Lecture 07

IIT Kanpur July 201
20 followers
Feb 26, 2024
[Music]
hello and welcome to the course
introduction to the psychology of
bilingualism and multilingualism I am Dr
arwma from the Department of cognitive
Sciences at ID kpur this is the second
week of the course and we are sort of
moving to the second lecture now you
might remember that in the previous
lecture we were talking about
categorical perception and the ability
to perceive fmic contrasts across uh
different languages and also in the
native language we said that this is
very very fundamental to the infant's
ability to perceive speech and acquire
language eventually in the current
lecture we will begin with discussing
the consequences of the same for infants
uh who are bilingual uh either in a
sequential bilingual sense for example
uh they are learning their first
language and then after some time the
second language comes in or in a
simultaneous bilingual setting where
they are exposed to both their languages
starting from the birth itself let's go
ahead now if a bilingual child is
introduced to the sounds of a second
language at a slightly later age the
ability to perceive phonic contrast is
already on the vein it is shifting from
more of a language General to a language
specific ability uh with more focus on
the contrast of what it considers the
native language so basically what will
happen is that perceiving the fomic
contract us of uh you know the second
language or the third will become as
much more difficult the further uh you
know you take this introduction of the
second or third language so children
have sort of you know you know with
increasing age are becoming generally
less sensitive to the non-native fomic
contrast including contrast that may be
meaningful for the second language so
basically it presents a sort of a
challenge uh you know given how the
organization of phonic perception is
happening with with in the children and
therefore it'll become a bit harder for
them to acquire the sounds of the second
language the further it happens from the
first language let's let's go ahead with
that now if children have to in order
for them to learn the sounds of the
second language uh you know uh or say
basically uh you know they have it it is
it becomes part of their ability to
perceive non-native contrast will need
to be restored if they have to acquire
uh these phic contrast from the second
or the third language basically what
we'll need is we'll need a little bit of
a reversal of this you know decline uh
of uh you know language General fomic
contrast uh you know perceiving ability
and in this instance the gr for example
takes the you know Deo uh takes the
example of a japanese-born child at 12
months of age who could no longer
perceive the difference between the L
and R sounds which are instances of the
same for in Japanese but basically are
are treated as separate categories in
English so obviously the later the onset
of uh you know second language learning
uh will be it will be that much more
difficult for these children uh to
acquire the sounds of a given second or
a third
language now K and colleagues basically
examine the conditions which would lead
to the reversal of dis ability to
perceive non for you know non-native
homic contrasts so what they did was in
12 laboratory sessions of around 2 5
minutes across spread across four weeks
they exposed a group of 9 and 10mon old
American
infants from only English speaking
families to Mandarin Chinese during
these sessions four different Mandarin
Chinese speakers read children's books
to the infants and they played games
with them exposing them to around 25,000
syllables from mandering Chinese whereas
for a control group the similar sessions
were conducted but by speakers of
English so here you're saying basically
is that you are uh while not explicitly
teaching language to Children you are
basically exposing them to the sounds of
a different language these are children
who are from only English speaking
American families but what is happening
is that during play during their
interaction they are getting exposed to
more than 25,000 syllables which are
part of Mandarin Chinese now post
training this ability to perceive um you
know uh contrast in Mandarin Chinese for
the two groups was tested using the head
turn Paradigm I hope you remember what
the head turn Paradigm is about if not
please go and refer to lecture five from
the first week now as experimental
stimuli during test two computer
synthesized speech sounds were chosen so
that they were contrastive in Mandarin
Chinese but not in English all right now
the results what did the results show
the results actually showed that infants
who were exposed to this interactive
sessions of mandarine Chinese across uh
you know the 12 across the uh 12
laboratory sessions of 25 minutes each
were able to actually perceive the
contrast in mandarine Chinese almost as
well as native Chinese infants so you
can see here that this is still a very
plastic ability it is still and because
we are talking about these very young
infants it is still an ability that can
be restored the ability to to perceive
non-native fomic contrast can still be
restored through practice and
instruction and interactive games and so
on now if you look at these findings
from a distance these findings actually
Provide support to the notion that the
decline in this ability to perceive
nonnative contrast can obviously be
reversed through uh you know systematic
exposure to that language so it is not
like if something is lost at least at
this very early age it is lost
permanently now in another similar
experiment which expose children to
audio visual material or just the audio
material actually it was found that the
same effect did not happen so the same
reversal of this ability of uh you know
perceiving you know this the reversal of
this inability of perceiving non-native
phic contrast could not be achieved when
the interactive sessions were replaced
by Audio material let's say just sounds
uh on a recorder were played or even
audio visual material where videos of
Chinese speakers were played so this is
also something very very interesting
that you know we need to keep in mind
that when we are imparting language
lessons to children or let us say when
children are themselves acquiring
language they do so much better in an
interactive true to life kind of
scenario rather than picking up language
from only uh you know audios or only
videos now let us look at how this might
play out in bilinguals okay and there
can be a few questions we can ask about
this for example what would happen in
the case of infants who are exposed to
to a bilingual environment from birth
what would be considered as the native
language which contrast the infants will
prioritize with the perspective of
learning now imagine say for example uh
you know in case of simultaneous
bilingual uh and I as I keep giving the
example suppose a child is born to
parents who speak let's say English and
French you know one parent speaks
English the other parent speaks French
and uh the language environment at home
would be constantly uh you know uh of
both these languages together okay now
what would happen in that case what is
the native language here which phonic
system will the child prioritize over
the other one uh and how will they
actually learn will there be an order of
learning here as as what happens in the
nonnative as what happens in the
sequential bilingualism uh you know
system so researchers who were trying to
look at these questions focused more on
the role of the statistical distribution
of the information available to the
infants from the the two or more
languages that are available to them so
the idea is that these infants are
probably paying attention to the
statistical properties of the language
input and remember the language input is
mixed there are both languages which are
coming in so what the infants are
probably doing here is they are paying
attention to patterns of you know speech
sounds in uh when they are being spoken
to in English or when they are being
spoken to in French or Chinese or
whichever the second language is now uh
for instance May and colleagues uh you
know in 2006 in a study with six and
8-month-old infants actually
demonstrated that infants were able to
exploit the statistical
distribution statistical distribution
information or speech sounds from the
language input to build these different
phonetic categories more specifically
what they did was that they exposed
these infants to either a bodal
distribution say for example wherein two
types of fonetic categories were you
know equally frequently presented or a
unimodal you know distribution system
wherein say for example only one type of
phon category was presented again and
again and what did they find they
actually find that infants in the former
condition were able to develop two
distinct ftic categories whereas infants
in the latter condition where they were
presented only with a unimodal
distribution were keep were able to you
know develop only one prominent category
so this sort of tells us that language
input actually plays a very very
important role when you're talking about
how are infants picking up language from
their environment the nature of language
input the regularities in the language
input although uh you know uh in the
previous lecture we were talking about
that categorical perception is innate
and this is something that we are Bor
with you can see here uh you know this
is a very solid example of the fact that
although the basic abilities might be
innate the input actually plays a very
very important role the statistical prop
properties of the input plays a very
very important role in how the child and
to what extent and to you know what will
be the nature of uh uh how children will
pick up language from their
environment let's go further now infants
who are raised in a simultaneous
bilingual environment are actually
exposed to two speech sound systems at
the same time now these sound systems
May either be very similar to each other
or they may differ from each other in a
number of different ways say for example
example in the carving out of fmic
categories based on you know voice onet
time values or maybe Rhythm or maybe
some other things so in that respect you
know my and uh colleagues demonstration
that infants are sensitive to
statistical distribution of speech
sounds from the input may serve as a
mechanism you know it may be a candidate
mechanism for acquiring speech sounds uh
uh you know basically in a way that you
know that that becomes the fundamental
source of information for children but
in case of simultaneous bilinguals we
can ask this question that okay how are
you know how are children actually doing
this which speech which regularities in
their input they are actually paying
attention to are they sort of doing
something like they're paying more
attention to First English and then they
SEC then uh French or let's say first
they L1 and then they L2 or in which
order and how are they actually
extracting this information because uh
obviously the input is mixed the
statistical patterns are probably too
many to decipher from so given that you
know the linguistic environment of uh
the simultaneous bilinguals is
considerably more complex than of
monolinguals or even sequential
bilinguals because in sequential
bilinguals also uh you're exposed to the
first language first and you know you
might be learning uh aspects of the
first language the second language only
comes later in simultaneous bilingual
you are having input from both languages
at the same time starting from birth and
it becomes that much more difficult for
people to sort of you know or for these
infants to maybe acquire these
categories from the two languages
simultaneously are these competing sound
systems or are they you know very
similar sound systems both will have
their own
challenges so uh using the head turn
procedure Bosch and Sebastian Gul uh
actually looked at the development of
Vel contrasts in catle and Spanish
infants uh for two groups of four and
8mon olds so when I'm saying catle and
Spanish infants we are talking about
infants who uh you know who are speaking
uh whose environment consists of both
Catalin and Spanish input and uh you
will see that the choice of age groups
is also interesting because remember uh
prior to 6.5 months the ability is more
language General so for one group of in
infants which is the 4month old infants
it is still that more language General
and for the other group of infants which
is after 7.5 months which is around 8
months old it is gradually moving to
towards more language specific
capabilities so uh as a control group
now as a control group four and 6month
old Spanish and catle and monolingual
children were also tested and the idea
was to be able to compare their
developmental pace and trajectory uh
with these infants which are four and
8mth olds of you know cattle in Spanish
uh bilingual infants so technically what
we have here in this experiment just to
sort of recount is that we have six
groups we have uh Spanish bilinguals 4
months and 8 months so two groups here
we have four month Spanish monolinguals
uh and 6 month Spanish monolinguals and
we have 4month Catalan monolinguals and
six Monon Catalan monolinguals so in all
what we have is we have six groups of
infants which are being sort of uh you
know uh compared with each other uh at
each age
group now the Vel contrast that the
experimenal chose to study was the
contrast between a and uh as in bait and
a as in bet now this is this vowel pair
is contrastive in Catalin uh so in
Catalin uh bit a sound or bet a sound is
actually different phonic categories but
not really in Spanish in Spanish they
basically instantiate the same vowel
category and is not really distinguished
so for example for Spanish adults uh
these two will be the same sounds
whereas for cattle and adults these two
will be slightly different sounds and
and infants were basically what we how
they were testing is that infants were
exposed to different instances of these
two sounds embedded in dbic nonwords and
why are they using nonword because a
they are you know uh sort of they don't
have any other familiarity uh clue of
familiarity to these children and uh
they were basically diic nonwords and
they were embedded in words and not
produced in an isolated fashion so that
things are slightly closer to to how
these fimes are actually encountered in
natural speech more specifically what
they did was that during the
familiarization half of the infants in
each age group were presented with a as
in bait uh and half of them were
presented with a as in bet whereas in
the test trials they were presented with
either the same one that were presented
in the test in the familiarization phase
or a different sound to that was
presented in the familiarization phase
so the Discrimination they have to make
again remember the head term procedure
is being used if I am correct yeah head
term procedure is being used they are
basically asking them to make these
discriminations and what we are sort of
going to see is whether the Catal in
Spanish bilinguals make this
discrimination well or whether the
Catalan monolist obviously should be
able to make this disc discrimination
Spanish monol linguist might not be able
to make this discrimination so we have
these three
predictions okay uh but around 4 months
of age the authors predicted that for at
least four months of age all the three
groups will be able to perceive this
monic contrast because you know there is
this ability there is this language
General ability of perceiving pic
contrast that is there up four four five
months of age however uh Spanish infants
at 8 months of age might have lost this
ability might have lost this generic
ability of making these pic contras so
technically they should not be expected
to be able to make this difference okay
so by this time they would have lost the
ability to pursue these Catalin specific
contrasts since these are not relevant
to
Spanish uh interestingly if you look at
the possibility for bilingual simult
these for these simultaneous bilingual
kids if only continued exposure to both
languages is important uh they should
have developed distinct categories for a
uh you know as in bait and e as in uh B
uh and should be able to perceive the
contrast so let us see alternatively it
it could also be possible that due to
the distributional overlap between the
cattle and a and uh you know in bit
versus a as in bet the and the acoustic
acoustically intermediate category that
happens in Spanish they could fuse
everything together and just come up
with a single phonic category let us see
what the results tell us the results
actually showed that for all the three
groups of 4month olds they were able to
perceive the phic contrast in a versus a
bait versus B uh which as uh you know we
were reasoning is perfectly plausible
also 8-month-old Catalin but not
8-month-old Spanish people could
perceive the contrast uh which is also
uh you know perfectly plausible because
we know that 8-month-old Spanish monol
lingal infants uh uh you know would not
be able to do this because they are sort
of moving more towards the language
specific uh fomic contrast than the
language General fomic contrast the
interesting pattern for results was
found with bilingual infants you know
the results were interesting because
they seem seem to have lost the ability
to perceive this monic contrast and they
behaved much like the Spanish
monolingual of the same age how is this
happening why is this happening will the
ability be restored again why are they
not being able to perceive the phonic
contrast that are specific to Catalin
because they are also Catal in Spanish
bilingu eventually by when would this be
restored if at all so to answer this
question uh they you know tested another
group of slightly older 12mon old catle
in Spanish simultaneous bilingual
infants and actually found that the
ability to perceive fomic contrast in
Catalin were restored by this age so
what we are seeing here is for
simultaneous bilinguals this ability of
being able to perceive the contrasts in
uh you know given language uh may be
slightly delayed you know it it may take
some time for the system to you know
pick up the contrast of one language and
uh then pick up the contrast of the
other language and then sort of you know
reach some kind of a Reconciliation
where they are able to keep these two
categories separate what might have
happened and again I'm speculating what
might have happened in this earlier age
is that they might have sort of uh you
know uh fused these categories together
and it is because it is not relevant in
Spanish uh they are not being able to
make these differences but as they grow
older as they move towards the
capability of developing language
specific contrast in Catalin and
language specific contrast in Spanish
they get back this ability to make these
uh you know distinction between the bait
and the bet a sound this is something
very similar to the author's conclusion
so they said that due to the
simultaneous exposure to both languages
Catalan and Spanish these infants may
have initially developed a single phonic
category of these a sounds and uh you
know as in bait and bet and hence uh
initially they failed to distinguish
between them however gradually has the
ability to perceive these contrast uh
you know becomes better uh they become
better in processing language specific
input uh they get this ability back a
bit with a delay of a few months so uh
you know interesting because it sort of
tells us that how is this mixing of two
languages since birth is impacting the
development of phonic perception in
these children similar demonstrations
were also documented in later Studies by
Bosch and GS uh in 2003 and 2005 and
based on these findings it could then be
concluded that cross language
distribution overlap indeed delays
slightly the building of language
specific contrastive categories in
simultaneous bilinguals it does not uh
impact it in a sense that they do not
have that ability it just delays it
basically you know you can say that the
system is buying time for itself to
acquire the language specific uh you
know uh you know fmic contrast for both
of their known languages now an
additional interesting proposal was made
by Sund and colleagues in 2008 who
proposed that infants are not only
sensitive to just the distributional
characteristics of the speet sound but
also to the frequency with which these
speech sounds occur in these respective
languages and uh how will they test
their hypothesis so what they did was
they compared the speech perception
ability of monolingual French English
and bilingual French English infants and
they basically compared them on their
ability to distinguish between the
French the and the English duh now in
French the sound is Du uh you know do
something or something like that and
where is English it is the very specific
du you can also see that in French the
DU is basically pronounced by you know
keeping the tongue just at the back of
the teeth whereas the English du is
pronounced by keeping the tongue in the
alular rich you know in the pal
so uh while the French and English du
sounds or the DU sounds are differently
produced you know the place of
articulation is different and this
difference can be phonic in some some
languages but interestingly what happens
in English is that these two sounds
instantiate the same fmic category so
technically while you're saying the DU
or the DU it is actually instantiating
the same phic category irrespective of
whether it is English or whether is
Dutch because there is a lot of overlap
still when you
test English adults are able to
distinguish between the French the and
the English du but French adults are not
capable of distinguishing between these
two so these given that these two
versions also given that these two
versions are very frequent in French and
English respectively uh it could be
expected that French English bilingual
infants might be expected to follow the
same strategy as adults would do so
basically they should be able to you
know uh make this distinction uh as uh
you know as their monolingual English
controls now for the experiment what did
they do was they actually had a visual
Hab fixation habituation procedure and
three groups of six uh to 8mon old uh
infants and three groups of 10 to
12month Old infants were tested uh
during the habituation phase half of the
participants half of the infants were
presented uh uh with the with the French
the tokens and half of them were
presented which were produced by French
monolingual speakers and the other half
were exposed to the repetition of
English dir tokens which were produced
by ver excis you know uh from uh
monolingual English speakers so they
were basically half of them were
familiarized with the half of them were
familiarized with du and again we were
basically asking them to make the uh
same different
decision the results actually showed
that since language specific percep
ction has not really developed by around
6 to 8 months of age infants from all
three groups French monolinguals English
monolinguals and French English
bilinguals were capable to make this
distinction between the French the and
the English du sound interestingly since
language per specific perception has
started to crystallize by around 10 to
12 months of age uh English and French
monolinguals actually showed very
similar patterns and were not able to
perceive the difference again something
very interesting uh as to you know what
we saw in the previous study now for the
results of uh bilinguals to be you know
these infant simultaneous bilinguals 10
to 12 month old if only cross language
distributional overlap would determine
the time course of phonetic development
these bilinguals should actually fail to
uh perceive the contrast uh you know and
they would basically be grouping the
incenses of of French and English du and
du together on the contrary IF frequency
really matters and because in English du
is much more frequent than the these
people should follow the example of
English adults and be able to
distinguish between the DU and the DU
sounds and indeed this is what happened
because uh the these bilingual kids were
actually able to detect the difference
much like English monolingual infants
and adults so we can see here that not
only the statistical distribution but
also the frequency you know uh of use of
these particular sounds in a given
language and which basically would
reflect in the input that they would
have received actually also
matters so uh to conclude this uh the if
you consider the joint results of both
types of studies uh uh you know from
sequential bilingual to simultaneous
bilinguals you could it it could
basically say that simultaneous
bilinguals develop some contrast at the
same pace and with the same trajectory
as their monolingual peers although the
development of few other contrasts may
be slightly delayed okay and may follow
a different route further both cross
language distributional overlap as well
as the frequency of occurrence of
specific speech sounds in these
different languages actually plays a
very important role in how these
individuals are going to acquire these
speech sounds all right so this is all
about uh you know how in uh you know uh
these phic contrasts are acquired uh in
sequential bilinguals and simultaneous
bilinguals and what are the specific
characteristics in the next lecture I'm
going to talk to you more about words
and how things are sort of how children
start you know to segment the words
stream thank
[Music]
you
[Music]
[Music]
