Lecture 5: Language Development Flashcards
SPEECH PERCEPTION
High Amplitude Sucking Procedure
- Used to test infants from birth to 4 months of age
- Relies on infants sucking reflex
- Infants hear a sound stimulus every time they produce a strong/ high-amplitude suck on a pacifier –> child quickly learns that strong suck produces noise.
- The number of strong sucks is an indicator of the infant’s interest
- More strong sucks = more interest
- 2 variations of procedure:
- Discrimination
- Preference
High Amplitude Sucking Procedure - Discrimination (similar to habituation)
Discrimination:
* Used to test whether infants can tell the difference between two auditory stimuli
* Variation of visual habituation paradigm
* Habituation phase: Each time infant produces a strong suck, a sound is played
- Continues until sucking has declined significantly (e.g. by 20%)
* Test phase: Hears new speech stimulus every time produces a strong suck
* If can distinguish between stimuli, sucking behaviour should increase –> expect dishabituation = increase in sucking
High Amplitude Sucking Procedure - Preference
Preference:
* Used to test infants’ preference for different stimuli
* 2 different stimuli are played on alternating minutes each time a strong suck is produced
- i.e., minute 1 = Stimulus A, minute 2 = Stimulus B, minute 3 = Stimulus A
* Number of strong sucks produced during presentation of each stimulus type is compared
* Preference = infants suck more during one stimulus minute type than the other –> this shows preference for this stimuli
Preferential Listening Procedure
- Speaker on either side of infant’s head (speakers blink at the start to orient children to understand where noise is coming from)
- When looks at speaker, a recording of speech plays
- Different speech from each speaker
- How long an infant spends looking in a particular direction/ listening to a particular sound indicates how much they like it
- Familiarity effect: Will listen longer to sounds they recognize
- Novelty effect: If first habituated to a sound, will listen longer to new sound
Note: only used with 4 months old and older - needs to be able to control neck movement
Speech Perception in Infancy
- Using high amplitude sucking paradigms, research has shown that newborns:
- Prefer to listen to speech sounds over artificial/electronic sounds
- Prefer mother’s voice over another woman’s voice
- Prefer to listen to native language vs. other language
- Suggests that language learning starts in utero - starts before birth!
Categorical Perception of Speech
- Adults perceive speech sounds as distinct categories even though the differences between speech sounds is gradual
- The way the 2 sounds are produced by our mouth is the same (super similar). They just have different voice onset times (VOT) (b/ vs p/).
- Anything that has a shorter VOT than 25 is perceived as b/ anything longer than 25 VOT is perceived as p/. There is a continuum but we dont hear that we hear 2 distinct categories.
Categorical Perception of Speech
- Categorical perception is useful because focuses listeners on sounds that are linguistically meaningful while ignoring meaningless differences
* E.g. difference between a 10ms VOT /b/ vs. 20ms VOT /b/ is meaningless in English
Infant Categorical Perception of Speech
- Do infants perceive the same speech categories as adults?
- Classic study by Eimaset al., 1971:
- Tested 1 month old infants learning English
- Used High amplitude sucking paradigm to test discrimination between /ba/ and /pa/
-
2 groups:
- Different speech sounds: Infants habituated to /ba/ (20 msVOT) and then tested with /pa/ (40 msVOT)
- Same speech sounds: Infants habituated to 60 msVOT /pa/ and then tested with 80 msVOT /pa
If infants can distinguish between ba and pa, how should they behave?
Increase sucking when sound from new category (‘pa’)
Infant Categorical Perception of Speech
- Newborns have same categorical perception of speech as adult
- They seem to have a similar category boundary as adults.
Results:
* Increased sucking when sound from new category.
* No change in sucking when sound from same category.
* Conclude that at one month old, we have the same categorical perception as adults (innate).
Infant Cross-Language Speech Perception
- Infants make more distinctions between speech sounds than adults
- Adults have difficulty perceiving differences between speech sounds that are not important in their native language
- E.g. In French, difference between /ou/ and /u/ is meaningful, but not in English
- What about infants?
Infant Cross-Language Speech Perception
- Classic study by Werker et al., 1988:
- Tested 6 month olds American infants learning English
- High amplitude sucking paradigm to see if they can discriminate between Hindi /Ta/ and /ta/
- English speaking adults struggle to distinguish between these 2 sounds
Infants Cross-Language Speech Perception (results)
Results:
* After habituating to one of these Hindi speech sounds, increased sucking
when heard other speech sounds
* i.e., if habituated to /Ta/, then increased sucking when tested with /ta/
Implications
- Infants discriminate between speech sounds they have never heard
before (i.e., speech sounds not found in their native language) - Infants are biologically ready to learn any of the world’s languages
Perceptual Narrowing of Speech Perception
- Infants’ ability to easily distinguish between non-native speech sounds diminishes around 8 months
- Not impossible to learn to distinguish betweeen these sounds but just harder.
- Perceptual narrowing is not a bad thing, its a good thing, makes the perception of these sounds more efficient. - By 10-12 months, infants’ perceptual abilities are narrowed to those sounds that are relevant to their native language
- Improves perception of speech sounds in native language
Word Segmentation
-
Word segmentation: knowing where words begin and end in fluent speech
- Begins around 7 months of age
- Infants’ statistical learning enables them to segment words in a stream of speech (enables them to pick out words from a stream of speech)
- Stress-patterning
- Distribution of speech sounds
Stress Patterning
- Different languages place stress on different parts of a word
- English: stress usually on first syllable
- French: stress usually on last syllable (babies growing up in french household will learn that the stress on last syllable indicates end of a word).
- Infants pick up on the stress patterning in their language and use it as clues to for word segmentation
Distribution of Speech Sounds
Babies learn that:
* Sounds that appear together often are likely to be words
* Sounds that don’t appear together often are more likely to be boundaries between words.
* Example: “happy baby”
- “ba” and “by” occur together often because make word “baby”
- “ha” and “ppy” occur together often because make word “happy”
- “ppy” and “ba” occur together less often because don’t make a word and many different words can come before “baby” (“happy”, “little” ) and many words can come after “happy” (“birthday”, “baby”, “puppy”)
Distribution of Speech Sounds (study)
- Study: Preferential listening procedure (turn their head towards sound they prefer)
-
Habituation: 8-month-olds listened to a stream of syllables for a long time (2 mins)
- Some syllables always occurred together
- Others rarely or never occurred together - Test: Presented with a syllable sequence that always co-occurred (“tokibu”) vs. syllable sequences that rarely co-occurred (“bagopi”)
- none of these were real words to ensure they had never heard it before.
-
Habituation: 8-month-olds listened to a stream of syllables for a long time (2 mins)
Distribution of Speech Sounds (Results)
- Results: Listened longer to rarely occurring sequence
- Shows that infants understood word boundaries by detecting the likelihood of syllables belonging together
Summary of Speech Perception
- Speech perception is studied with high amplitude sucking procedure and preferential listening paradigm
- Infants have remarkable speech perception abilities
- From birth, show adult-like categorical perception of speech for sounds that are physically similar and able to distinguish between speech sounds not found in their native language
- Speech perception is narrowed to sounds found in native language by 10-12 months of age
- Infants are sensitive to the patterns of language and use it to segment words from speech beginning around 7 months of age
THE JOURNEY TO FIRST WORDS
Certain important milestones!
* Important as indicator of developmental delays
* Can give information to parents.
Cooing
- Start around 2 months of age
- Drawn out vowel sounds, like “ooooohhh” and “aahhhh”
- Helps infants gain motor control over their vocalizations
- Elicits reactions from caregivers leading to back-and-forth cooing with caregivers
- Serves as social function - elicits reaction from caregivers, mimics what a conversation is like.
Babbling
- Start around 7 months of age (6-10 months of age)
- Repetitive consonant-vowel syllables, like “papapa” and “babababa”
- Speech sounds not necessarily from native language
- Infant babbling is very similar across languages
-
Manual babbling: Deaf infants that are exposed to sign language babble with repetitive hand movements made up of pieces of full signs
- Deaf children do not verbally babble
- Evidence that language exposure is critical for babbling
babling includes consonants and vowels, cooing is just vowels
Functions of babbling
-
Social function: Practicing turn-taking in a dialogue
- Infant babbling elicits caregiver reactions which in turn elicit more babbling -
Learning function: Signal that the infant is alert and ready to learn
- Infants learn more when an adult labels a new object just after they babble vs. learning the word in the absence of babbling
- caregiver tells the new word after the baby has babbled = more likely for them to learn that word than if the caregiver tells them after they have not babbled.
Understanding the words precedes production
- Infants appear to understand high-frequency words around 6 months of age
- they start to understand words before they learn to produce them.
- In lab studies, when 6-month-olds are presented with pictures of common items and hear one of the pictures being named, they look to the correct picture more often than chance- But cannot yet name these items themselves
- Shows that infants understand more words than they can produce
- Shows that infants understand more words than their caregivers realize
First words
- First words are produced around 12 months of age (10-15 months)
- First word: Any specific utterance/vocalization consistently used to refer to a particular meaning or thing.
- Can be tricky to identify:
- Babbling can sound like words (babbling can sound like a first word - they arent actually referring to their mom)
- E.g. “mamamama”
- Meaning of a first word can differ from its standard meaning
- E.g. “woof woof” referring to “dog”
- they are not using the word dog but they have come up with some other label to say dog.
Mispronunciations of First Words
- Often mispronounced in predictable ways:
- Omit difficult parts of words:
- “Banana” –> “nana” - Substitute difficult sounds for easier sounds:
- “Rabbit” –> “wabbit” - Re-order sounds to put easy sound first
- “Spaghetti” –> “pisketti”
Earliest words in 3 languages
Bold = 10 most common words in these 3 languages
First Words
- Usually refer to family members, pets, or important objects (things of interest to the baby)
- Meaning of first words are very similar across cultures
- Suggests that infants around the world have similar interests and priorities
Limitations of First Words
- Infants express themselves initially with only one-word utterances so cannot clearly communicate what they want to say
-
Overextension: using a word in a broader context than is appropriate
- E.g. “dog” refers to any 4 legged animal
- Does not mean that they don’t understand what the word refers to
-
Underextension: using a word in a more limited context than appropriate
- “cat” only refers to the family’s pet cat
Learning more words
-
18 months of age:
- Knows about 50 words
- Vocabulary spurt: Rate of word learning accelerates
WORD LEARNING
Shape bias
- Children will apply a noun to a new object of the same shape, even if that object is very different in size, colour, or texture
Grammatical Form
- Grammatical form of a word influences whether it’s interpreted as a noun, verb, or adjective
- toddlers are highly attuned to the gramatical form
Cross-Situational Word Learning
- Determining word meanings by tracking the correlations between labels and meanings across contexts
- Statistical learning plays a big role in learning language.
Pragmatic Cues – Tone of Voice
- If an adult uses a word that conflicts with child’s word for that object, they will learn the new word if it is said with confidence
IDS and Early Word Recognition
- Study:
- 7-8 month old infants were introduced to new words in IDS or regular adult speech
- Recognition of words tested 24 hours later using preferential listening procedure (familiarity)
- Infants were better at recognizing (looked longer at) words introduced in IDS than adult speech
- Indicates that IDS facilitates word learning
Quantity of Speech and SES
They had data up until 36 months and then used statistical modelling to predict what a child ability would be as they entered kindergarden
* average low SES = 13 million wods
* Middle SES = 25 million words
* High SES = 45 million
Known as the 30 million word gap (about 1/3 of language exposure compared to the high SES group.
Learning grammar
-
Age 5: Mastered basics of grammar
- Allows children to express and understand more complex ideas - We know that children have learned the grammar of their language when they:
- Can apply a grammatical rule to a new word/context
- E.g. Adding “s” to makes a word plural
- Clear sign that they internalized grammar, not copying someone
- Over-regularization errors
- Can apply a grammatical rule to a new word/context
Evidence: Recovery after Brain Damage
- Children that sustain brain damage to language areas (brocas and wenickes) usually recover full language capability
- Children’s brains are highly plastic; other parts of the developing brain can take over language functions - Teenagers and adults that sustain brain damage to language areas are more likely to suffer permanent language impairment
- More mature brain is much less plastic
- ie: when you have a stroke, brain is less plastic because you are older so can’t fix it
Evidence: Deaf Individuals
Researchers tested 2 groups of deaf adults:
1. No exposure to language during early childhood
2. Learned spoken language during early childhood
* Both groups began learning ASL in school between ages of 9-15
* Results: Those with exposure to language in infancy, even though spoken, performed better on language task than those with no language exposure
Evidence: Deaf Individuals
- Follow-up study tested deaf adults that had exposure to ASL in early childhood
- Performance of deaf adults with early exposure to ASL was the same as deaf adults with exposure to spoken language
- Shows that exposure to language, regardless of modality, in infancy is critical for full language development
- no early exposure = poor language ability
Evidence: Second Language Learners
- Performance on an English test by Chinese and Korean immigrants was related to the age at which they first arrived in the USA
- Shows that language proficiency is related to first age of exposure to that language
- Language performance is highly variable when a language is learned after puberty
Sensitive period applies to all languages (not just native language).
Bilingualism in Utero (results)
Results:
* English monolinguals had a preference for English (kids born to the monolingual moms have a preference for english)
* English-Tagalog bilinguals showed no consistent preference for either language
* Suggests that bilingual infants start learning about two native languages prebirth (treat both of the languages as native)
Possible implication: could be that they are confused and that is why they are not showing a preference.
Bilingualism in the Womb (results)
Results:
* Both bilingual babies (newborns) and monolingual babies differentiated between Tagalog and English
* Shows that bilingual infants can differentiate between native languages despite showing similar preference for both languages
* shows that the newborns can already tell the difference between two languages. The bilingual babies are treating both of these as native because they are showing an equal preference to them.
