20:25

utilitarian. So in one of

20:27

these areas of research, scientists

20:30

were asking bilinguals to make moral decisions

20:32

in either their native language or their

20:34

second language. And I'll give you an

20:36

example. You may have heard of the

20:38

very famous Strolli dilemma. And

20:41

there are different versions of it, but in

20:43

this particular version of the Strolli dilemma, a

20:45

Strolli is coming down the tracks and there

20:47

are five workers working on the tracks and

20:49

the Strolli is about to run over them

20:51

and kill them. And you are standing on

20:54

the foot bridge above

20:56

those tracks next to a large

20:58

person with a large backpack.

21:01

And if you push the person off

21:03

the footbridge, the person will

21:05

die, but that will stop the

21:07

Strolli and will save the lives

21:09

of these five people. So

21:12

the question is, is it permissible to

21:15

sacrifice the life of one person to

21:17

save the lives of five people? How

21:19

would you answer that? I

21:21

hate that dilemma. I don't know

21:23

what I would do. I think I would not be

21:26

able to push somebody off a bridge. I

21:28

hate the dilemma, too. And I know what you

21:30

mean. And your answer is really

21:32

just at the end. You can change. A

21:34

lot of people would say, yeah, I can

21:36

see how it makes more sense to sacrifice

21:38

one person for five, but I would never

21:42

push a person off the bridge or kill a person.

21:45

Some people would be more likely to do

21:47

it if all they have to do is push a button

21:49

versus push a person. And then it

21:51

depends with five people versus five thousand

21:53

people. There are so many versions of

21:56

this problem. But interestingly,

22:00

bilingual to make this decision

22:02

in their native language versus

22:04

their second language, you

22:06

see differences in how they answer. They

22:08

are much more likely to be guided

22:11

by what's known as geontological values, to

22:13

what's inherently morally right or wrong in

22:15

their native language and

22:17

say that they are

22:21

less likely to sacrifice the life, to kill

22:23

one person, to sacrifice the life of one

22:25

person for five when they're speaking their native

22:27

language versus when they're speaking their second language.

22:30

When they're speaking their second language, they are

22:32

more likely to make a utilitarian decision and

22:36

decide that you

22:40

go with the decision

22:42

that has the greater good. So this

22:44

is one set of experiments, but people

22:46

have looked at the likelihood of cheating,

22:48

at medical decisions, at all sorts of

22:51

decisions, and they are finding that bilinguals

22:53

tend to make different decisions in one

22:55

language versus another. And

22:58

it does seem that

23:00

the first language just tends to

23:02

be more closely tied to emotion.

23:05

And I think many individuals who speak

23:07

two or more languages often experience it

23:09

themselves, that they feel a different reaction

23:12

when someone says, I love you to

23:14

them in their native language, or when

23:16

they say it themselves in their native

23:18

language versus their second language. So that's

23:21

the tip out there for your listeners

23:23

who may be in the relationship with

23:25

bilinguals or multilinguals to learn some of

23:28

these key phrases, emotional phrases in

23:30

the native language of the person you are

23:32

in a relationship with to help build a

23:34

deeper connection. And

23:37

it's also curse words. People there, bilinguals

23:39

differ in their likelihood of being able

23:42

to curse or how worked up to

23:44

get when they hear curse words in

23:46

their native language versus second language. Many

23:50

bilinguals, myself included, will tell you that they

23:53

could never curse in

23:55

certain, you

23:58

know, powerful in their native language. and

24:00

have no problem doing that in their

24:03

second, third, fourth language where you don't

24:05

have just such strong connection and emotional

24:09

reaction. So

24:13

there is this relationship between the language you

24:15

speak and how you feel, what you remember,

24:17

the emotion. We could talk

24:19

about this for so long because there is just so

24:22

much interesting research on

24:25

this topic. And

24:27

I go into a lot of it in the book, but

24:29

I'm happy to talk

24:32

about some of the examples here. I

24:35

want to ask something a little bit out

24:37

of left field maybe. We did a podcast

24:39

not that long ago about people who can

24:41

speak dozens of languages. And I'm wondering in

24:44

all the research you have done, is

24:46

there a limit to the number of

24:48

languages that a human brain can

24:50

learn? Not that we know of.

24:53

And I think this is one of

24:55

the reasons why I'm so interested in

24:58

multilingualism. And the core

25:00

of it, I'm interested in language

25:02

and thought, language and mind, how

25:04

closely related language and

25:06

thought are, what does language learning

25:08

tell us about learning in general

25:10

because there does seem to be

25:12

a relationship

25:15

between, you know,

25:17

when each new language you learn becomes

25:19

easier to learn additional language. So

25:22

it's this virtuous cycle. The more you

25:24

learn, the easier it becomes to learn

25:26

more things in

25:28

languages. So there are indeed

25:30

multilinguals who speak 2, 6, 20. There

25:36

have been claims of over

25:38

100 languages. And it

25:40

of course also depends on what you define

25:42

as a language and what you define as

25:45

fluency. But the

25:48

human mind seems to be able to

25:52

learn languages,

25:54

you know, ad infinitum really.

25:57

And it's really interesting to

25:59

compare. artificial intelligence and human

26:01

intelligence and use multilingualism as

26:04

a way to study the

26:08

human potential. What

26:11

is language, if I can ask

26:13

that, how do you define it?

26:15

Is computer code language? Is math

26:17

language? Is music language? I'm

26:20

glad you asked it because I think

26:22

most people use a definitional language in

26:26

a somewhat narrower sense than I tend to

26:28

think about it. Usually

26:30

when most people think about language,

26:32

they think about natural languages like

26:34

English or Spanish or French or

26:37

Hindi, a language that a group

26:39

of people can speak and communicate,

26:42

an old sign language. If

26:46

we broaden the definition of language to

26:51

use it to refer to a

26:53

symbolic system, so language

26:55

is something that uses symbols to

26:59

communicate information, then we

27:01

can go beyond these natural human

27:03

languages. It is

27:05

an app definition because as we mentioned

27:07

earlier, what languages do is they use

27:09

symbols, words,

27:12

or notations to

27:15

encode the thought and idea

27:17

of concepts transmitted over space

27:19

and time for something

27:21

at the other end to decode it. If

27:27

we use that broader definition, they're

27:29

not just natural languages, they're not

27:31

just sign

27:34

languages, which are

27:36

also natural languages, but also computer

27:40

languages, artificial languages, and

27:42

math. Music can also be

27:44

conceived of as symbolic systems and languages.

27:47

Many will tell you that math

27:49

is a language. It's one of the

27:51

most powerful languages out there. Math

27:54

is the queen of sciences that allows

27:57

us to encode really complex ideas.

28:00

that have allowed humanity to

28:02

make discoveries and build things

28:04

that we wouldn't be able

28:06

to do without math. So,

28:09

if we

28:11

broaden the definition of language to

28:13

symbolic systems, then most of our minds,

28:16

even if we only speak one natural

28:18

language, human natural language, have

28:21

multiple symbolic systems at our disposal because

28:23

we can read math and

28:25

use math. Perhaps we can read music. And

28:28

each of those symbolic systems changes

28:30

the way we think. If

28:32

you have math and you can

28:36

read about very

28:38

big numbers that we couldn't

28:41

read about centuries ago, or

28:44

you have very small

28:47

numbers, again, you can

28:50

use math

28:52

and language to now talk about

28:54

things that the eye

28:56

cannot see and

28:59

create mental concepts that you would

29:02

not be able to create otherwise.

29:04

So, language

29:07

is key

29:10

to scientific discovery, to

29:12

thought, to technology, to science, to

29:14

human advancement. It's one of the

29:16

most powerful tools we have at

29:19

our disposal. We

29:21

all know that it's easy or easier

29:24

for kids to pick up a language

29:26

compared to adults. But

29:29

can a person learn a language fluently

29:31

at any age? We

29:36

used to think that there is this critical

29:38

age period. And if you

29:40

miss this critical age period, too

29:42

bad. You won't be able to learn a

29:44

language. And you won't be able to

29:46

learn a second or third language. You

29:50

can learn a second language, or third

29:52

language, or multiple languages. And

29:54

you can learn them fluently at any age.

29:57

Now, you may have an accent.

30:00

learn your second or third language later

30:02

in life after puberty after

30:04

your articulatory system is set. Not always. Some

30:06

people don't have accents, but many do. You

30:08

can hear me speak with an accent in

30:11

English even though I would consider myself fluent

30:13

in English. I do

30:15

have an accent because I learned

30:17

English later in life, but

30:20

I'm fluent in it and you can learn

30:22

a language to fluency at any age, but

30:24

even if you don't learn a language to

30:27

fluency, I would almost say it

30:30

helps to remove that pressure of yourself

30:32

and say, well, I can't be fluent. I will

30:34

always have an accent. I'll never be fluent in

30:36

another language, so maybe I shouldn't even bother. I

30:39

think that's the wrong approach. You can see a

30:41

lot of consequences and

30:44

benefits of learning another

30:46

language even without being

30:48

fluent. Most

30:50

of all, it can actually improve your

30:52

quality of life in many different ways

30:55

depending on how you choose to learn

30:57

another language, but if you choose

31:00

to learn another language by engaging

31:02

in activities that you enjoy that

31:04

bring you pleasure and joy and

31:06

enrich your life and you find

31:08

a way to learn another language

31:11

that makes you happy, be it

31:13

by exchanging language lessons with another

31:15

person. We know that social interaction

31:17

is so beneficial for mental health,

31:19

physical health, a lot of positive

31:21

consequences, or by immersion,

31:23

by traveling, living in another country,

31:26

and you enjoy that, then the quality

31:28

of your life will be improved by

31:30

that, or maybe you enjoy learning

31:34

using apps. Many of

31:36

the apps right now capitalize

31:38

on the knowledge we have from neuroscience

31:42

by having this reward circuit in

31:44

the brain activated all the serotonin.

31:46

You play this game and you

31:48

get a reward of some sort,

31:50

whereas the badge or sound and you move

31:52

through levels and you enjoy

31:54

doing that. So whatever it is, maybe you enjoy

31:56

listening to music in another language, or watching a

31:58

movie in another language. So, there

32:01

are still many reasons to learn

32:04

another language that may make your

32:06

life more enjoyable and

32:08

better that even if you

32:10

don't become fluent is

32:14

worthwhile. I

32:16

want to talk for a minute about an op-ed

32:18

that you wrote that was

32:20

published in the Washington Post not

32:22

long ago that talked about how

32:24

artificial intelligence may decrease language diversity

32:26

and multilingualism in the future. Why

32:30

do you believe that? So,

32:33

I will pull

32:36

back by saying that

32:39

most artificial intelligence that

32:41

is now making the

32:43

news as it is so

32:45

much relies on what's known as large language

32:47

model. And these large

32:49

language models follow the same principles

32:51

that linguists and psycholinguists, psychologists of

32:53

language have been studying for decades.

32:57

I don't think that many people

33:00

anticipated that the advancement of the

33:02

large language models will happen as

33:04

fast as it has and that

33:06

they will be able to do

33:08

it as much as they can

33:10

right now already based on essentially

33:13

learning and using statistical

33:15

probabilities and their input at the

33:18

base. But

33:20

now that AI is there and it's

33:23

developing so rapidly and we have this

33:25

evolution of AI, evolution of knowledge at

33:27

a very rapid pace faster

33:30

than previous generations, it's

33:35

everyone's guess what will

33:37

happen next. And people vary in the

33:39

predictions they make and their

33:42

predictions are informed based

33:44

on their knowledge if you're a computer scientist,

33:46

if you're a learning theorist, I'm

33:48

a psycholinguist so my predictions are based

33:51

on what I know about language. And

33:56

what I will say to that is

33:58

that our initial predictions, my initial expectations

34:00

are expectation and reaction would be to think

34:02

that having this

34:05

LLM will be beneficial to language

34:07

diversity because we think, well, now

34:09

we can use online translation tools

34:12

that can immediately translate text, that

34:14

can translate language input. Here's my

34:16

phone and I can speak into

34:19

it and will immediately translate when

34:21

I'm traveling into the language of

34:24

my listener. So these are all

34:26

wonderful things and certainly great benefits.

34:28

You can also now document, record

34:31

and save many languages that are

34:33

dying. So the initial

34:37

thought is that it will be beneficial and we

34:39

do have those beneficial components. But

34:42

I think long term my

34:45

prognologist for the diversity of natural

34:48

languages is the

34:50

result of AI is

34:53

grim. And here's

34:55

why. There are more than 7,000

34:58

languages spoken in the world today. We

35:01

lose a few every year. In

35:04

theory, you could have LLMs,

35:07

you could have large language models based

35:09

on all of these 7,000 class

35:11

natural languages. However, the

35:13

way LLMs work is they

35:16

learn based on the

35:18

digital footprint that the language has

35:20

online. So the more information, the

35:22

larger the database that exists online,

35:24

the better those language models will

35:26

be. And we do have a

35:28

huge linguistic footprint in English, in

35:31

Mandarin, in French, in several

35:34

languages. So those

35:36

models learn on a

35:38

much larger input and

35:41

therefore they become better than large

35:43

language models that learn on

35:45

a much more limited input.

35:49

And predictions vary on how many

35:52

LLMs are likely to become really

35:55

powerful. Some conservative estimates

35:57

are at around 20. even

36:00

if we broaden it from 20

36:02

to let's say 200 or 2000, it's

36:06

not realistic to think 2000, but it's

36:09

still more 7000 plus. So

36:11

most likely, only a very

36:13

small relatively to the 7000 plus languages

36:19

will have a large

36:22

enough digital footprint to result in

36:24

very good strong LLM

36:28

language models. And you can think

36:30

of this almost as an invasive

36:32

species that then drives out the

36:34

small languages. And

36:36

of course, we're not talking about tomorrow

36:38

or a year from now, we are talking about

36:41

a few

36:43

generations. But, and

36:45

this is, and some might say much less

36:47

than a few generations, it remains to be

36:49

seen because human nature essentially, we will always

36:52

want to use the best, the strongest, we'll

36:54

always want to use the best model out

36:56

there. So eventually, a few models will win

36:58

out. So if we take

37:00

it, if we push this argument really

37:03

further, almost to the like a

37:05

science fiction, film

37:08

direction, you

37:10

can imagine a situation where a lot

37:12

of the small languages are disappearing at

37:15

a much faster pace. And

37:17

if language and thought are

37:19

indeed interconnected, with

37:21

this many languages dying

37:23

out, we'll lose some

37:26

diversity of thought, we use the culture,

37:28

the concept, the differences in the way

37:30

people think, as a result of these

37:32

dying languages. So we

37:35

will simultaneously potentially have

37:38

two competing things happening at the same

37:40

time. Diversity of languages,

37:42

natural languages disappearing, diversity of

37:44

human thought disappearing, maybe having

37:47

more uniform thinking as a

37:49

result of only several large

37:51

LLMs, while at the same

37:53

time as the decreased

37:55

diversity of languages and thoughts,

37:57

the artificial languages And

38:00

artificial intelligence continues to evolve

38:03

and become stronger. And

38:05

that's not a great scenario

38:07

to think about for us

38:10

humans. Now we

38:12

don't know what will happen. Things can go

38:16

very different ways right now.

38:18

In just this past week,

38:20

very recently the European Union

38:22

started putting some guidelines in

38:24

place on artificial intelligence. I'm

38:27

sure most other countries

38:29

will follow, but it's

38:32

very difficult, if not impossible,

38:34

to regulate AI in

38:36

this kind of research. It doesn't take a

38:39

lot to have a rotation or

38:41

an individual group of individuals to

38:43

really not want to follow those

38:45

regulations. So we

38:48

are on the precipice

38:50

of big changes in

38:52

the evolution of

38:54

knowledge. And I think the

38:57

psychology of language is likely

38:59

to play an important role there. It almost

39:02

makes me think of the movie Arrival and

39:05

how important the ability to communicate is

39:09

for humanity. One

39:11

last question to wrap up. What

39:14

are the big questions you're looking at now? What

39:16

would you like to be able to answer in

39:18

the next six months or a year? We

39:21

have about a dozen experiments in

39:23

the lab right now. For

39:27

the past more than 20 years, I've

39:30

been looking at co-activation and parallel processing

39:32

of languages and bilinguals and multilinguals. And

39:35

now I'm looking at the consequences

39:37

of parallel processing for higher

39:39

level cognition, for higher level cognitive

39:42

processes like memory, like creativity, like

39:44

decision making, like semantic organization.

39:46

So does the fact that our

39:49

mind processes multiple languages in

39:51

parallel change as higher

39:55

level cognition? So I just mentioned the memory

39:57

study that was just published in Science. advances

40:00

this year showing that people who speak different

40:02

languages remember different things. So

40:04

now we are looking at the other consequences

40:06

for creativity, decision

40:09

making. We're also looking at the relationship

40:11

between form and meaning. So

40:13

are the, is

40:16

it completely arbitrary, the form of words

40:18

and the meaning they carry,

40:21

which tells us about the relationship between

40:23

language and thought. And this is another

40:26

question that has been around for centuries.

40:29

We're looking at, we

40:31

have a study right now on parents and children

40:33

in different countries who

40:36

are bilingual. And we are finding

40:39

that parents and children, toddlers and

40:41

young children, four-year-olds and their moms

40:44

interact with each other differently depending

40:46

upon which language they speak. So

40:48

as early as a preschool age,

40:52

children interact differently with

40:55

their caregivers and the caregivers interact differently

40:57

with the children depending on the languages

40:59

that they speak. So

41:02

we have multiple projects in

41:04

the lab. We're looking at

41:07

emotion processing. We're looking at

41:11

perception of truth and

41:16

susceptibility to misinformation across

41:18

languages. We have so many interesting

41:20

projects and there is enough work

41:23

for lifetime, many more lifetimes. I

41:25

think it's a

41:30

field that has been understudied

41:32

up until relatively very recently in

41:34

the history of science. Dr.

41:36

Marion, I want to thank you for joining me

41:39

today. This has been really fascinating. I appreciate your

41:41

taking the time. Thank you for

41:43

having me. I really enjoy talking

41:45

to you and I enjoy talking

41:47

about this issue. And I hope

41:49

that people who speak

41:51

two or more languages and even those who don't

41:54

enjoy the conversation we have

41:57

and want To dig in

41:59

deeper. Look at some of the

42:01

studies, read the book and that, and find

42:03

out more about the relation between language and

42:05

thought. I was all go

42:07

out there and learn a new

42:09

language. Absolutely. You can find previous

42:11

episodes of Speaking of Psychology on

42:13

our website at www.speaking of psychology.org

42:15

or on Apple's Spot of Ice

42:17

You Tube or where ever you

42:19

get your podcasts and if you

42:21

like when you've heard please leave

42:23

us a review. If you have

42:25

com answer ideas for future podcast

42:28

you can email us at Speaking

42:30

of Psychology at eighty a.org Speaking

42:32

of Psychologist. Gymnast.