Module 7: How Babies Understand Communication Flashcards
Opening observation:
Babies are Social
- From the time babies are born they are
immersed in a rich social life and make
connections within this world very early on. - Social behaviours engage in social
behaviours within the first few weeks of life
which supports that these connections are
being made. - For example, engage in eye contact,
smiling, and other communicative
behaviours like pointing, imitating, or
helping and sharing behaviours. - Aside from being social agents themselves
babies also have a lot of opportunities to
observe and learn from others social
interactions. - How do babies figure out what these
behaviours mean? Why and how they can
do that themselves? What do people do,
mean, and try to achieve when others use
language? These are all puzzles’ babies
need to figure out. - A big part of this puzzle involves
communication.
Developing an Understanding of Communication
*How do babies make sense of language,
what people mean and do when they do
not use it themselves? How do they begin
to learn the nuances of language? Where
can they start?
• Detecting communicative signals
o Signals from their environment are used
to signal and identify communicative
interactions.
o Are used even if infants do not know how
to decipher the meaning of words from
speech patterns.
• Used to understand the structure of
communicative interactions
• Use communication as a window into
other minds
o Understanding meaning and intentions of
others
Example 1: Preference for Speech
High Amplitude Sucking Study
• infants have a fully developed auditory
system in the third trimester of pregnancy; a
useful way to identify these signals is to
have a preference or attentional bias
towards sounds of speech.
• If spoken language is the primary currency
of communication, then paying attention to
the sound of speech even if you do not
understand its meaning can aid babies kick
start children’s understanding of language
as a communication tool.
Experiment 1:
Do babies show a preference for listening to speech?
(Vouloumanos & Werker, 2004)
Method:
High amplitude sucking method
- Babies as young as one day old.
- It utilitizes a frequent behaviour in babies
(sucking; innate skills because it helps them
drink milk and gain nutrients)
- Place a pacifier in newborn babies’ mouth,
they will reflexively begin to suck on it.
- The pacifier is hooked up to a pressure
transducer which tells researchers the rate
of sucking babies is doing.
- This method allows us to teach babies that
if they suck on the pacifier (pattern) they
can be reinforced with a stimulus they like.
In this example, it was an auditory stimulus.
- Sounds:
• Speech
• Non-speech sounds:
Like the speech audio but
distorted; synthesized speech or
sign wave speech; preserves the
acoustic properties, rhythm etc, of
speech but is synthesized i.e., not
produced by a human
- Do babies prefer human made speech?
- Steps:
• Train babies to change their rate of sucking
to be positively reinforced (80% or higher
rate; increase) with a sound.
• The audio changed every minute, with each
audio change we are seeing whether
infants will keep up the high sucking rate.
Whether babies were rewarded with
speech or sign waves speech sounds. This
is continued for 8 min or till the baby gets
bored.
• Babies learn that rewards are contingent on
their behaviour. Will they keep up the high
sucking rate to listen to the speech sounds
more than non-speech sounds?
Results:
• (1) = first 4 minuets of study and (2) = last 4
min
• Babies sucking rate in (1) is not different
between speech and non-speech sounds
because babies are still learning the
contingency.
• However, in the second block we see that
babies differentiate their sucking rate
between speech and non-speech sounds.
For example, babies continue the high
sucking rate for speech sounds from trial
one but reduce their sucking rates for non-
speech sounds.
Conclusion:
• This tells us that even 1–4-day old infants that have minimal exposure to speech outside the womb already prefer human generated speech over acoustically matched analogous non-speech sounds.
Summary:
Detecting Communicative Signals:
Preference for speech
• Is this an innate bias they have or is it
something they’ve learnt within the
womb?
• How do we determine which
explanation is correct?
• It’s difficult because they have some
exposure outside of the womb but since
their auditory system is fully developed
by the third trimester, they have prenatal
exposure as well.
*We can only conclude from this study that
there is an early bias (not innate bias).
Example 2:
(Vouloumanos & Werker, 2007)
Do newborns prefer speech to all other sounds? Or is it just synthetic sound they don’t like? What properties of speech sounds do babies like?
- One possibility is that they like
biologically produced sound (human
sounds and human speech).
Method:
- We compared two types of sound to see
what babies prefer to listen to (non-speech
sound biologically made or speech). For
example, the mating call of a monkey and
Japanese speech (novel speech sound;
control for familiarity effect).
Results:
- The results of this high amplitude sucking
experiment we see that babies show no
preference for listening to monkey
vocalizations or human speech. They
preferred human speech and monkey
vocalizations over sine-wave non-speech!
Conclusion:
- This tells us that babies have an early
attentional bias to biologically produced
speech. At least early on this early
communicative detection system is not
specific to human speech (general).
Example 3:
Research Question:
Is babies’ speech preference (communicative detecting system) always general to biological produced speech? Or does it specialize to human made speech later in development? Focus on human made speech that will be more beneficial to language acquisition.
Method:
• Adopt a different method now that we are
testing older children with more control over
their behaviour.
• At three-months have their detection
system for speech become more
specialized?
• We used a sequential preferential looking
procedure.
- Babies sit on their caregiver’s lap in front of
a screen. Can use preferential looking
because babies have more control over
their head, gaze and can sit upright with
support. We see whether babies can learn
that they can control what they hear with
their looking behaviour. Do they change
their behaviour to show you what they like
to hear?
- Babies sit on their caregiver’s lap Infront of a
checkerboard screen. Babies like high
contrast images like this. We can teach
children that every time they turn to look at
the checkerboard they get rewarded with a
sound (speech or non-speech sound).
Looking away causes the sound to stop and
looking back plays the next sound.
- This study allows for us to test the
developmental progression of speech
preference (as a communicative signal dev
ice) to see whether it becomes more
specialized over time.
• Results: - babies differentiate their looking time, they look longer at the speech sounds vs biologically non-speech sounds. - Unfamiliar human speech sound is preferred over monkey sounds. - Children look away faster when their behaviour is being rewarded by monkey sounds.
• Conclusion:
- There is a tuning process or perceptual
narrowing where their auditory preference
becomes more specialsed over time. They
are born with a generalized system which
gets fine-tuned over time with experience to
facilitate their learning.
- For example, 3 months of emersion in a
linguistically rich environment with little to
no monkey sounds. Salience of human
speech bias their attentional system to focus
on human made speech over non-speech.
Problem/Next Question: even at 3-months is this specialized to human speech or human-
non-speech sounds as well? Broader bias to human produced sound?
Example 4A:
(Vouloumanos et al., 2010)
• Research Question:
- Do babies prefer human speech or
non-human speech sounds at 3-
months?
- what were the results?
Example 4B:
(Vouloumanos et al., 2010)
• Research Question:
- Do infants show a preference for
human-speech over other
communicative sounds?
• Results:
- 3-month-old babies like speech over
sine-waves sounds, speech over
monkey sounds.
• Method:
- Using a preferential looking paradigm.
- Compare looking time between human
speech, human communicative sounds
(surprise, disgust and agreement-non-
speech like communicative but not
speech and human no-communicative
sounds, sneezes or yawns; with human
speech sounds).
• Results:
- 3-month-olds paid more attention to get
reinforced with human speech sounds
over human communicative non-
speech sounds or human-non-
communicative non-speech sounds.
• Conclusion:
- This doesn’t mean that speech is the
only important sound to babies. There
is merely a specific bias towards human
speech sounds over other human
produced sounds and other biologically
and synthetic sounds.
What does a bias for speech do for a baby?
*Paying attention to human speech- why is this important or helpful?
• Speech sounds can highlight how the world
is organized
- Applies to sign language as well. Any
speech channel highlights how the world is
organized.
- Speech provides labels to categories
information and objects in their physical
world; helps them make sense of their
world which is filled with so much stimuli in
their environment and can help them feel
like the world is organized.
- These perceptual systems help filter out
stimuli, what is important to focus our
attention on.
- Language and communication is an
organizing tool adults use to make sense of
their world. Once we understand language
we can identify objects, what objects go in
what categories, why things happen, social
learning.
- It’s a medium which allows use to learn
about language and the world around us
and how it works.
• How can speech sounds be useful for
babies when they do not know the meaning
of the sounds they hear?
- Difference between the form of speech and
its content. So, babies recognize its
structure, organization, and form without
understanding what the words mean!
Making use of communicative signals
(A) Speech for Categorization in Babies
Making use of communicative signals
Speech for Categorization in Babies
• What are these? All dogs. All fruit. Using
consistent speech sounds to refer to all
objects within the category. There is
variation within the category, but they use
the same label to unite them all these
exemplars together. The label “dog” gives
us information about what things fall within
the category even if I do not know what it
means. Labels help children learn the
boundaries of what entities fall within the
category. Saying no, that’s fruit for example
is a mechanism which allows babies to
figure out what objects are in a category
and what are not.
ExperimentA:
• Research Question:
- Does speech help babies form categories,
like adults do?
- Bias for speech in babies allows them to
categories objects in their world.
• Method:
- 3-month-old infants. Does using a
consistent speech sound (label) help them
notice what a category is?
- They use the same word “talma” and pair it
with different members of a category.
Repeated, look at the talma do you see the
talma? When showing them different
colored and shaped dinosaurs.
- In the testing phase, babies are shown a
novel dinosaur (not paired with word) and a
fish (novel; non-category member). Will
babies look longer at the new dinosaur
over the fish which doesn’t fit their
category.
• Results: - Babies differentiated between the two objects, they looked longer at the dinosaur than the fish. New category member of non-category member when presented side by side.
• Conclusion:
- This suggests that infants must notice
some characteristics or features of the
dinosaur to determine that it is a member
of the talma family (paired with that sound)
even if they word has no meaning (content)
to them.
- Words guide their attention and structure
their world even if the word has no
meaning to them.
ExperimentB:
(Ferry et al., 2010)
Now looking at the tone column: the comparison group
• Research Question
- Do babies form categories using only
consistent speech sounds or is it
generalizable to any consistent sound?
• Method:
- The same method was repeated with non-
speech sounds (i.e., specific sequence of
sounds, beep-beep-beeep-beep-beeep).
• Results:
- Babies in the tone experimental condition,
in the test phase showed no difference in
looking time between the dinosaur and fish
(results).
• Conclusion:
- This indicates that babies use consistent
human speech sounds to help notice
characteristics from prototypes to build
categories about objects in their world.
- Words are strongly linked to categorization
as early on 3-months, brings their attention
to similarities between items when paired
with a label to build a category in their
mind in ways that tones do not.
• Side Note:
- A third experimental condition in this study
looked at whether it is language or
consistent language that facilitates this
process?
Talma on every trial or different label for each example.
We see that variable words used (inconsistent) means that babies do not build categories about the objects and look equally between the dinosaur and fish.
This suggests that it is consistent label use that is important!
Experiment:
What does a bias for speech do for a baby?
• Speech sounds can highlight how the
world is organized
• Speech sounds can highlight how the world
is organized
o Link between speech they hear and
objects they see in their physical world to
organize what they see into categories.
o Is this an innate association, learnt or
developmentally progressive?
o We can test this with using monkey calls
(biologically made, unfamiliar non-speech
sounds vs human speech)
• Example:
o Replicating dinosaur study but this time we
pair each exemplar with a lemur call.
o We find that at 3-months babies
differentiate between the fish and the
dinosaur and look longer at the dinosaur
even if it was consistently paired with a
lemur cal.
o This tells us that there is system specificity
as they age: 3-month-olds can form
categories with consistent lemur calls or
human speech. However, 6-month-old
babies only from categories with
consistent human speech and NOT lemur
calls.
o = tuning process, where they have an
innate attentional system which is fine
tuned with their experiential learning over
time.
o At three months monkey calls still help
infants’ attention and categorization of
information even if they prefer to listen to
human speech BUT this effect disappears
by 6-months.
Experiment:
(Ferguson & Waxman, 2016)
Communicative Tones vs Noncommunicative Tones
*What is it about tones/sounds that helps
babies learn?
• Is it being human produced or something
else?
• One explanation is that it is not specific to
speech and can include communicative
actions or sounds (that they have more
experience with language being used as
communication mechanisms than other
variables).
• Experiment:
• Method:
• Gave non-speech sounds and given them
communicative meaning to see if babies
will use these sounds to organize their
world.
• Communicative tone:
- Two people communicating with one
another. One member of the pair talks with
normal speech, but the other’s voice is
replaced with beeps. However, they are
responded to as if the sounds have
meaning and they are having a
conversation.
- Tones are being used to respond to other
people’s speech in a social manner.
• Noncommunicative tones:
- Where the conversation voice over
(between the beeps and human speech) is
not meaningfully linked. There behaviour
doesn’t indicate that the sounds have
meaning in their social interaction.
- Tones are not paired with person, or social
interaction, not being used to respond to
someone in a meaningful social way!
• Similarities between conditions:
- Two people
- Social interaction (eye contact, mouth
moves, actions)
- Same auditory track (speech and non-
speech sounds
• Research Question:
- If we give tones a meaningful context in
the learning phase; will this later impact
their performance on the categorization
task? i.e., treat communicative tone the
same as human speech? (in consistent
communicative and not noncommunicative
tone).
- Were 6-month-old infants able to
differentiate their looking time between fish
and dinosaur when it was paired with a
consistent communicative tone or
uncommunicative tone)
• Results:
- Their results indicate that 6-month-olds
form categories from communicative tones.
- They looked longer at novel category
member (dinosaur) than to the fish.
- This suggests that it is not the fact that
speech is human produced that caused
this affect but simply that sounds are used
in a communicative way.
- No difference in looking time when non-
communicative tone is used (i.e., treat tone
different to speech like they normally do)
• Conclusion:
- Something about how tones are
embedded in a communicated situation
which leads them to treat it like speech
What does a bias for speech do for a baby?
• Bias is not specific to human speech.
• It is sounds that are used as communicated
signals, embedded in meaningful social
interactions, can be used to categorized
information in their environment and works.
• Could include, other sounds, gestures or
actions used by social agents in a
communicative social setting will have the
same effect.
(B) Speech Sounds may help babies orient
themselves too potential communicative
partners
What else do babies pay attention to other than speech? The person who is speaking and can help you learn.
Experiment 1:
Matching speech to its source
Matching speech to its source
(Vouloumanos et al., 2009)
• Research Question: - Do babies think about the source of speech? Can they match speech to its source? Who is producing these communicative sounds?
• Method:
- Using a similar sequential preferential
looking procedure called the sequential
audio-visual mapping procedure (sequential
because the images are not placed side by
side; they are done one after the other;
multiple trails)
- This procedure 5-month-old infants are sat
in their caregivers’ lap Infront of a screen.
We are testing whether infants look longer
at the screen which the audio matches the
visual stimuli.
- They see a human face, likely to produce
human speech or sound, or other beings.
Do they differentiate between human and
non-human animals about being social
agents responsible for speech? Using novel
Japanese speech or novel monkey sounds.
• Results:
- We see that 5-month-old babies look
significantly longer at the human face
(match) relative to the monkey (mismatch)
when human speech sound is played.
- 5-month-old’s look significantly longer a
monkey faces (match) than human faces
(mismatch) when the monkey sound is
played.
• Conclusion:
- We already know that by age 5-months
babies prefer human speech over monkey
speech. However, we now also know that
this is because babies pay attention to the
source of the sound/speech and prefer to
look at faces that match what they hear.
- They must know speech correspond with
humans, they are the typical sound of
speech in their environment and orients
them towards other social beings in their
environment that they can
communicate/interact with.
*5-month-olds seem to identify humans as
the source of speech. Together this shows
speech helps them both understand how the
world is organized and orient towards humans as communicative social partners. Understand where communicative interactions fit within the world, when they occur, how, why and what they mean.
Summary
Detecting communicative signals
• Infants show a bias for speech over other
sounds
- “tuned” from biological vocalizations to
speech by 3 months
• Infants make use of communicative signals
(like speech) to make sense of objects and
categories by 3 months
• Infants match speech to its source
(humans) by 5 months
Understanding the structure of communicative interactions
• How do babies come to understand what
communication is?
• What is the structure of communicative
interaction? How does it unfold?
Communicative Interactions:
• Common feature is that people talk and do
other communicative actions when there
are other people present and you are not
alone.
• Do babies understand how social
interaction unfold (despite being a part of it
or not). Are babies using communicative
signals decontextualized or do they use
them in reference to the social interaction
that they are embedded within?
Experiment:
(Legerstee et al. 2000)
• Research Question:
- Do infants understand that when someone
is producing a communicative action or
sound that there is typically another social
agent (recipient) present? Do babies pick
up on this important social aspect?
• Method:
o Habituation:
- 6-month-old babies were habituated to a
video of an actor who was talking
(speaking whilst looking behind a curtain)
or acting (physical manual action whilst
looking behind the curtain).
o Test phase:
- The idea was to see if babies understand
that speaking is a communicative action
that is more likely to be directed towards
another person and actions which are
more likely to be directed towards objects.
- Do they look longer at an outcome video
where the curtain (occlude) is moved to
reveal an object rather than a human when
habituated to talking?
• Results:
- Babies habituated to talking event, babies
looked longer at (dishabituated) when the
curtain was pulled back and an object was
revealed.
- This suggests that babies understand that
talking is more likely to be directed to
humans than inanimate objects. They were
surprised (VOE) when humans were talking
to objects and not another human.
- Babies habituated to the action event,
looked longer when the curtain was pulled
back and they saw a human back there
instead of an object. Dishabituation
indicates that 6-month-olds understand
were shocked that actions were directed at
a human rather than an object.
• Conclusion:
- This suggests that by 6-months babies
have a rudimentary understanding of the
social context that actions like speech are
occurring within. Furthermore, the purpose
that speech is used for in social
interactions. Even when it is not directed to
the infant themselves.
Communicative interactions
What other features do infants notice or understand about communicative interactions?
• It is common for people to make eye
contact with one another when they are
talking/interacting with one another.
• What other social engagement signs are
used? Paying attention to one another!
• Do infants understand that when you are
communicating with others you are
typically engaged with the person you are
communicating with?
Experiment:
(Beier & Spelke, 2011)
• Research Question:
- When infants begin to understand that
people who produce speech are mutually
engage with one another in any given
social interaction (mutual engagement).
• Method:
- 9-10-month-old infants were habituated to
videos of two actors who were talking to
one another with eye-contact.
• Test phase:
- Watched one of two outcome videos (in
both videos the people had switched
places with one another). In one test video,
the still made eye contact with one another
and were interacting with each other. In the
second video, they not looking at each
other and talking (conceptually different).
- Do babies understand that people who are
communicating with one another typically
face each other (mutually engagement).
• Results:
- 9-month-old infants did not dishabituate to
the video, looked equally between the two
outcome videos which indicates that they
were not able to identify that the videos
were different (speech with mutual gaze or
averted gaze).
- 10-month-old infants dishabituate and look
longer at the video where they have
averted gaze. This indicates that they are
able to understand that these videos are
conceptually different and that it is
unexpected for people to be talking but
not looking at one another.
• Conclusion:
- Somewhere between 9-10 months babies
are able to pick up on this particular aspect
of communicative interactions. That people
mutually engage with one another (eye
gaze) when speaking to one another. That
it is an important component of
communicative interaction.
